public inbox for [email protected]
help / color / mirror / Atom feed[PATCH] Support pg_walfile_name on standby
127+ messages / 23 participants
[nested] [flat]
* [PATCH] Support pg_walfile_name on standby
@ 2019-07-23 15:28 Jehan-Guillaume de Rorthais <[email protected]>
0 siblings, 0 replies; 127+ messages in thread
From: Jehan-Guillaume de Rorthais @ 2019-07-23 15:28 UTC (permalink / raw)
Support executing both SQL functions pg_walfile_name() and
pg_walfile_name_offset() on a standby.
---
src/backend/access/transam/xlogfuncs.c | 39 +++++++++++++++++---------
1 file changed, 25 insertions(+), 14 deletions(-)
diff --git a/src/backend/access/transam/xlogfuncs.c b/src/backend/access/transam/xlogfuncs.c
index b35043bf71..86c4d8382b 100644
--- a/src/backend/access/transam/xlogfuncs.c
+++ b/src/backend/access/transam/xlogfuncs.c
@@ -460,13 +460,7 @@ pg_walfile_name_offset(PG_FUNCTION_ARGS)
TupleDesc resultTupleDesc;
HeapTuple resultHeapTuple;
Datum result;
-
- if (RecoveryInProgress())
- ereport(ERROR,
- (errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
- errmsg("recovery is in progress"),
- errhint("%s cannot be executed during recovery.",
- "pg_walfile_name_offset()")));
+ TimeLineID tl;
/*
* Construct a tuple descriptor for the result row. This must match this
@@ -480,11 +474,24 @@ pg_walfile_name_offset(PG_FUNCTION_ARGS)
resultTupleDesc = BlessTupleDesc(resultTupleDesc);
+ if (RecoveryInProgress())
+ {
+ GetWalRcvWriteRecPtr(NULL, &tl);
+
+ if (!tl)
+ {
+ isnull[0] = isnull[1] = true;
+ goto result;
+ }
+ }
+ else
+ tl = ThisTimeLineID;
+
/*
* xlogfilename
*/
XLByteToPrevSeg(locationpoint, xlogsegno, wal_segment_size);
- XLogFileName(xlogfilename, ThisTimeLineID, xlogsegno, wal_segment_size);
+ XLogFileName(xlogfilename, tl, xlogsegno, wal_segment_size);
values[0] = CStringGetTextDatum(xlogfilename);
isnull[0] = false;
@@ -497,6 +504,7 @@ pg_walfile_name_offset(PG_FUNCTION_ARGS)
values[1] = UInt32GetDatum(xrecoff);
isnull[1] = false;
+result:
/*
* Tuple jam: Having first prepared your Datums, then squash together
*/
@@ -517,16 +525,19 @@ pg_walfile_name(PG_FUNCTION_ARGS)
XLogSegNo xlogsegno;
XLogRecPtr locationpoint = PG_GETARG_LSN(0);
char xlogfilename[MAXFNAMELEN];
+ TimeLineID tl;
if (RecoveryInProgress())
- ereport(ERROR,
- (errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
- errmsg("recovery is in progress"),
- errhint("%s cannot be executed during recovery.",
- "pg_walfile_name()")));
+ {
+ GetWalRcvWriteRecPtr(NULL, &tl);
+ if (!tl)
+ PG_RETURN_NULL();
+ }
+ else
+ tl = ThisTimeLineID;
XLByteToPrevSeg(locationpoint, xlogsegno, wal_segment_size);
- XLogFileName(xlogfilename, ThisTimeLineID, xlogsegno, wal_segment_size);
+ XLogFileName(xlogfilename, tl, xlogsegno, wal_segment_size);
PG_RETURN_TEXT_P(cstring_to_text(xlogfilename));
}
--
2.20.1
--MP_/ioOWfsqs/5fRg3CRPtZ8uQl--
^ permalink raw reply [nested|flat] 127+ messages in thread
* [PATCH] Support pg_walfile_name on standby
@ 2019-07-23 15:28 Jehan-Guillaume de Rorthais <[email protected]>
0 siblings, 0 replies; 127+ messages in thread
From: Jehan-Guillaume de Rorthais @ 2019-07-23 15:28 UTC (permalink / raw)
Support executing both SQL functions pg_walfile_name() and
pg_walfile_name_offset() on a standby.
---
src/backend/access/transam/xlogfuncs.c | 22 ++++++++++------------
1 file changed, 10 insertions(+), 12 deletions(-)
diff --git a/src/backend/access/transam/xlogfuncs.c b/src/backend/access/transam/xlogfuncs.c
index b35043bf71..a8184a20c4 100644
--- a/src/backend/access/transam/xlogfuncs.c
+++ b/src/backend/access/transam/xlogfuncs.c
@@ -460,13 +460,12 @@ pg_walfile_name_offset(PG_FUNCTION_ARGS)
TupleDesc resultTupleDesc;
HeapTuple resultHeapTuple;
Datum result;
+ TimeLineID tl;
if (RecoveryInProgress())
- ereport(ERROR,
- (errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
- errmsg("recovery is in progress"),
- errhint("%s cannot be executed during recovery.",
- "pg_walfile_name_offset()")));
+ GetWalRcvWriteRecPtr(NULL, &tl);
+ else
+ tl = ThisTimeLineID;
/*
* Construct a tuple descriptor for the result row. This must match this
@@ -484,7 +483,7 @@ pg_walfile_name_offset(PG_FUNCTION_ARGS)
* xlogfilename
*/
XLByteToPrevSeg(locationpoint, xlogsegno, wal_segment_size);
- XLogFileName(xlogfilename, ThisTimeLineID, xlogsegno, wal_segment_size);
+ XLogFileName(xlogfilename, tl, xlogsegno, wal_segment_size);
values[0] = CStringGetTextDatum(xlogfilename);
isnull[0] = false;
@@ -517,16 +516,15 @@ pg_walfile_name(PG_FUNCTION_ARGS)
XLogSegNo xlogsegno;
XLogRecPtr locationpoint = PG_GETARG_LSN(0);
char xlogfilename[MAXFNAMELEN];
+ TimeLineID tl;
if (RecoveryInProgress())
- ereport(ERROR,
- (errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
- errmsg("recovery is in progress"),
- errhint("%s cannot be executed during recovery.",
- "pg_walfile_name()")));
+ GetWalRcvWriteRecPtr(NULL, &tl);
+ else
+ tl = ThisTimeLineID;
XLByteToPrevSeg(locationpoint, xlogsegno, wal_segment_size);
- XLogFileName(xlogfilename, ThisTimeLineID, xlogsegno, wal_segment_size);
+ XLogFileName(xlogfilename, tl, xlogsegno, wal_segment_size);
PG_RETURN_TEXT_P(cstring_to_text(xlogfilename));
}
--
2.20.1
--MP_/btmtf9HMzmlizXPI4PTvbk4--
^ permalink raw reply [nested|flat] 127+ messages in thread
* logical decoding and replication of sequences
@ 2021-06-07 22:28 Tomas Vondra <[email protected]>
0 siblings, 3 replies; 127+ messages in thread
From: Tomas Vondra @ 2021-06-07 22:28 UTC (permalink / raw)
To: PostgreSQL Hackers <[email protected]>; +Cc: Petr Jelinek <[email protected]>
Hi,
One of the existing limitations of logical decoding / replication is
that it does no care about sequences. The annoying consequence is that
after a failover to logical replica, all the table data may be
replicated but the sequences are still at the initial values, requiring
some custom solution that moves the sequences forward enough to prevent
duplicities.
There have been attempts to address this in the past, most recently [1],
but none of them got in due to various issues.
This is an attempt, based on [1] (but with many significant parts added
or reworked), aiming to deal with this. The primary purpose of sharing
it is getting feedback and opinions on the design decisions. It's still
a WIP - it works fine AFAICS, but some of the bits may be a bit hackish.
The overall goal is to have the same sequence data on the primary and
logical replica, or something sufficiently close to that, so that the
replica after a failover does not generate duplicate values.
This patch does a couple basic things:
1) extends the logical decoding to handle sequences. It adds a new
callback, similarly to what we have for messages. There's a bit of
complexity with transactional and non-transactional behavior, more
about that later
2) extends test_decoding to support this new callback, printing the
sequence increments (the decoded WAL records)
3) extends built-in replication to support sequences, so publications
may contain both tables and sequences, etc., sequences data sync
when creating subscriptions, etc.
transactional vs. non-transactional
-----------------------------------
The first part (extending logical decoding) is simple in principle. We
simply decode the sequence updates, but then comes a challenge - should
we just treat it transactionally and stash it in reorder buffer, or
just pass it to the output plugin right-away?
For messages, this can be specified as a flag when adding the message,
so the user can decide depending on the message purpose. For sequences,
all we do is nextval() and it depends on the context in which it's used,
we can't just pick one of those approaches.
Consider this, for example:
CREATE SEQUENCE s;
BEGIN;
SELECT nextval('s') FROM generate_series(1,1000) s(i);
ROLLBACK;
If we handle this "transactionally", we'd stash the "nextval" increment
into the transaction, and then discard it due to the rollback, so the
output plugin (and replica) would never get it. So this is an argument
for non-transactional behavior.
On the other hand, consider this:
CREATE SEQUENCE s;
BEGIN;
ALTER SEQUENCE s RESTART WITH 2000;
SELECT nextval('s') FROM generate_series(1,1000) s(i);
ROLLBACK;
In this case the ALTER creates a new relfilenode, and the ROLLBACK does
discard it including the effects of the nextval calls. So here we should
treat it transactionally, stash the increment(s) in the transaction and
just discard it all on rollback.
A somewhat similar example is this
BEGIN;
CREATE SEQUENCE s;
SELECT nextval('s') FROM generate_series(1,1000) s(i);
COMMIT;
Again - the decoded nextval needs to be handled transactionally, because
otherwise it's going to be very difficult for custom plugins to combine
this with DDL replication.
So the patch does a fairly simple thing:
1) By default, sequences are treated non-transactionally, i.e. sent to
the output plugin right away.
2) We track sequences created in running (sub)transactions, and those
are handled transactionally. This includes ALTER SEQUENCE cases,
which create a new relfilenode, which is used as an identifier.
It's a bit more complex, because of cases like this:
BEGIN;
CREATE SEQUENCE s;
SAVEPOINT a;
SELECT nextval('s') FROM generate_series(1,1000) s(i);
ROLLBACK TO a;
COMMIT;
because we must not discard the nextval changes - this is handled by
always stashing the nextval changes to the subxact where the sequence
relfilenode was created.
The tracking is a bit cumbersome - there's a hash table with relfilenode
mapped to XID in which it was created. AFAIK that works, but might be
an issue with many sequences created in running transactions. Not sure.
detecting sequence creation
---------------------------
Detection that a sequence (or rather the relfilenode) was created is
done by adding a "created" flag into the xl_seq_rec, and setting it to
"true" in the first WAL record after the creation. There might be some
other way, but this seemed simple enough.
applying the sequence (ResetSequence2)
--------------------------------------
The decoding pretty much just extracts log_value, log_cnt and is_called
from the sequence, and passes them to the output plugin. On the replica
we extract those from the message, and write them to the local sequence
using a new ResetSequence2 function.
It's possible we don't really need log_cnt and is_called. After all,
log_cnt is zero most of the time anyway, and the worst thing that could
happen if we ignore it is we skip a couple values (which seems fine).
syncing sequences in a subscription
-----------------------------------
After creating a subscription, the sequences get syncronized just like
tables. This part ia a bit hacked together, and there's definitely room
for improvement - e.g. a new bgworker is started for each sequence, as
we simply treat both tabels and sequences as "relation". But all we need
to do for sequences is copying the (last_value, log_cnt, is_called) and
calling ResetSequence2, so maybe we could sync all sequences in a single
worker, or something like that.
new "sequence" publication action
---------------------------------
The publications now have a new "sequence" publication action, which is
enabled by default. This determines whether the publication decodes
sequences or what.
FOR ALL SEQUENCES
-----------------
It should be possible to create FOR ALL SEQUENCES publications, just
like we have FOR ALL TABLES. But this produces shift/reduce conflicts
in the grammar, and I didn't bother dealing with that. So for now it's
required to do ALTER PUBLICATION ... [ADD | DROP] SEQUENCE ...
no streaming support yet
------------------------
There's no supoprt for streaming of in-progress transactions yet, but
should be trivial to add.
GetCurrentTransactionId() in nextval
------------------------------------
There's a bit annoying behavior of nextval() - if you do this:
BEGIN;
CREATE SEQUENCE s;
SAVEPOINT a;
SELECT nextval('s') FROM generate_series(1,100) s(i);
COMMIT;
then the WAL record for nextval (right after the savepoint) will have
XID 0 (easy to see in pg_waldump). That's kinda strange, and it causes
problems in DecodeSequence() when calling
SnapBuildProcessChange(builder, xid, buf->origptr)
for transactional changes, because that expects a valid XID. Fixing
this required adding GetCurrentTransactionId() to nextval() and two
other functions, which were only doing
if (RelationNeedsWAL(seqrel))
GetTopTransactionId();
so far. I'm not sure if this has some particularly bad consequences.
regards
[1]
https://www.postgresql.org/message-id/flat/1710ed7e13b.cd7177461430746.3372264562543607781%40highgo....
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
Attachments:
[text/x-patch] sequence-decoding-20210608.patch (113.9K, ../../[email protected]/2-sequence-decoding-20210608.patch)
download | inline diff:
diff --git a/contrib/test_decoding/Makefile b/contrib/test_decoding/Makefile
index 9a31e0b879..56ddc3abae 100644
--- a/contrib/test_decoding/Makefile
+++ b/contrib/test_decoding/Makefile
@@ -5,7 +5,8 @@ PGFILEDESC = "test_decoding - example of a logical decoding output plugin"
REGRESS = ddl xact rewrite toast permissions decoding_in_xact \
decoding_into_rel binary prepared replorigin time messages \
- spill slot truncate stream stats twophase twophase_stream
+ spill slot truncate stream stats twophase twophase_stream \
+ sequence
ISOLATION = mxact delayed_startup ondisk_startup concurrent_ddl_dml \
oldest_xmin snapshot_transfer subxact_without_top concurrent_stream \
twophase_snapshot
diff --git a/contrib/test_decoding/expected/sequence.out b/contrib/test_decoding/expected/sequence.out
new file mode 100644
index 0000000000..07f3e3e92c
--- /dev/null
+++ b/contrib/test_decoding/expected/sequence.out
@@ -0,0 +1,327 @@
+-- predictability
+SET synchronous_commit = on;
+SELECT 'init' FROM pg_create_logical_replication_slot('regression_slot', 'test_decoding');
+ ?column?
+----------
+ init
+(1 row)
+
+CREATE SEQUENCE test_sequence;
+-- test the sequence changes by several nextval() calls
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 1
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 2
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 4
+(1 row)
+
+-- test the sequence changes by several ALTER commands
+ALTER SEQUENCE test_sequence INCREMENT BY 10;
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 14
+(1 row)
+
+ALTER SEQUENCE test_sequence START WITH 3000;
+ALTER SEQUENCE test_sequence MAXVALUE 10000;
+ALTER SEQUENCE test_sequence RESTART WITH 4000;
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 4000
+(1 row)
+
+-- test the sequence changes by several setval() calls
+SELECT setval('test_sequence', 3500);
+ setval
+--------
+ 3500
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3510
+(1 row)
+
+SELECT setval('test_sequence', 3500, true);
+ setval
+--------
+ 3500
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3510
+(1 row)
+
+SELECT setval('test_sequence', 3500, false);
+ setval
+--------
+ 3500
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3500
+(1 row)
+
+-- show results and drop sequence
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+ data
+------------------------------------------------------------------------------------------------------
+ BEGIN
+ sequence: public.test_sequence transactional: 1 created: 1 last_value: 1, log_cnt: 0 is_called: 0
+ COMMIT
+ sequence: public.test_sequence transactional: 0 created: 0 last_value: 33, log_cnt: 0 is_called: 1
+ BEGIN
+ sequence: public.test_sequence transactional: 1 created: 1 last_value: 4, log_cnt: 0 is_called: 1
+ COMMIT
+ sequence: public.test_sequence transactional: 0 created: 0 last_value: 334, log_cnt: 0 is_called: 1
+ BEGIN
+ sequence: public.test_sequence transactional: 1 created: 1 last_value: 14, log_cnt: 32 is_called: 1
+ COMMIT
+ BEGIN
+ sequence: public.test_sequence transactional: 1 created: 1 last_value: 14, log_cnt: 0 is_called: 1
+ COMMIT
+ BEGIN
+ sequence: public.test_sequence transactional: 1 created: 1 last_value: 4000, log_cnt: 0 is_called: 0
+ COMMIT
+ sequence: public.test_sequence transactional: 0 created: 0 last_value: 4320, log_cnt: 0 is_called: 1
+ sequence: public.test_sequence transactional: 0 created: 0 last_value: 3500, log_cnt: 0 is_called: 1
+ sequence: public.test_sequence transactional: 0 created: 0 last_value: 3830, log_cnt: 0 is_called: 1
+ sequence: public.test_sequence transactional: 0 created: 0 last_value: 3500, log_cnt: 0 is_called: 1
+ sequence: public.test_sequence transactional: 0 created: 0 last_value: 3830, log_cnt: 0 is_called: 1
+ sequence: public.test_sequence transactional: 0 created: 0 last_value: 3500, log_cnt: 0 is_called: 0
+ sequence: public.test_sequence transactional: 0 created: 0 last_value: 3820, log_cnt: 0 is_called: 1
+(24 rows)
+
+DROP SEQUENCE test_sequence;
+-- rollback on sequence creation and update
+BEGIN;
+CREATE SEQUENCE test_sequence;
+CREATE TABLE test_table (a INT);
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 1
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 2
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3
+(1 row)
+
+SELECT setval('test_sequence', 3000);
+ setval
+--------
+ 3000
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3001
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3002
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3003
+(1 row)
+
+ALTER SEQUENCE test_sequence RESTART WITH 6000;
+INSERT INTO test_table VALUES( (SELECT nextval('test_sequence')) );
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 6001
+(1 row)
+
+ROLLBACK;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'skip-sequences', '0');
+ data
+------
+(0 rows)
+
+-- rollback on table creation with serial column
+BEGIN;
+CREATE TABLE test_table (a SERIAL, b INT);
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+INSERT INTO test_table (b) VALUES (300);
+SELECT setval('test_table_a_seq', 3000);
+ setval
+--------
+ 3000
+(1 row)
+
+INSERT INTO test_table (b) VALUES (400);
+INSERT INTO test_table (b) VALUES (500);
+INSERT INTO test_table (b) VALUES (600);
+ALTER SEQUENCE test_table_a_seq RESTART WITH 6000;
+INSERT INTO test_table (b) VALUES (700);
+INSERT INTO test_table (b) VALUES (800);
+INSERT INTO test_table (b) VALUES (900);
+ROLLBACK;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'skip-sequences', '0');
+ data
+------
+(0 rows)
+
+-- rollback on table with serial column
+CREATE TABLE test_table (a SERIAL, b INT);
+BEGIN;
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+INSERT INTO test_table (b) VALUES (300);
+SELECT setval('test_table_a_seq', 3000);
+ setval
+--------
+ 3000
+(1 row)
+
+INSERT INTO test_table (b) VALUES (400);
+INSERT INTO test_table (b) VALUES (500);
+INSERT INTO test_table (b) VALUES (600);
+ALTER SEQUENCE test_table_a_seq RESTART WITH 6000;
+INSERT INTO test_table (b) VALUES (700);
+INSERT INTO test_table (b) VALUES (800);
+INSERT INTO test_table (b) VALUES (900);
+ROLLBACK;
+-- check table and sequence values after rollback
+SELECT * from test_table_a_seq;
+ last_value | log_cnt | is_called
+------------+---------+-----------
+ 3003 | 30 | t
+(1 row)
+
+SELECT nextval('test_table_a_seq');
+ nextval
+---------
+ 3004
+(1 row)
+
+DROP TABLE test_table;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+ data
+---------------------------------------------------------------------------------------------------------
+ BEGIN
+ sequence: public.test_table_a_seq transactional: 1 created: 1 last_value: 1, log_cnt: 0 is_called: 0
+ COMMIT
+ sequence: public.test_table_a_seq transactional: 0 created: 0 last_value: 33, log_cnt: 0 is_called: 1
+ sequence: public.test_table_a_seq transactional: 0 created: 0 last_value: 3000, log_cnt: 0 is_called: 1
+ sequence: public.test_table_a_seq transactional: 0 created: 0 last_value: 3033, log_cnt: 0 is_called: 1
+ BEGIN
+ COMMIT
+(8 rows)
+
+-- savepoint test on table with serial column
+BEGIN;
+CREATE TABLE test_table (a SERIAL, b INT);
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+SAVEPOINT a;
+INSERT INTO test_table (b) VALUES (300);
+ROLLBACK TO SAVEPOINT a;
+DROP TABLE test_table;
+COMMIT;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+ data
+-------------------------------------------------------------------------------------------------------
+ BEGIN
+ sequence: public.test_table_a_seq transactional: 1 created: 1 last_value: 1, log_cnt: 0 is_called: 0
+ sequence: public.test_table_a_seq transactional: 1 created: 0 last_value: 33, log_cnt: 0 is_called: 1
+ table public.test_table: INSERT: a[integer]:1 b[integer]:100
+ table public.test_table: INSERT: a[integer]:2 b[integer]:200
+ COMMIT
+(6 rows)
+
+-- savepoint test on table with serial column
+BEGIN;
+CREATE SEQUENCE test_sequence;
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 1
+(1 row)
+
+SELECT setval('test_sequence', 3000);
+ setval
+--------
+ 3000
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3001
+(1 row)
+
+SAVEPOINT a;
+ALTER SEQUENCE test_sequence START WITH 7000;
+SELECT setval('test_sequence', 5000);
+ setval
+--------
+ 5000
+(1 row)
+
+ROLLBACK TO SAVEPOINT a;
+SELECT * FROM test_sequence;
+ last_value | log_cnt | is_called
+------------+---------+-----------
+ 3001 | 32 | t
+(1 row)
+
+DROP SEQUENCE test_sequence;
+COMMIT;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+ data
+------------------------------------------------------------------------------------------------------
+ BEGIN
+ sequence: public.test_sequence transactional: 1 created: 1 last_value: 1, log_cnt: 0 is_called: 0
+ sequence: public.test_sequence transactional: 1 created: 0 last_value: 33, log_cnt: 0 is_called: 1
+ sequence: public.test_sequence transactional: 1 created: 0 last_value: 3000, log_cnt: 0 is_called: 1
+ sequence: public.test_sequence transactional: 1 created: 0 last_value: 3033, log_cnt: 0 is_called: 1
+ COMMIT
+(6 rows)
+
+SELECT pg_drop_replication_slot('regression_slot');
+ pg_drop_replication_slot
+--------------------------
+
+(1 row)
+
diff --git a/contrib/test_decoding/sql/sequence.sql b/contrib/test_decoding/sql/sequence.sql
new file mode 100644
index 0000000000..d8a34738f3
--- /dev/null
+++ b/contrib/test_decoding/sql/sequence.sql
@@ -0,0 +1,119 @@
+-- predictability
+SET synchronous_commit = on;
+SELECT 'init' FROM pg_create_logical_replication_slot('regression_slot', 'test_decoding');
+
+CREATE SEQUENCE test_sequence;
+
+-- test the sequence changes by several nextval() calls
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+
+-- test the sequence changes by several ALTER commands
+ALTER SEQUENCE test_sequence INCREMENT BY 10;
+SELECT nextval('test_sequence');
+
+ALTER SEQUENCE test_sequence START WITH 3000;
+ALTER SEQUENCE test_sequence MAXVALUE 10000;
+ALTER SEQUENCE test_sequence RESTART WITH 4000;
+SELECT nextval('test_sequence');
+
+-- test the sequence changes by several setval() calls
+SELECT setval('test_sequence', 3500);
+SELECT nextval('test_sequence');
+SELECT setval('test_sequence', 3500, true);
+SELECT nextval('test_sequence');
+SELECT setval('test_sequence', 3500, false);
+SELECT nextval('test_sequence');
+
+-- show results and drop sequence
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+DROP SEQUENCE test_sequence;
+
+-- rollback on sequence creation and update
+BEGIN;
+CREATE SEQUENCE test_sequence;
+CREATE TABLE test_table (a INT);
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT setval('test_sequence', 3000);
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+ALTER SEQUENCE test_sequence RESTART WITH 6000;
+INSERT INTO test_table VALUES( (SELECT nextval('test_sequence')) );
+SELECT nextval('test_sequence');
+ROLLBACK;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'skip-sequences', '0');
+
+-- rollback on table creation with serial column
+BEGIN;
+CREATE TABLE test_table (a SERIAL, b INT);
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+INSERT INTO test_table (b) VALUES (300);
+SELECT setval('test_table_a_seq', 3000);
+INSERT INTO test_table (b) VALUES (400);
+INSERT INTO test_table (b) VALUES (500);
+INSERT INTO test_table (b) VALUES (600);
+ALTER SEQUENCE test_table_a_seq RESTART WITH 6000;
+INSERT INTO test_table (b) VALUES (700);
+INSERT INTO test_table (b) VALUES (800);
+INSERT INTO test_table (b) VALUES (900);
+ROLLBACK;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'skip-sequences', '0');
+
+-- rollback on table with serial column
+CREATE TABLE test_table (a SERIAL, b INT);
+
+BEGIN;
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+INSERT INTO test_table (b) VALUES (300);
+SELECT setval('test_table_a_seq', 3000);
+INSERT INTO test_table (b) VALUES (400);
+INSERT INTO test_table (b) VALUES (500);
+INSERT INTO test_table (b) VALUES (600);
+ALTER SEQUENCE test_table_a_seq RESTART WITH 6000;
+INSERT INTO test_table (b) VALUES (700);
+INSERT INTO test_table (b) VALUES (800);
+INSERT INTO test_table (b) VALUES (900);
+ROLLBACK;
+
+-- check table and sequence values after rollback
+SELECT * from test_table_a_seq;
+SELECT nextval('test_table_a_seq');
+
+DROP TABLE test_table;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+
+-- savepoint test on table with serial column
+BEGIN;
+CREATE TABLE test_table (a SERIAL, b INT);
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+SAVEPOINT a;
+INSERT INTO test_table (b) VALUES (300);
+ROLLBACK TO SAVEPOINT a;
+DROP TABLE test_table;
+COMMIT;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+
+-- savepoint test on table with serial column
+BEGIN;
+CREATE SEQUENCE test_sequence;
+SELECT nextval('test_sequence');
+SELECT setval('test_sequence', 3000);
+SELECT nextval('test_sequence');
+SAVEPOINT a;
+ALTER SEQUENCE test_sequence START WITH 7000;
+SELECT setval('test_sequence', 5000);
+ROLLBACK TO SAVEPOINT a;
+SELECT * FROM test_sequence;
+DROP SEQUENCE test_sequence;
+COMMIT;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+
+SELECT pg_drop_replication_slot('regression_slot');
diff --git a/contrib/test_decoding/test_decoding.c b/contrib/test_decoding/test_decoding.c
index de1b692658..162e3c18f0 100644
--- a/contrib/test_decoding/test_decoding.c
+++ b/contrib/test_decoding/test_decoding.c
@@ -35,6 +35,7 @@ typedef struct
bool include_timestamp;
bool skip_empty_xacts;
bool only_local;
+ bool skip_sequences;
} TestDecodingData;
/*
@@ -76,6 +77,10 @@ static void pg_decode_message(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn, XLogRecPtr message_lsn,
bool transactional, const char *prefix,
Size sz, const char *message);
+static void pg_decode_sequence(LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn, XLogRecPtr sequence_lsn,
+ Relation rel, bool transactional, bool created,
+ int64 last_value, int64 log_cnt, bool is_called);
static bool pg_decode_filter_prepare(LogicalDecodingContext *ctx,
TransactionId xid,
const char *gid);
@@ -141,6 +146,7 @@ _PG_output_plugin_init(OutputPluginCallbacks *cb)
cb->filter_by_origin_cb = pg_decode_filter;
cb->shutdown_cb = pg_decode_shutdown;
cb->message_cb = pg_decode_message;
+ cb->sequence_cb = pg_decode_sequence;
cb->filter_prepare_cb = pg_decode_filter_prepare;
cb->begin_prepare_cb = pg_decode_begin_prepare_txn;
cb->prepare_cb = pg_decode_prepare_txn;
@@ -175,6 +181,9 @@ pg_decode_startup(LogicalDecodingContext *ctx, OutputPluginOptions *opt,
data->skip_empty_xacts = false;
data->only_local = false;
+ /* skip sequences by default for backwards compatibility */
+ data->skip_sequences = true;
+
ctx->output_plugin_private = data;
opt->output_type = OUTPUT_PLUGIN_TEXTUAL_OUTPUT;
@@ -265,6 +274,17 @@ pg_decode_startup(LogicalDecodingContext *ctx, OutputPluginOptions *opt,
errmsg("could not parse value \"%s\" for parameter \"%s\"",
strVal(elem->arg), elem->defname)));
}
+ else if (strcmp(elem->defname, "skip-sequences") == 0)
+ {
+
+ if (elem->arg == NULL)
+ continue; /* true by default */
+ else if (!parse_bool(strVal(elem->arg), &data->skip_sequences))
+ ereport(ERROR,
+ (errcode(ERRCODE_INVALID_PARAMETER_VALUE),
+ errmsg("could not parse value \"%s\" for parameter \"%s\"",
+ strVal(elem->arg), elem->defname)));
+ }
else
{
ereport(ERROR,
@@ -744,6 +764,26 @@ pg_decode_message(LogicalDecodingContext *ctx,
OutputPluginWrite(ctx, true);
}
+static void
+pg_decode_sequence(LogicalDecodingContext *ctx, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ bool transactional, bool created,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ TestDecodingData *data = ctx->output_plugin_private;
+ char *nspname = get_namespace_name(RelationGetNamespace(rel));
+
+ /* return if requested to skip_sequences */
+ if (data->skip_sequences)
+ return;
+
+ OutputPluginPrepareWrite(ctx, true);
+ appendStringInfo(ctx->out, "sequence: %s.%s transactional: %d created: %d last_value: %zu, log_cnt: %zu is_called: %d",
+ nspname, RelationGetRelationName(rel),
+ transactional, created, last_value, log_cnt, is_called);
+ OutputPluginWrite(ctx, true);
+}
+
static void
pg_decode_stream_start(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn)
diff --git a/doc/src/sgml/catalogs.sgml b/doc/src/sgml/catalogs.sgml
index 16493209c6..ff8a6f102c 100644
--- a/doc/src/sgml/catalogs.sgml
+++ b/doc/src/sgml/catalogs.sgml
@@ -9429,6 +9429,11 @@ SCRAM-SHA-256$<replaceable><iteration count></replaceable>:<replaceable>&l
<entry>prepared transactions</entry>
</row>
+ <row>
+ <entry><link linkend="view-pg-publication-sequences"><structname>pg_publication_sequences</structname></link></entry>
+ <entry>publications and their associated sequences</entry>
+ </row>
+
<row>
<entry><link linkend="view-pg-publication-tables"><structname>pg_publication_tables</structname></link></entry>
<entry>publications and their associated tables</entry>
@@ -11263,6 +11268,72 @@ SELECT * FROM pg_locks pl LEFT JOIN pg_prepared_xacts ppx
</sect1>
+ <sect1 id="view-pg-publication-sequences">
+ <title><structname>pg_publication_sequences</structname></title>
+
+ <indexterm zone="view-pg-publication-sequences">
+ <primary>pg_publication_sequences</primary>
+ </indexterm>
+
+ <para>
+ The view <structname>pg_publication_sequences</structname> provides
+ information about the mapping between publications and the sequences they
+ contain. Unlike the underlying catalog
+ <link linkend="catalog-pg-publication-rel"><structname>pg_publication_rel</structname></link>,
+ this view expands
+ publications defined as <literal>FOR ALL SEQUENCES</literal>, so for such
+ publications there will be a row for each eligible sequence.
+ </para>
+
+ <table>
+ <title><structname>pg_publication_sequences</structname> Columns</title>
+ <tgroup cols="1">
+ <thead>
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ Column Type
+ </para>
+ <para>
+ Description
+ </para></entry>
+ </row>
+ </thead>
+
+ <tbody>
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ <structfield>pubname</structfield> <type>name</type>
+ (references <link linkend="catalog-pg-publication"><structname>pg_publication</structname></link>.<structfield>pubname</structfield>)
+ </para>
+ <para>
+ Name of publication
+ </para></entry>
+ </row>
+
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ <structfield>schemaname</structfield> <type>name</type>
+ (references <link linkend="catalog-pg-namespace"><structname>pg_namespace</structname></link>.<structfield>nspname</structfield>)
+ </para>
+ <para>
+ Name of schema containing sequence
+ </para></entry>
+ </row>
+
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ <structfield>sequencename</structfield> <type>name</type>
+ (references <link linkend="catalog-pg-class"><structname>pg_class</structname></link>.<structfield>relname</structfield>)
+ </para>
+ <para>
+ Name of sequence
+ </para></entry>
+ </row>
+ </tbody>
+ </tgroup>
+ </table>
+ </sect1>
+
<sect1 id="view-pg-publication-tables">
<title><structname>pg_publication_tables</structname></title>
diff --git a/doc/src/sgml/ref/alter_publication.sgml b/doc/src/sgml/ref/alter_publication.sgml
index faa114b2c6..c68a1573de 100644
--- a/doc/src/sgml/ref/alter_publication.sgml
+++ b/doc/src/sgml/ref/alter_publication.sgml
@@ -24,6 +24,9 @@ PostgreSQL documentation
ALTER PUBLICATION <replaceable class="parameter">name</replaceable> ADD TABLE [ ONLY ] <replaceable class="parameter">table_name</replaceable> [ * ] [, ...]
ALTER PUBLICATION <replaceable class="parameter">name</replaceable> SET TABLE [ ONLY ] <replaceable class="parameter">table_name</replaceable> [ * ] [, ...]
ALTER PUBLICATION <replaceable class="parameter">name</replaceable> DROP TABLE [ ONLY ] <replaceable class="parameter">table_name</replaceable> [ * ] [, ...]
+ALTER PUBLICATION <replaceable class="parameter">name</replaceable> ADD SEQUENCE <replaceable class="parameter">sequence_name</replaceable> [ * ] [, ...]
+ALTER PUBLICATION <replaceable class="parameter">name</replaceable> SET SEQUENCE <replaceable class="parameter">sequence_name</replaceable> [ * ] [, ...]
+ALTER PUBLICATION <replaceable class="parameter">name</replaceable> DROP SEQUENCE <replaceable class="parameter">sequence_name</replaceable> [ * ] [, ...]
ALTER PUBLICATION <replaceable class="parameter">name</replaceable> SET ( <replaceable class="parameter">publication_parameter</replaceable> [= <replaceable class="parameter">value</replaceable>] [, ... ] )
ALTER PUBLICATION <replaceable class="parameter">name</replaceable> OWNER TO { <replaceable>new_owner</replaceable> | CURRENT_ROLE | CURRENT_USER | SESSION_USER }
ALTER PUBLICATION <replaceable class="parameter">name</replaceable> RENAME TO <replaceable>new_name</replaceable>
@@ -50,7 +53,18 @@ ALTER PUBLICATION <replaceable class="parameter">name</replaceable> RENAME TO <r
</para>
<para>
- The fourth variant of this command listed in the synopsis can change
+ The next three variants change which sequences are part of the publication.
+ The <literal>SET SEQUENCE</literal> clause will replace the list of sequences
+ in the publication with the specified one. The <literal>ADD SEQUENCE</literal>
+ and <literal>DROP SEQUENCE</literal> clauses will add and remove one or more
+ sequences from the publication. Note that adding sequences to a publication
+ that is already subscribed to will require a <literal>ALTER SUBSCRIPTION
+ ... REFRESH PUBLICATION</literal> action on the subscribing side in order
+ to become effective.
+ </para>
+
+ <para>
+ The seventh variant of this command listed in the synopsis can change
all of the publication properties specified in
<xref linkend="sql-createpublication"/>. Properties not mentioned in the
command retain their previous settings.
@@ -62,7 +76,8 @@ ALTER PUBLICATION <replaceable class="parameter">name</replaceable> RENAME TO <r
<para>
You must own the publication to use <command>ALTER PUBLICATION</command>.
- Adding a table to a publication additionally requires owning that table.
+ Adding a table to a publication additionally requires owning that table,
+ and the same requirement applies to sequences.
To alter the owner, you must also be a direct or indirect member of the new
owning role. The new owner must have <literal>CREATE</literal> privilege on
the database. Also, the new owner of a <literal>FOR ALL TABLES</literal>
@@ -97,6 +112,15 @@ ALTER PUBLICATION <replaceable class="parameter">name</replaceable> RENAME TO <r
</listitem>
</varlistentry>
+ <varlistentry>
+ <term><replaceable class="parameter">sequence_name</replaceable></term>
+ <listitem>
+ <para>
+ Name of an existing sequence.
+ </para>
+ </listitem>
+ </varlistentry>
+
<varlistentry>
<term><literal>SET ( <replaceable class="parameter">publication_parameter</replaceable> [= <replaceable class="parameter">value</replaceable>] [, ... ] )</literal></term>
<listitem>
diff --git a/doc/src/sgml/ref/alter_subscription.sgml b/doc/src/sgml/ref/alter_subscription.sgml
index 367ac814f4..3412ca68e9 100644
--- a/doc/src/sgml/ref/alter_subscription.sgml
+++ b/doc/src/sgml/ref/alter_subscription.sgml
@@ -138,8 +138,8 @@ ALTER SUBSCRIPTION <replaceable class="parameter">name</replaceable> RENAME TO <
<term><literal>REFRESH PUBLICATION</literal></term>
<listitem>
<para>
- Fetch missing table information from publisher. This will start
- replication of tables that were added to the subscribed-to publications
+ Fetch missing table and sequence information from publisher. This will start
+ replication of tables and sequences that were added to the subscribed-to publications
since the last invocation of <command>REFRESH PUBLICATION</command> or
since <command>CREATE SUBSCRIPTION</command>.
</para>
@@ -156,7 +156,7 @@ ALTER SUBSCRIPTION <replaceable class="parameter">name</replaceable> RENAME TO <
Specifies whether the existing data in the publications that are
being subscribed to should be copied once the replication starts.
The default is <literal>true</literal>. (Previously subscribed
- tables are not copied.)
+ tables and sequences are not copied.)
</para>
</listitem>
</varlistentry>
diff --git a/src/backend/catalog/pg_publication.c b/src/backend/catalog/pg_publication.c
index 86e415af89..427582f36a 100644
--- a/src/backend/catalog/pg_publication.c
+++ b/src/backend/catalog/pg_publication.c
@@ -51,26 +51,27 @@ check_publication_add_relation(Relation targetrel)
{
/* Must be a regular or partitioned table */
if (RelationGetForm(targetrel)->relkind != RELKIND_RELATION &&
- RelationGetForm(targetrel)->relkind != RELKIND_PARTITIONED_TABLE)
+ RelationGetForm(targetrel)->relkind != RELKIND_PARTITIONED_TABLE &&
+ RelationGetForm(targetrel)->relkind != RELKIND_SEQUENCE)
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
- errmsg("\"%s\" is not a table",
+ errmsg("\"%s\" is not a table or sequence",
RelationGetRelationName(targetrel)),
- errdetail("Only tables can be added to publications.")));
+ errdetail("Only tables and sequences can be added to publications.")));
- /* Can't be system table */
+ /* Can't be system table/sequence */
if (IsCatalogRelation(targetrel))
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
- errmsg("\"%s\" is a system table",
+ errmsg("\"%s\" is a system table or sequence",
RelationGetRelationName(targetrel)),
- errdetail("System tables cannot be added to publications.")));
+ errdetail("System tables / sequences cannot be added to publications.")));
/* UNLOGGED and TEMP relations cannot be part of publication. */
if (!RelationIsPermanent(targetrel))
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
- errmsg("table \"%s\" cannot be replicated",
+ errmsg("table or sequence \"%s\" cannot be replicated",
RelationGetRelationName(targetrel)),
errdetail("Temporary and unlogged relations cannot be replicated.")));
}
@@ -98,7 +99,8 @@ static bool
is_publishable_class(Oid relid, Form_pg_class reltuple)
{
return (reltuple->relkind == RELKIND_RELATION ||
- reltuple->relkind == RELKIND_PARTITIONED_TABLE) &&
+ reltuple->relkind == RELKIND_PARTITIONED_TABLE ||
+ reltuple->relkind == RELKIND_SEQUENCE) &&
!IsCatalogRelationOid(relid) &&
reltuple->relpersistence == RELPERSISTENCE_PERMANENT &&
relid >= FirstNormalObjectId;
@@ -271,6 +273,10 @@ GetPublicationRelations(Oid pubid, PublicationPartOpt pub_partopt)
pubrel = (Form_pg_publication_rel) GETSTRUCT(tup);
+ /* skip sequences here */
+ if (get_rel_relkind(pubrel->prrelid) == RELKIND_SEQUENCE)
+ continue;
+
if (get_rel_relkind(pubrel->prrelid) == RELKIND_PARTITIONED_TABLE &&
pub_partopt != PUBLICATION_PART_ROOT)
{
@@ -304,6 +310,49 @@ GetPublicationRelations(Oid pubid, PublicationPartOpt pub_partopt)
return result;
}
+/*
+ * Gets list of relation oids for a publication (sequences only).
+ *
+ * This should only be used for normal publications, the FOR ALL TABLES
+ * should use GetAllSequencesPublicationRelations().
+ */
+List *
+GetPublicationSequenceRelations(Oid pubid)
+{
+ List *result;
+ Relation pubrelsrel;
+ ScanKeyData scankey;
+ SysScanDesc scan;
+ HeapTuple tup;
+
+ /* Find all publications associated with the relation. */
+ pubrelsrel = table_open(PublicationRelRelationId, AccessShareLock);
+
+ ScanKeyInit(&scankey,
+ Anum_pg_publication_rel_prpubid,
+ BTEqualStrategyNumber, F_OIDEQ,
+ ObjectIdGetDatum(pubid));
+
+ scan = systable_beginscan(pubrelsrel, PublicationRelPrrelidPrpubidIndexId,
+ true, NULL, 1, &scankey);
+
+ result = NIL;
+ while (HeapTupleIsValid(tup = systable_getnext(scan)))
+ {
+ Form_pg_publication_rel pubrel;
+
+ pubrel = (Form_pg_publication_rel) GETSTRUCT(tup);
+
+ if (get_rel_relkind(pubrel->prrelid) == RELKIND_SEQUENCE)
+ result = lappend_oid(result, pubrel->prrelid);
+ }
+
+ systable_endscan(scan);
+ table_close(pubrelsrel, AccessShareLock);
+
+ return result;
+}
+
/*
* Gets list of publication oids for publications marked as FOR ALL TABLES.
*/
@@ -404,6 +453,46 @@ GetAllTablesPublicationRelations(bool pubviaroot)
return result;
}
+/*
+ * Gets list of all relation published by FOR ALL TABLES publication(s).
+ *
+ * If the publication publishes partition changes via their respective root
+ * partitioned tables, we must exclude partitions in favor of including the
+ * root partitioned tables.
+ */
+List *
+GetAllSequencesPublicationRelations(void)
+{
+ Relation classRel;
+ ScanKeyData key[1];
+ TableScanDesc scan;
+ HeapTuple tuple;
+ List *result = NIL;
+
+ classRel = table_open(RelationRelationId, AccessShareLock);
+
+ ScanKeyInit(&key[0],
+ Anum_pg_class_relkind,
+ BTEqualStrategyNumber, F_CHAREQ,
+ CharGetDatum(RELKIND_SEQUENCE));
+
+ scan = table_beginscan_catalog(classRel, 1, key);
+
+ while ((tuple = heap_getnext(scan, ForwardScanDirection)) != NULL)
+ {
+ Form_pg_class relForm = (Form_pg_class) GETSTRUCT(tuple);
+ Oid relid = relForm->oid;
+
+ if (is_publishable_class(relid, relForm))
+ result = lappend_oid(result, relid);
+ }
+
+ table_endscan(scan);
+
+ table_close(classRel, AccessShareLock);
+ return result;
+}
+
/*
* Get publication using oid
*
@@ -426,10 +515,12 @@ GetPublication(Oid pubid)
pub->oid = pubid;
pub->name = pstrdup(NameStr(pubform->pubname));
pub->alltables = pubform->puballtables;
+ pub->allsequences = pubform->puballsequences;
pub->pubactions.pubinsert = pubform->pubinsert;
pub->pubactions.pubupdate = pubform->pubupdate;
pub->pubactions.pubdelete = pubform->pubdelete;
pub->pubactions.pubtruncate = pubform->pubtruncate;
+ pub->pubactions.pubsequence = pubform->pubsequence;
pub->pubviaroot = pubform->pubviaroot;
ReleaseSysCache(tup);
@@ -555,3 +646,56 @@ pg_get_publication_tables(PG_FUNCTION_ARGS)
SRF_RETURN_DONE(funcctx);
}
+
+/*
+ * Returns Oids of sequences in a publication.
+ */
+Datum
+pg_get_publication_sequences(PG_FUNCTION_ARGS)
+{
+ FuncCallContext *funcctx;
+ char *pubname = text_to_cstring(PG_GETARG_TEXT_PP(0));
+ Publication *publication;
+ List *sequences;
+
+ /* stuff done only on the first call of the function */
+ if (SRF_IS_FIRSTCALL())
+ {
+ MemoryContext oldcontext;
+
+ /* create a function context for cross-call persistence */
+ funcctx = SRF_FIRSTCALL_INIT();
+
+ /* switch to memory context appropriate for multiple function calls */
+ oldcontext = MemoryContextSwitchTo(funcctx->multi_call_memory_ctx);
+
+ publication = GetPublicationByName(pubname, false);
+
+ /*
+ * Publications support partitioned tables, although all changes are
+ * replicated using leaf partition identity and schema, so we only
+ * need those.
+ */
+ if (publication->allsequences)
+ sequences = GetAllSequencesPublicationRelations();
+ else
+ sequences = GetPublicationSequenceRelations(publication->oid);
+
+ funcctx->user_fctx = (void *) sequences;
+
+ MemoryContextSwitchTo(oldcontext);
+ }
+
+ /* stuff done on every call of the function */
+ funcctx = SRF_PERCALL_SETUP();
+ sequences = (List *) funcctx->user_fctx;
+
+ if (funcctx->call_cntr < list_length(sequences))
+ {
+ Oid relid = list_nth_oid(sequences, funcctx->call_cntr);
+
+ SRF_RETURN_NEXT(funcctx, ObjectIdGetDatum(relid));
+ }
+
+ SRF_RETURN_DONE(funcctx);
+}
diff --git a/src/backend/catalog/system_views.sql b/src/backend/catalog/system_views.sql
index 999d984068..16f050023a 100644
--- a/src/backend/catalog/system_views.sql
+++ b/src/backend/catalog/system_views.sql
@@ -372,6 +372,16 @@ CREATE VIEW pg_publication_tables AS
pg_class C JOIN pg_namespace N ON (N.oid = C.relnamespace)
WHERE C.oid = GPT.relid;
+CREATE VIEW pg_publication_sequences AS
+ SELECT
+ P.pubname AS pubname,
+ N.nspname AS schemaname,
+ C.relname AS sequencename
+ FROM pg_publication P,
+ LATERAL pg_get_publication_sequences(P.pubname) GPT,
+ pg_class C JOIN pg_namespace N ON (N.oid = C.relnamespace)
+ WHERE C.oid = GPT.relid;
+
CREATE VIEW pg_locks AS
SELECT * FROM pg_lock_status() AS L;
diff --git a/src/backend/commands/publicationcmds.c b/src/backend/commands/publicationcmds.c
index 95c253c8e0..5e5c5feab4 100644
--- a/src/backend/commands/publicationcmds.c
+++ b/src/backend/commands/publicationcmds.c
@@ -16,6 +16,7 @@
#include "access/genam.h"
#include "access/htup_details.h"
+#include "access/relation.h"
#include "access/table.h"
#include "access/xact.h"
#include "catalog/catalog.h"
@@ -54,6 +55,12 @@ static void PublicationAddTables(Oid pubid, List *rels, bool if_not_exists,
AlterPublicationStmt *stmt);
static void PublicationDropTables(Oid pubid, List *rels, bool missing_ok);
+static List *OpenSequenceList(List *sequences);
+static void CloseSequenceList(List *rels);
+static void PublicationAddSequences(Oid pubid, List *rels, bool if_not_exists,
+ AlterPublicationStmt *stmt);
+static void PublicationDropSequences(Oid pubid, List *rels, bool missing_ok);
+
static void
parse_publication_options(List *options,
bool *publish_given,
@@ -71,6 +78,7 @@ parse_publication_options(List *options,
pubactions->pubupdate = true;
pubactions->pubdelete = true;
pubactions->pubtruncate = true;
+ pubactions->pubsequence = true;
*publish_via_partition_root = false;
/* Parse options */
@@ -97,6 +105,7 @@ parse_publication_options(List *options,
pubactions->pubupdate = false;
pubactions->pubdelete = false;
pubactions->pubtruncate = false;
+ pubactions->pubsequence = false;
*publish_given = true;
publish = defGetString(defel);
@@ -119,6 +128,8 @@ parse_publication_options(List *options,
pubactions->pubdelete = true;
else if (strcmp(publish_opt, "truncate") == 0)
pubactions->pubtruncate = true;
+ else if (strcmp(publish_opt, "sequence") == 0)
+ pubactions->pubsequence = true;
else
ereport(ERROR,
(errcode(ERRCODE_SYNTAX_ERROR),
@@ -210,6 +221,8 @@ CreatePublication(CreatePublicationStmt *stmt)
BoolGetDatum(pubactions.pubdelete);
values[Anum_pg_publication_pubtruncate - 1] =
BoolGetDatum(pubactions.pubtruncate);
+ values[Anum_pg_publication_pubsequence - 1] =
+ BoolGetDatum(pubactions.pubsequence);
values[Anum_pg_publication_pubviaroot - 1] =
BoolGetDatum(publish_via_partition_root);
@@ -374,9 +387,9 @@ AlterPublicationTables(AlterPublicationStmt *stmt, Relation rel,
rels = OpenTableList(stmt->tables);
- if (stmt->tableAction == DEFELEM_ADD)
+ if (stmt->action == DEFELEM_ADD)
PublicationAddTables(pubid, rels, false, stmt);
- else if (stmt->tableAction == DEFELEM_DROP)
+ else if (stmt->action == DEFELEM_DROP)
PublicationDropTables(pubid, rels, false);
else /* DEFELEM_SET */
{
@@ -427,6 +440,82 @@ AlterPublicationTables(AlterPublicationStmt *stmt, Relation rel,
CloseTableList(rels);
}
+/*
+ * Add or remove sequence to/from publication.
+ */
+static void
+AlterPublicationSequences(AlterPublicationStmt *stmt, Relation rel,
+ HeapTuple tup)
+{
+ List *rels = NIL;
+ Form_pg_publication pubform = (Form_pg_publication) GETSTRUCT(tup);
+ Oid pubid = pubform->oid;
+
+ /* Check that user is allowed to manipulate the publication tables. */
+ if (pubform->puballsequences)
+ ereport(ERROR,
+ (errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
+ errmsg("publication \"%s\" is defined as FOR ALL SEQUENCES",
+ NameStr(pubform->pubname)),
+ errdetail("Sequences cannot be added to or dropped from FOR ALL TABLES publications.")));
+
+ Assert(list_length(stmt->sequences) > 0);
+
+ rels = OpenSequenceList(stmt->sequences);
+
+ if (stmt->action == DEFELEM_ADD)
+ PublicationAddSequences(pubid, rels, false, stmt);
+ else if (stmt->action == DEFELEM_DROP)
+ PublicationDropSequences(pubid, rels, false);
+ else /* DEFELEM_SET */
+ {
+ List *oldrelids = GetPublicationRelations(pubid,
+ PUBLICATION_PART_ROOT);
+ List *delrels = NIL;
+ ListCell *oldlc;
+
+ /* Calculate which relations to drop. */
+ foreach(oldlc, oldrelids)
+ {
+ Oid oldrelid = lfirst_oid(oldlc);
+ ListCell *newlc;
+ bool found = false;
+
+ foreach(newlc, rels)
+ {
+ Relation newrel = (Relation) lfirst(newlc);
+
+ if (RelationGetRelid(newrel) == oldrelid)
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ Relation oldrel = relation_open(oldrelid,
+ ShareUpdateExclusiveLock);
+
+ delrels = lappend(delrels, oldrel);
+ }
+ }
+
+ /* And drop them. */
+ PublicationDropSequences(pubid, delrels, true);
+
+ /*
+ * Don't bother calculating the difference for adding, we'll catch and
+ * skip existing ones when doing catalog update.
+ */
+ PublicationAddSequences(pubid, rels, true, stmt);
+
+ CloseSequenceList(delrels);
+ }
+
+ CloseSequenceList(rels);
+}
+
/*
* Alter the existing publication.
*
@@ -460,8 +549,12 @@ AlterPublication(AlterPublicationStmt *stmt)
if (stmt->options)
AlterPublicationOptions(stmt, rel, tup);
- else
+ else if (stmt->tables)
AlterPublicationTables(stmt, rel, tup);
+ else if (stmt->sequences)
+ AlterPublicationSequences(stmt, rel, tup);
+ else
+ Assert(false);
/* Cleanup. */
heap_freetuple(tup);
@@ -666,6 +759,139 @@ PublicationDropTables(Oid pubid, List *rels, bool missing_ok)
}
}
+/*
+ * Open relations specified by a RangeVar list.
+ * The returned tables are locked in ShareUpdateExclusiveLock mode in order to
+ * add them to a publication.
+ */
+static List *
+OpenSequenceList(List *sequences)
+{
+ List *relids = NIL;
+ List *rels = NIL;
+ ListCell *lc;
+
+ /*
+ * Open, share-lock, and check all the explicitly-specified relations
+ */
+ foreach(lc, sequences)
+ {
+ RangeVar *rv = castNode(RangeVar, lfirst(lc));
+ Relation rel;
+ Oid myrelid;
+
+ /* Allow query cancel in case this takes a long time */
+ CHECK_FOR_INTERRUPTS();
+
+ rel = relation_openrv(rv, ShareUpdateExclusiveLock);
+ myrelid = RelationGetRelid(rel);
+
+ /*
+ * Filter out duplicates if user specifies "foo, foo".
+ *
+ * Note that this algorithm is known to not be very efficient (O(N^2))
+ * but given that it only works on list of sequences given to us by user
+ * it's deemed acceptable.
+ */
+ if (list_member_oid(relids, myrelid))
+ {
+ relation_close(rel, ShareUpdateExclusiveLock);
+ continue;
+ }
+
+ rels = lappend(rels, rel);
+ relids = lappend_oid(relids, myrelid);
+ }
+
+ list_free(relids);
+
+ return rels;
+}
+
+/*
+ * Close all relations in the list.
+ */
+static void
+CloseSequenceList(List *rels)
+{
+ ListCell *lc;
+
+ foreach(lc, rels)
+ {
+ Relation rel = (Relation) lfirst(lc);
+
+ relation_close(rel, NoLock);
+ }
+}
+
+/*
+ * Add listed tables to the publication.
+ */
+static void
+PublicationAddSequences(Oid pubid, List *rels, bool if_not_exists,
+ AlterPublicationStmt *stmt)
+{
+ ListCell *lc;
+
+ Assert(!stmt || !stmt->for_all_sequences);
+
+ foreach(lc, rels)
+ {
+ Relation rel = (Relation) lfirst(lc);
+ ObjectAddress obj;
+
+ /* Must be owner of the table or superuser. */
+ if (!pg_class_ownercheck(RelationGetRelid(rel), GetUserId()))
+ aclcheck_error(ACLCHECK_NOT_OWNER, get_relkind_objtype(rel->rd_rel->relkind),
+ RelationGetRelationName(rel));
+
+ obj = publication_add_relation(pubid, rel, if_not_exists);
+ if (stmt)
+ {
+ EventTriggerCollectSimpleCommand(obj, InvalidObjectAddress,
+ (Node *) stmt);
+
+ InvokeObjectPostCreateHook(PublicationRelRelationId,
+ obj.objectId, 0);
+ }
+ }
+}
+
+/*
+ * Remove listed sequences from the publication.
+ */
+static void
+PublicationDropSequences(Oid pubid, List *rels, bool missing_ok)
+{
+ ObjectAddress obj;
+ ListCell *lc;
+ Oid prid;
+
+ foreach(lc, rels)
+ {
+ Relation rel = (Relation) lfirst(lc);
+ Oid relid = RelationGetRelid(rel);
+
+ prid = GetSysCacheOid2(PUBLICATIONRELMAP, Anum_pg_publication_rel_oid,
+ ObjectIdGetDatum(relid),
+ ObjectIdGetDatum(pubid));
+ if (!OidIsValid(prid))
+ {
+ if (missing_ok)
+ continue;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_UNDEFINED_OBJECT),
+ errmsg("relation \"%s\" is not part of the publication",
+ RelationGetRelationName(rel))));
+ }
+
+ ObjectAddressSet(obj, PublicationRelRelationId, prid);
+ performDeletion(&obj, DROP_CASCADE, 0);
+ }
+}
+
+
/*
* Internal workhorse for changing a publication owner
*/
diff --git a/src/backend/commands/sequence.c b/src/backend/commands/sequence.c
index 0415df9ccb..6898a45365 100644
--- a/src/backend/commands/sequence.c
+++ b/src/backend/commands/sequence.c
@@ -94,7 +94,7 @@ static HTAB *seqhashtab = NULL; /* hash table for SeqTable items */
*/
static SeqTableData *last_used_seq = NULL;
-static void fill_seq_with_data(Relation rel, HeapTuple tuple);
+static void fill_seq_with_data(Relation rel, HeapTuple tuple, bool create);
static Relation lock_and_open_sequence(SeqTable seq);
static void create_seq_hashtable(void);
static void init_sequence(Oid relid, SeqTable *p_elm, Relation *p_rel);
@@ -222,7 +222,7 @@ DefineSequence(ParseState *pstate, CreateSeqStmt *seq)
/* now initialize the sequence's data */
tuple = heap_form_tuple(tupDesc, value, null);
- fill_seq_with_data(rel, tuple);
+ fill_seq_with_data(rel, tuple, true);
/* process OWNED BY if given */
if (owned_by)
@@ -327,7 +327,86 @@ ResetSequence(Oid seq_relid)
/*
* Insert the modified tuple into the new storage file.
*/
- fill_seq_with_data(seq_rel, tuple);
+ fill_seq_with_data(seq_rel, tuple, true);
+
+ /* Clear local cache so that we don't think we have cached numbers */
+ /* Note that we do not change the currval() state */
+ elm->cached = elm->last;
+
+ relation_close(seq_rel, NoLock);
+}
+
+/*
+ * Reset a sequence to its initial value.
+ *
+ * The change is made transactionally, so that on failure of the current
+ * transaction, the sequence will be restored to its previous state.
+ * We do that by creating a whole new relfilenode for the sequence; so this
+ * works much like the rewriting forms of ALTER TABLE.
+ *
+ * Caller is assumed to have acquired AccessExclusiveLock on the sequence,
+ * which must not be released until end of transaction. Caller is also
+ * responsible for permissions checking.
+ */
+void
+ResetSequence2(Oid seq_relid, int64 last_value, int64 log_cnt, bool is_called)
+{
+ Relation seq_rel;
+ SeqTable elm;
+ Form_pg_sequence_data seq;
+ Buffer buf;
+ HeapTupleData seqdatatuple;
+ HeapTuple tuple;
+
+ /*
+ * Read the old sequence. This does a bit more work than really
+ * necessary, but it's simple, and we do want to double-check that it's
+ * indeed a sequence.
+ */
+ init_sequence(seq_relid, &elm, &seq_rel);
+ (void) read_seq_tuple(seq_rel, &buf, &seqdatatuple);
+
+ /*
+ * Copy the existing sequence tuple.
+ */
+ tuple = heap_copytuple(&seqdatatuple);
+
+ /* Now we're done with the old page */
+ UnlockReleaseBuffer(buf);
+
+ /*
+ * Modify the copied tuple to execute the restart (compare the RESTART
+ * action in AlterSequence)
+ */
+ seq = (Form_pg_sequence_data) GETSTRUCT(tuple);
+ seq->last_value = last_value;
+ seq->is_called = is_called;
+ seq->log_cnt = log_cnt;
+
+ /*
+ * Create a new storage file for the sequence.
+ */
+ RelationSetNewRelfilenode(seq_rel, seq_rel->rd_rel->relpersistence);
+
+ /*
+ * Ensure sequence's relfrozenxid is at 0, since it won't contain any
+ * unfrozen XIDs. Same with relminmxid, since a sequence will never
+ * contain multixacts.
+ */
+ Assert(seq_rel->rd_rel->relfrozenxid == InvalidTransactionId);
+ Assert(seq_rel->rd_rel->relminmxid == InvalidMultiXactId);
+
+ /*
+ * Insert the modified tuple into the new storage file.
+ *
+ * XXX Maybe this should also use created=true, just like the other places
+ * calling fill_seq_with_data. That's probably needed for correct cascading
+ * replication.
+ *
+ * XXX That'd mean all fill_seq_with_data callers use created=true, making
+ * the parameter unnecessary.
+ */
+ fill_seq_with_data(seq_rel, tuple, false);
/* Clear local cache so that we don't think we have cached numbers */
/* Note that we do not change the currval() state */
@@ -340,7 +419,7 @@ ResetSequence(Oid seq_relid)
* Initialize a sequence's relation with the specified tuple as content
*/
static void
-fill_seq_with_data(Relation rel, HeapTuple tuple)
+fill_seq_with_data(Relation rel, HeapTuple tuple, bool create)
{
Buffer buf;
Page page;
@@ -378,8 +457,21 @@ fill_seq_with_data(Relation rel, HeapTuple tuple)
/* check the comment above nextval_internal()'s equivalent call. */
if (RelationNeedsWAL(rel))
+ {
GetTopTransactionId();
+ /*
+ * Ensure we have a proper XID, which will be included in the XLOG
+ * record by XLogRecordAssemble. Otherwise the first nextval() in
+ * a subxact (without any preceding changes) would get XID 0,
+ * and it'd be impossible to decide which top xact it belongs to.
+ * It'd also trigger assert in DecodeSequence.
+ *
+ * XXX Not sure if this is the best solution.
+ */
+ GetCurrentTransactionId();
+ }
+
START_CRIT_SECTION();
MarkBufferDirty(buf);
@@ -399,6 +491,7 @@ fill_seq_with_data(Relation rel, HeapTuple tuple)
XLogRegisterBuffer(0, buf, REGBUF_WILL_INIT);
xlrec.node = rel->rd_node;
+ xlrec.created = create;
XLogRegisterData((char *) &xlrec, sizeof(xl_seq_rec));
XLogRegisterData((char *) tuple->t_data, tuple->t_len);
@@ -502,7 +595,7 @@ AlterSequence(ParseState *pstate, AlterSeqStmt *stmt)
/*
* Insert the modified tuple into the new storage file.
*/
- fill_seq_with_data(seqrel, newdatatuple);
+ fill_seq_with_data(seqrel, newdatatuple, true);
}
/* process OWNED BY if given */
@@ -766,8 +859,21 @@ nextval_internal(Oid relid, bool check_permissions)
* (Have to do that here, so we're outside the critical section)
*/
if (logit && RelationNeedsWAL(seqrel))
+ {
GetTopTransactionId();
+ /*
+ * Ensure we have a proper XID, which will be included in the XLOG
+ * record by XLogRecordAssemble. Otherwise the first nextval() in
+ * a subxact (without any preceding changes) would get XID 0,
+ * and it'd be impossible to decide which top xact it belongs to.
+ * It'd also trigger assert in DecodeSequence.
+ *
+ * XXX Not sure if this is the best solution.
+ */
+ GetCurrentTransactionId();
+ }
+
/* ready to change the on-disk (or really, in-buffer) tuple */
START_CRIT_SECTION();
@@ -803,6 +909,7 @@ nextval_internal(Oid relid, bool check_permissions)
seq->log_cnt = 0;
xlrec.node = seqrel->rd_node;
+ xlrec.created = false;
XLogRegisterData((char *) &xlrec, sizeof(xl_seq_rec));
XLogRegisterData((char *) seqdatatuple.t_data, seqdatatuple.t_len);
@@ -977,8 +1084,21 @@ do_setval(Oid relid, int64 next, bool iscalled)
/* check the comment above nextval_internal()'s equivalent call. */
if (RelationNeedsWAL(seqrel))
+ {
GetTopTransactionId();
+ /*
+ * Ensure we have a proper XID, which will be included in the XLOG
+ * record by XLogRecordAssemble. Otherwise the first nextval() in
+ * a subxact (without any preceding changes) would get XID 0,
+ * and it'd be impossible to decide which top xact it belongs to.
+ * It'd also trigger assert in DecodeSequence.
+ *
+ * XXX Not sure if this is the best solution.
+ */
+ GetCurrentTransactionId();
+ }
+
/* ready to change the on-disk (or really, in-buffer) tuple */
START_CRIT_SECTION();
@@ -999,6 +1119,8 @@ do_setval(Oid relid, int64 next, bool iscalled)
XLogRegisterBuffer(0, buf, REGBUF_WILL_INIT);
xlrec.node = seqrel->rd_node;
+ xlrec.created = false;
+
XLogRegisterData((char *) &xlrec, sizeof(xl_seq_rec));
XLogRegisterData((char *) seqdatatuple.t_data, seqdatatuple.t_len);
diff --git a/src/backend/commands/subscriptioncmds.c b/src/backend/commands/subscriptioncmds.c
index 8aa6de1785..2112be663b 100644
--- a/src/backend/commands/subscriptioncmds.c
+++ b/src/backend/commands/subscriptioncmds.c
@@ -47,6 +47,7 @@
#include "utils/syscache.h"
static List *fetch_table_list(WalReceiverConn *wrconn, List *publications);
+static List *fetch_sequence_list(WalReceiverConn *wrconn, List *publications);
static void check_duplicates_in_publist(List *publist, Datum *datums);
static List *merge_publications(List *oldpublist, List *newpublist, bool addpub, const char *subname);
static void ReportSlotConnectionError(List *rstates, Oid subid, char *slotname, char *err);
@@ -461,6 +462,7 @@ CreateSubscription(CreateSubscriptionStmt *stmt, bool isTopLevel)
char *err;
WalReceiverConn *wrconn;
List *tables;
+ List *sequences;
ListCell *lc;
char table_state;
@@ -498,6 +500,26 @@ CreateSubscription(CreateSubscriptionStmt *stmt, bool isTopLevel)
InvalidXLogRecPtr);
}
+ /*
+ * Get the sequence list from publisher and build local sequence
+ * status info.
+ */
+ sequences = fetch_sequence_list(wrconn, publications);
+ foreach(lc, sequences)
+ {
+ RangeVar *rv = (RangeVar *) lfirst(lc);
+ Oid relid;
+
+ relid = RangeVarGetRelid(rv, AccessShareLock, false);
+
+ /* Check for supported relkind. */
+ CheckSubscriptionRelkind(get_rel_relkind(relid),
+ rv->schemaname, rv->relname);
+
+ AddSubscriptionRelState(subid, relid, table_state,
+ InvalidXLogRecPtr);
+ }
+
/*
* If requested, create permanent slot for the subscription. We
* won't use the initial snapshot for anything, so no need to
@@ -644,6 +666,10 @@ AlterSubscription_refresh(Subscription *sub, bool copy_data)
{
Oid relid = subrel_local_oids[off];
+ /* XXX ignore sequences - maybe do this in GetSubscriptionRelations? */
+ if (get_rel_relkind(relid) == RELKIND_SEQUENCE)
+ continue;
+
if (!bsearch(&relid, pubrel_local_oids,
list_length(pubrel_names), sizeof(Oid), oid_cmp))
{
@@ -735,6 +761,185 @@ AlterSubscription_refresh(Subscription *sub, bool copy_data)
ReplicationSlotDropAtPubNode(wrconn, syncslotname, true);
}
}
+
+ /*
+ * XXX now do the same thing for sequences, maybe before the preceding
+ * block, or earlier?
+ */
+
+ /* Get the table list from publisher. */
+ pubrel_names = fetch_sequence_list(wrconn, sub->publications);
+
+ /* Get local table list. */
+ subrel_states = GetSubscriptionRelations(sub->oid);
+
+ /*
+ * Build qsorted array of local table oids for faster lookup. This can
+ * potentially contain all tables in the database so speed of lookup
+ * is important.
+ */
+ subrel_local_oids = palloc(list_length(subrel_states) * sizeof(Oid));
+ off = 0;
+ foreach(lc, subrel_states)
+ {
+ SubscriptionRelState *relstate = (SubscriptionRelState *) lfirst(lc);
+
+ subrel_local_oids[off++] = relstate->relid;
+ }
+ qsort(subrel_local_oids, list_length(subrel_states),
+ sizeof(Oid), oid_cmp);
+
+ /*
+ * Rels that we want to remove from subscription and drop any slots
+ * and origins corresponding to them.
+ */
+ sub_remove_rels = palloc(list_length(subrel_states) * sizeof(SubRemoveRels));
+
+ /*
+ * Walk over the remote tables and try to match them to locally known
+ * tables. If the table is not known locally create a new state for
+ * it.
+ *
+ * Also builds array of local oids of remote tables for the next step.
+ */
+ off = 0;
+ pubrel_local_oids = palloc(list_length(pubrel_names) * sizeof(Oid));
+
+ foreach(lc, pubrel_names)
+ {
+ RangeVar *rv = (RangeVar *) lfirst(lc);
+ Oid relid;
+
+ relid = RangeVarGetRelid(rv, AccessShareLock, false);
+
+ /* Check for supported relkind. */
+ CheckSubscriptionRelkind(get_rel_relkind(relid),
+ rv->schemaname, rv->relname);
+
+ pubrel_local_oids[off++] = relid;
+
+ if (!bsearch(&relid, subrel_local_oids,
+ list_length(subrel_states), sizeof(Oid), oid_cmp))
+ {
+ AddSubscriptionRelState(sub->oid, relid,
+ copy_data ? SUBREL_STATE_INIT : SUBREL_STATE_READY,
+ InvalidXLogRecPtr);
+ ereport(DEBUG1,
+ (errmsg_internal("table \"%s.%s\" added to subscription \"%s\"",
+ rv->schemaname, rv->relname, sub->name)));
+ }
+ }
+
+ /*
+ * Next remove state for tables we should not care about anymore using
+ * the data we collected above
+ */
+ qsort(pubrel_local_oids, list_length(pubrel_names),
+ sizeof(Oid), oid_cmp);
+
+ remove_rel_len = 0;
+ for (off = 0; off < list_length(subrel_states); off++)
+ {
+ Oid relid = subrel_local_oids[off];
+
+ /* XXX ignore non-sequences - maybe do this in GetSubscriptionRelations? */
+ if (get_rel_relkind(relid) != RELKIND_SEQUENCE)
+ continue;
+
+ if (!bsearch(&relid, pubrel_local_oids,
+ list_length(pubrel_names), sizeof(Oid), oid_cmp))
+ {
+ char state;
+ XLogRecPtr statelsn;
+
+ /*
+ * Lock pg_subscription_rel with AccessExclusiveLock to
+ * prevent any race conditions with the apply worker
+ * re-launching workers at the same time this code is trying
+ * to remove those tables.
+ *
+ * Even if new worker for this particular rel is restarted it
+ * won't be able to make any progress as we hold exclusive
+ * lock on subscription_rel till the transaction end. It will
+ * simply exit as there is no corresponding rel entry.
+ *
+ * This locking also ensures that the state of rels won't
+ * change till we are done with this refresh operation.
+ */
+ if (!rel)
+ rel = table_open(SubscriptionRelRelationId, AccessExclusiveLock);
+
+ /* Last known rel state. */
+ state = GetSubscriptionRelState(sub->oid, relid, &statelsn);
+
+ sub_remove_rels[remove_rel_len].relid = relid;
+ sub_remove_rels[remove_rel_len++].state = state;
+
+ elog(WARNING, "B: remove rel %d", relid);
+
+ RemoveSubscriptionRel(sub->oid, relid);
+
+ logicalrep_worker_stop(sub->oid, relid);
+
+ /*
+ * For READY state, we would have already dropped the
+ * tablesync origin.
+ */
+ if (state != SUBREL_STATE_READY)
+ {
+ char originname[NAMEDATALEN];
+
+ /*
+ * Drop the tablesync's origin tracking if exists.
+ *
+ * It is possible that the origin is not yet created for
+ * tablesync worker, this can happen for the states before
+ * SUBREL_STATE_FINISHEDCOPY. The apply worker can also
+ * concurrently try to drop the origin and by this time
+ * the origin might be already removed. For these reasons,
+ * passing missing_ok = true.
+ */
+ ReplicationOriginNameForTablesync(sub->oid, relid, originname,
+ sizeof(originname));
+ replorigin_drop_by_name(originname, true, false);
+ }
+
+ ereport(DEBUG1,
+ (errmsg_internal("table \"%s.%s\" removed from subscription \"%s\"",
+ get_namespace_name(get_rel_namespace(relid)),
+ get_rel_name(relid),
+ sub->name)));
+ }
+ }
+
+ /*
+ * Drop the tablesync slots associated with removed tables. This has
+ * to be at the end because otherwise if there is an error while doing
+ * the database operations we won't be able to rollback dropped slots.
+ */
+ for (off = 0; off < remove_rel_len; off++)
+ {
+ if (sub_remove_rels[off].state != SUBREL_STATE_READY &&
+ sub_remove_rels[off].state != SUBREL_STATE_SYNCDONE)
+ {
+ char syncslotname[NAMEDATALEN] = {0};
+
+ /*
+ * For READY/SYNCDONE states we know the tablesync slot has
+ * already been dropped by the tablesync worker.
+ *
+ * For other states, there is no certainty, maybe the slot
+ * does not exist yet. Also, if we fail after removing some of
+ * the slots, next time, it will again try to drop already
+ * dropped slots and fail. For these reasons, we allow
+ * missing_ok = true for the drop.
+ */
+ ReplicationSlotNameForTablesync(sub->oid, sub_remove_rels[off].relid,
+ syncslotname, sizeof(syncslotname));
+ ReplicationSlotDropAtPubNode(wrconn, syncslotname, true);
+ }
+ }
+
}
PG_FINALLY();
{
@@ -1534,6 +1739,75 @@ fetch_table_list(WalReceiverConn *wrconn, List *publications)
return tablelist;
}
+/*
+ * Get the list of sequences which belong to specified publications on the
+ * publisher connection.
+ */
+static List *
+fetch_sequence_list(WalReceiverConn *wrconn, List *publications)
+{
+ WalRcvExecResult *res;
+ StringInfoData cmd;
+ TupleTableSlot *slot;
+ Oid tableRow[2] = {TEXTOID, TEXTOID};
+ ListCell *lc;
+ bool first;
+ List *tablelist = NIL;
+
+ Assert(list_length(publications) > 0);
+
+ initStringInfo(&cmd);
+ appendStringInfoString(&cmd, "SELECT DISTINCT s.schemaname, s.tablename\n"
+ " FROM pg_catalog.pg_publication_sequences s\n"
+ " WHERE s.pubname IN (");
+ first = true;
+ foreach(lc, publications)
+ {
+ char *pubname = strVal(lfirst(lc));
+
+ if (first)
+ first = false;
+ else
+ appendStringInfoString(&cmd, ", ");
+
+ appendStringInfoString(&cmd, quote_literal_cstr(pubname));
+ }
+ appendStringInfoChar(&cmd, ')');
+
+ res = walrcv_exec(wrconn, cmd.data, 2, tableRow);
+ pfree(cmd.data);
+
+ if (res->status != WALRCV_OK_TUPLES)
+ ereport(ERROR,
+ (errmsg("could not receive list of replicated tables from the publisher: %s",
+ res->err)));
+
+ /* Process tables. */
+ slot = MakeSingleTupleTableSlot(res->tupledesc, &TTSOpsMinimalTuple);
+ while (tuplestore_gettupleslot(res->tuplestore, true, false, slot))
+ {
+ char *nspname;
+ char *relname;
+ bool isnull;
+ RangeVar *rv;
+
+ nspname = TextDatumGetCString(slot_getattr(slot, 1, &isnull));
+ Assert(!isnull);
+ relname = TextDatumGetCString(slot_getattr(slot, 2, &isnull));
+ Assert(!isnull);
+
+ rv = makeRangeVar(nspname, relname, -1);
+ tablelist = lappend(tablelist, rv);
+
+ ExecClearTuple(slot);
+ }
+ ExecDropSingleTupleTableSlot(slot);
+
+ walrcv_clear_result(res);
+
+ return tablelist;
+}
+
/*
* This is to report the connection failure while dropping replication slots.
* Here, we report the WARNING for all tablesync slots so that user can drop
diff --git a/src/backend/executor/execReplication.c b/src/backend/executor/execReplication.c
index 1e285e0349..a1c7880be6 100644
--- a/src/backend/executor/execReplication.c
+++ b/src/backend/executor/execReplication.c
@@ -619,11 +619,11 @@ CheckSubscriptionRelkind(char relkind, const char *nspname,
errdetail("\"%s.%s\" is a foreign table.",
nspname, relname)));
- if (relkind != RELKIND_RELATION && relkind != RELKIND_PARTITIONED_TABLE)
+ if (relkind != RELKIND_RELATION && relkind != RELKIND_PARTITIONED_TABLE && relkind != RELKIND_SEQUENCE)
ereport(ERROR,
(errcode(ERRCODE_WRONG_OBJECT_TYPE),
errmsg("cannot use relation \"%s.%s\" as logical replication target",
nspname, relname),
- errdetail("\"%s.%s\" is not a table.",
+ errdetail("\"%s.%s\" is not a table or a sequence.",
nspname, relname)));
}
diff --git a/src/backend/nodes/copyfuncs.c b/src/backend/nodes/copyfuncs.c
index 621f7ce068..b98c1b2f6e 100644
--- a/src/backend/nodes/copyfuncs.c
+++ b/src/backend/nodes/copyfuncs.c
@@ -4831,8 +4831,10 @@ _copyAlterPublicationStmt(const AlterPublicationStmt *from)
COPY_STRING_FIELD(pubname);
COPY_NODE_FIELD(options);
COPY_NODE_FIELD(tables);
+ COPY_NODE_FIELD(sequences);
COPY_SCALAR_FIELD(for_all_tables);
- COPY_SCALAR_FIELD(tableAction);
+ COPY_SCALAR_FIELD(for_all_sequences);
+ COPY_SCALAR_FIELD(action);
return newnode;
}
diff --git a/src/backend/nodes/equalfuncs.c b/src/backend/nodes/equalfuncs.c
index 3033c1934c..b72a5831e3 100644
--- a/src/backend/nodes/equalfuncs.c
+++ b/src/backend/nodes/equalfuncs.c
@@ -2300,8 +2300,10 @@ _equalAlterPublicationStmt(const AlterPublicationStmt *a,
COMPARE_STRING_FIELD(pubname);
COMPARE_NODE_FIELD(options);
COMPARE_NODE_FIELD(tables);
+ COMPARE_NODE_FIELD(sequences);
COMPARE_SCALAR_FIELD(for_all_tables);
- COMPARE_SCALAR_FIELD(tableAction);
+ COMPARE_SCALAR_FIELD(for_all_sequences);
+ COMPARE_SCALAR_FIELD(action);
return true;
}
diff --git a/src/backend/parser/gram.y b/src/backend/parser/gram.y
index 9ee90e3f13..880883e3b8 100644
--- a/src/backend/parser/gram.y
+++ b/src/backend/parser/gram.y
@@ -434,6 +434,7 @@ static Node *makeRecursiveViewSelect(char *relname, List *aliases, Node *query);
%type <node> group_by_item empty_grouping_set rollup_clause cube_clause
%type <node> grouping_sets_clause
%type <node> opt_publication_for_tables publication_for_tables
+/* %type <node> opt_publication_for_sequences publication_for_sequences */
%type <list> opt_fdw_options fdw_options
%type <defelt> fdw_option
@@ -9612,6 +9613,7 @@ AlterOwnerStmt: ALTER AGGREGATE aggregate_with_argtypes OWNER TO RoleSpec
*****************************************************************************/
CreatePublicationStmt:
+ /* CREATE PUBLICATION name opt_publication_for_tables opt_publication_for_sequences opt_definition */
CREATE PUBLICATION name opt_publication_for_tables opt_definition
{
CreatePublicationStmt *n = makeNode(CreatePublicationStmt);
@@ -9646,6 +9648,12 @@ publication_for_tables:
}
;
+/*
+ * FIXME
+ *
+ * opt_publication_for_sequences and publication_for_sequences should be
+ * copies for sequences
+ */
/*****************************************************************************
*
@@ -9657,6 +9665,12 @@ publication_for_tables:
*
* ALTER PUBLICATION name SET TABLE table [, table2]
*
+ * ALTER PUBLICATION name ADD SEQUENCE sequence [, sequence2]
+ *
+ * ALTER PUBLICATION name DROP SEQUENCE sequence [, sequence2]
+ *
+ * ALTER PUBLICATION name SET SEQUENCE sequence [, sequence2]
+ *
*****************************************************************************/
AlterPublicationStmt:
@@ -9672,7 +9686,7 @@ AlterPublicationStmt:
AlterPublicationStmt *n = makeNode(AlterPublicationStmt);
n->pubname = $3;
n->tables = $6;
- n->tableAction = DEFELEM_ADD;
+ n->action = DEFELEM_ADD;
$$ = (Node *)n;
}
| ALTER PUBLICATION name SET TABLE relation_expr_list
@@ -9680,7 +9694,7 @@ AlterPublicationStmt:
AlterPublicationStmt *n = makeNode(AlterPublicationStmt);
n->pubname = $3;
n->tables = $6;
- n->tableAction = DEFELEM_SET;
+ n->action = DEFELEM_SET;
$$ = (Node *)n;
}
| ALTER PUBLICATION name DROP TABLE relation_expr_list
@@ -9688,7 +9702,31 @@ AlterPublicationStmt:
AlterPublicationStmt *n = makeNode(AlterPublicationStmt);
n->pubname = $3;
n->tables = $6;
- n->tableAction = DEFELEM_DROP;
+ n->action = DEFELEM_DROP;
+ $$ = (Node *)n;
+ }
+ | ALTER PUBLICATION name ADD_P SEQUENCE relation_expr_list
+ {
+ AlterPublicationStmt *n = makeNode(AlterPublicationStmt);
+ n->pubname = $3;
+ n->sequences = $6;
+ n->action = DEFELEM_ADD;
+ $$ = (Node *)n;
+ }
+ | ALTER PUBLICATION name SET SEQUENCE relation_expr_list
+ {
+ AlterPublicationStmt *n = makeNode(AlterPublicationStmt);
+ n->pubname = $3;
+ n->sequences = $6;
+ n->action = DEFELEM_SET;
+ $$ = (Node *)n;
+ }
+ | ALTER PUBLICATION name DROP SEQUENCE relation_expr_list
+ {
+ AlterPublicationStmt *n = makeNode(AlterPublicationStmt);
+ n->pubname = $3;
+ n->sequences = $6;
+ n->action = DEFELEM_DROP;
$$ = (Node *)n;
}
;
diff --git a/src/backend/replication/logical/decode.c b/src/backend/replication/logical/decode.c
index 70670169ac..f6296941d9 100644
--- a/src/backend/replication/logical/decode.c
+++ b/src/backend/replication/logical/decode.c
@@ -42,6 +42,7 @@
#include "replication/reorderbuffer.h"
#include "replication/snapbuild.h"
#include "storage/standby.h"
+#include "commands/sequence.h"
typedef struct XLogRecordBuffer
{
@@ -74,10 +75,11 @@ static void DecodeAbort(LogicalDecodingContext *ctx, XLogRecordBuffer *buf,
bool two_phase);
static void DecodePrepare(LogicalDecodingContext *ctx, XLogRecordBuffer *buf,
xl_xact_parsed_prepare *parsed);
-
+static void DecodeSequence(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
/* common function to decode tuples */
static void DecodeXLogTuple(char *data, Size len, ReorderBufferTupleBuf *tup);
+static void DecodeSeqTuple(char *data, Size len, ReorderBufferTupleBuf *tuple);
/* helper functions for decoding transactions */
static inline bool FilterPrepare(LogicalDecodingContext *ctx,
@@ -158,6 +160,10 @@ LogicalDecodingProcessRecord(LogicalDecodingContext *ctx, XLogReaderState *recor
DecodeLogicalMsgOp(ctx, &buf);
break;
+ case RM_SEQ_ID:
+ DecodeSequence(ctx, &buf);
+ break;
+
/*
* Rmgrs irrelevant for logical decoding; they describe stuff not
* represented in logical decoding. Add new rmgrs in rmgrlist.h's
@@ -173,7 +179,6 @@ LogicalDecodingProcessRecord(LogicalDecodingContext *ctx, XLogReaderState *recor
case RM_HASH_ID:
case RM_GIN_ID:
case RM_GIST_ID:
- case RM_SEQ_ID:
case RM_SPGIST_ID:
case RM_BRIN_ID:
case RM_COMMIT_TS_ID:
@@ -1315,3 +1320,125 @@ DecodeTXNNeedSkip(LogicalDecodingContext *ctx, XLogRecordBuffer *buf,
(txn_dbid != InvalidOid && txn_dbid != ctx->slot->data.database) ||
ctx->fast_forward || FilterByOrigin(ctx, origin_id));
}
+
+/*
+ * Decode Sequence Tuple
+ */
+static void
+DecodeSeqTuple(char *data, Size len, ReorderBufferTupleBuf *tuple)
+{
+ int datalen = len - sizeof(xl_seq_rec) - SizeofHeapTupleHeader;
+
+ Assert(datalen >= 0);
+
+ tuple->tuple.t_len = datalen + SizeofHeapTupleHeader;;
+
+ ItemPointerSetInvalid(&tuple->tuple.t_self);
+
+ tuple->tuple.t_tableOid = InvalidOid;
+
+ memcpy(((char *) tuple->tuple.t_data),
+ data + sizeof(xl_seq_rec),
+ SizeofHeapTupleHeader);
+
+ memcpy(((char *) tuple->tuple.t_data) + SizeofHeapTupleHeader,
+ data + sizeof(xl_seq_rec) + SizeofHeapTupleHeader,
+ datalen);
+}
+
+/*
+ * Handle sequence decode
+ *
+ * Decoding sequences is a bit tricky, because while most sequence actions
+ * are non-transactional (not subject to rollback), some need to be handled
+ * as transactional.
+ *
+ * By default, a sequence increment is non-transactional - we must not queue
+ * it in a transaction as other changes, because the transaction might get
+ * rolled back and we'd discard the increment. The downstream would not be
+ * notified about the increment, which is wrong.
+ *
+ * On the other hand, the sequence may be created in a transaction. In this
+ * case we *should* queue the change as other changes in the transaction,
+ * because we don't want to send the increments for unknown sequence to the
+ * plugin - it might get confused about which sequence it's related to etc.
+ */
+static void
+DecodeSequence(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
+{
+ SnapBuild *builder = ctx->snapshot_builder;
+ ReorderBufferTupleBuf *tuplebuf;
+ RelFileNode target_node;
+ XLogReaderState *r = buf->record;
+ char *tupledata = NULL;
+ Size tuplelen;
+ Size datalen = 0;
+ TransactionId xid = XLogRecGetXid(r);
+ uint8 info = XLogRecGetInfo(buf->record) & ~XLR_INFO_MASK;
+ xl_seq_rec *xlrec;
+ Snapshot snapshot;
+ RepOriginId origin_id = XLogRecGetOrigin(r);
+ bool transactional;
+
+ /* only decode changes flagged with XLOG_SEQ_LOG */
+ if (info != XLOG_SEQ_LOG)
+ elog(ERROR, "unexpected RM_SEQ_ID record type: %u", info);
+
+ /*
+ * If we don't have snapshot or we are just fast-forwarding, there is no
+ * point in decoding messages.
+ */
+ if (SnapBuildCurrentState(builder) < SNAPBUILD_FULL_SNAPSHOT ||
+ ctx->fast_forward)
+ return;
+
+ /* only interested in our database */
+ XLogRecGetBlockTag(r, 0, &target_node, NULL, NULL);
+ if (target_node.dbNode != ctx->slot->data.database)
+ return;
+
+ /* output plugin doesn't look for this origin, no need to queue */
+ if (FilterByOrigin(ctx, XLogRecGetOrigin(r)))
+ return;
+
+ tupledata = XLogRecGetData(r);
+ datalen = XLogRecGetDataLen(r);
+ tuplelen = datalen - SizeOfHeapHeader - sizeof(xl_seq_rec);
+
+ /* extract the WAL record, with "created" flag */
+ xlrec = (xl_seq_rec *) XLogRecGetData(r);
+
+ /* XXX how could we have sequence change without data? */
+ if(!datalen || !tupledata)
+ return;
+
+ tuplebuf = ReorderBufferGetTupleBuf(ctx->reorder, tuplelen);
+ DecodeSeqTuple(tupledata, datalen, tuplebuf);
+
+ /*
+ * Should we handle the sequence increment as transactional or not?
+ *
+ * If the sequence was created in a still-running transaction, treat
+ * it as transactional and queue the increments. Otherwise it needs
+ * to be treated as non-transactional, in which case we send it to
+ * the plugin right away.
+ */
+ transactional = ReorderBufferSequenceIsTransactional(ctx->reorder,
+ target_node,
+ xlrec->created);
+
+ /* Skip the change if already processed (per the snapshot). */
+ if (transactional &&
+ !SnapBuildProcessChange(builder, xid, buf->origptr))
+ return;
+ else if (!transactional &&
+ (SnapBuildCurrentState(builder) != SNAPBUILD_CONSISTENT ||
+ SnapBuildXactNeedsSkip(builder, buf->origptr)))
+ return;
+
+ /* Queue the increment (or send immediately if not transactional). */
+ snapshot = SnapBuildGetOrBuildSnapshot(builder, xid);
+ ReorderBufferQueueSequence(ctx->reorder, xid, snapshot, buf->endptr,
+ origin_id, target_node, transactional,
+ xlrec->created, tuplebuf);
+}
diff --git a/src/backend/replication/logical/logical.c b/src/backend/replication/logical/logical.c
index ffc6160e9f..66a4a4603a 100644
--- a/src/backend/replication/logical/logical.c
+++ b/src/backend/replication/logical/logical.c
@@ -73,6 +73,10 @@ static void truncate_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
static void message_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
XLogRecPtr message_lsn, bool transactional,
const char *prefix, Size message_size, const char *message);
+static void sequence_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ bool transactional, bool created,
+ int64 last_value, int64 log_cnt, bool is_called);
/* streaming callbacks */
static void stream_start_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
@@ -217,6 +221,7 @@ StartupDecodingContext(List *output_plugin_options,
ctx->reorder->apply_truncate = truncate_cb_wrapper;
ctx->reorder->commit = commit_cb_wrapper;
ctx->reorder->message = message_cb_wrapper;
+ ctx->reorder->sequence = sequence_cb_wrapper;
/*
* To support streaming, we require start/stop/abort/commit/change
@@ -1179,6 +1184,43 @@ message_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
error_context_stack = errcallback.previous;
}
+static void
+sequence_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ bool transactional, bool created,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ LogicalDecodingContext *ctx = cache->private_data;
+ LogicalErrorCallbackState state;
+ ErrorContextCallback errcallback;
+
+ Assert(!ctx->fast_forward);
+
+ if (ctx->callbacks.sequence_cb == NULL)
+ return;
+
+ /* Push callback + info on the error context stack */
+ state.ctx = ctx;
+ state.callback_name = "sequence";
+ state.report_location = sequence_lsn;
+ errcallback.callback = output_plugin_error_callback;
+ errcallback.arg = (void *) &state;
+ errcallback.previous = error_context_stack;
+ error_context_stack = &errcallback;
+
+ /* set output state */
+ ctx->accept_writes = true;
+ ctx->write_xid = txn != NULL ? txn->xid : InvalidTransactionId;
+ ctx->write_location = sequence_lsn;
+
+ /* do the actual work: call callback */
+ ctx->callbacks.sequence_cb(ctx, txn, sequence_lsn, rel, transactional,
+ created, last_value, log_cnt, is_called);
+
+ /* Pop the error context stack */
+ error_context_stack = errcallback.previous;
+}
+
static void
stream_start_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
XLogRecPtr first_lsn)
diff --git a/src/backend/replication/logical/proto.c b/src/backend/replication/logical/proto.c
index 1cf59e0fb0..f9e928f906 100644
--- a/src/backend/replication/logical/proto.c
+++ b/src/backend/replication/logical/proto.c
@@ -389,6 +389,58 @@ logicalrep_write_message(StringInfo out, TransactionId xid, XLogRecPtr lsn,
pq_sendbytes(out, message, sz);
}
+/*
+ * Write SEQUENCE to stream
+ */
+void
+logicalrep_write_sequence(StringInfo out, Relation rel, TransactionId xid,
+ XLogRecPtr lsn, bool transactional, bool created,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ uint8 flags = 0;
+ char *relname;
+
+ pq_sendbyte(out, LOGICAL_REP_MSG_SEQUENCE);
+
+ /* transaction ID (if not valid, we're not streaming) */
+ if (TransactionIdIsValid(xid))
+ pq_sendint32(out, xid);
+
+ pq_sendint8(out, flags);
+ pq_sendint64(out, lsn);
+
+ logicalrep_write_namespace(out, RelationGetNamespace(rel));
+ relname = RelationGetRelationName(rel);
+ pq_sendstring(out, relname);
+
+ pq_sendint8(out, transactional);
+ pq_sendint8(out, created);
+ pq_sendint64(out, last_value);
+ pq_sendint64(out, log_cnt);
+ pq_sendint8(out, is_called);
+}
+
+/*
+ * Read SEQUENCE from the stream.
+ */
+void
+logicalrep_read_sequence(StringInfo in, LogicalRepSequence *seqdata)
+{
+ /* XXX skipping flags and lsn */
+ pq_getmsgint(in, 1);
+ pq_getmsgint64(in);
+
+ /* Read relation name from stream */
+ seqdata->nspname = pstrdup(logicalrep_read_namespace(in));
+ seqdata->seqname = pstrdup(pq_getmsgstring(in));
+
+ seqdata->transactional = pq_getmsgint(in, 1);
+ seqdata->created = pq_getmsgint(in, 1);
+ seqdata->last_value = pq_getmsgint64(in);
+ seqdata->log_cnt = pq_getmsgint64(in);
+ seqdata->is_called = pq_getmsgint(in, 1);
+}
+
/*
* Write relation description to the output stream.
*/
diff --git a/src/backend/replication/logical/reorderbuffer.c b/src/backend/replication/logical/reorderbuffer.c
index 2d9e1279bb..94b0290d78 100644
--- a/src/backend/replication/logical/reorderbuffer.c
+++ b/src/backend/replication/logical/reorderbuffer.c
@@ -116,6 +116,13 @@ typedef struct ReorderBufferTXNByIdEnt
ReorderBufferTXN *txn;
} ReorderBufferTXNByIdEnt;
+/* entry for hash table we use to track sequences created in running xacts */
+typedef struct ReorderBufferSequenceEnt
+{
+ RelFileNode rnode;
+ TransactionId xid;
+} ReorderBufferSequenceEnt;
+
/* data structures for (relfilenode, ctid) => (cmin, cmax) mapping */
typedef struct ReorderBufferTupleCidKey
{
@@ -337,6 +344,14 @@ ReorderBufferAllocate(void)
buffer->by_txn = hash_create("ReorderBufferByXid", 1000, &hash_ctl,
HASH_ELEM | HASH_BLOBS | HASH_CONTEXT);
+ /* hash table of sequences, mapping relfilenode to XID of transaction */
+ hash_ctl.keysize = sizeof(RelFileNode);
+ hash_ctl.entrysize = sizeof(ReorderBufferSequenceEnt);
+ hash_ctl.hcxt = buffer->context;
+
+ buffer->sequences = hash_create("ReorderBufferSequenceHash", 1000, &hash_ctl,
+ HASH_ELEM | HASH_BLOBS | HASH_CONTEXT);
+
buffer->by_txn_last_xid = InvalidTransactionId;
buffer->by_txn_last_txn = NULL;
@@ -523,6 +538,13 @@ ReorderBufferReturnChange(ReorderBuffer *rb, ReorderBufferChange *change,
case REORDER_BUFFER_CHANGE_INTERNAL_COMMAND_ID:
case REORDER_BUFFER_CHANGE_INTERNAL_TUPLECID:
break;
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
+ if (change->data.sequence.tuple)
+ {
+ ReorderBufferReturnTupleBuf(rb, change->data.sequence.tuple);
+ change->data.sequence.tuple = NULL;
+ }
+ break;
}
pfree(change);
@@ -850,6 +872,212 @@ ReorderBufferQueueMessage(ReorderBuffer *rb, TransactionId xid,
}
}
+/*
+ * Treat the sequence increment as transactional?
+ */
+bool
+ReorderBufferSequenceIsTransactional(ReorderBuffer *rb,
+ RelFileNode rnode, bool created)
+{
+ bool found = false;
+
+ if (created)
+ return true;
+
+ hash_search(rb->sequences,
+ (void *) &rnode,
+ HASH_FIND,
+ &found);
+
+ return found;
+}
+
+
+/*
+ * A transactional sequence increment is queued to be processed upon commit
+ * and a non-transactional increment gets processed immediately.
+ *
+ * A sequence update may be both transactional and non-transactional. When
+ * created in a running transaction, treat it as transactional and queue
+ * the change in it. Otherwise treat it as non-transactional, so that we
+ * don't forget the increment in case of a rollback.
+ */
+void
+ReorderBufferQueueSequence(ReorderBuffer *rb, TransactionId xid,
+ Snapshot snapshot, XLogRecPtr lsn, RepOriginId origin_id,
+ RelFileNode rnode, bool transactional, bool created,
+ ReorderBufferTupleBuf *tuplebuf)
+{
+ /*
+ * Change needs to be handled as transactional, because the sequence was
+ * created in a transaction that is still running. In that case all the
+ * changes need to be queued in that transaction, we must not send them
+ * to the downstream until the transaction commits.
+ *
+ * There's a bit of a trouble with subtransactions - we can't queue it
+ * into the subxact, because it might be rolled back and we'd lose the
+ * increment. We need to queue it into the same (sub)xact that created
+ * the sequence, which is why we track the XID in the hash table.
+ */
+ if (transactional)
+ {
+ MemoryContext oldcontext;
+ ReorderBufferChange *change;
+
+ /* lookup sequence by relfilenode */
+ ReorderBufferSequenceEnt *ent;
+ bool found;
+
+ /* transactional changes require a transaction */
+ Assert(xid != InvalidTransactionId);
+
+ /* search the lookup table (we ignore the return value, found is enough) */
+ ent = hash_search(rb->sequences,
+ (void *) &rnode,
+ created ? HASH_ENTER : HASH_FIND,
+ &found);
+
+ /*
+ * If this is the "create" increment, we must not have found any
+ * pre-existing entry in the hash table (i.e. there must not be
+ * any conflicting sequence).
+ */
+ Assert(!(created && found));
+
+ /* But we must have either created or found an existing entry. */
+ Assert(created || found);
+
+ /*
+ * When creating the sequence, remember the XID of the transaction
+ * that created id.
+ */
+ if (created)
+ ent->xid = xid;
+
+ /* XXX Maybe check that we're still in the same top-level xact? */
+
+ /* OK, allocate and queue the change */
+ oldcontext = MemoryContextSwitchTo(rb->context);
+
+ change = ReorderBufferGetChange(rb);
+
+ change->action = REORDER_BUFFER_CHANGE_SEQUENCE;
+ change->origin_id = origin_id;
+
+ memcpy(&change->data.sequence.relnode, &rnode, sizeof(RelFileNode));
+
+ change->data.sequence.created = created;
+ change->data.sequence.tuple = tuplebuf;
+
+ /* add it to the same subxact that created the sequence */
+ ReorderBufferQueueChange(rb, ent->xid, lsn, change, false);
+
+ MemoryContextSwitchTo(oldcontext);
+ }
+ else
+ {
+ /*
+ * This increment is for a sequence that was not created in any
+ * running transaction, so we treat it as non-transactional and
+ * just send it to the output plugin directly.
+ */
+ ReorderBufferTXN *txn = NULL;
+ volatile Snapshot snapshot_now = snapshot;
+ bool using_subtxn;
+
+#ifdef USE_ASSERT_CHECKING
+ /* All "creates" have to be handled as transactional. */
+ Assert(!created);
+
+ /* Make sure the sequence is not in the hash table. */
+ {
+ bool found;
+ hash_search(rb->sequences,
+ (void *) &rnode,
+ HASH_FIND, &found);
+ Assert(!found);
+ }
+#endif
+
+ if (xid != InvalidTransactionId)
+ txn = ReorderBufferTXNByXid(rb, xid, true, NULL, lsn, true);
+
+ /* setup snapshot to allow catalog access */
+ SetupHistoricSnapshot(snapshot_now, NULL);
+
+ /*
+ * Decoding needs access to syscaches et al., which in turn use
+ * heavyweight locks and such. Thus we need to have enough state around to
+ * keep track of those. The easiest way is to simply use a transaction
+ * internally. That also allows us to easily enforce that nothing writes
+ * to the database by checking for xid assignments.
+ *
+ * When we're called via the SQL SRF there's already a transaction
+ * started, so start an explicit subtransaction there.
+ */
+ using_subtxn = IsTransactionOrTransactionBlock();
+
+ PG_TRY();
+ {
+ Relation relation;
+ HeapTuple tuple;
+ bool isnull;
+ int64 last_value, log_cnt;
+ bool is_called;
+ Oid reloid;
+
+ if (using_subtxn)
+ BeginInternalSubTransaction("sequence");
+ else
+ StartTransactionCommand();
+
+ reloid = RelidByRelfilenode(rnode.spcNode, rnode.relNode);
+
+ if (reloid == InvalidOid)
+ elog(ERROR, "could not map filenode \"%s\" to relation OID",
+ relpathperm(rnode,
+ MAIN_FORKNUM));
+
+ relation = RelationIdGetRelation(reloid);
+ tuple = &tuplebuf->tuple;
+
+ /*
+ * Extract the internal sequence values, describing the state.
+ *
+ * XXX Seems a bit strange to access it directly. Maybe there's
+ * a better / more correct way?
+ */
+ last_value = heap_getattr(tuple, 1, RelationGetDescr(relation), &isnull);
+ log_cnt = heap_getattr(tuple, 2, RelationGetDescr(relation), &isnull);
+ is_called = heap_getattr(tuple, 3, RelationGetDescr(relation), &isnull);
+
+ rb->sequence(rb, txn, lsn, relation, transactional, created,
+ last_value, log_cnt, is_called);
+
+ RelationClose(relation);
+
+ TeardownHistoricSnapshot(false);
+
+ AbortCurrentTransaction();
+
+ if (using_subtxn)
+ RollbackAndReleaseCurrentSubTransaction();
+ }
+ PG_CATCH();
+ {
+ TeardownHistoricSnapshot(true);
+
+ AbortCurrentTransaction();
+
+ if (using_subtxn)
+ RollbackAndReleaseCurrentSubTransaction();
+
+ PG_RE_THROW();
+ }
+ PG_END_TRY();
+ }
+}
+
/*
* AssertTXNLsnOrder
* Verify LSN ordering of transaction lists in the reorderbuffer
@@ -1526,6 +1754,31 @@ ReorderBufferCleanupTXN(ReorderBuffer *rb, ReorderBufferTXN *txn)
&found);
Assert(found);
+ /*
+ * Remove sequences created in this transaction (if any).
+ *
+ * There's no way to search by XID, so we simply do a seqscan of all
+ * the entries in the hash table. Hopefully there are only a couple
+ * entries in most cases - people generally don't create many new
+ * sequences over and over.
+ */
+ {
+ HASH_SEQ_STATUS scan_status;
+ ReorderBufferSequenceEnt *ent;
+
+ hash_seq_init(&scan_status, rb->sequences);
+ while ((ent = (ReorderBufferSequenceEnt *) hash_seq_search(&scan_status)) != NULL)
+ {
+ /* skip sequences not from this transaction */
+ if (ent->xid != txn->xid)
+ continue;
+
+ (void) hash_search(rb->sequences,
+ (void *) &(ent->rnode),
+ HASH_REMOVE, NULL);
+ }
+ }
+
/* remove entries spilled to disk */
if (rbtxn_is_serialized(txn))
ReorderBufferRestoreCleanup(rb, txn);
@@ -1941,6 +2194,39 @@ ReorderBufferApplyMessage(ReorderBuffer *rb, ReorderBufferTXN *txn,
change->data.msg.message);
}
+/*
+ * Helper function for ReorderBufferProcessTXN for applying sequences.
+ */
+static inline void
+ReorderBufferApplySequence(ReorderBuffer *rb, ReorderBufferTXN *txn,
+ Relation relation, ReorderBufferChange *change,
+ bool streaming)
+{
+ HeapTuple tuple;
+ bool isnull;
+ int64 last_value, log_cnt;
+ bool is_called;
+
+ /* FIXME support streaming */
+ Assert(!streaming);
+
+ tuple = &change->data.sequence.tuple->tuple;
+
+ /*
+ * Extract the internal sequence values, describing the state.
+ *
+ * XXX Seems a bit strange to access it directly. Maybe there's
+ * a better / more correct way?
+ */
+ last_value = heap_getattr(tuple, 1, RelationGetDescr(relation), &isnull);
+ log_cnt = heap_getattr(tuple, 2, RelationGetDescr(relation), &isnull);
+ is_called = heap_getattr(tuple, 3, RelationGetDescr(relation), &isnull);
+
+ rb->sequence(rb, txn, change->lsn, relation, true, /* gotta be transactional */
+ change->data.sequence.created,
+ last_value, log_cnt, is_called);
+}
+
/*
* Function to store the command id and snapshot at the end of the current
* stream so that we can reuse the same while sending the next stream.
@@ -2357,6 +2643,31 @@ ReorderBufferProcessTXN(ReorderBuffer *rb, ReorderBufferTXN *txn,
case REORDER_BUFFER_CHANGE_INTERNAL_TUPLECID:
elog(ERROR, "tuplecid value in changequeue");
break;
+
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
+ Assert(snapshot_now);
+
+ reloid = RelidByRelfilenode(change->data.sequence.relnode.spcNode,
+ change->data.sequence.relnode.relNode);
+
+ if (reloid == InvalidOid)
+ elog(ERROR, "could not map filenode \"%s\" to relation OID",
+ relpathperm(change->data.sequence.relnode,
+ MAIN_FORKNUM));
+
+ relation = RelationIdGetRelation(reloid);
+
+ if (!RelationIsValid(relation))
+ elog(ERROR, "could not open relation with OID %u (for filenode \"%s\")",
+ reloid,
+ relpathperm(change->data.sequence.relnode,
+ MAIN_FORKNUM));
+
+ if (RelationIsLogicallyLogged(relation))
+ ReorderBufferApplySequence(rb, txn, relation, change, streaming);
+
+ RelationClose(relation);
+ break;
}
}
@@ -3751,6 +4062,39 @@ ReorderBufferSerializeChange(ReorderBuffer *rb, ReorderBufferTXN *txn,
memcpy(data, change->data.truncate.relids, size);
data += size;
+ break;
+ }
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
+ {
+ char *data;
+ ReorderBufferTupleBuf *tup;
+ Size len = 0;
+
+ tup = change->data.sequence.tuple;
+
+ if (tup)
+ {
+ sz += sizeof(HeapTupleData);
+ len = tup->tuple.t_len;
+ sz += len;
+ }
+
+ /* make sure we have enough space */
+ ReorderBufferSerializeReserve(rb, sz);
+
+ data = ((char *) rb->outbuf) + sizeof(ReorderBufferDiskChange);
+ /* might have been reallocated above */
+ ondisk = (ReorderBufferDiskChange *) rb->outbuf;
+
+ if (len)
+ {
+ memcpy(data, &tup->tuple, sizeof(HeapTupleData));
+ data += sizeof(HeapTupleData);
+
+ memcpy(data, tup->tuple.t_data, len);
+ data += len;
+ }
+
break;
}
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM:
@@ -4014,6 +4358,22 @@ ReorderBufferChangeSize(ReorderBufferChange *change)
{
sz += sizeof(Oid) * change->data.truncate.nrelids;
+ break;
+ }
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
+ {
+ ReorderBufferTupleBuf *tup;
+ Size len = 0;
+
+ tup = change->data.sequence.tuple;
+
+ if (tup)
+ {
+ sz += sizeof(HeapTupleData);
+ len = tup->tuple.t_len;
+ sz += len;
+ }
+
break;
}
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM:
@@ -4314,6 +4674,30 @@ ReorderBufferRestoreChange(ReorderBuffer *rb, ReorderBufferTXN *txn,
break;
}
+
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
+ if (change->data.sequence.tuple)
+ {
+ uint32 tuplelen = ((HeapTuple) data)->t_len;
+
+ change->data.sequence.tuple =
+ ReorderBufferGetTupleBuf(rb, tuplelen - SizeofHeapTupleHeader);
+
+ /* restore ->tuple */
+ memcpy(&change->data.sequence.tuple->tuple, data,
+ sizeof(HeapTupleData));
+ data += sizeof(HeapTupleData);
+
+ /* reset t_data pointer into the new tuplebuf */
+ change->data.sequence.tuple->tuple.t_data =
+ ReorderBufferTupleBufData(change->data.sequence.tuple);
+
+ /* restore tuple data itself */
+ memcpy(change->data.sequence.tuple->tuple.t_data, data, tuplelen);
+ data += tuplelen;
+ }
+ break;
+
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM:
case REORDER_BUFFER_CHANGE_INTERNAL_COMMAND_ID:
case REORDER_BUFFER_CHANGE_INTERNAL_TUPLECID:
diff --git a/src/backend/replication/logical/tablesync.c b/src/backend/replication/logical/tablesync.c
index 67f907cdd9..d1ba2a9645 100644
--- a/src/backend/replication/logical/tablesync.c
+++ b/src/backend/replication/logical/tablesync.c
@@ -99,6 +99,7 @@
#include "catalog/pg_subscription_rel.h"
#include "catalog/pg_type.h"
#include "commands/copy.h"
+#include "commands/sequence.h"
#include "miscadmin.h"
#include "parser/parse_relation.h"
#include "pgstat.h"
@@ -353,6 +354,12 @@ process_syncing_tables_for_sync(XLogRecPtr current_lsn)
*
* If the synchronization position is reached (SYNCDONE), then the table can
* be marked as READY and is no longer tracked.
+ *
+ * XXX This needs to handle sequences too - after AlterSubscription_refresh
+ * starts caring about sequences, GetSubscriptionNotReadyRelations won't
+ * return just tables, and we'll have to sync them here. Not sure it's worth
+ * creating a new "sync" worker per sequence, maybe we should just sync them
+ * in the current process (it's pretty light-weight).
*/
static void
process_syncing_tables_for_apply(XLogRecPtr current_lsn)
@@ -870,6 +877,99 @@ copy_table(Relation rel)
logicalrep_rel_close(relmapentry, NoLock);
}
+
+
+/*
+ * FIXME add comment
+ */
+static void
+fetch_sequence_data(char *nspname, char *relname,
+ int64 *last_value, int64 *log_cnt, bool *is_called)
+{
+ WalRcvExecResult *res;
+ StringInfoData cmd;
+ TupleTableSlot *slot;
+ Oid tableRow[3] = {INT8OID, INT8OID, BOOLOID};
+
+ initStringInfo(&cmd);
+ appendStringInfo(&cmd, "SELECT last_value, log_cnt, is_called\n"
+ " FROM %s", quote_qualified_identifier(nspname, relname));
+
+ res = walrcv_exec(LogRepWorkerWalRcvConn, cmd.data, 3, tableRow);
+ pfree(cmd.data);
+
+ if (res->status != WALRCV_OK_TUPLES)
+ ereport(ERROR,
+ (errmsg("could not receive list of replicated tables from the publisher: %s",
+ res->err)));
+
+ /* Process the sequence. */
+ slot = MakeSingleTupleTableSlot(res->tupledesc, &TTSOpsMinimalTuple);
+ while (tuplestore_gettupleslot(res->tuplestore, true, false, slot))
+ {
+ bool isnull;
+
+ *last_value = DatumGetInt64(slot_getattr(slot, 1, &isnull));
+ Assert(!isnull);
+
+ *log_cnt = DatumGetInt64(slot_getattr(slot, 2, &isnull));
+ Assert(!isnull);
+
+ *is_called = DatumGetBool(slot_getattr(slot, 3, &isnull));
+ Assert(!isnull);
+
+ ExecClearTuple(slot);
+ }
+ ExecDropSingleTupleTableSlot(slot);
+
+ walrcv_clear_result(res);
+}
+
+/*
+ * Copy existing data of a sequence from publisher.
+ *
+ * Caller is responsible for locking the local relation.
+ */
+static void
+copy_sequence(Relation rel)
+{
+ LogicalRepRelMapEntry *relmapentry;
+ LogicalRepRelation lrel;
+ StringInfoData cmd;
+ int64 last_value = 0,
+ log_cnt = 0;
+ bool is_called = 0;
+
+ /* Get the publisher relation info. */
+ fetch_remote_table_info(get_namespace_name(RelationGetNamespace(rel)),
+ RelationGetRelationName(rel), &lrel);
+
+ /* Put the relation into relmap. */
+ logicalrep_relmap_update(&lrel);
+
+ /* Map the publisher relation to local one. */
+ relmapentry = logicalrep_rel_open(lrel.remoteid, NoLock);
+ Assert(rel == relmapentry->localrel);
+
+ /* Start copy on the publisher. */
+ initStringInfo(&cmd);
+
+ Assert(lrel.relkind == RELKIND_SEQUENCE);
+
+ fetch_sequence_data(lrel.nspname, lrel.relname, &last_value, &log_cnt, &is_called);
+
+ elog(WARNING, "sequence %s info last_value %ld log_cnt %ld is_called %d",
+ quote_qualified_identifier(lrel.nspname, lrel.relname),
+ last_value, log_cnt, is_called);
+
+ ResetSequence2(RelationGetRelid(rel), last_value, log_cnt, is_called);
+
+ logicalrep_rel_close(relmapentry, NoLock);
+}
+
+
+
+
/*
* Determine the tablesync slot name.
*
@@ -1102,10 +1202,20 @@ LogicalRepSyncTableStart(XLogRecPtr *origin_startpos)
originname)));
}
- /* Now do the initial data copy */
- PushActiveSnapshot(GetTransactionSnapshot());
- copy_table(rel);
- PopActiveSnapshot();
+ if (get_rel_relkind(RelationGetRelid(rel)) == RELKIND_SEQUENCE)
+ {
+ /* Now do the initial sequence copy */
+ PushActiveSnapshot(GetTransactionSnapshot());
+ copy_sequence(rel);
+ PopActiveSnapshot();
+ }
+ else
+ {
+ /* Now do the initial data copy */
+ PushActiveSnapshot(GetTransactionSnapshot());
+ copy_table(rel);
+ PopActiveSnapshot();
+ }
res = walrcv_exec(LogRepWorkerWalRcvConn, "COMMIT", 0, NULL);
if (res->status != WALRCV_OK_COMMAND)
diff --git a/src/backend/replication/logical/worker.c b/src/backend/replication/logical/worker.c
index 6ba447ea97..1e85a65890 100644
--- a/src/backend/replication/logical/worker.c
+++ b/src/backend/replication/logical/worker.c
@@ -70,6 +70,7 @@
#include "catalog/pg_tablespace.h"
#include "commands/tablecmds.h"
#include "commands/tablespace.h"
+#include "commands/sequence.h"
#include "commands/trigger.h"
#include "executor/executor.h"
#include "executor/execPartition.h"
@@ -813,6 +814,36 @@ apply_handle_origin(StringInfo s)
errmsg("ORIGIN message sent out of order")));
}
+/*
+ * Handle SEQUENCE message.
+ */
+static void
+apply_handle_sequence(StringInfo s)
+{
+ LogicalRepSequence seq;
+ Oid relid;
+
+ // FIXME
+ // if (handle_streamed_transaction(LOGICAL_REP_MSG_SEQUENCE, s))
+ // return;
+ ensure_transaction();
+
+ logicalrep_read_sequence(s, &seq);
+
+ relid = RangeVarGetRelid(makeRangeVar(seq.nspname,
+ seq.seqname, -1),
+ RowExclusiveLock, false);
+
+ elog(WARNING, "applying sequence transactional %d created %d last_value %ld log_cnt %ld is_called %d",
+ seq.transactional, seq.created, seq.last_value, seq.log_cnt, seq.is_called);
+
+ /* apply the sequence change */
+ ResetSequence2(relid, seq.last_value, seq.log_cnt, seq.is_called);
+
+ /* Commit the per-stream transaction */
+ CommitTransactionCommand();
+}
+
/*
* Handle STREAM START message.
*/
@@ -2017,6 +2048,10 @@ apply_dispatch(StringInfo s)
*/
return;
+ case LOGICAL_REP_MSG_SEQUENCE:
+ apply_handle_sequence(s);
+ return;
+
case LOGICAL_REP_MSG_STREAM_START:
apply_handle_stream_start(s);
return;
diff --git a/src/backend/replication/pgoutput/pgoutput.c b/src/backend/replication/pgoutput/pgoutput.c
index fe12d08a94..37aeb42d22 100644
--- a/src/backend/replication/pgoutput/pgoutput.c
+++ b/src/backend/replication/pgoutput/pgoutput.c
@@ -49,6 +49,10 @@ static void pgoutput_message(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn, XLogRecPtr message_lsn,
bool transactional, const char *prefix,
Size sz, const char *message);
+static void pgoutput_sequence(LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn, XLogRecPtr sequence_lsn,
+ Relation rel, bool transactional, bool created,
+ int64 last_value, int64 log_cnt, bool is_called);
static bool pgoutput_origin_filter(LogicalDecodingContext *ctx,
RepOriginId origin_id);
static void pgoutput_stream_start(struct LogicalDecodingContext *ctx,
@@ -144,6 +148,7 @@ _PG_output_plugin_init(OutputPluginCallbacks *cb)
cb->change_cb = pgoutput_change;
cb->truncate_cb = pgoutput_truncate;
cb->message_cb = pgoutput_message;
+ cb->sequence_cb = pgoutput_sequence;
cb->commit_cb = pgoutput_commit_txn;
cb->filter_by_origin_cb = pgoutput_origin_filter;
cb->shutdown_cb = pgoutput_shutdown;
@@ -166,11 +171,13 @@ parse_output_parameters(List *options, PGOutputData *data)
bool publication_names_given = false;
bool binary_option_given = false;
bool messages_option_given = false;
+ bool sequences_option_given = false;
bool streaming_given = false;
data->binary = false;
data->streaming = false;
data->messages = false;
+ data->sequences = true;
foreach(lc, options)
{
@@ -236,6 +243,16 @@ parse_output_parameters(List *options, PGOutputData *data)
data->messages = defGetBoolean(defel);
}
+ else if (strcmp(defel->defname, "sequences") == 0)
+ {
+ if (sequences_option_given)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("conflicting or redundant options")));
+ sequences_option_given = true;
+
+ data->sequences = defGetBoolean(defel);
+ }
else if (strcmp(defel->defname, "streaming") == 0)
{
if (streaming_given)
@@ -753,6 +770,47 @@ pgoutput_message(LogicalDecodingContext *ctx, ReorderBufferTXN *txn,
OutputPluginWrite(ctx, true);
}
+static void
+pgoutput_sequence(LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn, XLogRecPtr sequence_lsn,
+ Relation rel, bool transactional, bool created,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ PGOutputData *data = (PGOutputData *) ctx->output_plugin_private;
+ TransactionId xid = InvalidTransactionId;
+ RelationSyncEntry *relentry;
+
+ if (!data->sequences)
+ return;
+
+ if (!is_publishable_relation(rel))
+ return;
+
+ /* XXX handle (in_streaming) here */
+
+ relentry = get_rel_sync_entry(data, RelationGetRelid(rel));
+
+ /*
+ * First check the sequence filter.
+ *
+ * We handle just REORDER_BUFFER_CHANGE_SEQUENCE here.
+ */
+ if (!relentry->pubactions.pubsequence)
+ return;
+
+ OutputPluginPrepareWrite(ctx, true);
+ logicalrep_write_sequence(ctx->out,
+ rel,
+ xid,
+ sequence_lsn,
+ transactional,
+ created,
+ last_value,
+ log_cnt,
+ is_called);
+ OutputPluginWrite(ctx, true);
+}
+
/*
* Currently we always forward.
*/
@@ -1026,7 +1084,8 @@ get_rel_sync_entry(PGOutputData *data, Oid relid)
entry->streamed_txns = NIL;
entry->replicate_valid = false;
entry->pubactions.pubinsert = entry->pubactions.pubupdate =
- entry->pubactions.pubdelete = entry->pubactions.pubtruncate = false;
+ entry->pubactions.pubdelete = entry->pubactions.pubtruncate =
+ entry->pubactions.pubsequence = false;
entry->publish_as_relid = InvalidOid;
entry->map = NULL; /* will be set by maybe_send_schema() if needed */
}
@@ -1037,6 +1096,7 @@ get_rel_sync_entry(PGOutputData *data, Oid relid)
List *pubids = GetRelationPublications(relid);
ListCell *lc;
Oid publish_as_relid = relid;
+ bool is_sequence = (get_rel_relkind(relid) == RELKIND_SEQUENCE);
/* Reload publications if needed before use. */
if (!publications_valid)
@@ -1060,12 +1120,23 @@ get_rel_sync_entry(PGOutputData *data, Oid relid)
Publication *pub = lfirst(lc);
bool publish = false;
- if (pub->alltables)
+ if (pub->alltables && (!is_sequence))
{
publish = true;
if (pub->pubviaroot && am_partition)
publish_as_relid = llast_oid(get_partition_ancestors(relid));
}
+ else if (pub->allsequences && is_sequence)
+ {
+ publish = true;
+ }
+
+ /* if a sequence, just cross-check the list of publications */
+ if (!publish && is_sequence)
+ {
+ if (list_member_oid(pubids, pub->oid))
+ publish = true;
+ }
if (!publish)
{
@@ -1116,10 +1187,12 @@ get_rel_sync_entry(PGOutputData *data, Oid relid)
entry->pubactions.pubupdate |= pub->pubactions.pubupdate;
entry->pubactions.pubdelete |= pub->pubactions.pubdelete;
entry->pubactions.pubtruncate |= pub->pubactions.pubtruncate;
+ entry->pubactions.pubsequence |= pub->pubactions.pubsequence;
}
if (entry->pubactions.pubinsert && entry->pubactions.pubupdate &&
- entry->pubactions.pubdelete && entry->pubactions.pubtruncate)
+ entry->pubactions.pubdelete && entry->pubactions.pubtruncate &&
+ entry->pubactions.pubsequence)
break;
}
@@ -1264,5 +1337,6 @@ rel_sync_cache_publication_cb(Datum arg, int cacheid, uint32 hashvalue)
entry->pubactions.pubupdate = false;
entry->pubactions.pubdelete = false;
entry->pubactions.pubtruncate = false;
+ entry->pubactions.pubsequence = false;
}
}
diff --git a/src/backend/utils/cache/relcache.c b/src/backend/utils/cache/relcache.c
index fd05615e76..890829bc0c 100644
--- a/src/backend/utils/cache/relcache.c
+++ b/src/backend/utils/cache/relcache.c
@@ -5487,6 +5487,7 @@ GetRelationPublicationActions(Relation relation)
pubactions->pubupdate |= pubform->pubupdate;
pubactions->pubdelete |= pubform->pubdelete;
pubactions->pubtruncate |= pubform->pubtruncate;
+ pubactions->pubsequence |= pubform->pubsequence;
ReleaseSysCache(tup);
@@ -5495,7 +5496,8 @@ GetRelationPublicationActions(Relation relation)
* other publications.
*/
if (pubactions->pubinsert && pubactions->pubupdate &&
- pubactions->pubdelete && pubactions->pubtruncate)
+ pubactions->pubdelete && pubactions->pubtruncate &&
+ pubactions->pubsequence)
break;
}
diff --git a/src/bin/psql/tab-complete.c b/src/bin/psql/tab-complete.c
index 32c1bdfdca..147745d94c 100644
--- a/src/bin/psql/tab-complete.c
+++ b/src/bin/psql/tab-complete.c
@@ -1640,7 +1640,7 @@ psql_completion(const char *text, int start, int end)
/* ALTER PUBLICATION <name> */
else if (Matches("ALTER", "PUBLICATION", MatchAny))
- COMPLETE_WITH("ADD TABLE", "DROP TABLE", "OWNER TO", "RENAME TO", "SET");
+ COMPLETE_WITH("ADD TABLE", "DROP TABLE", "ADD SEQUENCE", "DROP SEQUENCE", "OWNER TO", "RENAME TO", "SET");
/* ALTER PUBLICATION <name> SET */
else if (Matches("ALTER", "PUBLICATION", MatchAny, "SET"))
COMPLETE_WITH("(", "TABLE");
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index acbcae4607..e3043e0f63 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -11442,6 +11442,11 @@
provolatile => 's', prorettype => 'oid', proargtypes => 'text',
proallargtypes => '{text,oid}', proargmodes => '{i,o}',
proargnames => '{pubname,relid}', prosrc => 'pg_get_publication_tables' },
+{ oid => '8000', descr => 'get OIDs of sequences in a publication',
+ proname => 'pg_get_publication_sequences', prorows => '1000', proretset => 't',
+ provolatile => 's', prorettype => 'oid', proargtypes => 'text',
+ proallargtypes => '{text,oid}', proargmodes => '{i,o}',
+ proargnames => '{pubname,relid}', prosrc => 'pg_get_publication_sequences' },
{ oid => '6121',
descr => 'returns whether a relation can be part of a publication',
proname => 'pg_relation_is_publishable', provolatile => 's',
diff --git a/src/include/catalog/pg_publication.h b/src/include/catalog/pg_publication.h
index 1b31fee9e3..d1c9a18b05 100644
--- a/src/include/catalog/pg_publication.h
+++ b/src/include/catalog/pg_publication.h
@@ -40,6 +40,12 @@ CATALOG(pg_publication,6104,PublicationRelationId)
*/
bool puballtables;
+ /*
+ * indicates that this is special publication which should encompass all
+ * sequences in the database (except for the unlogged and temp ones)
+ */
+ bool puballsequences;
+
/* true if inserts are published */
bool pubinsert;
@@ -52,6 +58,9 @@ CATALOG(pg_publication,6104,PublicationRelationId)
/* true if truncates are published */
bool pubtruncate;
+ /* true if sequences are published */
+ bool pubsequence;
+
/* true if partition changes are published using root schema */
bool pubviaroot;
} FormData_pg_publication;
@@ -74,6 +83,7 @@ typedef struct PublicationActions
bool pubupdate;
bool pubdelete;
bool pubtruncate;
+ bool pubsequence;
} PublicationActions;
typedef struct Publication
@@ -81,6 +91,7 @@ typedef struct Publication
Oid oid;
char *name;
bool alltables;
+ bool allsequences;
bool pubviaroot;
PublicationActions pubactions;
} Publication;
@@ -109,6 +120,9 @@ extern List *GetPublicationRelations(Oid pubid, PublicationPartOpt pub_partopt);
extern List *GetAllTablesPublications(void);
extern List *GetAllTablesPublicationRelations(bool pubviaroot);
+extern List *GetAllSequencesPublicationRelations(void);
+extern List *GetPublicationSequenceRelations(Oid pubid);
+
extern bool is_publishable_relation(Relation rel);
extern ObjectAddress publication_add_relation(Oid pubid, Relation targetrel,
bool if_not_exists);
diff --git a/src/include/commands/sequence.h b/src/include/commands/sequence.h
index 40544dd4c7..c28e8695cb 100644
--- a/src/include/commands/sequence.h
+++ b/src/include/commands/sequence.h
@@ -48,6 +48,7 @@ typedef FormData_pg_sequence_data *Form_pg_sequence_data;
typedef struct xl_seq_rec
{
RelFileNode node;
+ bool created; /* is this a CREATE SEQUENCE */
/* SEQUENCE TUPLE DATA FOLLOWS AT THE END */
} xl_seq_rec;
@@ -59,6 +60,7 @@ extern ObjectAddress DefineSequence(ParseState *pstate, CreateSeqStmt *stmt);
extern ObjectAddress AlterSequence(ParseState *pstate, AlterSeqStmt *stmt);
extern void DeleteSequenceTuple(Oid relid);
extern void ResetSequence(Oid seq_relid);
+extern void ResetSequence2(Oid seq_relid, int64 last_value, int64 log_cnt, bool is_called);
extern void ResetSequenceCaches(void);
extern void seq_redo(XLogReaderState *rptr);
diff --git a/src/include/nodes/parsenodes.h b/src/include/nodes/parsenodes.h
index ef73342019..b45c103b6c 100644
--- a/src/include/nodes/parsenodes.h
+++ b/src/include/nodes/parsenodes.h
@@ -3624,8 +3624,10 @@ typedef struct AlterPublicationStmt
/* parameters used for ALTER PUBLICATION ... ADD/DROP TABLE */
List *tables; /* List of tables to add/drop */
+ List *sequences; /* List of sequences to add/drop */
bool for_all_tables; /* Special publication for all tables in db */
- DefElemAction tableAction; /* What action to perform with the tables */
+ bool for_all_sequences; /* Special publication for all tables in db */
+ DefElemAction action; /* What action to perform with the tables/sequences */
} AlterPublicationStmt;
typedef struct CreateSubscriptionStmt
diff --git a/src/include/replication/logicalproto.h b/src/include/replication/logicalproto.h
index 55b90c03ea..f3a7d18fca 100644
--- a/src/include/replication/logicalproto.h
+++ b/src/include/replication/logicalproto.h
@@ -55,6 +55,7 @@ typedef enum LogicalRepMsgType
LOGICAL_REP_MSG_RELATION = 'R',
LOGICAL_REP_MSG_TYPE = 'Y',
LOGICAL_REP_MSG_MESSAGE = 'M',
+ LOGICAL_REP_MSG_SEQUENCE = 'X', /* FIXME change */
LOGICAL_REP_MSG_STREAM_START = 'S',
LOGICAL_REP_MSG_STREAM_END = 'E',
LOGICAL_REP_MSG_STREAM_COMMIT = 'c',
@@ -107,6 +108,19 @@ typedef struct LogicalRepTyp
char *typname; /* name of the remote type */
} LogicalRepTyp;
+/* Sequence info */
+typedef struct LogicalRepSequence
+{
+ Oid remoteid; /* unique id of the remote sequence */
+ char *nspname; /* schema name of remote sequence */
+ char *seqname; /* name of the remote sequence */
+ bool transactional;
+ bool created;
+ int64 last_value;
+ int64 log_cnt; /* XXX probably not needed? */
+ bool is_called; /* XXX probably not needed? */
+} LogicalRepSequence;
+
/* Transaction info */
typedef struct LogicalRepBeginData
{
@@ -154,6 +168,12 @@ extern List *logicalrep_read_truncate(StringInfo in,
bool *cascade, bool *restart_seqs);
extern void logicalrep_write_message(StringInfo out, TransactionId xid, XLogRecPtr lsn,
bool transactional, const char *prefix, Size sz, const char *message);
+extern void logicalrep_write_sequence(StringInfo out, Relation rel,
+ TransactionId xid, XLogRecPtr lsn,
+ bool transactional, bool created,
+ int64 last_value, int64 log_cnt,
+ bool is_called);
+extern void logicalrep_read_sequence(StringInfo in, LogicalRepSequence *seqdata);
extern void logicalrep_write_rel(StringInfo out, TransactionId xid,
Relation rel);
extern LogicalRepRelation *logicalrep_read_rel(StringInfo in);
diff --git a/src/include/replication/output_plugin.h b/src/include/replication/output_plugin.h
index 810495ed0e..e3e1b03f32 100644
--- a/src/include/replication/output_plugin.h
+++ b/src/include/replication/output_plugin.h
@@ -88,6 +88,19 @@ typedef void (*LogicalDecodeMessageCB) (struct LogicalDecodingContext *ctx,
Size message_size,
const char *message);
+/*
+ * Called for the generic logical decoding sequences.
+ */
+typedef void (*LogicalDecodeSequenceCB) (struct LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional,
+ bool created,
+ int64 last_value,
+ int64 log_cnt,
+ bool is_called);
+
/*
* Filter changes by origin.
*/
@@ -219,6 +232,7 @@ typedef struct OutputPluginCallbacks
LogicalDecodeTruncateCB truncate_cb;
LogicalDecodeCommitCB commit_cb;
LogicalDecodeMessageCB message_cb;
+ LogicalDecodeSequenceCB sequence_cb;
LogicalDecodeFilterByOriginCB filter_by_origin_cb;
LogicalDecodeShutdownCB shutdown_cb;
diff --git a/src/include/replication/pgoutput.h b/src/include/replication/pgoutput.h
index 51e7c0348d..a5aec2928a 100644
--- a/src/include/replication/pgoutput.h
+++ b/src/include/replication/pgoutput.h
@@ -27,6 +27,7 @@ typedef struct PGOutputData
bool binary;
bool streaming;
bool messages;
+ bool sequences;
} PGOutputData;
#endif /* PGOUTPUT_H */
diff --git a/src/include/replication/reorderbuffer.h b/src/include/replication/reorderbuffer.h
index 0c6e9d1cb9..5190cdf196 100644
--- a/src/include/replication/reorderbuffer.h
+++ b/src/include/replication/reorderbuffer.h
@@ -63,7 +63,8 @@ enum ReorderBufferChangeType
REORDER_BUFFER_CHANGE_INTERNAL_TUPLECID,
REORDER_BUFFER_CHANGE_INTERNAL_SPEC_INSERT,
REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM,
- REORDER_BUFFER_CHANGE_TRUNCATE
+ REORDER_BUFFER_CHANGE_TRUNCATE,
+ REORDER_BUFFER_CHANGE_SEQUENCE,
};
/* forward declaration */
@@ -157,6 +158,14 @@ typedef struct ReorderBufferChange
uint32 ninvalidations; /* Number of messages */
SharedInvalidationMessage *invalidations; /* invalidation message */
} inval;
+
+ /* Context data for Sequence changes */
+ struct
+ {
+ RelFileNode relnode;
+ bool created;
+ ReorderBufferTupleBuf *tuple;
+ } sequence;
} data;
/*
@@ -425,6 +434,15 @@ typedef void (*ReorderBufferMessageCB) (ReorderBuffer *rb,
const char *prefix, Size sz,
const char *message);
+/* sequence callback signature */
+typedef void (*ReorderBufferSequenceCB) (ReorderBuffer *rb,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional, bool created,
+ int64 last_value, int64 log_cnt,
+ bool is_called);
+
/* begin prepare callback signature */
typedef void (*ReorderBufferBeginPrepareCB) (ReorderBuffer *rb,
ReorderBufferTXN *txn);
@@ -506,6 +524,12 @@ struct ReorderBuffer
*/
HTAB *by_txn;
+ /*
+ * relfilenode => XID lookup table for sequences created in a transaction
+ * (also includes altered sequences, which assigns new relfilenode)
+ */
+ HTAB *sequences;
+
/*
* Transactions that could be a toplevel xact, ordered by LSN of the first
* record bearing that xid.
@@ -536,6 +560,7 @@ struct ReorderBuffer
ReorderBufferApplyTruncateCB apply_truncate;
ReorderBufferCommitCB commit;
ReorderBufferMessageCB message;
+ ReorderBufferSequenceCB sequence;
/*
* Callbacks to be called when streaming a transaction at prepare time.
@@ -630,6 +655,10 @@ void ReorderBufferQueueChange(ReorderBuffer *, TransactionId,
void ReorderBufferQueueMessage(ReorderBuffer *, TransactionId, Snapshot snapshot, XLogRecPtr lsn,
bool transactional, const char *prefix,
Size message_size, const char *message);
+void ReorderBufferQueueSequence(ReorderBuffer *rb, TransactionId xid,
+ Snapshot snapshot, XLogRecPtr lsn, RepOriginId origin_id,
+ RelFileNode rnode, bool transactional, bool created,
+ ReorderBufferTupleBuf *tuplebuf);
void ReorderBufferCommit(ReorderBuffer *, TransactionId,
XLogRecPtr commit_lsn, XLogRecPtr end_lsn,
TimestampTz commit_time, RepOriginId origin_id, XLogRecPtr origin_lsn);
@@ -677,4 +706,7 @@ void ReorderBufferSetRestartPoint(ReorderBuffer *, XLogRecPtr ptr);
void StartupReorderBuffer(void);
+bool ReorderBufferSequenceIsTransactional(ReorderBuffer *rb,
+ RelFileNode rnode, bool created);
+
#endif
diff --git a/src/test/regress/expected/publication.out b/src/test/regress/expected/publication.out
index 63d6ab7a4e..2cbad49f2e 100644
--- a/src/test/regress/expected/publication.out
+++ b/src/test/regress/expected/publication.out
@@ -158,8 +158,8 @@ DROP TABLE testpub_parted1;
DROP PUBLICATION testpub_forparted, testpub_forparted1;
-- fail - view
CREATE PUBLICATION testpub_fortbl FOR TABLE testpub_view;
-ERROR: "testpub_view" is not a table
-DETAIL: Only tables can be added to publications.
+ERROR: "testpub_view" is not a table or sequence
+DETAIL: Only tables and sequences can be added to publications.
SET client_min_messages = 'ERROR';
CREATE PUBLICATION testpub_fortbl FOR TABLE testpub_tbl1, pub_test.testpub_nopk;
RESET client_min_messages;
@@ -180,8 +180,8 @@ Tables:
-- fail - view
ALTER PUBLICATION testpub_default ADD TABLE testpub_view;
-ERROR: "testpub_view" is not a table
-DETAIL: Only tables can be added to publications.
+ERROR: "testpub_view" is not a table or sequence
+DETAIL: Only tables and sequences can be added to publications.
ALTER PUBLICATION testpub_default ADD TABLE testpub_tbl1;
ALTER PUBLICATION testpub_default SET TABLE testpub_tbl1;
ALTER PUBLICATION testpub_default ADD TABLE pub_test.testpub_nopk;
diff --git a/src/test/regress/expected/rules.out b/src/test/regress/expected/rules.out
index e5ab11275d..393be871be 100644
--- a/src/test/regress/expected/rules.out
+++ b/src/test/regress/expected/rules.out
@@ -1451,6 +1451,14 @@ pg_prepared_xacts| SELECT p.transaction,
FROM ((pg_prepared_xact() p(transaction, gid, prepared, ownerid, dbid)
LEFT JOIN pg_authid u ON ((p.ownerid = u.oid)))
LEFT JOIN pg_database d ON ((p.dbid = d.oid)));
+pg_publication_sequences| SELECT p.pubname,
+ n.nspname AS schemaname,
+ c.relname AS sequencename
+ FROM pg_publication p,
+ LATERAL pg_get_publication_sequences((p.pubname)::text) gpt(relid),
+ (pg_class c
+ JOIN pg_namespace n ON ((n.oid = c.relnamespace)))
+ WHERE (c.oid = gpt.relid);
pg_publication_tables| SELECT p.pubname,
n.nspname AS schemaname,
c.relname AS tablename
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2021-06-13 21:15 Tomas Vondra <[email protected]>
parent: Tomas Vondra <[email protected]>
2 siblings, 1 reply; 127+ messages in thread
From: Tomas Vondra @ 2021-06-13 21:15 UTC (permalink / raw)
To: PostgreSQL Hackers <[email protected]>; +Cc: Petr Jelinek <[email protected]>
Hi,
Seems the cfbot was not entirely happy with the patch on some platforms,
so here's a fixed version. There was a bogus call to ensure_transaction
function (which does not exist at all) and a silly bug in transaction
management in apply_handle_sequence.
regards
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2021-06-23 14:14 Tomas Vondra <[email protected]>
parent: Tomas Vondra <[email protected]>
0 siblings, 1 reply; 127+ messages in thread
From: Tomas Vondra @ 2021-06-23 14:14 UTC (permalink / raw)
To: PostgreSQL Hackers <[email protected]>; +Cc: Petr Jelinek <[email protected]>
A rebased patch, addressing a minor bitrot due to 4daa140a2f5.
regards
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2021-06-23 14:25 Tomas Vondra <[email protected]>
parent: Tomas Vondra <[email protected]>
0 siblings, 1 reply; 127+ messages in thread
From: Tomas Vondra @ 2021-06-23 14:25 UTC (permalink / raw)
To: PostgreSQL Hackers <[email protected]>; +Cc: Petr Jelinek <[email protected]>
On 6/23/21 4:14 PM, Tomas Vondra wrote:
> A rebased patch, addressing a minor bitrot due to 4daa140a2f5.
>
Meh, forgot to attach the patch as usual, of course ...
regards
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2021-07-15 12:17 vignesh C <[email protected]>
parent: Tomas Vondra <[email protected]>
0 siblings, 1 reply; 127+ messages in thread
From: vignesh C @ 2021-07-15 12:17 UTC (permalink / raw)
To: Tomas Vondra <[email protected]>; +Cc: PostgreSQL Hackers <[email protected]>; Petr Jelinek <[email protected]>
On Wed, Jun 23, 2021 at 7:55 PM Tomas Vondra
<[email protected]> wrote:
>
> On 6/23/21 4:14 PM, Tomas Vondra wrote:
> > A rebased patch, addressing a minor bitrot due to 4daa140a2f5.
> >
>
> Meh, forgot to attach the patch as usual, of course ...
The patch does not apply on Head anymore, could you rebase and post a
patch. I'm changing the status to "Waiting for Author".
Regards,
Vignesh
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2021-07-20 15:30 Peter Eisentraut <[email protected]>
parent: Tomas Vondra <[email protected]>
2 siblings, 1 reply; 127+ messages in thread
From: Peter Eisentraut @ 2021-07-20 15:30 UTC (permalink / raw)
To: Tomas Vondra <[email protected]>; PostgreSQL Hackers <[email protected]>; +Cc: Petr Jelinek <[email protected]>
On 08.06.21 00:28, Tomas Vondra wrote:
>
> new "sequence" publication action
> ---------------------------------
>
> The publications now have a new "sequence" publication action, which is
> enabled by default. This determines whether the publication decodes
> sequences or what.
>
>
> FOR ALL SEQUENCES
> -----------------
>
> It should be possible to create FOR ALL SEQUENCES publications, just
> like we have FOR ALL TABLES. But this produces shift/reduce conflicts
> in the grammar, and I didn't bother dealing with that. So for now it's
> required to do ALTER PUBLICATION ... [ADD | DROP] SEQUENCE ...
I have been thinking about these DDL-level issues a bit. The most
common use case will be to have a bunch of tables with implicit
sequences, and you just want to replicate them from here to there
without too much effort. So ideally an implicit sequence should be
replicated by default if the table is part of a publication (unless
sequences are turned off by the publication option).
We already have support for things like that in
GetPublicationRelations(), where a partitioned table is expanded to
include the actual partitions. I think that logic could be reused. So
in general I would have GetPublicationRelations() include sequences and
don't have GetPublicationSequenceRelations() at all. Then sequences
could also be sent by pg_publication_tables(), maybe add a relkind
column. And then you also don't need so much duplicate DDL code, if you
just consider everything as a relation. For example, there doesn't seem
to be an actual need to have fetch_sequence_list() and subsequent
processing on the subscriber side. It does the same thing as
fetch_table_list(), so it might as well just all be one thing.
We do, however, probably need some checking that we don't replicate
tables to sequences or vice versa.
We probably also don't need a separate FOR ALL SEQUENCES option. What
users really want is a "for everything" option. We could think about
renaming or alternative syntax, but in principle I think FOR ALL TABLES
should include sequences by default.
Tests under src/test/subscription/ are needed.
I'm not sure why test_decoding needs a skip-sequences option. The
source code says it's for backward compatibility, but I don't see why we
need that.
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2021-07-20 20:27 Tomas Vondra <[email protected]>
parent: Peter Eisentraut <[email protected]>
0 siblings, 0 replies; 127+ messages in thread
From: Tomas Vondra @ 2021-07-20 20:27 UTC (permalink / raw)
To: Peter Eisentraut <[email protected]>; PostgreSQL Hackers <[email protected]>; +Cc: Petr Jelinek <[email protected]>
On 7/20/21 5:30 PM, Peter Eisentraut wrote:
> On 08.06.21 00:28, Tomas Vondra wrote:
>>
>> new "sequence" publication action
>> ---------------------------------
>>
>> The publications now have a new "sequence" publication action, which is
>> enabled by default. This determines whether the publication decodes
>> sequences or what.
>>
>>
>> FOR ALL SEQUENCES
>> -----------------
>>
>> It should be possible to create FOR ALL SEQUENCES publications, just
>> like we have FOR ALL TABLES. But this produces shift/reduce conflicts
>> in the grammar, and I didn't bother dealing with that. So for now it's
>> required to do ALTER PUBLICATION ... [ADD | DROP] SEQUENCE ...
>
> I have been thinking about these DDL-level issues a bit. The most
> common use case will be to have a bunch of tables with implicit
> sequences, and you just want to replicate them from here to there
> without too much effort. So ideally an implicit sequence should be
> replicated by default if the table is part of a publication (unless
> sequences are turned off by the publication option).
>
Agreed, that seems like a reasonable approach.
> We already have support for things like that in
> GetPublicationRelations(), where a partitioned table is expanded to
> include the actual partitions. I think that logic could be reused. So
> in general I would have GetPublicationRelations() include sequences and
> don't have GetPublicationSequenceRelations() at all. Then sequences
> could also be sent by pg_publication_tables(), maybe add a relkind
> column. And then you also don't need so much duplicate DDL code, if you
> just consider everything as a relation. For example, there doesn't seem
> to be an actual need to have fetch_sequence_list() and subsequent
> processing on the subscriber side. It does the same thing as
> fetch_table_list(), so it might as well just all be one thing.
>
Not sure. I agree with replicating implicit sequences by default,
without having to add them to the publication. But I think we should
allow adding other sequences too, and I think some of this code and
differentiation from tables will be needed.
FWIW I'm not claiming there are no duplicate parts - I've mostly
copy-pasted the table-handling code for sequences, and I'll look into
reusing some of it.
> We do, however, probably need some checking that we don't replicate
> tables to sequences or vice versa.
>
True. I haven't tried doing such silly things yet.
> We probably also don't need a separate FOR ALL SEQUENCES option. What
> users really want is a "for everything" option. We could think about
> renaming or alternative syntax, but in principle I think FOR ALL TABLES
> should include sequences by default.
>
> Tests under src/test/subscription/ are needed.
>
Yeah, true. At the moment there are just tests in test_decoding, mostly
because the previous patch versions did not add support for sequences to
built-in replication. Will fix.
> I'm not sure why test_decoding needs a skip-sequences option. The
> source code says it's for backward compatibility, but I don't see why we
> need that.
Hmmm, I'm a bit baffled by skip-sequences true. I think Cary added it to
limit chances in test_decoding tests, while the misleading comment about
backwards compatibility comes from me.
regards
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2021-07-20 21:41 Tomas Vondra <[email protected]>
parent: vignesh C <[email protected]>
0 siblings, 2 replies; 127+ messages in thread
From: Tomas Vondra @ 2021-07-20 21:41 UTC (permalink / raw)
To: vignesh C <[email protected]>; +Cc: PostgreSQL Hackers <[email protected]>; Petr Jelinek <[email protected]>
Here's a rebased version of the patch, no other changes.
I think the crucial aspect of this that needs discussion/feedback the
most is the transactional vs. non-transactional behavior. All the other
questions are less important / cosmetic.
regards
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2021-07-30 18:26 Tomas Vondra <[email protected]>
parent: Tomas Vondra <[email protected]>
1 sibling, 1 reply; 127+ messages in thread
From: Tomas Vondra @ 2021-07-30 18:26 UTC (permalink / raw)
To: vignesh C <[email protected]>; +Cc: PostgreSQL Hackers <[email protected]>; Petr Jelinek <[email protected]>
Hi,
Here's a an updated version of this patch - rebased to current master,
and fixing some of the issues raised in Peter's review.
Mainly, this adds a TAP test to src/test/subscription, focusing on
testing the various situations with transactional and non-transactional
behavior (with subtransactions and various ROLLBACK versions).
This new TAP test however uncovered an issue with wait_for_catchup(),
because that uses pg_current_wal_lsn() to wait for replication of all
the changes. But that does not work when the sequence gets advanced in a
transaction that is then aborted, e.g. like this:
BEGIN;
SELECT nextval('s') FROM generate_series(1,100);
ROLLBACK;
The root cause is that pg_current_wal_lsn() uses the LogwrtResult.Write,
which is updated by XLogFlush() - but only in RecordTransactionCommit.
Which makes sense, because only the committed stuff is "visible". But
the non-transactional behavior changes this, because now some of the
changes from aborted transactions may need to be replicated. Which means
the wait_for_catchup() ends up not waiting for the sequence change.
One option would be adding XLogFlush() to RecordTransactionAbort(), but
my guess is we don't do that intentionally (even though aborts should be
fairly rare in most workloads).
I'm not entirely sure changing this (replicating changes from aborted
xacts) is a good idea. Maybe it'd be better to replicate this "lazily" -
instead of replicating the advances right away, we might remember which
sequences were advanced in the transaction, and then replicate current
state for those sequences at commit time.
The idea is that if an increment is "invisible" we probably don't need
to replicate it, it's fine to replicate the next "visible" increment. So
for example given
BEGIN;
SELECT nextval('s');
ROLLBACK;
BEGIN;
SELECT nextval('s');
COMMIT;
we don't need to replicate the first change, but we need to replicate
the second one.
The trouble is we don't decode individual sequence advances, just those
that update the sequence tuple (so every 32 values or so). So we'd need
to remeber the first increment, in a way that is (a) persistent across
restarts and (b) shared by all backends.
The other challenge seems to be ordering of the changes - at the moment
we have no issues with this, because increments on the same sequence are
replicated immediately, in the WAL order. But postponing them to commit
time would affect this order.
I've also briefly looked at the code duplication - there's a couple
functions in the patch that I shamelessly copy-pasted and tweaked to
handle sequences instead of tables:
publicationcmds.c
-----------------
1) OpenTableList/CloseTableList - > OpenSequenceList/CloseSequenceList
Trivial differences, trivial to get rid of - the only difference is
pretty much just table_open vs. relation open.
2) AlterPublicationTables -> AlterPublicationSequences
This is a bit more complicated, because the tables also handle
inheritance (which I think does not apply to sequences). Other than
that, it's calling the functions from (1).
subscriptioncmds.c
------------------
1) fetch_table_list, fetch_sequence_list
Minimal differences, essentially just the catalog name.
2) AlterSubscription_refresh
A lot of duplication, but the code handling tables and sequences is
almost exactly the same and can be reused fairly easily (moved to a
separate function, called for tables and then sequences).
regards
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
Attachments:
[text/x-patch] 0001-Logical-decoding-replication-of-sequences-20210729.patch (117.8K, ../../[email protected]/2-0001-Logical-decoding-replication-of-sequences-20210729.patch)
download | inline diff:
diff --git a/contrib/test_decoding/Makefile b/contrib/test_decoding/Makefile
index 9a31e0b879..56ddc3abae 100644
--- a/contrib/test_decoding/Makefile
+++ b/contrib/test_decoding/Makefile
@@ -5,7 +5,8 @@ PGFILEDESC = "test_decoding - example of a logical decoding output plugin"
REGRESS = ddl xact rewrite toast permissions decoding_in_xact \
decoding_into_rel binary prepared replorigin time messages \
- spill slot truncate stream stats twophase twophase_stream
+ spill slot truncate stream stats twophase twophase_stream \
+ sequence
ISOLATION = mxact delayed_startup ondisk_startup concurrent_ddl_dml \
oldest_xmin snapshot_transfer subxact_without_top concurrent_stream \
twophase_snapshot
diff --git a/contrib/test_decoding/expected/sequence.out b/contrib/test_decoding/expected/sequence.out
new file mode 100644
index 0000000000..07f3e3e92c
--- /dev/null
+++ b/contrib/test_decoding/expected/sequence.out
@@ -0,0 +1,327 @@
+-- predictability
+SET synchronous_commit = on;
+SELECT 'init' FROM pg_create_logical_replication_slot('regression_slot', 'test_decoding');
+ ?column?
+----------
+ init
+(1 row)
+
+CREATE SEQUENCE test_sequence;
+-- test the sequence changes by several nextval() calls
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 1
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 2
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 4
+(1 row)
+
+-- test the sequence changes by several ALTER commands
+ALTER SEQUENCE test_sequence INCREMENT BY 10;
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 14
+(1 row)
+
+ALTER SEQUENCE test_sequence START WITH 3000;
+ALTER SEQUENCE test_sequence MAXVALUE 10000;
+ALTER SEQUENCE test_sequence RESTART WITH 4000;
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 4000
+(1 row)
+
+-- test the sequence changes by several setval() calls
+SELECT setval('test_sequence', 3500);
+ setval
+--------
+ 3500
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3510
+(1 row)
+
+SELECT setval('test_sequence', 3500, true);
+ setval
+--------
+ 3500
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3510
+(1 row)
+
+SELECT setval('test_sequence', 3500, false);
+ setval
+--------
+ 3500
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3500
+(1 row)
+
+-- show results and drop sequence
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+ data
+------------------------------------------------------------------------------------------------------
+ BEGIN
+ sequence: public.test_sequence transactional: 1 created: 1 last_value: 1, log_cnt: 0 is_called: 0
+ COMMIT
+ sequence: public.test_sequence transactional: 0 created: 0 last_value: 33, log_cnt: 0 is_called: 1
+ BEGIN
+ sequence: public.test_sequence transactional: 1 created: 1 last_value: 4, log_cnt: 0 is_called: 1
+ COMMIT
+ sequence: public.test_sequence transactional: 0 created: 0 last_value: 334, log_cnt: 0 is_called: 1
+ BEGIN
+ sequence: public.test_sequence transactional: 1 created: 1 last_value: 14, log_cnt: 32 is_called: 1
+ COMMIT
+ BEGIN
+ sequence: public.test_sequence transactional: 1 created: 1 last_value: 14, log_cnt: 0 is_called: 1
+ COMMIT
+ BEGIN
+ sequence: public.test_sequence transactional: 1 created: 1 last_value: 4000, log_cnt: 0 is_called: 0
+ COMMIT
+ sequence: public.test_sequence transactional: 0 created: 0 last_value: 4320, log_cnt: 0 is_called: 1
+ sequence: public.test_sequence transactional: 0 created: 0 last_value: 3500, log_cnt: 0 is_called: 1
+ sequence: public.test_sequence transactional: 0 created: 0 last_value: 3830, log_cnt: 0 is_called: 1
+ sequence: public.test_sequence transactional: 0 created: 0 last_value: 3500, log_cnt: 0 is_called: 1
+ sequence: public.test_sequence transactional: 0 created: 0 last_value: 3830, log_cnt: 0 is_called: 1
+ sequence: public.test_sequence transactional: 0 created: 0 last_value: 3500, log_cnt: 0 is_called: 0
+ sequence: public.test_sequence transactional: 0 created: 0 last_value: 3820, log_cnt: 0 is_called: 1
+(24 rows)
+
+DROP SEQUENCE test_sequence;
+-- rollback on sequence creation and update
+BEGIN;
+CREATE SEQUENCE test_sequence;
+CREATE TABLE test_table (a INT);
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 1
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 2
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3
+(1 row)
+
+SELECT setval('test_sequence', 3000);
+ setval
+--------
+ 3000
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3001
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3002
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3003
+(1 row)
+
+ALTER SEQUENCE test_sequence RESTART WITH 6000;
+INSERT INTO test_table VALUES( (SELECT nextval('test_sequence')) );
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 6001
+(1 row)
+
+ROLLBACK;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'skip-sequences', '0');
+ data
+------
+(0 rows)
+
+-- rollback on table creation with serial column
+BEGIN;
+CREATE TABLE test_table (a SERIAL, b INT);
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+INSERT INTO test_table (b) VALUES (300);
+SELECT setval('test_table_a_seq', 3000);
+ setval
+--------
+ 3000
+(1 row)
+
+INSERT INTO test_table (b) VALUES (400);
+INSERT INTO test_table (b) VALUES (500);
+INSERT INTO test_table (b) VALUES (600);
+ALTER SEQUENCE test_table_a_seq RESTART WITH 6000;
+INSERT INTO test_table (b) VALUES (700);
+INSERT INTO test_table (b) VALUES (800);
+INSERT INTO test_table (b) VALUES (900);
+ROLLBACK;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'skip-sequences', '0');
+ data
+------
+(0 rows)
+
+-- rollback on table with serial column
+CREATE TABLE test_table (a SERIAL, b INT);
+BEGIN;
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+INSERT INTO test_table (b) VALUES (300);
+SELECT setval('test_table_a_seq', 3000);
+ setval
+--------
+ 3000
+(1 row)
+
+INSERT INTO test_table (b) VALUES (400);
+INSERT INTO test_table (b) VALUES (500);
+INSERT INTO test_table (b) VALUES (600);
+ALTER SEQUENCE test_table_a_seq RESTART WITH 6000;
+INSERT INTO test_table (b) VALUES (700);
+INSERT INTO test_table (b) VALUES (800);
+INSERT INTO test_table (b) VALUES (900);
+ROLLBACK;
+-- check table and sequence values after rollback
+SELECT * from test_table_a_seq;
+ last_value | log_cnt | is_called
+------------+---------+-----------
+ 3003 | 30 | t
+(1 row)
+
+SELECT nextval('test_table_a_seq');
+ nextval
+---------
+ 3004
+(1 row)
+
+DROP TABLE test_table;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+ data
+---------------------------------------------------------------------------------------------------------
+ BEGIN
+ sequence: public.test_table_a_seq transactional: 1 created: 1 last_value: 1, log_cnt: 0 is_called: 0
+ COMMIT
+ sequence: public.test_table_a_seq transactional: 0 created: 0 last_value: 33, log_cnt: 0 is_called: 1
+ sequence: public.test_table_a_seq transactional: 0 created: 0 last_value: 3000, log_cnt: 0 is_called: 1
+ sequence: public.test_table_a_seq transactional: 0 created: 0 last_value: 3033, log_cnt: 0 is_called: 1
+ BEGIN
+ COMMIT
+(8 rows)
+
+-- savepoint test on table with serial column
+BEGIN;
+CREATE TABLE test_table (a SERIAL, b INT);
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+SAVEPOINT a;
+INSERT INTO test_table (b) VALUES (300);
+ROLLBACK TO SAVEPOINT a;
+DROP TABLE test_table;
+COMMIT;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+ data
+-------------------------------------------------------------------------------------------------------
+ BEGIN
+ sequence: public.test_table_a_seq transactional: 1 created: 1 last_value: 1, log_cnt: 0 is_called: 0
+ sequence: public.test_table_a_seq transactional: 1 created: 0 last_value: 33, log_cnt: 0 is_called: 1
+ table public.test_table: INSERT: a[integer]:1 b[integer]:100
+ table public.test_table: INSERT: a[integer]:2 b[integer]:200
+ COMMIT
+(6 rows)
+
+-- savepoint test on table with serial column
+BEGIN;
+CREATE SEQUENCE test_sequence;
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 1
+(1 row)
+
+SELECT setval('test_sequence', 3000);
+ setval
+--------
+ 3000
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3001
+(1 row)
+
+SAVEPOINT a;
+ALTER SEQUENCE test_sequence START WITH 7000;
+SELECT setval('test_sequence', 5000);
+ setval
+--------
+ 5000
+(1 row)
+
+ROLLBACK TO SAVEPOINT a;
+SELECT * FROM test_sequence;
+ last_value | log_cnt | is_called
+------------+---------+-----------
+ 3001 | 32 | t
+(1 row)
+
+DROP SEQUENCE test_sequence;
+COMMIT;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+ data
+------------------------------------------------------------------------------------------------------
+ BEGIN
+ sequence: public.test_sequence transactional: 1 created: 1 last_value: 1, log_cnt: 0 is_called: 0
+ sequence: public.test_sequence transactional: 1 created: 0 last_value: 33, log_cnt: 0 is_called: 1
+ sequence: public.test_sequence transactional: 1 created: 0 last_value: 3000, log_cnt: 0 is_called: 1
+ sequence: public.test_sequence transactional: 1 created: 0 last_value: 3033, log_cnt: 0 is_called: 1
+ COMMIT
+(6 rows)
+
+SELECT pg_drop_replication_slot('regression_slot');
+ pg_drop_replication_slot
+--------------------------
+
+(1 row)
+
diff --git a/contrib/test_decoding/sql/sequence.sql b/contrib/test_decoding/sql/sequence.sql
new file mode 100644
index 0000000000..d8a34738f3
--- /dev/null
+++ b/contrib/test_decoding/sql/sequence.sql
@@ -0,0 +1,119 @@
+-- predictability
+SET synchronous_commit = on;
+SELECT 'init' FROM pg_create_logical_replication_slot('regression_slot', 'test_decoding');
+
+CREATE SEQUENCE test_sequence;
+
+-- test the sequence changes by several nextval() calls
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+
+-- test the sequence changes by several ALTER commands
+ALTER SEQUENCE test_sequence INCREMENT BY 10;
+SELECT nextval('test_sequence');
+
+ALTER SEQUENCE test_sequence START WITH 3000;
+ALTER SEQUENCE test_sequence MAXVALUE 10000;
+ALTER SEQUENCE test_sequence RESTART WITH 4000;
+SELECT nextval('test_sequence');
+
+-- test the sequence changes by several setval() calls
+SELECT setval('test_sequence', 3500);
+SELECT nextval('test_sequence');
+SELECT setval('test_sequence', 3500, true);
+SELECT nextval('test_sequence');
+SELECT setval('test_sequence', 3500, false);
+SELECT nextval('test_sequence');
+
+-- show results and drop sequence
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+DROP SEQUENCE test_sequence;
+
+-- rollback on sequence creation and update
+BEGIN;
+CREATE SEQUENCE test_sequence;
+CREATE TABLE test_table (a INT);
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT setval('test_sequence', 3000);
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+ALTER SEQUENCE test_sequence RESTART WITH 6000;
+INSERT INTO test_table VALUES( (SELECT nextval('test_sequence')) );
+SELECT nextval('test_sequence');
+ROLLBACK;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'skip-sequences', '0');
+
+-- rollback on table creation with serial column
+BEGIN;
+CREATE TABLE test_table (a SERIAL, b INT);
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+INSERT INTO test_table (b) VALUES (300);
+SELECT setval('test_table_a_seq', 3000);
+INSERT INTO test_table (b) VALUES (400);
+INSERT INTO test_table (b) VALUES (500);
+INSERT INTO test_table (b) VALUES (600);
+ALTER SEQUENCE test_table_a_seq RESTART WITH 6000;
+INSERT INTO test_table (b) VALUES (700);
+INSERT INTO test_table (b) VALUES (800);
+INSERT INTO test_table (b) VALUES (900);
+ROLLBACK;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'skip-sequences', '0');
+
+-- rollback on table with serial column
+CREATE TABLE test_table (a SERIAL, b INT);
+
+BEGIN;
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+INSERT INTO test_table (b) VALUES (300);
+SELECT setval('test_table_a_seq', 3000);
+INSERT INTO test_table (b) VALUES (400);
+INSERT INTO test_table (b) VALUES (500);
+INSERT INTO test_table (b) VALUES (600);
+ALTER SEQUENCE test_table_a_seq RESTART WITH 6000;
+INSERT INTO test_table (b) VALUES (700);
+INSERT INTO test_table (b) VALUES (800);
+INSERT INTO test_table (b) VALUES (900);
+ROLLBACK;
+
+-- check table and sequence values after rollback
+SELECT * from test_table_a_seq;
+SELECT nextval('test_table_a_seq');
+
+DROP TABLE test_table;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+
+-- savepoint test on table with serial column
+BEGIN;
+CREATE TABLE test_table (a SERIAL, b INT);
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+SAVEPOINT a;
+INSERT INTO test_table (b) VALUES (300);
+ROLLBACK TO SAVEPOINT a;
+DROP TABLE test_table;
+COMMIT;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+
+-- savepoint test on table with serial column
+BEGIN;
+CREATE SEQUENCE test_sequence;
+SELECT nextval('test_sequence');
+SELECT setval('test_sequence', 3000);
+SELECT nextval('test_sequence');
+SAVEPOINT a;
+ALTER SEQUENCE test_sequence START WITH 7000;
+SELECT setval('test_sequence', 5000);
+ROLLBACK TO SAVEPOINT a;
+SELECT * FROM test_sequence;
+DROP SEQUENCE test_sequence;
+COMMIT;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+
+SELECT pg_drop_replication_slot('regression_slot');
diff --git a/contrib/test_decoding/test_decoding.c b/contrib/test_decoding/test_decoding.c
index e5cd84e85e..54d734cb85 100644
--- a/contrib/test_decoding/test_decoding.c
+++ b/contrib/test_decoding/test_decoding.c
@@ -35,6 +35,7 @@ typedef struct
bool include_timestamp;
bool skip_empty_xacts;
bool only_local;
+ bool skip_sequences;
} TestDecodingData;
/*
@@ -76,6 +77,10 @@ static void pg_decode_message(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn, XLogRecPtr message_lsn,
bool transactional, const char *prefix,
Size sz, const char *message);
+static void pg_decode_sequence(LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn, XLogRecPtr sequence_lsn,
+ Relation rel, bool transactional, bool created,
+ int64 last_value, int64 log_cnt, bool is_called);
static bool pg_decode_filter_prepare(LogicalDecodingContext *ctx,
TransactionId xid,
const char *gid);
@@ -141,6 +146,7 @@ _PG_output_plugin_init(OutputPluginCallbacks *cb)
cb->filter_by_origin_cb = pg_decode_filter;
cb->shutdown_cb = pg_decode_shutdown;
cb->message_cb = pg_decode_message;
+ cb->sequence_cb = pg_decode_sequence;
cb->filter_prepare_cb = pg_decode_filter_prepare;
cb->begin_prepare_cb = pg_decode_begin_prepare_txn;
cb->prepare_cb = pg_decode_prepare_txn;
@@ -175,6 +181,9 @@ pg_decode_startup(LogicalDecodingContext *ctx, OutputPluginOptions *opt,
data->skip_empty_xacts = false;
data->only_local = false;
+ /* skip sequences by default for backwards compatibility */
+ data->skip_sequences = true;
+
ctx->output_plugin_private = data;
opt->output_type = OUTPUT_PLUGIN_TEXTUAL_OUTPUT;
@@ -265,6 +274,17 @@ pg_decode_startup(LogicalDecodingContext *ctx, OutputPluginOptions *opt,
errmsg("could not parse value \"%s\" for parameter \"%s\"",
strVal(elem->arg), elem->defname)));
}
+ else if (strcmp(elem->defname, "skip-sequences") == 0)
+ {
+
+ if (elem->arg == NULL)
+ continue; /* true by default */
+ else if (!parse_bool(strVal(elem->arg), &data->skip_sequences))
+ ereport(ERROR,
+ (errcode(ERRCODE_INVALID_PARAMETER_VALUE),
+ errmsg("could not parse value \"%s\" for parameter \"%s\"",
+ strVal(elem->arg), elem->defname)));
+ }
else
{
ereport(ERROR,
@@ -744,6 +764,26 @@ pg_decode_message(LogicalDecodingContext *ctx,
OutputPluginWrite(ctx, true);
}
+static void
+pg_decode_sequence(LogicalDecodingContext *ctx, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ bool transactional, bool created,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ TestDecodingData *data = ctx->output_plugin_private;
+ char *nspname = get_namespace_name(RelationGetNamespace(rel));
+
+ /* return if requested to skip_sequences */
+ if (data->skip_sequences)
+ return;
+
+ OutputPluginPrepareWrite(ctx, true);
+ appendStringInfo(ctx->out, "sequence: %s.%s transactional: %d created: %d last_value: %zu, log_cnt: %zu is_called: %d",
+ nspname, RelationGetRelationName(rel),
+ transactional, created, last_value, log_cnt, is_called);
+ OutputPluginWrite(ctx, true);
+}
+
static void
pg_decode_stream_start(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn)
diff --git a/doc/src/sgml/catalogs.sgml b/doc/src/sgml/catalogs.sgml
index 2b2c70a26e..a342a54629 100644
--- a/doc/src/sgml/catalogs.sgml
+++ b/doc/src/sgml/catalogs.sgml
@@ -9430,6 +9430,11 @@ SCRAM-SHA-256$<replaceable><iteration count></replaceable>:<replaceable>&l
<entry>prepared transactions</entry>
</row>
+ <row>
+ <entry><link linkend="view-pg-publication-sequences"><structname>pg_publication_sequences</structname></link></entry>
+ <entry>publications and their associated sequences</entry>
+ </row>
+
<row>
<entry><link linkend="view-pg-publication-tables"><structname>pg_publication_tables</structname></link></entry>
<entry>publications and their associated tables</entry>
@@ -11264,6 +11269,72 @@ SELECT * FROM pg_locks pl LEFT JOIN pg_prepared_xacts ppx
</sect1>
+ <sect1 id="view-pg-publication-sequences">
+ <title><structname>pg_publication_sequences</structname></title>
+
+ <indexterm zone="view-pg-publication-sequences">
+ <primary>pg_publication_sequences</primary>
+ </indexterm>
+
+ <para>
+ The view <structname>pg_publication_sequences</structname> provides
+ information about the mapping between publications and the sequences they
+ contain. Unlike the underlying catalog
+ <link linkend="catalog-pg-publication-rel"><structname>pg_publication_rel</structname></link>,
+ this view expands
+ publications defined as <literal>FOR ALL SEQUENCES</literal>, so for such
+ publications there will be a row for each eligible sequence.
+ </para>
+
+ <table>
+ <title><structname>pg_publication_sequences</structname> Columns</title>
+ <tgroup cols="1">
+ <thead>
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ Column Type
+ </para>
+ <para>
+ Description
+ </para></entry>
+ </row>
+ </thead>
+
+ <tbody>
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ <structfield>pubname</structfield> <type>name</type>
+ (references <link linkend="catalog-pg-publication"><structname>pg_publication</structname></link>.<structfield>pubname</structfield>)
+ </para>
+ <para>
+ Name of publication
+ </para></entry>
+ </row>
+
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ <structfield>schemaname</structfield> <type>name</type>
+ (references <link linkend="catalog-pg-namespace"><structname>pg_namespace</structname></link>.<structfield>nspname</structfield>)
+ </para>
+ <para>
+ Name of schema containing sequence
+ </para></entry>
+ </row>
+
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ <structfield>sequencename</structfield> <type>name</type>
+ (references <link linkend="catalog-pg-class"><structname>pg_class</structname></link>.<structfield>relname</structfield>)
+ </para>
+ <para>
+ Name of sequence
+ </para></entry>
+ </row>
+ </tbody>
+ </tgroup>
+ </table>
+ </sect1>
+
<sect1 id="view-pg-publication-tables">
<title><structname>pg_publication_tables</structname></title>
diff --git a/doc/src/sgml/ref/alter_publication.sgml b/doc/src/sgml/ref/alter_publication.sgml
index faa114b2c6..c68a1573de 100644
--- a/doc/src/sgml/ref/alter_publication.sgml
+++ b/doc/src/sgml/ref/alter_publication.sgml
@@ -24,6 +24,9 @@ PostgreSQL documentation
ALTER PUBLICATION <replaceable class="parameter">name</replaceable> ADD TABLE [ ONLY ] <replaceable class="parameter">table_name</replaceable> [ * ] [, ...]
ALTER PUBLICATION <replaceable class="parameter">name</replaceable> SET TABLE [ ONLY ] <replaceable class="parameter">table_name</replaceable> [ * ] [, ...]
ALTER PUBLICATION <replaceable class="parameter">name</replaceable> DROP TABLE [ ONLY ] <replaceable class="parameter">table_name</replaceable> [ * ] [, ...]
+ALTER PUBLICATION <replaceable class="parameter">name</replaceable> ADD SEQUENCE <replaceable class="parameter">sequence_name</replaceable> [ * ] [, ...]
+ALTER PUBLICATION <replaceable class="parameter">name</replaceable> SET SEQUENCE <replaceable class="parameter">sequence_name</replaceable> [ * ] [, ...]
+ALTER PUBLICATION <replaceable class="parameter">name</replaceable> DROP SEQUENCE <replaceable class="parameter">sequence_name</replaceable> [ * ] [, ...]
ALTER PUBLICATION <replaceable class="parameter">name</replaceable> SET ( <replaceable class="parameter">publication_parameter</replaceable> [= <replaceable class="parameter">value</replaceable>] [, ... ] )
ALTER PUBLICATION <replaceable class="parameter">name</replaceable> OWNER TO { <replaceable>new_owner</replaceable> | CURRENT_ROLE | CURRENT_USER | SESSION_USER }
ALTER PUBLICATION <replaceable class="parameter">name</replaceable> RENAME TO <replaceable>new_name</replaceable>
@@ -50,7 +53,18 @@ ALTER PUBLICATION <replaceable class="parameter">name</replaceable> RENAME TO <r
</para>
<para>
- The fourth variant of this command listed in the synopsis can change
+ The next three variants change which sequences are part of the publication.
+ The <literal>SET SEQUENCE</literal> clause will replace the list of sequences
+ in the publication with the specified one. The <literal>ADD SEQUENCE</literal>
+ and <literal>DROP SEQUENCE</literal> clauses will add and remove one or more
+ sequences from the publication. Note that adding sequences to a publication
+ that is already subscribed to will require a <literal>ALTER SUBSCRIPTION
+ ... REFRESH PUBLICATION</literal> action on the subscribing side in order
+ to become effective.
+ </para>
+
+ <para>
+ The seventh variant of this command listed in the synopsis can change
all of the publication properties specified in
<xref linkend="sql-createpublication"/>. Properties not mentioned in the
command retain their previous settings.
@@ -62,7 +76,8 @@ ALTER PUBLICATION <replaceable class="parameter">name</replaceable> RENAME TO <r
<para>
You must own the publication to use <command>ALTER PUBLICATION</command>.
- Adding a table to a publication additionally requires owning that table.
+ Adding a table to a publication additionally requires owning that table,
+ and the same requirement applies to sequences.
To alter the owner, you must also be a direct or indirect member of the new
owning role. The new owner must have <literal>CREATE</literal> privilege on
the database. Also, the new owner of a <literal>FOR ALL TABLES</literal>
@@ -97,6 +112,15 @@ ALTER PUBLICATION <replaceable class="parameter">name</replaceable> RENAME TO <r
</listitem>
</varlistentry>
+ <varlistentry>
+ <term><replaceable class="parameter">sequence_name</replaceable></term>
+ <listitem>
+ <para>
+ Name of an existing sequence.
+ </para>
+ </listitem>
+ </varlistentry>
+
<varlistentry>
<term><literal>SET ( <replaceable class="parameter">publication_parameter</replaceable> [= <replaceable class="parameter">value</replaceable>] [, ... ] )</literal></term>
<listitem>
diff --git a/doc/src/sgml/ref/alter_subscription.sgml b/doc/src/sgml/ref/alter_subscription.sgml
index a6f994450d..ff1dcd3bfd 100644
--- a/doc/src/sgml/ref/alter_subscription.sgml
+++ b/doc/src/sgml/ref/alter_subscription.sgml
@@ -144,8 +144,8 @@ ALTER SUBSCRIPTION <replaceable class="parameter">name</replaceable> RENAME TO <
<term><literal>REFRESH PUBLICATION</literal></term>
<listitem>
<para>
- Fetch missing table information from publisher. This will start
- replication of tables that were added to the subscribed-to publications
+ Fetch missing table and sequence information from publisher. This will start
+ replication of tables and sequences that were added to the subscribed-to publications
since the last invocation of <command>REFRESH PUBLICATION</command> or
since <command>CREATE SUBSCRIPTION</command>.
</para>
@@ -162,7 +162,7 @@ ALTER SUBSCRIPTION <replaceable class="parameter">name</replaceable> RENAME TO <
Specifies whether the existing data in the publications that are
being subscribed to should be copied once the replication starts.
The default is <literal>true</literal>. (Previously subscribed
- tables are not copied.)
+ tables and sequences are not copied.)
</para>
</listitem>
</varlistentry>
diff --git a/src/backend/catalog/pg_publication.c b/src/backend/catalog/pg_publication.c
index 2a2fe03c13..a66a5a2aec 100644
--- a/src/backend/catalog/pg_publication.c
+++ b/src/backend/catalog/pg_publication.c
@@ -51,7 +51,8 @@ check_publication_add_relation(Relation targetrel)
{
/* Must be a regular or partitioned table */
if (RelationGetForm(targetrel)->relkind != RELKIND_RELATION &&
- RelationGetForm(targetrel)->relkind != RELKIND_PARTITIONED_TABLE)
+ RelationGetForm(targetrel)->relkind != RELKIND_PARTITIONED_TABLE &&
+ RelationGetForm(targetrel)->relkind != RELKIND_SEQUENCE)
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("cannot add relation \"%s\" to publication",
@@ -98,7 +99,8 @@ static bool
is_publishable_class(Oid relid, Form_pg_class reltuple)
{
return (reltuple->relkind == RELKIND_RELATION ||
- reltuple->relkind == RELKIND_PARTITIONED_TABLE) &&
+ reltuple->relkind == RELKIND_PARTITIONED_TABLE ||
+ reltuple->relkind == RELKIND_SEQUENCE) &&
!IsCatalogRelationOid(relid) &&
reltuple->relpersistence == RELPERSISTENCE_PERMANENT &&
relid >= FirstNormalObjectId;
@@ -271,6 +273,10 @@ GetPublicationRelations(Oid pubid, PublicationPartOpt pub_partopt)
pubrel = (Form_pg_publication_rel) GETSTRUCT(tup);
+ /* skip sequences here */
+ if (get_rel_relkind(pubrel->prrelid) == RELKIND_SEQUENCE)
+ continue;
+
if (get_rel_relkind(pubrel->prrelid) == RELKIND_PARTITIONED_TABLE &&
pub_partopt != PUBLICATION_PART_ROOT)
{
@@ -304,6 +310,49 @@ GetPublicationRelations(Oid pubid, PublicationPartOpt pub_partopt)
return result;
}
+/*
+ * Gets list of relation oids for a publication (sequences only).
+ *
+ * This should only be used for normal publications, the FOR ALL TABLES
+ * should use GetAllSequencesPublicationRelations().
+ */
+List *
+GetPublicationSequenceRelations(Oid pubid)
+{
+ List *result;
+ Relation pubrelsrel;
+ ScanKeyData scankey;
+ SysScanDesc scan;
+ HeapTuple tup;
+
+ /* Find all publications associated with the relation. */
+ pubrelsrel = table_open(PublicationRelRelationId, AccessShareLock);
+
+ ScanKeyInit(&scankey,
+ Anum_pg_publication_rel_prpubid,
+ BTEqualStrategyNumber, F_OIDEQ,
+ ObjectIdGetDatum(pubid));
+
+ scan = systable_beginscan(pubrelsrel, PublicationRelPrrelidPrpubidIndexId,
+ true, NULL, 1, &scankey);
+
+ result = NIL;
+ while (HeapTupleIsValid(tup = systable_getnext(scan)))
+ {
+ Form_pg_publication_rel pubrel;
+
+ pubrel = (Form_pg_publication_rel) GETSTRUCT(tup);
+
+ if (get_rel_relkind(pubrel->prrelid) == RELKIND_SEQUENCE)
+ result = lappend_oid(result, pubrel->prrelid);
+ }
+
+ systable_endscan(scan);
+ table_close(pubrelsrel, AccessShareLock);
+
+ return result;
+}
+
/*
* Gets list of publication oids for publications marked as FOR ALL TABLES.
*/
@@ -404,6 +453,46 @@ GetAllTablesPublicationRelations(bool pubviaroot)
return result;
}
+/*
+ * Gets list of all relation published by FOR ALL TABLES publication(s).
+ *
+ * If the publication publishes partition changes via their respective root
+ * partitioned tables, we must exclude partitions in favor of including the
+ * root partitioned tables.
+ */
+List *
+GetAllSequencesPublicationRelations(void)
+{
+ Relation classRel;
+ ScanKeyData key[1];
+ TableScanDesc scan;
+ HeapTuple tuple;
+ List *result = NIL;
+
+ classRel = table_open(RelationRelationId, AccessShareLock);
+
+ ScanKeyInit(&key[0],
+ Anum_pg_class_relkind,
+ BTEqualStrategyNumber, F_CHAREQ,
+ CharGetDatum(RELKIND_SEQUENCE));
+
+ scan = table_beginscan_catalog(classRel, 1, key);
+
+ while ((tuple = heap_getnext(scan, ForwardScanDirection)) != NULL)
+ {
+ Form_pg_class relForm = (Form_pg_class) GETSTRUCT(tuple);
+ Oid relid = relForm->oid;
+
+ if (is_publishable_class(relid, relForm))
+ result = lappend_oid(result, relid);
+ }
+
+ table_endscan(scan);
+
+ table_close(classRel, AccessShareLock);
+ return result;
+}
+
/*
* Get publication using oid
*
@@ -426,10 +515,12 @@ GetPublication(Oid pubid)
pub->oid = pubid;
pub->name = pstrdup(NameStr(pubform->pubname));
pub->alltables = pubform->puballtables;
+ pub->allsequences = pubform->puballsequences;
pub->pubactions.pubinsert = pubform->pubinsert;
pub->pubactions.pubupdate = pubform->pubupdate;
pub->pubactions.pubdelete = pubform->pubdelete;
pub->pubactions.pubtruncate = pubform->pubtruncate;
+ pub->pubactions.pubsequence = pubform->pubsequence;
pub->pubviaroot = pubform->pubviaroot;
ReleaseSysCache(tup);
@@ -555,3 +646,56 @@ pg_get_publication_tables(PG_FUNCTION_ARGS)
SRF_RETURN_DONE(funcctx);
}
+
+/*
+ * Returns Oids of sequences in a publication.
+ */
+Datum
+pg_get_publication_sequences(PG_FUNCTION_ARGS)
+{
+ FuncCallContext *funcctx;
+ char *pubname = text_to_cstring(PG_GETARG_TEXT_PP(0));
+ Publication *publication;
+ List *sequences;
+
+ /* stuff done only on the first call of the function */
+ if (SRF_IS_FIRSTCALL())
+ {
+ MemoryContext oldcontext;
+
+ /* create a function context for cross-call persistence */
+ funcctx = SRF_FIRSTCALL_INIT();
+
+ /* switch to memory context appropriate for multiple function calls */
+ oldcontext = MemoryContextSwitchTo(funcctx->multi_call_memory_ctx);
+
+ publication = GetPublicationByName(pubname, false);
+
+ /*
+ * Publications support partitioned tables, although all changes are
+ * replicated using leaf partition identity and schema, so we only
+ * need those.
+ */
+ if (publication->allsequences)
+ sequences = GetAllSequencesPublicationRelations();
+ else
+ sequences = GetPublicationSequenceRelations(publication->oid);
+
+ funcctx->user_fctx = (void *) sequences;
+
+ MemoryContextSwitchTo(oldcontext);
+ }
+
+ /* stuff done on every call of the function */
+ funcctx = SRF_PERCALL_SETUP();
+ sequences = (List *) funcctx->user_fctx;
+
+ if (funcctx->call_cntr < list_length(sequences))
+ {
+ Oid relid = list_nth_oid(sequences, funcctx->call_cntr);
+
+ SRF_RETURN_NEXT(funcctx, ObjectIdGetDatum(relid));
+ }
+
+ SRF_RETURN_DONE(funcctx);
+}
diff --git a/src/backend/catalog/system_views.sql b/src/backend/catalog/system_views.sql
index 55f6e3711d..9bec49bc3e 100644
--- a/src/backend/catalog/system_views.sql
+++ b/src/backend/catalog/system_views.sql
@@ -372,6 +372,16 @@ CREATE VIEW pg_publication_tables AS
pg_class C JOIN pg_namespace N ON (N.oid = C.relnamespace)
WHERE C.oid = GPT.relid;
+CREATE VIEW pg_publication_sequences AS
+ SELECT
+ P.pubname AS pubname,
+ N.nspname AS schemaname,
+ C.relname AS sequencename
+ FROM pg_publication P,
+ LATERAL pg_get_publication_sequences(P.pubname) GPT,
+ pg_class C JOIN pg_namespace N ON (N.oid = C.relnamespace)
+ WHERE C.oid = GPT.relid;
+
CREATE VIEW pg_locks AS
SELECT * FROM pg_lock_status() AS L;
diff --git a/src/backend/commands/publicationcmds.c b/src/backend/commands/publicationcmds.c
index 8487eeb7e6..90ad83d4a1 100644
--- a/src/backend/commands/publicationcmds.c
+++ b/src/backend/commands/publicationcmds.c
@@ -16,6 +16,7 @@
#include "access/genam.h"
#include "access/htup_details.h"
+#include "access/relation.h"
#include "access/table.h"
#include "access/xact.h"
#include "catalog/catalog.h"
@@ -54,6 +55,12 @@ static void PublicationAddTables(Oid pubid, List *rels, bool if_not_exists,
AlterPublicationStmt *stmt);
static void PublicationDropTables(Oid pubid, List *rels, bool missing_ok);
+static List *OpenSequenceList(List *sequences);
+static void CloseSequenceList(List *rels);
+static void PublicationAddSequences(Oid pubid, List *rels, bool if_not_exists,
+ AlterPublicationStmt *stmt);
+static void PublicationDropSequences(Oid pubid, List *rels, bool missing_ok);
+
static void
parse_publication_options(ParseState *pstate,
List *options,
@@ -72,6 +79,7 @@ parse_publication_options(ParseState *pstate,
pubactions->pubupdate = true;
pubactions->pubdelete = true;
pubactions->pubtruncate = true;
+ pubactions->pubsequence = true;
*publish_via_partition_root = false;
/* Parse options */
@@ -96,6 +104,7 @@ parse_publication_options(ParseState *pstate,
pubactions->pubupdate = false;
pubactions->pubdelete = false;
pubactions->pubtruncate = false;
+ pubactions->pubsequence = false;
*publish_given = true;
publish = defGetString(defel);
@@ -118,6 +127,8 @@ parse_publication_options(ParseState *pstate,
pubactions->pubdelete = true;
else if (strcmp(publish_opt, "truncate") == 0)
pubactions->pubtruncate = true;
+ else if (strcmp(publish_opt, "sequence") == 0)
+ pubactions->pubsequence = true;
else
ereport(ERROR,
(errcode(ERRCODE_SYNTAX_ERROR),
@@ -208,6 +219,8 @@ CreatePublication(ParseState *pstate, CreatePublicationStmt *stmt)
BoolGetDatum(pubactions.pubdelete);
values[Anum_pg_publication_pubtruncate - 1] =
BoolGetDatum(pubactions.pubtruncate);
+ values[Anum_pg_publication_pubsequence - 1] =
+ BoolGetDatum(pubactions.pubsequence);
values[Anum_pg_publication_pubviaroot - 1] =
BoolGetDatum(publish_via_partition_root);
@@ -373,9 +386,9 @@ AlterPublicationTables(AlterPublicationStmt *stmt, Relation rel,
rels = OpenTableList(stmt->tables);
- if (stmt->tableAction == DEFELEM_ADD)
+ if (stmt->action == DEFELEM_ADD)
PublicationAddTables(pubid, rels, false, stmt);
- else if (stmt->tableAction == DEFELEM_DROP)
+ else if (stmt->action == DEFELEM_DROP)
PublicationDropTables(pubid, rels, false);
else /* DEFELEM_SET */
{
@@ -426,6 +439,82 @@ AlterPublicationTables(AlterPublicationStmt *stmt, Relation rel,
CloseTableList(rels);
}
+/*
+ * Add or remove sequence to/from publication.
+ */
+static void
+AlterPublicationSequences(AlterPublicationStmt *stmt, Relation rel,
+ HeapTuple tup)
+{
+ List *rels = NIL;
+ Form_pg_publication pubform = (Form_pg_publication) GETSTRUCT(tup);
+ Oid pubid = pubform->oid;
+
+ /* Check that user is allowed to manipulate the publication tables. */
+ if (pubform->puballsequences)
+ ereport(ERROR,
+ (errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
+ errmsg("publication \"%s\" is defined as FOR ALL SEQUENCES",
+ NameStr(pubform->pubname)),
+ errdetail("Sequences cannot be added to or dropped from FOR ALL TABLES publications.")));
+
+ Assert(list_length(stmt->sequences) > 0);
+
+ rels = OpenSequenceList(stmt->sequences);
+
+ if (stmt->action == DEFELEM_ADD)
+ PublicationAddSequences(pubid, rels, false, stmt);
+ else if (stmt->action == DEFELEM_DROP)
+ PublicationDropSequences(pubid, rels, false);
+ else /* DEFELEM_SET */
+ {
+ List *oldrelids = GetPublicationRelations(pubid,
+ PUBLICATION_PART_ROOT);
+ List *delrels = NIL;
+ ListCell *oldlc;
+
+ /* Calculate which relations to drop. */
+ foreach(oldlc, oldrelids)
+ {
+ Oid oldrelid = lfirst_oid(oldlc);
+ ListCell *newlc;
+ bool found = false;
+
+ foreach(newlc, rels)
+ {
+ Relation newrel = (Relation) lfirst(newlc);
+
+ if (RelationGetRelid(newrel) == oldrelid)
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ Relation oldrel = relation_open(oldrelid,
+ ShareUpdateExclusiveLock);
+
+ delrels = lappend(delrels, oldrel);
+ }
+ }
+
+ /* And drop them. */
+ PublicationDropSequences(pubid, delrels, true);
+
+ /*
+ * Don't bother calculating the difference for adding, we'll catch and
+ * skip existing ones when doing catalog update.
+ */
+ PublicationAddSequences(pubid, rels, true, stmt);
+
+ CloseSequenceList(delrels);
+ }
+
+ CloseSequenceList(rels);
+}
+
/*
* Alter the existing publication.
*
@@ -459,8 +548,12 @@ AlterPublication(ParseState *pstate, AlterPublicationStmt *stmt)
if (stmt->options)
AlterPublicationOptions(pstate, stmt, rel, tup);
- else
+ else if (stmt->tables)
AlterPublicationTables(stmt, rel, tup);
+ else if (stmt->sequences)
+ AlterPublicationSequences(stmt, rel, tup);
+ else
+ Assert(false);
/* Cleanup. */
heap_freetuple(tup);
@@ -665,6 +758,139 @@ PublicationDropTables(Oid pubid, List *rels, bool missing_ok)
}
}
+/*
+ * Open relations specified by a RangeVar list.
+ * The returned tables are locked in ShareUpdateExclusiveLock mode in order to
+ * add them to a publication.
+ */
+static List *
+OpenSequenceList(List *sequences)
+{
+ List *relids = NIL;
+ List *rels = NIL;
+ ListCell *lc;
+
+ /*
+ * Open, share-lock, and check all the explicitly-specified relations
+ */
+ foreach(lc, sequences)
+ {
+ RangeVar *rv = castNode(RangeVar, lfirst(lc));
+ Relation rel;
+ Oid myrelid;
+
+ /* Allow query cancel in case this takes a long time */
+ CHECK_FOR_INTERRUPTS();
+
+ rel = relation_openrv(rv, ShareUpdateExclusiveLock);
+ myrelid = RelationGetRelid(rel);
+
+ /*
+ * Filter out duplicates if user specifies "foo, foo".
+ *
+ * Note that this algorithm is known to not be very efficient (O(N^2))
+ * but given that it only works on list of sequences given to us by user
+ * it's deemed acceptable.
+ */
+ if (list_member_oid(relids, myrelid))
+ {
+ relation_close(rel, ShareUpdateExclusiveLock);
+ continue;
+ }
+
+ rels = lappend(rels, rel);
+ relids = lappend_oid(relids, myrelid);
+ }
+
+ list_free(relids);
+
+ return rels;
+}
+
+/*
+ * Close all relations in the list.
+ */
+static void
+CloseSequenceList(List *rels)
+{
+ ListCell *lc;
+
+ foreach(lc, rels)
+ {
+ Relation rel = (Relation) lfirst(lc);
+
+ relation_close(rel, NoLock);
+ }
+}
+
+/*
+ * Add listed tables to the publication.
+ */
+static void
+PublicationAddSequences(Oid pubid, List *rels, bool if_not_exists,
+ AlterPublicationStmt *stmt)
+{
+ ListCell *lc;
+
+ Assert(!stmt || !stmt->for_all_sequences);
+
+ foreach(lc, rels)
+ {
+ Relation rel = (Relation) lfirst(lc);
+ ObjectAddress obj;
+
+ /* Must be owner of the table or superuser. */
+ if (!pg_class_ownercheck(RelationGetRelid(rel), GetUserId()))
+ aclcheck_error(ACLCHECK_NOT_OWNER, get_relkind_objtype(rel->rd_rel->relkind),
+ RelationGetRelationName(rel));
+
+ obj = publication_add_relation(pubid, rel, if_not_exists);
+ if (stmt)
+ {
+ EventTriggerCollectSimpleCommand(obj, InvalidObjectAddress,
+ (Node *) stmt);
+
+ InvokeObjectPostCreateHook(PublicationRelRelationId,
+ obj.objectId, 0);
+ }
+ }
+}
+
+/*
+ * Remove listed sequences from the publication.
+ */
+static void
+PublicationDropSequences(Oid pubid, List *rels, bool missing_ok)
+{
+ ObjectAddress obj;
+ ListCell *lc;
+ Oid prid;
+
+ foreach(lc, rels)
+ {
+ Relation rel = (Relation) lfirst(lc);
+ Oid relid = RelationGetRelid(rel);
+
+ prid = GetSysCacheOid2(PUBLICATIONRELMAP, Anum_pg_publication_rel_oid,
+ ObjectIdGetDatum(relid),
+ ObjectIdGetDatum(pubid));
+ if (!OidIsValid(prid))
+ {
+ if (missing_ok)
+ continue;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_UNDEFINED_OBJECT),
+ errmsg("relation \"%s\" is not part of the publication",
+ RelationGetRelationName(rel))));
+ }
+
+ ObjectAddressSet(obj, PublicationRelRelationId, prid);
+ performDeletion(&obj, DROP_CASCADE, 0);
+ }
+}
+
+
/*
* Internal workhorse for changing a publication owner
*/
diff --git a/src/backend/commands/sequence.c b/src/backend/commands/sequence.c
index 72bfdc07a4..2fc644c74c 100644
--- a/src/backend/commands/sequence.c
+++ b/src/backend/commands/sequence.c
@@ -94,7 +94,7 @@ static HTAB *seqhashtab = NULL; /* hash table for SeqTable items */
*/
static SeqTableData *last_used_seq = NULL;
-static void fill_seq_with_data(Relation rel, HeapTuple tuple);
+static void fill_seq_with_data(Relation rel, HeapTuple tuple, bool create);
static Relation lock_and_open_sequence(SeqTable seq);
static void create_seq_hashtable(void);
static void init_sequence(Oid relid, SeqTable *p_elm, Relation *p_rel);
@@ -222,7 +222,7 @@ DefineSequence(ParseState *pstate, CreateSeqStmt *seq)
/* now initialize the sequence's data */
tuple = heap_form_tuple(tupDesc, value, null);
- fill_seq_with_data(rel, tuple);
+ fill_seq_with_data(rel, tuple, true);
/* process OWNED BY if given */
if (owned_by)
@@ -327,7 +327,86 @@ ResetSequence(Oid seq_relid)
/*
* Insert the modified tuple into the new storage file.
*/
- fill_seq_with_data(seq_rel, tuple);
+ fill_seq_with_data(seq_rel, tuple, true);
+
+ /* Clear local cache so that we don't think we have cached numbers */
+ /* Note that we do not change the currval() state */
+ elm->cached = elm->last;
+
+ relation_close(seq_rel, NoLock);
+}
+
+/*
+ * Reset a sequence to its initial value.
+ *
+ * The change is made transactionally, so that on failure of the current
+ * transaction, the sequence will be restored to its previous state.
+ * We do that by creating a whole new relfilenode for the sequence; so this
+ * works much like the rewriting forms of ALTER TABLE.
+ *
+ * Caller is assumed to have acquired AccessExclusiveLock on the sequence,
+ * which must not be released until end of transaction. Caller is also
+ * responsible for permissions checking.
+ */
+void
+ResetSequence2(Oid seq_relid, int64 last_value, int64 log_cnt, bool is_called)
+{
+ Relation seq_rel;
+ SeqTable elm;
+ Form_pg_sequence_data seq;
+ Buffer buf;
+ HeapTupleData seqdatatuple;
+ HeapTuple tuple;
+
+ /*
+ * Read the old sequence. This does a bit more work than really
+ * necessary, but it's simple, and we do want to double-check that it's
+ * indeed a sequence.
+ */
+ init_sequence(seq_relid, &elm, &seq_rel);
+ (void) read_seq_tuple(seq_rel, &buf, &seqdatatuple);
+
+ /*
+ * Copy the existing sequence tuple.
+ */
+ tuple = heap_copytuple(&seqdatatuple);
+
+ /* Now we're done with the old page */
+ UnlockReleaseBuffer(buf);
+
+ /*
+ * Modify the copied tuple to execute the restart (compare the RESTART
+ * action in AlterSequence)
+ */
+ seq = (Form_pg_sequence_data) GETSTRUCT(tuple);
+ seq->last_value = last_value;
+ seq->is_called = is_called;
+ seq->log_cnt = log_cnt;
+
+ /*
+ * Create a new storage file for the sequence.
+ */
+ RelationSetNewRelfilenode(seq_rel, seq_rel->rd_rel->relpersistence);
+
+ /*
+ * Ensure sequence's relfrozenxid is at 0, since it won't contain any
+ * unfrozen XIDs. Same with relminmxid, since a sequence will never
+ * contain multixacts.
+ */
+ Assert(seq_rel->rd_rel->relfrozenxid == InvalidTransactionId);
+ Assert(seq_rel->rd_rel->relminmxid == InvalidMultiXactId);
+
+ /*
+ * Insert the modified tuple into the new storage file.
+ *
+ * XXX Maybe this should also use created=true, just like the other places
+ * calling fill_seq_with_data. That's probably needed for correct cascading
+ * replication.
+ *
+ * XXX That'd mean all fill_seq_with_data callers use created=true, making
+ * the parameter unnecessary.
+ */
+ fill_seq_with_data(seq_rel, tuple, false);
/* Clear local cache so that we don't think we have cached numbers */
/* Note that we do not change the currval() state */
@@ -340,7 +419,7 @@ ResetSequence(Oid seq_relid)
* Initialize a sequence's relation with the specified tuple as content
*/
static void
-fill_seq_with_data(Relation rel, HeapTuple tuple)
+fill_seq_with_data(Relation rel, HeapTuple tuple, bool create)
{
Buffer buf;
Page page;
@@ -378,8 +457,21 @@ fill_seq_with_data(Relation rel, HeapTuple tuple)
/* check the comment above nextval_internal()'s equivalent call. */
if (RelationNeedsWAL(rel))
+ {
GetTopTransactionId();
+ /*
+ * Ensure we have a proper XID, which will be included in the XLOG
+ * record by XLogRecordAssemble. Otherwise the first nextval() in
+ * a subxact (without any preceding changes) would get XID 0,
+ * and it'd be impossible to decide which top xact it belongs to.
+ * It'd also trigger assert in DecodeSequence.
+ *
+ * XXX Not sure if this is the best solution.
+ */
+ GetCurrentTransactionId();
+ }
+
START_CRIT_SECTION();
MarkBufferDirty(buf);
@@ -399,6 +491,7 @@ fill_seq_with_data(Relation rel, HeapTuple tuple)
XLogRegisterBuffer(0, buf, REGBUF_WILL_INIT);
xlrec.node = rel->rd_node;
+ xlrec.created = create;
XLogRegisterData((char *) &xlrec, sizeof(xl_seq_rec));
XLogRegisterData((char *) tuple->t_data, tuple->t_len);
@@ -502,7 +595,7 @@ AlterSequence(ParseState *pstate, AlterSeqStmt *stmt)
/*
* Insert the modified tuple into the new storage file.
*/
- fill_seq_with_data(seqrel, newdatatuple);
+ fill_seq_with_data(seqrel, newdatatuple, true);
}
/* process OWNED BY if given */
@@ -766,8 +859,21 @@ nextval_internal(Oid relid, bool check_permissions)
* (Have to do that here, so we're outside the critical section)
*/
if (logit && RelationNeedsWAL(seqrel))
+ {
GetTopTransactionId();
+ /*
+ * Ensure we have a proper XID, which will be included in the XLOG
+ * record by XLogRecordAssemble. Otherwise the first nextval() in
+ * a subxact (without any preceding changes) would get XID 0,
+ * and it'd be impossible to decide which top xact it belongs to.
+ * It'd also trigger assert in DecodeSequence.
+ *
+ * XXX Not sure if this is the best solution.
+ */
+ GetCurrentTransactionId();
+ }
+
/* ready to change the on-disk (or really, in-buffer) tuple */
START_CRIT_SECTION();
@@ -803,6 +909,7 @@ nextval_internal(Oid relid, bool check_permissions)
seq->log_cnt = 0;
xlrec.node = seqrel->rd_node;
+ xlrec.created = false;
XLogRegisterData((char *) &xlrec, sizeof(xl_seq_rec));
XLogRegisterData((char *) seqdatatuple.t_data, seqdatatuple.t_len);
@@ -977,8 +1084,21 @@ do_setval(Oid relid, int64 next, bool iscalled)
/* check the comment above nextval_internal()'s equivalent call. */
if (RelationNeedsWAL(seqrel))
+ {
GetTopTransactionId();
+ /*
+ * Ensure we have a proper XID, which will be included in the XLOG
+ * record by XLogRecordAssemble. Otherwise the first nextval() in
+ * a subxact (without any preceding changes) would get XID 0,
+ * and it'd be impossible to decide which top xact it belongs to.
+ * It'd also trigger assert in DecodeSequence.
+ *
+ * XXX Not sure if this is the best solution.
+ */
+ GetCurrentTransactionId();
+ }
+
/* ready to change the on-disk (or really, in-buffer) tuple */
START_CRIT_SECTION();
@@ -999,6 +1119,8 @@ do_setval(Oid relid, int64 next, bool iscalled)
XLogRegisterBuffer(0, buf, REGBUF_WILL_INIT);
xlrec.node = seqrel->rd_node;
+ xlrec.created = false;
+
XLogRegisterData((char *) &xlrec, sizeof(xl_seq_rec));
XLogRegisterData((char *) seqdatatuple.t_data, seqdatatuple.t_len);
diff --git a/src/backend/commands/subscriptioncmds.c b/src/backend/commands/subscriptioncmds.c
index 22ae982328..2d03cf011e 100644
--- a/src/backend/commands/subscriptioncmds.c
+++ b/src/backend/commands/subscriptioncmds.c
@@ -84,6 +84,7 @@ typedef struct SubOpts
} SubOpts;
static List *fetch_table_list(WalReceiverConn *wrconn, List *publications);
+static List *fetch_sequence_list(WalReceiverConn *wrconn, List *publications);
static void check_duplicates_in_publist(List *publist, Datum *datums);
static List *merge_publications(List *oldpublist, List *newpublist, bool addpub, const char *subname);
static void ReportSlotConnectionError(List *rstates, Oid subid, char *slotname, char *err);
@@ -521,6 +522,7 @@ CreateSubscription(ParseState *pstate, CreateSubscriptionStmt *stmt,
char *err;
WalReceiverConn *wrconn;
List *tables;
+ List *sequences;
ListCell *lc;
char table_state;
@@ -559,6 +561,26 @@ CreateSubscription(ParseState *pstate, CreateSubscriptionStmt *stmt,
InvalidXLogRecPtr);
}
+ /*
+ * Get the sequence list from publisher and build local sequence
+ * status info.
+ */
+ sequences = fetch_sequence_list(wrconn, publications);
+ foreach(lc, sequences)
+ {
+ RangeVar *rv = (RangeVar *) lfirst(lc);
+ Oid relid;
+
+ relid = RangeVarGetRelid(rv, AccessShareLock, false);
+
+ /* Check for supported relkind. */
+ CheckSubscriptionRelkind(get_rel_relkind(relid),
+ rv->schemaname, rv->relname);
+
+ AddSubscriptionRelState(subid, relid, table_state,
+ InvalidXLogRecPtr);
+ }
+
/*
* If requested, create permanent slot for the subscription. We
* won't use the initial snapshot for anything, so no need to
@@ -731,6 +753,10 @@ AlterSubscription_refresh(Subscription *sub, bool copy_data)
{
Oid relid = subrel_local_oids[off];
+ /* XXX ignore sequences - maybe do this in GetSubscriptionRelations? */
+ if (get_rel_relkind(relid) == RELKIND_SEQUENCE)
+ continue;
+
if (!bsearch(&relid, pubrel_local_oids,
list_length(pubrel_names), sizeof(Oid), oid_cmp))
{
@@ -822,6 +848,183 @@ AlterSubscription_refresh(Subscription *sub, bool copy_data)
ReplicationSlotDropAtPubNode(wrconn, syncslotname, true);
}
}
+
+ /*
+ * XXX now do the same thing for sequences, maybe before the preceding
+ * block, or earlier?
+ */
+
+ /* Get the table list from publisher. */
+ pubrel_names = fetch_sequence_list(wrconn, sub->publications);
+
+ /* Get local table list. */
+ subrel_states = GetSubscriptionRelations(sub->oid);
+
+ /*
+ * Build qsorted array of local table oids for faster lookup. This can
+ * potentially contain all tables in the database so speed of lookup
+ * is important.
+ */
+ subrel_local_oids = palloc(list_length(subrel_states) * sizeof(Oid));
+ off = 0;
+ foreach(lc, subrel_states)
+ {
+ SubscriptionRelState *relstate = (SubscriptionRelState *) lfirst(lc);
+
+ subrel_local_oids[off++] = relstate->relid;
+ }
+ qsort(subrel_local_oids, list_length(subrel_states),
+ sizeof(Oid), oid_cmp);
+
+ /*
+ * Rels that we want to remove from subscription and drop any slots
+ * and origins corresponding to them.
+ */
+ sub_remove_rels = palloc(list_length(subrel_states) * sizeof(SubRemoveRels));
+
+ /*
+ * Walk over the remote tables and try to match them to locally known
+ * tables. If the table is not known locally create a new state for
+ * it.
+ *
+ * Also builds array of local oids of remote tables for the next step.
+ */
+ off = 0;
+ pubrel_local_oids = palloc(list_length(pubrel_names) * sizeof(Oid));
+
+ foreach(lc, pubrel_names)
+ {
+ RangeVar *rv = (RangeVar *) lfirst(lc);
+ Oid relid;
+
+ relid = RangeVarGetRelid(rv, AccessShareLock, false);
+
+ /* Check for supported relkind. */
+ CheckSubscriptionRelkind(get_rel_relkind(relid),
+ rv->schemaname, rv->relname);
+
+ pubrel_local_oids[off++] = relid;
+
+ if (!bsearch(&relid, subrel_local_oids,
+ list_length(subrel_states), sizeof(Oid), oid_cmp))
+ {
+ AddSubscriptionRelState(sub->oid, relid,
+ copy_data ? SUBREL_STATE_INIT : SUBREL_STATE_READY,
+ InvalidXLogRecPtr);
+ ereport(DEBUG1,
+ (errmsg_internal("table \"%s.%s\" added to subscription \"%s\"",
+ rv->schemaname, rv->relname, sub->name)));
+ }
+ }
+
+ /*
+ * Next remove state for tables we should not care about anymore using
+ * the data we collected above
+ */
+ qsort(pubrel_local_oids, list_length(pubrel_names),
+ sizeof(Oid), oid_cmp);
+
+ remove_rel_len = 0;
+ for (off = 0; off < list_length(subrel_states); off++)
+ {
+ Oid relid = subrel_local_oids[off];
+
+ /* XXX ignore non-sequences - maybe do this in GetSubscriptionRelations? */
+ if (get_rel_relkind(relid) != RELKIND_SEQUENCE)
+ continue;
+
+ if (!bsearch(&relid, pubrel_local_oids,
+ list_length(pubrel_names), sizeof(Oid), oid_cmp))
+ {
+ char state;
+ XLogRecPtr statelsn;
+
+ /*
+ * Lock pg_subscription_rel with AccessExclusiveLock to
+ * prevent any race conditions with the apply worker
+ * re-launching workers at the same time this code is trying
+ * to remove those tables.
+ *
+ * Even if new worker for this particular rel is restarted it
+ * won't be able to make any progress as we hold exclusive
+ * lock on subscription_rel till the transaction end. It will
+ * simply exit as there is no corresponding rel entry.
+ *
+ * This locking also ensures that the state of rels won't
+ * change till we are done with this refresh operation.
+ */
+ if (!rel)
+ rel = table_open(SubscriptionRelRelationId, AccessExclusiveLock);
+
+ /* Last known rel state. */
+ state = GetSubscriptionRelState(sub->oid, relid, &statelsn);
+
+ sub_remove_rels[remove_rel_len].relid = relid;
+ sub_remove_rels[remove_rel_len++].state = state;
+
+ RemoveSubscriptionRel(sub->oid, relid);
+
+ logicalrep_worker_stop(sub->oid, relid);
+
+ /*
+ * For READY state, we would have already dropped the
+ * tablesync origin.
+ */
+ if (state != SUBREL_STATE_READY)
+ {
+ char originname[NAMEDATALEN];
+
+ /*
+ * Drop the tablesync's origin tracking if exists.
+ *
+ * It is possible that the origin is not yet created for
+ * tablesync worker, this can happen for the states before
+ * SUBREL_STATE_FINISHEDCOPY. The apply worker can also
+ * concurrently try to drop the origin and by this time
+ * the origin might be already removed. For these reasons,
+ * passing missing_ok = true.
+ */
+ ReplicationOriginNameForTablesync(sub->oid, relid, originname,
+ sizeof(originname));
+ replorigin_drop_by_name(originname, true, false);
+ }
+
+ ereport(DEBUG1,
+ (errmsg_internal("table \"%s.%s\" removed from subscription \"%s\"",
+ get_namespace_name(get_rel_namespace(relid)),
+ get_rel_name(relid),
+ sub->name)));
+ }
+ }
+
+ /*
+ * Drop the tablesync slots associated with removed tables. This has
+ * to be at the end because otherwise if there is an error while doing
+ * the database operations we won't be able to rollback dropped slots.
+ */
+ for (off = 0; off < remove_rel_len; off++)
+ {
+ if (sub_remove_rels[off].state != SUBREL_STATE_READY &&
+ sub_remove_rels[off].state != SUBREL_STATE_SYNCDONE)
+ {
+ char syncslotname[NAMEDATALEN] = {0};
+
+ /*
+ * For READY/SYNCDONE states we know the tablesync slot has
+ * already been dropped by the tablesync worker.
+ *
+ * For other states, there is no certainty, maybe the slot
+ * does not exist yet. Also, if we fail after removing some of
+ * the slots, next time, it will again try to drop already
+ * dropped slots and fail. For these reasons, we allow
+ * missing_ok = true for the drop.
+ */
+ ReplicationSlotNameForTablesync(sub->oid, sub_remove_rels[off].relid,
+ syncslotname, sizeof(syncslotname));
+ ReplicationSlotDropAtPubNode(wrconn, syncslotname, true);
+ }
+ }
+
}
PG_FINALLY();
{
@@ -1635,6 +1838,75 @@ fetch_table_list(WalReceiverConn *wrconn, List *publications)
return tablelist;
}
+/*
+ * Get the list of sequences which belong to specified publications on the
+ * publisher connection.
+ */
+static List *
+fetch_sequence_list(WalReceiverConn *wrconn, List *publications)
+{
+ WalRcvExecResult *res;
+ StringInfoData cmd;
+ TupleTableSlot *slot;
+ Oid tableRow[2] = {TEXTOID, TEXTOID};
+ ListCell *lc;
+ bool first;
+ List *tablelist = NIL;
+
+ Assert(list_length(publications) > 0);
+
+ initStringInfo(&cmd);
+ appendStringInfoString(&cmd, "SELECT DISTINCT s.schemaname, s.sequencename\n"
+ " FROM pg_catalog.pg_publication_sequences s\n"
+ " WHERE s.pubname IN (");
+ first = true;
+ foreach(lc, publications)
+ {
+ char *pubname = strVal(lfirst(lc));
+
+ if (first)
+ first = false;
+ else
+ appendStringInfoString(&cmd, ", ");
+
+ appendStringInfoString(&cmd, quote_literal_cstr(pubname));
+ }
+ appendStringInfoChar(&cmd, ')');
+
+ res = walrcv_exec(wrconn, cmd.data, 2, tableRow);
+ pfree(cmd.data);
+
+ if (res->status != WALRCV_OK_TUPLES)
+ ereport(ERROR,
+ (errmsg("could not receive list of replicated tables from the publisher: %s",
+ res->err)));
+
+ /* Process tables. */
+ slot = MakeSingleTupleTableSlot(res->tupledesc, &TTSOpsMinimalTuple);
+ while (tuplestore_gettupleslot(res->tuplestore, true, false, slot))
+ {
+ char *nspname;
+ char *relname;
+ bool isnull;
+ RangeVar *rv;
+
+ nspname = TextDatumGetCString(slot_getattr(slot, 1, &isnull));
+ Assert(!isnull);
+ relname = TextDatumGetCString(slot_getattr(slot, 2, &isnull));
+ Assert(!isnull);
+
+ rv = makeRangeVar(nspname, relname, -1);
+ tablelist = lappend(tablelist, rv);
+
+ ExecClearTuple(slot);
+ }
+ ExecDropSingleTupleTableSlot(slot);
+
+ walrcv_clear_result(res);
+
+ return tablelist;
+}
+
/*
* This is to report the connection failure while dropping replication slots.
* Here, we report the WARNING for all tablesync slots so that user can drop
diff --git a/src/backend/executor/execReplication.c b/src/backend/executor/execReplication.c
index 574d7d27fd..6f0d42d551 100644
--- a/src/backend/executor/execReplication.c
+++ b/src/backend/executor/execReplication.c
@@ -608,7 +608,7 @@ void
CheckSubscriptionRelkind(char relkind, const char *nspname,
const char *relname)
{
- if (relkind != RELKIND_RELATION && relkind != RELKIND_PARTITIONED_TABLE)
+ if (relkind != RELKIND_RELATION && relkind != RELKIND_PARTITIONED_TABLE && relkind != RELKIND_SEQUENCE)
ereport(ERROR,
(errcode(ERRCODE_WRONG_OBJECT_TYPE),
errmsg("cannot use relation \"%s.%s\" as logical replication target",
diff --git a/src/backend/nodes/copyfuncs.c b/src/backend/nodes/copyfuncs.c
index 29020c908e..04bb9a27a8 100644
--- a/src/backend/nodes/copyfuncs.c
+++ b/src/backend/nodes/copyfuncs.c
@@ -4834,8 +4834,10 @@ _copyAlterPublicationStmt(const AlterPublicationStmt *from)
COPY_STRING_FIELD(pubname);
COPY_NODE_FIELD(options);
COPY_NODE_FIELD(tables);
+ COPY_NODE_FIELD(sequences);
COPY_SCALAR_FIELD(for_all_tables);
- COPY_SCALAR_FIELD(tableAction);
+ COPY_SCALAR_FIELD(for_all_sequences);
+ COPY_SCALAR_FIELD(action);
return newnode;
}
diff --git a/src/backend/nodes/equalfuncs.c b/src/backend/nodes/equalfuncs.c
index 8a1762000c..ebf48f88f1 100644
--- a/src/backend/nodes/equalfuncs.c
+++ b/src/backend/nodes/equalfuncs.c
@@ -2308,8 +2308,10 @@ _equalAlterPublicationStmt(const AlterPublicationStmt *a,
COMPARE_STRING_FIELD(pubname);
COMPARE_NODE_FIELD(options);
COMPARE_NODE_FIELD(tables);
+ COMPARE_NODE_FIELD(sequences);
COMPARE_SCALAR_FIELD(for_all_tables);
- COMPARE_SCALAR_FIELD(tableAction);
+ COMPARE_SCALAR_FIELD(for_all_sequences);
+ COMPARE_SCALAR_FIELD(action);
return true;
}
diff --git a/src/backend/parser/gram.y b/src/backend/parser/gram.y
index 39a2849eba..d42be08c0c 100644
--- a/src/backend/parser/gram.y
+++ b/src/backend/parser/gram.y
@@ -434,6 +434,7 @@ static Node *makeRecursiveViewSelect(char *relname, List *aliases, Node *query);
%type <node> group_by_item empty_grouping_set rollup_clause cube_clause
%type <node> grouping_sets_clause
%type <node> opt_publication_for_tables publication_for_tables
+/* %type <node> opt_publication_for_sequences publication_for_sequences */
%type <list> opt_fdw_options fdw_options
%type <defelt> fdw_option
@@ -9596,6 +9597,7 @@ AlterOwnerStmt: ALTER AGGREGATE aggregate_with_argtypes OWNER TO RoleSpec
*****************************************************************************/
CreatePublicationStmt:
+ /* CREATE PUBLICATION name opt_publication_for_tables opt_publication_for_sequences opt_definition */
CREATE PUBLICATION name opt_publication_for_tables opt_definition
{
CreatePublicationStmt *n = makeNode(CreatePublicationStmt);
@@ -9630,6 +9632,12 @@ publication_for_tables:
}
;
+/*
+ * FIXME
+ *
+ * opt_publication_for_sequences and publication_for_sequences should be
+ * copies for sequences
+ */
/*****************************************************************************
*
@@ -9641,6 +9649,12 @@ publication_for_tables:
*
* ALTER PUBLICATION name SET TABLE table [, table2]
*
+ * ALTER PUBLICATION name ADD SEQUENCE sequence [, sequence2]
+ *
+ * ALTER PUBLICATION name DROP SEQUENCE sequence [, sequence2]
+ *
+ * ALTER PUBLICATION name SET SEQUENCE sequence [, sequence2]
+ *
*****************************************************************************/
AlterPublicationStmt:
@@ -9656,7 +9670,7 @@ AlterPublicationStmt:
AlterPublicationStmt *n = makeNode(AlterPublicationStmt);
n->pubname = $3;
n->tables = $6;
- n->tableAction = DEFELEM_ADD;
+ n->action = DEFELEM_ADD;
$$ = (Node *)n;
}
| ALTER PUBLICATION name SET TABLE relation_expr_list
@@ -9664,7 +9678,7 @@ AlterPublicationStmt:
AlterPublicationStmt *n = makeNode(AlterPublicationStmt);
n->pubname = $3;
n->tables = $6;
- n->tableAction = DEFELEM_SET;
+ n->action = DEFELEM_SET;
$$ = (Node *)n;
}
| ALTER PUBLICATION name DROP TABLE relation_expr_list
@@ -9672,7 +9686,31 @@ AlterPublicationStmt:
AlterPublicationStmt *n = makeNode(AlterPublicationStmt);
n->pubname = $3;
n->tables = $6;
- n->tableAction = DEFELEM_DROP;
+ n->action = DEFELEM_DROP;
+ $$ = (Node *)n;
+ }
+ | ALTER PUBLICATION name ADD_P SEQUENCE relation_expr_list
+ {
+ AlterPublicationStmt *n = makeNode(AlterPublicationStmt);
+ n->pubname = $3;
+ n->sequences = $6;
+ n->action = DEFELEM_ADD;
+ $$ = (Node *)n;
+ }
+ | ALTER PUBLICATION name SET SEQUENCE relation_expr_list
+ {
+ AlterPublicationStmt *n = makeNode(AlterPublicationStmt);
+ n->pubname = $3;
+ n->sequences = $6;
+ n->action = DEFELEM_SET;
+ $$ = (Node *)n;
+ }
+ | ALTER PUBLICATION name DROP SEQUENCE relation_expr_list
+ {
+ AlterPublicationStmt *n = makeNode(AlterPublicationStmt);
+ n->pubname = $3;
+ n->sequences = $6;
+ n->action = DEFELEM_DROP;
$$ = (Node *)n;
}
;
diff --git a/src/backend/replication/logical/decode.c b/src/backend/replication/logical/decode.c
index 2874dc0612..98d0edefb0 100644
--- a/src/backend/replication/logical/decode.c
+++ b/src/backend/replication/logical/decode.c
@@ -42,6 +42,7 @@
#include "replication/reorderbuffer.h"
#include "replication/snapbuild.h"
#include "storage/standby.h"
+#include "commands/sequence.h"
typedef struct XLogRecordBuffer
{
@@ -74,10 +75,11 @@ static void DecodeAbort(LogicalDecodingContext *ctx, XLogRecordBuffer *buf,
bool two_phase);
static void DecodePrepare(LogicalDecodingContext *ctx, XLogRecordBuffer *buf,
xl_xact_parsed_prepare *parsed);
-
+static void DecodeSequence(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
/* common function to decode tuples */
static void DecodeXLogTuple(char *data, Size len, ReorderBufferTupleBuf *tup);
+static void DecodeSeqTuple(char *data, Size len, ReorderBufferTupleBuf *tuple);
/* helper functions for decoding transactions */
static inline bool FilterPrepare(LogicalDecodingContext *ctx,
@@ -158,6 +160,10 @@ LogicalDecodingProcessRecord(LogicalDecodingContext *ctx, XLogReaderState *recor
DecodeLogicalMsgOp(ctx, &buf);
break;
+ case RM_SEQ_ID:
+ DecodeSequence(ctx, &buf);
+ break;
+
/*
* Rmgrs irrelevant for logical decoding; they describe stuff not
* represented in logical decoding. Add new rmgrs in rmgrlist.h's
@@ -173,7 +179,6 @@ LogicalDecodingProcessRecord(LogicalDecodingContext *ctx, XLogReaderState *recor
case RM_HASH_ID:
case RM_GIN_ID:
case RM_GIST_ID:
- case RM_SEQ_ID:
case RM_SPGIST_ID:
case RM_BRIN_ID:
case RM_COMMIT_TS_ID:
@@ -1312,3 +1317,125 @@ DecodeTXNNeedSkip(LogicalDecodingContext *ctx, XLogRecordBuffer *buf,
(txn_dbid != InvalidOid && txn_dbid != ctx->slot->data.database) ||
ctx->fast_forward || FilterByOrigin(ctx, origin_id));
}
+
+/*
+ * Decode Sequence Tuple
+ */
+static void
+DecodeSeqTuple(char *data, Size len, ReorderBufferTupleBuf *tuple)
+{
+ int datalen = len - sizeof(xl_seq_rec) - SizeofHeapTupleHeader;
+
+ Assert(datalen >= 0);
+
+ tuple->tuple.t_len = datalen + SizeofHeapTupleHeader;;
+
+ ItemPointerSetInvalid(&tuple->tuple.t_self);
+
+ tuple->tuple.t_tableOid = InvalidOid;
+
+ memcpy(((char *) tuple->tuple.t_data),
+ data + sizeof(xl_seq_rec),
+ SizeofHeapTupleHeader);
+
+ memcpy(((char *) tuple->tuple.t_data) + SizeofHeapTupleHeader,
+ data + sizeof(xl_seq_rec) + SizeofHeapTupleHeader,
+ datalen);
+}
+
+/*
+ * Handle sequence decode
+ *
+ * Decoding sequences is a bit tricky, because while most sequence actions
+ * are non-transactional (not subject to rollback), some need to be handled
+ * as transactional.
+ *
+ * By default, a sequence increment is non-transactional - we must not queue
+ * it in a transaction as other changes, because the transaction might get
+ * rolled back and we'd discard the increment. The downstream would not be
+ * notified about the increment, which is wrong.
+ *
+ * On the other hand, the sequence may be created in a transaction. In this
+ * case we *should* queue the change as other changes in the transaction,
+ * because we don't want to send the increments for unknown sequence to the
+ * plugin - it might get confused about which sequence it's related to etc.
+ */
+static void
+DecodeSequence(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
+{
+ SnapBuild *builder = ctx->snapshot_builder;
+ ReorderBufferTupleBuf *tuplebuf;
+ RelFileNode target_node;
+ XLogReaderState *r = buf->record;
+ char *tupledata = NULL;
+ Size tuplelen;
+ Size datalen = 0;
+ TransactionId xid = XLogRecGetXid(r);
+ uint8 info = XLogRecGetInfo(buf->record) & ~XLR_INFO_MASK;
+ xl_seq_rec *xlrec;
+ Snapshot snapshot;
+ RepOriginId origin_id = XLogRecGetOrigin(r);
+ bool transactional;
+
+ /* only decode changes flagged with XLOG_SEQ_LOG */
+ if (info != XLOG_SEQ_LOG)
+ elog(ERROR, "unexpected RM_SEQ_ID record type: %u", info);
+
+ /*
+ * If we don't have snapshot or we are just fast-forwarding, there is no
+ * point in decoding messages.
+ */
+ if (SnapBuildCurrentState(builder) < SNAPBUILD_FULL_SNAPSHOT ||
+ ctx->fast_forward)
+ return;
+
+ /* only interested in our database */
+ XLogRecGetBlockTag(r, 0, &target_node, NULL, NULL);
+ if (target_node.dbNode != ctx->slot->data.database)
+ return;
+
+ /* output plugin doesn't look for this origin, no need to queue */
+ if (FilterByOrigin(ctx, XLogRecGetOrigin(r)))
+ return;
+
+ tupledata = XLogRecGetData(r);
+ datalen = XLogRecGetDataLen(r);
+ tuplelen = datalen - SizeOfHeapHeader - sizeof(xl_seq_rec);
+
+ /* extract the WAL record, with "created" flag */
+ xlrec = (xl_seq_rec *) XLogRecGetData(r);
+
+ /* XXX how could we have sequence change without data? */
+ if(!datalen || !tupledata)
+ return;
+
+ tuplebuf = ReorderBufferGetTupleBuf(ctx->reorder, tuplelen);
+ DecodeSeqTuple(tupledata, datalen, tuplebuf);
+
+ /*
+ * Should we handle the sequence increment as transactional or not?
+ *
+ * If the sequence was created in a still-running transaction, treat
+ * it as transactional and queue the increments. Otherwise it needs
+ * to be treated as non-transactional, in which case we send it to
+ * the plugin right away.
+ */
+ transactional = ReorderBufferSequenceIsTransactional(ctx->reorder,
+ target_node,
+ xlrec->created);
+
+ /* Skip the change if already processed (per the snapshot). */
+ if (transactional &&
+ !SnapBuildProcessChange(builder, xid, buf->origptr))
+ return;
+ else if (!transactional &&
+ (SnapBuildCurrentState(builder) != SNAPBUILD_CONSISTENT ||
+ SnapBuildXactNeedsSkip(builder, buf->origptr)))
+ return;
+
+ /* Queue the increment (or send immediately if not transactional). */
+ snapshot = SnapBuildGetOrBuildSnapshot(builder, xid);
+ ReorderBufferQueueSequence(ctx->reorder, xid, snapshot, buf->endptr,
+ origin_id, target_node, transactional,
+ xlrec->created, tuplebuf);
+}
diff --git a/src/backend/replication/logical/logical.c b/src/backend/replication/logical/logical.c
index 8dcb564af8..a74dd75fde 100644
--- a/src/backend/replication/logical/logical.c
+++ b/src/backend/replication/logical/logical.c
@@ -73,6 +73,10 @@ static void truncate_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
static void message_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
XLogRecPtr message_lsn, bool transactional,
const char *prefix, Size message_size, const char *message);
+static void sequence_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ bool transactional, bool created,
+ int64 last_value, int64 log_cnt, bool is_called);
/* streaming callbacks */
static void stream_start_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
@@ -217,6 +221,7 @@ StartupDecodingContext(List *output_plugin_options,
ctx->reorder->apply_truncate = truncate_cb_wrapper;
ctx->reorder->commit = commit_cb_wrapper;
ctx->reorder->message = message_cb_wrapper;
+ ctx->reorder->sequence = sequence_cb_wrapper;
/*
* To support streaming, we require start/stop/abort/commit/change
@@ -1198,6 +1203,43 @@ message_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
error_context_stack = errcallback.previous;
}
+static void
+sequence_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ bool transactional, bool created,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ LogicalDecodingContext *ctx = cache->private_data;
+ LogicalErrorCallbackState state;
+ ErrorContextCallback errcallback;
+
+ Assert(!ctx->fast_forward);
+
+ if (ctx->callbacks.sequence_cb == NULL)
+ return;
+
+ /* Push callback + info on the error context stack */
+ state.ctx = ctx;
+ state.callback_name = "sequence";
+ state.report_location = sequence_lsn;
+ errcallback.callback = output_plugin_error_callback;
+ errcallback.arg = (void *) &state;
+ errcallback.previous = error_context_stack;
+ error_context_stack = &errcallback;
+
+ /* set output state */
+ ctx->accept_writes = true;
+ ctx->write_xid = txn != NULL ? txn->xid : InvalidTransactionId;
+ ctx->write_location = sequence_lsn;
+
+ /* do the actual work: call callback */
+ ctx->callbacks.sequence_cb(ctx, txn, sequence_lsn, rel, transactional,
+ created, last_value, log_cnt, is_called);
+
+ /* Pop the error context stack */
+ error_context_stack = errcallback.previous;
+}
+
static void
stream_start_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
XLogRecPtr first_lsn)
diff --git a/src/backend/replication/logical/proto.c b/src/backend/replication/logical/proto.c
index 2d774567e0..d42c749cfa 100644
--- a/src/backend/replication/logical/proto.c
+++ b/src/backend/replication/logical/proto.c
@@ -622,6 +622,58 @@ logicalrep_write_message(StringInfo out, TransactionId xid, XLogRecPtr lsn,
pq_sendbytes(out, message, sz);
}
+/*
+ * Write SEQUENCE to stream
+ */
+void
+logicalrep_write_sequence(StringInfo out, Relation rel, TransactionId xid,
+ XLogRecPtr lsn, bool transactional, bool created,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ uint8 flags = 0;
+ char *relname;
+
+ pq_sendbyte(out, LOGICAL_REP_MSG_SEQUENCE);
+
+ /* transaction ID (if not valid, we're not streaming) */
+ if (TransactionIdIsValid(xid))
+ pq_sendint32(out, xid);
+
+ pq_sendint8(out, flags);
+ pq_sendint64(out, lsn);
+
+ logicalrep_write_namespace(out, RelationGetNamespace(rel));
+ relname = RelationGetRelationName(rel);
+ pq_sendstring(out, relname);
+
+ pq_sendint8(out, transactional);
+ pq_sendint8(out, created);
+ pq_sendint64(out, last_value);
+ pq_sendint64(out, log_cnt);
+ pq_sendint8(out, is_called);
+}
+
+/*
+ * Read SEQUENCE from the stream.
+ */
+void
+logicalrep_read_sequence(StringInfo in, LogicalRepSequence *seqdata)
+{
+ /* XXX skipping flags and lsn */
+ pq_getmsgint(in, 1);
+ pq_getmsgint64(in);
+
+ /* Read relation name from stream */
+ seqdata->nspname = pstrdup(logicalrep_read_namespace(in));
+ seqdata->seqname = pstrdup(pq_getmsgstring(in));
+
+ seqdata->transactional = pq_getmsgint(in, 1);
+ seqdata->created = pq_getmsgint(in, 1);
+ seqdata->last_value = pq_getmsgint64(in);
+ seqdata->log_cnt = pq_getmsgint64(in);
+ seqdata->is_called = pq_getmsgint(in, 1);
+}
+
/*
* Write relation description to the output stream.
*/
diff --git a/src/backend/replication/logical/reorderbuffer.c b/src/backend/replication/logical/reorderbuffer.c
index 7378beb684..f99e45f22d 100644
--- a/src/backend/replication/logical/reorderbuffer.c
+++ b/src/backend/replication/logical/reorderbuffer.c
@@ -116,6 +116,13 @@ typedef struct ReorderBufferTXNByIdEnt
ReorderBufferTXN *txn;
} ReorderBufferTXNByIdEnt;
+/* entry for hash table we use to track sequences created in running xacts */
+typedef struct ReorderBufferSequenceEnt
+{
+ RelFileNode rnode;
+ TransactionId xid;
+} ReorderBufferSequenceEnt;
+
/* data structures for (relfilenode, ctid) => (cmin, cmax) mapping */
typedef struct ReorderBufferTupleCidKey
{
@@ -338,6 +345,14 @@ ReorderBufferAllocate(void)
buffer->by_txn = hash_create("ReorderBufferByXid", 1000, &hash_ctl,
HASH_ELEM | HASH_BLOBS | HASH_CONTEXT);
+ /* hash table of sequences, mapping relfilenode to XID of transaction */
+ hash_ctl.keysize = sizeof(RelFileNode);
+ hash_ctl.entrysize = sizeof(ReorderBufferSequenceEnt);
+ hash_ctl.hcxt = buffer->context;
+
+ buffer->sequences = hash_create("ReorderBufferSequenceHash", 1000, &hash_ctl,
+ HASH_ELEM | HASH_BLOBS | HASH_CONTEXT);
+
buffer->by_txn_last_xid = InvalidTransactionId;
buffer->by_txn_last_txn = NULL;
@@ -528,6 +543,13 @@ ReorderBufferReturnChange(ReorderBuffer *rb, ReorderBufferChange *change,
case REORDER_BUFFER_CHANGE_INTERNAL_COMMAND_ID:
case REORDER_BUFFER_CHANGE_INTERNAL_TUPLECID:
break;
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
+ if (change->data.sequence.tuple)
+ {
+ ReorderBufferReturnTupleBuf(rb, change->data.sequence.tuple);
+ change->data.sequence.tuple = NULL;
+ }
+ break;
}
pfree(change);
@@ -856,6 +878,212 @@ ReorderBufferQueueMessage(ReorderBuffer *rb, TransactionId xid,
}
}
+/*
+ * Treat the sequence increment as transactional?
+ */
+bool
+ReorderBufferSequenceIsTransactional(ReorderBuffer *rb,
+ RelFileNode rnode, bool created)
+{
+ bool found = false;
+
+ if (created)
+ return true;
+
+ hash_search(rb->sequences,
+ (void *) &rnode,
+ HASH_FIND,
+ &found);
+
+ return found;
+}
+
+
+/*
+ * A transactional sequence increment is queued to be processed upon commit
+ * and a non-transactional increment gets processed immediately.
+ *
+ * A sequence update may be both transactional and non-transactional. When
+ * created in a running transaction, treat it as transactional and queue
+ * the change in it. Otherwise treat it as non-transactional, so that we
+ * don't forget the increment in case of a rollback.
+ */
+void
+ReorderBufferQueueSequence(ReorderBuffer *rb, TransactionId xid,
+ Snapshot snapshot, XLogRecPtr lsn, RepOriginId origin_id,
+ RelFileNode rnode, bool transactional, bool created,
+ ReorderBufferTupleBuf *tuplebuf)
+{
+ /*
+ * Change needs to be handled as transactional, because the sequence was
+ * created in a transaction that is still running. In that case all the
+ * changes need to be queued in that transaction, we must not send them
+ * to the downstream until the transaction commits.
+ *
+ * There's a bit of a trouble with subtransactions - we can't queue it
+ * into the subxact, because it might be rolled back and we'd lose the
+ * increment. We need to queue it into the same (sub)xact that created
+ * the sequence, which is why we track the XID in the hash table.
+ */
+ if (transactional)
+ {
+ MemoryContext oldcontext;
+ ReorderBufferChange *change;
+
+ /* lookup sequence by relfilenode */
+ ReorderBufferSequenceEnt *ent;
+ bool found;
+
+ /* transactional changes require a transaction */
+ Assert(xid != InvalidTransactionId);
+
+ /* search the lookup table (we ignore the return value, found is enough) */
+ ent = hash_search(rb->sequences,
+ (void *) &rnode,
+ created ? HASH_ENTER : HASH_FIND,
+ &found);
+
+ /*
+ * If this is the "create" increment, we must not have found any
+ * pre-existing entry in the hash table (i.e. there must not be
+ * any conflicting sequence).
+ */
+ Assert(!(created && found));
+
+ /* But we must have either created or found an existing entry. */
+ Assert(created || found);
+
+ /*
+ * When creating the sequence, remember the XID of the transaction
+ * that created id.
+ */
+ if (created)
+ ent->xid = xid;
+
+ /* XXX Maybe check that we're still in the same top-level xact? */
+
+ /* OK, allocate and queue the change */
+ oldcontext = MemoryContextSwitchTo(rb->context);
+
+ change = ReorderBufferGetChange(rb);
+
+ change->action = REORDER_BUFFER_CHANGE_SEQUENCE;
+ change->origin_id = origin_id;
+
+ memcpy(&change->data.sequence.relnode, &rnode, sizeof(RelFileNode));
+
+ change->data.sequence.created = created;
+ change->data.sequence.tuple = tuplebuf;
+
+ /* add it to the same subxact that created the sequence */
+ ReorderBufferQueueChange(rb, ent->xid, lsn, change, false);
+
+ MemoryContextSwitchTo(oldcontext);
+ }
+ else
+ {
+ /*
+ * This increment is for a sequence that was not created in any
+ * running transaction, so we treat it as non-transactional and
+ * just send it to the output plugin directly.
+ */
+ ReorderBufferTXN *txn = NULL;
+ volatile Snapshot snapshot_now = snapshot;
+ bool using_subtxn;
+
+#ifdef USE_ASSERT_CHECKING
+ /* All "creates" have to be handled as transactional. */
+ Assert(!created);
+
+ /* Make sure the sequence is not in the hash table. */
+ {
+ bool found;
+ hash_search(rb->sequences,
+ (void *) &rnode,
+ HASH_FIND, &found);
+ Assert(!found);
+ }
+#endif
+
+ if (xid != InvalidTransactionId)
+ txn = ReorderBufferTXNByXid(rb, xid, true, NULL, lsn, true);
+
+ /* setup snapshot to allow catalog access */
+ SetupHistoricSnapshot(snapshot_now, NULL);
+
+ /*
+ * Decoding needs access to syscaches et al., which in turn use
+ * heavyweight locks and such. Thus we need to have enough state around to
+ * keep track of those. The easiest way is to simply use a transaction
+ * internally. That also allows us to easily enforce that nothing writes
+ * to the database by checking for xid assignments.
+ *
+ * When we're called via the SQL SRF there's already a transaction
+ * started, so start an explicit subtransaction there.
+ */
+ using_subtxn = IsTransactionOrTransactionBlock();
+
+ PG_TRY();
+ {
+ Relation relation;
+ HeapTuple tuple;
+ bool isnull;
+ int64 last_value, log_cnt;
+ bool is_called;
+ Oid reloid;
+
+ if (using_subtxn)
+ BeginInternalSubTransaction("sequence");
+ else
+ StartTransactionCommand();
+
+ reloid = RelidByRelfilenode(rnode.spcNode, rnode.relNode);
+
+ if (reloid == InvalidOid)
+ elog(ERROR, "could not map filenode \"%s\" to relation OID",
+ relpathperm(rnode,
+ MAIN_FORKNUM));
+
+ relation = RelationIdGetRelation(reloid);
+ tuple = &tuplebuf->tuple;
+
+ /*
+ * Extract the internal sequence values, describing the state.
+ *
+ * XXX Seems a bit strange to access it directly. Maybe there's
+ * a better / more correct way?
+ */
+ last_value = heap_getattr(tuple, 1, RelationGetDescr(relation), &isnull);
+ log_cnt = heap_getattr(tuple, 2, RelationGetDescr(relation), &isnull);
+ is_called = heap_getattr(tuple, 3, RelationGetDescr(relation), &isnull);
+
+ rb->sequence(rb, txn, lsn, relation, transactional, created,
+ last_value, log_cnt, is_called);
+
+ RelationClose(relation);
+
+ TeardownHistoricSnapshot(false);
+
+ AbortCurrentTransaction();
+
+ if (using_subtxn)
+ RollbackAndReleaseCurrentSubTransaction();
+ }
+ PG_CATCH();
+ {
+ TeardownHistoricSnapshot(true);
+
+ AbortCurrentTransaction();
+
+ if (using_subtxn)
+ RollbackAndReleaseCurrentSubTransaction();
+
+ PG_RE_THROW();
+ }
+ PG_END_TRY();
+ }
+}
+
/*
* AssertTXNLsnOrder
* Verify LSN ordering of transaction lists in the reorderbuffer
@@ -1532,6 +1760,31 @@ ReorderBufferCleanupTXN(ReorderBuffer *rb, ReorderBufferTXN *txn)
&found);
Assert(found);
+ /*
+ * Remove sequences created in this transaction (if any).
+ *
+ * There's no way to search by XID, so we simply do a seqscan of all
+ * the entries in the hash table. Hopefully there are only a couple
+ * entries in most cases - people generally don't create many new
+ * sequences over and over.
+ */
+ {
+ HASH_SEQ_STATUS scan_status;
+ ReorderBufferSequenceEnt *ent;
+
+ hash_seq_init(&scan_status, rb->sequences);
+ while ((ent = (ReorderBufferSequenceEnt *) hash_seq_search(&scan_status)) != NULL)
+ {
+ /* skip sequences not from this transaction */
+ if (ent->xid != txn->xid)
+ continue;
+
+ (void) hash_search(rb->sequences,
+ (void *) &(ent->rnode),
+ HASH_REMOVE, NULL);
+ }
+ }
+
/* remove entries spilled to disk */
if (rbtxn_is_serialized(txn))
ReorderBufferRestoreCleanup(rb, txn);
@@ -1947,6 +2200,39 @@ ReorderBufferApplyMessage(ReorderBuffer *rb, ReorderBufferTXN *txn,
change->data.msg.message);
}
+/*
+ * Helper function for ReorderBufferProcessTXN for applying sequences.
+ */
+static inline void
+ReorderBufferApplySequence(ReorderBuffer *rb, ReorderBufferTXN *txn,
+ Relation relation, ReorderBufferChange *change,
+ bool streaming)
+{
+ HeapTuple tuple;
+ bool isnull;
+ int64 last_value, log_cnt;
+ bool is_called;
+
+ /* FIXME support streaming */
+ Assert(!streaming);
+
+ tuple = &change->data.sequence.tuple->tuple;
+
+ /*
+ * Extract the internal sequence values, describing the state.
+ *
+ * XXX Seems a bit strange to access it directly. Maybe there's
+ * a better / more correct way?
+ */
+ last_value = heap_getattr(tuple, 1, RelationGetDescr(relation), &isnull);
+ log_cnt = heap_getattr(tuple, 2, RelationGetDescr(relation), &isnull);
+ is_called = heap_getattr(tuple, 3, RelationGetDescr(relation), &isnull);
+
+ rb->sequence(rb, txn, change->lsn, relation, true, /* gotta be transactional */
+ change->data.sequence.created,
+ last_value, log_cnt, is_called);
+}
+
/*
* Function to store the command id and snapshot at the end of the current
* stream so that we can reuse the same while sending the next stream.
@@ -2389,6 +2675,31 @@ ReorderBufferProcessTXN(ReorderBuffer *rb, ReorderBufferTXN *txn,
case REORDER_BUFFER_CHANGE_INTERNAL_TUPLECID:
elog(ERROR, "tuplecid value in changequeue");
break;
+
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
+ Assert(snapshot_now);
+
+ reloid = RelidByRelfilenode(change->data.sequence.relnode.spcNode,
+ change->data.sequence.relnode.relNode);
+
+ if (reloid == InvalidOid)
+ elog(ERROR, "could not map filenode \"%s\" to relation OID",
+ relpathperm(change->data.sequence.relnode,
+ MAIN_FORKNUM));
+
+ relation = RelationIdGetRelation(reloid);
+
+ if (!RelationIsValid(relation))
+ elog(ERROR, "could not open relation with OID %u (for filenode \"%s\")",
+ reloid,
+ relpathperm(change->data.sequence.relnode,
+ MAIN_FORKNUM));
+
+ if (RelationIsLogicallyLogged(relation))
+ ReorderBufferApplySequence(rb, txn, relation, change, streaming);
+
+ RelationClose(relation);
+ break;
}
}
@@ -3776,6 +4087,39 @@ ReorderBufferSerializeChange(ReorderBuffer *rb, ReorderBufferTXN *txn,
memcpy(data, change->data.truncate.relids, size);
data += size;
+ break;
+ }
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
+ {
+ char *data;
+ ReorderBufferTupleBuf *tup;
+ Size len = 0;
+
+ tup = change->data.sequence.tuple;
+
+ if (tup)
+ {
+ sz += sizeof(HeapTupleData);
+ len = tup->tuple.t_len;
+ sz += len;
+ }
+
+ /* make sure we have enough space */
+ ReorderBufferSerializeReserve(rb, sz);
+
+ data = ((char *) rb->outbuf) + sizeof(ReorderBufferDiskChange);
+ /* might have been reallocated above */
+ ondisk = (ReorderBufferDiskChange *) rb->outbuf;
+
+ if (len)
+ {
+ memcpy(data, &tup->tuple, sizeof(HeapTupleData));
+ data += sizeof(HeapTupleData);
+
+ memcpy(data, tup->tuple.t_data, len);
+ data += len;
+ }
+
break;
}
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM:
@@ -4040,6 +4384,22 @@ ReorderBufferChangeSize(ReorderBufferChange *change)
{
sz += sizeof(Oid) * change->data.truncate.nrelids;
+ break;
+ }
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
+ {
+ ReorderBufferTupleBuf *tup;
+ Size len = 0;
+
+ tup = change->data.sequence.tuple;
+
+ if (tup)
+ {
+ sz += sizeof(HeapTupleData);
+ len = tup->tuple.t_len;
+ sz += len;
+ }
+
break;
}
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM:
@@ -4341,6 +4701,30 @@ ReorderBufferRestoreChange(ReorderBuffer *rb, ReorderBufferTXN *txn,
break;
}
+
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
+ if (change->data.sequence.tuple)
+ {
+ uint32 tuplelen = ((HeapTuple) data)->t_len;
+
+ change->data.sequence.tuple =
+ ReorderBufferGetTupleBuf(rb, tuplelen - SizeofHeapTupleHeader);
+
+ /* restore ->tuple */
+ memcpy(&change->data.sequence.tuple->tuple, data,
+ sizeof(HeapTupleData));
+ data += sizeof(HeapTupleData);
+
+ /* reset t_data pointer into the new tuplebuf */
+ change->data.sequence.tuple->tuple.t_data =
+ ReorderBufferTupleBufData(change->data.sequence.tuple);
+
+ /* restore tuple data itself */
+ memcpy(change->data.sequence.tuple->tuple.t_data, data, tuplelen);
+ data += tuplelen;
+ }
+ break;
+
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM:
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_ABORT:
case REORDER_BUFFER_CHANGE_INTERNAL_COMMAND_ID:
diff --git a/src/backend/replication/logical/tablesync.c b/src/backend/replication/logical/tablesync.c
index f07983a43c..24369a1522 100644
--- a/src/backend/replication/logical/tablesync.c
+++ b/src/backend/replication/logical/tablesync.c
@@ -100,6 +100,7 @@
#include "catalog/pg_subscription_rel.h"
#include "catalog/pg_type.h"
#include "commands/copy.h"
+#include "commands/sequence.h"
#include "miscadmin.h"
#include "parser/parse_relation.h"
#include "pgstat.h"
@@ -357,6 +358,12 @@ process_syncing_tables_for_sync(XLogRecPtr current_lsn)
*
* If the synchronization position is reached (SYNCDONE), then the table can
* be marked as READY and is no longer tracked.
+ *
+ * XXX This needs to handle sequences too - after AlterSubscription_refresh
+ * starts caring about sequences, GetSubscriptionNotReadyRelations won't
+ * return just tables, and we'll have to sync them here. Not sure it's worth
+ * creating a new "sync" worker per sequence, maybe we should just sync them
+ * in the current process (it's pretty light-weight).
*/
static void
process_syncing_tables_for_apply(XLogRecPtr current_lsn)
@@ -871,6 +878,99 @@ copy_table(Relation rel)
logicalrep_rel_close(relmapentry, NoLock);
}
+
+
+/*
+ * FIXME add comment
+ */
+static void
+fetch_sequence_data(char *nspname, char *relname,
+ int64 *last_value, int64 *log_cnt, bool *is_called)
+{
+ WalRcvExecResult *res;
+ StringInfoData cmd;
+ TupleTableSlot *slot;
+ Oid tableRow[3] = {INT8OID, INT8OID, BOOLOID};
+
+ initStringInfo(&cmd);
+ appendStringInfo(&cmd, "SELECT last_value, log_cnt, is_called\n"
+ " FROM %s", quote_qualified_identifier(nspname, relname));
+
+ res = walrcv_exec(LogRepWorkerWalRcvConn, cmd.data, 3, tableRow);
+ pfree(cmd.data);
+
+ if (res->status != WALRCV_OK_TUPLES)
+ ereport(ERROR,
+ (errmsg("could not receive list of replicated tables from the publisher: %s",
+ res->err)));
+
+ /* Process the sequence. */
+ slot = MakeSingleTupleTableSlot(res->tupledesc, &TTSOpsMinimalTuple);
+ while (tuplestore_gettupleslot(res->tuplestore, true, false, slot))
+ {
+ bool isnull;
+
+ *last_value = DatumGetInt64(slot_getattr(slot, 1, &isnull));
+ Assert(!isnull);
+
+ *log_cnt = DatumGetInt64(slot_getattr(slot, 2, &isnull));
+ Assert(!isnull);
+
+ *is_called = DatumGetBool(slot_getattr(slot, 3, &isnull));
+ Assert(!isnull);
+
+ ExecClearTuple(slot);
+ }
+ ExecDropSingleTupleTableSlot(slot);
+
+ walrcv_clear_result(res);
+}
+
+/*
+ * Copy existing data of a sequence from publisher.
+ *
+ * Caller is responsible for locking the local relation.
+ */
+static void
+copy_sequence(Relation rel)
+{
+ LogicalRepRelMapEntry *relmapentry;
+ LogicalRepRelation lrel;
+ StringInfoData cmd;
+ int64 last_value = 0,
+ log_cnt = 0;
+ bool is_called = 0;
+
+ /* Get the publisher relation info. */
+ fetch_remote_table_info(get_namespace_name(RelationGetNamespace(rel)),
+ RelationGetRelationName(rel), &lrel);
+
+ /* Put the relation into relmap. */
+ logicalrep_relmap_update(&lrel);
+
+ /* Map the publisher relation to local one. */
+ relmapentry = logicalrep_rel_open(lrel.remoteid, NoLock);
+ Assert(rel == relmapentry->localrel);
+
+ /* Start copy on the publisher. */
+ initStringInfo(&cmd);
+
+ Assert(lrel.relkind == RELKIND_SEQUENCE);
+
+ fetch_sequence_data(lrel.nspname, lrel.relname, &last_value, &log_cnt, &is_called);
+
+ elog(WARNING, "sequence %s info last_value %ld log_cnt %ld is_called %d",
+ quote_qualified_identifier(lrel.nspname, lrel.relname),
+ last_value, log_cnt, is_called);
+
+ ResetSequence2(RelationGetRelid(rel), last_value, log_cnt, is_called);
+
+ logicalrep_rel_close(relmapentry, NoLock);
+}
+
+
+
+
/*
* Determine the tablesync slot name.
*
@@ -1106,10 +1206,20 @@ LogicalRepSyncTableStart(XLogRecPtr *origin_startpos)
originname)));
}
- /* Now do the initial data copy */
- PushActiveSnapshot(GetTransactionSnapshot());
- copy_table(rel);
- PopActiveSnapshot();
+ if (get_rel_relkind(RelationGetRelid(rel)) == RELKIND_SEQUENCE)
+ {
+ /* Now do the initial sequence copy */
+ PushActiveSnapshot(GetTransactionSnapshot());
+ copy_sequence(rel);
+ PopActiveSnapshot();
+ }
+ else
+ {
+ /* Now do the initial data copy */
+ PushActiveSnapshot(GetTransactionSnapshot());
+ copy_table(rel);
+ PopActiveSnapshot();
+ }
res = walrcv_exec(LogRepWorkerWalRcvConn, "COMMIT", 0, NULL);
if (res->status != WALRCV_OK_COMMAND)
diff --git a/src/backend/replication/logical/worker.c b/src/backend/replication/logical/worker.c
index 3f499b11f7..f4dbf5ed71 100644
--- a/src/backend/replication/logical/worker.c
+++ b/src/backend/replication/logical/worker.c
@@ -144,6 +144,7 @@
#include "catalog/pg_tablespace.h"
#include "commands/tablecmds.h"
#include "commands/tablespace.h"
+#include "commands/sequence.h"
#include "commands/trigger.h"
#include "executor/executor.h"
#include "executor/execPartition.h"
@@ -1073,6 +1074,61 @@ apply_handle_origin(StringInfo s)
errmsg_internal("ORIGIN message sent out of order")));
}
+/*
+ * Handle SEQUENCE message.
+ */
+static void
+apply_handle_sequence(StringInfo s)
+{
+ LogicalRepSequence seq;
+ Oid relid;
+
+ if (handle_streamed_transaction(LOGICAL_REP_MSG_SEQUENCE, s))
+ return;
+
+ logicalrep_read_sequence(s, &seq);
+
+ /*
+ * Non-transactional sequence updates should not be part of a remote
+ * transaction. There should not be any running transaction.
+ */
+ Assert((!seq.transactional) || in_remote_transaction);
+ Assert(!(!seq.transactional && in_remote_transaction));
+ Assert(!(!seq.transactional && IsTransactionState()));
+
+ /*
+ * Make sure we're in a transaction (needed by ResetSequence2). For
+ * non-transactional updates we're guaranteed to start a new one,
+ * and we'll commit it at the end.
+ */
+ if (!IsTransactionState())
+ {
+ StartTransactionCommand();
+ maybe_reread_subscription();
+ }
+
+ relid = RangeVarGetRelid(makeRangeVar(seq.nspname,
+ seq.seqname, -1),
+ RowExclusiveLock, false);
+
+ /* lock the sequence in AccessExclusiveLock, as expected by ResetSequence2 */
+ elog(WARNING, "locking sequence %d in exclusive mode", relid);
+ LockRelationOid(relid, AccessExclusiveLock);
+
+ elog(WARNING, "applying sequence %s.%s transactional %d created %d last_value %ld log_cnt %ld is_called %d",
+ seq.nspname, seq.seqname, seq.transactional, seq.created, seq.last_value, seq.log_cnt, seq.is_called);
+
+ /* apply the sequence change */
+ ResetSequence2(relid, seq.last_value, seq.log_cnt, seq.is_called);
+
+ /*
+ * Commit the per-stream transaction (we only do this when not in
+ * remote transaction, i.e. for non-transactional sequence updates.
+ */
+ if (!in_remote_transaction)
+ CommitTransactionCommand();
+}
+
/*
* Handle STREAM START message.
*/
@@ -2327,6 +2383,10 @@ apply_dispatch(StringInfo s)
*/
return;
+ case LOGICAL_REP_MSG_SEQUENCE:
+ apply_handle_sequence(s);
+ return;
+
case LOGICAL_REP_MSG_STREAM_START:
apply_handle_stream_start(s);
return;
diff --git a/src/backend/replication/pgoutput/pgoutput.c b/src/backend/replication/pgoutput/pgoutput.c
index e4314af13a..4366c275dc 100644
--- a/src/backend/replication/pgoutput/pgoutput.c
+++ b/src/backend/replication/pgoutput/pgoutput.c
@@ -49,6 +49,10 @@ static void pgoutput_message(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn, XLogRecPtr message_lsn,
bool transactional, const char *prefix,
Size sz, const char *message);
+static void pgoutput_sequence(LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn, XLogRecPtr sequence_lsn,
+ Relation rel, bool transactional, bool created,
+ int64 last_value, int64 log_cnt, bool is_called);
static bool pgoutput_origin_filter(LogicalDecodingContext *ctx,
RepOriginId origin_id);
static void pgoutput_begin_prepare_txn(LogicalDecodingContext *ctx,
@@ -157,6 +161,7 @@ _PG_output_plugin_init(OutputPluginCallbacks *cb)
cb->change_cb = pgoutput_change;
cb->truncate_cb = pgoutput_truncate;
cb->message_cb = pgoutput_message;
+ cb->sequence_cb = pgoutput_sequence;
cb->commit_cb = pgoutput_commit_txn;
cb->begin_prepare_cb = pgoutput_begin_prepare_txn;
@@ -186,6 +191,7 @@ parse_output_parameters(List *options, PGOutputData *data)
bool publication_names_given = false;
bool binary_option_given = false;
bool messages_option_given = false;
+ bool sequences_option_given = false;
bool streaming_given = false;
bool two_phase_option_given = false;
@@ -193,6 +199,7 @@ parse_output_parameters(List *options, PGOutputData *data)
data->streaming = false;
data->messages = false;
data->two_phase = false;
+ data->sequences = true;
foreach(lc, options)
{
@@ -258,6 +265,16 @@ parse_output_parameters(List *options, PGOutputData *data)
data->messages = defGetBoolean(defel);
}
+ else if (strcmp(defel->defname, "sequences") == 0)
+ {
+ if (sequences_option_given)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("conflicting or redundant options")));
+ sequences_option_given = true;
+
+ data->sequences = defGetBoolean(defel);
+ }
else if (strcmp(defel->defname, "streaming") == 0)
{
if (streaming_given)
@@ -865,6 +882,47 @@ pgoutput_message(LogicalDecodingContext *ctx, ReorderBufferTXN *txn,
OutputPluginWrite(ctx, true);
}
+static void
+pgoutput_sequence(LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn, XLogRecPtr sequence_lsn,
+ Relation rel, bool transactional, bool created,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ PGOutputData *data = (PGOutputData *) ctx->output_plugin_private;
+ TransactionId xid = InvalidTransactionId;
+ RelationSyncEntry *relentry;
+
+ if (!data->sequences)
+ return;
+
+ if (!is_publishable_relation(rel))
+ return;
+
+ /* XXX handle (in_streaming) here */
+
+ relentry = get_rel_sync_entry(data, RelationGetRelid(rel));
+
+ /*
+ * First check the sequence filter.
+ *
+ * We handle just REORDER_BUFFER_CHANGE_SEQUENCE here.
+ */
+ if (!relentry->pubactions.pubsequence)
+ return;
+
+ OutputPluginPrepareWrite(ctx, true);
+ logicalrep_write_sequence(ctx->out,
+ rel,
+ xid,
+ sequence_lsn,
+ transactional,
+ created,
+ last_value,
+ log_cnt,
+ is_called);
+ OutputPluginWrite(ctx, true);
+}
+
/*
* Currently we always forward.
*/
@@ -1128,7 +1186,8 @@ get_rel_sync_entry(PGOutputData *data, Oid relid)
entry->streamed_txns = NIL;
entry->replicate_valid = false;
entry->pubactions.pubinsert = entry->pubactions.pubupdate =
- entry->pubactions.pubdelete = entry->pubactions.pubtruncate = false;
+ entry->pubactions.pubdelete = entry->pubactions.pubtruncate =
+ entry->pubactions.pubsequence = false;
entry->publish_as_relid = InvalidOid;
entry->map = NULL; /* will be set by maybe_send_schema() if
* needed */
@@ -1140,6 +1199,7 @@ get_rel_sync_entry(PGOutputData *data, Oid relid)
List *pubids = GetRelationPublications(relid);
ListCell *lc;
Oid publish_as_relid = relid;
+ bool is_sequence = (get_rel_relkind(relid) == RELKIND_SEQUENCE);
/* Reload publications if needed before use. */
if (!publications_valid)
@@ -1163,12 +1223,23 @@ get_rel_sync_entry(PGOutputData *data, Oid relid)
Publication *pub = lfirst(lc);
bool publish = false;
- if (pub->alltables)
+ if (pub->alltables && (!is_sequence))
{
publish = true;
if (pub->pubviaroot && am_partition)
publish_as_relid = llast_oid(get_partition_ancestors(relid));
}
+ else if (pub->allsequences && is_sequence)
+ {
+ publish = true;
+ }
+
+ /* if a sequence, just cross-check the list of publications */
+ if (!publish && is_sequence)
+ {
+ if (list_member_oid(pubids, pub->oid))
+ publish = true;
+ }
if (!publish)
{
@@ -1219,10 +1290,12 @@ get_rel_sync_entry(PGOutputData *data, Oid relid)
entry->pubactions.pubupdate |= pub->pubactions.pubupdate;
entry->pubactions.pubdelete |= pub->pubactions.pubdelete;
entry->pubactions.pubtruncate |= pub->pubactions.pubtruncate;
+ entry->pubactions.pubsequence |= pub->pubactions.pubsequence;
}
if (entry->pubactions.pubinsert && entry->pubactions.pubupdate &&
- entry->pubactions.pubdelete && entry->pubactions.pubtruncate)
+ entry->pubactions.pubdelete && entry->pubactions.pubtruncate &&
+ entry->pubactions.pubsequence)
break;
}
@@ -1367,6 +1440,7 @@ rel_sync_cache_publication_cb(Datum arg, int cacheid, uint32 hashvalue)
entry->pubactions.pubupdate = false;
entry->pubactions.pubdelete = false;
entry->pubactions.pubtruncate = false;
+ entry->pubactions.pubsequence = false;
}
}
diff --git a/src/backend/utils/cache/relcache.c b/src/backend/utils/cache/relcache.c
index 13d9994af3..3dd39054e6 100644
--- a/src/backend/utils/cache/relcache.c
+++ b/src/backend/utils/cache/relcache.c
@@ -5495,6 +5495,7 @@ GetRelationPublicationActions(Relation relation)
pubactions->pubupdate |= pubform->pubupdate;
pubactions->pubdelete |= pubform->pubdelete;
pubactions->pubtruncate |= pubform->pubtruncate;
+ pubactions->pubsequence |= pubform->pubsequence;
ReleaseSysCache(tup);
@@ -5503,7 +5504,8 @@ GetRelationPublicationActions(Relation relation)
* other publications.
*/
if (pubactions->pubinsert && pubactions->pubupdate &&
- pubactions->pubdelete && pubactions->pubtruncate)
+ pubactions->pubdelete && pubactions->pubtruncate &&
+ pubactions->pubsequence)
break;
}
diff --git a/src/bin/psql/tab-complete.c b/src/bin/psql/tab-complete.c
index 064892bade..fac6f9a563 100644
--- a/src/bin/psql/tab-complete.c
+++ b/src/bin/psql/tab-complete.c
@@ -1640,7 +1640,7 @@ psql_completion(const char *text, int start, int end)
/* ALTER PUBLICATION <name> */
else if (Matches("ALTER", "PUBLICATION", MatchAny))
- COMPLETE_WITH("ADD TABLE", "DROP TABLE", "OWNER TO", "RENAME TO", "SET");
+ COMPLETE_WITH("ADD TABLE", "DROP TABLE", "ADD SEQUENCE", "DROP SEQUENCE", "OWNER TO", "RENAME TO", "SET");
/* ALTER PUBLICATION <name> SET */
else if (Matches("ALTER", "PUBLICATION", MatchAny, "SET"))
COMPLETE_WITH("(", "TABLE");
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 8cd0252082..8700a72dea 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -11445,6 +11445,11 @@
provolatile => 's', prorettype => 'oid', proargtypes => 'text',
proallargtypes => '{text,oid}', proargmodes => '{i,o}',
proargnames => '{pubname,relid}', prosrc => 'pg_get_publication_tables' },
+{ oid => '8000', descr => 'get OIDs of sequences in a publication',
+ proname => 'pg_get_publication_sequences', prorows => '1000', proretset => 't',
+ provolatile => 's', prorettype => 'oid', proargtypes => 'text',
+ proallargtypes => '{text,oid}', proargmodes => '{i,o}',
+ proargnames => '{pubname,relid}', prosrc => 'pg_get_publication_sequences' },
{ oid => '6121',
descr => 'returns whether a relation can be part of a publication',
proname => 'pg_relation_is_publishable', provolatile => 's',
diff --git a/src/include/catalog/pg_publication.h b/src/include/catalog/pg_publication.h
index f332bad4d4..150b8922cf 100644
--- a/src/include/catalog/pg_publication.h
+++ b/src/include/catalog/pg_publication.h
@@ -40,6 +40,12 @@ CATALOG(pg_publication,6104,PublicationRelationId)
*/
bool puballtables;
+ /*
+ * indicates that this is special publication which should encompass all
+ * sequences in the database (except for the unlogged and temp ones)
+ */
+ bool puballsequences;
+
/* true if inserts are published */
bool pubinsert;
@@ -52,6 +58,9 @@ CATALOG(pg_publication,6104,PublicationRelationId)
/* true if truncates are published */
bool pubtruncate;
+ /* true if sequences are published */
+ bool pubsequence;
+
/* true if partition changes are published using root schema */
bool pubviaroot;
} FormData_pg_publication;
@@ -72,6 +81,7 @@ typedef struct PublicationActions
bool pubupdate;
bool pubdelete;
bool pubtruncate;
+ bool pubsequence;
} PublicationActions;
typedef struct Publication
@@ -79,6 +89,7 @@ typedef struct Publication
Oid oid;
char *name;
bool alltables;
+ bool allsequences;
bool pubviaroot;
PublicationActions pubactions;
} Publication;
@@ -107,6 +118,9 @@ extern List *GetPublicationRelations(Oid pubid, PublicationPartOpt pub_partopt);
extern List *GetAllTablesPublications(void);
extern List *GetAllTablesPublicationRelations(bool pubviaroot);
+extern List *GetAllSequencesPublicationRelations(void);
+extern List *GetPublicationSequenceRelations(Oid pubid);
+
extern bool is_publishable_relation(Relation rel);
extern ObjectAddress publication_add_relation(Oid pubid, Relation targetrel,
bool if_not_exists);
diff --git a/src/include/commands/sequence.h b/src/include/commands/sequence.h
index 40544dd4c7..c28e8695cb 100644
--- a/src/include/commands/sequence.h
+++ b/src/include/commands/sequence.h
@@ -48,6 +48,7 @@ typedef FormData_pg_sequence_data *Form_pg_sequence_data;
typedef struct xl_seq_rec
{
RelFileNode node;
+ bool created; /* is this a CREATE SEQUENCE */
/* SEQUENCE TUPLE DATA FOLLOWS AT THE END */
} xl_seq_rec;
@@ -59,6 +60,7 @@ extern ObjectAddress DefineSequence(ParseState *pstate, CreateSeqStmt *stmt);
extern ObjectAddress AlterSequence(ParseState *pstate, AlterSeqStmt *stmt);
extern void DeleteSequenceTuple(Oid relid);
extern void ResetSequence(Oid seq_relid);
+extern void ResetSequence2(Oid seq_relid, int64 last_value, int64 log_cnt, bool is_called);
extern void ResetSequenceCaches(void);
extern void seq_redo(XLogReaderState *rptr);
diff --git a/src/include/nodes/parsenodes.h b/src/include/nodes/parsenodes.h
index e28248af32..6e76c40f98 100644
--- a/src/include/nodes/parsenodes.h
+++ b/src/include/nodes/parsenodes.h
@@ -3644,8 +3644,10 @@ typedef struct AlterPublicationStmt
/* parameters used for ALTER PUBLICATION ... ADD/DROP TABLE */
List *tables; /* List of tables to add/drop */
+ List *sequences; /* List of sequences to add/drop */
bool for_all_tables; /* Special publication for all tables in db */
- DefElemAction tableAction; /* What action to perform with the tables */
+ bool for_all_sequences; /* Special publication for all tables in db */
+ DefElemAction action; /* What action to perform with the tables/sequences */
} AlterPublicationStmt;
typedef struct CreateSubscriptionStmt
diff --git a/src/include/replication/logicalproto.h b/src/include/replication/logicalproto.h
index 63de90d94a..931bc0722a 100644
--- a/src/include/replication/logicalproto.h
+++ b/src/include/replication/logicalproto.h
@@ -60,6 +60,7 @@ typedef enum LogicalRepMsgType
LOGICAL_REP_MSG_RELATION = 'R',
LOGICAL_REP_MSG_TYPE = 'Y',
LOGICAL_REP_MSG_MESSAGE = 'M',
+ LOGICAL_REP_MSG_SEQUENCE = 'X', /* FIXME change */
LOGICAL_REP_MSG_BEGIN_PREPARE = 'b',
LOGICAL_REP_MSG_PREPARE = 'P',
LOGICAL_REP_MSG_COMMIT_PREPARED = 'K',
@@ -116,6 +117,19 @@ typedef struct LogicalRepTyp
char *typname; /* name of the remote type */
} LogicalRepTyp;
+/* Sequence info */
+typedef struct LogicalRepSequence
+{
+ Oid remoteid; /* unique id of the remote sequence */
+ char *nspname; /* schema name of remote sequence */
+ char *seqname; /* name of the remote sequence */
+ bool transactional;
+ bool created;
+ int64 last_value;
+ int64 log_cnt; /* XXX probably not needed? */
+ bool is_called; /* XXX probably not needed? */
+} LogicalRepSequence;
+
/* Transaction info */
typedef struct LogicalRepBeginData
{
@@ -223,6 +237,12 @@ extern List *logicalrep_read_truncate(StringInfo in,
bool *cascade, bool *restart_seqs);
extern void logicalrep_write_message(StringInfo out, TransactionId xid, XLogRecPtr lsn,
bool transactional, const char *prefix, Size sz, const char *message);
+extern void logicalrep_write_sequence(StringInfo out, Relation rel,
+ TransactionId xid, XLogRecPtr lsn,
+ bool transactional, bool created,
+ int64 last_value, int64 log_cnt,
+ bool is_called);
+extern void logicalrep_read_sequence(StringInfo in, LogicalRepSequence *seqdata);
extern void logicalrep_write_rel(StringInfo out, TransactionId xid,
Relation rel);
extern LogicalRepRelation *logicalrep_read_rel(StringInfo in);
diff --git a/src/include/replication/output_plugin.h b/src/include/replication/output_plugin.h
index 810495ed0e..e3e1b03f32 100644
--- a/src/include/replication/output_plugin.h
+++ b/src/include/replication/output_plugin.h
@@ -88,6 +88,19 @@ typedef void (*LogicalDecodeMessageCB) (struct LogicalDecodingContext *ctx,
Size message_size,
const char *message);
+/*
+ * Called for the generic logical decoding sequences.
+ */
+typedef void (*LogicalDecodeSequenceCB) (struct LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional,
+ bool created,
+ int64 last_value,
+ int64 log_cnt,
+ bool is_called);
+
/*
* Filter changes by origin.
*/
@@ -219,6 +232,7 @@ typedef struct OutputPluginCallbacks
LogicalDecodeTruncateCB truncate_cb;
LogicalDecodeCommitCB commit_cb;
LogicalDecodeMessageCB message_cb;
+ LogicalDecodeSequenceCB sequence_cb;
LogicalDecodeFilterByOriginCB filter_by_origin_cb;
LogicalDecodeShutdownCB shutdown_cb;
diff --git a/src/include/replication/pgoutput.h b/src/include/replication/pgoutput.h
index 0dc460fb70..97eb2a7ef1 100644
--- a/src/include/replication/pgoutput.h
+++ b/src/include/replication/pgoutput.h
@@ -28,6 +28,7 @@ typedef struct PGOutputData
bool streaming;
bool messages;
bool two_phase;
+ bool sequences;
} PGOutputData;
#endif /* PGOUTPUT_H */
diff --git a/src/include/replication/reorderbuffer.h b/src/include/replication/reorderbuffer.h
index 5b40ff75f7..4512ed679b 100644
--- a/src/include/replication/reorderbuffer.h
+++ b/src/include/replication/reorderbuffer.h
@@ -64,7 +64,8 @@ enum ReorderBufferChangeType
REORDER_BUFFER_CHANGE_INTERNAL_SPEC_INSERT,
REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM,
REORDER_BUFFER_CHANGE_INTERNAL_SPEC_ABORT,
- REORDER_BUFFER_CHANGE_TRUNCATE
+ REORDER_BUFFER_CHANGE_TRUNCATE,
+ REORDER_BUFFER_CHANGE_SEQUENCE
};
/* forward declaration */
@@ -158,6 +159,14 @@ typedef struct ReorderBufferChange
uint32 ninvalidations; /* Number of messages */
SharedInvalidationMessage *invalidations; /* invalidation message */
} inval;
+
+ /* Context data for Sequence changes */
+ struct
+ {
+ RelFileNode relnode;
+ bool created;
+ ReorderBufferTupleBuf *tuple;
+ } sequence;
} data;
/*
@@ -430,6 +439,15 @@ typedef void (*ReorderBufferMessageCB) (ReorderBuffer *rb,
const char *prefix, Size sz,
const char *message);
+/* sequence callback signature */
+typedef void (*ReorderBufferSequenceCB) (ReorderBuffer *rb,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional, bool created,
+ int64 last_value, int64 log_cnt,
+ bool is_called);
+
/* begin prepare callback signature */
typedef void (*ReorderBufferBeginPrepareCB) (ReorderBuffer *rb,
ReorderBufferTXN *txn);
@@ -511,6 +529,12 @@ struct ReorderBuffer
*/
HTAB *by_txn;
+ /*
+ * relfilenode => XID lookup table for sequences created in a transaction
+ * (also includes altered sequences, which assigns new relfilenode)
+ */
+ HTAB *sequences;
+
/*
* Transactions that could be a toplevel xact, ordered by LSN of the first
* record bearing that xid.
@@ -541,6 +565,7 @@ struct ReorderBuffer
ReorderBufferApplyTruncateCB apply_truncate;
ReorderBufferCommitCB commit;
ReorderBufferMessageCB message;
+ ReorderBufferSequenceCB sequence;
/*
* Callbacks to be called when streaming a transaction at prepare time.
@@ -635,6 +660,10 @@ void ReorderBufferQueueChange(ReorderBuffer *, TransactionId,
void ReorderBufferQueueMessage(ReorderBuffer *, TransactionId, Snapshot snapshot, XLogRecPtr lsn,
bool transactional, const char *prefix,
Size message_size, const char *message);
+void ReorderBufferQueueSequence(ReorderBuffer *rb, TransactionId xid,
+ Snapshot snapshot, XLogRecPtr lsn, RepOriginId origin_id,
+ RelFileNode rnode, bool transactional, bool created,
+ ReorderBufferTupleBuf *tuplebuf);
void ReorderBufferCommit(ReorderBuffer *, TransactionId,
XLogRecPtr commit_lsn, XLogRecPtr end_lsn,
TimestampTz commit_time, RepOriginId origin_id, XLogRecPtr origin_lsn);
@@ -682,4 +711,7 @@ void ReorderBufferSetRestartPoint(ReorderBuffer *, XLogRecPtr ptr);
void StartupReorderBuffer(void);
+bool ReorderBufferSequenceIsTransactional(ReorderBuffer *rb,
+ RelFileNode rnode, bool created);
+
#endif
diff --git a/src/test/regress/expected/rules.out b/src/test/regress/expected/rules.out
index e5ab11275d..393be871be 100644
--- a/src/test/regress/expected/rules.out
+++ b/src/test/regress/expected/rules.out
@@ -1451,6 +1451,14 @@ pg_prepared_xacts| SELECT p.transaction,
FROM ((pg_prepared_xact() p(transaction, gid, prepared, ownerid, dbid)
LEFT JOIN pg_authid u ON ((p.ownerid = u.oid)))
LEFT JOIN pg_database d ON ((p.dbid = d.oid)));
+pg_publication_sequences| SELECT p.pubname,
+ n.nspname AS schemaname,
+ c.relname AS sequencename
+ FROM pg_publication p,
+ LATERAL pg_get_publication_sequences((p.pubname)::text) gpt(relid),
+ (pg_class c
+ JOIN pg_namespace n ON ((n.oid = c.relnamespace)))
+ WHERE (c.oid = gpt.relid);
pg_publication_tables| SELECT p.pubname,
n.nspname AS schemaname,
c.relname AS tablename
diff --git a/src/test/subscription/t/023_sequences.pl b/src/test/subscription/t/023_sequences.pl
new file mode 100644
index 0000000000..1f7768acd6
--- /dev/null
+++ b/src/test/subscription/t/023_sequences.pl
@@ -0,0 +1,196 @@
+
+# Copyright (c) 2021, PostgreSQL Global Development Group
+
+# This tests that sequences are replicated correctly by logical replication
+use strict;
+use warnings;
+use PostgresNode;
+use TestLib;
+use Test::More tests => 6;
+
+# Initialize publisher node
+my $node_publisher = PostgresNode->new('publisher');
+$node_publisher->init(allows_streaming => 'logical');
+$node_publisher->start;
+
+# Create subscriber node
+my $node_subscriber = PostgresNode->new('subscriber');
+$node_subscriber->init(allows_streaming => 'logical');
+$node_subscriber->start;
+
+# Create some preexisting content on publisher
+my $ddl = qq(
+ CREATE SEQUENCE s;
+);
+
+# Setup structure on the publisher
+$node_publisher->safe_psql('postgres', $ddl);
+
+# Create some the same structure on subscriber, and an extra sequence that
+# we'll create on the publisher later
+$ddl = qq(
+ CREATE SEQUENCE s;
+ CREATE SEQUENCE s2;
+);
+
+$node_subscriber->safe_psql('postgres', $ddl);
+
+# Setup logical replication
+my $publisher_connstr = $node_publisher->connstr . ' dbname=postgres';
+$node_publisher->safe_psql('postgres',
+ "CREATE PUBLICATION seq_pub");
+
+$node_publisher->safe_psql('postgres',
+ "ALTER PUBLICATION seq_pub ADD SEQUENCE s");
+
+$node_subscriber->safe_psql('postgres',
+ "CREATE SUBSCRIPTION seq_sub CONNECTION '$publisher_connstr' PUBLICATION seq_pub WITH (slot_name = seq_sub_slot)"
+);
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Wait for initial sync to finish as well
+my $synced_query =
+ "SELECT count(1) = 0 FROM pg_subscription_rel WHERE srsubstate NOT IN ('s', 'r');";
+$node_subscriber->poll_query_until('postgres', $synced_query)
+ or die "Timed out while waiting for subscriber to synchronize data";
+
+# Insert initial test data
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ -- generate a number of values using the sequence
+ SELECT nextval('s') FROM generate_series(1,100);
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+my $result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s;
+));
+
+is( $result, '132|0|t',
+ 'check replicated sequence values on subscriber');
+
+
+# advance the sequence in a rolled-back transaction - should be replicated
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ SELECT nextval('s') FROM generate_series(1,100);
+ ROLLBACK;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s;
+));
+
+is( $result, '231|0|t',
+ 'check replicated sequence values on subscriber');
+
+
+# create a new sequence and roll it back - should not be replicated, due to
+# the transactional behavior
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ CREATE SEQUENCE s2;
+ ALTER PUBLICATION seq_pub ADD SEQUENCE s2;
+ SELECT nextval('s2') FROM generate_series(1,100);
+ ROLLBACK;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s2;
+));
+
+is( $result, '1|0|f',
+ 'check replicated sequence values on subscriber');
+
+
+# create a new sequence, advance it in a rolled-back transaction, but commit
+# the create - the advance should be replicated nevertheless
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ CREATE SEQUENCE s2;
+ ALTER PUBLICATION seq_pub ADD SEQUENCE s2;
+ SAVEPOINT sp1;
+ SELECT nextval('s2') FROM generate_series(1,100);
+ ROLLBACK TO sp1;
+ COMMIT;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Wait for sync of the second sequence we just added to finish
+$synced_query =
+ "SELECT count(1) = 0 FROM pg_subscription_rel WHERE srsubstate NOT IN ('s', 'r');";
+$node_subscriber->poll_query_until('postgres', $synced_query)
+ or die "Timed out while waiting for subscriber to synchronize data";
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s2;
+));
+
+is( $result, '132|0|t',
+ 'check replicated sequence values on subscriber');
+
+
+# advance the new sequence in a transaction, and roll it back - in this case
+# it should be replicated as the behavior is non-transactional
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ SELECT nextval('s2') FROM generate_series(1,100);
+ ROLLBACK;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s2;
+));
+
+is( $result, '231|0|t',
+ 'check replicated sequence values on subscriber');
+
+
+# advance the sequence in a subtransaction - the subtransaction gets rolled
+# back, but commit the main one - the changes should still be replicated
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ SAVEPOINT s1;
+ SELECT nextval('s2') FROM generate_series(1,100);
+ ROLLBACK TO s1;
+ COMMIT;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s2;
+));
+
+is( $result, '330|0|t',
+ 'check replicated sequence values on subscriber');
+
+
+$node_subscriber->stop('fast');
+$node_publisher->stop('fast');
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2021-07-30 18:58 Robert Haas <[email protected]>
parent: Tomas Vondra <[email protected]>
1 sibling, 0 replies; 127+ messages in thread
From: Robert Haas @ 2021-07-30 18:58 UTC (permalink / raw)
To: Tomas Vondra <[email protected]>; +Cc: vignesh C <[email protected]>; PostgreSQL Hackers <[email protected]>; Petr Jelinek <[email protected]>
On Tue, Jul 20, 2021 at 5:41 PM Tomas Vondra
<[email protected]> wrote:
> I think the crucial aspect of this that needs discussion/feedback the
> most is the transactional vs. non-transactional behavior. All the other
> questions are less important / cosmetic.
Yeah, it seems really tricky to me to get this right. The hard part
is, I think, mostly figuring out what the right behavior really is.
DDL in PostgreSQL is transactional. Non-DDL operations on sequences
are non-transactional. If a given transaction does only one of those
things, it seems clear enough what to do, but when the same
(sub)transaction does both, it gets messy. I'd be tempted to think
about something like:
1. When a transaction performs only non-transactional operations on
sequences, they are emitted immediately.
2. If a transaction performs transactional operations on sequences,
the decoded operations acquire a dependency on the transaction and
cannot be emitted until that transaction is fully decoded. When commit
or abort of that XID is reached, emit the postponed non-transactional
operations at that point.
I think this is similar to what you've designed, but I'm not sure that
it's exactly equivalent. I think in particular that it may be better
to insist that all of these operations are non-transactional and that
the debate is only about when they can be sent, rather than trying to
sort of convert them into an equivalent series of transactional
operations. That approach seems confusing especially in the case where
some (sub)transactions abort.
But this is just my $0.02.
--
Robert Haas
EDB: http://www.enterprisedb.com
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2021-09-23 10:27 Peter Eisentraut <[email protected]>
parent: Tomas Vondra <[email protected]>
0 siblings, 1 reply; 127+ messages in thread
From: Peter Eisentraut @ 2021-09-23 10:27 UTC (permalink / raw)
To: Tomas Vondra <[email protected]>; vignesh C <[email protected]>; +Cc: PostgreSQL Hackers <[email protected]>; Petr Jelinek <[email protected]>
On 30.07.21 20:26, Tomas Vondra wrote:
> Here's a an updated version of this patch - rebased to current master,
> and fixing some of the issues raised in Peter's review.
This patch needs an update, as various conflicts have arisen now.
As was discussed before, it might be better to present a separate patch
for just the logical decoding part for now, since the replication and
DDL stuff has the potential to conflict heavily with other patches being
discussed right now. It looks like cutting this patch in two should be
doable easily.
I looked through the test cases in test_decoding again. It all looks
pretty sensible. If anyone can think of any other tricky or dubious
cases, we can add them there. It's easiest to discuss these things with
concrete test cases rather than in theory.
One slightly curious issue is that this can make sequence values go
backwards, when seen by the logical decoding consumer, like in the test
case:
+ BEGIN
+ sequence: public.test_sequence transactional: 1 created: 1 last_value:
1, log_cnt: 0 is_called: 0
+ COMMIT
+ sequence: public.test_sequence transactional: 0 created: 0 last_value:
33, log_cnt: 0 is_called: 1
+ BEGIN
+ sequence: public.test_sequence transactional: 1 created: 1 last_value:
4, log_cnt: 0 is_called: 1
+ COMMIT
+ sequence: public.test_sequence transactional: 0 created: 0 last_value:
334, log_cnt: 0 is_called: 1
I suppose that's okay, since it's not really the intention that someone
is concurrently consuming sequence values on the subscriber. Maybe
something for the future. Fixing that would require changing the way
transactional sequence DDL updates these values, so it's not directly
the job of the decoding to address this.
A small thing I found: Maybe the text that test_decoding produces for
sequences can be made to look more consistent with the one for tables.
For example, in
+ BEGIN
+ sequence: public.test_table_a_seq transactional: 1 created: 1
last_value: 1, log_cnt: 0 is_called: 0
+ sequence: public.test_table_a_seq transactional: 1 created: 0
last_value: 33, log_cnt: 0 is_called: 1
+ table public.test_table: INSERT: a[integer]:1 b[integer]:100
+ table public.test_table: INSERT: a[integer]:2 b[integer]:200
+ COMMIT
note how the punctuation is different.
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2021-09-24 19:16 Tomas Vondra <[email protected]>
parent: Peter Eisentraut <[email protected]>
0 siblings, 1 reply; 127+ messages in thread
From: Tomas Vondra @ 2021-09-24 19:16 UTC (permalink / raw)
To: Peter Eisentraut <[email protected]>; vignesh C <[email protected]>; +Cc: PostgreSQL Hackers <[email protected]>; Petr Jelinek <[email protected]>
Hi,
On 9/23/21 12:27 PM, Peter Eisentraut wrote:
> On 30.07.21 20:26, Tomas Vondra wrote:
>> Here's a an updated version of this patch - rebased to current master,
>> and fixing some of the issues raised in Peter's review.
>
> This patch needs an update, as various conflicts have arisen now.
>
> As was discussed before, it might be better to present a separate patch
> for just the logical decoding part for now, since the replication and
> DDL stuff has the potential to conflict heavily with other patches being
> discussed right now. It looks like cutting this patch in two should be
> doable easily.
>
Attached is the rebased patch, split into three parts:
1) basic decoding infrastructure (decoding, reorderbuffer etc.)
2) support for sequences in test_decoding
3) support for sequences in built-in replication (catalogs, syntax, ...)
The last part is the largest one - I'm sure we'll have discussions about
the grammar, adding sequences automatically, etc. But as you said, let's
focus on the first part, which deals with the required decoding stuff.
I've added a couple comments, explaining how we track sequences, why we
need the XID in nextval() etc. I've also added streaming support.
> I looked through the test cases in test_decoding again. It all looks
> pretty sensible. If anyone can think of any other tricky or dubious
> cases, we can add them there. It's easiest to discuss these things with
> concrete test cases rather than in theory.
>
> One slightly curious issue is that this can make sequence values go
> backwards, when seen by the logical decoding consumer, like in the test
> case:
>
> + BEGIN
> + sequence: public.test_sequence transactional: 1 created: 1 last_value:
> 1, log_cnt: 0 is_called: 0
> + COMMIT
> + sequence: public.test_sequence transactional: 0 created: 0 last_value:
> 33, log_cnt: 0 is_called: 1
> + BEGIN
> + sequence: public.test_sequence transactional: 1 created: 1 last_value:
> 4, log_cnt: 0 is_called: 1
> + COMMIT
> + sequence: public.test_sequence transactional: 0 created: 0 last_value:
> 334, log_cnt: 0 is_called: 1
>
> I suppose that's okay, since it's not really the intention that someone
> is concurrently consuming sequence values on the subscriber. Maybe
> something for the future. Fixing that would require changing the way
> transactional sequence DDL updates these values, so it's not directly
> the job of the decoding to address this.
>
Yeah, that's due to how ALTER SEQUENCE does things, and I agree redoing
that seems well out of scope for this patch. What seems a bit suspicious
is that some of the ALTER SEQUENCE changes have "created: 1" - it's
probably correct, though, because those ALTER SEQUENCE statements can be
rolled-back, so we see it as if a new sequence is created. The flag name
might be a bit confusing, though.
> A small thing I found: Maybe the text that test_decoding produces for
> sequences can be made to look more consistent with the one for tables.
> For example, in
>
> + BEGIN
> + sequence: public.test_table_a_seq transactional: 1 created: 1
> last_value: 1, log_cnt: 0 is_called: 0
> + sequence: public.test_table_a_seq transactional: 1 created: 0
> last_value: 33, log_cnt: 0 is_called: 1
> + table public.test_table: INSERT: a[integer]:1 b[integer]:100
> + table public.test_table: INSERT: a[integer]:2 b[integer]:200
> + COMMIT
>
> note how the punctuation is different.
I did tweak this a bit, hopefully it's more consistent.
regards
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2021-09-25 20:05 Hannu Krosing <[email protected]>
parent: Tomas Vondra <[email protected]>
0 siblings, 2 replies; 127+ messages in thread
From: Hannu Krosing @ 2021-09-25 20:05 UTC (permalink / raw)
To: Tomas Vondra <[email protected]>; +Cc: Peter Eisentraut <[email protected]>; vignesh C <[email protected]>; PostgreSQL Hackers <[email protected]>; Petr Jelinek <[email protected]>
Just a note for some design decisions
> 1) By default, sequences are treated non-transactionally, i.e. sent to the output plugin right away.
If our aim is just to make sure that all user-visible data in
*transactional* tables is consistent with sequence state then one
very much simplified approach to this could be to track the results of
nextval() calls in a transaction at COMMIT put the latest sequence
value in WAL (or just track the sequences affected and put the latest
sequence state in WAL at commit which needs extra read of sequence but
protects against race conditions with parallel transactions which get
rolled back later)
This avoids sending redundant changes for multiple nextval() calls
(like loading a million-row table with sequence-generated id column)
And one can argue that we can safely ignore anything in ROLLBACKED
sequences. This is assuming that even if we did advance the sequence
paste the last value sent by the latest COMMITTED transaction it does
not matter for database consistency.
It can matter if customers just call nextval() in rolled-back
transactions and somehow expect these values to be replicated based on
reasoning along "sequences are not transactional - so rollbacks should
not matter" .
Or we may get away with most in-detail sequence tracking on the source
if we just keep track of the xmin of the sequence and send the
sequence info over at commit if it == current_transaction_id ?
-----
Hannu Krosing
Google Cloud - We have a long list of planned contributions and we are hiring.
Contact me if interested.
On Fri, Sep 24, 2021 at 9:16 PM Tomas Vondra
<[email protected]> wrote:
>
> Hi,
>
> On 9/23/21 12:27 PM, Peter Eisentraut wrote:
> > On 30.07.21 20:26, Tomas Vondra wrote:
> >> Here's a an updated version of this patch - rebased to current master,
> >> and fixing some of the issues raised in Peter's review.
> >
> > This patch needs an update, as various conflicts have arisen now.
> >
> > As was discussed before, it might be better to present a separate patch
> > for just the logical decoding part for now, since the replication and
> > DDL stuff has the potential to conflict heavily with other patches being
> > discussed right now. It looks like cutting this patch in two should be
> > doable easily.
> >
>
> Attached is the rebased patch, split into three parts:
>
> 1) basic decoding infrastructure (decoding, reorderbuffer etc.)
> 2) support for sequences in test_decoding
> 3) support for sequences in built-in replication (catalogs, syntax, ...)
>
> The last part is the largest one - I'm sure we'll have discussions about
> the grammar, adding sequences automatically, etc. But as you said, let's
> focus on the first part, which deals with the required decoding stuff.
>
> I've added a couple comments, explaining how we track sequences, why we
> need the XID in nextval() etc. I've also added streaming support.
>
>
> > I looked through the test cases in test_decoding again. It all looks
> > pretty sensible. If anyone can think of any other tricky or dubious
> > cases, we can add them there. It's easiest to discuss these things with
> > concrete test cases rather than in theory.
> >
> > One slightly curious issue is that this can make sequence values go
> > backwards, when seen by the logical decoding consumer, like in the test
> > case:
> >
> > + BEGIN
> > + sequence: public.test_sequence transactional: 1 created: 1 last_value:
> > 1, log_cnt: 0 is_called: 0
> > + COMMIT
> > + sequence: public.test_sequence transactional: 0 created: 0 last_value:
> > 33, log_cnt: 0 is_called: 1
> > + BEGIN
> > + sequence: public.test_sequence transactional: 1 created: 1 last_value:
> > 4, log_cnt: 0 is_called: 1
> > + COMMIT
> > + sequence: public.test_sequence transactional: 0 created: 0 last_value:
> > 334, log_cnt: 0 is_called: 1
> >
> > I suppose that's okay, since it's not really the intention that someone
> > is concurrently consuming sequence values on the subscriber. Maybe
> > something for the future. Fixing that would require changing the way
> > transactional sequence DDL updates these values, so it's not directly
> > the job of the decoding to address this.
> >
>
> Yeah, that's due to how ALTER SEQUENCE does things, and I agree redoing
> that seems well out of scope for this patch. What seems a bit suspicious
> is that some of the ALTER SEQUENCE changes have "created: 1" - it's
> probably correct, though, because those ALTER SEQUENCE statements can be
> rolled-back, so we see it as if a new sequence is created. The flag name
> might be a bit confusing, though.
>
> > A small thing I found: Maybe the text that test_decoding produces for
> > sequences can be made to look more consistent with the one for tables.
> > For example, in
> >
> > + BEGIN
> > + sequence: public.test_table_a_seq transactional: 1 created: 1
> > last_value: 1, log_cnt: 0 is_called: 0
> > + sequence: public.test_table_a_seq transactional: 1 created: 0
> > last_value: 33, log_cnt: 0 is_called: 1
> > + table public.test_table: INSERT: a[integer]:1 b[integer]:100
> > + table public.test_table: INSERT: a[integer]:2 b[integer]:200
> > + COMMIT
> >
> > note how the punctuation is different.
>
> I did tweak this a bit, hopefully it's more consistent.
>
>
> regards
>
> --
> Tomas Vondra
> EnterpriseDB: http://www.enterprisedb.com
> The Enterprise PostgreSQL Company
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2021-10-31 20:44 Tomas Vondra <[email protected]>
parent: Hannu Krosing <[email protected]>
1 sibling, 1 reply; 127+ messages in thread
From: Tomas Vondra @ 2021-10-31 20:44 UTC (permalink / raw)
To: Hannu Krosing <[email protected]>; +Cc: Peter Eisentraut <[email protected]>; vignesh C <[email protected]>; PostgreSQL Hackers <[email protected]>; Petr Jelinek <[email protected]>
On 9/25/21 22:05, Hannu Krosing wrote:
> Just a note for some design decisions
>
>> 1) By default, sequences are treated non-transactionally, i.e. sent to the output plugin right away.
>
> If our aim is just to make sure that all user-visible data in
> *transactional* tables is consistent with sequence state then one
> very much simplified approach to this could be to track the results of
> nextval() calls in a transaction at COMMIT put the latest sequence
> value in WAL (or just track the sequences affected and put the latest
> sequence state in WAL at commit which needs extra read of sequence but
> protects against race conditions with parallel transactions which get
> rolled back later)
>
Not sure. TBH I feel rather uneasy about adding more stuff in COMMIT.
> This avoids sending redundant changes for multiple nextval() calls
> (like loading a million-row table with sequence-generated id column)
>
Yeah, it'd be nice to have to optimize this a bit, somehow. But I'd bet
it's a negligible amount of data / changes, compared to the table.
> And one can argue that we can safely ignore anything in ROLLBACKED
> sequences. This is assuming that even if we did advance the sequence
> paste the last value sent by the latest COMMITTED transaction it does
> not matter for database consistency.
>
I don't think we can ignore aborted (ROLLBACK) transactions, in the
sense that you can't just discard the increments. Imagine you have this
sequence of transactions:
BEGIN;
SELECT nextval('s'); -- allocates new chunk of values
ROLLBACK;
BEGIN;
SELECT nextval('s'); -- returns one of the cached values
COMMIT;
If you ignore the aborted transaction, then the sequence increment won't
be replicated -- but that's wrong, because user now has a visible
sequence value from that chunk.
So I guess we'd have to maintain a cache of sequences incremented in the
current session, do nothing in aborted transactions (i.e. keep the
contents but don't log anything) and log/reset at commit.
I wonder if multiple sessions make this even more problematic (e.g. due
to session just disconnecting mid transansaction, without writing the
abort record at all). But AFAICS that's not an issue, because the other
session has a separate cache for the sequence.
> It can matter if customers just call nextval() in rolled-back
> transactions and somehow expect these values to be replicated based on
> reasoning along "sequences are not transactional - so rollbacks should
> not matter" .
>
I don't think we guarantee anything for data in transactions that did
not commit, so this seems like a non-issue. I.e. we don't need to go out
of our way to guarantee something we never promised.
> Or we may get away with most in-detail sequence tracking on the source
> if we just keep track of the xmin of the sequence and send the
> sequence info over at commit if it == current_transaction_id ?
>
Not sure I understand this proposal. Can you explain?
regards
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2021-11-19 19:54 Tomas Vondra <[email protected]>
parent: Tomas Vondra <[email protected]>
0 siblings, 1 reply; 127+ messages in thread
From: Tomas Vondra @ 2021-11-19 19:54 UTC (permalink / raw)
To: Hannu Krosing <[email protected]>; +Cc: Peter Eisentraut <[email protected]>; vignesh C <[email protected]>; PostgreSQL Hackers <[email protected]>; Petr Jelinek <[email protected]>
Hi,
I've spent a bit of time exploring the alternative approach outlined by
Hannu, i.e. tracking sequences accessed by the transaction, and logging
the final state just once at COMMIT. Attached is an experimental version
of the patch series doing that - 0001 does the original approach
(decoding the sequence updates from WAL) and then 0002 reworks it to
this alternative solution. The 0003 and 0004 stay mostly the same,
except for minor fixes. Some of the tests in 0003/0004 fail, because
0002 changes the semantics in various ways (more about that later).
The original approach (0001) may seem complex at first, but in principle
it just decodes changes to the sequence relation, and either stashes
them into transaction (just like other changes) or applies them right
away. I'd say that's the most complicated part - deciding whether the
change is transactional or not.
0002 reworks that so that it doesn't decode the existing WAL records,
but tracks sequences which have been modified (updated on-disk state)
and then accessed in the current transaction. And then at COMMIT time we
write a new WAL message with info about the sequence.
I realized we already cache sequences for each session - seqhashtab in
sequence.c. It doesn't have any concept of a transaction, but it seems
fairly easy to make that possible. I did this by adding two flags
- needs_log - means the seesion advanced the sequence (on disk)
- touched - true if the current xact called nextval() etc.
The idea is that what matters is updates to on-disk state, so whenever
we do that we set needs_log. But it only matters when the changes are
made visible in a committed transaction. Consider for example this:
BEGIN;
SELECT nextval('s') FROM generate_series(1,10000) s(i);
ROLLBACK;
SELECT nextval('s');
The first nextval() call certainly sets both flags to true, at least for
default sequences caching 32 values. But the values are not confirmed to
the user because of the rollback - this resets 'touched' flag, but
leaves 'needs_log' set to true.
And then the next nextval() - which may easily be just from cache - sets
touched=true again, and logs the sequence state at (implicit) commit.
Which resets both flags again.
The logging/cleanup happens in AtEOXact_Sequences() which gets called
before commit/abort. This walks all cached sequences and writes the
state for those with both flags true (or resets flag for abort).
The cache also keeps info about the last "sequence state" in the
session, which is then used when writing into into WAL.
To write the sequence state into WAL, I've added a new WAL record
xl_logical_sequence to RM_LOGICALMSG_ID, next to the xl_logical_message.
It's a bit arbitrary, maybe it should be part of RM_SEQ_ID, but it does
the trick. I don't think this is the main issue and it's easy enough to
move it elsewhere if needed.
So, that seems fairly straight-forward and it may reduce the number of
replication messages for large transactions. Unfortunately, it's not
much simpler compared to the first approach - the amount of code is
about the same, and there's a bunch of other issues.
The main issue seems to be about ordering. Consider multiple sessions
all advancing the sequence. With the "old" approach this was naturally
ordered - the order in which the increments were written to WAL made
sense. But the sessions may advance the sequences in one order and then
commit in a different order, which mixes the updates. Consider for
example this scenario with two concurrent transactions:
T1: nextval('s') -> allocates values [1,32]
T2: nextval('s') -> allocates values [33,64]
T2: commit -> logs [33,64]
T1: commit -> logs [1,32]
The result is the sequence on the replica diverted because it replayed
the increments in the opposite order.
I can think of two ways to fix this. Firstly, we could "merge" the
increments in some smart way, e.g. by discarding values considered
"stale" (like decrements). But that seems pretty fragile, because the
sequence may be altered in various ways, reset, etc. And it seems more
like transferring responsibility to someone else instead of actually
solving the issue.
The other fix is simply reading the current sequence state from disk at
commit and logging that (instead of the values cached from the last
increment). But I'm rather skeptical about doing such things right
before COMMIT.
regards
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
Attachments:
[text/x-patch] 0004-Add-support-for-decoding-sequences-to-built-20211118.patch (77.1K, ../../[email protected]/2-0004-Add-support-for-decoding-sequences-to-built-20211118.patch)
download | inline diff:
From 24c30875c9b024b830a03b780fedbcd130395c25 Mon Sep 17 00:00:00 2001
From: Tomas Vondra <[email protected]>
Date: Fri, 19 Nov 2021 01:12:23 +0100
Subject: [PATCH 4/4] Add support for decoding sequences to built-in
replication
---
doc/src/sgml/catalogs.sgml | 71 ++++
doc/src/sgml/ref/alter_publication.sgml | 24 +-
doc/src/sgml/ref/alter_subscription.sgml | 4 +-
src/backend/catalog/pg_publication.c | 149 ++++++++-
src/backend/catalog/system_views.sql | 10 +
src/backend/commands/publicationcmds.c | 350 +++++++++++++++++++-
src/backend/commands/sequence.c | 79 +++++
src/backend/commands/subscriptioncmds.c | 272 +++++++++++++++
src/backend/executor/execReplication.c | 2 +-
src/backend/nodes/copyfuncs.c | 1 +
src/backend/nodes/equalfuncs.c | 1 +
src/backend/parser/gram.y | 32 ++
src/backend/replication/logical/proto.c | 54 +++
src/backend/replication/logical/tablesync.c | 118 ++++++-
src/backend/replication/logical/worker.c | 60 ++++
src/backend/replication/pgoutput/pgoutput.c | 86 ++++-
src/backend/utils/cache/relcache.c | 4 +-
src/bin/psql/tab-complete.c | 6 +-
src/include/catalog/pg_proc.dat | 5 +
src/include/catalog/pg_publication.h | 14 +
src/include/commands/sequence.h | 1 +
src/include/nodes/parsenodes.h | 6 +
src/include/replication/logicalproto.h | 20 ++
src/include/replication/pgoutput.h | 1 +
src/test/regress/expected/rules.out | 8 +
src/test/subscription/t/023_sequences.pl | 196 +++++++++++
26 files changed, 1542 insertions(+), 32 deletions(-)
create mode 100644 src/test/subscription/t/023_sequences.pl
diff --git a/doc/src/sgml/catalogs.sgml b/doc/src/sgml/catalogs.sgml
index c1d11be73f..37b9b52b29 100644
--- a/doc/src/sgml/catalogs.sgml
+++ b/doc/src/sgml/catalogs.sgml
@@ -9498,6 +9498,11 @@ SCRAM-SHA-256$<replaceable><iteration count></replaceable>:<replaceable>&l
<entry>prepared transactions</entry>
</row>
+ <row>
+ <entry><link linkend="view-pg-publication-sequences"><structname>pg_publication_sequences</structname></link></entry>
+ <entry>publications and their associated sequences</entry>
+ </row>
+
<row>
<entry><link linkend="view-pg-publication-tables"><structname>pg_publication_tables</structname></link></entry>
<entry>publications and their associated tables</entry>
@@ -11333,6 +11338,72 @@ SELECT * FROM pg_locks pl LEFT JOIN pg_prepared_xacts ppx
</sect1>
+ <sect1 id="view-pg-publication-sequences">
+ <title><structname>pg_publication_sequences</structname></title>
+
+ <indexterm zone="view-pg-publication-sequences">
+ <primary>pg_publication_sequences</primary>
+ </indexterm>
+
+ <para>
+ The view <structname>pg_publication_sequences</structname> provides
+ information about the mapping between publications and the sequences they
+ contain. Unlike the underlying catalog
+ <link linkend="catalog-pg-publication-rel"><structname>pg_publication_rel</structname></link>,
+ this view expands
+ publications defined as <literal>FOR ALL SEQUENCES</literal>, so for such
+ publications there will be a row for each eligible sequence.
+ </para>
+
+ <table>
+ <title><structname>pg_publication_sequences</structname> Columns</title>
+ <tgroup cols="1">
+ <thead>
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ Column Type
+ </para>
+ <para>
+ Description
+ </para></entry>
+ </row>
+ </thead>
+
+ <tbody>
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ <structfield>pubname</structfield> <type>name</type>
+ (references <link linkend="catalog-pg-publication"><structname>pg_publication</structname></link>.<structfield>pubname</structfield>)
+ </para>
+ <para>
+ Name of publication
+ </para></entry>
+ </row>
+
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ <structfield>schemaname</structfield> <type>name</type>
+ (references <link linkend="catalog-pg-namespace"><structname>pg_namespace</structname></link>.<structfield>nspname</structfield>)
+ </para>
+ <para>
+ Name of schema containing sequence
+ </para></entry>
+ </row>
+
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ <structfield>sequencename</structfield> <type>name</type>
+ (references <link linkend="catalog-pg-class"><structname>pg_class</structname></link>.<structfield>relname</structfield>)
+ </para>
+ <para>
+ Name of sequence
+ </para></entry>
+ </row>
+ </tbody>
+ </tgroup>
+ </table>
+ </sect1>
+
<sect1 id="view-pg-publication-tables">
<title><structname>pg_publication_tables</structname></title>
diff --git a/doc/src/sgml/ref/alter_publication.sgml b/doc/src/sgml/ref/alter_publication.sgml
index bb4ef5e5e2..4d166ad3f9 100644
--- a/doc/src/sgml/ref/alter_publication.sgml
+++ b/doc/src/sgml/ref/alter_publication.sgml
@@ -31,7 +31,9 @@ ALTER PUBLICATION <replaceable class="parameter">name</replaceable> RENAME TO <r
<phrase>where <replaceable class="parameter">publication_object</replaceable> is one of:</phrase>
TABLE [ ONLY ] <replaceable class="parameter">table_name</replaceable> [ * ] [, ... ]
+ SEQUENCE <replaceable class="parameter">sequence_name</replaceable> [ * ] [, ... ]
ALL TABLES IN SCHEMA { <replaceable class="parameter">schema_name</replaceable> | CURRENT_SCHEMA } [, ... ]
+ ALL SEQUENCE IN SCHEMA { <replaceable class="parameter">schema_name</replaceable> | CURRENT_SCHEMA } [, ... ]
</synopsis>
</refsynopsisdiv>
@@ -56,7 +58,18 @@ ALTER PUBLICATION <replaceable class="parameter">name</replaceable> RENAME TO <r
</para>
<para>
- The fourth variant of this command listed in the synopsis can change
+ The next three variants change which sequences are part of the publication.
+ The <literal>SET SEQUENCE</literal> clause will replace the list of sequences
+ in the publication with the specified one. The <literal>ADD SEQUENCE</literal>
+ and <literal>DROP SEQUENCE</literal> clauses will add and remove one or more
+ sequences from the publication. Note that adding sequences to a publication
+ that is already subscribed to will require a <literal>ALTER SUBSCRIPTION
+ ... REFRESH PUBLICATION</literal> action on the subscribing side in order
+ to become effective.
+ </para>
+
+ <para>
+ The seventh variant of this command listed in the synopsis can change
all of the publication properties specified in
<xref linkend="sql-createpublication"/>. Properties not mentioned in the
command retain their previous settings.
@@ -123,6 +136,15 @@ ALTER PUBLICATION <replaceable class="parameter">name</replaceable> RENAME TO <r
</listitem>
</varlistentry>
+ <varlistentry>
+ <term><replaceable class="parameter">sequence_name</replaceable></term>
+ <listitem>
+ <para>
+ Name of an existing sequence.
+ </para>
+ </listitem>
+ </varlistentry>
+
<varlistentry>
<term><literal>SET ( <replaceable class="parameter">publication_parameter</replaceable> [= <replaceable class="parameter">value</replaceable>] [, ... ] )</literal></term>
<listitem>
diff --git a/doc/src/sgml/ref/alter_subscription.sgml b/doc/src/sgml/ref/alter_subscription.sgml
index 0b027cc346..8f28cf03f4 100644
--- a/doc/src/sgml/ref/alter_subscription.sgml
+++ b/doc/src/sgml/ref/alter_subscription.sgml
@@ -147,7 +147,7 @@ ALTER SUBSCRIPTION <replaceable class="parameter">name</replaceable> RENAME TO <
<listitem>
<para>
Fetch missing table information from publisher. This will start
- replication of tables that were added to the subscribed-to publications
+ replication of tables and sequences that were added to the subscribed-to publications
since <command>CREATE SUBSCRIPTION</command> or
the last invocation of <command>REFRESH PUBLICATION</command>.
</para>
@@ -164,7 +164,7 @@ ALTER SUBSCRIPTION <replaceable class="parameter">name</replaceable> RENAME TO <
Specifies whether to copy pre-existing data in the publications
that are being subscribed to when the replication starts.
The default is <literal>true</literal>. (Previously-subscribed
- tables are not copied.)
+ tables and sequences are not copied.)
</para>
</listitem>
</varlistentry>
diff --git a/src/backend/catalog/pg_publication.c b/src/backend/catalog/pg_publication.c
index 63579b2f82..2d5bc73a96 100644
--- a/src/backend/catalog/pg_publication.c
+++ b/src/backend/catalog/pg_publication.c
@@ -54,7 +54,8 @@ check_publication_add_relation(Relation targetrel)
{
/* Must be a regular or partitioned table */
if (RelationGetForm(targetrel)->relkind != RELKIND_RELATION &&
- RelationGetForm(targetrel)->relkind != RELKIND_PARTITIONED_TABLE)
+ RelationGetForm(targetrel)->relkind != RELKIND_PARTITIONED_TABLE &&
+ RelationGetForm(targetrel)->relkind != RELKIND_SEQUENCE)
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("cannot add relation \"%s\" to publication",
@@ -131,7 +132,8 @@ static bool
is_publishable_class(Oid relid, Form_pg_class reltuple)
{
return (reltuple->relkind == RELKIND_RELATION ||
- reltuple->relkind == RELKIND_PARTITIONED_TABLE) &&
+ reltuple->relkind == RELKIND_PARTITIONED_TABLE ||
+ reltuple->relkind == RELKIND_SEQUENCE) &&
!IsCatalogRelationOid(relid) &&
reltuple->relpersistence == RELPERSISTENCE_PERMANENT &&
relid >= FirstNormalObjectId;
@@ -480,6 +482,11 @@ GetPublicationRelations(Oid pubid, PublicationPartOpt pub_partopt)
Form_pg_publication_rel pubrel;
pubrel = (Form_pg_publication_rel) GETSTRUCT(tup);
+
+ /* skip sequences here */
+ if (get_rel_relkind(pubrel->prrelid) == RELKIND_SEQUENCE)
+ continue;
+
result = GetPubPartitionOptionRelations(result, pub_partopt,
pubrel->prrelid);
}
@@ -490,6 +497,49 @@ GetPublicationRelations(Oid pubid, PublicationPartOpt pub_partopt)
return result;
}
+/*
+ * Gets list of relation oids for a publication (sequences only).
+ *
+ * This should only be used for normal publications, the FOR ALL TABLES
+ * should use GetAllSequencesPublicationRelations().
+ */
+List *
+GetPublicationSequenceRelations(Oid pubid)
+{
+ List *result;
+ Relation pubrelsrel;
+ ScanKeyData scankey;
+ SysScanDesc scan;
+ HeapTuple tup;
+
+ /* Find all publications associated with the relation. */
+ pubrelsrel = table_open(PublicationRelRelationId, AccessShareLock);
+
+ ScanKeyInit(&scankey,
+ Anum_pg_publication_rel_prpubid,
+ BTEqualStrategyNumber, F_OIDEQ,
+ ObjectIdGetDatum(pubid));
+
+ scan = systable_beginscan(pubrelsrel, PublicationRelPrrelidPrpubidIndexId,
+ true, NULL, 1, &scankey);
+
+ result = NIL;
+ while (HeapTupleIsValid(tup = systable_getnext(scan)))
+ {
+ Form_pg_publication_rel pubrel;
+
+ pubrel = (Form_pg_publication_rel) GETSTRUCT(tup);
+
+ if (get_rel_relkind(pubrel->prrelid) == RELKIND_SEQUENCE)
+ result = lappend_oid(result, pubrel->prrelid);
+ }
+
+ systable_endscan(scan);
+ table_close(pubrelsrel, AccessShareLock);
+
+ return result;
+}
+
/*
* Gets list of publication oids for publications marked as FOR ALL TABLES.
*/
@@ -735,6 +785,46 @@ GetAllSchemaPublicationRelations(Oid pubid, PublicationPartOpt pub_partopt)
return result;
}
+/*
+ * Gets list of all relation published by FOR ALL TABLES publication(s).
+ *
+ * If the publication publishes partition changes via their respective root
+ * partitioned tables, we must exclude partitions in favor of including the
+ * root partitioned tables.
+ */
+List *
+GetAllSequencesPublicationRelations(void)
+{
+ Relation classRel;
+ ScanKeyData key[1];
+ TableScanDesc scan;
+ HeapTuple tuple;
+ List *result = NIL;
+
+ classRel = table_open(RelationRelationId, AccessShareLock);
+
+ ScanKeyInit(&key[0],
+ Anum_pg_class_relkind,
+ BTEqualStrategyNumber, F_CHAREQ,
+ CharGetDatum(RELKIND_SEQUENCE));
+
+ scan = table_beginscan_catalog(classRel, 1, key);
+
+ while ((tuple = heap_getnext(scan, ForwardScanDirection)) != NULL)
+ {
+ Form_pg_class relForm = (Form_pg_class) GETSTRUCT(tuple);
+ Oid relid = relForm->oid;
+
+ if (is_publishable_class(relid, relForm))
+ result = lappend_oid(result, relid);
+ }
+
+ table_endscan(scan);
+
+ table_close(classRel, AccessShareLock);
+ return result;
+}
+
/*
* Get publication using oid
*
@@ -757,10 +847,12 @@ GetPublication(Oid pubid)
pub->oid = pubid;
pub->name = pstrdup(NameStr(pubform->pubname));
pub->alltables = pubform->puballtables;
+ pub->allsequences = pubform->puballsequences;
pub->pubactions.pubinsert = pubform->pubinsert;
pub->pubactions.pubupdate = pubform->pubupdate;
pub->pubactions.pubdelete = pubform->pubdelete;
pub->pubactions.pubtruncate = pubform->pubtruncate;
+ pub->pubactions.pubsequence = pubform->pubsequence;
pub->pubviaroot = pubform->pubviaroot;
ReleaseSysCache(tup);
@@ -915,3 +1007,56 @@ pg_get_publication_tables(PG_FUNCTION_ARGS)
SRF_RETURN_DONE(funcctx);
}
+
+/*
+ * Returns Oids of sequences in a publication.
+ */
+Datum
+pg_get_publication_sequences(PG_FUNCTION_ARGS)
+{
+ FuncCallContext *funcctx;
+ char *pubname = text_to_cstring(PG_GETARG_TEXT_PP(0));
+ Publication *publication;
+ List *sequences;
+
+ /* stuff done only on the first call of the function */
+ if (SRF_IS_FIRSTCALL())
+ {
+ MemoryContext oldcontext;
+
+ /* create a function context for cross-call persistence */
+ funcctx = SRF_FIRSTCALL_INIT();
+
+ /* switch to memory context appropriate for multiple function calls */
+ oldcontext = MemoryContextSwitchTo(funcctx->multi_call_memory_ctx);
+
+ publication = GetPublicationByName(pubname, false);
+
+ /*
+ * Publications support partitioned tables, although all changes are
+ * replicated using leaf partition identity and schema, so we only
+ * need those.
+ */
+ if (publication->allsequences)
+ sequences = GetAllSequencesPublicationRelations();
+ else
+ sequences = GetPublicationSequenceRelations(publication->oid);
+
+ funcctx->user_fctx = (void *) sequences;
+
+ MemoryContextSwitchTo(oldcontext);
+ }
+
+ /* stuff done on every call of the function */
+ funcctx = SRF_PERCALL_SETUP();
+ sequences = (List *) funcctx->user_fctx;
+
+ if (funcctx->call_cntr < list_length(sequences))
+ {
+ Oid relid = list_nth_oid(sequences, funcctx->call_cntr);
+
+ SRF_RETURN_NEXT(funcctx, ObjectIdGetDatum(relid));
+ }
+
+ SRF_RETURN_DONE(funcctx);
+}
diff --git a/src/backend/catalog/system_views.sql b/src/backend/catalog/system_views.sql
index eb560955cd..2c12d373c1 100644
--- a/src/backend/catalog/system_views.sql
+++ b/src/backend/catalog/system_views.sql
@@ -372,6 +372,16 @@ CREATE VIEW pg_publication_tables AS
pg_class C JOIN pg_namespace N ON (N.oid = C.relnamespace)
WHERE C.oid = GPT.relid;
+CREATE VIEW pg_publication_sequences AS
+ SELECT
+ P.pubname AS pubname,
+ N.nspname AS schemaname,
+ C.relname AS sequencename
+ FROM pg_publication P,
+ LATERAL pg_get_publication_sequences(P.pubname) GPT,
+ pg_class C JOIN pg_namespace N ON (N.oid = C.relnamespace)
+ WHERE C.oid = GPT.relid;
+
CREATE VIEW pg_locks AS
SELECT * FROM pg_lock_status() AS L;
diff --git a/src/backend/commands/publicationcmds.c b/src/backend/commands/publicationcmds.c
index 7d4a0e95f6..e8dfb1a068 100644
--- a/src/backend/commands/publicationcmds.c
+++ b/src/backend/commands/publicationcmds.c
@@ -16,6 +16,7 @@
#include "access/genam.h"
#include "access/htup_details.h"
+#include "access/relation.h"
#include "access/table.h"
#include "access/xact.h"
#include "catalog/catalog.h"
@@ -59,6 +60,12 @@ static void PublicationAddSchemas(Oid pubid, List *schemas, bool if_not_exists,
AlterPublicationStmt *stmt);
static void PublicationDropSchemas(Oid pubid, List *schemas, bool missing_ok);
+static List *OpenSequenceList(List *sequences);
+static void CloseSequenceList(List *rels);
+static void PublicationAddSequences(Oid pubid, List *rels, bool if_not_exists,
+ AlterPublicationStmt *stmt);
+static void PublicationDropSequences(Oid pubid, List *rels, bool missing_ok);
+
static void
parse_publication_options(ParseState *pstate,
List *options,
@@ -77,6 +84,7 @@ parse_publication_options(ParseState *pstate,
pubactions->pubupdate = true;
pubactions->pubdelete = true;
pubactions->pubtruncate = true;
+ pubactions->pubsequence = true;
*publish_via_partition_root = false;
/* Parse options */
@@ -101,6 +109,7 @@ parse_publication_options(ParseState *pstate,
pubactions->pubupdate = false;
pubactions->pubdelete = false;
pubactions->pubtruncate = false;
+ pubactions->pubsequence = false;
*publish_given = true;
publish = defGetString(defel);
@@ -123,6 +132,8 @@ parse_publication_options(ParseState *pstate,
pubactions->pubdelete = true;
else if (strcmp(publish_opt, "truncate") == 0)
pubactions->pubtruncate = true;
+ else if (strcmp(publish_opt, "sequence") == 0)
+ pubactions->pubsequence = true;
else
ereport(ERROR,
(errcode(ERRCODE_SYNTAX_ERROR),
@@ -149,7 +160,9 @@ parse_publication_options(ParseState *pstate,
*/
static void
ObjectsInPublicationToOids(List *pubobjspec_list, ParseState *pstate,
- List **rels, List **schemas)
+ List **tables, List **sequences,
+ List **tables_schemas, List **sequences_schemas,
+ List **schemas)
{
ListCell *cell;
PublicationObjSpec *pubobj;
@@ -167,12 +180,23 @@ ObjectsInPublicationToOids(List *pubobjspec_list, ParseState *pstate,
switch (pubobj->pubobjtype)
{
case PUBLICATIONOBJ_TABLE:
- *rels = lappend(*rels, pubobj->pubtable);
+ *tables = lappend(*tables, pubobj->pubtable);
+ break;
+ case PUBLICATIONOBJ_SEQUENCE:
+ *sequences = lappend(*sequences, pubobj->pubtable);
break;
case PUBLICATIONOBJ_TABLE_IN_SCHEMA:
schemaid = get_namespace_oid(pubobj->name, false);
/* Filter out duplicates if user specifies "sch1, sch1" */
+ *tables_schemas = list_append_unique_oid(*tables_schemas, schemaid);
+ *schemas = list_append_unique_oid(*schemas, schemaid);
+ break;
+ case PUBLICATIONOBJ_SEQUENCE_IN_SCHEMA:
+ schemaid = get_namespace_oid(pubobj->name, false);
+
+ /* Filter out duplicates if user specifies "sch1, sch1" */
+ *sequences_schemas = list_append_unique_oid(*sequences_schemas, schemaid);
*schemas = list_append_unique_oid(*schemas, schemaid);
break;
case PUBLICATIONOBJ_TABLE_IN_CUR_SCHEMA:
@@ -186,6 +210,21 @@ ObjectsInPublicationToOids(List *pubobjspec_list, ParseState *pstate,
list_free(search_path);
/* Filter out duplicates if user specifies "sch1, sch1" */
+ *tables_schemas = list_append_unique_oid(*tables_schemas, schemaid);
+ *schemas = list_append_unique_oid(*schemas, schemaid);
+ break;
+ case PUBLICATIONOBJ_SEQUENCE_IN_CUR_SCHEMA:
+ search_path = fetch_search_path(false);
+ if (search_path == NIL) /* nothing valid in search_path? */
+ ereport(ERROR,
+ errcode(ERRCODE_UNDEFINED_SCHEMA),
+ errmsg("no schema has been selected for CURRENT_SCHEMA"));
+
+ schemaid = linitial_oid(search_path);
+ list_free(search_path);
+
+ /* Filter out duplicates if user specifies "sch1, sch1" */
+ *sequences_schemas = list_append_unique_oid(*sequences_schemas, schemaid);
*schemas = list_append_unique_oid(*schemas, schemaid);
break;
default:
@@ -251,7 +290,10 @@ CreatePublication(ParseState *pstate, CreatePublicationStmt *stmt)
bool publish_via_partition_root_given;
bool publish_via_partition_root;
AclResult aclresult;
- List *relations = NIL;
+ List *tables = NIL;
+ List *sequences = NIL;
+ List *tables_schemaidlist = NIL;
+ List *sequences_schemaidlist = NIL;
List *schemaidlist = NIL;
/* must have CREATE privilege on database */
@@ -306,6 +348,8 @@ CreatePublication(ParseState *pstate, CreatePublicationStmt *stmt)
BoolGetDatum(pubactions.pubdelete);
values[Anum_pg_publication_pubtruncate - 1] =
BoolGetDatum(pubactions.pubtruncate);
+ values[Anum_pg_publication_pubsequence - 1] =
+ BoolGetDatum(pubactions.pubsequence);
values[Anum_pg_publication_pubviaroot - 1] =
BoolGetDatum(publish_via_partition_root);
@@ -330,26 +374,40 @@ CreatePublication(ParseState *pstate, CreatePublicationStmt *stmt)
}
else
{
- ObjectsInPublicationToOids(stmt->pubobjects, pstate, &relations,
+ ObjectsInPublicationToOids(stmt->pubobjects, pstate,
+ &tables, &sequences,
+ &tables_schemaidlist,
+ &sequences_schemaidlist,
&schemaidlist);
/* FOR ALL TABLES IN SCHEMA requires superuser */
- if (list_length(schemaidlist) > 0 && !superuser())
+ if (list_length(tables_schemaidlist) > 0 && !superuser())
ereport(ERROR,
errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
errmsg("must be superuser to create FOR ALL TABLES IN SCHEMA publication"));
- if (list_length(relations) > 0)
+ if (list_length(tables) > 0)
{
List *rels;
- rels = OpenTableList(relations);
- CheckObjSchemaNotAlreadyInPublication(rels, schemaidlist,
+ rels = OpenTableList(tables);
+ CheckObjSchemaNotAlreadyInPublication(rels, tables_schemaidlist,
PUBLICATIONOBJ_TABLE);
PublicationAddTables(puboid, rels, true, NULL);
CloseTableList(rels);
}
+ if (list_length(sequences) > 0)
+ {
+ List *rels;
+
+ rels = OpenSequenceList(sequences);
+ CheckObjSchemaNotAlreadyInPublication(rels, sequences_schemaidlist,
+ PUBLICATIONOBJ_SEQUENCE);
+ PublicationAddTables(puboid, rels, true, NULL);
+ CloseSequenceList(rels);
+ }
+
if (list_length(schemaidlist) > 0)
{
/*
@@ -651,12 +709,13 @@ AlterPublicationSchemas(AlterPublicationStmt *stmt,
*/
static void
CheckAlterPublication(AlterPublicationStmt *stmt, HeapTuple tup,
- List *tables, List *schemaidlist)
+ List *tables, List *tables_schemaidlist,
+ List *sequences, List *sequences_schemaidlist)
{
Form_pg_publication pubform = (Form_pg_publication) GETSTRUCT(tup);
if ((stmt->action == DEFELEM_ADD || stmt->action == DEFELEM_SET) &&
- schemaidlist && !superuser())
+ (tables_schemaidlist || sequences_schemaidlist) && !superuser())
ereport(ERROR,
(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
errmsg("must be superuser to add or set schemas")));
@@ -665,13 +724,24 @@ CheckAlterPublication(AlterPublicationStmt *stmt, HeapTuple tup,
* Check that user is allowed to manipulate the publication tables in
* schema
*/
- if (schemaidlist && pubform->puballtables)
+ if (tables_schemaidlist && pubform->puballtables)
ereport(ERROR,
(errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
errmsg("publication \"%s\" is defined as FOR ALL TABLES",
NameStr(pubform->pubname)),
errdetail("Tables from schema cannot be added to, dropped from, or set on FOR ALL TABLES publications.")));
+ /*
+ * Check that user is allowed to manipulate the publication sequences in
+ * schema
+ */
+ if (sequences_schemaidlist && pubform->puballsequences)
+ ereport(ERROR,
+ (errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
+ errmsg("publication \"%s\" is defined as FOR ALL SEQUENCES",
+ NameStr(pubform->pubname)),
+ errdetail("Sequences from schema cannot be added to, dropped from, or set on FOR ALL SEQUENCES publications.")));
+
/* Check that user is allowed to manipulate the publication tables. */
if (tables && pubform->puballtables)
ereport(ERROR,
@@ -679,6 +749,108 @@ CheckAlterPublication(AlterPublicationStmt *stmt, HeapTuple tup,
errmsg("publication \"%s\" is defined as FOR ALL TABLES",
NameStr(pubform->pubname)),
errdetail("Tables cannot be added to or dropped from FOR ALL TABLES publications.")));
+
+ /* Check that user is allowed to manipulate the publication tables. */
+ if (sequences && pubform->puballsequences)
+ ereport(ERROR,
+ (errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
+ errmsg("publication \"%s\" is defined as FOR ALL SEQUENCES",
+ NameStr(pubform->pubname)),
+ errdetail("Sequences cannot be added to or dropped from FOR ALL SEQUENCES publications.")));
+}
+
+/*
+ * Add or remove sequence to/from publication.
+ */
+static void
+AlterPublicationSequences(AlterPublicationStmt *stmt, HeapTuple tup,
+ List *sequences, List *schemaidlist)
+{
+ List *rels = NIL;
+ Form_pg_publication pubform = (Form_pg_publication) GETSTRUCT(tup);
+ Oid pubid = pubform->oid;
+
+ /*
+ * It is quite possible that for the SET case user has not specified any
+ * tables in which case we need to remove all the existing tables.
+ */
+ if (!sequences && stmt->action != DEFELEM_SET)
+ return;
+
+ rels = OpenSequenceList(sequences);
+
+ if (stmt->action == DEFELEM_ADD)
+ {
+ List *schemas = NIL;
+
+ /*
+ * Check if the relation is member of the existing schema in the
+ * publication or member of the schema list specified.
+ */
+ schemas = list_concat_copy(schemaidlist, GetPublicationSchemas(pubid));
+ CheckObjSchemaNotAlreadyInPublication(rels, schemas,
+ PUBLICATIONOBJ_SEQUENCE);
+ PublicationAddSequences(pubid, rels, false, stmt);
+ }
+ else if (stmt->action == DEFELEM_DROP)
+ PublicationDropSequences(pubid, rels, false);
+ else /* DEFELEM_SET */
+ {
+ List *oldrelids = GetPublicationRelations(pubid,
+ PUBLICATION_PART_ROOT);
+ List *delrels = NIL;
+ ListCell *oldlc;
+
+ CheckObjSchemaNotAlreadyInPublication(rels, schemaidlist,
+ PUBLICATIONOBJ_SEQUENCE);
+
+ /* Calculate which relations to drop. */
+ foreach(oldlc, oldrelids)
+ {
+ Oid oldrelid = lfirst_oid(oldlc);
+ ListCell *newlc;
+ bool found = false;
+
+ foreach(newlc, rels)
+ {
+ PublicationRelInfo *newpubrel;
+
+ newpubrel = (PublicationRelInfo *) lfirst(newlc);
+ if (RelationGetRelid(newpubrel->relation) == oldrelid)
+ {
+ found = true;
+ break;
+ }
+ }
+ /* Not yet in the list, open it and add to the list */
+ if (!found)
+ {
+ Relation oldrel;
+ PublicationRelInfo *pubrel;
+
+ /* Wrap relation into PublicationRelInfo */
+ oldrel = table_open(oldrelid, ShareUpdateExclusiveLock);
+
+ pubrel = palloc(sizeof(PublicationRelInfo));
+ pubrel->relation = oldrel;
+
+ delrels = lappend(delrels, pubrel);
+ }
+ }
+
+ /* And drop them. */
+ PublicationDropSequences(pubid, delrels, true);
+
+ /*
+ * Don't bother calculating the difference for adding, we'll catch and
+ * skip existing ones when doing catalog update.
+ */
+ PublicationAddSequences(pubid, rels, true, stmt);
+
+ CloseSequenceList(delrels);
+ }
+
+ CloseSequenceList(rels);
}
/*
@@ -716,13 +888,21 @@ AlterPublication(ParseState *pstate, AlterPublicationStmt *stmt)
AlterPublicationOptions(pstate, stmt, rel, tup);
else
{
- List *relations = NIL;
+ List *tables = NIL;
+ List *sequences = NIL;
+ List *tables_schemaidlist = NIL;
+ List *sequences_schemaidlist = NIL;
List *schemaidlist = NIL;
- ObjectsInPublicationToOids(stmt->pubobjects, pstate, &relations,
+ ObjectsInPublicationToOids(stmt->pubobjects, pstate,
+ &tables, &sequences,
+ &tables_schemaidlist,
+ &sequences_schemaidlist,
&schemaidlist);
- CheckAlterPublication(stmt, tup, relations, schemaidlist);
+ CheckAlterPublication(stmt, tup,
+ tables, tables_schemaidlist,
+ sequences, sequences_schemaidlist);
/*
* Lock the publication so nobody else can do anything with it. This
@@ -747,7 +927,9 @@ AlterPublication(ParseState *pstate, AlterPublicationStmt *stmt)
errmsg("publication \"%s\" does not exist",
stmt->pubname));
- AlterPublicationTables(stmt, tup, relations, schemaidlist);
+ AlterPublicationTables(stmt, tup, tables, tables_schemaidlist);
+ AlterPublicationSequences(stmt, tup, sequences, sequences_schemaidlist);
+
AlterPublicationSchemas(stmt, tup, schemaidlist);
}
@@ -1155,6 +1337,144 @@ PublicationDropSchemas(Oid pubid, List *schemas, bool missing_ok)
}
}
+/*
+ * Open relations specified by a PublicationTable list.
+ * In the returned list of PublicationRelInfo, tables are locked
+ * in ShareUpdateExclusiveLock mode in order to add them to a publication.
+ */
+static List *
+OpenSequenceList(List *sequences)
+{
+ List *relids = NIL;
+ List *rels = NIL;
+ ListCell *lc;
+
+ /*
+ * Open, share-lock, and check all the explicitly-specified relations
+ */
+ foreach(lc, sequences)
+ {
+ PublicationTable *s = lfirst_node(PublicationTable, lc);
+ Relation rel;
+ Oid myrelid;
+ PublicationRelInfo *pub_rel;
+
+ /* Allow query cancel in case this takes a long time */
+ CHECK_FOR_INTERRUPTS();
+
+ rel = table_openrv(s->relation, ShareUpdateExclusiveLock);
+ myrelid = RelationGetRelid(rel);
+
+ /*
+ * Filter out duplicates if user specifies "foo, foo".
+ *
+ * Note that this algorithm is known to not be very efficient (O(N^2))
+ * but given that it only works on list of tables given to us by user
+ * it's deemed acceptable.
+ */
+ if (list_member_oid(relids, myrelid))
+ {
+ table_close(rel, ShareUpdateExclusiveLock);
+ continue;
+ }
+
+ pub_rel = palloc(sizeof(PublicationRelInfo));
+ pub_rel->relation = rel;
+ rels = lappend(rels, pub_rel);
+ relids = lappend_oid(relids, myrelid);
+ }
+
+ list_free(relids);
+
+ return rels;
+}
+
+/*
+ * Close all relations in the list.
+ */
+static void
+CloseSequenceList(List *rels)
+{
+ ListCell *lc;
+
+ foreach(lc, rels)
+ {
+ PublicationRelInfo *pub_rel;
+
+ pub_rel = (PublicationRelInfo *) lfirst(lc);
+ table_close(pub_rel->relation, NoLock);
+ }
+}
+
+/*
+ * Add listed tables to the publication.
+ */
+static void
+PublicationAddSequences(Oid pubid, List *rels, bool if_not_exists,
+ AlterPublicationStmt *stmt)
+{
+ ListCell *lc;
+
+ Assert(!stmt || !stmt->for_all_sequences);
+
+ foreach(lc, rels)
+ {
+ PublicationRelInfo *pub_rel = (PublicationRelInfo *) lfirst(lc);
+ Relation rel = pub_rel->relation;
+ ObjectAddress obj;
+
+ /* Must be owner of the sequence or superuser. */
+ if (!pg_class_ownercheck(RelationGetRelid(rel), GetUserId()))
+ aclcheck_error(ACLCHECK_NOT_OWNER, get_relkind_objtype(rel->rd_rel->relkind),
+ RelationGetRelationName(rel));
+
+ obj = publication_add_relation(pubid, pub_rel, if_not_exists);
+ if (stmt)
+ {
+ EventTriggerCollectSimpleCommand(obj, InvalidObjectAddress,
+ (Node *) stmt);
+
+ InvokeObjectPostCreateHook(PublicationRelRelationId,
+ obj.objectId, 0);
+ }
+ }
+}
+
+/*
+ * Remove listed sequences from the publication.
+ */
+static void
+PublicationDropSequences(Oid pubid, List *rels, bool missing_ok)
+{
+ ObjectAddress obj;
+ ListCell *lc;
+ Oid prid;
+
+ foreach(lc, rels)
+ {
+ Relation rel = (Relation) lfirst(lc);
+ Oid relid = RelationGetRelid(rel);
+
+ prid = GetSysCacheOid2(PUBLICATIONRELMAP, Anum_pg_publication_rel_oid,
+ ObjectIdGetDatum(relid),
+ ObjectIdGetDatum(pubid));
+ if (!OidIsValid(prid))
+ {
+ if (missing_ok)
+ continue;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_UNDEFINED_OBJECT),
+ errmsg("relation \"%s\" is not part of the publication",
+ RelationGetRelationName(rel))));
+ }
+
+ ObjectAddressSet(obj, PublicationRelRelationId, prid);
+ performDeletion(&obj, DROP_CASCADE, 0);
+ }
+}
+
+
/*
* Internal workhorse for changing a publication owner
*/
diff --git a/src/backend/commands/sequence.c b/src/backend/commands/sequence.c
index 92d1966f34..b22f83ea82 100644
--- a/src/backend/commands/sequence.c
+++ b/src/backend/commands/sequence.c
@@ -347,6 +347,85 @@ ResetSequence(Oid seq_relid)
relation_close(seq_rel, NoLock);
}
+/*
+ * Reset a sequence to its initial value.
+ *
+ * The change is made transactionally, so that on failure of the current
+ * transaction, the sequence will be restored to its previous state.
+ * We do that by creating a whole new relfilenode for the sequence; so this
+ * works much like the rewriting forms of ALTER TABLE.
+ *
+ * Caller is assumed to have acquired AccessExclusiveLock on the sequence,
+ * which must not be released until end of transaction. Caller is also
+ * responsible for permissions checking.
+ */
+void
+ResetSequence2(Oid seq_relid, int64 last_value, int64 log_cnt, bool is_called)
+{
+ Relation seq_rel;
+ SeqTable elm;
+ Form_pg_sequence_data seq;
+ Buffer buf;
+ HeapTupleData seqdatatuple;
+ HeapTuple tuple;
+
+ /*
+ * Read the old sequence. This does a bit more work than really
+ * necessary, but it's simple, and we do want to double-check that it's
+ * indeed a sequence.
+ */
+ init_sequence(seq_relid, &elm, &seq_rel);
+ (void) read_seq_tuple(seq_rel, &buf, &seqdatatuple);
+
+ /*
+ * Copy the existing sequence tuple.
+ */
+ tuple = heap_copytuple(&seqdatatuple);
+
+ /* Now we're done with the old page */
+ UnlockReleaseBuffer(buf);
+
+ /*
+ * Modify the copied tuple to execute the restart (compare the RESTART
+ * action in AlterSequence)
+ */
+ seq = (Form_pg_sequence_data) GETSTRUCT(tuple);
+ seq->last_value = last_value;
+ seq->is_called = is_called;
+ seq->log_cnt = log_cnt;
+
+ /*
+ * Create a new storage file for the sequence.
+ */
+ RelationSetNewRelfilenode(seq_rel, seq_rel->rd_rel->relpersistence);
+
+ /*
+ * Ensure sequence's relfrozenxid is at 0, since it won't contain any
+ * unfrozen XIDs. Same with relminmxid, since a sequence will never
+ * contain multixacts.
+ */
+ Assert(seq_rel->rd_rel->relfrozenxid == InvalidTransactionId);
+ Assert(seq_rel->rd_rel->relminmxid == InvalidMultiXactId);
+
+ /*
+ * Insert the modified tuple into the new storage file.
+ *
+ * XXX Maybe this should also use created=true, just like the other places
+ * calling fill_seq_with_data. That's probably needed for correct cascading
+ * replication.
+ *
+ * XXX That'd mean all fill_seq_with_data callers use created=true, making
+ * the parameter unnecessary.
+ */
+ fill_seq_with_data(seq_rel, tuple);
+
+ /* Clear local cache so that we don't think we have cached numbers */
+ /* Note that we do not change the currval() state */
+ elm->cached = elm->last;
+
+ relation_close(seq_rel, NoLock);
+}
+
/*
* Initialize a sequence's relation with the specified tuple as content
*/
diff --git a/src/backend/commands/subscriptioncmds.c b/src/backend/commands/subscriptioncmds.c
index c47ba26369..5086526904 100644
--- a/src/backend/commands/subscriptioncmds.c
+++ b/src/backend/commands/subscriptioncmds.c
@@ -84,6 +84,7 @@ typedef struct SubOpts
} SubOpts;
static List *fetch_table_list(WalReceiverConn *wrconn, List *publications);
+static List *fetch_sequence_list(WalReceiverConn *wrconn, List *publications);
static void check_duplicates_in_publist(List *publist, Datum *datums);
static List *merge_publications(List *oldpublist, List *newpublist, bool addpub, const char *subname);
static void ReportSlotConnectionError(List *rstates, Oid subid, char *slotname, char *err);
@@ -502,6 +503,7 @@ CreateSubscription(ParseState *pstate, CreateSubscriptionStmt *stmt,
char *err;
WalReceiverConn *wrconn;
List *tables;
+ List *sequences;
ListCell *lc;
char table_state;
@@ -540,6 +542,26 @@ CreateSubscription(ParseState *pstate, CreateSubscriptionStmt *stmt,
InvalidXLogRecPtr);
}
+ /*
+ * Get the sequence list from publisher and build local sequence
+ * status info.
+ */
+ sequences = fetch_sequence_list(wrconn, publications);
+ foreach(lc, sequences)
+ {
+ RangeVar *rv = (RangeVar *) lfirst(lc);
+ Oid relid;
+
+ relid = RangeVarGetRelid(rv, AccessShareLock, false);
+
+ /* Check for supported relkind. */
+ CheckSubscriptionRelkind(get_rel_relkind(relid),
+ rv->schemaname, rv->relname);
+
+ AddSubscriptionRelState(subid, relid, table_state,
+ InvalidXLogRecPtr);
+ }
+
/*
* If requested, create permanent slot for the subscription. We
* won't use the initial snapshot for anything, so no need to
@@ -712,6 +734,10 @@ AlterSubscription_refresh(Subscription *sub, bool copy_data)
{
Oid relid = subrel_local_oids[off];
+ /* XXX ignore sequences - maybe do this in GetSubscriptionRelations? */
+ if (get_rel_relkind(relid) == RELKIND_SEQUENCE)
+ continue;
+
if (!bsearch(&relid, pubrel_local_oids,
list_length(pubrel_names), sizeof(Oid), oid_cmp))
{
@@ -803,6 +829,183 @@ AlterSubscription_refresh(Subscription *sub, bool copy_data)
ReplicationSlotDropAtPubNode(wrconn, syncslotname, true);
}
}
+
+ /*
+ * XXX now do the same thing for sequences, maybe before the preceding
+ * block, or earlier?
+ */
+
+ /* Get the table list from publisher. */
+ pubrel_names = fetch_sequence_list(wrconn, sub->publications);
+
+ /* Get local table list. */
+ subrel_states = GetSubscriptionRelations(sub->oid);
+
+ /*
+ * Build qsorted array of local table oids for faster lookup. This can
+ * potentially contain all tables in the database so speed of lookup
+ * is important.
+ */
+ subrel_local_oids = palloc(list_length(subrel_states) * sizeof(Oid));
+ off = 0;
+ foreach(lc, subrel_states)
+ {
+ SubscriptionRelState *relstate = (SubscriptionRelState *) lfirst(lc);
+
+ subrel_local_oids[off++] = relstate->relid;
+ }
+ qsort(subrel_local_oids, list_length(subrel_states),
+ sizeof(Oid), oid_cmp);
+
+ /*
+ * Rels that we want to remove from subscription and drop any slots
+ * and origins corresponding to them.
+ */
+ sub_remove_rels = palloc(list_length(subrel_states) * sizeof(SubRemoveRels));
+
+ /*
+ * Walk over the remote tables and try to match them to locally known
+ * tables. If the table is not known locally create a new state for
+ * it.
+ *
+ * Also builds array of local oids of remote tables for the next step.
+ */
+ off = 0;
+ pubrel_local_oids = palloc(list_length(pubrel_names) * sizeof(Oid));
+
+ foreach(lc, pubrel_names)
+ {
+ RangeVar *rv = (RangeVar *) lfirst(lc);
+ Oid relid;
+
+ relid = RangeVarGetRelid(rv, AccessShareLock, false);
+
+ /* Check for supported relkind. */
+ CheckSubscriptionRelkind(get_rel_relkind(relid),
+ rv->schemaname, rv->relname);
+
+ pubrel_local_oids[off++] = relid;
+
+ if (!bsearch(&relid, subrel_local_oids,
+ list_length(subrel_states), sizeof(Oid), oid_cmp))
+ {
+ AddSubscriptionRelState(sub->oid, relid,
+ copy_data ? SUBREL_STATE_INIT : SUBREL_STATE_READY,
+ InvalidXLogRecPtr);
+ ereport(DEBUG1,
+ (errmsg_internal("table \"%s.%s\" added to subscription \"%s\"",
+ rv->schemaname, rv->relname, sub->name)));
+ }
+ }
+
+ /*
+ * Next remove state for tables we should not care about anymore using
+ * the data we collected above
+ */
+ qsort(pubrel_local_oids, list_length(pubrel_names),
+ sizeof(Oid), oid_cmp);
+
+ remove_rel_len = 0;
+ for (off = 0; off < list_length(subrel_states); off++)
+ {
+ Oid relid = subrel_local_oids[off];
+
+ /* XXX ignore non-sequences - maybe do this in GetSubscriptionRelations? */
+ if (get_rel_relkind(relid) != RELKIND_SEQUENCE)
+ continue;
+
+ if (!bsearch(&relid, pubrel_local_oids,
+ list_length(pubrel_names), sizeof(Oid), oid_cmp))
+ {
+ char state;
+ XLogRecPtr statelsn;
+
+ /*
+ * Lock pg_subscription_rel with AccessExclusiveLock to
+ * prevent any race conditions with the apply worker
+ * re-launching workers at the same time this code is trying
+ * to remove those tables.
+ *
+ * Even if new worker for this particular rel is restarted it
+ * won't be able to make any progress as we hold exclusive
+ * lock on subscription_rel till the transaction end. It will
+ * simply exit as there is no corresponding rel entry.
+ *
+ * This locking also ensures that the state of rels won't
+ * change till we are done with this refresh operation.
+ */
+ if (!rel)
+ rel = table_open(SubscriptionRelRelationId, AccessExclusiveLock);
+
+ /* Last known rel state. */
+ state = GetSubscriptionRelState(sub->oid, relid, &statelsn);
+
+ sub_remove_rels[remove_rel_len].relid = relid;
+ sub_remove_rels[remove_rel_len++].state = state;
+
+ RemoveSubscriptionRel(sub->oid, relid);
+
+ logicalrep_worker_stop(sub->oid, relid);
+
+ /*
+ * For READY state, we would have already dropped the
+ * tablesync origin.
+ */
+ if (state != SUBREL_STATE_READY)
+ {
+ char originname[NAMEDATALEN];
+
+ /*
+ * Drop the tablesync's origin tracking if exists.
+ *
+ * It is possible that the origin is not yet created for
+ * tablesync worker, this can happen for the states before
+ * SUBREL_STATE_FINISHEDCOPY. The apply worker can also
+ * concurrently try to drop the origin and by this time
+ * the origin might be already removed. For these reasons,
+ * passing missing_ok = true.
+ */
+ ReplicationOriginNameForTablesync(sub->oid, relid, originname,
+ sizeof(originname));
+ replorigin_drop_by_name(originname, true, false);
+ }
+
+ ereport(DEBUG1,
+ (errmsg_internal("table \"%s.%s\" removed from subscription \"%s\"",
+ get_namespace_name(get_rel_namespace(relid)),
+ get_rel_name(relid),
+ sub->name)));
+ }
+ }
+
+ /*
+ * Drop the tablesync slots associated with removed tables. This has
+ * to be at the end because otherwise if there is an error while doing
+ * the database operations we won't be able to rollback dropped slots.
+ */
+ for (off = 0; off < remove_rel_len; off++)
+ {
+ if (sub_remove_rels[off].state != SUBREL_STATE_READY &&
+ sub_remove_rels[off].state != SUBREL_STATE_SYNCDONE)
+ {
+ char syncslotname[NAMEDATALEN] = {0};
+
+ /*
+ * For READY/SYNCDONE states we know the tablesync slot has
+ * already been dropped by the tablesync worker.
+ *
+ * For other states, there is no certainty, maybe the slot
+ * does not exist yet. Also, if we fail after removing some of
+ * the slots, next time, it will again try to drop already
+ * dropped slots and fail. For these reasons, we allow
+ * missing_ok = true for the drop.
+ */
+ ReplicationSlotNameForTablesync(sub->oid, sub_remove_rels[off].relid,
+ syncslotname, sizeof(syncslotname));
+ ReplicationSlotDropAtPubNode(wrconn, syncslotname, true);
+ }
+ }
+
}
PG_FINALLY();
{
@@ -1607,6 +1810,75 @@ fetch_table_list(WalReceiverConn *wrconn, List *publications)
return tablelist;
}
+/*
+ * Get the list of sequences which belong to specified publications on the
+ * publisher connection.
+ */
+static List *
+fetch_sequence_list(WalReceiverConn *wrconn, List *publications)
+{
+ WalRcvExecResult *res;
+ StringInfoData cmd;
+ TupleTableSlot *slot;
+ Oid tableRow[2] = {TEXTOID, TEXTOID};
+ ListCell *lc;
+ bool first;
+ List *tablelist = NIL;
+
+ Assert(list_length(publications) > 0);
+
+ initStringInfo(&cmd);
+ appendStringInfoString(&cmd, "SELECT DISTINCT s.schemaname, s.sequencename\n"
+ " FROM pg_catalog.pg_publication_sequences s\n"
+ " WHERE s.pubname IN (");
+ first = true;
+ foreach(lc, publications)
+ {
+ char *pubname = strVal(lfirst(lc));
+
+ if (first)
+ first = false;
+ else
+ appendStringInfoString(&cmd, ", ");
+
+ appendStringInfoString(&cmd, quote_literal_cstr(pubname));
+ }
+ appendStringInfoChar(&cmd, ')');
+
+ res = walrcv_exec(wrconn, cmd.data, 2, tableRow);
+ pfree(cmd.data);
+
+ if (res->status != WALRCV_OK_TUPLES)
+ ereport(ERROR,
+ (errmsg("could not receive list of replicated tables from the publisher: %s",
+ res->err)));
+
+ /* Process tables. */
+ slot = MakeSingleTupleTableSlot(res->tupledesc, &TTSOpsMinimalTuple);
+ while (tuplestore_gettupleslot(res->tuplestore, true, false, slot))
+ {
+ char *nspname;
+ char *relname;
+ bool isnull;
+ RangeVar *rv;
+
+ nspname = TextDatumGetCString(slot_getattr(slot, 1, &isnull));
+ Assert(!isnull);
+ relname = TextDatumGetCString(slot_getattr(slot, 2, &isnull));
+ Assert(!isnull);
+
+ rv = makeRangeVar(nspname, relname, -1);
+ tablelist = lappend(tablelist, rv);
+
+ ExecClearTuple(slot);
+ }
+ ExecDropSingleTupleTableSlot(slot);
+
+ walrcv_clear_result(res);
+
+ return tablelist;
+}
+
/*
* This is to report the connection failure while dropping replication slots.
* Here, we report the WARNING for all tablesync slots so that user can drop
diff --git a/src/backend/executor/execReplication.c b/src/backend/executor/execReplication.c
index 574d7d27fd..6f0d42d551 100644
--- a/src/backend/executor/execReplication.c
+++ b/src/backend/executor/execReplication.c
@@ -608,7 +608,7 @@ void
CheckSubscriptionRelkind(char relkind, const char *nspname,
const char *relname)
{
- if (relkind != RELKIND_RELATION && relkind != RELKIND_PARTITIONED_TABLE)
+ if (relkind != RELKIND_RELATION && relkind != RELKIND_PARTITIONED_TABLE && relkind != RELKIND_SEQUENCE)
ereport(ERROR,
(errcode(ERRCODE_WRONG_OBJECT_TYPE),
errmsg("cannot use relation \"%s.%s\" as logical replication target",
diff --git a/src/backend/nodes/copyfuncs.c b/src/backend/nodes/copyfuncs.c
index ad1ea2ff2f..3dc5eeed91 100644
--- a/src/backend/nodes/copyfuncs.c
+++ b/src/backend/nodes/copyfuncs.c
@@ -4843,6 +4843,7 @@ _copyCreatePublicationStmt(const CreatePublicationStmt *from)
COPY_NODE_FIELD(options);
COPY_NODE_FIELD(pubobjects);
COPY_SCALAR_FIELD(for_all_tables);
+ COPY_SCALAR_FIELD(for_all_sequences);
return newnode;
}
diff --git a/src/backend/nodes/equalfuncs.c b/src/backend/nodes/equalfuncs.c
index f537d3eb96..2317185daf 100644
--- a/src/backend/nodes/equalfuncs.c
+++ b/src/backend/nodes/equalfuncs.c
@@ -2336,6 +2336,7 @@ _equalAlterPublicationStmt(const AlterPublicationStmt *a,
COMPARE_NODE_FIELD(options);
COMPARE_NODE_FIELD(pubobjects);
COMPARE_SCALAR_FIELD(for_all_tables);
+ COMPARE_SCALAR_FIELD(for_all_sequences);
COMPARE_SCALAR_FIELD(action);
return true;
diff --git a/src/backend/parser/gram.y b/src/backend/parser/gram.y
index a6d0cefa6b..5c4e8bf3d7 100644
--- a/src/backend/parser/gram.y
+++ b/src/backend/parser/gram.y
@@ -9672,6 +9672,26 @@ PublicationObjSpec:
$$->pubobjtype = PUBLICATIONOBJ_TABLE_IN_CUR_SCHEMA;
$$->location = @5;
}
+ | SEQUENCE relation_expr
+ {
+ $$ = makeNode(PublicationObjSpec);
+ $$->pubobjtype = PUBLICATIONOBJ_SEQUENCE;
+ $$->pubtable = makeNode(PublicationTable);
+ $$->pubtable->relation = $2;
+ }
+ | ALL SEQUENCE IN_P SCHEMA ColId
+ {
+ $$ = makeNode(PublicationObjSpec);
+ $$->pubobjtype = PUBLICATIONOBJ_SEQUENCE_IN_SCHEMA;
+ $$->name = $5;
+ $$->location = @5;
+ }
+ | ALL SEQUENCES IN_P SCHEMA CURRENT_SCHEMA
+ {
+ $$ = makeNode(PublicationObjSpec);
+ $$->pubobjtype = PUBLICATIONOBJ_SEQUENCE_IN_CUR_SCHEMA;
+ $$->location = @5;
+ }
| ColId
{
$$ = makeNode(PublicationObjSpec);
@@ -10007,6 +10027,12 @@ UnlistenStmt:
}
;
+/*
+ * FIXME
+ *
+ * opt_publication_for_sequences and publication_for_sequences should be
+ * copies for sequences
+ */
/*****************************************************************************
*
@@ -10015,6 +10041,12 @@ UnlistenStmt:
* BEGIN / COMMIT / ROLLBACK
* (also older versions END / ABORT)
*
+ * ALTER PUBLICATION name ADD SEQUENCE sequence [, sequence2]
+ *
+ * ALTER PUBLICATION name DROP SEQUENCE sequence [, sequence2]
+ *
+ * ALTER PUBLICATION name SET SEQUENCE sequence [, sequence2]
+ *
*****************************************************************************/
TransactionStmt:
diff --git a/src/backend/replication/logical/proto.c b/src/backend/replication/logical/proto.c
index 9f5bf4b639..9f53b4c47b 100644
--- a/src/backend/replication/logical/proto.c
+++ b/src/backend/replication/logical/proto.c
@@ -647,6 +647,58 @@ logicalrep_write_message(StringInfo out, TransactionId xid, XLogRecPtr lsn,
pq_sendbytes(out, message, sz);
}
+/*
+ * Write SEQUENCE to stream
+ */
+void
+logicalrep_write_sequence(StringInfo out, Relation rel, TransactionId xid,
+ XLogRecPtr lsn, bool transactional, bool created,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ uint8 flags = 0;
+ char *relname;
+
+ pq_sendbyte(out, LOGICAL_REP_MSG_SEQUENCE);
+
+ /* transaction ID (if not valid, we're not streaming) */
+ if (TransactionIdIsValid(xid))
+ pq_sendint32(out, xid);
+
+ pq_sendint8(out, flags);
+ pq_sendint64(out, lsn);
+
+ logicalrep_write_namespace(out, RelationGetNamespace(rel));
+ relname = RelationGetRelationName(rel);
+ pq_sendstring(out, relname);
+
+ pq_sendint8(out, transactional);
+ pq_sendint8(out, created);
+ pq_sendint64(out, last_value);
+ pq_sendint64(out, log_cnt);
+ pq_sendint8(out, is_called);
+}
+
+/*
+ * Read SEQUENCE from the stream.
+ */
+void
+logicalrep_read_sequence(StringInfo in, LogicalRepSequence *seqdata)
+{
+ /* XXX skipping flags and lsn */
+ pq_getmsgint(in, 1);
+ pq_getmsgint64(in);
+
+ /* Read relation name from stream */
+ seqdata->nspname = pstrdup(logicalrep_read_namespace(in));
+ seqdata->seqname = pstrdup(pq_getmsgstring(in));
+
+ seqdata->transactional = pq_getmsgint(in, 1);
+ seqdata->created = pq_getmsgint(in, 1);
+ seqdata->last_value = pq_getmsgint64(in);
+ seqdata->log_cnt = pq_getmsgint64(in);
+ seqdata->is_called = pq_getmsgint(in, 1);
+}
+
/*
* Write relation description to the output stream.
*/
@@ -1203,6 +1255,8 @@ logicalrep_message_type(LogicalRepMsgType action)
return "STREAM ABORT";
case LOGICAL_REP_MSG_STREAM_PREPARE:
return "STREAM PREPARE";
+ case LOGICAL_REP_MSG_SEQUENCE:
+ return "SEQUENCE";
}
elog(ERROR, "invalid logical replication message type \"%c\"", action);
diff --git a/src/backend/replication/logical/tablesync.c b/src/backend/replication/logical/tablesync.c
index f07983a43c..24369a1522 100644
--- a/src/backend/replication/logical/tablesync.c
+++ b/src/backend/replication/logical/tablesync.c
@@ -100,6 +100,7 @@
#include "catalog/pg_subscription_rel.h"
#include "catalog/pg_type.h"
#include "commands/copy.h"
+#include "commands/sequence.h"
#include "miscadmin.h"
#include "parser/parse_relation.h"
#include "pgstat.h"
@@ -357,6 +358,12 @@ process_syncing_tables_for_sync(XLogRecPtr current_lsn)
*
* If the synchronization position is reached (SYNCDONE), then the table can
* be marked as READY and is no longer tracked.
+ *
+ * XXX This needs to handle sequences too - after AlterSubscription_refresh
+ * starts caring about sequences, GetSubscriptionNotReadyRelations won't
+ * return just tables, and we'll have to sync them here. Not sure it's worth
+ * creating a new "sync" worker per sequence, maybe we should just sync them
+ * in the current process (it's pretty light-weight).
*/
static void
process_syncing_tables_for_apply(XLogRecPtr current_lsn)
@@ -871,6 +878,99 @@ copy_table(Relation rel)
logicalrep_rel_close(relmapentry, NoLock);
}
+
+
+/*
+ * FIXME add comment
+ */
+static void
+fetch_sequence_data(char *nspname, char *relname,
+ int64 *last_value, int64 *log_cnt, bool *is_called)
+{
+ WalRcvExecResult *res;
+ StringInfoData cmd;
+ TupleTableSlot *slot;
+ Oid tableRow[3] = {INT8OID, INT8OID, BOOLOID};
+
+ initStringInfo(&cmd);
+ appendStringInfo(&cmd, "SELECT last_value, log_cnt, is_called\n"
+ " FROM %s", quote_qualified_identifier(nspname, relname));
+
+ res = walrcv_exec(LogRepWorkerWalRcvConn, cmd.data, 3, tableRow);
+ pfree(cmd.data);
+
+ if (res->status != WALRCV_OK_TUPLES)
+ ereport(ERROR,
+ (errmsg("could not receive list of replicated tables from the publisher: %s",
+ res->err)));
+
+ /* Process the sequence. */
+ slot = MakeSingleTupleTableSlot(res->tupledesc, &TTSOpsMinimalTuple);
+ while (tuplestore_gettupleslot(res->tuplestore, true, false, slot))
+ {
+ bool isnull;
+
+ *last_value = DatumGetInt64(slot_getattr(slot, 1, &isnull));
+ Assert(!isnull);
+
+ *log_cnt = DatumGetInt64(slot_getattr(slot, 2, &isnull));
+ Assert(!isnull);
+
+ *is_called = DatumGetBool(slot_getattr(slot, 3, &isnull));
+ Assert(!isnull);
+
+ ExecClearTuple(slot);
+ }
+ ExecDropSingleTupleTableSlot(slot);
+
+ walrcv_clear_result(res);
+}
+
+/*
+ * Copy existing data of a sequence from publisher.
+ *
+ * Caller is responsible for locking the local relation.
+ */
+static void
+copy_sequence(Relation rel)
+{
+ LogicalRepRelMapEntry *relmapentry;
+ LogicalRepRelation lrel;
+ StringInfoData cmd;
+ int64 last_value = 0,
+ log_cnt = 0;
+ bool is_called = 0;
+
+ /* Get the publisher relation info. */
+ fetch_remote_table_info(get_namespace_name(RelationGetNamespace(rel)),
+ RelationGetRelationName(rel), &lrel);
+
+ /* Put the relation into relmap. */
+ logicalrep_relmap_update(&lrel);
+
+ /* Map the publisher relation to local one. */
+ relmapentry = logicalrep_rel_open(lrel.remoteid, NoLock);
+ Assert(rel == relmapentry->localrel);
+
+ /* Start copy on the publisher. */
+ initStringInfo(&cmd);
+
+ Assert(lrel.relkind == RELKIND_SEQUENCE);
+
+ fetch_sequence_data(lrel.nspname, lrel.relname, &last_value, &log_cnt, &is_called);
+
+ elog(WARNING, "sequence %s info last_value %ld log_cnt %ld is_called %d",
+ quote_qualified_identifier(lrel.nspname, lrel.relname),
+ last_value, log_cnt, is_called);
+
+ ResetSequence2(RelationGetRelid(rel), last_value, log_cnt, is_called);
+
+ logicalrep_rel_close(relmapentry, NoLock);
+}
+
+
+
+
/*
* Determine the tablesync slot name.
*
@@ -1106,10 +1206,20 @@ LogicalRepSyncTableStart(XLogRecPtr *origin_startpos)
originname)));
}
- /* Now do the initial data copy */
- PushActiveSnapshot(GetTransactionSnapshot());
- copy_table(rel);
- PopActiveSnapshot();
+ if (get_rel_relkind(RelationGetRelid(rel)) == RELKIND_SEQUENCE)
+ {
+ /* Now do the initial sequence copy */
+ PushActiveSnapshot(GetTransactionSnapshot());
+ copy_sequence(rel);
+ PopActiveSnapshot();
+ }
+ else
+ {
+ /* Now do the initial data copy */
+ PushActiveSnapshot(GetTransactionSnapshot());
+ copy_table(rel);
+ PopActiveSnapshot();
+ }
res = walrcv_exec(LogRepWorkerWalRcvConn, "COMMIT", 0, NULL);
if (res->status != WALRCV_OK_COMMAND)
diff --git a/src/backend/replication/logical/worker.c b/src/backend/replication/logical/worker.c
index ae1b391bda..55d48ce5be 100644
--- a/src/backend/replication/logical/worker.c
+++ b/src/backend/replication/logical/worker.c
@@ -144,6 +144,7 @@
#include "catalog/pg_tablespace.h"
#include "commands/tablecmds.h"
#include "commands/tablespace.h"
+#include "commands/sequence.h"
#include "commands/trigger.h"
#include "executor/executor.h"
#include "executor/execPartition.h"
@@ -1091,6 +1092,61 @@ apply_handle_origin(StringInfo s)
errmsg_internal("ORIGIN message sent out of order")));
}
+/*
+ * Handle SEQUENCE message.
+ */
+static void
+apply_handle_sequence(StringInfo s)
+{
+ LogicalRepSequence seq;
+ Oid relid;
+
+ if (handle_streamed_transaction(LOGICAL_REP_MSG_SEQUENCE, s))
+ return;
+
+ logicalrep_read_sequence(s, &seq);
+
+ /*
+ * Non-transactional sequence updates should not be part of a remote
+ * transaction. There should not be any running transaction.
+ */
+ Assert((!seq.transactional) || in_remote_transaction);
+ Assert(!(!seq.transactional && in_remote_transaction));
+ Assert(!(!seq.transactional && IsTransactionState()));
+
+ /*
+ * Make sure we're in a transaction (needed by ResetSequence2). For
+ * non-transactional updates we're guaranteed to start a new one,
+ * and we'll commit it at the end.
+ */
+ if (!IsTransactionState())
+ {
+ StartTransactionCommand();
+ maybe_reread_subscription();
+ }
+
+ relid = RangeVarGetRelid(makeRangeVar(seq.nspname,
+ seq.seqname, -1),
+ RowExclusiveLock, false);
+
+ /* lock the sequence in AccessExclusiveLock, as expected by ResetSequence2 */
+ elog(WARNING, "locking sequence %d in exclusive mode", relid);
+ LockRelationOid(relid, AccessExclusiveLock);
+
+ elog(WARNING, "applying sequence %s.%s transactional %d created %d last_value %ld log_cnt %ld is_called %d",
+ seq.nspname, seq.seqname, seq.transactional, seq.created, seq.last_value, seq.log_cnt, seq.is_called);
+
+ /* apply the sequence change */
+ ResetSequence2(relid, seq.last_value, seq.log_cnt, seq.is_called);
+
+ /*
+ * Commit the per-stream transaction (we only do this when not in
+ * remote transaction, i.e. for non-transactional sequence updates.
+ */
+ if (!in_remote_transaction)
+ CommitTransactionCommand();
+}
+
/*
* Handle STREAM START message.
*/
@@ -2374,6 +2430,10 @@ apply_dispatch(StringInfo s)
*/
break;
+ case LOGICAL_REP_MSG_SEQUENCE:
+ apply_handle_sequence(s);
+ return;
+
case LOGICAL_REP_MSG_STREAM_START:
apply_handle_stream_start(s);
break;
diff --git a/src/backend/replication/pgoutput/pgoutput.c b/src/backend/replication/pgoutput/pgoutput.c
index 6f6a203dea..fef568b5b2 100644
--- a/src/backend/replication/pgoutput/pgoutput.c
+++ b/src/backend/replication/pgoutput/pgoutput.c
@@ -49,6 +49,10 @@ static void pgoutput_message(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn, XLogRecPtr message_lsn,
bool transactional, const char *prefix,
Size sz, const char *message);
+static void pgoutput_sequence(LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn, XLogRecPtr sequence_lsn,
+ Relation rel, bool transactional, bool created,
+ int64 last_value, int64 log_cnt, bool is_called);
static bool pgoutput_origin_filter(LogicalDecodingContext *ctx,
RepOriginId origin_id);
static void pgoutput_begin_prepare_txn(LogicalDecodingContext *ctx,
@@ -159,6 +163,7 @@ _PG_output_plugin_init(OutputPluginCallbacks *cb)
cb->change_cb = pgoutput_change;
cb->truncate_cb = pgoutput_truncate;
cb->message_cb = pgoutput_message;
+ cb->sequence_cb = pgoutput_sequence;
cb->commit_cb = pgoutput_commit_txn;
cb->begin_prepare_cb = pgoutput_begin_prepare_txn;
@@ -175,6 +180,7 @@ _PG_output_plugin_init(OutputPluginCallbacks *cb)
cb->stream_commit_cb = pgoutput_stream_commit;
cb->stream_change_cb = pgoutput_change;
cb->stream_message_cb = pgoutput_message;
+ cb->stream_sequence_cb = pgoutput_sequence;
cb->stream_truncate_cb = pgoutput_truncate;
/* transaction streaming - two-phase commit */
cb->stream_prepare_cb = pgoutput_stream_prepare_txn;
@@ -188,6 +194,7 @@ parse_output_parameters(List *options, PGOutputData *data)
bool publication_names_given = false;
bool binary_option_given = false;
bool messages_option_given = false;
+ bool sequences_option_given = false;
bool streaming_given = false;
bool two_phase_option_given = false;
@@ -195,6 +202,7 @@ parse_output_parameters(List *options, PGOutputData *data)
data->streaming = false;
data->messages = false;
data->two_phase = false;
+ data->sequences = true;
foreach(lc, options)
{
@@ -260,6 +268,16 @@ parse_output_parameters(List *options, PGOutputData *data)
data->messages = defGetBoolean(defel);
}
+ else if (strcmp(defel->defname, "sequences") == 0)
+ {
+ if (sequences_option_given)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("conflicting or redundant options")));
+ sequences_option_given = true;
+
+ data->sequences = defGetBoolean(defel);
+ }
else if (strcmp(defel->defname, "streaming") == 0)
{
if (streaming_given)
@@ -856,6 +874,52 @@ pgoutput_message(LogicalDecodingContext *ctx, ReorderBufferTXN *txn,
OutputPluginWrite(ctx, true);
}
+static void
+pgoutput_sequence(LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn, XLogRecPtr sequence_lsn,
+ Relation rel, bool transactional, bool created,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ PGOutputData *data = (PGOutputData *) ctx->output_plugin_private;
+ TransactionId xid = InvalidTransactionId;
+ RelationSyncEntry *relentry;
+
+ if (!data->sequences)
+ return;
+
+ if (!is_publishable_relation(rel))
+ return;
+
+ /*
+ * Remember the xid for the message in streaming mode. See
+ * pgoutput_change.
+ */
+ if (in_streaming)
+ xid = txn->xid;
+
+ relentry = get_rel_sync_entry(data, RelationGetRelid(rel));
+
+ /*
+ * First check the sequence filter.
+ *
+ * We handle just REORDER_BUFFER_CHANGE_SEQUENCE here.
+ */
+ if (!relentry->pubactions.pubsequence)
+ return;
+
+ OutputPluginPrepareWrite(ctx, true);
+ logicalrep_write_sequence(ctx->out,
+ rel,
+ xid,
+ sequence_lsn,
+ transactional,
+ created,
+ last_value,
+ log_cnt,
+ is_called);
+ OutputPluginWrite(ctx, true);
+}
+
/*
* Currently we always forward.
*/
@@ -1140,7 +1204,8 @@ get_rel_sync_entry(PGOutputData *data, Oid relid)
entry->streamed_txns = NIL;
entry->replicate_valid = false;
entry->pubactions.pubinsert = entry->pubactions.pubupdate =
- entry->pubactions.pubdelete = entry->pubactions.pubtruncate = false;
+ entry->pubactions.pubdelete = entry->pubactions.pubtruncate =
+ entry->pubactions.pubsequence = false;
entry->publish_as_relid = InvalidOid;
entry->map = NULL; /* will be set by maybe_send_schema() if
* needed */
@@ -1160,6 +1225,7 @@ get_rel_sync_entry(PGOutputData *data, Oid relid)
List *schemaPubids = GetSchemaPublications(schemaId);
ListCell *lc;
Oid publish_as_relid = relid;
+ bool is_sequence = (get_rel_relkind(relid) == RELKIND_SEQUENCE);
/* Reload publications if needed before use. */
if (!publications_valid)
@@ -1183,12 +1249,23 @@ get_rel_sync_entry(PGOutputData *data, Oid relid)
Publication *pub = lfirst(lc);
bool publish = false;
- if (pub->alltables)
+ if (pub->alltables && (!is_sequence))
{
publish = true;
if (pub->pubviaroot && am_partition)
publish_as_relid = llast_oid(get_partition_ancestors(relid));
}
+ else if (pub->allsequences && is_sequence)
+ {
+ publish = true;
+ }
+
+ /* if a sequence, just cross-check the list of publications */
+ if (!publish && is_sequence)
+ {
+ if (list_member_oid(pubids, pub->oid))
+ publish = true;
+ }
if (!publish)
{
@@ -1243,10 +1320,12 @@ get_rel_sync_entry(PGOutputData *data, Oid relid)
entry->pubactions.pubupdate |= pub->pubactions.pubupdate;
entry->pubactions.pubdelete |= pub->pubactions.pubdelete;
entry->pubactions.pubtruncate |= pub->pubactions.pubtruncate;
+ entry->pubactions.pubsequence |= pub->pubactions.pubsequence;
}
if (entry->pubactions.pubinsert && entry->pubactions.pubupdate &&
- entry->pubactions.pubdelete && entry->pubactions.pubtruncate)
+ entry->pubactions.pubdelete && entry->pubactions.pubtruncate &&
+ entry->pubactions.pubsequence)
break;
}
@@ -1391,6 +1470,7 @@ rel_sync_cache_publication_cb(Datum arg, int cacheid, uint32 hashvalue)
entry->pubactions.pubupdate = false;
entry->pubactions.pubdelete = false;
entry->pubactions.pubtruncate = false;
+ entry->pubactions.pubsequence = false;
}
}
diff --git a/src/backend/utils/cache/relcache.c b/src/backend/utils/cache/relcache.c
index 9fa9e671a1..1c3286a38c 100644
--- a/src/backend/utils/cache/relcache.c
+++ b/src/backend/utils/cache/relcache.c
@@ -5605,6 +5605,7 @@ GetRelationPublicationActions(Relation relation)
pubactions->pubupdate |= pubform->pubupdate;
pubactions->pubdelete |= pubform->pubdelete;
pubactions->pubtruncate |= pubform->pubtruncate;
+ pubactions->pubsequence |= pubform->pubsequence;
ReleaseSysCache(tup);
@@ -5613,7 +5614,8 @@ GetRelationPublicationActions(Relation relation)
* other publications.
*/
if (pubactions->pubinsert && pubactions->pubupdate &&
- pubactions->pubdelete && pubactions->pubtruncate)
+ pubactions->pubdelete && pubactions->pubtruncate &&
+ pubactions->pubsequence)
break;
}
diff --git a/src/bin/psql/tab-complete.c b/src/bin/psql/tab-complete.c
index 4f724e4428..59a3502ae2 100644
--- a/src/bin/psql/tab-complete.c
+++ b/src/bin/psql/tab-complete.c
@@ -1647,13 +1647,13 @@ psql_completion(const char *text, int start, int end)
COMPLETE_WITH("ADD", "DROP", "OWNER TO", "RENAME TO", "SET");
/* ALTER PUBLICATION <name> ADD */
else if (Matches("ALTER", "PUBLICATION", MatchAny, "ADD"))
- COMPLETE_WITH("ALL TABLES IN SCHEMA", "TABLE");
+ COMPLETE_WITH("ALL TABLES IN SCHEMA", "TABLE|SEQUENCE");
/* ALTER PUBLICATION <name> DROP */
else if (Matches("ALTER", "PUBLICATION", MatchAny, "DROP"))
- COMPLETE_WITH("ALL TABLES IN SCHEMA", "TABLE");
+ COMPLETE_WITH("ALL TABLES IN SCHEMA", "TABLE|SEQUENCE");
/* ALTER PUBLICATION <name> SET */
else if (Matches("ALTER", "PUBLICATION", MatchAny, "SET"))
- COMPLETE_WITH("(", "ALL TABLES IN SCHEMA", "TABLE");
+ COMPLETE_WITH("(", "ALL TABLES IN SCHEMA", "TABLE|SEQUENCE");
else if (Matches("ALTER", "PUBLICATION", MatchAny, "ADD|DROP|SET", "ALL", "TABLES", "IN", "SCHEMA"))
COMPLETE_WITH_QUERY(Query_for_list_of_schemas
" AND nspname != 'pg_catalog' "
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 6412f369f1..be5e3b4d40 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -11513,6 +11513,11 @@
provolatile => 's', prorettype => 'oid', proargtypes => 'text',
proallargtypes => '{text,oid}', proargmodes => '{i,o}',
proargnames => '{pubname,relid}', prosrc => 'pg_get_publication_tables' },
+{ oid => '8000', descr => 'get OIDs of sequences in a publication',
+ proname => 'pg_get_publication_sequences', prorows => '1000', proretset => 't',
+ provolatile => 's', prorettype => 'oid', proargtypes => 'text',
+ proallargtypes => '{text,oid}', proargmodes => '{i,o}',
+ proargnames => '{pubname,relid}', prosrc => 'pg_get_publication_sequences' },
{ oid => '6121',
descr => 'returns whether a relation can be part of a publication',
proname => 'pg_relation_is_publishable', provolatile => 's',
diff --git a/src/include/catalog/pg_publication.h b/src/include/catalog/pg_publication.h
index 1ae439e6f3..0a05fb885c 100644
--- a/src/include/catalog/pg_publication.h
+++ b/src/include/catalog/pg_publication.h
@@ -40,6 +40,12 @@ CATALOG(pg_publication,6104,PublicationRelationId)
*/
bool puballtables;
+ /*
+ * indicates that this is special publication which should encompass all
+ * sequences in the database (except for the unlogged and temp ones)
+ */
+ bool puballsequences;
+
/* true if inserts are published */
bool pubinsert;
@@ -52,6 +58,9 @@ CATALOG(pg_publication,6104,PublicationRelationId)
/* true if truncates are published */
bool pubtruncate;
+ /* true if sequences are published */
+ bool pubsequence;
+
/* true if partition changes are published using root schema */
bool pubviaroot;
} FormData_pg_publication;
@@ -72,6 +81,7 @@ typedef struct PublicationActions
bool pubupdate;
bool pubdelete;
bool pubtruncate;
+ bool pubsequence;
} PublicationActions;
typedef struct Publication
@@ -79,6 +89,7 @@ typedef struct Publication
Oid oid;
char *name;
bool alltables;
+ bool allsequences;
bool pubviaroot;
PublicationActions pubactions;
} Publication;
@@ -121,6 +132,9 @@ extern List *GetPubPartitionOptionRelations(List *result,
PublicationPartOpt pub_partopt,
Oid relid);
+extern List *GetAllSequencesPublicationRelations(void);
+extern List *GetPublicationSequenceRelations(Oid pubid);
+
extern bool is_publishable_relation(Relation rel);
extern ObjectAddress publication_add_relation(Oid pubid, PublicationRelInfo *targetrel,
bool if_not_exists);
diff --git a/src/include/commands/sequence.h b/src/include/commands/sequence.h
index ddb36b5b17..00038fcea3 100644
--- a/src/include/commands/sequence.h
+++ b/src/include/commands/sequence.h
@@ -60,6 +60,7 @@ extern ObjectAddress DefineSequence(ParseState *pstate, CreateSeqStmt *stmt);
extern ObjectAddress AlterSequence(ParseState *pstate, AlterSeqStmt *stmt);
extern void DeleteSequenceTuple(Oid relid);
extern void ResetSequence(Oid seq_relid);
+extern void ResetSequence2(Oid seq_relid, int64 last_value, int64 log_cnt, bool is_called);
extern void ResetSequenceCaches(void);
extern void AtEOXact_Sequences(bool isCommit);
diff --git a/src/include/nodes/parsenodes.h b/src/include/nodes/parsenodes.h
index 067138e6b5..44e596fd12 100644
--- a/src/include/nodes/parsenodes.h
+++ b/src/include/nodes/parsenodes.h
@@ -3652,6 +3652,10 @@ typedef enum PublicationObjSpecType
PUBLICATIONOBJ_TABLE_IN_SCHEMA, /* Tables in schema type */
PUBLICATIONOBJ_TABLE_IN_CUR_SCHEMA, /* Get the first element from
* search_path */
+ PUBLICATIONOBJ_SEQUENCE, /* Sequence type */
+ PUBLICATIONOBJ_SEQUENCE_IN_SCHEMA, /* Sequences in schema type */
+ PUBLICATIONOBJ_SEQUENCE_IN_CUR_SCHEMA, /* Get the first element from
+ * search_path */
PUBLICATIONOBJ_CONTINUATION /* Continuation of previous type */
} PublicationObjSpecType;
@@ -3671,6 +3675,7 @@ typedef struct CreatePublicationStmt
List *options; /* List of DefElem nodes */
List *pubobjects; /* Optional list of publication objects */
bool for_all_tables; /* Special publication for all tables in db */
+ bool for_all_sequences; /* Special publication for all sequences in db */
} CreatePublicationStmt;
typedef struct AlterPublicationStmt
@@ -3687,6 +3692,7 @@ typedef struct AlterPublicationStmt
*/
List *pubobjects; /* Optional list of publication objects */
bool for_all_tables; /* Special publication for all tables in db */
+ bool for_all_sequences; /* Special publication for all sequences in db */
DefElemAction action; /* What action to perform with the
* tables/schemas */
} AlterPublicationStmt;
diff --git a/src/include/replication/logicalproto.h b/src/include/replication/logicalproto.h
index 83741dcf42..d468c3ccfb 100644
--- a/src/include/replication/logicalproto.h
+++ b/src/include/replication/logicalproto.h
@@ -60,6 +60,7 @@ typedef enum LogicalRepMsgType
LOGICAL_REP_MSG_RELATION = 'R',
LOGICAL_REP_MSG_TYPE = 'Y',
LOGICAL_REP_MSG_MESSAGE = 'M',
+ LOGICAL_REP_MSG_SEQUENCE = 'X', /* FIXME change */
LOGICAL_REP_MSG_BEGIN_PREPARE = 'b',
LOGICAL_REP_MSG_PREPARE = 'P',
LOGICAL_REP_MSG_COMMIT_PREPARED = 'K',
@@ -117,6 +118,19 @@ typedef struct LogicalRepTyp
char *typname; /* name of the remote type */
} LogicalRepTyp;
+/* Sequence info */
+typedef struct LogicalRepSequence
+{
+ Oid remoteid; /* unique id of the remote sequence */
+ char *nspname; /* schema name of remote sequence */
+ char *seqname; /* name of the remote sequence */
+ bool transactional;
+ bool created;
+ int64 last_value;
+ int64 log_cnt; /* XXX probably not needed? */
+ bool is_called; /* XXX probably not needed? */
+} LogicalRepSequence;
+
/* Transaction info */
typedef struct LogicalRepBeginData
{
@@ -227,6 +241,12 @@ extern List *logicalrep_read_truncate(StringInfo in,
bool *cascade, bool *restart_seqs);
extern void logicalrep_write_message(StringInfo out, TransactionId xid, XLogRecPtr lsn,
bool transactional, const char *prefix, Size sz, const char *message);
+extern void logicalrep_write_sequence(StringInfo out, Relation rel,
+ TransactionId xid, XLogRecPtr lsn,
+ bool transactional, bool created,
+ int64 last_value, int64 log_cnt,
+ bool is_called);
+extern void logicalrep_read_sequence(StringInfo in, LogicalRepSequence *seqdata);
extern void logicalrep_write_rel(StringInfo out, TransactionId xid,
Relation rel);
extern LogicalRepRelation *logicalrep_read_rel(StringInfo in);
diff --git a/src/include/replication/pgoutput.h b/src/include/replication/pgoutput.h
index 0dc460fb70..97eb2a7ef1 100644
--- a/src/include/replication/pgoutput.h
+++ b/src/include/replication/pgoutput.h
@@ -28,6 +28,7 @@ typedef struct PGOutputData
bool streaming;
bool messages;
bool two_phase;
+ bool sequences;
} PGOutputData;
#endif /* PGOUTPUT_H */
diff --git a/src/test/regress/expected/rules.out b/src/test/regress/expected/rules.out
index 2fa00a3c29..b32a1e6729 100644
--- a/src/test/regress/expected/rules.out
+++ b/src/test/regress/expected/rules.out
@@ -1451,6 +1451,14 @@ pg_prepared_xacts| SELECT p.transaction,
FROM ((pg_prepared_xact() p(transaction, gid, prepared, ownerid, dbid)
LEFT JOIN pg_authid u ON ((p.ownerid = u.oid)))
LEFT JOIN pg_database d ON ((p.dbid = d.oid)));
+pg_publication_sequences| SELECT p.pubname,
+ n.nspname AS schemaname,
+ c.relname AS sequencename
+ FROM pg_publication p,
+ LATERAL pg_get_publication_sequences((p.pubname)::text) gpt(relid),
+ (pg_class c
+ JOIN pg_namespace n ON ((n.oid = c.relnamespace)))
+ WHERE (c.oid = gpt.relid);
pg_publication_tables| SELECT p.pubname,
n.nspname AS schemaname,
c.relname AS tablename
diff --git a/src/test/subscription/t/023_sequences.pl b/src/test/subscription/t/023_sequences.pl
new file mode 100644
index 0000000000..c6c438cdbf
--- /dev/null
+++ b/src/test/subscription/t/023_sequences.pl
@@ -0,0 +1,196 @@
+
+# Copyright (c) 2021, PostgreSQL Global Development Group
+
+# This tests that sequences are replicated correctly by logical replication
+use strict;
+use warnings;
+use PostgreSQL::Test::Cluster;
+use PostgreSQL::Test::Utils;
+use Test::More tests => 6;
+
+# Initialize publisher node
+my $node_publisher = PostgreSQL::Test::Cluster->new('publisher');
+$node_publisher->init(allows_streaming => 'logical');
+$node_publisher->start;
+
+# Create subscriber node
+my $node_subscriber = PostgreSQL::Test::Cluster->new('subscriber');
+$node_subscriber->init(allows_streaming => 'logical');
+$node_subscriber->start;
+
+# Create some preexisting content on publisher
+my $ddl = qq(
+ CREATE SEQUENCE s;
+);
+
+# Setup structure on the publisher
+$node_publisher->safe_psql('postgres', $ddl);
+
+# Create some the same structure on subscriber, and an extra sequence that
+# we'll create on the publisher later
+$ddl = qq(
+ CREATE SEQUENCE s;
+ CREATE SEQUENCE s2;
+);
+
+$node_subscriber->safe_psql('postgres', $ddl);
+
+# Setup logical replication
+my $publisher_connstr = $node_publisher->connstr . ' dbname=postgres';
+$node_publisher->safe_psql('postgres',
+ "CREATE PUBLICATION seq_pub");
+
+$node_publisher->safe_psql('postgres',
+ "ALTER PUBLICATION seq_pub ADD SEQUENCE s");
+
+$node_subscriber->safe_psql('postgres',
+ "CREATE SUBSCRIPTION seq_sub CONNECTION '$publisher_connstr' PUBLICATION seq_pub WITH (slot_name = seq_sub_slot)"
+);
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Wait for initial sync to finish as well
+my $synced_query =
+ "SELECT count(1) = 0 FROM pg_subscription_rel WHERE srsubstate NOT IN ('s', 'r');";
+$node_subscriber->poll_query_until('postgres', $synced_query)
+ or die "Timed out while waiting for subscriber to synchronize data";
+
+# Insert initial test data
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ -- generate a number of values using the sequence
+ SELECT nextval('s') FROM generate_series(1,100);
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+my $result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s;
+));
+
+is( $result, '132|0|t',
+ 'check replicated sequence values on subscriber');
+
+
+# advance the sequence in a rolled-back transaction - should be replicated
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ SELECT nextval('s') FROM generate_series(1,100);
+ ROLLBACK;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s;
+));
+
+is( $result, '231|0|t',
+ 'check replicated sequence values on subscriber');
+
+
+# create a new sequence and roll it back - should not be replicated, due to
+# the transactional behavior
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ CREATE SEQUENCE s2;
+ ALTER PUBLICATION seq_pub ADD SEQUENCE s2;
+ SELECT nextval('s2') FROM generate_series(1,100);
+ ROLLBACK;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s2;
+));
+
+is( $result, '1|0|f',
+ 'check replicated sequence values on subscriber');
+
+
+# create a new sequence, advance it in a rolled-back transaction, but commit
+# the create - the advance should be replicated nevertheless
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ CREATE SEQUENCE s2;
+ ALTER PUBLICATION seq_pub ADD SEQUENCE s2;
+ SAVEPOINT sp1;
+ SELECT nextval('s2') FROM generate_series(1,100);
+ ROLLBACK TO sp1;
+ COMMIT;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Wait for sync of the second sequence we just added to finish
+$synced_query =
+ "SELECT count(1) = 0 FROM pg_subscription_rel WHERE srsubstate NOT IN ('s', 'r');";
+$node_subscriber->poll_query_until('postgres', $synced_query)
+ or die "Timed out while waiting for subscriber to synchronize data";
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s2;
+));
+
+is( $result, '132|0|t',
+ 'check replicated sequence values on subscriber');
+
+
+# advance the new sequence in a transaction, and roll it back - in this case
+# it should be replicated as the behavior is non-transactional
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ SELECT nextval('s2') FROM generate_series(1,100);
+ ROLLBACK;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s2;
+));
+
+is( $result, '231|0|t',
+ 'check replicated sequence values on subscriber');
+
+
+# advance the sequence in a subtransaction - the subtransaction gets rolled
+# back, but commit the main one - the changes should still be replicated
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ SAVEPOINT s1;
+ SELECT nextval('s2') FROM generate_series(1,100);
+ ROLLBACK TO s1;
+ COMMIT;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s2;
+));
+
+is( $result, '330|0|t',
+ 'check replicated sequence values on subscriber');
+
+
+$node_subscriber->stop('fast');
+$node_publisher->stop('fast');
--
2.31.1
[text/x-patch] 0003-Add-support-for-decoding-sequences-to-test_-20211118.patch (20.7K, ../../[email protected]/3-0003-Add-support-for-decoding-sequences-to-test_-20211118.patch)
download | inline diff:
From 0666c3a34955fc2edb811b36c081d271855d569e Mon Sep 17 00:00:00 2001
From: Tomas Vondra <[email protected]>
Date: Fri, 24 Sep 2021 00:42:04 +0200
Subject: [PATCH 3/4] Add support for decoding sequences to test_decoding
---
contrib/test_decoding/Makefile | 3 +-
contrib/test_decoding/expected/sequence.out | 327 ++++++++++++++++++++
contrib/test_decoding/sql/sequence.sql | 119 +++++++
contrib/test_decoding/test_decoding.c | 69 +++++
4 files changed, 517 insertions(+), 1 deletion(-)
create mode 100644 contrib/test_decoding/expected/sequence.out
create mode 100644 contrib/test_decoding/sql/sequence.sql
diff --git a/contrib/test_decoding/Makefile b/contrib/test_decoding/Makefile
index 9a31e0b879..56ddc3abae 100644
--- a/contrib/test_decoding/Makefile
+++ b/contrib/test_decoding/Makefile
@@ -5,7 +5,8 @@ PGFILEDESC = "test_decoding - example of a logical decoding output plugin"
REGRESS = ddl xact rewrite toast permissions decoding_in_xact \
decoding_into_rel binary prepared replorigin time messages \
- spill slot truncate stream stats twophase twophase_stream
+ spill slot truncate stream stats twophase twophase_stream \
+ sequence
ISOLATION = mxact delayed_startup ondisk_startup concurrent_ddl_dml \
oldest_xmin snapshot_transfer subxact_without_top concurrent_stream \
twophase_snapshot
diff --git a/contrib/test_decoding/expected/sequence.out b/contrib/test_decoding/expected/sequence.out
new file mode 100644
index 0000000000..beecc3e0c8
--- /dev/null
+++ b/contrib/test_decoding/expected/sequence.out
@@ -0,0 +1,327 @@
+-- predictability
+SET synchronous_commit = on;
+SELECT 'init' FROM pg_create_logical_replication_slot('regression_slot', 'test_decoding');
+ ?column?
+----------
+ init
+(1 row)
+
+CREATE SEQUENCE test_sequence;
+-- test the sequence changes by several nextval() calls
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 1
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 2
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 4
+(1 row)
+
+-- test the sequence changes by several ALTER commands
+ALTER SEQUENCE test_sequence INCREMENT BY 10;
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 14
+(1 row)
+
+ALTER SEQUENCE test_sequence START WITH 3000;
+ALTER SEQUENCE test_sequence MAXVALUE 10000;
+ALTER SEQUENCE test_sequence RESTART WITH 4000;
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 4000
+(1 row)
+
+-- test the sequence changes by several setval() calls
+SELECT setval('test_sequence', 3500);
+ setval
+--------
+ 3500
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3510
+(1 row)
+
+SELECT setval('test_sequence', 3500, true);
+ setval
+--------
+ 3500
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3510
+(1 row)
+
+SELECT setval('test_sequence', 3500, false);
+ setval
+--------
+ 3500
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3500
+(1 row)
+
+-- show results and drop sequence
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+ data
+------------------------------------------------------------------------------------------------
+ BEGIN
+ sequence public.test_sequence: transactional:1 created:1 last_value:1 log_cnt:0 is_called:0
+ COMMIT
+ sequence public.test_sequence: transactional:0 created:0 last_value:33 log_cnt:0 is_called:1
+ BEGIN
+ sequence public.test_sequence: transactional:1 created:1 last_value:4 log_cnt:0 is_called:1
+ COMMIT
+ sequence public.test_sequence: transactional:0 created:0 last_value:334 log_cnt:0 is_called:1
+ BEGIN
+ sequence public.test_sequence: transactional:1 created:1 last_value:14 log_cnt:32 is_called:1
+ COMMIT
+ BEGIN
+ sequence public.test_sequence: transactional:1 created:1 last_value:14 log_cnt:0 is_called:1
+ COMMIT
+ BEGIN
+ sequence public.test_sequence: transactional:1 created:1 last_value:4000 log_cnt:0 is_called:0
+ COMMIT
+ sequence public.test_sequence: transactional:0 created:0 last_value:4320 log_cnt:0 is_called:1
+ sequence public.test_sequence: transactional:0 created:0 last_value:3500 log_cnt:0 is_called:1
+ sequence public.test_sequence: transactional:0 created:0 last_value:3830 log_cnt:0 is_called:1
+ sequence public.test_sequence: transactional:0 created:0 last_value:3500 log_cnt:0 is_called:1
+ sequence public.test_sequence: transactional:0 created:0 last_value:3830 log_cnt:0 is_called:1
+ sequence public.test_sequence: transactional:0 created:0 last_value:3500 log_cnt:0 is_called:0
+ sequence public.test_sequence: transactional:0 created:0 last_value:3820 log_cnt:0 is_called:1
+(24 rows)
+
+DROP SEQUENCE test_sequence;
+-- rollback on sequence creation and update
+BEGIN;
+CREATE SEQUENCE test_sequence;
+CREATE TABLE test_table (a INT);
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 1
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 2
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3
+(1 row)
+
+SELECT setval('test_sequence', 3000);
+ setval
+--------
+ 3000
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3001
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3002
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3003
+(1 row)
+
+ALTER SEQUENCE test_sequence RESTART WITH 6000;
+INSERT INTO test_table VALUES( (SELECT nextval('test_sequence')) );
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 6001
+(1 row)
+
+ROLLBACK;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'skip-sequences', '0');
+ data
+------
+(0 rows)
+
+-- rollback on table creation with serial column
+BEGIN;
+CREATE TABLE test_table (a SERIAL, b INT);
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+INSERT INTO test_table (b) VALUES (300);
+SELECT setval('test_table_a_seq', 3000);
+ setval
+--------
+ 3000
+(1 row)
+
+INSERT INTO test_table (b) VALUES (400);
+INSERT INTO test_table (b) VALUES (500);
+INSERT INTO test_table (b) VALUES (600);
+ALTER SEQUENCE test_table_a_seq RESTART WITH 6000;
+INSERT INTO test_table (b) VALUES (700);
+INSERT INTO test_table (b) VALUES (800);
+INSERT INTO test_table (b) VALUES (900);
+ROLLBACK;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'skip-sequences', '0');
+ data
+------
+(0 rows)
+
+-- rollback on table with serial column
+CREATE TABLE test_table (a SERIAL, b INT);
+BEGIN;
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+INSERT INTO test_table (b) VALUES (300);
+SELECT setval('test_table_a_seq', 3000);
+ setval
+--------
+ 3000
+(1 row)
+
+INSERT INTO test_table (b) VALUES (400);
+INSERT INTO test_table (b) VALUES (500);
+INSERT INTO test_table (b) VALUES (600);
+ALTER SEQUENCE test_table_a_seq RESTART WITH 6000;
+INSERT INTO test_table (b) VALUES (700);
+INSERT INTO test_table (b) VALUES (800);
+INSERT INTO test_table (b) VALUES (900);
+ROLLBACK;
+-- check table and sequence values after rollback
+SELECT * from test_table_a_seq;
+ last_value | log_cnt | is_called
+------------+---------+-----------
+ 3003 | 30 | t
+(1 row)
+
+SELECT nextval('test_table_a_seq');
+ nextval
+---------
+ 3004
+(1 row)
+
+DROP TABLE test_table;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+ data
+---------------------------------------------------------------------------------------------------
+ BEGIN
+ sequence public.test_table_a_seq: transactional:1 created:1 last_value:1 log_cnt:0 is_called:0
+ COMMIT
+ sequence public.test_table_a_seq: transactional:0 created:0 last_value:33 log_cnt:0 is_called:1
+ sequence public.test_table_a_seq: transactional:0 created:0 last_value:3000 log_cnt:0 is_called:1
+ sequence public.test_table_a_seq: transactional:0 created:0 last_value:3033 log_cnt:0 is_called:1
+ BEGIN
+ COMMIT
+(8 rows)
+
+-- savepoint test on table with serial column
+BEGIN;
+CREATE TABLE test_table (a SERIAL, b INT);
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+SAVEPOINT a;
+INSERT INTO test_table (b) VALUES (300);
+ROLLBACK TO SAVEPOINT a;
+DROP TABLE test_table;
+COMMIT;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+ data
+-------------------------------------------------------------------------------------------------
+ BEGIN
+ sequence public.test_table_a_seq: transactional:1 created:1 last_value:1 log_cnt:0 is_called:0
+ sequence public.test_table_a_seq: transactional:1 created:0 last_value:33 log_cnt:0 is_called:1
+ table public.test_table: INSERT: a[integer]:1 b[integer]:100
+ table public.test_table: INSERT: a[integer]:2 b[integer]:200
+ COMMIT
+(6 rows)
+
+-- savepoint test on table with serial column
+BEGIN;
+CREATE SEQUENCE test_sequence;
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 1
+(1 row)
+
+SELECT setval('test_sequence', 3000);
+ setval
+--------
+ 3000
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3001
+(1 row)
+
+SAVEPOINT a;
+ALTER SEQUENCE test_sequence START WITH 7000;
+SELECT setval('test_sequence', 5000);
+ setval
+--------
+ 5000
+(1 row)
+
+ROLLBACK TO SAVEPOINT a;
+SELECT * FROM test_sequence;
+ last_value | log_cnt | is_called
+------------+---------+-----------
+ 3001 | 32 | t
+(1 row)
+
+DROP SEQUENCE test_sequence;
+COMMIT;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+ data
+------------------------------------------------------------------------------------------------
+ BEGIN
+ sequence public.test_sequence: transactional:1 created:1 last_value:1 log_cnt:0 is_called:0
+ sequence public.test_sequence: transactional:1 created:0 last_value:33 log_cnt:0 is_called:1
+ sequence public.test_sequence: transactional:1 created:0 last_value:3000 log_cnt:0 is_called:1
+ sequence public.test_sequence: transactional:1 created:0 last_value:3033 log_cnt:0 is_called:1
+ COMMIT
+(6 rows)
+
+SELECT pg_drop_replication_slot('regression_slot');
+ pg_drop_replication_slot
+--------------------------
+
+(1 row)
+
diff --git a/contrib/test_decoding/sql/sequence.sql b/contrib/test_decoding/sql/sequence.sql
new file mode 100644
index 0000000000..d8a34738f3
--- /dev/null
+++ b/contrib/test_decoding/sql/sequence.sql
@@ -0,0 +1,119 @@
+-- predictability
+SET synchronous_commit = on;
+SELECT 'init' FROM pg_create_logical_replication_slot('regression_slot', 'test_decoding');
+
+CREATE SEQUENCE test_sequence;
+
+-- test the sequence changes by several nextval() calls
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+
+-- test the sequence changes by several ALTER commands
+ALTER SEQUENCE test_sequence INCREMENT BY 10;
+SELECT nextval('test_sequence');
+
+ALTER SEQUENCE test_sequence START WITH 3000;
+ALTER SEQUENCE test_sequence MAXVALUE 10000;
+ALTER SEQUENCE test_sequence RESTART WITH 4000;
+SELECT nextval('test_sequence');
+
+-- test the sequence changes by several setval() calls
+SELECT setval('test_sequence', 3500);
+SELECT nextval('test_sequence');
+SELECT setval('test_sequence', 3500, true);
+SELECT nextval('test_sequence');
+SELECT setval('test_sequence', 3500, false);
+SELECT nextval('test_sequence');
+
+-- show results and drop sequence
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+DROP SEQUENCE test_sequence;
+
+-- rollback on sequence creation and update
+BEGIN;
+CREATE SEQUENCE test_sequence;
+CREATE TABLE test_table (a INT);
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT setval('test_sequence', 3000);
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+ALTER SEQUENCE test_sequence RESTART WITH 6000;
+INSERT INTO test_table VALUES( (SELECT nextval('test_sequence')) );
+SELECT nextval('test_sequence');
+ROLLBACK;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'skip-sequences', '0');
+
+-- rollback on table creation with serial column
+BEGIN;
+CREATE TABLE test_table (a SERIAL, b INT);
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+INSERT INTO test_table (b) VALUES (300);
+SELECT setval('test_table_a_seq', 3000);
+INSERT INTO test_table (b) VALUES (400);
+INSERT INTO test_table (b) VALUES (500);
+INSERT INTO test_table (b) VALUES (600);
+ALTER SEQUENCE test_table_a_seq RESTART WITH 6000;
+INSERT INTO test_table (b) VALUES (700);
+INSERT INTO test_table (b) VALUES (800);
+INSERT INTO test_table (b) VALUES (900);
+ROLLBACK;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'skip-sequences', '0');
+
+-- rollback on table with serial column
+CREATE TABLE test_table (a SERIAL, b INT);
+
+BEGIN;
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+INSERT INTO test_table (b) VALUES (300);
+SELECT setval('test_table_a_seq', 3000);
+INSERT INTO test_table (b) VALUES (400);
+INSERT INTO test_table (b) VALUES (500);
+INSERT INTO test_table (b) VALUES (600);
+ALTER SEQUENCE test_table_a_seq RESTART WITH 6000;
+INSERT INTO test_table (b) VALUES (700);
+INSERT INTO test_table (b) VALUES (800);
+INSERT INTO test_table (b) VALUES (900);
+ROLLBACK;
+
+-- check table and sequence values after rollback
+SELECT * from test_table_a_seq;
+SELECT nextval('test_table_a_seq');
+
+DROP TABLE test_table;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+
+-- savepoint test on table with serial column
+BEGIN;
+CREATE TABLE test_table (a SERIAL, b INT);
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+SAVEPOINT a;
+INSERT INTO test_table (b) VALUES (300);
+ROLLBACK TO SAVEPOINT a;
+DROP TABLE test_table;
+COMMIT;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+
+-- savepoint test on table with serial column
+BEGIN;
+CREATE SEQUENCE test_sequence;
+SELECT nextval('test_sequence');
+SELECT setval('test_sequence', 3000);
+SELECT nextval('test_sequence');
+SAVEPOINT a;
+ALTER SEQUENCE test_sequence START WITH 7000;
+SELECT setval('test_sequence', 5000);
+ROLLBACK TO SAVEPOINT a;
+SELECT * FROM test_sequence;
+DROP SEQUENCE test_sequence;
+COMMIT;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+
+SELECT pg_drop_replication_slot('regression_slot');
diff --git a/contrib/test_decoding/test_decoding.c b/contrib/test_decoding/test_decoding.c
index e5cd84e85e..45765d299e 100644
--- a/contrib/test_decoding/test_decoding.c
+++ b/contrib/test_decoding/test_decoding.c
@@ -35,6 +35,7 @@ typedef struct
bool include_timestamp;
bool skip_empty_xacts;
bool only_local;
+ bool skip_sequences;
} TestDecodingData;
/*
@@ -76,6 +77,10 @@ static void pg_decode_message(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn, XLogRecPtr message_lsn,
bool transactional, const char *prefix,
Size sz, const char *message);
+static void pg_decode_sequence(LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn, XLogRecPtr sequence_lsn,
+ Relation rel, bool transactional, bool created,
+ int64 last_value, int64 log_cnt, bool is_called);
static bool pg_decode_filter_prepare(LogicalDecodingContext *ctx,
TransactionId xid,
const char *gid);
@@ -116,6 +121,10 @@ static void pg_decode_stream_message(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn, XLogRecPtr message_lsn,
bool transactional, const char *prefix,
Size sz, const char *message);
+static void pg_decode_stream_sequence(LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn, XLogRecPtr sequence_lsn,
+ Relation rel, bool transactional, bool created,
+ int64 last_value, int64 log_cnt, bool is_called);
static void pg_decode_stream_truncate(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn,
int nrelations, Relation relations[],
@@ -141,6 +150,7 @@ _PG_output_plugin_init(OutputPluginCallbacks *cb)
cb->filter_by_origin_cb = pg_decode_filter;
cb->shutdown_cb = pg_decode_shutdown;
cb->message_cb = pg_decode_message;
+ cb->sequence_cb = pg_decode_sequence;
cb->filter_prepare_cb = pg_decode_filter_prepare;
cb->begin_prepare_cb = pg_decode_begin_prepare_txn;
cb->prepare_cb = pg_decode_prepare_txn;
@@ -153,6 +163,7 @@ _PG_output_plugin_init(OutputPluginCallbacks *cb)
cb->stream_commit_cb = pg_decode_stream_commit;
cb->stream_change_cb = pg_decode_stream_change;
cb->stream_message_cb = pg_decode_stream_message;
+ cb->stream_sequence_cb = pg_decode_stream_sequence;
cb->stream_truncate_cb = pg_decode_stream_truncate;
}
@@ -175,6 +186,9 @@ pg_decode_startup(LogicalDecodingContext *ctx, OutputPluginOptions *opt,
data->skip_empty_xacts = false;
data->only_local = false;
+ /* skip sequences by default for backwards compatibility */
+ data->skip_sequences = true;
+
ctx->output_plugin_private = data;
opt->output_type = OUTPUT_PLUGIN_TEXTUAL_OUTPUT;
@@ -265,6 +279,17 @@ pg_decode_startup(LogicalDecodingContext *ctx, OutputPluginOptions *opt,
errmsg("could not parse value \"%s\" for parameter \"%s\"",
strVal(elem->arg), elem->defname)));
}
+ else if (strcmp(elem->defname, "skip-sequences") == 0)
+ {
+
+ if (elem->arg == NULL)
+ continue; /* true by default */
+ else if (!parse_bool(strVal(elem->arg), &data->skip_sequences))
+ ereport(ERROR,
+ (errcode(ERRCODE_INVALID_PARAMETER_VALUE),
+ errmsg("could not parse value \"%s\" for parameter \"%s\"",
+ strVal(elem->arg), elem->defname)));
+ }
else
{
ereport(ERROR,
@@ -744,6 +769,28 @@ pg_decode_message(LogicalDecodingContext *ctx,
OutputPluginWrite(ctx, true);
}
+static void
+pg_decode_sequence(LogicalDecodingContext *ctx, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ bool transactional, bool created,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ TestDecodingData *data = ctx->output_plugin_private;
+
+ /* return if requested to skip_sequences */
+ if (data->skip_sequences)
+ return;
+
+ OutputPluginPrepareWrite(ctx, true);
+ appendStringInfoString(ctx->out, "sequence ");
+ appendStringInfoString(ctx->out,
+ quote_qualified_identifier(get_namespace_name(get_rel_namespace(RelationGetRelid(rel))),
+ RelationGetRelationName(rel)));
+ appendStringInfo(ctx->out, ": transactional:%d created:%d last_value:%zu log_cnt:%zu is_called:%d",
+ transactional, created, last_value, log_cnt, is_called);
+ OutputPluginWrite(ctx, true);
+}
+
static void
pg_decode_stream_start(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn)
@@ -943,6 +990,28 @@ pg_decode_stream_message(LogicalDecodingContext *ctx,
OutputPluginWrite(ctx, true);
}
+static void
+pg_decode_stream_sequence(LogicalDecodingContext *ctx, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ bool transactional, bool created,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ TestDecodingData *data = ctx->output_plugin_private;
+
+ /* return if requested to skip_sequences */
+ if (data->skip_sequences)
+ return;
+
+ OutputPluginPrepareWrite(ctx, true);
+ appendStringInfoString(ctx->out, "streaming sequence ");
+ appendStringInfoString(ctx->out,
+ quote_qualified_identifier(get_namespace_name(get_rel_namespace(RelationGetRelid(rel))),
+ RelationGetRelationName(rel)));
+ appendStringInfo(ctx->out, ": transactional:%d created:%d last_value:%zu log_cnt:%zu is_called:%d",
+ transactional, created, last_value, log_cnt, is_called);
+ OutputPluginWrite(ctx, true);
+}
+
/*
* In streaming mode, we don't display the detailed information of Truncate.
* See pg_decode_stream_change.
--
2.31.1
[text/x-patch] 0002-rework-20211118.patch (35.4K, ../../[email protected]/4-0002-rework-20211118.patch)
download | inline diff:
From 43f63100b0e792a3f6e96c9c4f2218b27a731f52 Mon Sep 17 00:00:00 2001
From: Tomas Vondra <[email protected]>
Date: Thu, 18 Nov 2021 19:48:30 +0100
Subject: [PATCH 2/4] rework
---
src/backend/access/rmgrdesc/logicalmsgdesc.c | 13 ++
src/backend/access/transam/xact.c | 9 +
src/backend/commands/sequence.c | 180 +++++++++------
src/backend/replication/logical/decode.c | 104 +++------
src/backend/replication/logical/message.c | 3 +-
.../replication/logical/reorderbuffer.c | 213 ++++--------------
src/include/commands/sequence.h | 2 +
src/include/replication/message.h | 15 ++
src/include/replication/reorderbuffer.h | 10 +-
9 files changed, 232 insertions(+), 317 deletions(-)
diff --git a/src/backend/access/rmgrdesc/logicalmsgdesc.c b/src/backend/access/rmgrdesc/logicalmsgdesc.c
index d64ce2e7ef..c84628cd54 100644
--- a/src/backend/access/rmgrdesc/logicalmsgdesc.c
+++ b/src/backend/access/rmgrdesc/logicalmsgdesc.c
@@ -40,6 +40,16 @@ logicalmsg_desc(StringInfo buf, XLogReaderState *record)
sep = " ";
}
}
+ else if (info == XLOG_LOGICAL_SEQUENCE)
+ {
+ xl_logical_sequence *xlrec = (xl_logical_sequence *) rec;
+
+ appendStringInfo(buf, "rel %u/%u/%u last: %lu log_cnt: %lu is_called: %d",
+ xlrec->node.spcNode,
+ xlrec->node.dbNode,
+ xlrec->node.relNode,
+ xlrec->last, xlrec->log_cnt, xlrec->is_called);
+ }
}
const char *
@@ -48,5 +58,8 @@ logicalmsg_identify(uint8 info)
if ((info & ~XLR_INFO_MASK) == XLOG_LOGICAL_MESSAGE)
return "MESSAGE";
+ if ((info & ~XLR_INFO_MASK) == XLOG_LOGICAL_SEQUENCE)
+ return "SEQUENCE";
+
return NULL;
}
diff --git a/src/backend/access/transam/xact.c b/src/backend/access/transam/xact.c
index 8e35c432f5..eac8180250 100644
--- a/src/backend/access/transam/xact.c
+++ b/src/backend/access/transam/xact.c
@@ -36,6 +36,7 @@
#include "catalog/storage.h"
#include "commands/async.h"
#include "commands/tablecmds.h"
+#include "commands/sequence.h"
#include "commands/trigger.h"
#include "executor/spi.h"
#include "libpq/be-fsstubs.h"
@@ -2220,6 +2221,9 @@ CommitTransaction(void)
/* Prevent cancel/die interrupt while cleaning up */
HOLD_INTERRUPTS();
+ /* Write state of sequences to WAL */
+ AtEOXact_Sequences(true);
+
/* Commit updates to the relation map --- do this as late as possible */
AtEOXact_RelationMap(true, is_parallel_worker);
@@ -2378,6 +2382,8 @@ CommitTransaction(void)
* PrepareTransaction
*
* NB: if you change this routine, better look at CommitTransaction too!
+ *
+ * XXX Does this need to do something about logging of sequences?
*/
static void
PrepareTransaction(void)
@@ -2776,6 +2782,9 @@ AbortTransaction(void)
AtEOXact_RelationMap(false, is_parallel_worker);
AtAbort_Twophase();
+ /* XXX not sure we need to do anything about sequences at abort */
+ AtEOXact_Sequences(false);
+
/*
* Advertise the fact that we aborted in pg_xact (assuming that we got as
* far as assigning an XID to advertise). But if we're inside a parallel
diff --git a/src/backend/commands/sequence.c b/src/backend/commands/sequence.c
index a98fcc2e97..92d1966f34 100644
--- a/src/backend/commands/sequence.c
+++ b/src/backend/commands/sequence.c
@@ -37,6 +37,7 @@
#include "miscadmin.h"
#include "nodes/makefuncs.h"
#include "parser/parse_type.h"
+#include "replication/message.h"
#include "storage/lmgr.h"
#include "storage/proc.h"
#include "storage/smgr.h"
@@ -75,6 +76,7 @@ typedef struct SeqTableData
{
Oid relid; /* pg_class OID of this sequence (hash key) */
Oid filenode; /* last seen relfilenode of this sequence */
+ Oid tablespace; /* last seen tablespace of this sequence */
LocalTransactionId lxid; /* xact in which we last did a seq op */
bool last_valid; /* do we have a valid "last" value? */
int64 last; /* value last returned by nextval */
@@ -82,6 +84,10 @@ typedef struct SeqTableData
/* if last != cached, we have not used up all the cached values */
int64 increment; /* copy of sequence's increment field */
/* note that increment is zero until we first do nextval_internal() */
+ bool need_log; /* should be written to WAL at commit? */
+ bool touched;
+ int64 log_cnt;
+ bool is_called;
} SeqTableData;
typedef SeqTableData *SeqTable;
@@ -94,7 +100,7 @@ static HTAB *seqhashtab = NULL; /* hash table for SeqTable items */
*/
static SeqTableData *last_used_seq = NULL;
-static void fill_seq_with_data(Relation rel, HeapTuple tuple, bool create);
+static void fill_seq_with_data(Relation rel, HeapTuple tuple);
static Relation lock_and_open_sequence(SeqTable seq);
static void create_seq_hashtable(void);
static void init_sequence(Oid relid, SeqTable *p_elm, Relation *p_rel);
@@ -222,7 +228,7 @@ DefineSequence(ParseState *pstate, CreateSeqStmt *seq)
/* now initialize the sequence's data */
tuple = heap_form_tuple(tupDesc, value, null);
- fill_seq_with_data(rel, tuple, true);
+ fill_seq_with_data(rel, tuple);
/* process OWNED BY if given */
if (owned_by)
@@ -327,12 +333,17 @@ ResetSequence(Oid seq_relid)
/*
* Insert the modified tuple into the new storage file.
*/
- fill_seq_with_data(seq_rel, tuple, true);
+ fill_seq_with_data(seq_rel, tuple);
/* Clear local cache so that we don't think we have cached numbers */
/* Note that we do not change the currval() state */
elm->cached = elm->last;
+ /* Remember we need to write the sequence to WAL at commit. */
+ elm->need_log = true;
+ elm->is_called = false;
+ elm->log_cnt = 0;
+
relation_close(seq_rel, NoLock);
}
@@ -340,7 +351,7 @@ ResetSequence(Oid seq_relid)
* Initialize a sequence's relation with the specified tuple as content
*/
static void
-fill_seq_with_data(Relation rel, HeapTuple tuple, bool create)
+fill_seq_with_data(Relation rel, HeapTuple tuple)
{
Buffer buf;
Page page;
@@ -380,27 +391,7 @@ fill_seq_with_data(Relation rel, HeapTuple tuple, bool create)
if (RelationNeedsWAL(rel))
{
GetTopTransactionId();
-
- /*
- * Ensure we have a proper XID, which will be included in the XLOG
- * record by XLogRecordAssemble. Otherwise the first nextval() in
- * a subxact (without any preceding changes) would get XID 0,
- * and it'd be impossible to decide which top xact it belongs to.
- * It'd also trigger assert in DecodeSequence.
- *
- * XXX This might seem unnecessary, because if there's no XID the
- * transaction couldn't have done anything important yet, e.g. it
- * could not have created a sequence. But that's incorrect, as it
- * ignores subtransactions. The current subtransaction might not
- * have done anything yet (thus no XID), but an earlier one might
- * have created the sequence.
- *
- * XXX Not sure if this is the best solution. Maybe do this only
- * with wal_level=logical to minimize the overhead. OTOH advancing
- * the sequence is likely followed by using the value(s) for some
- * other activity, which assigns the XID.
- */
- GetCurrentTransactionId();
+ // GetCurrentTransactionId();
}
START_CRIT_SECTION();
@@ -422,7 +413,6 @@ fill_seq_with_data(Relation rel, HeapTuple tuple, bool create)
XLogRegisterBuffer(0, buf, REGBUF_WILL_INIT);
xlrec.node = rel->rd_node;
- xlrec.created = create;
XLogRegisterData((char *) &xlrec, sizeof(xl_seq_rec));
XLogRegisterData((char *) tuple->t_data, tuple->t_len);
@@ -526,7 +516,11 @@ AlterSequence(ParseState *pstate, AlterSeqStmt *stmt)
/*
* Insert the modified tuple into the new storage file.
*/
- fill_seq_with_data(seqrel, newdatatuple, true);
+ fill_seq_with_data(seqrel, newdatatuple);
+
+ elm->need_log = true;
+ elm->is_called = false;
+ elm->log_cnt = 0;
}
/* process OWNED BY if given */
@@ -792,27 +786,7 @@ nextval_internal(Oid relid, bool check_permissions)
if (logit && RelationNeedsWAL(seqrel))
{
GetTopTransactionId();
-
- /*
- * Ensure we have a proper XID, which will be included in the XLOG
- * record by XLogRecordAssemble. Otherwise the first nextval() in
- * a subxact (without any preceding changes) would get XID 0,
- * and it'd be impossible to decide which top xact it belongs to.
- * It'd also trigger assert in DecodeSequence.
- *
- * XXX This might seem unnecessary, because if there's no XID the
- * transaction couldn't have done anything important yet, e.g. it
- * could not have created a sequence. But that's incorrect, as it
- * ignores subtransactions. The current subtransaction might not
- * have done anything yet (thus no XID), but an earlier one might
- * have created the sequence.
- *
- * XXX Not sure if this is the best solution. Maybe do this only
- * with wal_level=logical to minimize the overhead. OTOH advancing
- * the sequence is likely followed by using the value(s) for some
- * other activity, which assigns the XID.
- */
- GetCurrentTransactionId();
+ // GetCurrentTransactionId();
}
/* ready to change the on-disk (or really, in-buffer) tuple */
@@ -850,7 +824,6 @@ nextval_internal(Oid relid, bool check_permissions)
seq->log_cnt = 0;
xlrec.node = seqrel->rd_node;
- xlrec.created = false;
XLogRegisterData((char *) &xlrec, sizeof(xl_seq_rec));
XLogRegisterData((char *) seqdatatuple.t_data, seqdatatuple.t_len);
@@ -858,6 +831,10 @@ nextval_internal(Oid relid, bool check_permissions)
recptr = XLogInsert(RM_SEQ_ID, XLOG_SEQ_LOG);
PageSetLSN(page, recptr);
+
+ elm->need_log = true;
+ elm->is_called = true;
+ elm->log_cnt = 0;
}
/* Now update sequence tuple to the intended final state */
@@ -1027,27 +1004,7 @@ do_setval(Oid relid, int64 next, bool iscalled)
if (RelationNeedsWAL(seqrel))
{
GetTopTransactionId();
-
- /*
- * Ensure we have a proper XID, which will be included in the XLOG
- * record by XLogRecordAssemble. Otherwise the first nextval() in
- * a subxact (without any preceding changes) would get XID 0,
- * and it'd be impossible to decide which top xact it belongs to.
- * It'd also trigger assert in DecodeSequence.
- *
- * XXX This might seem unnecessary, because if there's no XID the
- * transaction couldn't have done anything important yet, e.g. it
- * could not have created a sequence. But that's incorrect, as it
- * ignores subtransactions. The current subtransaction might not
- * have done anything yet (thus no XID), but an earlier one might
- * have created the sequence.
- *
- * XXX Not sure if this is the best solution. Maybe do this only
- * with wal_level=logical to minimize the overhead. OTOH advancing
- * the sequence is likely followed by using the value(s) for some
- * other activity, which assigns the XID.
- */
- GetCurrentTransactionId();
+ // GetCurrentTransactionId();
}
/* ready to change the on-disk (or really, in-buffer) tuple */
@@ -1070,14 +1027,16 @@ do_setval(Oid relid, int64 next, bool iscalled)
XLogRegisterBuffer(0, buf, REGBUF_WILL_INIT);
xlrec.node = seqrel->rd_node;
- xlrec.created = false;
-
XLogRegisterData((char *) &xlrec, sizeof(xl_seq_rec));
XLogRegisterData((char *) seqdatatuple.t_data, seqdatatuple.t_len);
recptr = XLogInsert(RM_SEQ_ID, XLOG_SEQ_LOG);
PageSetLSN(page, recptr);
+
+ elm->need_log = true;
+ elm->is_called = false;
+ elm->log_cnt = 0;
}
END_CRIT_SECTION();
@@ -1196,11 +1155,14 @@ init_sequence(Oid relid, SeqTable *p_elm, Relation *p_rel)
{
/* relid already filled in */
elm->filenode = InvalidOid;
+ elm->tablespace = InvalidOid;
elm->lxid = InvalidLocalTransactionId;
elm->last_valid = false;
elm->last = elm->cached = 0;
}
+ elm->touched = true;
+
/*
* Open the sequence relation.
*/
@@ -1219,6 +1181,7 @@ init_sequence(Oid relid, SeqTable *p_elm, Relation *p_rel)
*/
if (seqrel->rd_rel->relfilenode != elm->filenode)
{
+ elm->tablespace = seqrel->rd_rel->reltablespace;
elm->filenode = seqrel->rd_rel->relfilenode;
elm->cached = elm->last;
}
@@ -1977,6 +1940,8 @@ seq_redo(XLogReaderState *record)
/*
* Flush cached sequence information.
+ *
+ * XXX Do we need to WAL-log the entries based on need_log?
*/
void
ResetSequenceCaches(void)
@@ -2000,3 +1965,74 @@ seq_mask(char *page, BlockNumber blkno)
mask_unused_space(page);
}
+
+/* XXX Do this only for wal_level = logical, probably? */
+void
+AtEOXact_Sequences(bool isCommit)
+{
+ SeqTable entry;
+ HASH_SEQ_STATUS scan;
+
+ if (!seqhashtab)
+ return;
+
+ /*
+ * XXX If we run just nextval() that returns cached value, the xact may
+ * not have XID, but we expect it in reorderbuffer. Maybe treat that as
+ * transactional?
+ */
+ // GetTopTransactionId();
+ // GetCurrentTransactionId();
+
+ hash_seq_init(&scan, seqhashtab);
+
+ while ((entry = (SeqTable) hash_seq_search(&scan)))
+ {
+ RelFileNode rnode;
+ xl_logical_sequence xlrec;
+
+ if (!isCommit)
+ entry->touched = false;
+
+ /*
+ * If not touched in the current transaction, don't log anything.
+ * We leave needs_log set, so that if future transactions touch
+ * the sequence we'll log it properly.
+ */
+ if (!entry->touched)
+ continue;
+
+ /*
+ * Likewise, if the sequence does not need logging, we're done.
+ */
+ if (!entry->need_log)
+ continue;
+
+ /* if this is commit, we'll log the */
+ entry->need_log = false;
+ entry->touched = false;
+
+ rnode.spcNode = entry->tablespace; /* tablespace */
+ rnode.dbNode = MyDatabaseId; /* database */
+ rnode.relNode = entry->filenode; /* relation */
+
+ xlrec.node = rnode;
+ xlrec.reloid = entry->relid;
+
+ /* XXX is it good enough to log values we have in cache? seems
+ * wrong and we may need to re-read that. */
+ // xlrec.dbId = MyDatabaseId;
+ // xlrec.relId = entry->relid;
+ xlrec.last = entry->last;
+ xlrec.log_cnt = entry->log_cnt;
+ xlrec.is_called = entry->is_called;
+
+ XLogBeginInsert();
+ XLogRegisterData((char *) &xlrec, SizeOfLogicalSequence);
+
+ /* allow origin filtering */
+ XLogSetRecordFlags(XLOG_INCLUDE_ORIGIN);
+
+ (void) XLogInsert(RM_LOGICALMSG_ID, XLOG_LOGICAL_SEQUENCE);
+ }
+}
diff --git a/src/backend/replication/logical/decode.c b/src/backend/replication/logical/decode.c
index b61f2e264a..d29844adf9 100644
--- a/src/backend/replication/logical/decode.c
+++ b/src/backend/replication/logical/decode.c
@@ -59,6 +59,9 @@ static void DecodeXactOp(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
static void DecodeStandbyOp(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
static void DecodeLogicalMsgOp(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
+static void DecodeLogicalMessage(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
+static void DecodeLogicalSequence(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
+
/* individual record(group)'s handlers */
static void DecodeInsert(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
static void DecodeUpdate(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
@@ -75,11 +78,9 @@ static void DecodeAbort(LogicalDecodingContext *ctx, XLogRecordBuffer *buf,
bool two_phase);
static void DecodePrepare(LogicalDecodingContext *ctx, XLogRecordBuffer *buf,
xl_xact_parsed_prepare *parsed);
-static void DecodeSequence(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
/* common function to decode tuples */
static void DecodeXLogTuple(char *data, Size len, ReorderBufferTupleBuf *tup);
-static void DecodeSeqTuple(char *data, Size len, ReorderBufferTupleBuf *tuple);
/* helper functions for decoding transactions */
static inline bool FilterPrepare(LogicalDecodingContext *ctx,
@@ -160,10 +161,6 @@ LogicalDecodingProcessRecord(LogicalDecodingContext *ctx, XLogReaderState *recor
DecodeLogicalMsgOp(ctx, &buf);
break;
- case RM_SEQ_ID:
- DecodeSequence(ctx, &buf);
- break;
-
/*
* Rmgrs irrelevant for logical decoding; they describe stuff not
* represented in logical decoding. Add new rmgrs in rmgrlist.h's
@@ -626,6 +623,27 @@ FilterByOrigin(LogicalDecodingContext *ctx, RepOriginId origin_id)
*/
static void
DecodeLogicalMsgOp(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
+{
+ XLogReaderState *r = buf->record;
+ uint8 info = XLogRecGetInfo(r) & ~XLR_INFO_MASK;
+
+ switch (info)
+ {
+ case XLOG_LOGICAL_MESSAGE:
+ DecodeLogicalMessage(ctx, buf);
+ break;
+
+ case XLOG_LOGICAL_SEQUENCE:
+ DecodeLogicalSequence(ctx, buf);
+ break;
+
+ default:
+ elog(ERROR, "unexpected RM_LOGICALMSG_ID record type: %u", info);
+ }
+}
+
+static void
+DecodeLogicalMessage(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
{
SnapBuild *builder = ctx->snapshot_builder;
XLogReaderState *r = buf->record;
@@ -1319,31 +1337,6 @@ DecodeTXNNeedSkip(LogicalDecodingContext *ctx, XLogRecordBuffer *buf,
ctx->fast_forward || FilterByOrigin(ctx, origin_id));
}
-/*
- * Decode Sequence Tuple
- */
-static void
-DecodeSeqTuple(char *data, Size len, ReorderBufferTupleBuf *tuple)
-{
- int datalen = len - sizeof(xl_seq_rec) - SizeofHeapTupleHeader;
-
- Assert(datalen >= 0);
-
- tuple->tuple.t_len = datalen + SizeofHeapTupleHeader;;
-
- ItemPointerSetInvalid(&tuple->tuple.t_self);
-
- tuple->tuple.t_tableOid = InvalidOid;
-
- memcpy(((char *) tuple->tuple.t_data),
- data + sizeof(xl_seq_rec),
- SizeofHeapTupleHeader);
-
- memcpy(((char *) tuple->tuple.t_data) + SizeofHeapTupleHeader,
- data + sizeof(xl_seq_rec) + SizeofHeapTupleHeader,
- datalen);
-}
-
/*
* Handle sequence decode
*
@@ -1362,24 +1355,20 @@ DecodeSeqTuple(char *data, Size len, ReorderBufferTupleBuf *tuple)
* plugin - it might get confused about which sequence it's related to etc.
*/
static void
-DecodeSequence(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
+DecodeLogicalSequence(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
{
SnapBuild *builder = ctx->snapshot_builder;
- ReorderBufferTupleBuf *tuplebuf;
RelFileNode target_node;
XLogReaderState *r = buf->record;
- char *tupledata = NULL;
- Size tuplelen;
- Size datalen = 0;
TransactionId xid = XLogRecGetXid(r);
uint8 info = XLogRecGetInfo(buf->record) & ~XLR_INFO_MASK;
- xl_seq_rec *xlrec;
+ xl_logical_sequence *xlrec;
Snapshot snapshot;
RepOriginId origin_id = XLogRecGetOrigin(r);
- bool transactional;
+ bool transactional = TransactionIdIsValid(xid);
/* only decode changes flagged with XLOG_SEQ_LOG */
- if (info != XLOG_SEQ_LOG)
+ if (info != XLOG_LOGICAL_SEQUENCE)
elog(ERROR, "unexpected RM_SEQ_ID record type: %u", info);
/*
@@ -1390,8 +1379,11 @@ DecodeSequence(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
ctx->fast_forward)
return;
+ /* extract the WAL record, with "created" flag */
+ xlrec = (xl_logical_sequence *) XLogRecGetData(r);
+
/* only interested in our database */
- XLogRecGetBlockTag(r, 0, &target_node, NULL, NULL);
+ target_node = xlrec->node;
if (target_node.dbNode != ctx->slot->data.database)
return;
@@ -1399,44 +1391,18 @@ DecodeSequence(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
if (FilterByOrigin(ctx, XLogRecGetOrigin(r)))
return;
- tupledata = XLogRecGetData(r);
- datalen = XLogRecGetDataLen(r);
- tuplelen = datalen - SizeOfHeapHeader - sizeof(xl_seq_rec);
-
- /* extract the WAL record, with "created" flag */
- xlrec = (xl_seq_rec *) XLogRecGetData(r);
-
- /* XXX how could we have sequence change without data? */
- if(!datalen || !tupledata)
- return;
-
- tuplebuf = ReorderBufferGetTupleBuf(ctx->reorder, tuplelen);
- DecodeSeqTuple(tupledata, datalen, tuplebuf);
-
- /*
- * Should we handle the sequence increment as transactional or not?
- *
- * If the sequence was created in a still-running transaction, treat
- * it as transactional and queue the increments. Otherwise it needs
- * to be treated as non-transactional, in which case we send it to
- * the plugin right away.
- */
- transactional = ReorderBufferSequenceIsTransactional(ctx->reorder,
- target_node,
- xlrec->created);
-
/* Skip the change if already processed (per the snapshot). */
if (transactional &&
!SnapBuildProcessChange(builder, xid, buf->origptr))
return;
else if (!transactional &&
(SnapBuildCurrentState(builder) != SNAPBUILD_CONSISTENT ||
- SnapBuildXactNeedsSkip(builder, buf->origptr)))
+ SnapBuildXactNeedsSkip(builder, buf->origptr)))
return;
/* Queue the increment (or send immediately if not transactional). */
snapshot = SnapBuildGetOrBuildSnapshot(builder, xid);
ReorderBufferQueueSequence(ctx->reorder, xid, snapshot, buf->endptr,
- origin_id, target_node, transactional,
- xlrec->created, tuplebuf);
+ origin_id, xlrec->reloid, target_node,
+ xlrec->last, xlrec->log_cnt, xlrec->is_called);
}
diff --git a/src/backend/replication/logical/message.c b/src/backend/replication/logical/message.c
index 93bd372421..3f0ac57fc6 100644
--- a/src/backend/replication/logical/message.c
+++ b/src/backend/replication/logical/message.c
@@ -81,7 +81,8 @@ logicalmsg_redo(XLogReaderState *record)
{
uint8 info = XLogRecGetInfo(record) & ~XLR_INFO_MASK;
- if (info != XLOG_LOGICAL_MESSAGE)
+ if (info != XLOG_LOGICAL_MESSAGE &&
+ info != XLOG_LOGICAL_SEQUENCE)
elog(PANIC, "logicalmsg_redo: unknown op code %u", info);
/* This is only interesting for logical decoding, see decode.c. */
diff --git a/src/backend/replication/logical/reorderbuffer.c b/src/backend/replication/logical/reorderbuffer.c
index 434926459f..1d9a39ac4b 100644
--- a/src/backend/replication/logical/reorderbuffer.c
+++ b/src/backend/replication/logical/reorderbuffer.c
@@ -569,17 +569,11 @@ ReorderBufferReturnChange(ReorderBuffer *rb, ReorderBufferChange *change,
change->data.truncate.relids = NULL;
}
break;
- case REORDER_BUFFER_CHANGE_SEQUENCE:
- if (change->data.sequence.tuple)
- {
- ReorderBufferReturnTupleBuf(rb, change->data.sequence.tuple);
- change->data.sequence.tuple = NULL;
- }
- break;
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM:
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_ABORT:
case REORDER_BUFFER_CHANGE_INTERNAL_COMMAND_ID:
case REORDER_BUFFER_CHANGE_INTERNAL_TUPLECID:
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
break;
}
@@ -973,9 +967,10 @@ ReorderBufferSequenceCleanup(ReorderBuffer *rb, TransactionId xid)
void
ReorderBufferQueueSequence(ReorderBuffer *rb, TransactionId xid,
Snapshot snapshot, XLogRecPtr lsn, RepOriginId origin_id,
- RelFileNode rnode, bool transactional, bool created,
- ReorderBufferTupleBuf *tuplebuf)
+ Oid reloid, RelFileNode rnode, int64 last, int64 log_cnt, bool is_called)
{
+ bool transactional = TransactionIdIsValid(xid);
+
/*
* Change needs to be handled as transactional, because the sequence was
* created in a transaction that is still running. In that case all the
@@ -992,38 +987,6 @@ ReorderBufferQueueSequence(ReorderBuffer *rb, TransactionId xid,
MemoryContext oldcontext;
ReorderBufferChange *change;
- /* lookup sequence by relfilenode */
- ReorderBufferSequenceEnt *ent;
- bool found;
-
- /* transactional changes require a transaction */
- Assert(xid != InvalidTransactionId);
-
- /* search the lookup table (we ignore the return value, found is enough) */
- ent = hash_search(rb->sequences,
- (void *) &rnode,
- created ? HASH_ENTER : HASH_FIND,
- &found);
-
- /*
- * If this is the "create" increment, we must not have found any
- * pre-existing entry in the hash table (i.e. there must not be
- * any conflicting sequence).
- */
- Assert(!(created && found));
-
- /* But we must have either created or found an existing entry. */
- Assert(created || found);
-
- /*
- * When creating the sequence, remember the XID of the transaction
- * that created id.
- */
- if (created)
- ent->xid = xid;
-
- /* XXX Maybe check that we're still in the same top-level xact? */
-
/* OK, allocate and queue the change */
oldcontext = MemoryContextSwitchTo(rb->context);
@@ -1034,38 +997,26 @@ ReorderBufferQueueSequence(ReorderBuffer *rb, TransactionId xid,
memcpy(&change->data.sequence.relnode, &rnode, sizeof(RelFileNode));
- change->data.sequence.created = created;
- change->data.sequence.tuple = tuplebuf;
+ change->data.sequence.reloid = reloid;
+ change->data.sequence.last = last;
+ change->data.sequence.log_cnt = log_cnt;
+ change->data.sequence.is_called = is_called;
/* add it to the same subxact that created the sequence */
- ReorderBufferQueueChange(rb, ent->xid, lsn, change, false);
+ ReorderBufferQueueChange(rb, xid, lsn, change, false);
MemoryContextSwitchTo(oldcontext);
}
else
{
/*
- * This increment is for a sequence that was not created in any
- * running transaction, so we treat it as non-transactional and
- * just send it to the output plugin directly.
- */
+ * This increment is for a sequence that was not created in any
+ * running transaction, so we treat it as non-transactional and
+ * just send it to the output plugin directly.
+ */
ReorderBufferTXN *txn = NULL;
volatile Snapshot snapshot_now = snapshot;
- bool using_subtxn;
-
-#ifdef USE_ASSERT_CHECKING
- /* All "creates" have to be handled as transactional. */
- Assert(!created);
-
- /* Make sure the sequence is not in the hash table. */
- {
- bool found;
- hash_search(rb->sequences,
- (void *) &rnode,
- HASH_FIND, &found);
- Assert(!found);
- }
-#endif
+ bool using_subtxn;
if (xid != InvalidTransactionId)
txn = ReorderBufferTXNByXid(rb, xid, true, NULL, lsn, true);
@@ -1074,40 +1025,35 @@ ReorderBufferQueueSequence(ReorderBuffer *rb, TransactionId xid,
SetupHistoricSnapshot(snapshot_now, NULL);
/*
- * Decoding needs access to syscaches et al., which in turn use
- * heavyweight locks and such. Thus we need to have enough state around to
- * keep track of those. The easiest way is to simply use a transaction
- * internally. That also allows us to easily enforce that nothing writes
- * to the database by checking for xid assignments.
- *
- * When we're called via the SQL SRF there's already a transaction
- * started, so start an explicit subtransaction there.
- */
+ * Decoding needs access to syscaches et al., which in turn use
+ * heavyweight locks and such. Thus we need to have enough state around to
+ * keep track of those. The easiest way is to simply use a transaction
+ * internally. That also allows us to easily enforce that nothing writes
+ * to the database by checking for xid assignments.
+ *
+ * When we're called via the SQL SRF there's already a transaction
+ * started, so start an explicit subtransaction there.
+ */
using_subtxn = IsTransactionOrTransactionBlock();
PG_TRY();
{
Relation relation;
- HeapTuple tuple;
- bool isnull;
- int64 last_value, log_cnt;
- bool is_called;
- Oid reloid;
+// Oid reloid;
if (using_subtxn)
BeginInternalSubTransaction("sequence");
else
StartTransactionCommand();
-
+/*
reloid = RelidByRelfilenode(rnode.spcNode, rnode.relNode);
if (reloid == InvalidOid)
elog(ERROR, "could not map filenode \"%s\" to relation OID",
- relpathperm(rnode,
- MAIN_FORKNUM));
-
+ relpathperm(rnode,
+ MAIN_FORKNUM));
+*/
relation = RelationIdGetRelation(reloid);
- tuple = &tuplebuf->tuple;
/*
* Extract the internal sequence values, describing the state.
@@ -1115,12 +1061,8 @@ ReorderBufferQueueSequence(ReorderBuffer *rb, TransactionId xid,
* XXX Seems a bit strange to access it directly. Maybe there's
* a better / more correct way?
*/
- last_value = heap_getattr(tuple, 1, RelationGetDescr(relation), &isnull);
- log_cnt = heap_getattr(tuple, 2, RelationGetDescr(relation), &isnull);
- is_called = heap_getattr(tuple, 3, RelationGetDescr(relation), &isnull);
-
- rb->sequence(rb, txn, lsn, relation, transactional, created,
- last_value, log_cnt, is_called);
+ rb->sequence(rb, txn, lsn, relation, transactional, false,
+ last, log_cnt, is_called);
RelationClose(relation);
@@ -2248,31 +2190,27 @@ ReorderBufferApplySequence(ReorderBuffer *rb, ReorderBufferTXN *txn,
Relation relation, ReorderBufferChange *change,
bool streaming)
{
- HeapTuple tuple;
- bool isnull;
int64 last_value, log_cnt;
bool is_called;
- tuple = &change->data.sequence.tuple->tuple;
-
/*
* Extract the internal sequence values, describing the state.
*
* XXX Seems a bit strange to access it directly. Maybe there's
* a better / more correct way?
*/
- last_value = heap_getattr(tuple, 1, RelationGetDescr(relation), &isnull);
- log_cnt = heap_getattr(tuple, 2, RelationGetDescr(relation), &isnull);
- is_called = heap_getattr(tuple, 3, RelationGetDescr(relation), &isnull);
+ last_value = change->data.sequence.last;
+ log_cnt = change->data.sequence.log_cnt;
+ is_called = change->data.sequence.is_called;
/* Only ever called from ReorderBufferApplySequence, so transational. */
if (streaming)
rb->stream_sequence(rb, txn, change->lsn, relation, true,
- change->data.sequence.created,
+ false, /* FIXME */
last_value, log_cnt, is_called);
else
rb->sequence(rb, txn, change->lsn, relation, true,
- change->data.sequence.created,
+ false, /* FIXME */
last_value, log_cnt, is_called);
}
@@ -2722,15 +2660,17 @@ ReorderBufferProcessTXN(ReorderBuffer *rb, ReorderBufferTXN *txn,
case REORDER_BUFFER_CHANGE_SEQUENCE:
Assert(snapshot_now);
+ /*
reloid = RelidByRelfilenode(change->data.sequence.relnode.spcNode,
change->data.sequence.relnode.relNode);
if (reloid == InvalidOid)
- elog(ERROR, "could not map filenode \"%s\" to relation OID",
+ elog(ERROR, "zz could not map filenode \"%s\" to relation OID",
relpathperm(change->data.sequence.relnode,
MAIN_FORKNUM));
- relation = RelationIdGetRelation(reloid);
+ */
+ relation = RelationIdGetRelation(change->data.sequence.reloid);
if (!RelationIsValid(relation))
elog(ERROR, "could not open relation with OID %u (for filenode \"%s\")",
@@ -4127,45 +4067,13 @@ ReorderBufferSerializeChange(ReorderBuffer *rb, ReorderBufferTXN *txn,
memcpy(data, change->data.truncate.relids, size);
data += size;
- break;
- }
- case REORDER_BUFFER_CHANGE_SEQUENCE:
- {
- char *data;
- ReorderBufferTupleBuf *tup;
- Size len = 0;
-
- tup = change->data.sequence.tuple;
-
- if (tup)
- {
- sz += sizeof(HeapTupleData);
- len = tup->tuple.t_len;
- sz += len;
- }
-
- /* make sure we have enough space */
- ReorderBufferSerializeReserve(rb, sz);
-
- data = ((char *) rb->outbuf) + sizeof(ReorderBufferDiskChange);
- /* might have been reallocated above */
- ondisk = (ReorderBufferDiskChange *) rb->outbuf;
-
- if (len)
- {
- memcpy(data, &tup->tuple, sizeof(HeapTupleData));
- data += sizeof(HeapTupleData);
-
- memcpy(data, tup->tuple.t_data, len);
- data += len;
- }
-
break;
}
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM:
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_ABORT:
case REORDER_BUFFER_CHANGE_INTERNAL_COMMAND_ID:
case REORDER_BUFFER_CHANGE_INTERNAL_TUPLECID:
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
/* ReorderBufferChange contains everything important */
break;
}
@@ -4424,28 +4332,13 @@ ReorderBufferChangeSize(ReorderBufferChange *change)
{
sz += sizeof(Oid) * change->data.truncate.nrelids;
- break;
- }
- case REORDER_BUFFER_CHANGE_SEQUENCE:
- {
- ReorderBufferTupleBuf *tup;
- Size len = 0;
-
- tup = change->data.sequence.tuple;
-
- if (tup)
- {
- sz += sizeof(HeapTupleData);
- len = tup->tuple.t_len;
- sz += len;
- }
-
break;
}
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM:
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_ABORT:
case REORDER_BUFFER_CHANGE_INTERNAL_COMMAND_ID:
case REORDER_BUFFER_CHANGE_INTERNAL_TUPLECID:
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
/* ReorderBufferChange contains everything important */
break;
}
@@ -4742,33 +4635,11 @@ ReorderBufferRestoreChange(ReorderBuffer *rb, ReorderBufferTXN *txn,
break;
}
- case REORDER_BUFFER_CHANGE_SEQUENCE:
- if (change->data.sequence.tuple)
- {
- uint32 tuplelen = ((HeapTuple) data)->t_len;
-
- change->data.sequence.tuple =
- ReorderBufferGetTupleBuf(rb, tuplelen - SizeofHeapTupleHeader);
-
- /* restore ->tuple */
- memcpy(&change->data.sequence.tuple->tuple, data,
- sizeof(HeapTupleData));
- data += sizeof(HeapTupleData);
-
- /* reset t_data pointer into the new tuplebuf */
- change->data.sequence.tuple->tuple.t_data =
- ReorderBufferTupleBufData(change->data.sequence.tuple);
-
- /* restore tuple data itself */
- memcpy(change->data.sequence.tuple->tuple.t_data, data, tuplelen);
- data += tuplelen;
- }
- break;
-
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM:
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_ABORT:
case REORDER_BUFFER_CHANGE_INTERNAL_COMMAND_ID:
case REORDER_BUFFER_CHANGE_INTERNAL_TUPLECID:
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
break;
}
diff --git a/src/include/commands/sequence.h b/src/include/commands/sequence.h
index 5919fb90ee..ddb36b5b17 100644
--- a/src/include/commands/sequence.h
+++ b/src/include/commands/sequence.h
@@ -62,6 +62,8 @@ extern void DeleteSequenceTuple(Oid relid);
extern void ResetSequence(Oid seq_relid);
extern void ResetSequenceCaches(void);
+extern void AtEOXact_Sequences(bool isCommit);
+
extern void seq_redo(XLogReaderState *rptr);
extern void seq_desc(StringInfo buf, XLogReaderState *rptr);
extern const char *seq_identify(uint8 info);
diff --git a/src/include/replication/message.h b/src/include/replication/message.h
index d3fb324c81..a39e8e984d 100644
--- a/src/include/replication/message.h
+++ b/src/include/replication/message.h
@@ -27,13 +27,28 @@ typedef struct xl_logical_message
char message[FLEXIBLE_ARRAY_MEMBER];
} xl_logical_message;
+/*
+ * Generic logical decoding sequence wal record.
+ */
+typedef struct xl_logical_sequence
+{
+ RelFileNode node;
+ Oid reloid;
+ int64 last; /* last value emitted for sequence */
+ int64 log_cnt; /* last value cached for sequence */
+ bool is_called;
+} xl_logical_sequence;
+
#define SizeOfLogicalMessage (offsetof(xl_logical_message, message))
+#define SizeOfLogicalSequence (sizeof(xl_logical_sequence))
extern XLogRecPtr LogLogicalMessage(const char *prefix, const char *message,
size_t size, bool transactional);
/* RMGR API*/
#define XLOG_LOGICAL_MESSAGE 0x00
+#define XLOG_LOGICAL_SEQUENCE 0x10
+
void logicalmsg_redo(XLogReaderState *record);
void logicalmsg_desc(StringInfo buf, XLogReaderState *record);
const char *logicalmsg_identify(uint8 info);
diff --git a/src/include/replication/reorderbuffer.h b/src/include/replication/reorderbuffer.h
index 63e6ed037b..eb78d5dff4 100644
--- a/src/include/replication/reorderbuffer.h
+++ b/src/include/replication/reorderbuffer.h
@@ -164,8 +164,10 @@ typedef struct ReorderBufferChange
struct
{
RelFileNode relnode;
- bool created;
- ReorderBufferTupleBuf *tuple;
+ Oid reloid;
+ int64 last;
+ int64 log_cnt;
+ bool is_called;
} sequence;
} data;
@@ -672,8 +674,8 @@ void ReorderBufferQueueMessage(ReorderBuffer *, TransactionId, Snapshot snapsho
Size message_size, const char *message);
void ReorderBufferQueueSequence(ReorderBuffer *rb, TransactionId xid,
Snapshot snapshot, XLogRecPtr lsn, RepOriginId origin_id,
- RelFileNode rnode, bool transactional, bool created,
- ReorderBufferTupleBuf *tuplebuf);
+ Oid reloid, RelFileNode rnode,
+ int64 last, int64 log_cnt, bool is_called);
void ReorderBufferCommit(ReorderBuffer *, TransactionId,
XLogRecPtr commit_lsn, XLogRecPtr end_lsn,
TimestampTz commit_time, RepOriginId origin_id, XLogRecPtr origin_lsn);
--
2.31.1
[text/x-patch] 0001-Logical-decoding-of-sequences-20211118.patch (43.5K, ../../[email protected]/5-0001-Logical-decoding-of-sequences-20211118.patch)
download | inline diff:
From e61a7c9994133d1a0dc25c369c026205602b3314 Mon Sep 17 00:00:00 2001
From: Tomas Vondra <[email protected]>
Date: Fri, 24 Sep 2021 00:41:33 +0200
Subject: [PATCH 1/4] Logical decoding of sequences
This adds the infrastructure for logical decoding of sequences. Support
for built-in logical replication and test_decoding is added separately.
When decoding sequences, we differentiate between a sequences created in
a (running) transaction, and sequences created in other (already
committed) transactions. Changes for sequences created in the same
top-level transaction are treated as "transactional" i.e. just like any
other change from that transaction (and discarded in case of a
rollback). Changes for sequences created earlier are treated as not
transactional - are processed immediately, as if performed outside any
transaction (and thus not rolled back).
This mixed behavior is necessary - sequences are non-transactional (e.g.
ROLLBACK does not undo the sequence increments). But for new sequences,
we need to handle them in a transactional way, because if we ever get
some DDL support, the sequence won't exist until the transaction gets
applied. So we need to ensure the increments don't happen until the
sequence gets created.
To differentiate which sequences are "old" and which were created in a
still-running transaction, we track sequences created in running
transactions in a hash table. Each sequence is identified by it's
relfilenode, and at transaction commit we discard all sequences created
in that particular transaction. For each sequence we track the XID of
the (sub)transaction that created it, and we cleanup the sequences for
each subtransaction when it completes (commit/rollback).
We don't use the XID to check if it's the same top-level transaction.
It's enough to know it was created in an in-progress transaction, and we
know it must be the current one because otherwise it wouldn't see the
sequence object.
The main changes in this patch are:
1) ensure WAL-logging of all necessary info for sequence advances
We need to be able to associate the advance with a XID, but until now
sequence advance might have XID 0 if it was the first thing the
transaction did. So we ensure the transaction has XID.
Note: This is needed because of subxacts. A XID 0 might still have the
sequence created in a different subxact of the same top-level xact.
Then there's a "created" flag added to the xl_seq_rec, which is
necessary to differentiate WAL for created/reset sequences.
2) decoding / queueing of sequences into ReorderBuffer
This is mostly copy-paste of existing code for other decoded events.
3) tracking sequences created in in-progress transactions
We use a simple hash table, indexed by relfilenode.
---
src/backend/commands/sequence.c | 83 +++-
src/backend/replication/logical/decode.c | 131 +++++-
src/backend/replication/logical/logical.c | 89 ++++
.../replication/logical/reorderbuffer.c | 424 ++++++++++++++++++
src/include/commands/sequence.h | 1 +
src/include/replication/output_plugin.h | 29 ++
src/include/replication/reorderbuffer.h | 44 +-
7 files changed, 793 insertions(+), 8 deletions(-)
diff --git a/src/backend/commands/sequence.c b/src/backend/commands/sequence.c
index 72bfdc07a4..a98fcc2e97 100644
--- a/src/backend/commands/sequence.c
+++ b/src/backend/commands/sequence.c
@@ -94,7 +94,7 @@ static HTAB *seqhashtab = NULL; /* hash table for SeqTable items */
*/
static SeqTableData *last_used_seq = NULL;
-static void fill_seq_with_data(Relation rel, HeapTuple tuple);
+static void fill_seq_with_data(Relation rel, HeapTuple tuple, bool create);
static Relation lock_and_open_sequence(SeqTable seq);
static void create_seq_hashtable(void);
static void init_sequence(Oid relid, SeqTable *p_elm, Relation *p_rel);
@@ -222,7 +222,7 @@ DefineSequence(ParseState *pstate, CreateSeqStmt *seq)
/* now initialize the sequence's data */
tuple = heap_form_tuple(tupDesc, value, null);
- fill_seq_with_data(rel, tuple);
+ fill_seq_with_data(rel, tuple, true);
/* process OWNED BY if given */
if (owned_by)
@@ -327,7 +327,7 @@ ResetSequence(Oid seq_relid)
/*
* Insert the modified tuple into the new storage file.
*/
- fill_seq_with_data(seq_rel, tuple);
+ fill_seq_with_data(seq_rel, tuple, true);
/* Clear local cache so that we don't think we have cached numbers */
/* Note that we do not change the currval() state */
@@ -340,7 +340,7 @@ ResetSequence(Oid seq_relid)
* Initialize a sequence's relation with the specified tuple as content
*/
static void
-fill_seq_with_data(Relation rel, HeapTuple tuple)
+fill_seq_with_data(Relation rel, HeapTuple tuple, bool create)
{
Buffer buf;
Page page;
@@ -378,8 +378,31 @@ fill_seq_with_data(Relation rel, HeapTuple tuple)
/* check the comment above nextval_internal()'s equivalent call. */
if (RelationNeedsWAL(rel))
+ {
GetTopTransactionId();
+ /*
+ * Ensure we have a proper XID, which will be included in the XLOG
+ * record by XLogRecordAssemble. Otherwise the first nextval() in
+ * a subxact (without any preceding changes) would get XID 0,
+ * and it'd be impossible to decide which top xact it belongs to.
+ * It'd also trigger assert in DecodeSequence.
+ *
+ * XXX This might seem unnecessary, because if there's no XID the
+ * transaction couldn't have done anything important yet, e.g. it
+ * could not have created a sequence. But that's incorrect, as it
+ * ignores subtransactions. The current subtransaction might not
+ * have done anything yet (thus no XID), but an earlier one might
+ * have created the sequence.
+ *
+ * XXX Not sure if this is the best solution. Maybe do this only
+ * with wal_level=logical to minimize the overhead. OTOH advancing
+ * the sequence is likely followed by using the value(s) for some
+ * other activity, which assigns the XID.
+ */
+ GetCurrentTransactionId();
+ }
+
START_CRIT_SECTION();
MarkBufferDirty(buf);
@@ -399,6 +422,7 @@ fill_seq_with_data(Relation rel, HeapTuple tuple)
XLogRegisterBuffer(0, buf, REGBUF_WILL_INIT);
xlrec.node = rel->rd_node;
+ xlrec.created = create;
XLogRegisterData((char *) &xlrec, sizeof(xl_seq_rec));
XLogRegisterData((char *) tuple->t_data, tuple->t_len);
@@ -502,7 +526,7 @@ AlterSequence(ParseState *pstate, AlterSeqStmt *stmt)
/*
* Insert the modified tuple into the new storage file.
*/
- fill_seq_with_data(seqrel, newdatatuple);
+ fill_seq_with_data(seqrel, newdatatuple, true);
}
/* process OWNED BY if given */
@@ -766,8 +790,31 @@ nextval_internal(Oid relid, bool check_permissions)
* (Have to do that here, so we're outside the critical section)
*/
if (logit && RelationNeedsWAL(seqrel))
+ {
GetTopTransactionId();
+ /*
+ * Ensure we have a proper XID, which will be included in the XLOG
+ * record by XLogRecordAssemble. Otherwise the first nextval() in
+ * a subxact (without any preceding changes) would get XID 0,
+ * and it'd be impossible to decide which top xact it belongs to.
+ * It'd also trigger assert in DecodeSequence.
+ *
+ * XXX This might seem unnecessary, because if there's no XID the
+ * transaction couldn't have done anything important yet, e.g. it
+ * could not have created a sequence. But that's incorrect, as it
+ * ignores subtransactions. The current subtransaction might not
+ * have done anything yet (thus no XID), but an earlier one might
+ * have created the sequence.
+ *
+ * XXX Not sure if this is the best solution. Maybe do this only
+ * with wal_level=logical to minimize the overhead. OTOH advancing
+ * the sequence is likely followed by using the value(s) for some
+ * other activity, which assigns the XID.
+ */
+ GetCurrentTransactionId();
+ }
+
/* ready to change the on-disk (or really, in-buffer) tuple */
START_CRIT_SECTION();
@@ -803,6 +850,7 @@ nextval_internal(Oid relid, bool check_permissions)
seq->log_cnt = 0;
xlrec.node = seqrel->rd_node;
+ xlrec.created = false;
XLogRegisterData((char *) &xlrec, sizeof(xl_seq_rec));
XLogRegisterData((char *) seqdatatuple.t_data, seqdatatuple.t_len);
@@ -977,8 +1025,31 @@ do_setval(Oid relid, int64 next, bool iscalled)
/* check the comment above nextval_internal()'s equivalent call. */
if (RelationNeedsWAL(seqrel))
+ {
GetTopTransactionId();
+ /*
+ * Ensure we have a proper XID, which will be included in the XLOG
+ * record by XLogRecordAssemble. Otherwise the first nextval() in
+ * a subxact (without any preceding changes) would get XID 0,
+ * and it'd be impossible to decide which top xact it belongs to.
+ * It'd also trigger assert in DecodeSequence.
+ *
+ * XXX This might seem unnecessary, because if there's no XID the
+ * transaction couldn't have done anything important yet, e.g. it
+ * could not have created a sequence. But that's incorrect, as it
+ * ignores subtransactions. The current subtransaction might not
+ * have done anything yet (thus no XID), but an earlier one might
+ * have created the sequence.
+ *
+ * XXX Not sure if this is the best solution. Maybe do this only
+ * with wal_level=logical to minimize the overhead. OTOH advancing
+ * the sequence is likely followed by using the value(s) for some
+ * other activity, which assigns the XID.
+ */
+ GetCurrentTransactionId();
+ }
+
/* ready to change the on-disk (or really, in-buffer) tuple */
START_CRIT_SECTION();
@@ -999,6 +1070,8 @@ do_setval(Oid relid, int64 next, bool iscalled)
XLogRegisterBuffer(0, buf, REGBUF_WILL_INIT);
xlrec.node = seqrel->rd_node;
+ xlrec.created = false;
+
XLogRegisterData((char *) &xlrec, sizeof(xl_seq_rec));
XLogRegisterData((char *) seqdatatuple.t_data, seqdatatuple.t_len);
diff --git a/src/backend/replication/logical/decode.c b/src/backend/replication/logical/decode.c
index a2b69511b4..b61f2e264a 100644
--- a/src/backend/replication/logical/decode.c
+++ b/src/backend/replication/logical/decode.c
@@ -42,6 +42,7 @@
#include "replication/reorderbuffer.h"
#include "replication/snapbuild.h"
#include "storage/standby.h"
+#include "commands/sequence.h"
typedef struct XLogRecordBuffer
{
@@ -74,10 +75,11 @@ static void DecodeAbort(LogicalDecodingContext *ctx, XLogRecordBuffer *buf,
bool two_phase);
static void DecodePrepare(LogicalDecodingContext *ctx, XLogRecordBuffer *buf,
xl_xact_parsed_prepare *parsed);
-
+static void DecodeSequence(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
/* common function to decode tuples */
static void DecodeXLogTuple(char *data, Size len, ReorderBufferTupleBuf *tup);
+static void DecodeSeqTuple(char *data, Size len, ReorderBufferTupleBuf *tuple);
/* helper functions for decoding transactions */
static inline bool FilterPrepare(LogicalDecodingContext *ctx,
@@ -158,6 +160,10 @@ LogicalDecodingProcessRecord(LogicalDecodingContext *ctx, XLogReaderState *recor
DecodeLogicalMsgOp(ctx, &buf);
break;
+ case RM_SEQ_ID:
+ DecodeSequence(ctx, &buf);
+ break;
+
/*
* Rmgrs irrelevant for logical decoding; they describe stuff not
* represented in logical decoding. Add new rmgrs in rmgrlist.h's
@@ -173,7 +179,6 @@ LogicalDecodingProcessRecord(LogicalDecodingContext *ctx, XLogReaderState *recor
case RM_HASH_ID:
case RM_GIN_ID:
case RM_GIST_ID:
- case RM_SEQ_ID:
case RM_SPGIST_ID:
case RM_BRIN_ID:
case RM_COMMIT_TS_ID:
@@ -1313,3 +1318,125 @@ DecodeTXNNeedSkip(LogicalDecodingContext *ctx, XLogRecordBuffer *buf,
(txn_dbid != InvalidOid && txn_dbid != ctx->slot->data.database) ||
ctx->fast_forward || FilterByOrigin(ctx, origin_id));
}
+
+/*
+ * Decode Sequence Tuple
+ */
+static void
+DecodeSeqTuple(char *data, Size len, ReorderBufferTupleBuf *tuple)
+{
+ int datalen = len - sizeof(xl_seq_rec) - SizeofHeapTupleHeader;
+
+ Assert(datalen >= 0);
+
+ tuple->tuple.t_len = datalen + SizeofHeapTupleHeader;;
+
+ ItemPointerSetInvalid(&tuple->tuple.t_self);
+
+ tuple->tuple.t_tableOid = InvalidOid;
+
+ memcpy(((char *) tuple->tuple.t_data),
+ data + sizeof(xl_seq_rec),
+ SizeofHeapTupleHeader);
+
+ memcpy(((char *) tuple->tuple.t_data) + SizeofHeapTupleHeader,
+ data + sizeof(xl_seq_rec) + SizeofHeapTupleHeader,
+ datalen);
+}
+
+/*
+ * Handle sequence decode
+ *
+ * Decoding sequences is a bit tricky, because while most sequence actions
+ * are non-transactional (not subject to rollback), some need to be handled
+ * as transactional.
+ *
+ * By default, a sequence increment is non-transactional - we must not queue
+ * it in a transaction as other changes, because the transaction might get
+ * rolled back and we'd discard the increment. The downstream would not be
+ * notified about the increment, which is wrong.
+ *
+ * On the other hand, the sequence may be created in a transaction. In this
+ * case we *should* queue the change as other changes in the transaction,
+ * because we don't want to send the increments for unknown sequence to the
+ * plugin - it might get confused about which sequence it's related to etc.
+ */
+static void
+DecodeSequence(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
+{
+ SnapBuild *builder = ctx->snapshot_builder;
+ ReorderBufferTupleBuf *tuplebuf;
+ RelFileNode target_node;
+ XLogReaderState *r = buf->record;
+ char *tupledata = NULL;
+ Size tuplelen;
+ Size datalen = 0;
+ TransactionId xid = XLogRecGetXid(r);
+ uint8 info = XLogRecGetInfo(buf->record) & ~XLR_INFO_MASK;
+ xl_seq_rec *xlrec;
+ Snapshot snapshot;
+ RepOriginId origin_id = XLogRecGetOrigin(r);
+ bool transactional;
+
+ /* only decode changes flagged with XLOG_SEQ_LOG */
+ if (info != XLOG_SEQ_LOG)
+ elog(ERROR, "unexpected RM_SEQ_ID record type: %u", info);
+
+ /*
+ * If we don't have snapshot or we are just fast-forwarding, there is no
+ * point in decoding messages.
+ */
+ if (SnapBuildCurrentState(builder) < SNAPBUILD_FULL_SNAPSHOT ||
+ ctx->fast_forward)
+ return;
+
+ /* only interested in our database */
+ XLogRecGetBlockTag(r, 0, &target_node, NULL, NULL);
+ if (target_node.dbNode != ctx->slot->data.database)
+ return;
+
+ /* output plugin doesn't look for this origin, no need to queue */
+ if (FilterByOrigin(ctx, XLogRecGetOrigin(r)))
+ return;
+
+ tupledata = XLogRecGetData(r);
+ datalen = XLogRecGetDataLen(r);
+ tuplelen = datalen - SizeOfHeapHeader - sizeof(xl_seq_rec);
+
+ /* extract the WAL record, with "created" flag */
+ xlrec = (xl_seq_rec *) XLogRecGetData(r);
+
+ /* XXX how could we have sequence change without data? */
+ if(!datalen || !tupledata)
+ return;
+
+ tuplebuf = ReorderBufferGetTupleBuf(ctx->reorder, tuplelen);
+ DecodeSeqTuple(tupledata, datalen, tuplebuf);
+
+ /*
+ * Should we handle the sequence increment as transactional or not?
+ *
+ * If the sequence was created in a still-running transaction, treat
+ * it as transactional and queue the increments. Otherwise it needs
+ * to be treated as non-transactional, in which case we send it to
+ * the plugin right away.
+ */
+ transactional = ReorderBufferSequenceIsTransactional(ctx->reorder,
+ target_node,
+ xlrec->created);
+
+ /* Skip the change if already processed (per the snapshot). */
+ if (transactional &&
+ !SnapBuildProcessChange(builder, xid, buf->origptr))
+ return;
+ else if (!transactional &&
+ (SnapBuildCurrentState(builder) != SNAPBUILD_CONSISTENT ||
+ SnapBuildXactNeedsSkip(builder, buf->origptr)))
+ return;
+
+ /* Queue the increment (or send immediately if not transactional). */
+ snapshot = SnapBuildGetOrBuildSnapshot(builder, xid);
+ ReorderBufferQueueSequence(ctx->reorder, xid, snapshot, buf->endptr,
+ origin_id, target_node, transactional,
+ xlrec->created, tuplebuf);
+}
diff --git a/src/backend/replication/logical/logical.c b/src/backend/replication/logical/logical.c
index b7d9521576..811d9912a1 100644
--- a/src/backend/replication/logical/logical.c
+++ b/src/backend/replication/logical/logical.c
@@ -73,6 +73,10 @@ static void truncate_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
static void message_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
XLogRecPtr message_lsn, bool transactional,
const char *prefix, Size message_size, const char *message);
+static void sequence_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ bool transactional, bool created,
+ int64 last_value, int64 log_cnt, bool is_called);
/* streaming callbacks */
static void stream_start_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
@@ -90,6 +94,10 @@ static void stream_change_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn
static void stream_message_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
XLogRecPtr message_lsn, bool transactional,
const char *prefix, Size message_size, const char *message);
+static void stream_sequence_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ bool transactional, bool created,
+ int64 last_value, int64 log_cnt, bool is_called);
static void stream_truncate_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
int nrelations, Relation relations[], ReorderBufferChange *change);
@@ -217,6 +225,7 @@ StartupDecodingContext(List *output_plugin_options,
ctx->reorder->apply_truncate = truncate_cb_wrapper;
ctx->reorder->commit = commit_cb_wrapper;
ctx->reorder->message = message_cb_wrapper;
+ ctx->reorder->sequence = sequence_cb_wrapper;
/*
* To support streaming, we require start/stop/abort/commit/change
@@ -233,6 +242,7 @@ StartupDecodingContext(List *output_plugin_options,
(ctx->callbacks.stream_commit_cb != NULL) ||
(ctx->callbacks.stream_change_cb != NULL) ||
(ctx->callbacks.stream_message_cb != NULL) ||
+ (ctx->callbacks.stream_sequence_cb != NULL) ||
(ctx->callbacks.stream_truncate_cb != NULL);
/*
@@ -250,6 +260,7 @@ StartupDecodingContext(List *output_plugin_options,
ctx->reorder->stream_commit = stream_commit_cb_wrapper;
ctx->reorder->stream_change = stream_change_cb_wrapper;
ctx->reorder->stream_message = stream_message_cb_wrapper;
+ ctx->reorder->stream_sequence = stream_sequence_cb_wrapper;
ctx->reorder->stream_truncate = stream_truncate_cb_wrapper;
@@ -1202,6 +1213,43 @@ message_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
error_context_stack = errcallback.previous;
}
+static void
+sequence_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ bool transactional, bool created,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ LogicalDecodingContext *ctx = cache->private_data;
+ LogicalErrorCallbackState state;
+ ErrorContextCallback errcallback;
+
+ Assert(!ctx->fast_forward);
+
+ if (ctx->callbacks.sequence_cb == NULL)
+ return;
+
+ /* Push callback + info on the error context stack */
+ state.ctx = ctx;
+ state.callback_name = "sequence";
+ state.report_location = sequence_lsn;
+ errcallback.callback = output_plugin_error_callback;
+ errcallback.arg = (void *) &state;
+ errcallback.previous = error_context_stack;
+ error_context_stack = &errcallback;
+
+ /* set output state */
+ ctx->accept_writes = true;
+ ctx->write_xid = txn != NULL ? txn->xid : InvalidTransactionId;
+ ctx->write_location = sequence_lsn;
+
+ /* do the actual work: call callback */
+ ctx->callbacks.sequence_cb(ctx, txn, sequence_lsn, rel, transactional,
+ created, last_value, log_cnt, is_called);
+
+ /* Pop the error context stack */
+ error_context_stack = errcallback.previous;
+}
+
static void
stream_start_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
XLogRecPtr first_lsn)
@@ -1507,6 +1555,47 @@ stream_message_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
error_context_stack = errcallback.previous;
}
+static void
+stream_sequence_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ bool transactional, bool created,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ LogicalDecodingContext *ctx = cache->private_data;
+ LogicalErrorCallbackState state;
+ ErrorContextCallback errcallback;
+
+ Assert(!ctx->fast_forward);
+
+ /* We're only supposed to call this when streaming is supported. */
+ Assert(ctx->streaming);
+
+ /* this callback is optional */
+ if (ctx->callbacks.stream_sequence_cb == NULL)
+ return;
+
+ /* Push callback + info on the error context stack */
+ state.ctx = ctx;
+ state.callback_name = "stream_sequence";
+ state.report_location = sequence_lsn;
+ errcallback.callback = output_plugin_error_callback;
+ errcallback.arg = (void *) &state;
+ errcallback.previous = error_context_stack;
+ error_context_stack = &errcallback;
+
+ /* set output state */
+ ctx->accept_writes = true;
+ ctx->write_xid = txn != NULL ? txn->xid : InvalidTransactionId;
+ ctx->write_location = sequence_lsn;
+
+ /* do the actual work: call callback */
+ ctx->callbacks.sequence_cb(ctx, txn, sequence_lsn, rel, transactional,
+ created, last_value, log_cnt, is_called);
+
+ /* Pop the error context stack */
+ error_context_stack = errcallback.previous;
+}
+
static void
stream_truncate_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
int nrelations, Relation relations[],
diff --git a/src/backend/replication/logical/reorderbuffer.c b/src/backend/replication/logical/reorderbuffer.c
index 46e66608cf..434926459f 100644
--- a/src/backend/replication/logical/reorderbuffer.c
+++ b/src/backend/replication/logical/reorderbuffer.c
@@ -77,6 +77,35 @@
* a bit more memory to the oldest subtransactions, because it's likely
* they are the source for the next sequence of changes.
*
+ * When decoding sequences, we differentiate between a sequences created
+ * in a (running) transaction, and sequences created in other (already
+ * committed) transactions. Changes for sequences created in the same
+ * top-level transaction are treated as "transactional" i.e. just like
+ * any other change from that transaction (and discarded in case of a
+ * rollback). Changes for sequences created earlier are treated as not
+ * transactional - are processed immediately, as if performed outside
+ * any transaction (and thus not rolled back).
+ *
+ * This mixed behavior is necessary - sequences are non-transactional
+ * (e.g. ROLLBACK does not undo the sequence increments). But for new
+ * sequences, we need to handle them in a transactional way, because if
+ * we ever get some DDL support, the sequence won't exist until the
+ * transaction gets applied. So we need to ensure the increments don't
+ * happen until the sequence gets created.
+ *
+ * To differentiate which sequences are "old" and which were created
+ * in a still-running transaction, we track sequences created in running
+ * transactions in a hash table. Each sequence is identified by it's
+ * relfilenode, and at transaction commit we discard all sequences
+ * created in that particular transaction. For each sequence we track
+ * the XID of the (sub)transaction that created it, and we cleanup the
+ * sequences for each subtransaction when it completes (commit/rollback).
+ *
+ * We don't use the XID to check if it's the same top-level transaction.
+ * It's enough to know it was created in an in-progress transaction,
+ * and we know it must be the current one because otherwise it wouldn't
+ * see the sequence object.
+ *
* -------------------------------------------------------------------------
*/
#include "postgres.h"
@@ -116,6 +145,13 @@ typedef struct ReorderBufferTXNByIdEnt
ReorderBufferTXN *txn;
} ReorderBufferTXNByIdEnt;
+/* entry for hash table we use to track sequences created in running xacts */
+typedef struct ReorderBufferSequenceEnt
+{
+ RelFileNode rnode;
+ TransactionId xid;
+} ReorderBufferSequenceEnt;
+
/* data structures for (relfilenode, ctid) => (cmin, cmax) mapping */
typedef struct ReorderBufferTupleCidKey
{
@@ -339,6 +375,14 @@ ReorderBufferAllocate(void)
buffer->by_txn = hash_create("ReorderBufferByXid", 1000, &hash_ctl,
HASH_ELEM | HASH_BLOBS | HASH_CONTEXT);
+ /* hash table of sequences, mapping relfilenode to XID of transaction */
+ hash_ctl.keysize = sizeof(RelFileNode);
+ hash_ctl.entrysize = sizeof(ReorderBufferSequenceEnt);
+ hash_ctl.hcxt = buffer->context;
+
+ buffer->sequences = hash_create("ReorderBufferSequenceHash", 1000, &hash_ctl,
+ HASH_ELEM | HASH_BLOBS | HASH_CONTEXT);
+
buffer->by_txn_last_xid = InvalidTransactionId;
buffer->by_txn_last_txn = NULL;
@@ -525,6 +569,13 @@ ReorderBufferReturnChange(ReorderBuffer *rb, ReorderBufferChange *change,
change->data.truncate.relids = NULL;
}
break;
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
+ if (change->data.sequence.tuple)
+ {
+ ReorderBufferReturnTupleBuf(rb, change->data.sequence.tuple);
+ change->data.sequence.tuple = NULL;
+ }
+ break;
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM:
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_ABORT:
case REORDER_BUFFER_CHANGE_INTERNAL_COMMAND_ID:
@@ -859,6 +910,242 @@ ReorderBufferQueueMessage(ReorderBuffer *rb, TransactionId xid,
}
}
+/*
+ * Treat the sequence increment as transactional?
+ *
+ * The hash table tracks all sequences created in in-progress transactions,
+ * so we simply do a lookup (the sequence is identified by relfilende). If
+ * we find a match, the increment should be handled as transactional.
+ */
+bool
+ReorderBufferSequenceIsTransactional(ReorderBuffer *rb,
+ RelFileNode rnode, bool created)
+{
+ bool found = false;
+
+ if (created)
+ return true;
+
+ hash_search(rb->sequences,
+ (void *) &rnode,
+ HASH_FIND,
+ &found);
+
+ return found;
+}
+
+/*
+ * Cleanup sequences created in in-progress transactions.
+ *
+ * There's no way to search by XID, so we simply do a seqscan of all
+ * the entries in the hash table. Hopefully there are only a couple
+ * entries in most cases - people generally don't create many new
+ * sequences over and over.
+ */
+static void
+ReorderBufferSequenceCleanup(ReorderBuffer *rb, TransactionId xid)
+{
+ HASH_SEQ_STATUS scan_status;
+ ReorderBufferSequenceEnt *ent;
+
+ hash_seq_init(&scan_status, rb->sequences);
+ while ((ent = (ReorderBufferSequenceEnt *) hash_seq_search(&scan_status)) != NULL)
+ {
+ /* skip sequences not from this transaction */
+ if (ent->xid != xid)
+ continue;
+
+ (void) hash_search(rb->sequences,
+ (void *) &(ent->rnode),
+ HASH_REMOVE, NULL);
+ }
+}
+
+/*
+ * A transactional sequence increment is queued to be processed upon commit
+ * and a non-transactional increment gets processed immediately.
+ *
+ * A sequence update may be both transactional and non-transactional. When
+ * created in a running transaction, treat it as transactional and queue
+ * the change in it. Otherwise treat it as non-transactional, so that we
+ * don't forget the increment in case of a rollback.
+ */
+void
+ReorderBufferQueueSequence(ReorderBuffer *rb, TransactionId xid,
+ Snapshot snapshot, XLogRecPtr lsn, RepOriginId origin_id,
+ RelFileNode rnode, bool transactional, bool created,
+ ReorderBufferTupleBuf *tuplebuf)
+{
+ /*
+ * Change needs to be handled as transactional, because the sequence was
+ * created in a transaction that is still running. In that case all the
+ * changes need to be queued in that transaction, we must not send them
+ * to the downstream until the transaction commits.
+ *
+ * There's a bit of a trouble with subtransactions - we can't queue it
+ * into the subxact, because it might be rolled back and we'd lose the
+ * increment. We need to queue it into the same (sub)xact that created
+ * the sequence, which is why we track the XID in the hash table.
+ */
+ if (transactional)
+ {
+ MemoryContext oldcontext;
+ ReorderBufferChange *change;
+
+ /* lookup sequence by relfilenode */
+ ReorderBufferSequenceEnt *ent;
+ bool found;
+
+ /* transactional changes require a transaction */
+ Assert(xid != InvalidTransactionId);
+
+ /* search the lookup table (we ignore the return value, found is enough) */
+ ent = hash_search(rb->sequences,
+ (void *) &rnode,
+ created ? HASH_ENTER : HASH_FIND,
+ &found);
+
+ /*
+ * If this is the "create" increment, we must not have found any
+ * pre-existing entry in the hash table (i.e. there must not be
+ * any conflicting sequence).
+ */
+ Assert(!(created && found));
+
+ /* But we must have either created or found an existing entry. */
+ Assert(created || found);
+
+ /*
+ * When creating the sequence, remember the XID of the transaction
+ * that created id.
+ */
+ if (created)
+ ent->xid = xid;
+
+ /* XXX Maybe check that we're still in the same top-level xact? */
+
+ /* OK, allocate and queue the change */
+ oldcontext = MemoryContextSwitchTo(rb->context);
+
+ change = ReorderBufferGetChange(rb);
+
+ change->action = REORDER_BUFFER_CHANGE_SEQUENCE;
+ change->origin_id = origin_id;
+
+ memcpy(&change->data.sequence.relnode, &rnode, sizeof(RelFileNode));
+
+ change->data.sequence.created = created;
+ change->data.sequence.tuple = tuplebuf;
+
+ /* add it to the same subxact that created the sequence */
+ ReorderBufferQueueChange(rb, ent->xid, lsn, change, false);
+
+ MemoryContextSwitchTo(oldcontext);
+ }
+ else
+ {
+ /*
+ * This increment is for a sequence that was not created in any
+ * running transaction, so we treat it as non-transactional and
+ * just send it to the output plugin directly.
+ */
+ ReorderBufferTXN *txn = NULL;
+ volatile Snapshot snapshot_now = snapshot;
+ bool using_subtxn;
+
+#ifdef USE_ASSERT_CHECKING
+ /* All "creates" have to be handled as transactional. */
+ Assert(!created);
+
+ /* Make sure the sequence is not in the hash table. */
+ {
+ bool found;
+ hash_search(rb->sequences,
+ (void *) &rnode,
+ HASH_FIND, &found);
+ Assert(!found);
+ }
+#endif
+
+ if (xid != InvalidTransactionId)
+ txn = ReorderBufferTXNByXid(rb, xid, true, NULL, lsn, true);
+
+ /* setup snapshot to allow catalog access */
+ SetupHistoricSnapshot(snapshot_now, NULL);
+
+ /*
+ * Decoding needs access to syscaches et al., which in turn use
+ * heavyweight locks and such. Thus we need to have enough state around to
+ * keep track of those. The easiest way is to simply use a transaction
+ * internally. That also allows us to easily enforce that nothing writes
+ * to the database by checking for xid assignments.
+ *
+ * When we're called via the SQL SRF there's already a transaction
+ * started, so start an explicit subtransaction there.
+ */
+ using_subtxn = IsTransactionOrTransactionBlock();
+
+ PG_TRY();
+ {
+ Relation relation;
+ HeapTuple tuple;
+ bool isnull;
+ int64 last_value, log_cnt;
+ bool is_called;
+ Oid reloid;
+
+ if (using_subtxn)
+ BeginInternalSubTransaction("sequence");
+ else
+ StartTransactionCommand();
+
+ reloid = RelidByRelfilenode(rnode.spcNode, rnode.relNode);
+
+ if (reloid == InvalidOid)
+ elog(ERROR, "could not map filenode \"%s\" to relation OID",
+ relpathperm(rnode,
+ MAIN_FORKNUM));
+
+ relation = RelationIdGetRelation(reloid);
+ tuple = &tuplebuf->tuple;
+
+ /*
+ * Extract the internal sequence values, describing the state.
+ *
+ * XXX Seems a bit strange to access it directly. Maybe there's
+ * a better / more correct way?
+ */
+ last_value = heap_getattr(tuple, 1, RelationGetDescr(relation), &isnull);
+ log_cnt = heap_getattr(tuple, 2, RelationGetDescr(relation), &isnull);
+ is_called = heap_getattr(tuple, 3, RelationGetDescr(relation), &isnull);
+
+ rb->sequence(rb, txn, lsn, relation, transactional, created,
+ last_value, log_cnt, is_called);
+
+ RelationClose(relation);
+
+ TeardownHistoricSnapshot(false);
+
+ AbortCurrentTransaction();
+
+ if (using_subtxn)
+ RollbackAndReleaseCurrentSubTransaction();
+ }
+ PG_CATCH();
+ {
+ TeardownHistoricSnapshot(true);
+
+ AbortCurrentTransaction();
+
+ if (using_subtxn)
+ RollbackAndReleaseCurrentSubTransaction();
+
+ PG_RE_THROW();
+ }
+ PG_END_TRY();
+ }
+}
+
/*
* AssertTXNLsnOrder
* Verify LSN ordering of transaction lists in the reorderbuffer
@@ -1535,6 +1822,9 @@ ReorderBufferCleanupTXN(ReorderBuffer *rb, ReorderBufferTXN *txn)
&found);
Assert(found);
+ /* Remove sequences created in this transaction (if any). */
+ ReorderBufferSequenceCleanup(rb, txn->xid);
+
/* remove entries spilled to disk */
if (rbtxn_is_serialized(txn))
ReorderBufferRestoreCleanup(rb, txn);
@@ -1950,6 +2240,42 @@ ReorderBufferApplyMessage(ReorderBuffer *rb, ReorderBufferTXN *txn,
change->data.msg.message);
}
+/*
+ * Helper function for ReorderBufferProcessTXN for applying sequences.
+ */
+static inline void
+ReorderBufferApplySequence(ReorderBuffer *rb, ReorderBufferTXN *txn,
+ Relation relation, ReorderBufferChange *change,
+ bool streaming)
+{
+ HeapTuple tuple;
+ bool isnull;
+ int64 last_value, log_cnt;
+ bool is_called;
+
+ tuple = &change->data.sequence.tuple->tuple;
+
+ /*
+ * Extract the internal sequence values, describing the state.
+ *
+ * XXX Seems a bit strange to access it directly. Maybe there's
+ * a better / more correct way?
+ */
+ last_value = heap_getattr(tuple, 1, RelationGetDescr(relation), &isnull);
+ log_cnt = heap_getattr(tuple, 2, RelationGetDescr(relation), &isnull);
+ is_called = heap_getattr(tuple, 3, RelationGetDescr(relation), &isnull);
+
+ /* Only ever called from ReorderBufferApplySequence, so transational. */
+ if (streaming)
+ rb->stream_sequence(rb, txn, change->lsn, relation, true,
+ change->data.sequence.created,
+ last_value, log_cnt, is_called);
+ else
+ rb->sequence(rb, txn, change->lsn, relation, true,
+ change->data.sequence.created,
+ last_value, log_cnt, is_called);
+}
+
/*
* Function to store the command id and snapshot at the end of the current
* stream so that we can reuse the same while sending the next stream.
@@ -2392,6 +2718,31 @@ ReorderBufferProcessTXN(ReorderBuffer *rb, ReorderBufferTXN *txn,
case REORDER_BUFFER_CHANGE_INTERNAL_TUPLECID:
elog(ERROR, "tuplecid value in changequeue");
break;
+
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
+ Assert(snapshot_now);
+
+ reloid = RelidByRelfilenode(change->data.sequence.relnode.spcNode,
+ change->data.sequence.relnode.relNode);
+
+ if (reloid == InvalidOid)
+ elog(ERROR, "could not map filenode \"%s\" to relation OID",
+ relpathperm(change->data.sequence.relnode,
+ MAIN_FORKNUM));
+
+ relation = RelationIdGetRelation(reloid);
+
+ if (!RelationIsValid(relation))
+ elog(ERROR, "could not open relation with OID %u (for filenode \"%s\")",
+ reloid,
+ relpathperm(change->data.sequence.relnode,
+ MAIN_FORKNUM));
+
+ if (RelationIsLogicallyLogged(relation))
+ ReorderBufferApplySequence(rb, txn, relation, change, streaming);
+
+ RelationClose(relation);
+ break;
}
}
@@ -3776,6 +4127,39 @@ ReorderBufferSerializeChange(ReorderBuffer *rb, ReorderBufferTXN *txn,
memcpy(data, change->data.truncate.relids, size);
data += size;
+ break;
+ }
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
+ {
+ char *data;
+ ReorderBufferTupleBuf *tup;
+ Size len = 0;
+
+ tup = change->data.sequence.tuple;
+
+ if (tup)
+ {
+ sz += sizeof(HeapTupleData);
+ len = tup->tuple.t_len;
+ sz += len;
+ }
+
+ /* make sure we have enough space */
+ ReorderBufferSerializeReserve(rb, sz);
+
+ data = ((char *) rb->outbuf) + sizeof(ReorderBufferDiskChange);
+ /* might have been reallocated above */
+ ondisk = (ReorderBufferDiskChange *) rb->outbuf;
+
+ if (len)
+ {
+ memcpy(data, &tup->tuple, sizeof(HeapTupleData));
+ data += sizeof(HeapTupleData);
+
+ memcpy(data, tup->tuple.t_data, len);
+ data += len;
+ }
+
break;
}
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM:
@@ -4040,6 +4424,22 @@ ReorderBufferChangeSize(ReorderBufferChange *change)
{
sz += sizeof(Oid) * change->data.truncate.nrelids;
+ break;
+ }
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
+ {
+ ReorderBufferTupleBuf *tup;
+ Size len = 0;
+
+ tup = change->data.sequence.tuple;
+
+ if (tup)
+ {
+ sz += sizeof(HeapTupleData);
+ len = tup->tuple.t_len;
+ sz += len;
+ }
+
break;
}
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM:
@@ -4341,6 +4741,30 @@ ReorderBufferRestoreChange(ReorderBuffer *rb, ReorderBufferTXN *txn,
break;
}
+
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
+ if (change->data.sequence.tuple)
+ {
+ uint32 tuplelen = ((HeapTuple) data)->t_len;
+
+ change->data.sequence.tuple =
+ ReorderBufferGetTupleBuf(rb, tuplelen - SizeofHeapTupleHeader);
+
+ /* restore ->tuple */
+ memcpy(&change->data.sequence.tuple->tuple, data,
+ sizeof(HeapTupleData));
+ data += sizeof(HeapTupleData);
+
+ /* reset t_data pointer into the new tuplebuf */
+ change->data.sequence.tuple->tuple.t_data =
+ ReorderBufferTupleBufData(change->data.sequence.tuple);
+
+ /* restore tuple data itself */
+ memcpy(change->data.sequence.tuple->tuple.t_data, data, tuplelen);
+ data += tuplelen;
+ }
+ break;
+
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM:
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_ABORT:
case REORDER_BUFFER_CHANGE_INTERNAL_COMMAND_ID:
diff --git a/src/include/commands/sequence.h b/src/include/commands/sequence.h
index 40544dd4c7..5919fb90ee 100644
--- a/src/include/commands/sequence.h
+++ b/src/include/commands/sequence.h
@@ -48,6 +48,7 @@ typedef FormData_pg_sequence_data *Form_pg_sequence_data;
typedef struct xl_seq_rec
{
RelFileNode node;
+ bool created; /* is this a CREATE SEQUENCE */
/* SEQUENCE TUPLE DATA FOLLOWS AT THE END */
} xl_seq_rec;
diff --git a/src/include/replication/output_plugin.h b/src/include/replication/output_plugin.h
index 810495ed0e..57bc13f11c 100644
--- a/src/include/replication/output_plugin.h
+++ b/src/include/replication/output_plugin.h
@@ -88,6 +88,19 @@ typedef void (*LogicalDecodeMessageCB) (struct LogicalDecodingContext *ctx,
Size message_size,
const char *message);
+/*
+ * Called for the generic logical decoding sequences.
+ */
+typedef void (*LogicalDecodeSequenceCB) (struct LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional,
+ bool created,
+ int64 last_value,
+ int64 log_cnt,
+ bool is_called);
+
/*
* Filter changes by origin.
*/
@@ -199,6 +212,20 @@ typedef void (*LogicalDecodeStreamMessageCB) (struct LogicalDecodingContext *ctx
Size message_size,
const char *message);
+/*
+ * Called for the streaming generic logical decoding sequences from in-progress
+ * transactions.
+ */
+typedef void (*LogicalDecodeStreamSequenceCB) (struct LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional,
+ bool created,
+ int64 last_value,
+ int64 log_cnt,
+ bool is_called);
+
/*
* Callback for streaming truncates from in-progress transactions.
*/
@@ -219,6 +246,7 @@ typedef struct OutputPluginCallbacks
LogicalDecodeTruncateCB truncate_cb;
LogicalDecodeCommitCB commit_cb;
LogicalDecodeMessageCB message_cb;
+ LogicalDecodeSequenceCB sequence_cb;
LogicalDecodeFilterByOriginCB filter_by_origin_cb;
LogicalDecodeShutdownCB shutdown_cb;
@@ -237,6 +265,7 @@ typedef struct OutputPluginCallbacks
LogicalDecodeStreamCommitCB stream_commit_cb;
LogicalDecodeStreamChangeCB stream_change_cb;
LogicalDecodeStreamMessageCB stream_message_cb;
+ LogicalDecodeStreamSequenceCB stream_sequence_cb;
LogicalDecodeStreamTruncateCB stream_truncate_cb;
} OutputPluginCallbacks;
diff --git a/src/include/replication/reorderbuffer.h b/src/include/replication/reorderbuffer.h
index 5b40ff75f7..63e6ed037b 100644
--- a/src/include/replication/reorderbuffer.h
+++ b/src/include/replication/reorderbuffer.h
@@ -64,7 +64,8 @@ enum ReorderBufferChangeType
REORDER_BUFFER_CHANGE_INTERNAL_SPEC_INSERT,
REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM,
REORDER_BUFFER_CHANGE_INTERNAL_SPEC_ABORT,
- REORDER_BUFFER_CHANGE_TRUNCATE
+ REORDER_BUFFER_CHANGE_TRUNCATE,
+ REORDER_BUFFER_CHANGE_SEQUENCE
};
/* forward declaration */
@@ -158,6 +159,14 @@ typedef struct ReorderBufferChange
uint32 ninvalidations; /* Number of messages */
SharedInvalidationMessage *invalidations; /* invalidation message */
} inval;
+
+ /* Context data for Sequence changes */
+ struct
+ {
+ RelFileNode relnode;
+ bool created;
+ ReorderBufferTupleBuf *tuple;
+ } sequence;
} data;
/*
@@ -430,6 +439,15 @@ typedef void (*ReorderBufferMessageCB) (ReorderBuffer *rb,
const char *prefix, Size sz,
const char *message);
+/* sequence callback signature */
+typedef void (*ReorderBufferSequenceCB) (ReorderBuffer *rb,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional, bool created,
+ int64 last_value, int64 log_cnt,
+ bool is_called);
+
/* begin prepare callback signature */
typedef void (*ReorderBufferBeginPrepareCB) (ReorderBuffer *rb,
ReorderBufferTXN *txn);
@@ -496,6 +514,15 @@ typedef void (*ReorderBufferStreamMessageCB) (
const char *prefix, Size sz,
const char *message);
+/* stream sequence callback signature */
+typedef void (*ReorderBufferStreamSequenceCB) (ReorderBuffer *rb,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional, bool created,
+ int64 last_value, int64 log_cnt,
+ bool is_called);
+
/* stream truncate callback signature */
typedef void (*ReorderBufferStreamTruncateCB) (
ReorderBuffer *rb,
@@ -511,6 +538,12 @@ struct ReorderBuffer
*/
HTAB *by_txn;
+ /*
+ * relfilenode => XID lookup table for sequences created in a transaction
+ * (also includes altered sequences, which assigns new relfilenode)
+ */
+ HTAB *sequences;
+
/*
* Transactions that could be a toplevel xact, ordered by LSN of the first
* record bearing that xid.
@@ -541,6 +574,7 @@ struct ReorderBuffer
ReorderBufferApplyTruncateCB apply_truncate;
ReorderBufferCommitCB commit;
ReorderBufferMessageCB message;
+ ReorderBufferSequenceCB sequence;
/*
* Callbacks to be called when streaming a transaction at prepare time.
@@ -560,6 +594,7 @@ struct ReorderBuffer
ReorderBufferStreamCommitCB stream_commit;
ReorderBufferStreamChangeCB stream_change;
ReorderBufferStreamMessageCB stream_message;
+ ReorderBufferStreamSequenceCB stream_sequence;
ReorderBufferStreamTruncateCB stream_truncate;
/*
@@ -635,6 +670,10 @@ void ReorderBufferQueueChange(ReorderBuffer *, TransactionId,
void ReorderBufferQueueMessage(ReorderBuffer *, TransactionId, Snapshot snapshot, XLogRecPtr lsn,
bool transactional, const char *prefix,
Size message_size, const char *message);
+void ReorderBufferQueueSequence(ReorderBuffer *rb, TransactionId xid,
+ Snapshot snapshot, XLogRecPtr lsn, RepOriginId origin_id,
+ RelFileNode rnode, bool transactional, bool created,
+ ReorderBufferTupleBuf *tuplebuf);
void ReorderBufferCommit(ReorderBuffer *, TransactionId,
XLogRecPtr commit_lsn, XLogRecPtr end_lsn,
TimestampTz commit_time, RepOriginId origin_id, XLogRecPtr origin_lsn);
@@ -682,4 +721,7 @@ void ReorderBufferSetRestartPoint(ReorderBuffer *, XLogRecPtr ptr);
void StartupReorderBuffer(void);
+bool ReorderBufferSequenceIsTransactional(ReorderBuffer *rb,
+ RelFileNode rnode, bool created);
+
#endif
--
2.31.1
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2021-11-22 00:47 Tomas Vondra <[email protected]>
parent: Tomas Vondra <[email protected]>
0 siblings, 2 replies; 127+ messages in thread
From: Tomas Vondra @ 2021-11-22 00:47 UTC (permalink / raw)
To: Hannu Krosing <[email protected]>; +Cc: Peter Eisentraut <[email protected]>; vignesh C <[email protected]>; PostgreSQL Hackers <[email protected]>; Petr Jelinek <[email protected]>
Hi,
Here's a slightly improved version of the patch, fixing a couple issues
I failed to notice. It also addresses a couple of the issues described
in the last message, although mostly to show what would need to be done.
1) Handle sequences dropped in the xact by calling try_relation_open,
and just doing nothing if not found. Otherwise we'd get a failure in
reorderbuffer, when decoding the change.
2) Fixed nextval_internal() to log the correct last_value (the one we
write into WAL).
3) Reread the sequence state in AtEOXact_Sequences, to deal with the
ordering issue described before. This makes (2) somewhat pointless,
because we just read whatever is on disk at that point. But having both
makes it easier to experiment / see what'd happen.
4) Log the stats in DefineSequence() - Without this we'd not have the
initial sequence state in the WAL, because only nextval/setval etc. do
the logging. The old approach (decoding the sequence tuple) does not
have this issue.
The (3) changes the behavior in a somewhat strange way. Consider this
case with two concurrent transactions:
T1: BEGIN;
T2: BEGIN;
T1: SELECT nextval('s') FROM generate_series(1,100) s(i);
T2: SELECT nextval('s') FROM generate_series(1,100) s(i);
T1: COMMIT;
T2: COMMIT;
The result is that both transactions have used the same sequence, and so
will re-read the state from disk. But at that point the state is exactly
the same, so we'll log the same thing twice.
There's a much deeper issue, though. The current patch only logs the
sequence if the session generated WAL when incrementing the sequence
(which happens every 32 values). But other sessions may already use
values from this range, so consider for example this:
T1: BEGIN;
T1: SELECT nextval('s') FROM generate_series(1,100) s(i);
T2: BEGIN;
T2: SELECT nextval('s');
T2: COMMIT;
T1: ROLLBACK;
Which unfortunately means T2 already used a value, but the increment may
not be logged at that time (or ever). This seems like a fatal issue,
because it means we need to log *all* sequences the transaction touches,
not just those that wrote the increment to WAL. That might still work
for large transactions consuming many sequence values, but it's pretty
inefficient for small OLTP transactions that only need one or two values
from the sequence.
So I think just decoding the sequence tuples is a better solution - for
large transactions (consuming many values from the sequence) it may be
more expensive (i.e. send more records to replica). But I doubt that
matters too much - it's likely negligible compared to other data for
large transactions.
regards
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
Attachments:
[text/x-patch] 0004-Add-support-for-decoding-sequences-to-built-20211121.patch (77.1K, ../../[email protected]/2-0004-Add-support-for-decoding-sequences-to-built-20211121.patch)
download | inline diff:
From 90d8dde242b9b88573fd09bb7ebc6da66876b674 Mon Sep 17 00:00:00 2001
From: Tomas Vondra <[email protected]>
Date: Fri, 19 Nov 2021 01:12:23 +0100
Subject: [PATCH 4/4] Add support for decoding sequences to built-in
replication
---
doc/src/sgml/catalogs.sgml | 71 ++++
doc/src/sgml/ref/alter_publication.sgml | 24 +-
doc/src/sgml/ref/alter_subscription.sgml | 4 +-
src/backend/catalog/pg_publication.c | 149 ++++++++-
src/backend/catalog/system_views.sql | 10 +
src/backend/commands/publicationcmds.c | 350 +++++++++++++++++++-
src/backend/commands/sequence.c | 79 +++++
src/backend/commands/subscriptioncmds.c | 272 +++++++++++++++
src/backend/executor/execReplication.c | 2 +-
src/backend/nodes/copyfuncs.c | 1 +
src/backend/nodes/equalfuncs.c | 1 +
src/backend/parser/gram.y | 32 ++
src/backend/replication/logical/proto.c | 54 +++
src/backend/replication/logical/tablesync.c | 118 ++++++-
src/backend/replication/logical/worker.c | 60 ++++
src/backend/replication/pgoutput/pgoutput.c | 86 ++++-
src/backend/utils/cache/relcache.c | 4 +-
src/bin/psql/tab-complete.c | 6 +-
src/include/catalog/pg_proc.dat | 5 +
src/include/catalog/pg_publication.h | 14 +
src/include/commands/sequence.h | 1 +
src/include/nodes/parsenodes.h | 6 +
src/include/replication/logicalproto.h | 20 ++
src/include/replication/pgoutput.h | 1 +
src/test/regress/expected/rules.out | 8 +
src/test/subscription/t/023_sequences.pl | 196 +++++++++++
26 files changed, 1542 insertions(+), 32 deletions(-)
create mode 100644 src/test/subscription/t/023_sequences.pl
diff --git a/doc/src/sgml/catalogs.sgml b/doc/src/sgml/catalogs.sgml
index c1d11be73f..37b9b52b29 100644
--- a/doc/src/sgml/catalogs.sgml
+++ b/doc/src/sgml/catalogs.sgml
@@ -9498,6 +9498,11 @@ SCRAM-SHA-256$<replaceable><iteration count></replaceable>:<replaceable>&l
<entry>prepared transactions</entry>
</row>
+ <row>
+ <entry><link linkend="view-pg-publication-sequences"><structname>pg_publication_sequences</structname></link></entry>
+ <entry>publications and their associated sequences</entry>
+ </row>
+
<row>
<entry><link linkend="view-pg-publication-tables"><structname>pg_publication_tables</structname></link></entry>
<entry>publications and their associated tables</entry>
@@ -11333,6 +11338,72 @@ SELECT * FROM pg_locks pl LEFT JOIN pg_prepared_xacts ppx
</sect1>
+ <sect1 id="view-pg-publication-sequences">
+ <title><structname>pg_publication_sequences</structname></title>
+
+ <indexterm zone="view-pg-publication-sequences">
+ <primary>pg_publication_sequences</primary>
+ </indexterm>
+
+ <para>
+ The view <structname>pg_publication_sequences</structname> provides
+ information about the mapping between publications and the sequences they
+ contain. Unlike the underlying catalog
+ <link linkend="catalog-pg-publication-rel"><structname>pg_publication_rel</structname></link>,
+ this view expands
+ publications defined as <literal>FOR ALL SEQUENCES</literal>, so for such
+ publications there will be a row for each eligible sequence.
+ </para>
+
+ <table>
+ <title><structname>pg_publication_sequences</structname> Columns</title>
+ <tgroup cols="1">
+ <thead>
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ Column Type
+ </para>
+ <para>
+ Description
+ </para></entry>
+ </row>
+ </thead>
+
+ <tbody>
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ <structfield>pubname</structfield> <type>name</type>
+ (references <link linkend="catalog-pg-publication"><structname>pg_publication</structname></link>.<structfield>pubname</structfield>)
+ </para>
+ <para>
+ Name of publication
+ </para></entry>
+ </row>
+
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ <structfield>schemaname</structfield> <type>name</type>
+ (references <link linkend="catalog-pg-namespace"><structname>pg_namespace</structname></link>.<structfield>nspname</structfield>)
+ </para>
+ <para>
+ Name of schema containing sequence
+ </para></entry>
+ </row>
+
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ <structfield>sequencename</structfield> <type>name</type>
+ (references <link linkend="catalog-pg-class"><structname>pg_class</structname></link>.<structfield>relname</structfield>)
+ </para>
+ <para>
+ Name of sequence
+ </para></entry>
+ </row>
+ </tbody>
+ </tgroup>
+ </table>
+ </sect1>
+
<sect1 id="view-pg-publication-tables">
<title><structname>pg_publication_tables</structname></title>
diff --git a/doc/src/sgml/ref/alter_publication.sgml b/doc/src/sgml/ref/alter_publication.sgml
index bb4ef5e5e2..4d166ad3f9 100644
--- a/doc/src/sgml/ref/alter_publication.sgml
+++ b/doc/src/sgml/ref/alter_publication.sgml
@@ -31,7 +31,9 @@ ALTER PUBLICATION <replaceable class="parameter">name</replaceable> RENAME TO <r
<phrase>where <replaceable class="parameter">publication_object</replaceable> is one of:</phrase>
TABLE [ ONLY ] <replaceable class="parameter">table_name</replaceable> [ * ] [, ... ]
+ SEQUENCE <replaceable class="parameter">sequence_name</replaceable> [ * ] [, ... ]
ALL TABLES IN SCHEMA { <replaceable class="parameter">schema_name</replaceable> | CURRENT_SCHEMA } [, ... ]
+ ALL SEQUENCE IN SCHEMA { <replaceable class="parameter">schema_name</replaceable> | CURRENT_SCHEMA } [, ... ]
</synopsis>
</refsynopsisdiv>
@@ -56,7 +58,18 @@ ALTER PUBLICATION <replaceable class="parameter">name</replaceable> RENAME TO <r
</para>
<para>
- The fourth variant of this command listed in the synopsis can change
+ The next three variants change which sequences are part of the publication.
+ The <literal>SET SEQUENCE</literal> clause will replace the list of sequences
+ in the publication with the specified one. The <literal>ADD SEQUENCE</literal>
+ and <literal>DROP SEQUENCE</literal> clauses will add and remove one or more
+ sequences from the publication. Note that adding sequences to a publication
+ that is already subscribed to will require a <literal>ALTER SUBSCRIPTION
+ ... REFRESH PUBLICATION</literal> action on the subscribing side in order
+ to become effective.
+ </para>
+
+ <para>
+ The seventh variant of this command listed in the synopsis can change
all of the publication properties specified in
<xref linkend="sql-createpublication"/>. Properties not mentioned in the
command retain their previous settings.
@@ -123,6 +136,15 @@ ALTER PUBLICATION <replaceable class="parameter">name</replaceable> RENAME TO <r
</listitem>
</varlistentry>
+ <varlistentry>
+ <term><replaceable class="parameter">sequence_name</replaceable></term>
+ <listitem>
+ <para>
+ Name of an existing sequence.
+ </para>
+ </listitem>
+ </varlistentry>
+
<varlistentry>
<term><literal>SET ( <replaceable class="parameter">publication_parameter</replaceable> [= <replaceable class="parameter">value</replaceable>] [, ... ] )</literal></term>
<listitem>
diff --git a/doc/src/sgml/ref/alter_subscription.sgml b/doc/src/sgml/ref/alter_subscription.sgml
index 0b027cc346..8f28cf03f4 100644
--- a/doc/src/sgml/ref/alter_subscription.sgml
+++ b/doc/src/sgml/ref/alter_subscription.sgml
@@ -147,7 +147,7 @@ ALTER SUBSCRIPTION <replaceable class="parameter">name</replaceable> RENAME TO <
<listitem>
<para>
Fetch missing table information from publisher. This will start
- replication of tables that were added to the subscribed-to publications
+ replication of tables and sequences that were added to the subscribed-to publications
since <command>CREATE SUBSCRIPTION</command> or
the last invocation of <command>REFRESH PUBLICATION</command>.
</para>
@@ -164,7 +164,7 @@ ALTER SUBSCRIPTION <replaceable class="parameter">name</replaceable> RENAME TO <
Specifies whether to copy pre-existing data in the publications
that are being subscribed to when the replication starts.
The default is <literal>true</literal>. (Previously-subscribed
- tables are not copied.)
+ tables and sequences are not copied.)
</para>
</listitem>
</varlistentry>
diff --git a/src/backend/catalog/pg_publication.c b/src/backend/catalog/pg_publication.c
index 63579b2f82..2d5bc73a96 100644
--- a/src/backend/catalog/pg_publication.c
+++ b/src/backend/catalog/pg_publication.c
@@ -54,7 +54,8 @@ check_publication_add_relation(Relation targetrel)
{
/* Must be a regular or partitioned table */
if (RelationGetForm(targetrel)->relkind != RELKIND_RELATION &&
- RelationGetForm(targetrel)->relkind != RELKIND_PARTITIONED_TABLE)
+ RelationGetForm(targetrel)->relkind != RELKIND_PARTITIONED_TABLE &&
+ RelationGetForm(targetrel)->relkind != RELKIND_SEQUENCE)
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("cannot add relation \"%s\" to publication",
@@ -131,7 +132,8 @@ static bool
is_publishable_class(Oid relid, Form_pg_class reltuple)
{
return (reltuple->relkind == RELKIND_RELATION ||
- reltuple->relkind == RELKIND_PARTITIONED_TABLE) &&
+ reltuple->relkind == RELKIND_PARTITIONED_TABLE ||
+ reltuple->relkind == RELKIND_SEQUENCE) &&
!IsCatalogRelationOid(relid) &&
reltuple->relpersistence == RELPERSISTENCE_PERMANENT &&
relid >= FirstNormalObjectId;
@@ -480,6 +482,11 @@ GetPublicationRelations(Oid pubid, PublicationPartOpt pub_partopt)
Form_pg_publication_rel pubrel;
pubrel = (Form_pg_publication_rel) GETSTRUCT(tup);
+
+ /* skip sequences here */
+ if (get_rel_relkind(pubrel->prrelid) == RELKIND_SEQUENCE)
+ continue;
+
result = GetPubPartitionOptionRelations(result, pub_partopt,
pubrel->prrelid);
}
@@ -490,6 +497,49 @@ GetPublicationRelations(Oid pubid, PublicationPartOpt pub_partopt)
return result;
}
+/*
+ * Gets list of relation oids for a publication (sequences only).
+ *
+ * This should only be used for normal publications, the FOR ALL TABLES
+ * should use GetAllSequencesPublicationRelations().
+ */
+List *
+GetPublicationSequenceRelations(Oid pubid)
+{
+ List *result;
+ Relation pubrelsrel;
+ ScanKeyData scankey;
+ SysScanDesc scan;
+ HeapTuple tup;
+
+ /* Find all publications associated with the relation. */
+ pubrelsrel = table_open(PublicationRelRelationId, AccessShareLock);
+
+ ScanKeyInit(&scankey,
+ Anum_pg_publication_rel_prpubid,
+ BTEqualStrategyNumber, F_OIDEQ,
+ ObjectIdGetDatum(pubid));
+
+ scan = systable_beginscan(pubrelsrel, PublicationRelPrrelidPrpubidIndexId,
+ true, NULL, 1, &scankey);
+
+ result = NIL;
+ while (HeapTupleIsValid(tup = systable_getnext(scan)))
+ {
+ Form_pg_publication_rel pubrel;
+
+ pubrel = (Form_pg_publication_rel) GETSTRUCT(tup);
+
+ if (get_rel_relkind(pubrel->prrelid) == RELKIND_SEQUENCE)
+ result = lappend_oid(result, pubrel->prrelid);
+ }
+
+ systable_endscan(scan);
+ table_close(pubrelsrel, AccessShareLock);
+
+ return result;
+}
+
/*
* Gets list of publication oids for publications marked as FOR ALL TABLES.
*/
@@ -735,6 +785,46 @@ GetAllSchemaPublicationRelations(Oid pubid, PublicationPartOpt pub_partopt)
return result;
}
+/*
+ * Gets list of all relation published by FOR ALL TABLES publication(s).
+ *
+ * If the publication publishes partition changes via their respective root
+ * partitioned tables, we must exclude partitions in favor of including the
+ * root partitioned tables.
+ */
+List *
+GetAllSequencesPublicationRelations(void)
+{
+ Relation classRel;
+ ScanKeyData key[1];
+ TableScanDesc scan;
+ HeapTuple tuple;
+ List *result = NIL;
+
+ classRel = table_open(RelationRelationId, AccessShareLock);
+
+ ScanKeyInit(&key[0],
+ Anum_pg_class_relkind,
+ BTEqualStrategyNumber, F_CHAREQ,
+ CharGetDatum(RELKIND_SEQUENCE));
+
+ scan = table_beginscan_catalog(classRel, 1, key);
+
+ while ((tuple = heap_getnext(scan, ForwardScanDirection)) != NULL)
+ {
+ Form_pg_class relForm = (Form_pg_class) GETSTRUCT(tuple);
+ Oid relid = relForm->oid;
+
+ if (is_publishable_class(relid, relForm))
+ result = lappend_oid(result, relid);
+ }
+
+ table_endscan(scan);
+
+ table_close(classRel, AccessShareLock);
+ return result;
+}
+
/*
* Get publication using oid
*
@@ -757,10 +847,12 @@ GetPublication(Oid pubid)
pub->oid = pubid;
pub->name = pstrdup(NameStr(pubform->pubname));
pub->alltables = pubform->puballtables;
+ pub->allsequences = pubform->puballsequences;
pub->pubactions.pubinsert = pubform->pubinsert;
pub->pubactions.pubupdate = pubform->pubupdate;
pub->pubactions.pubdelete = pubform->pubdelete;
pub->pubactions.pubtruncate = pubform->pubtruncate;
+ pub->pubactions.pubsequence = pubform->pubsequence;
pub->pubviaroot = pubform->pubviaroot;
ReleaseSysCache(tup);
@@ -915,3 +1007,56 @@ pg_get_publication_tables(PG_FUNCTION_ARGS)
SRF_RETURN_DONE(funcctx);
}
+
+/*
+ * Returns Oids of sequences in a publication.
+ */
+Datum
+pg_get_publication_sequences(PG_FUNCTION_ARGS)
+{
+ FuncCallContext *funcctx;
+ char *pubname = text_to_cstring(PG_GETARG_TEXT_PP(0));
+ Publication *publication;
+ List *sequences;
+
+ /* stuff done only on the first call of the function */
+ if (SRF_IS_FIRSTCALL())
+ {
+ MemoryContext oldcontext;
+
+ /* create a function context for cross-call persistence */
+ funcctx = SRF_FIRSTCALL_INIT();
+
+ /* switch to memory context appropriate for multiple function calls */
+ oldcontext = MemoryContextSwitchTo(funcctx->multi_call_memory_ctx);
+
+ publication = GetPublicationByName(pubname, false);
+
+ /*
+ * Publications support partitioned tables, although all changes are
+ * replicated using leaf partition identity and schema, so we only
+ * need those.
+ */
+ if (publication->allsequences)
+ sequences = GetAllSequencesPublicationRelations();
+ else
+ sequences = GetPublicationSequenceRelations(publication->oid);
+
+ funcctx->user_fctx = (void *) sequences;
+
+ MemoryContextSwitchTo(oldcontext);
+ }
+
+ /* stuff done on every call of the function */
+ funcctx = SRF_PERCALL_SETUP();
+ sequences = (List *) funcctx->user_fctx;
+
+ if (funcctx->call_cntr < list_length(sequences))
+ {
+ Oid relid = list_nth_oid(sequences, funcctx->call_cntr);
+
+ SRF_RETURN_NEXT(funcctx, ObjectIdGetDatum(relid));
+ }
+
+ SRF_RETURN_DONE(funcctx);
+}
diff --git a/src/backend/catalog/system_views.sql b/src/backend/catalog/system_views.sql
index eb560955cd..2c12d373c1 100644
--- a/src/backend/catalog/system_views.sql
+++ b/src/backend/catalog/system_views.sql
@@ -372,6 +372,16 @@ CREATE VIEW pg_publication_tables AS
pg_class C JOIN pg_namespace N ON (N.oid = C.relnamespace)
WHERE C.oid = GPT.relid;
+CREATE VIEW pg_publication_sequences AS
+ SELECT
+ P.pubname AS pubname,
+ N.nspname AS schemaname,
+ C.relname AS sequencename
+ FROM pg_publication P,
+ LATERAL pg_get_publication_sequences(P.pubname) GPT,
+ pg_class C JOIN pg_namespace N ON (N.oid = C.relnamespace)
+ WHERE C.oid = GPT.relid;
+
CREATE VIEW pg_locks AS
SELECT * FROM pg_lock_status() AS L;
diff --git a/src/backend/commands/publicationcmds.c b/src/backend/commands/publicationcmds.c
index 7d4a0e95f6..e8dfb1a068 100644
--- a/src/backend/commands/publicationcmds.c
+++ b/src/backend/commands/publicationcmds.c
@@ -16,6 +16,7 @@
#include "access/genam.h"
#include "access/htup_details.h"
+#include "access/relation.h"
#include "access/table.h"
#include "access/xact.h"
#include "catalog/catalog.h"
@@ -59,6 +60,12 @@ static void PublicationAddSchemas(Oid pubid, List *schemas, bool if_not_exists,
AlterPublicationStmt *stmt);
static void PublicationDropSchemas(Oid pubid, List *schemas, bool missing_ok);
+static List *OpenSequenceList(List *sequences);
+static void CloseSequenceList(List *rels);
+static void PublicationAddSequences(Oid pubid, List *rels, bool if_not_exists,
+ AlterPublicationStmt *stmt);
+static void PublicationDropSequences(Oid pubid, List *rels, bool missing_ok);
+
static void
parse_publication_options(ParseState *pstate,
List *options,
@@ -77,6 +84,7 @@ parse_publication_options(ParseState *pstate,
pubactions->pubupdate = true;
pubactions->pubdelete = true;
pubactions->pubtruncate = true;
+ pubactions->pubsequence = true;
*publish_via_partition_root = false;
/* Parse options */
@@ -101,6 +109,7 @@ parse_publication_options(ParseState *pstate,
pubactions->pubupdate = false;
pubactions->pubdelete = false;
pubactions->pubtruncate = false;
+ pubactions->pubsequence = false;
*publish_given = true;
publish = defGetString(defel);
@@ -123,6 +132,8 @@ parse_publication_options(ParseState *pstate,
pubactions->pubdelete = true;
else if (strcmp(publish_opt, "truncate") == 0)
pubactions->pubtruncate = true;
+ else if (strcmp(publish_opt, "sequence") == 0)
+ pubactions->pubsequence = true;
else
ereport(ERROR,
(errcode(ERRCODE_SYNTAX_ERROR),
@@ -149,7 +160,9 @@ parse_publication_options(ParseState *pstate,
*/
static void
ObjectsInPublicationToOids(List *pubobjspec_list, ParseState *pstate,
- List **rels, List **schemas)
+ List **tables, List **sequences,
+ List **tables_schemas, List **sequences_schemas,
+ List **schemas)
{
ListCell *cell;
PublicationObjSpec *pubobj;
@@ -167,12 +180,23 @@ ObjectsInPublicationToOids(List *pubobjspec_list, ParseState *pstate,
switch (pubobj->pubobjtype)
{
case PUBLICATIONOBJ_TABLE:
- *rels = lappend(*rels, pubobj->pubtable);
+ *tables = lappend(*tables, pubobj->pubtable);
+ break;
+ case PUBLICATIONOBJ_SEQUENCE:
+ *sequences = lappend(*sequences, pubobj->pubtable);
break;
case PUBLICATIONOBJ_TABLE_IN_SCHEMA:
schemaid = get_namespace_oid(pubobj->name, false);
/* Filter out duplicates if user specifies "sch1, sch1" */
+ *tables_schemas = list_append_unique_oid(*tables_schemas, schemaid);
+ *schemas = list_append_unique_oid(*schemas, schemaid);
+ break;
+ case PUBLICATIONOBJ_SEQUENCE_IN_SCHEMA:
+ schemaid = get_namespace_oid(pubobj->name, false);
+
+ /* Filter out duplicates if user specifies "sch1, sch1" */
+ *sequences_schemas = list_append_unique_oid(*sequences_schemas, schemaid);
*schemas = list_append_unique_oid(*schemas, schemaid);
break;
case PUBLICATIONOBJ_TABLE_IN_CUR_SCHEMA:
@@ -186,6 +210,21 @@ ObjectsInPublicationToOids(List *pubobjspec_list, ParseState *pstate,
list_free(search_path);
/* Filter out duplicates if user specifies "sch1, sch1" */
+ *tables_schemas = list_append_unique_oid(*tables_schemas, schemaid);
+ *schemas = list_append_unique_oid(*schemas, schemaid);
+ break;
+ case PUBLICATIONOBJ_SEQUENCE_IN_CUR_SCHEMA:
+ search_path = fetch_search_path(false);
+ if (search_path == NIL) /* nothing valid in search_path? */
+ ereport(ERROR,
+ errcode(ERRCODE_UNDEFINED_SCHEMA),
+ errmsg("no schema has been selected for CURRENT_SCHEMA"));
+
+ schemaid = linitial_oid(search_path);
+ list_free(search_path);
+
+ /* Filter out duplicates if user specifies "sch1, sch1" */
+ *sequences_schemas = list_append_unique_oid(*sequences_schemas, schemaid);
*schemas = list_append_unique_oid(*schemas, schemaid);
break;
default:
@@ -251,7 +290,10 @@ CreatePublication(ParseState *pstate, CreatePublicationStmt *stmt)
bool publish_via_partition_root_given;
bool publish_via_partition_root;
AclResult aclresult;
- List *relations = NIL;
+ List *tables = NIL;
+ List *sequences = NIL;
+ List *tables_schemaidlist = NIL;
+ List *sequences_schemaidlist = NIL;
List *schemaidlist = NIL;
/* must have CREATE privilege on database */
@@ -306,6 +348,8 @@ CreatePublication(ParseState *pstate, CreatePublicationStmt *stmt)
BoolGetDatum(pubactions.pubdelete);
values[Anum_pg_publication_pubtruncate - 1] =
BoolGetDatum(pubactions.pubtruncate);
+ values[Anum_pg_publication_pubsequence - 1] =
+ BoolGetDatum(pubactions.pubsequence);
values[Anum_pg_publication_pubviaroot - 1] =
BoolGetDatum(publish_via_partition_root);
@@ -330,26 +374,40 @@ CreatePublication(ParseState *pstate, CreatePublicationStmt *stmt)
}
else
{
- ObjectsInPublicationToOids(stmt->pubobjects, pstate, &relations,
+ ObjectsInPublicationToOids(stmt->pubobjects, pstate,
+ &tables, &sequences,
+ &tables_schemaidlist,
+ &sequences_schemaidlist,
&schemaidlist);
/* FOR ALL TABLES IN SCHEMA requires superuser */
- if (list_length(schemaidlist) > 0 && !superuser())
+ if (list_length(tables_schemaidlist) > 0 && !superuser())
ereport(ERROR,
errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
errmsg("must be superuser to create FOR ALL TABLES IN SCHEMA publication"));
- if (list_length(relations) > 0)
+ if (list_length(tables) > 0)
{
List *rels;
- rels = OpenTableList(relations);
- CheckObjSchemaNotAlreadyInPublication(rels, schemaidlist,
+ rels = OpenTableList(tables);
+ CheckObjSchemaNotAlreadyInPublication(rels, tables_schemaidlist,
PUBLICATIONOBJ_TABLE);
PublicationAddTables(puboid, rels, true, NULL);
CloseTableList(rels);
}
+ if (list_length(sequences) > 0)
+ {
+ List *rels;
+
+ rels = OpenSequenceList(sequences);
+ CheckObjSchemaNotAlreadyInPublication(rels, sequences_schemaidlist,
+ PUBLICATIONOBJ_SEQUENCE);
+ PublicationAddTables(puboid, rels, true, NULL);
+ CloseSequenceList(rels);
+ }
+
if (list_length(schemaidlist) > 0)
{
/*
@@ -651,12 +709,13 @@ AlterPublicationSchemas(AlterPublicationStmt *stmt,
*/
static void
CheckAlterPublication(AlterPublicationStmt *stmt, HeapTuple tup,
- List *tables, List *schemaidlist)
+ List *tables, List *tables_schemaidlist,
+ List *sequences, List *sequences_schemaidlist)
{
Form_pg_publication pubform = (Form_pg_publication) GETSTRUCT(tup);
if ((stmt->action == DEFELEM_ADD || stmt->action == DEFELEM_SET) &&
- schemaidlist && !superuser())
+ (tables_schemaidlist || sequences_schemaidlist) && !superuser())
ereport(ERROR,
(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
errmsg("must be superuser to add or set schemas")));
@@ -665,13 +724,24 @@ CheckAlterPublication(AlterPublicationStmt *stmt, HeapTuple tup,
* Check that user is allowed to manipulate the publication tables in
* schema
*/
- if (schemaidlist && pubform->puballtables)
+ if (tables_schemaidlist && pubform->puballtables)
ereport(ERROR,
(errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
errmsg("publication \"%s\" is defined as FOR ALL TABLES",
NameStr(pubform->pubname)),
errdetail("Tables from schema cannot be added to, dropped from, or set on FOR ALL TABLES publications.")));
+ /*
+ * Check that user is allowed to manipulate the publication sequences in
+ * schema
+ */
+ if (sequences_schemaidlist && pubform->puballsequences)
+ ereport(ERROR,
+ (errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
+ errmsg("publication \"%s\" is defined as FOR ALL SEQUENCES",
+ NameStr(pubform->pubname)),
+ errdetail("Sequences from schema cannot be added to, dropped from, or set on FOR ALL SEQUENCES publications.")));
+
/* Check that user is allowed to manipulate the publication tables. */
if (tables && pubform->puballtables)
ereport(ERROR,
@@ -679,6 +749,108 @@ CheckAlterPublication(AlterPublicationStmt *stmt, HeapTuple tup,
errmsg("publication \"%s\" is defined as FOR ALL TABLES",
NameStr(pubform->pubname)),
errdetail("Tables cannot be added to or dropped from FOR ALL TABLES publications.")));
+
+ /* Check that user is allowed to manipulate the publication tables. */
+ if (sequences && pubform->puballsequences)
+ ereport(ERROR,
+ (errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
+ errmsg("publication \"%s\" is defined as FOR ALL SEQUENCES",
+ NameStr(pubform->pubname)),
+ errdetail("Sequences cannot be added to or dropped from FOR ALL SEQUENCES publications.")));
+}
+
+/*
+ * Add or remove sequence to/from publication.
+ */
+static void
+AlterPublicationSequences(AlterPublicationStmt *stmt, HeapTuple tup,
+ List *sequences, List *schemaidlist)
+{
+ List *rels = NIL;
+ Form_pg_publication pubform = (Form_pg_publication) GETSTRUCT(tup);
+ Oid pubid = pubform->oid;
+
+ /*
+ * It is quite possible that for the SET case user has not specified any
+ * tables in which case we need to remove all the existing tables.
+ */
+ if (!sequences && stmt->action != DEFELEM_SET)
+ return;
+
+ rels = OpenSequenceList(sequences);
+
+ if (stmt->action == DEFELEM_ADD)
+ {
+ List *schemas = NIL;
+
+ /*
+ * Check if the relation is member of the existing schema in the
+ * publication or member of the schema list specified.
+ */
+ schemas = list_concat_copy(schemaidlist, GetPublicationSchemas(pubid));
+ CheckObjSchemaNotAlreadyInPublication(rels, schemas,
+ PUBLICATIONOBJ_SEQUENCE);
+ PublicationAddSequences(pubid, rels, false, stmt);
+ }
+ else if (stmt->action == DEFELEM_DROP)
+ PublicationDropSequences(pubid, rels, false);
+ else /* DEFELEM_SET */
+ {
+ List *oldrelids = GetPublicationRelations(pubid,
+ PUBLICATION_PART_ROOT);
+ List *delrels = NIL;
+ ListCell *oldlc;
+
+ CheckObjSchemaNotAlreadyInPublication(rels, schemaidlist,
+ PUBLICATIONOBJ_SEQUENCE);
+
+ /* Calculate which relations to drop. */
+ foreach(oldlc, oldrelids)
+ {
+ Oid oldrelid = lfirst_oid(oldlc);
+ ListCell *newlc;
+ bool found = false;
+
+ foreach(newlc, rels)
+ {
+ PublicationRelInfo *newpubrel;
+
+ newpubrel = (PublicationRelInfo *) lfirst(newlc);
+ if (RelationGetRelid(newpubrel->relation) == oldrelid)
+ {
+ found = true;
+ break;
+ }
+ }
+ /* Not yet in the list, open it and add to the list */
+ if (!found)
+ {
+ Relation oldrel;
+ PublicationRelInfo *pubrel;
+
+ /* Wrap relation into PublicationRelInfo */
+ oldrel = table_open(oldrelid, ShareUpdateExclusiveLock);
+
+ pubrel = palloc(sizeof(PublicationRelInfo));
+ pubrel->relation = oldrel;
+
+ delrels = lappend(delrels, pubrel);
+ }
+ }
+
+ /* And drop them. */
+ PublicationDropSequences(pubid, delrels, true);
+
+ /*
+ * Don't bother calculating the difference for adding, we'll catch and
+ * skip existing ones when doing catalog update.
+ */
+ PublicationAddSequences(pubid, rels, true, stmt);
+
+ CloseSequenceList(delrels);
+ }
+
+ CloseSequenceList(rels);
}
/*
@@ -716,13 +888,21 @@ AlterPublication(ParseState *pstate, AlterPublicationStmt *stmt)
AlterPublicationOptions(pstate, stmt, rel, tup);
else
{
- List *relations = NIL;
+ List *tables = NIL;
+ List *sequences = NIL;
+ List *tables_schemaidlist = NIL;
+ List *sequences_schemaidlist = NIL;
List *schemaidlist = NIL;
- ObjectsInPublicationToOids(stmt->pubobjects, pstate, &relations,
+ ObjectsInPublicationToOids(stmt->pubobjects, pstate,
+ &tables, &sequences,
+ &tables_schemaidlist,
+ &sequences_schemaidlist,
&schemaidlist);
- CheckAlterPublication(stmt, tup, relations, schemaidlist);
+ CheckAlterPublication(stmt, tup,
+ tables, tables_schemaidlist,
+ sequences, sequences_schemaidlist);
/*
* Lock the publication so nobody else can do anything with it. This
@@ -747,7 +927,9 @@ AlterPublication(ParseState *pstate, AlterPublicationStmt *stmt)
errmsg("publication \"%s\" does not exist",
stmt->pubname));
- AlterPublicationTables(stmt, tup, relations, schemaidlist);
+ AlterPublicationTables(stmt, tup, tables, tables_schemaidlist);
+ AlterPublicationSequences(stmt, tup, sequences, sequences_schemaidlist);
+
AlterPublicationSchemas(stmt, tup, schemaidlist);
}
@@ -1155,6 +1337,144 @@ PublicationDropSchemas(Oid pubid, List *schemas, bool missing_ok)
}
}
+/*
+ * Open relations specified by a PublicationTable list.
+ * In the returned list of PublicationRelInfo, tables are locked
+ * in ShareUpdateExclusiveLock mode in order to add them to a publication.
+ */
+static List *
+OpenSequenceList(List *sequences)
+{
+ List *relids = NIL;
+ List *rels = NIL;
+ ListCell *lc;
+
+ /*
+ * Open, share-lock, and check all the explicitly-specified relations
+ */
+ foreach(lc, sequences)
+ {
+ PublicationTable *s = lfirst_node(PublicationTable, lc);
+ Relation rel;
+ Oid myrelid;
+ PublicationRelInfo *pub_rel;
+
+ /* Allow query cancel in case this takes a long time */
+ CHECK_FOR_INTERRUPTS();
+
+ rel = table_openrv(s->relation, ShareUpdateExclusiveLock);
+ myrelid = RelationGetRelid(rel);
+
+ /*
+ * Filter out duplicates if user specifies "foo, foo".
+ *
+ * Note that this algorithm is known to not be very efficient (O(N^2))
+ * but given that it only works on list of tables given to us by user
+ * it's deemed acceptable.
+ */
+ if (list_member_oid(relids, myrelid))
+ {
+ table_close(rel, ShareUpdateExclusiveLock);
+ continue;
+ }
+
+ pub_rel = palloc(sizeof(PublicationRelInfo));
+ pub_rel->relation = rel;
+ rels = lappend(rels, pub_rel);
+ relids = lappend_oid(relids, myrelid);
+ }
+
+ list_free(relids);
+
+ return rels;
+}
+
+/*
+ * Close all relations in the list.
+ */
+static void
+CloseSequenceList(List *rels)
+{
+ ListCell *lc;
+
+ foreach(lc, rels)
+ {
+ PublicationRelInfo *pub_rel;
+
+ pub_rel = (PublicationRelInfo *) lfirst(lc);
+ table_close(pub_rel->relation, NoLock);
+ }
+}
+
+/*
+ * Add listed tables to the publication.
+ */
+static void
+PublicationAddSequences(Oid pubid, List *rels, bool if_not_exists,
+ AlterPublicationStmt *stmt)
+{
+ ListCell *lc;
+
+ Assert(!stmt || !stmt->for_all_sequences);
+
+ foreach(lc, rels)
+ {
+ PublicationRelInfo *pub_rel = (PublicationRelInfo *) lfirst(lc);
+ Relation rel = pub_rel->relation;
+ ObjectAddress obj;
+
+ /* Must be owner of the sequence or superuser. */
+ if (!pg_class_ownercheck(RelationGetRelid(rel), GetUserId()))
+ aclcheck_error(ACLCHECK_NOT_OWNER, get_relkind_objtype(rel->rd_rel->relkind),
+ RelationGetRelationName(rel));
+
+ obj = publication_add_relation(pubid, pub_rel, if_not_exists);
+ if (stmt)
+ {
+ EventTriggerCollectSimpleCommand(obj, InvalidObjectAddress,
+ (Node *) stmt);
+
+ InvokeObjectPostCreateHook(PublicationRelRelationId,
+ obj.objectId, 0);
+ }
+ }
+}
+
+/*
+ * Remove listed sequences from the publication.
+ */
+static void
+PublicationDropSequences(Oid pubid, List *rels, bool missing_ok)
+{
+ ObjectAddress obj;
+ ListCell *lc;
+ Oid prid;
+
+ foreach(lc, rels)
+ {
+ Relation rel = (Relation) lfirst(lc);
+ Oid relid = RelationGetRelid(rel);
+
+ prid = GetSysCacheOid2(PUBLICATIONRELMAP, Anum_pg_publication_rel_oid,
+ ObjectIdGetDatum(relid),
+ ObjectIdGetDatum(pubid));
+ if (!OidIsValid(prid))
+ {
+ if (missing_ok)
+ continue;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_UNDEFINED_OBJECT),
+ errmsg("relation \"%s\" is not part of the publication",
+ RelationGetRelationName(rel))));
+ }
+
+ ObjectAddressSet(obj, PublicationRelRelationId, prid);
+ performDeletion(&obj, DROP_CASCADE, 0);
+ }
+}
+
+
/*
* Internal workhorse for changing a publication owner
*/
diff --git a/src/backend/commands/sequence.c b/src/backend/commands/sequence.c
index e45e4042a7..c3e1825056 100644
--- a/src/backend/commands/sequence.c
+++ b/src/backend/commands/sequence.c
@@ -374,6 +374,85 @@ ResetSequence(Oid seq_relid)
relation_close(seq_rel, NoLock);
}
+/*
+ * Reset a sequence to its initial value.
+ *
+ * The change is made transactionally, so that on failure of the current
+ * transaction, the sequence will be restored to its previous state.
+ * We do that by creating a whole new relfilenode for the sequence; so this
+ * works much like the rewriting forms of ALTER TABLE.
+ *
+ * Caller is assumed to have acquired AccessExclusiveLock on the sequence,
+ * which must not be released until end of transaction. Caller is also
+ * responsible for permissions checking.
+ */
+void
+ResetSequence2(Oid seq_relid, int64 last_value, int64 log_cnt, bool is_called)
+{
+ Relation seq_rel;
+ SeqTable elm;
+ Form_pg_sequence_data seq;
+ Buffer buf;
+ HeapTupleData seqdatatuple;
+ HeapTuple tuple;
+
+ /*
+ * Read the old sequence. This does a bit more work than really
+ * necessary, but it's simple, and we do want to double-check that it's
+ * indeed a sequence.
+ */
+ init_sequence(seq_relid, &elm, &seq_rel);
+ (void) read_seq_tuple(seq_rel, &buf, &seqdatatuple);
+
+ /*
+ * Copy the existing sequence tuple.
+ */
+ tuple = heap_copytuple(&seqdatatuple);
+
+ /* Now we're done with the old page */
+ UnlockReleaseBuffer(buf);
+
+ /*
+ * Modify the copied tuple to execute the restart (compare the RESTART
+ * action in AlterSequence)
+ */
+ seq = (Form_pg_sequence_data) GETSTRUCT(tuple);
+ seq->last_value = last_value;
+ seq->is_called = is_called;
+ seq->log_cnt = log_cnt;
+
+ /*
+ * Create a new storage file for the sequence.
+ */
+ RelationSetNewRelfilenode(seq_rel, seq_rel->rd_rel->relpersistence);
+
+ /*
+ * Ensure sequence's relfrozenxid is at 0, since it won't contain any
+ * unfrozen XIDs. Same with relminmxid, since a sequence will never
+ * contain multixacts.
+ */
+ Assert(seq_rel->rd_rel->relfrozenxid == InvalidTransactionId);
+ Assert(seq_rel->rd_rel->relminmxid == InvalidMultiXactId);
+
+ /*
+ * Insert the modified tuple into the new storage file.
+ *
+ * XXX Maybe this should also use created=true, just like the other places
+ * calling fill_seq_with_data. That's probably needed for correct cascading
+ * replication.
+ *
+ * XXX That'd mean all fill_seq_with_data callers use created=true, making
+ * the parameter unnecessary.
+ */
+ fill_seq_with_data(seq_rel, tuple);
+
+ /* Clear local cache so that we don't think we have cached numbers */
+ /* Note that we do not change the currval() state */
+ elm->cached = elm->last;
+
+ relation_close(seq_rel, NoLock);
+}
+
/*
* Initialize a sequence's relation with the specified tuple as content
*/
diff --git a/src/backend/commands/subscriptioncmds.c b/src/backend/commands/subscriptioncmds.c
index c47ba26369..5086526904 100644
--- a/src/backend/commands/subscriptioncmds.c
+++ b/src/backend/commands/subscriptioncmds.c
@@ -84,6 +84,7 @@ typedef struct SubOpts
} SubOpts;
static List *fetch_table_list(WalReceiverConn *wrconn, List *publications);
+static List *fetch_sequence_list(WalReceiverConn *wrconn, List *publications);
static void check_duplicates_in_publist(List *publist, Datum *datums);
static List *merge_publications(List *oldpublist, List *newpublist, bool addpub, const char *subname);
static void ReportSlotConnectionError(List *rstates, Oid subid, char *slotname, char *err);
@@ -502,6 +503,7 @@ CreateSubscription(ParseState *pstate, CreateSubscriptionStmt *stmt,
char *err;
WalReceiverConn *wrconn;
List *tables;
+ List *sequences;
ListCell *lc;
char table_state;
@@ -540,6 +542,26 @@ CreateSubscription(ParseState *pstate, CreateSubscriptionStmt *stmt,
InvalidXLogRecPtr);
}
+ /*
+ * Get the sequence list from publisher and build local sequence
+ * status info.
+ */
+ sequences = fetch_sequence_list(wrconn, publications);
+ foreach(lc, sequences)
+ {
+ RangeVar *rv = (RangeVar *) lfirst(lc);
+ Oid relid;
+
+ relid = RangeVarGetRelid(rv, AccessShareLock, false);
+
+ /* Check for supported relkind. */
+ CheckSubscriptionRelkind(get_rel_relkind(relid),
+ rv->schemaname, rv->relname);
+
+ AddSubscriptionRelState(subid, relid, table_state,
+ InvalidXLogRecPtr);
+ }
+
/*
* If requested, create permanent slot for the subscription. We
* won't use the initial snapshot for anything, so no need to
@@ -712,6 +734,10 @@ AlterSubscription_refresh(Subscription *sub, bool copy_data)
{
Oid relid = subrel_local_oids[off];
+ /* XXX ignore sequences - maybe do this in GetSubscriptionRelations? */
+ if (get_rel_relkind(relid) == RELKIND_SEQUENCE)
+ continue;
+
if (!bsearch(&relid, pubrel_local_oids,
list_length(pubrel_names), sizeof(Oid), oid_cmp))
{
@@ -803,6 +829,183 @@ AlterSubscription_refresh(Subscription *sub, bool copy_data)
ReplicationSlotDropAtPubNode(wrconn, syncslotname, true);
}
}
+
+ /*
+ * XXX now do the same thing for sequences, maybe before the preceding
+ * block, or earlier?
+ */
+
+ /* Get the table list from publisher. */
+ pubrel_names = fetch_sequence_list(wrconn, sub->publications);
+
+ /* Get local table list. */
+ subrel_states = GetSubscriptionRelations(sub->oid);
+
+ /*
+ * Build qsorted array of local table oids for faster lookup. This can
+ * potentially contain all tables in the database so speed of lookup
+ * is important.
+ */
+ subrel_local_oids = palloc(list_length(subrel_states) * sizeof(Oid));
+ off = 0;
+ foreach(lc, subrel_states)
+ {
+ SubscriptionRelState *relstate = (SubscriptionRelState *) lfirst(lc);
+
+ subrel_local_oids[off++] = relstate->relid;
+ }
+ qsort(subrel_local_oids, list_length(subrel_states),
+ sizeof(Oid), oid_cmp);
+
+ /*
+ * Rels that we want to remove from subscription and drop any slots
+ * and origins corresponding to them.
+ */
+ sub_remove_rels = palloc(list_length(subrel_states) * sizeof(SubRemoveRels));
+
+ /*
+ * Walk over the remote tables and try to match them to locally known
+ * tables. If the table is not known locally create a new state for
+ * it.
+ *
+ * Also builds array of local oids of remote tables for the next step.
+ */
+ off = 0;
+ pubrel_local_oids = palloc(list_length(pubrel_names) * sizeof(Oid));
+
+ foreach(lc, pubrel_names)
+ {
+ RangeVar *rv = (RangeVar *) lfirst(lc);
+ Oid relid;
+
+ relid = RangeVarGetRelid(rv, AccessShareLock, false);
+
+ /* Check for supported relkind. */
+ CheckSubscriptionRelkind(get_rel_relkind(relid),
+ rv->schemaname, rv->relname);
+
+ pubrel_local_oids[off++] = relid;
+
+ if (!bsearch(&relid, subrel_local_oids,
+ list_length(subrel_states), sizeof(Oid), oid_cmp))
+ {
+ AddSubscriptionRelState(sub->oid, relid,
+ copy_data ? SUBREL_STATE_INIT : SUBREL_STATE_READY,
+ InvalidXLogRecPtr);
+ ereport(DEBUG1,
+ (errmsg_internal("table \"%s.%s\" added to subscription \"%s\"",
+ rv->schemaname, rv->relname, sub->name)));
+ }
+ }
+
+ /*
+ * Next remove state for tables we should not care about anymore using
+ * the data we collected above
+ */
+ qsort(pubrel_local_oids, list_length(pubrel_names),
+ sizeof(Oid), oid_cmp);
+
+ remove_rel_len = 0;
+ for (off = 0; off < list_length(subrel_states); off++)
+ {
+ Oid relid = subrel_local_oids[off];
+
+ /* XXX ignore non-sequences - maybe do this in GetSubscriptionRelations? */
+ if (get_rel_relkind(relid) != RELKIND_SEQUENCE)
+ continue;
+
+ if (!bsearch(&relid, pubrel_local_oids,
+ list_length(pubrel_names), sizeof(Oid), oid_cmp))
+ {
+ char state;
+ XLogRecPtr statelsn;
+
+ /*
+ * Lock pg_subscription_rel with AccessExclusiveLock to
+ * prevent any race conditions with the apply worker
+ * re-launching workers at the same time this code is trying
+ * to remove those tables.
+ *
+ * Even if new worker for this particular rel is restarted it
+ * won't be able to make any progress as we hold exclusive
+ * lock on subscription_rel till the transaction end. It will
+ * simply exit as there is no corresponding rel entry.
+ *
+ * This locking also ensures that the state of rels won't
+ * change till we are done with this refresh operation.
+ */
+ if (!rel)
+ rel = table_open(SubscriptionRelRelationId, AccessExclusiveLock);
+
+ /* Last known rel state. */
+ state = GetSubscriptionRelState(sub->oid, relid, &statelsn);
+
+ sub_remove_rels[remove_rel_len].relid = relid;
+ sub_remove_rels[remove_rel_len++].state = state;
+
+ RemoveSubscriptionRel(sub->oid, relid);
+
+ logicalrep_worker_stop(sub->oid, relid);
+
+ /*
+ * For READY state, we would have already dropped the
+ * tablesync origin.
+ */
+ if (state != SUBREL_STATE_READY)
+ {
+ char originname[NAMEDATALEN];
+
+ /*
+ * Drop the tablesync's origin tracking if exists.
+ *
+ * It is possible that the origin is not yet created for
+ * tablesync worker, this can happen for the states before
+ * SUBREL_STATE_FINISHEDCOPY. The apply worker can also
+ * concurrently try to drop the origin and by this time
+ * the origin might be already removed. For these reasons,
+ * passing missing_ok = true.
+ */
+ ReplicationOriginNameForTablesync(sub->oid, relid, originname,
+ sizeof(originname));
+ replorigin_drop_by_name(originname, true, false);
+ }
+
+ ereport(DEBUG1,
+ (errmsg_internal("table \"%s.%s\" removed from subscription \"%s\"",
+ get_namespace_name(get_rel_namespace(relid)),
+ get_rel_name(relid),
+ sub->name)));
+ }
+ }
+
+ /*
+ * Drop the tablesync slots associated with removed tables. This has
+ * to be at the end because otherwise if there is an error while doing
+ * the database operations we won't be able to rollback dropped slots.
+ */
+ for (off = 0; off < remove_rel_len; off++)
+ {
+ if (sub_remove_rels[off].state != SUBREL_STATE_READY &&
+ sub_remove_rels[off].state != SUBREL_STATE_SYNCDONE)
+ {
+ char syncslotname[NAMEDATALEN] = {0};
+
+ /*
+ * For READY/SYNCDONE states we know the tablesync slot has
+ * already been dropped by the tablesync worker.
+ *
+ * For other states, there is no certainty, maybe the slot
+ * does not exist yet. Also, if we fail after removing some of
+ * the slots, next time, it will again try to drop already
+ * dropped slots and fail. For these reasons, we allow
+ * missing_ok = true for the drop.
+ */
+ ReplicationSlotNameForTablesync(sub->oid, sub_remove_rels[off].relid,
+ syncslotname, sizeof(syncslotname));
+ ReplicationSlotDropAtPubNode(wrconn, syncslotname, true);
+ }
+ }
+
}
PG_FINALLY();
{
@@ -1607,6 +1810,75 @@ fetch_table_list(WalReceiverConn *wrconn, List *publications)
return tablelist;
}
+/*
+ * Get the list of sequences which belong to specified publications on the
+ * publisher connection.
+ */
+static List *
+fetch_sequence_list(WalReceiverConn *wrconn, List *publications)
+{
+ WalRcvExecResult *res;
+ StringInfoData cmd;
+ TupleTableSlot *slot;
+ Oid tableRow[2] = {TEXTOID, TEXTOID};
+ ListCell *lc;
+ bool first;
+ List *tablelist = NIL;
+
+ Assert(list_length(publications) > 0);
+
+ initStringInfo(&cmd);
+ appendStringInfoString(&cmd, "SELECT DISTINCT s.schemaname, s.sequencename\n"
+ " FROM pg_catalog.pg_publication_sequences s\n"
+ " WHERE s.pubname IN (");
+ first = true;
+ foreach(lc, publications)
+ {
+ char *pubname = strVal(lfirst(lc));
+
+ if (first)
+ first = false;
+ else
+ appendStringInfoString(&cmd, ", ");
+
+ appendStringInfoString(&cmd, quote_literal_cstr(pubname));
+ }
+ appendStringInfoChar(&cmd, ')');
+
+ res = walrcv_exec(wrconn, cmd.data, 2, tableRow);
+ pfree(cmd.data);
+
+ if (res->status != WALRCV_OK_TUPLES)
+ ereport(ERROR,
+ (errmsg("could not receive list of replicated tables from the publisher: %s",
+ res->err)));
+
+ /* Process tables. */
+ slot = MakeSingleTupleTableSlot(res->tupledesc, &TTSOpsMinimalTuple);
+ while (tuplestore_gettupleslot(res->tuplestore, true, false, slot))
+ {
+ char *nspname;
+ char *relname;
+ bool isnull;
+ RangeVar *rv;
+
+ nspname = TextDatumGetCString(slot_getattr(slot, 1, &isnull));
+ Assert(!isnull);
+ relname = TextDatumGetCString(slot_getattr(slot, 2, &isnull));
+ Assert(!isnull);
+
+ rv = makeRangeVar(nspname, relname, -1);
+ tablelist = lappend(tablelist, rv);
+
+ ExecClearTuple(slot);
+ }
+ ExecDropSingleTupleTableSlot(slot);
+
+ walrcv_clear_result(res);
+
+ return tablelist;
+}
+
/*
* This is to report the connection failure while dropping replication slots.
* Here, we report the WARNING for all tablesync slots so that user can drop
diff --git a/src/backend/executor/execReplication.c b/src/backend/executor/execReplication.c
index 574d7d27fd..6f0d42d551 100644
--- a/src/backend/executor/execReplication.c
+++ b/src/backend/executor/execReplication.c
@@ -608,7 +608,7 @@ void
CheckSubscriptionRelkind(char relkind, const char *nspname,
const char *relname)
{
- if (relkind != RELKIND_RELATION && relkind != RELKIND_PARTITIONED_TABLE)
+ if (relkind != RELKIND_RELATION && relkind != RELKIND_PARTITIONED_TABLE && relkind != RELKIND_SEQUENCE)
ereport(ERROR,
(errcode(ERRCODE_WRONG_OBJECT_TYPE),
errmsg("cannot use relation \"%s.%s\" as logical replication target",
diff --git a/src/backend/nodes/copyfuncs.c b/src/backend/nodes/copyfuncs.c
index ad1ea2ff2f..3dc5eeed91 100644
--- a/src/backend/nodes/copyfuncs.c
+++ b/src/backend/nodes/copyfuncs.c
@@ -4843,6 +4843,7 @@ _copyCreatePublicationStmt(const CreatePublicationStmt *from)
COPY_NODE_FIELD(options);
COPY_NODE_FIELD(pubobjects);
COPY_SCALAR_FIELD(for_all_tables);
+ COPY_SCALAR_FIELD(for_all_sequences);
return newnode;
}
diff --git a/src/backend/nodes/equalfuncs.c b/src/backend/nodes/equalfuncs.c
index f537d3eb96..2317185daf 100644
--- a/src/backend/nodes/equalfuncs.c
+++ b/src/backend/nodes/equalfuncs.c
@@ -2336,6 +2336,7 @@ _equalAlterPublicationStmt(const AlterPublicationStmt *a,
COMPARE_NODE_FIELD(options);
COMPARE_NODE_FIELD(pubobjects);
COMPARE_SCALAR_FIELD(for_all_tables);
+ COMPARE_SCALAR_FIELD(for_all_sequences);
COMPARE_SCALAR_FIELD(action);
return true;
diff --git a/src/backend/parser/gram.y b/src/backend/parser/gram.y
index a6d0cefa6b..5c4e8bf3d7 100644
--- a/src/backend/parser/gram.y
+++ b/src/backend/parser/gram.y
@@ -9672,6 +9672,26 @@ PublicationObjSpec:
$$->pubobjtype = PUBLICATIONOBJ_TABLE_IN_CUR_SCHEMA;
$$->location = @5;
}
+ | SEQUENCE relation_expr
+ {
+ $$ = makeNode(PublicationObjSpec);
+ $$->pubobjtype = PUBLICATIONOBJ_SEQUENCE;
+ $$->pubtable = makeNode(PublicationTable);
+ $$->pubtable->relation = $2;
+ }
+ | ALL SEQUENCE IN_P SCHEMA ColId
+ {
+ $$ = makeNode(PublicationObjSpec);
+ $$->pubobjtype = PUBLICATIONOBJ_SEQUENCE_IN_SCHEMA;
+ $$->name = $5;
+ $$->location = @5;
+ }
+ | ALL SEQUENCES IN_P SCHEMA CURRENT_SCHEMA
+ {
+ $$ = makeNode(PublicationObjSpec);
+ $$->pubobjtype = PUBLICATIONOBJ_SEQUENCE_IN_CUR_SCHEMA;
+ $$->location = @5;
+ }
| ColId
{
$$ = makeNode(PublicationObjSpec);
@@ -10007,6 +10027,12 @@ UnlistenStmt:
}
;
+/*
+ * FIXME
+ *
+ * opt_publication_for_sequences and publication_for_sequences should be
+ * copies for sequences
+ */
/*****************************************************************************
*
@@ -10015,6 +10041,12 @@ UnlistenStmt:
* BEGIN / COMMIT / ROLLBACK
* (also older versions END / ABORT)
*
+ * ALTER PUBLICATION name ADD SEQUENCE sequence [, sequence2]
+ *
+ * ALTER PUBLICATION name DROP SEQUENCE sequence [, sequence2]
+ *
+ * ALTER PUBLICATION name SET SEQUENCE sequence [, sequence2]
+ *
*****************************************************************************/
TransactionStmt:
diff --git a/src/backend/replication/logical/proto.c b/src/backend/replication/logical/proto.c
index 9f5bf4b639..9f53b4c47b 100644
--- a/src/backend/replication/logical/proto.c
+++ b/src/backend/replication/logical/proto.c
@@ -647,6 +647,58 @@ logicalrep_write_message(StringInfo out, TransactionId xid, XLogRecPtr lsn,
pq_sendbytes(out, message, sz);
}
+/*
+ * Write SEQUENCE to stream
+ */
+void
+logicalrep_write_sequence(StringInfo out, Relation rel, TransactionId xid,
+ XLogRecPtr lsn, bool transactional, bool created,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ uint8 flags = 0;
+ char *relname;
+
+ pq_sendbyte(out, LOGICAL_REP_MSG_SEQUENCE);
+
+ /* transaction ID (if not valid, we're not streaming) */
+ if (TransactionIdIsValid(xid))
+ pq_sendint32(out, xid);
+
+ pq_sendint8(out, flags);
+ pq_sendint64(out, lsn);
+
+ logicalrep_write_namespace(out, RelationGetNamespace(rel));
+ relname = RelationGetRelationName(rel);
+ pq_sendstring(out, relname);
+
+ pq_sendint8(out, transactional);
+ pq_sendint8(out, created);
+ pq_sendint64(out, last_value);
+ pq_sendint64(out, log_cnt);
+ pq_sendint8(out, is_called);
+}
+
+/*
+ * Read SEQUENCE from the stream.
+ */
+void
+logicalrep_read_sequence(StringInfo in, LogicalRepSequence *seqdata)
+{
+ /* XXX skipping flags and lsn */
+ pq_getmsgint(in, 1);
+ pq_getmsgint64(in);
+
+ /* Read relation name from stream */
+ seqdata->nspname = pstrdup(logicalrep_read_namespace(in));
+ seqdata->seqname = pstrdup(pq_getmsgstring(in));
+
+ seqdata->transactional = pq_getmsgint(in, 1);
+ seqdata->created = pq_getmsgint(in, 1);
+ seqdata->last_value = pq_getmsgint64(in);
+ seqdata->log_cnt = pq_getmsgint64(in);
+ seqdata->is_called = pq_getmsgint(in, 1);
+}
+
/*
* Write relation description to the output stream.
*/
@@ -1203,6 +1255,8 @@ logicalrep_message_type(LogicalRepMsgType action)
return "STREAM ABORT";
case LOGICAL_REP_MSG_STREAM_PREPARE:
return "STREAM PREPARE";
+ case LOGICAL_REP_MSG_SEQUENCE:
+ return "SEQUENCE";
}
elog(ERROR, "invalid logical replication message type \"%c\"", action);
diff --git a/src/backend/replication/logical/tablesync.c b/src/backend/replication/logical/tablesync.c
index f07983a43c..24369a1522 100644
--- a/src/backend/replication/logical/tablesync.c
+++ b/src/backend/replication/logical/tablesync.c
@@ -100,6 +100,7 @@
#include "catalog/pg_subscription_rel.h"
#include "catalog/pg_type.h"
#include "commands/copy.h"
+#include "commands/sequence.h"
#include "miscadmin.h"
#include "parser/parse_relation.h"
#include "pgstat.h"
@@ -357,6 +358,12 @@ process_syncing_tables_for_sync(XLogRecPtr current_lsn)
*
* If the synchronization position is reached (SYNCDONE), then the table can
* be marked as READY and is no longer tracked.
+ *
+ * XXX This needs to handle sequences too - after AlterSubscription_refresh
+ * starts caring about sequences, GetSubscriptionNotReadyRelations won't
+ * return just tables, and we'll have to sync them here. Not sure it's worth
+ * creating a new "sync" worker per sequence, maybe we should just sync them
+ * in the current process (it's pretty light-weight).
*/
static void
process_syncing_tables_for_apply(XLogRecPtr current_lsn)
@@ -871,6 +878,99 @@ copy_table(Relation rel)
logicalrep_rel_close(relmapentry, NoLock);
}
+
+
+/*
+ * FIXME add comment
+ */
+static void
+fetch_sequence_data(char *nspname, char *relname,
+ int64 *last_value, int64 *log_cnt, bool *is_called)
+{
+ WalRcvExecResult *res;
+ StringInfoData cmd;
+ TupleTableSlot *slot;
+ Oid tableRow[3] = {INT8OID, INT8OID, BOOLOID};
+
+ initStringInfo(&cmd);
+ appendStringInfo(&cmd, "SELECT last_value, log_cnt, is_called\n"
+ " FROM %s", quote_qualified_identifier(nspname, relname));
+
+ res = walrcv_exec(LogRepWorkerWalRcvConn, cmd.data, 3, tableRow);
+ pfree(cmd.data);
+
+ if (res->status != WALRCV_OK_TUPLES)
+ ereport(ERROR,
+ (errmsg("could not receive list of replicated tables from the publisher: %s",
+ res->err)));
+
+ /* Process the sequence. */
+ slot = MakeSingleTupleTableSlot(res->tupledesc, &TTSOpsMinimalTuple);
+ while (tuplestore_gettupleslot(res->tuplestore, true, false, slot))
+ {
+ bool isnull;
+
+ *last_value = DatumGetInt64(slot_getattr(slot, 1, &isnull));
+ Assert(!isnull);
+
+ *log_cnt = DatumGetInt64(slot_getattr(slot, 2, &isnull));
+ Assert(!isnull);
+
+ *is_called = DatumGetBool(slot_getattr(slot, 3, &isnull));
+ Assert(!isnull);
+
+ ExecClearTuple(slot);
+ }
+ ExecDropSingleTupleTableSlot(slot);
+
+ walrcv_clear_result(res);
+}
+
+/*
+ * Copy existing data of a sequence from publisher.
+ *
+ * Caller is responsible for locking the local relation.
+ */
+static void
+copy_sequence(Relation rel)
+{
+ LogicalRepRelMapEntry *relmapentry;
+ LogicalRepRelation lrel;
+ StringInfoData cmd;
+ int64 last_value = 0,
+ log_cnt = 0;
+ bool is_called = 0;
+
+ /* Get the publisher relation info. */
+ fetch_remote_table_info(get_namespace_name(RelationGetNamespace(rel)),
+ RelationGetRelationName(rel), &lrel);
+
+ /* Put the relation into relmap. */
+ logicalrep_relmap_update(&lrel);
+
+ /* Map the publisher relation to local one. */
+ relmapentry = logicalrep_rel_open(lrel.remoteid, NoLock);
+ Assert(rel == relmapentry->localrel);
+
+ /* Start copy on the publisher. */
+ initStringInfo(&cmd);
+
+ Assert(lrel.relkind == RELKIND_SEQUENCE);
+
+ fetch_sequence_data(lrel.nspname, lrel.relname, &last_value, &log_cnt, &is_called);
+
+ elog(WARNING, "sequence %s info last_value %ld log_cnt %ld is_called %d",
+ quote_qualified_identifier(lrel.nspname, lrel.relname),
+ last_value, log_cnt, is_called);
+
+ ResetSequence2(RelationGetRelid(rel), last_value, log_cnt, is_called);
+
+ logicalrep_rel_close(relmapentry, NoLock);
+}
+
+
+
+
/*
* Determine the tablesync slot name.
*
@@ -1106,10 +1206,20 @@ LogicalRepSyncTableStart(XLogRecPtr *origin_startpos)
originname)));
}
- /* Now do the initial data copy */
- PushActiveSnapshot(GetTransactionSnapshot());
- copy_table(rel);
- PopActiveSnapshot();
+ if (get_rel_relkind(RelationGetRelid(rel)) == RELKIND_SEQUENCE)
+ {
+ /* Now do the initial sequence copy */
+ PushActiveSnapshot(GetTransactionSnapshot());
+ copy_sequence(rel);
+ PopActiveSnapshot();
+ }
+ else
+ {
+ /* Now do the initial data copy */
+ PushActiveSnapshot(GetTransactionSnapshot());
+ copy_table(rel);
+ PopActiveSnapshot();
+ }
res = walrcv_exec(LogRepWorkerWalRcvConn, "COMMIT", 0, NULL);
if (res->status != WALRCV_OK_COMMAND)
diff --git a/src/backend/replication/logical/worker.c b/src/backend/replication/logical/worker.c
index ae1b391bda..55d48ce5be 100644
--- a/src/backend/replication/logical/worker.c
+++ b/src/backend/replication/logical/worker.c
@@ -144,6 +144,7 @@
#include "catalog/pg_tablespace.h"
#include "commands/tablecmds.h"
#include "commands/tablespace.h"
+#include "commands/sequence.h"
#include "commands/trigger.h"
#include "executor/executor.h"
#include "executor/execPartition.h"
@@ -1091,6 +1092,61 @@ apply_handle_origin(StringInfo s)
errmsg_internal("ORIGIN message sent out of order")));
}
+/*
+ * Handle SEQUENCE message.
+ */
+static void
+apply_handle_sequence(StringInfo s)
+{
+ LogicalRepSequence seq;
+ Oid relid;
+
+ if (handle_streamed_transaction(LOGICAL_REP_MSG_SEQUENCE, s))
+ return;
+
+ logicalrep_read_sequence(s, &seq);
+
+ /*
+ * Non-transactional sequence updates should not be part of a remote
+ * transaction. There should not be any running transaction.
+ */
+ Assert((!seq.transactional) || in_remote_transaction);
+ Assert(!(!seq.transactional && in_remote_transaction));
+ Assert(!(!seq.transactional && IsTransactionState()));
+
+ /*
+ * Make sure we're in a transaction (needed by ResetSequence2). For
+ * non-transactional updates we're guaranteed to start a new one,
+ * and we'll commit it at the end.
+ */
+ if (!IsTransactionState())
+ {
+ StartTransactionCommand();
+ maybe_reread_subscription();
+ }
+
+ relid = RangeVarGetRelid(makeRangeVar(seq.nspname,
+ seq.seqname, -1),
+ RowExclusiveLock, false);
+
+ /* lock the sequence in AccessExclusiveLock, as expected by ResetSequence2 */
+ elog(WARNING, "locking sequence %d in exclusive mode", relid);
+ LockRelationOid(relid, AccessExclusiveLock);
+
+ elog(WARNING, "applying sequence %s.%s transactional %d created %d last_value %ld log_cnt %ld is_called %d",
+ seq.nspname, seq.seqname, seq.transactional, seq.created, seq.last_value, seq.log_cnt, seq.is_called);
+
+ /* apply the sequence change */
+ ResetSequence2(relid, seq.last_value, seq.log_cnt, seq.is_called);
+
+ /*
+ * Commit the per-stream transaction (we only do this when not in
+ * remote transaction, i.e. for non-transactional sequence updates.
+ */
+ if (!in_remote_transaction)
+ CommitTransactionCommand();
+}
+
/*
* Handle STREAM START message.
*/
@@ -2374,6 +2430,10 @@ apply_dispatch(StringInfo s)
*/
break;
+ case LOGICAL_REP_MSG_SEQUENCE:
+ apply_handle_sequence(s);
+ return;
+
case LOGICAL_REP_MSG_STREAM_START:
apply_handle_stream_start(s);
break;
diff --git a/src/backend/replication/pgoutput/pgoutput.c b/src/backend/replication/pgoutput/pgoutput.c
index 6f6a203dea..fef568b5b2 100644
--- a/src/backend/replication/pgoutput/pgoutput.c
+++ b/src/backend/replication/pgoutput/pgoutput.c
@@ -49,6 +49,10 @@ static void pgoutput_message(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn, XLogRecPtr message_lsn,
bool transactional, const char *prefix,
Size sz, const char *message);
+static void pgoutput_sequence(LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn, XLogRecPtr sequence_lsn,
+ Relation rel, bool transactional, bool created,
+ int64 last_value, int64 log_cnt, bool is_called);
static bool pgoutput_origin_filter(LogicalDecodingContext *ctx,
RepOriginId origin_id);
static void pgoutput_begin_prepare_txn(LogicalDecodingContext *ctx,
@@ -159,6 +163,7 @@ _PG_output_plugin_init(OutputPluginCallbacks *cb)
cb->change_cb = pgoutput_change;
cb->truncate_cb = pgoutput_truncate;
cb->message_cb = pgoutput_message;
+ cb->sequence_cb = pgoutput_sequence;
cb->commit_cb = pgoutput_commit_txn;
cb->begin_prepare_cb = pgoutput_begin_prepare_txn;
@@ -175,6 +180,7 @@ _PG_output_plugin_init(OutputPluginCallbacks *cb)
cb->stream_commit_cb = pgoutput_stream_commit;
cb->stream_change_cb = pgoutput_change;
cb->stream_message_cb = pgoutput_message;
+ cb->stream_sequence_cb = pgoutput_sequence;
cb->stream_truncate_cb = pgoutput_truncate;
/* transaction streaming - two-phase commit */
cb->stream_prepare_cb = pgoutput_stream_prepare_txn;
@@ -188,6 +194,7 @@ parse_output_parameters(List *options, PGOutputData *data)
bool publication_names_given = false;
bool binary_option_given = false;
bool messages_option_given = false;
+ bool sequences_option_given = false;
bool streaming_given = false;
bool two_phase_option_given = false;
@@ -195,6 +202,7 @@ parse_output_parameters(List *options, PGOutputData *data)
data->streaming = false;
data->messages = false;
data->two_phase = false;
+ data->sequences = true;
foreach(lc, options)
{
@@ -260,6 +268,16 @@ parse_output_parameters(List *options, PGOutputData *data)
data->messages = defGetBoolean(defel);
}
+ else if (strcmp(defel->defname, "sequences") == 0)
+ {
+ if (sequences_option_given)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("conflicting or redundant options")));
+ sequences_option_given = true;
+
+ data->sequences = defGetBoolean(defel);
+ }
else if (strcmp(defel->defname, "streaming") == 0)
{
if (streaming_given)
@@ -856,6 +874,52 @@ pgoutput_message(LogicalDecodingContext *ctx, ReorderBufferTXN *txn,
OutputPluginWrite(ctx, true);
}
+static void
+pgoutput_sequence(LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn, XLogRecPtr sequence_lsn,
+ Relation rel, bool transactional, bool created,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ PGOutputData *data = (PGOutputData *) ctx->output_plugin_private;
+ TransactionId xid = InvalidTransactionId;
+ RelationSyncEntry *relentry;
+
+ if (!data->sequences)
+ return;
+
+ if (!is_publishable_relation(rel))
+ return;
+
+ /*
+ * Remember the xid for the message in streaming mode. See
+ * pgoutput_change.
+ */
+ if (in_streaming)
+ xid = txn->xid;
+
+ relentry = get_rel_sync_entry(data, RelationGetRelid(rel));
+
+ /*
+ * First check the sequence filter.
+ *
+ * We handle just REORDER_BUFFER_CHANGE_SEQUENCE here.
+ */
+ if (!relentry->pubactions.pubsequence)
+ return;
+
+ OutputPluginPrepareWrite(ctx, true);
+ logicalrep_write_sequence(ctx->out,
+ rel,
+ xid,
+ sequence_lsn,
+ transactional,
+ created,
+ last_value,
+ log_cnt,
+ is_called);
+ OutputPluginWrite(ctx, true);
+}
+
/*
* Currently we always forward.
*/
@@ -1140,7 +1204,8 @@ get_rel_sync_entry(PGOutputData *data, Oid relid)
entry->streamed_txns = NIL;
entry->replicate_valid = false;
entry->pubactions.pubinsert = entry->pubactions.pubupdate =
- entry->pubactions.pubdelete = entry->pubactions.pubtruncate = false;
+ entry->pubactions.pubdelete = entry->pubactions.pubtruncate =
+ entry->pubactions.pubsequence = false;
entry->publish_as_relid = InvalidOid;
entry->map = NULL; /* will be set by maybe_send_schema() if
* needed */
@@ -1160,6 +1225,7 @@ get_rel_sync_entry(PGOutputData *data, Oid relid)
List *schemaPubids = GetSchemaPublications(schemaId);
ListCell *lc;
Oid publish_as_relid = relid;
+ bool is_sequence = (get_rel_relkind(relid) == RELKIND_SEQUENCE);
/* Reload publications if needed before use. */
if (!publications_valid)
@@ -1183,12 +1249,23 @@ get_rel_sync_entry(PGOutputData *data, Oid relid)
Publication *pub = lfirst(lc);
bool publish = false;
- if (pub->alltables)
+ if (pub->alltables && (!is_sequence))
{
publish = true;
if (pub->pubviaroot && am_partition)
publish_as_relid = llast_oid(get_partition_ancestors(relid));
}
+ else if (pub->allsequences && is_sequence)
+ {
+ publish = true;
+ }
+
+ /* if a sequence, just cross-check the list of publications */
+ if (!publish && is_sequence)
+ {
+ if (list_member_oid(pubids, pub->oid))
+ publish = true;
+ }
if (!publish)
{
@@ -1243,10 +1320,12 @@ get_rel_sync_entry(PGOutputData *data, Oid relid)
entry->pubactions.pubupdate |= pub->pubactions.pubupdate;
entry->pubactions.pubdelete |= pub->pubactions.pubdelete;
entry->pubactions.pubtruncate |= pub->pubactions.pubtruncate;
+ entry->pubactions.pubsequence |= pub->pubactions.pubsequence;
}
if (entry->pubactions.pubinsert && entry->pubactions.pubupdate &&
- entry->pubactions.pubdelete && entry->pubactions.pubtruncate)
+ entry->pubactions.pubdelete && entry->pubactions.pubtruncate &&
+ entry->pubactions.pubsequence)
break;
}
@@ -1391,6 +1470,7 @@ rel_sync_cache_publication_cb(Datum arg, int cacheid, uint32 hashvalue)
entry->pubactions.pubupdate = false;
entry->pubactions.pubdelete = false;
entry->pubactions.pubtruncate = false;
+ entry->pubactions.pubsequence = false;
}
}
diff --git a/src/backend/utils/cache/relcache.c b/src/backend/utils/cache/relcache.c
index 9fa9e671a1..1c3286a38c 100644
--- a/src/backend/utils/cache/relcache.c
+++ b/src/backend/utils/cache/relcache.c
@@ -5605,6 +5605,7 @@ GetRelationPublicationActions(Relation relation)
pubactions->pubupdate |= pubform->pubupdate;
pubactions->pubdelete |= pubform->pubdelete;
pubactions->pubtruncate |= pubform->pubtruncate;
+ pubactions->pubsequence |= pubform->pubsequence;
ReleaseSysCache(tup);
@@ -5613,7 +5614,8 @@ GetRelationPublicationActions(Relation relation)
* other publications.
*/
if (pubactions->pubinsert && pubactions->pubupdate &&
- pubactions->pubdelete && pubactions->pubtruncate)
+ pubactions->pubdelete && pubactions->pubtruncate &&
+ pubactions->pubsequence)
break;
}
diff --git a/src/bin/psql/tab-complete.c b/src/bin/psql/tab-complete.c
index fa2e19593c..a63ed8eea4 100644
--- a/src/bin/psql/tab-complete.c
+++ b/src/bin/psql/tab-complete.c
@@ -1647,13 +1647,13 @@ psql_completion(const char *text, int start, int end)
COMPLETE_WITH("ADD", "DROP", "OWNER TO", "RENAME TO", "SET");
/* ALTER PUBLICATION <name> ADD */
else if (Matches("ALTER", "PUBLICATION", MatchAny, "ADD"))
- COMPLETE_WITH("ALL TABLES IN SCHEMA", "TABLE");
+ COMPLETE_WITH("ALL TABLES IN SCHEMA", "TABLE|SEQUENCE");
/* ALTER PUBLICATION <name> DROP */
else if (Matches("ALTER", "PUBLICATION", MatchAny, "DROP"))
- COMPLETE_WITH("ALL TABLES IN SCHEMA", "TABLE");
+ COMPLETE_WITH("ALL TABLES IN SCHEMA", "TABLE|SEQUENCE");
/* ALTER PUBLICATION <name> SET */
else if (Matches("ALTER", "PUBLICATION", MatchAny, "SET"))
- COMPLETE_WITH("(", "ALL TABLES IN SCHEMA", "TABLE");
+ COMPLETE_WITH("(", "ALL TABLES IN SCHEMA", "TABLE|SEQUENCE");
else if (Matches("ALTER", "PUBLICATION", MatchAny, "ADD|DROP|SET", "ALL", "TABLES", "IN", "SCHEMA"))
COMPLETE_WITH_QUERY(Query_for_list_of_schemas
" AND nspname != 'pg_catalog' "
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 6412f369f1..be5e3b4d40 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -11513,6 +11513,11 @@
provolatile => 's', prorettype => 'oid', proargtypes => 'text',
proallargtypes => '{text,oid}', proargmodes => '{i,o}',
proargnames => '{pubname,relid}', prosrc => 'pg_get_publication_tables' },
+{ oid => '8000', descr => 'get OIDs of sequences in a publication',
+ proname => 'pg_get_publication_sequences', prorows => '1000', proretset => 't',
+ provolatile => 's', prorettype => 'oid', proargtypes => 'text',
+ proallargtypes => '{text,oid}', proargmodes => '{i,o}',
+ proargnames => '{pubname,relid}', prosrc => 'pg_get_publication_sequences' },
{ oid => '6121',
descr => 'returns whether a relation can be part of a publication',
proname => 'pg_relation_is_publishable', provolatile => 's',
diff --git a/src/include/catalog/pg_publication.h b/src/include/catalog/pg_publication.h
index 1ae439e6f3..0a05fb885c 100644
--- a/src/include/catalog/pg_publication.h
+++ b/src/include/catalog/pg_publication.h
@@ -40,6 +40,12 @@ CATALOG(pg_publication,6104,PublicationRelationId)
*/
bool puballtables;
+ /*
+ * indicates that this is special publication which should encompass all
+ * sequences in the database (except for the unlogged and temp ones)
+ */
+ bool puballsequences;
+
/* true if inserts are published */
bool pubinsert;
@@ -52,6 +58,9 @@ CATALOG(pg_publication,6104,PublicationRelationId)
/* true if truncates are published */
bool pubtruncate;
+ /* true if sequences are published */
+ bool pubsequence;
+
/* true if partition changes are published using root schema */
bool pubviaroot;
} FormData_pg_publication;
@@ -72,6 +81,7 @@ typedef struct PublicationActions
bool pubupdate;
bool pubdelete;
bool pubtruncate;
+ bool pubsequence;
} PublicationActions;
typedef struct Publication
@@ -79,6 +89,7 @@ typedef struct Publication
Oid oid;
char *name;
bool alltables;
+ bool allsequences;
bool pubviaroot;
PublicationActions pubactions;
} Publication;
@@ -121,6 +132,9 @@ extern List *GetPubPartitionOptionRelations(List *result,
PublicationPartOpt pub_partopt,
Oid relid);
+extern List *GetAllSequencesPublicationRelations(void);
+extern List *GetPublicationSequenceRelations(Oid pubid);
+
extern bool is_publishable_relation(Relation rel);
extern ObjectAddress publication_add_relation(Oid pubid, PublicationRelInfo *targetrel,
bool if_not_exists);
diff --git a/src/include/commands/sequence.h b/src/include/commands/sequence.h
index ddb36b5b17..00038fcea3 100644
--- a/src/include/commands/sequence.h
+++ b/src/include/commands/sequence.h
@@ -60,6 +60,7 @@ extern ObjectAddress DefineSequence(ParseState *pstate, CreateSeqStmt *stmt);
extern ObjectAddress AlterSequence(ParseState *pstate, AlterSeqStmt *stmt);
extern void DeleteSequenceTuple(Oid relid);
extern void ResetSequence(Oid seq_relid);
+extern void ResetSequence2(Oid seq_relid, int64 last_value, int64 log_cnt, bool is_called);
extern void ResetSequenceCaches(void);
extern void AtEOXact_Sequences(bool isCommit);
diff --git a/src/include/nodes/parsenodes.h b/src/include/nodes/parsenodes.h
index 067138e6b5..44e596fd12 100644
--- a/src/include/nodes/parsenodes.h
+++ b/src/include/nodes/parsenodes.h
@@ -3652,6 +3652,10 @@ typedef enum PublicationObjSpecType
PUBLICATIONOBJ_TABLE_IN_SCHEMA, /* Tables in schema type */
PUBLICATIONOBJ_TABLE_IN_CUR_SCHEMA, /* Get the first element from
* search_path */
+ PUBLICATIONOBJ_SEQUENCE, /* Sequence type */
+ PUBLICATIONOBJ_SEQUENCE_IN_SCHEMA, /* Sequences in schema type */
+ PUBLICATIONOBJ_SEQUENCE_IN_CUR_SCHEMA, /* Get the first element from
+ * search_path */
PUBLICATIONOBJ_CONTINUATION /* Continuation of previous type */
} PublicationObjSpecType;
@@ -3671,6 +3675,7 @@ typedef struct CreatePublicationStmt
List *options; /* List of DefElem nodes */
List *pubobjects; /* Optional list of publication objects */
bool for_all_tables; /* Special publication for all tables in db */
+ bool for_all_sequences; /* Special publication for all sequences in db */
} CreatePublicationStmt;
typedef struct AlterPublicationStmt
@@ -3687,6 +3692,7 @@ typedef struct AlterPublicationStmt
*/
List *pubobjects; /* Optional list of publication objects */
bool for_all_tables; /* Special publication for all tables in db */
+ bool for_all_sequences; /* Special publication for all sequences in db */
DefElemAction action; /* What action to perform with the
* tables/schemas */
} AlterPublicationStmt;
diff --git a/src/include/replication/logicalproto.h b/src/include/replication/logicalproto.h
index 83741dcf42..d468c3ccfb 100644
--- a/src/include/replication/logicalproto.h
+++ b/src/include/replication/logicalproto.h
@@ -60,6 +60,7 @@ typedef enum LogicalRepMsgType
LOGICAL_REP_MSG_RELATION = 'R',
LOGICAL_REP_MSG_TYPE = 'Y',
LOGICAL_REP_MSG_MESSAGE = 'M',
+ LOGICAL_REP_MSG_SEQUENCE = 'X', /* FIXME change */
LOGICAL_REP_MSG_BEGIN_PREPARE = 'b',
LOGICAL_REP_MSG_PREPARE = 'P',
LOGICAL_REP_MSG_COMMIT_PREPARED = 'K',
@@ -117,6 +118,19 @@ typedef struct LogicalRepTyp
char *typname; /* name of the remote type */
} LogicalRepTyp;
+/* Sequence info */
+typedef struct LogicalRepSequence
+{
+ Oid remoteid; /* unique id of the remote sequence */
+ char *nspname; /* schema name of remote sequence */
+ char *seqname; /* name of the remote sequence */
+ bool transactional;
+ bool created;
+ int64 last_value;
+ int64 log_cnt; /* XXX probably not needed? */
+ bool is_called; /* XXX probably not needed? */
+} LogicalRepSequence;
+
/* Transaction info */
typedef struct LogicalRepBeginData
{
@@ -227,6 +241,12 @@ extern List *logicalrep_read_truncate(StringInfo in,
bool *cascade, bool *restart_seqs);
extern void logicalrep_write_message(StringInfo out, TransactionId xid, XLogRecPtr lsn,
bool transactional, const char *prefix, Size sz, const char *message);
+extern void logicalrep_write_sequence(StringInfo out, Relation rel,
+ TransactionId xid, XLogRecPtr lsn,
+ bool transactional, bool created,
+ int64 last_value, int64 log_cnt,
+ bool is_called);
+extern void logicalrep_read_sequence(StringInfo in, LogicalRepSequence *seqdata);
extern void logicalrep_write_rel(StringInfo out, TransactionId xid,
Relation rel);
extern LogicalRepRelation *logicalrep_read_rel(StringInfo in);
diff --git a/src/include/replication/pgoutput.h b/src/include/replication/pgoutput.h
index 0dc460fb70..97eb2a7ef1 100644
--- a/src/include/replication/pgoutput.h
+++ b/src/include/replication/pgoutput.h
@@ -28,6 +28,7 @@ typedef struct PGOutputData
bool streaming;
bool messages;
bool two_phase;
+ bool sequences;
} PGOutputData;
#endif /* PGOUTPUT_H */
diff --git a/src/test/regress/expected/rules.out b/src/test/regress/expected/rules.out
index 2fa00a3c29..b32a1e6729 100644
--- a/src/test/regress/expected/rules.out
+++ b/src/test/regress/expected/rules.out
@@ -1451,6 +1451,14 @@ pg_prepared_xacts| SELECT p.transaction,
FROM ((pg_prepared_xact() p(transaction, gid, prepared, ownerid, dbid)
LEFT JOIN pg_authid u ON ((p.ownerid = u.oid)))
LEFT JOIN pg_database d ON ((p.dbid = d.oid)));
+pg_publication_sequences| SELECT p.pubname,
+ n.nspname AS schemaname,
+ c.relname AS sequencename
+ FROM pg_publication p,
+ LATERAL pg_get_publication_sequences((p.pubname)::text) gpt(relid),
+ (pg_class c
+ JOIN pg_namespace n ON ((n.oid = c.relnamespace)))
+ WHERE (c.oid = gpt.relid);
pg_publication_tables| SELECT p.pubname,
n.nspname AS schemaname,
c.relname AS tablename
diff --git a/src/test/subscription/t/023_sequences.pl b/src/test/subscription/t/023_sequences.pl
new file mode 100644
index 0000000000..734f428644
--- /dev/null
+++ b/src/test/subscription/t/023_sequences.pl
@@ -0,0 +1,196 @@
+
+# Copyright (c) 2021, PostgreSQL Global Development Group
+
+# This tests that sequences are replicated correctly by logical replication
+use strict;
+use warnings;
+use PostgreSQL::Test::Cluster;
+use PostgreSQL::Test::Utils;
+use Test::More tests => 6;
+
+# Initialize publisher node
+my $node_publisher = PostgreSQL::Test::Cluster->new('publisher');
+$node_publisher->init(allows_streaming => 'logical');
+$node_publisher->start;
+
+# Create subscriber node
+my $node_subscriber = PostgreSQL::Test::Cluster->new('subscriber');
+$node_subscriber->init(allows_streaming => 'logical');
+$node_subscriber->start;
+
+# Create some preexisting content on publisher
+my $ddl = qq(
+ CREATE SEQUENCE s;
+);
+
+# Setup structure on the publisher
+$node_publisher->safe_psql('postgres', $ddl);
+
+# Create some the same structure on subscriber, and an extra sequence that
+# we'll create on the publisher later
+$ddl = qq(
+ CREATE SEQUENCE s;
+ CREATE SEQUENCE s2;
+);
+
+$node_subscriber->safe_psql('postgres', $ddl);
+
+# Setup logical replication
+my $publisher_connstr = $node_publisher->connstr . ' dbname=postgres';
+$node_publisher->safe_psql('postgres',
+ "CREATE PUBLICATION seq_pub");
+
+$node_publisher->safe_psql('postgres',
+ "ALTER PUBLICATION seq_pub ADD SEQUENCE s");
+
+$node_subscriber->safe_psql('postgres',
+ "CREATE SUBSCRIPTION seq_sub CONNECTION '$publisher_connstr' PUBLICATION seq_pub WITH (slot_name = seq_sub_slot)"
+);
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Wait for initial sync to finish as well
+my $synced_query =
+ "SELECT count(1) = 0 FROM pg_subscription_rel WHERE srsubstate NOT IN ('s', 'r');";
+$node_subscriber->poll_query_until('postgres', $synced_query)
+ or die "Timed out while waiting for subscriber to synchronize data";
+
+# Insert initial test data
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ -- generate a number of values using the sequence
+ SELECT nextval('s') FROM generate_series(1,100);
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+my $result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s;
+));
+
+is( $result, '132|0|t',
+ 'check replicated sequence values on subscriber');
+
+
+# advance the sequence in a rolled-back transaction - should not be replicated
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ SELECT nextval('s') FROM generate_series(1,100);
+ ROLLBACK;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s;
+));
+
+is( $result, '132|0|t',
+ 'check replicated sequence values on subscriber');
+
+
+# create a new sequence and roll it back - should not be replicated, due to
+# the transactional behavior
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ CREATE SEQUENCE s2;
+ ALTER PUBLICATION seq_pub ADD SEQUENCE s2;
+ SELECT nextval('s2') FROM generate_series(1,100);
+ ROLLBACK;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s2;
+));
+
+is( $result, '1|0|f',
+ 'check replicated sequence values on subscriber');
+
+
+# create a new sequence, advance it in a rolled-back transaction, but commit
+# the create - the advance should be replicated nevertheless
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ CREATE SEQUENCE s2;
+ ALTER PUBLICATION seq_pub ADD SEQUENCE s2;
+ SAVEPOINT sp1;
+ SELECT nextval('s2') FROM generate_series(1,100);
+ ROLLBACK TO sp1;
+ COMMIT;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Wait for sync of the second sequence we just added to finish
+$synced_query =
+ "SELECT count(1) = 0 FROM pg_subscription_rel WHERE srsubstate NOT IN ('s', 'r');";
+$node_subscriber->poll_query_until('postgres', $synced_query)
+ or die "Timed out while waiting for subscriber to synchronize data";
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s2;
+));
+
+is( $result, '132|0|t',
+ 'check replicated sequence values on subscriber');
+
+
+# advance the new sequence in a transaction, and roll it back - in this case
+# it should not be replicated at commit
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ SELECT nextval('s2') FROM generate_series(1,100);
+ ROLLBACK;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s2;
+));
+
+is( $result, '132|0|t',
+ 'check replicated sequence values on subscriber');
+
+
+# advance the sequence in a subtransaction - the subtransaction gets rolled
+# back, but commit the main one - the changes should still be replicated
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ SAVEPOINT s1;
+ SELECT nextval('s2') FROM generate_series(1,100);
+ ROLLBACK TO s1;
+ COMMIT;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s2;
+));
+
+is( $result, '330|0|t',
+ 'check replicated sequence values on subscriber');
+
+
+$node_subscriber->stop('fast');
+$node_publisher->stop('fast');
--
2.31.1
[text/x-patch] 0003-Add-support-for-decoding-sequences-to-test_-20211121.patch (19.9K, ../../[email protected]/3-0003-Add-support-for-decoding-sequences-to-test_-20211121.patch)
download | inline diff:
From b4ac63f9f8d1739366021cee293b14231f753386 Mon Sep 17 00:00:00 2001
From: Tomas Vondra <[email protected]>
Date: Fri, 24 Sep 2021 00:42:04 +0200
Subject: [PATCH 3/4] Add support for decoding sequences to test_decoding
---
contrib/test_decoding/Makefile | 3 +-
contrib/test_decoding/expected/sequence.out | 321 ++++++++++++++++++++
contrib/test_decoding/sql/sequence.sql | 119 ++++++++
contrib/test_decoding/test_decoding.c | 69 +++++
4 files changed, 511 insertions(+), 1 deletion(-)
create mode 100644 contrib/test_decoding/expected/sequence.out
create mode 100644 contrib/test_decoding/sql/sequence.sql
diff --git a/contrib/test_decoding/Makefile b/contrib/test_decoding/Makefile
index 9a31e0b879..56ddc3abae 100644
--- a/contrib/test_decoding/Makefile
+++ b/contrib/test_decoding/Makefile
@@ -5,7 +5,8 @@ PGFILEDESC = "test_decoding - example of a logical decoding output plugin"
REGRESS = ddl xact rewrite toast permissions decoding_in_xact \
decoding_into_rel binary prepared replorigin time messages \
- spill slot truncate stream stats twophase twophase_stream
+ spill slot truncate stream stats twophase twophase_stream \
+ sequence
ISOLATION = mxact delayed_startup ondisk_startup concurrent_ddl_dml \
oldest_xmin snapshot_transfer subxact_without_top concurrent_stream \
twophase_snapshot
diff --git a/contrib/test_decoding/expected/sequence.out b/contrib/test_decoding/expected/sequence.out
new file mode 100644
index 0000000000..17c88990b1
--- /dev/null
+++ b/contrib/test_decoding/expected/sequence.out
@@ -0,0 +1,321 @@
+-- predictability
+SET synchronous_commit = on;
+SELECT 'init' FROM pg_create_logical_replication_slot('regression_slot', 'test_decoding');
+ ?column?
+----------
+ init
+(1 row)
+
+CREATE SEQUENCE test_sequence;
+-- test the sequence changes by several nextval() calls
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 1
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 2
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 4
+(1 row)
+
+-- test the sequence changes by several ALTER commands
+ALTER SEQUENCE test_sequence INCREMENT BY 10;
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 14
+(1 row)
+
+ALTER SEQUENCE test_sequence START WITH 3000;
+ALTER SEQUENCE test_sequence MAXVALUE 10000;
+ALTER SEQUENCE test_sequence RESTART WITH 4000;
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 4000
+(1 row)
+
+-- test the sequence changes by several setval() calls
+SELECT setval('test_sequence', 3500);
+ setval
+--------
+ 3500
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3510
+(1 row)
+
+SELECT setval('test_sequence', 3500, true);
+ setval
+--------
+ 3500
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3510
+(1 row)
+
+SELECT setval('test_sequence', 3500, false);
+ setval
+--------
+ 3500
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3500
+(1 row)
+
+-- show results and drop sequence
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+ data
+------------------------------------------------------------------------------------------------
+ BEGIN
+ sequence public.test_sequence: transactional:1 created:1 last_value:1 log_cnt:0 is_called:0
+ COMMIT
+ sequence public.test_sequence: transactional:0 created:0 last_value:33 log_cnt:0 is_called:1
+ BEGIN
+ sequence public.test_sequence: transactional:1 created:1 last_value:4 log_cnt:0 is_called:1
+ COMMIT
+ sequence public.test_sequence: transactional:0 created:0 last_value:334 log_cnt:0 is_called:1
+ BEGIN
+ sequence public.test_sequence: transactional:1 created:1 last_value:14 log_cnt:32 is_called:1
+ COMMIT
+ BEGIN
+ sequence public.test_sequence: transactional:1 created:1 last_value:14 log_cnt:0 is_called:1
+ COMMIT
+ BEGIN
+ sequence public.test_sequence: transactional:1 created:1 last_value:4000 log_cnt:0 is_called:0
+ COMMIT
+ sequence public.test_sequence: transactional:0 created:0 last_value:4320 log_cnt:0 is_called:1
+ sequence public.test_sequence: transactional:0 created:0 last_value:3500 log_cnt:0 is_called:1
+ sequence public.test_sequence: transactional:0 created:0 last_value:3830 log_cnt:0 is_called:1
+ sequence public.test_sequence: transactional:0 created:0 last_value:3500 log_cnt:0 is_called:1
+ sequence public.test_sequence: transactional:0 created:0 last_value:3830 log_cnt:0 is_called:1
+ sequence public.test_sequence: transactional:0 created:0 last_value:3500 log_cnt:0 is_called:0
+ sequence public.test_sequence: transactional:0 created:0 last_value:3820 log_cnt:0 is_called:1
+(24 rows)
+
+DROP SEQUENCE test_sequence;
+-- rollback on sequence creation and update
+BEGIN;
+CREATE SEQUENCE test_sequence;
+CREATE TABLE test_table (a INT);
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 1
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 2
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3
+(1 row)
+
+SELECT setval('test_sequence', 3000);
+ setval
+--------
+ 3000
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3001
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3002
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3003
+(1 row)
+
+ALTER SEQUENCE test_sequence RESTART WITH 6000;
+INSERT INTO test_table VALUES( (SELECT nextval('test_sequence')) );
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 6001
+(1 row)
+
+ROLLBACK;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'skip-sequences', '0');
+ data
+------
+(0 rows)
+
+-- rollback on table creation with serial column
+BEGIN;
+CREATE TABLE test_table (a SERIAL, b INT);
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+INSERT INTO test_table (b) VALUES (300);
+SELECT setval('test_table_a_seq', 3000);
+ setval
+--------
+ 3000
+(1 row)
+
+INSERT INTO test_table (b) VALUES (400);
+INSERT INTO test_table (b) VALUES (500);
+INSERT INTO test_table (b) VALUES (600);
+ALTER SEQUENCE test_table_a_seq RESTART WITH 6000;
+INSERT INTO test_table (b) VALUES (700);
+INSERT INTO test_table (b) VALUES (800);
+INSERT INTO test_table (b) VALUES (900);
+ROLLBACK;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'skip-sequences', '0');
+ data
+------
+(0 rows)
+
+-- rollback on table with serial column
+CREATE TABLE test_table (a SERIAL, b INT);
+BEGIN;
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+INSERT INTO test_table (b) VALUES (300);
+SELECT setval('test_table_a_seq', 3000);
+ setval
+--------
+ 3000
+(1 row)
+
+INSERT INTO test_table (b) VALUES (400);
+INSERT INTO test_table (b) VALUES (500);
+INSERT INTO test_table (b) VALUES (600);
+ALTER SEQUENCE test_table_a_seq RESTART WITH 6000;
+INSERT INTO test_table (b) VALUES (700);
+INSERT INTO test_table (b) VALUES (800);
+INSERT INTO test_table (b) VALUES (900);
+ROLLBACK;
+-- check table and sequence values after rollback
+SELECT * from test_table_a_seq;
+ last_value | log_cnt | is_called
+------------+---------+-----------
+ 3003 | 30 | t
+(1 row)
+
+SELECT nextval('test_table_a_seq');
+ nextval
+---------
+ 3004
+(1 row)
+
+DROP TABLE test_table;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+ data
+---------------------------------------------------------------------------------------------------
+ BEGIN
+ sequence public.test_table_a_seq: transactional:1 created:1 last_value:1 log_cnt:0 is_called:0
+ COMMIT
+ sequence public.test_table_a_seq: transactional:0 created:0 last_value:33 log_cnt:0 is_called:1
+ sequence public.test_table_a_seq: transactional:0 created:0 last_value:3000 log_cnt:0 is_called:1
+ sequence public.test_table_a_seq: transactional:0 created:0 last_value:3033 log_cnt:0 is_called:1
+ BEGIN
+ COMMIT
+(8 rows)
+
+-- savepoint test on table with serial column
+BEGIN;
+CREATE TABLE test_table (a SERIAL, b INT);
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+SAVEPOINT a;
+INSERT INTO test_table (b) VALUES (300);
+ROLLBACK TO SAVEPOINT a;
+DROP TABLE test_table;
+COMMIT;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+ data
+--------------------------------------------------------------
+ BEGIN
+ table public.test_table: INSERT: a[integer]:1 b[integer]:100
+ table public.test_table: INSERT: a[integer]:2 b[integer]:200
+ COMMIT
+(4 rows)
+
+-- savepoint test on table with serial column
+BEGIN;
+CREATE SEQUENCE test_sequence;
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 1
+(1 row)
+
+SELECT setval('test_sequence', 3000);
+ setval
+--------
+ 3000
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3001
+(1 row)
+
+SAVEPOINT a;
+ALTER SEQUENCE test_sequence START WITH 7000;
+SELECT setval('test_sequence', 5000);
+ setval
+--------
+ 5000
+(1 row)
+
+ROLLBACK TO SAVEPOINT a;
+SELECT * FROM test_sequence;
+ last_value | log_cnt | is_called
+------------+---------+-----------
+ 3001 | 32 | t
+(1 row)
+
+DROP SEQUENCE test_sequence;
+COMMIT;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+ data
+--------
+ BEGIN
+ COMMIT
+(2 rows)
+
+SELECT pg_drop_replication_slot('regression_slot');
+ pg_drop_replication_slot
+--------------------------
+
+(1 row)
+
diff --git a/contrib/test_decoding/sql/sequence.sql b/contrib/test_decoding/sql/sequence.sql
new file mode 100644
index 0000000000..d8a34738f3
--- /dev/null
+++ b/contrib/test_decoding/sql/sequence.sql
@@ -0,0 +1,119 @@
+-- predictability
+SET synchronous_commit = on;
+SELECT 'init' FROM pg_create_logical_replication_slot('regression_slot', 'test_decoding');
+
+CREATE SEQUENCE test_sequence;
+
+-- test the sequence changes by several nextval() calls
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+
+-- test the sequence changes by several ALTER commands
+ALTER SEQUENCE test_sequence INCREMENT BY 10;
+SELECT nextval('test_sequence');
+
+ALTER SEQUENCE test_sequence START WITH 3000;
+ALTER SEQUENCE test_sequence MAXVALUE 10000;
+ALTER SEQUENCE test_sequence RESTART WITH 4000;
+SELECT nextval('test_sequence');
+
+-- test the sequence changes by several setval() calls
+SELECT setval('test_sequence', 3500);
+SELECT nextval('test_sequence');
+SELECT setval('test_sequence', 3500, true);
+SELECT nextval('test_sequence');
+SELECT setval('test_sequence', 3500, false);
+SELECT nextval('test_sequence');
+
+-- show results and drop sequence
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+DROP SEQUENCE test_sequence;
+
+-- rollback on sequence creation and update
+BEGIN;
+CREATE SEQUENCE test_sequence;
+CREATE TABLE test_table (a INT);
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT setval('test_sequence', 3000);
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+ALTER SEQUENCE test_sequence RESTART WITH 6000;
+INSERT INTO test_table VALUES( (SELECT nextval('test_sequence')) );
+SELECT nextval('test_sequence');
+ROLLBACK;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'skip-sequences', '0');
+
+-- rollback on table creation with serial column
+BEGIN;
+CREATE TABLE test_table (a SERIAL, b INT);
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+INSERT INTO test_table (b) VALUES (300);
+SELECT setval('test_table_a_seq', 3000);
+INSERT INTO test_table (b) VALUES (400);
+INSERT INTO test_table (b) VALUES (500);
+INSERT INTO test_table (b) VALUES (600);
+ALTER SEQUENCE test_table_a_seq RESTART WITH 6000;
+INSERT INTO test_table (b) VALUES (700);
+INSERT INTO test_table (b) VALUES (800);
+INSERT INTO test_table (b) VALUES (900);
+ROLLBACK;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'skip-sequences', '0');
+
+-- rollback on table with serial column
+CREATE TABLE test_table (a SERIAL, b INT);
+
+BEGIN;
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+INSERT INTO test_table (b) VALUES (300);
+SELECT setval('test_table_a_seq', 3000);
+INSERT INTO test_table (b) VALUES (400);
+INSERT INTO test_table (b) VALUES (500);
+INSERT INTO test_table (b) VALUES (600);
+ALTER SEQUENCE test_table_a_seq RESTART WITH 6000;
+INSERT INTO test_table (b) VALUES (700);
+INSERT INTO test_table (b) VALUES (800);
+INSERT INTO test_table (b) VALUES (900);
+ROLLBACK;
+
+-- check table and sequence values after rollback
+SELECT * from test_table_a_seq;
+SELECT nextval('test_table_a_seq');
+
+DROP TABLE test_table;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+
+-- savepoint test on table with serial column
+BEGIN;
+CREATE TABLE test_table (a SERIAL, b INT);
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+SAVEPOINT a;
+INSERT INTO test_table (b) VALUES (300);
+ROLLBACK TO SAVEPOINT a;
+DROP TABLE test_table;
+COMMIT;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+
+-- savepoint test on table with serial column
+BEGIN;
+CREATE SEQUENCE test_sequence;
+SELECT nextval('test_sequence');
+SELECT setval('test_sequence', 3000);
+SELECT nextval('test_sequence');
+SAVEPOINT a;
+ALTER SEQUENCE test_sequence START WITH 7000;
+SELECT setval('test_sequence', 5000);
+ROLLBACK TO SAVEPOINT a;
+SELECT * FROM test_sequence;
+DROP SEQUENCE test_sequence;
+COMMIT;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+
+SELECT pg_drop_replication_slot('regression_slot');
diff --git a/contrib/test_decoding/test_decoding.c b/contrib/test_decoding/test_decoding.c
index e5cd84e85e..45765d299e 100644
--- a/contrib/test_decoding/test_decoding.c
+++ b/contrib/test_decoding/test_decoding.c
@@ -35,6 +35,7 @@ typedef struct
bool include_timestamp;
bool skip_empty_xacts;
bool only_local;
+ bool skip_sequences;
} TestDecodingData;
/*
@@ -76,6 +77,10 @@ static void pg_decode_message(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn, XLogRecPtr message_lsn,
bool transactional, const char *prefix,
Size sz, const char *message);
+static void pg_decode_sequence(LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn, XLogRecPtr sequence_lsn,
+ Relation rel, bool transactional, bool created,
+ int64 last_value, int64 log_cnt, bool is_called);
static bool pg_decode_filter_prepare(LogicalDecodingContext *ctx,
TransactionId xid,
const char *gid);
@@ -116,6 +121,10 @@ static void pg_decode_stream_message(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn, XLogRecPtr message_lsn,
bool transactional, const char *prefix,
Size sz, const char *message);
+static void pg_decode_stream_sequence(LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn, XLogRecPtr sequence_lsn,
+ Relation rel, bool transactional, bool created,
+ int64 last_value, int64 log_cnt, bool is_called);
static void pg_decode_stream_truncate(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn,
int nrelations, Relation relations[],
@@ -141,6 +150,7 @@ _PG_output_plugin_init(OutputPluginCallbacks *cb)
cb->filter_by_origin_cb = pg_decode_filter;
cb->shutdown_cb = pg_decode_shutdown;
cb->message_cb = pg_decode_message;
+ cb->sequence_cb = pg_decode_sequence;
cb->filter_prepare_cb = pg_decode_filter_prepare;
cb->begin_prepare_cb = pg_decode_begin_prepare_txn;
cb->prepare_cb = pg_decode_prepare_txn;
@@ -153,6 +163,7 @@ _PG_output_plugin_init(OutputPluginCallbacks *cb)
cb->stream_commit_cb = pg_decode_stream_commit;
cb->stream_change_cb = pg_decode_stream_change;
cb->stream_message_cb = pg_decode_stream_message;
+ cb->stream_sequence_cb = pg_decode_stream_sequence;
cb->stream_truncate_cb = pg_decode_stream_truncate;
}
@@ -175,6 +186,9 @@ pg_decode_startup(LogicalDecodingContext *ctx, OutputPluginOptions *opt,
data->skip_empty_xacts = false;
data->only_local = false;
+ /* skip sequences by default for backwards compatibility */
+ data->skip_sequences = true;
+
ctx->output_plugin_private = data;
opt->output_type = OUTPUT_PLUGIN_TEXTUAL_OUTPUT;
@@ -265,6 +279,17 @@ pg_decode_startup(LogicalDecodingContext *ctx, OutputPluginOptions *opt,
errmsg("could not parse value \"%s\" for parameter \"%s\"",
strVal(elem->arg), elem->defname)));
}
+ else if (strcmp(elem->defname, "skip-sequences") == 0)
+ {
+
+ if (elem->arg == NULL)
+ continue; /* true by default */
+ else if (!parse_bool(strVal(elem->arg), &data->skip_sequences))
+ ereport(ERROR,
+ (errcode(ERRCODE_INVALID_PARAMETER_VALUE),
+ errmsg("could not parse value \"%s\" for parameter \"%s\"",
+ strVal(elem->arg), elem->defname)));
+ }
else
{
ereport(ERROR,
@@ -744,6 +769,28 @@ pg_decode_message(LogicalDecodingContext *ctx,
OutputPluginWrite(ctx, true);
}
+static void
+pg_decode_sequence(LogicalDecodingContext *ctx, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ bool transactional, bool created,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ TestDecodingData *data = ctx->output_plugin_private;
+
+ /* return if requested to skip_sequences */
+ if (data->skip_sequences)
+ return;
+
+ OutputPluginPrepareWrite(ctx, true);
+ appendStringInfoString(ctx->out, "sequence ");
+ appendStringInfoString(ctx->out,
+ quote_qualified_identifier(get_namespace_name(get_rel_namespace(RelationGetRelid(rel))),
+ RelationGetRelationName(rel)));
+ appendStringInfo(ctx->out, ": transactional:%d created:%d last_value:%zu log_cnt:%zu is_called:%d",
+ transactional, created, last_value, log_cnt, is_called);
+ OutputPluginWrite(ctx, true);
+}
+
static void
pg_decode_stream_start(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn)
@@ -943,6 +990,28 @@ pg_decode_stream_message(LogicalDecodingContext *ctx,
OutputPluginWrite(ctx, true);
}
+static void
+pg_decode_stream_sequence(LogicalDecodingContext *ctx, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ bool transactional, bool created,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ TestDecodingData *data = ctx->output_plugin_private;
+
+ /* return if requested to skip_sequences */
+ if (data->skip_sequences)
+ return;
+
+ OutputPluginPrepareWrite(ctx, true);
+ appendStringInfoString(ctx->out, "streaming sequence ");
+ appendStringInfoString(ctx->out,
+ quote_qualified_identifier(get_namespace_name(get_rel_namespace(RelationGetRelid(rel))),
+ RelationGetRelationName(rel)));
+ appendStringInfo(ctx->out, ": transactional:%d created:%d last_value:%zu log_cnt:%zu is_called:%d",
+ transactional, created, last_value, log_cnt, is_called);
+ OutputPluginWrite(ctx, true);
+}
+
/*
* In streaming mode, we don't display the detailed information of Truncate.
* See pg_decode_stream_change.
--
2.31.1
[text/x-patch] 0002-rework-20211121.patch (37.2K, ../../[email protected]/4-0002-rework-20211121.patch)
download | inline diff:
From ce26bbc1a0a568e0e2946afa962966099daca997 Mon Sep 17 00:00:00 2001
From: Tomas Vondra <[email protected]>
Date: Thu, 18 Nov 2021 19:48:30 +0100
Subject: [PATCH 2/4] rework
---
src/backend/access/rmgrdesc/logicalmsgdesc.c | 13 +
src/backend/access/transam/xact.c | 9 +
src/backend/commands/sequence.c | 236 ++++++++++++------
src/backend/replication/logical/decode.c | 104 +++-----
src/backend/replication/logical/message.c | 3 +-
.../replication/logical/reorderbuffer.c | 213 ++++------------
src/include/commands/sequence.h | 2 +
src/include/replication/message.h | 15 ++
src/include/replication/reorderbuffer.h | 10 +-
9 files changed, 288 insertions(+), 317 deletions(-)
diff --git a/src/backend/access/rmgrdesc/logicalmsgdesc.c b/src/backend/access/rmgrdesc/logicalmsgdesc.c
index d64ce2e7ef..c84628cd54 100644
--- a/src/backend/access/rmgrdesc/logicalmsgdesc.c
+++ b/src/backend/access/rmgrdesc/logicalmsgdesc.c
@@ -40,6 +40,16 @@ logicalmsg_desc(StringInfo buf, XLogReaderState *record)
sep = " ";
}
}
+ else if (info == XLOG_LOGICAL_SEQUENCE)
+ {
+ xl_logical_sequence *xlrec = (xl_logical_sequence *) rec;
+
+ appendStringInfo(buf, "rel %u/%u/%u last: %lu log_cnt: %lu is_called: %d",
+ xlrec->node.spcNode,
+ xlrec->node.dbNode,
+ xlrec->node.relNode,
+ xlrec->last, xlrec->log_cnt, xlrec->is_called);
+ }
}
const char *
@@ -48,5 +58,8 @@ logicalmsg_identify(uint8 info)
if ((info & ~XLR_INFO_MASK) == XLOG_LOGICAL_MESSAGE)
return "MESSAGE";
+ if ((info & ~XLR_INFO_MASK) == XLOG_LOGICAL_SEQUENCE)
+ return "SEQUENCE";
+
return NULL;
}
diff --git a/src/backend/access/transam/xact.c b/src/backend/access/transam/xact.c
index 8e35c432f5..eac8180250 100644
--- a/src/backend/access/transam/xact.c
+++ b/src/backend/access/transam/xact.c
@@ -36,6 +36,7 @@
#include "catalog/storage.h"
#include "commands/async.h"
#include "commands/tablecmds.h"
+#include "commands/sequence.h"
#include "commands/trigger.h"
#include "executor/spi.h"
#include "libpq/be-fsstubs.h"
@@ -2220,6 +2221,9 @@ CommitTransaction(void)
/* Prevent cancel/die interrupt while cleaning up */
HOLD_INTERRUPTS();
+ /* Write state of sequences to WAL */
+ AtEOXact_Sequences(true);
+
/* Commit updates to the relation map --- do this as late as possible */
AtEOXact_RelationMap(true, is_parallel_worker);
@@ -2378,6 +2382,8 @@ CommitTransaction(void)
* PrepareTransaction
*
* NB: if you change this routine, better look at CommitTransaction too!
+ *
+ * XXX Does this need to do something about logging of sequences?
*/
static void
PrepareTransaction(void)
@@ -2776,6 +2782,9 @@ AbortTransaction(void)
AtEOXact_RelationMap(false, is_parallel_worker);
AtAbort_Twophase();
+ /* XXX not sure we need to do anything about sequences at abort */
+ AtEOXact_Sequences(false);
+
/*
* Advertise the fact that we aborted in pg_xact (assuming that we got as
* far as assigning an XID to advertise). But if we're inside a parallel
diff --git a/src/backend/commands/sequence.c b/src/backend/commands/sequence.c
index a98fcc2e97..e45e4042a7 100644
--- a/src/backend/commands/sequence.c
+++ b/src/backend/commands/sequence.c
@@ -37,6 +37,7 @@
#include "miscadmin.h"
#include "nodes/makefuncs.h"
#include "parser/parse_type.h"
+#include "replication/message.h"
#include "storage/lmgr.h"
#include "storage/proc.h"
#include "storage/smgr.h"
@@ -75,6 +76,7 @@ typedef struct SeqTableData
{
Oid relid; /* pg_class OID of this sequence (hash key) */
Oid filenode; /* last seen relfilenode of this sequence */
+ Oid tablespace; /* last seen tablespace of this sequence */
LocalTransactionId lxid; /* xact in which we last did a seq op */
bool last_valid; /* do we have a valid "last" value? */
int64 last; /* value last returned by nextval */
@@ -82,6 +84,11 @@ typedef struct SeqTableData
/* if last != cached, we have not used up all the cached values */
int64 increment; /* copy of sequence's increment field */
/* note that increment is zero until we first do nextval_internal() */
+ bool need_log; /* should be written to WAL at commit? */
+ bool touched;
+ int64 last_value;
+ int64 log_cnt;
+ bool is_called;
} SeqTableData;
typedef SeqTableData *SeqTable;
@@ -94,7 +101,7 @@ static HTAB *seqhashtab = NULL; /* hash table for SeqTable items */
*/
static SeqTableData *last_used_seq = NULL;
-static void fill_seq_with_data(Relation rel, HeapTuple tuple, bool create);
+static void fill_seq_with_data(Relation rel, HeapTuple tuple);
static Relation lock_and_open_sequence(SeqTable seq);
static void create_seq_hashtable(void);
static void init_sequence(Oid relid, SeqTable *p_elm, Relation *p_rel);
@@ -222,7 +229,7 @@ DefineSequence(ParseState *pstate, CreateSeqStmt *seq)
/* now initialize the sequence's data */
tuple = heap_form_tuple(tupDesc, value, null);
- fill_seq_with_data(rel, tuple, true);
+ fill_seq_with_data(rel, tuple);
/* process OWNED BY if given */
if (owned_by)
@@ -230,6 +237,31 @@ DefineSequence(ParseState *pstate, CreateSeqStmt *seq)
table_close(rel, NoLock);
+ /*
+ * Add the new sequence to local cache, so that we log it at commit.
+ */
+ {
+ Relation seqrel;
+ SeqTable elm;
+ Buffer buf;
+ Form_pg_sequence_data seq;
+ HeapTupleData seqdatatuple;
+
+ init_sequence(seqoid, &elm, &seqrel);
+
+ seq = read_seq_tuple(seqrel, &buf, &seqdatatuple);
+
+ elm->last_value = seq->last_value;
+ elm->is_called = seq->is_called;
+ elm->log_cnt = 0;
+ elm->touched = true;
+ elm->need_log = true;
+
+ relation_close(seqrel, NoLock);
+
+ UnlockReleaseBuffer(buf);
+ }
+
/* fill in pg_sequence */
rel = table_open(SequenceRelationId, RowExclusiveLock);
tupDesc = RelationGetDescr(rel);
@@ -327,12 +359,18 @@ ResetSequence(Oid seq_relid)
/*
* Insert the modified tuple into the new storage file.
*/
- fill_seq_with_data(seq_rel, tuple, true);
+ fill_seq_with_data(seq_rel, tuple);
/* Clear local cache so that we don't think we have cached numbers */
/* Note that we do not change the currval() state */
elm->cached = elm->last;
+ /* Remember we need to write the sequence to WAL at commit. */
+ elm->need_log = true;
+ elm->is_called = false;
+ elm->log_cnt = 0;
+ elm->touched = true;
+
relation_close(seq_rel, NoLock);
}
@@ -340,7 +378,7 @@ ResetSequence(Oid seq_relid)
* Initialize a sequence's relation with the specified tuple as content
*/
static void
-fill_seq_with_data(Relation rel, HeapTuple tuple, bool create)
+fill_seq_with_data(Relation rel, HeapTuple tuple)
{
Buffer buf;
Page page;
@@ -380,27 +418,7 @@ fill_seq_with_data(Relation rel, HeapTuple tuple, bool create)
if (RelationNeedsWAL(rel))
{
GetTopTransactionId();
-
- /*
- * Ensure we have a proper XID, which will be included in the XLOG
- * record by XLogRecordAssemble. Otherwise the first nextval() in
- * a subxact (without any preceding changes) would get XID 0,
- * and it'd be impossible to decide which top xact it belongs to.
- * It'd also trigger assert in DecodeSequence.
- *
- * XXX This might seem unnecessary, because if there's no XID the
- * transaction couldn't have done anything important yet, e.g. it
- * could not have created a sequence. But that's incorrect, as it
- * ignores subtransactions. The current subtransaction might not
- * have done anything yet (thus no XID), but an earlier one might
- * have created the sequence.
- *
- * XXX Not sure if this is the best solution. Maybe do this only
- * with wal_level=logical to minimize the overhead. OTOH advancing
- * the sequence is likely followed by using the value(s) for some
- * other activity, which assigns the XID.
- */
- GetCurrentTransactionId();
+ // GetCurrentTransactionId();
}
START_CRIT_SECTION();
@@ -422,7 +440,6 @@ fill_seq_with_data(Relation rel, HeapTuple tuple, bool create)
XLogRegisterBuffer(0, buf, REGBUF_WILL_INIT);
xlrec.node = rel->rd_node;
- xlrec.created = create;
XLogRegisterData((char *) &xlrec, sizeof(xl_seq_rec));
XLogRegisterData((char *) tuple->t_data, tuple->t_len);
@@ -526,7 +543,11 @@ AlterSequence(ParseState *pstate, AlterSeqStmt *stmt)
/*
* Insert the modified tuple into the new storage file.
*/
- fill_seq_with_data(seqrel, newdatatuple, true);
+ fill_seq_with_data(seqrel, newdatatuple);
+
+ elm->need_log = true;
+ elm->is_called = false;
+ elm->log_cnt = 0;
}
/* process OWNED BY if given */
@@ -626,6 +647,8 @@ nextval_internal(Oid relid, bool check_permissions)
/* open and lock sequence */
init_sequence(relid, &elm, &seqrel);
+ elm->touched = true;
+
if (check_permissions &&
pg_class_aclcheck(elm->relid, GetUserId(),
ACL_USAGE | ACL_UPDATE) != ACLCHECK_OK)
@@ -792,27 +815,7 @@ nextval_internal(Oid relid, bool check_permissions)
if (logit && RelationNeedsWAL(seqrel))
{
GetTopTransactionId();
-
- /*
- * Ensure we have a proper XID, which will be included in the XLOG
- * record by XLogRecordAssemble. Otherwise the first nextval() in
- * a subxact (without any preceding changes) would get XID 0,
- * and it'd be impossible to decide which top xact it belongs to.
- * It'd also trigger assert in DecodeSequence.
- *
- * XXX This might seem unnecessary, because if there's no XID the
- * transaction couldn't have done anything important yet, e.g. it
- * could not have created a sequence. But that's incorrect, as it
- * ignores subtransactions. The current subtransaction might not
- * have done anything yet (thus no XID), but an earlier one might
- * have created the sequence.
- *
- * XXX Not sure if this is the best solution. Maybe do this only
- * with wal_level=logical to minimize the overhead. OTOH advancing
- * the sequence is likely followed by using the value(s) for some
- * other activity, which assigns the XID.
- */
- GetCurrentTransactionId();
+ // GetCurrentTransactionId();
}
/* ready to change the on-disk (or really, in-buffer) tuple */
@@ -850,7 +853,6 @@ nextval_internal(Oid relid, bool check_permissions)
seq->log_cnt = 0;
xlrec.node = seqrel->rd_node;
- xlrec.created = false;
XLogRegisterData((char *) &xlrec, sizeof(xl_seq_rec));
XLogRegisterData((char *) seqdatatuple.t_data, seqdatatuple.t_len);
@@ -858,6 +860,11 @@ nextval_internal(Oid relid, bool check_permissions)
recptr = XLogInsert(RM_SEQ_ID, XLOG_SEQ_LOG);
PageSetLSN(page, recptr);
+
+ elm->need_log = true;
+ elm->is_called = true;
+ elm->last_value = next;
+ elm->log_cnt = 0;
}
/* Now update sequence tuple to the intended final state */
@@ -969,6 +976,8 @@ do_setval(Oid relid, int64 next, bool iscalled)
/* open and lock sequence */
init_sequence(relid, &elm, &seqrel);
+ elm->touched = true;
+
if (pg_class_aclcheck(elm->relid, GetUserId(), ACL_UPDATE) != ACLCHECK_OK)
ereport(ERROR,
(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
@@ -1027,27 +1036,7 @@ do_setval(Oid relid, int64 next, bool iscalled)
if (RelationNeedsWAL(seqrel))
{
GetTopTransactionId();
-
- /*
- * Ensure we have a proper XID, which will be included in the XLOG
- * record by XLogRecordAssemble. Otherwise the first nextval() in
- * a subxact (without any preceding changes) would get XID 0,
- * and it'd be impossible to decide which top xact it belongs to.
- * It'd also trigger assert in DecodeSequence.
- *
- * XXX This might seem unnecessary, because if there's no XID the
- * transaction couldn't have done anything important yet, e.g. it
- * could not have created a sequence. But that's incorrect, as it
- * ignores subtransactions. The current subtransaction might not
- * have done anything yet (thus no XID), but an earlier one might
- * have created the sequence.
- *
- * XXX Not sure if this is the best solution. Maybe do this only
- * with wal_level=logical to minimize the overhead. OTOH advancing
- * the sequence is likely followed by using the value(s) for some
- * other activity, which assigns the XID.
- */
- GetCurrentTransactionId();
+ // GetCurrentTransactionId();
}
/* ready to change the on-disk (or really, in-buffer) tuple */
@@ -1070,14 +1059,17 @@ do_setval(Oid relid, int64 next, bool iscalled)
XLogRegisterBuffer(0, buf, REGBUF_WILL_INIT);
xlrec.node = seqrel->rd_node;
- xlrec.created = false;
-
XLogRegisterData((char *) &xlrec, sizeof(xl_seq_rec));
XLogRegisterData((char *) seqdatatuple.t_data, seqdatatuple.t_len);
recptr = XLogInsert(RM_SEQ_ID, XLOG_SEQ_LOG);
PageSetLSN(page, recptr);
+
+ elm->need_log = true;
+ elm->is_called = false;
+ elm->last_value = next;
+ elm->log_cnt = 0;
}
END_CRIT_SECTION();
@@ -1196,6 +1188,7 @@ init_sequence(Oid relid, SeqTable *p_elm, Relation *p_rel)
{
/* relid already filled in */
elm->filenode = InvalidOid;
+ elm->tablespace = InvalidOid;
elm->lxid = InvalidLocalTransactionId;
elm->last_valid = false;
elm->last = elm->cached = 0;
@@ -1219,6 +1212,7 @@ init_sequence(Oid relid, SeqTable *p_elm, Relation *p_rel)
*/
if (seqrel->rd_rel->relfilenode != elm->filenode)
{
+ elm->tablespace = seqrel->rd_rel->reltablespace;
elm->filenode = seqrel->rd_rel->relfilenode;
elm->cached = elm->last;
}
@@ -1977,6 +1971,8 @@ seq_redo(XLogReaderState *record)
/*
* Flush cached sequence information.
+ *
+ * XXX Do we need to WAL-log the entries based on need_log?
*/
void
ResetSequenceCaches(void)
@@ -2000,3 +1996,99 @@ seq_mask(char *page, BlockNumber blkno)
mask_unused_space(page);
}
+
+
+static void
+read_sequence_info(Relation seqrel, int64 *last_value, int64 *log_cnt, bool *is_called)
+{
+ Buffer buf;
+ Form_pg_sequence_data seq;
+ HeapTupleData seqdatatuple;
+
+ seq = read_seq_tuple(seqrel, &buf, &seqdatatuple);
+
+ *last_value = seq->last_value;
+ *is_called = seq->is_called;
+ *log_cnt = seq->log_cnt;
+
+ UnlockReleaseBuffer(buf);
+}
+
+/* XXX Do this only for wal_level = logical, probably? */
+void
+AtEOXact_Sequences(bool isCommit)
+{
+ SeqTable entry;
+ HASH_SEQ_STATUS scan;
+
+ if (!seqhashtab)
+ return;
+
+ /*
+ * XXX If we run just nextval() that returns cached value, the xact may
+ * not have XID, but we expect it in reorderbuffer. Maybe treat that as
+ * transactional?
+ */
+ // GetTopTransactionId();
+ // GetCurrentTransactionId();
+
+ hash_seq_init(&scan, seqhashtab);
+
+ while ((entry = (SeqTable) hash_seq_search(&scan)))
+ {
+ Relation rel;
+ RelFileNode rnode;
+ xl_logical_sequence xlrec;
+
+ if (!isCommit)
+ entry->touched = false;
+
+ /*
+ * If not touched in the current transaction, don't log anything.
+ * We leave needs_log set, so that if future transactions touch
+ * the sequence we'll log it properly.
+ */
+ if (!entry->touched)
+ continue;
+
+ /*
+ * Likewise, if the sequence does not need logging, we're done.
+ */
+ if (!entry->need_log)
+ continue;
+
+ /* does the relation still exist? */
+ rel = try_relation_open(entry->relid, NoLock);
+
+ /* relation might have been dropped */
+ if (!rel)
+ continue;
+
+ /* if this is commit, we'll log the */
+ entry->need_log = false;
+ entry->touched = false;
+
+ rnode.spcNode = entry->tablespace; /* tablespace */
+ rnode.dbNode = MyDatabaseId; /* database */
+ rnode.relNode = entry->filenode; /* relation */
+
+ xlrec.node = rnode;
+ xlrec.reloid = entry->relid;
+
+ /* XXX is it good enough to log values we have in cache? seems
+ * wrong and we may need to re-read that. */
+ // xlrec.dbId = MyDatabaseId;
+ // xlrec.relId = entry->relid;
+ read_sequence_info(rel, &xlrec.last, &xlrec.log_cnt, &xlrec.is_called);
+
+ XLogBeginInsert();
+ XLogRegisterData((char *) &xlrec, SizeOfLogicalSequence);
+
+ /* allow origin filtering */
+ XLogSetRecordFlags(XLOG_INCLUDE_ORIGIN);
+
+ (void) XLogInsert(RM_LOGICALMSG_ID, XLOG_LOGICAL_SEQUENCE);
+
+ relation_close(rel, NoLock);
+ }
+}
diff --git a/src/backend/replication/logical/decode.c b/src/backend/replication/logical/decode.c
index b61f2e264a..d29844adf9 100644
--- a/src/backend/replication/logical/decode.c
+++ b/src/backend/replication/logical/decode.c
@@ -59,6 +59,9 @@ static void DecodeXactOp(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
static void DecodeStandbyOp(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
static void DecodeLogicalMsgOp(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
+static void DecodeLogicalMessage(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
+static void DecodeLogicalSequence(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
+
/* individual record(group)'s handlers */
static void DecodeInsert(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
static void DecodeUpdate(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
@@ -75,11 +78,9 @@ static void DecodeAbort(LogicalDecodingContext *ctx, XLogRecordBuffer *buf,
bool two_phase);
static void DecodePrepare(LogicalDecodingContext *ctx, XLogRecordBuffer *buf,
xl_xact_parsed_prepare *parsed);
-static void DecodeSequence(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
/* common function to decode tuples */
static void DecodeXLogTuple(char *data, Size len, ReorderBufferTupleBuf *tup);
-static void DecodeSeqTuple(char *data, Size len, ReorderBufferTupleBuf *tuple);
/* helper functions for decoding transactions */
static inline bool FilterPrepare(LogicalDecodingContext *ctx,
@@ -160,10 +161,6 @@ LogicalDecodingProcessRecord(LogicalDecodingContext *ctx, XLogReaderState *recor
DecodeLogicalMsgOp(ctx, &buf);
break;
- case RM_SEQ_ID:
- DecodeSequence(ctx, &buf);
- break;
-
/*
* Rmgrs irrelevant for logical decoding; they describe stuff not
* represented in logical decoding. Add new rmgrs in rmgrlist.h's
@@ -626,6 +623,27 @@ FilterByOrigin(LogicalDecodingContext *ctx, RepOriginId origin_id)
*/
static void
DecodeLogicalMsgOp(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
+{
+ XLogReaderState *r = buf->record;
+ uint8 info = XLogRecGetInfo(r) & ~XLR_INFO_MASK;
+
+ switch (info)
+ {
+ case XLOG_LOGICAL_MESSAGE:
+ DecodeLogicalMessage(ctx, buf);
+ break;
+
+ case XLOG_LOGICAL_SEQUENCE:
+ DecodeLogicalSequence(ctx, buf);
+ break;
+
+ default:
+ elog(ERROR, "unexpected RM_LOGICALMSG_ID record type: %u", info);
+ }
+}
+
+static void
+DecodeLogicalMessage(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
{
SnapBuild *builder = ctx->snapshot_builder;
XLogReaderState *r = buf->record;
@@ -1319,31 +1337,6 @@ DecodeTXNNeedSkip(LogicalDecodingContext *ctx, XLogRecordBuffer *buf,
ctx->fast_forward || FilterByOrigin(ctx, origin_id));
}
-/*
- * Decode Sequence Tuple
- */
-static void
-DecodeSeqTuple(char *data, Size len, ReorderBufferTupleBuf *tuple)
-{
- int datalen = len - sizeof(xl_seq_rec) - SizeofHeapTupleHeader;
-
- Assert(datalen >= 0);
-
- tuple->tuple.t_len = datalen + SizeofHeapTupleHeader;;
-
- ItemPointerSetInvalid(&tuple->tuple.t_self);
-
- tuple->tuple.t_tableOid = InvalidOid;
-
- memcpy(((char *) tuple->tuple.t_data),
- data + sizeof(xl_seq_rec),
- SizeofHeapTupleHeader);
-
- memcpy(((char *) tuple->tuple.t_data) + SizeofHeapTupleHeader,
- data + sizeof(xl_seq_rec) + SizeofHeapTupleHeader,
- datalen);
-}
-
/*
* Handle sequence decode
*
@@ -1362,24 +1355,20 @@ DecodeSeqTuple(char *data, Size len, ReorderBufferTupleBuf *tuple)
* plugin - it might get confused about which sequence it's related to etc.
*/
static void
-DecodeSequence(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
+DecodeLogicalSequence(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
{
SnapBuild *builder = ctx->snapshot_builder;
- ReorderBufferTupleBuf *tuplebuf;
RelFileNode target_node;
XLogReaderState *r = buf->record;
- char *tupledata = NULL;
- Size tuplelen;
- Size datalen = 0;
TransactionId xid = XLogRecGetXid(r);
uint8 info = XLogRecGetInfo(buf->record) & ~XLR_INFO_MASK;
- xl_seq_rec *xlrec;
+ xl_logical_sequence *xlrec;
Snapshot snapshot;
RepOriginId origin_id = XLogRecGetOrigin(r);
- bool transactional;
+ bool transactional = TransactionIdIsValid(xid);
/* only decode changes flagged with XLOG_SEQ_LOG */
- if (info != XLOG_SEQ_LOG)
+ if (info != XLOG_LOGICAL_SEQUENCE)
elog(ERROR, "unexpected RM_SEQ_ID record type: %u", info);
/*
@@ -1390,8 +1379,11 @@ DecodeSequence(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
ctx->fast_forward)
return;
+ /* extract the WAL record, with "created" flag */
+ xlrec = (xl_logical_sequence *) XLogRecGetData(r);
+
/* only interested in our database */
- XLogRecGetBlockTag(r, 0, &target_node, NULL, NULL);
+ target_node = xlrec->node;
if (target_node.dbNode != ctx->slot->data.database)
return;
@@ -1399,44 +1391,18 @@ DecodeSequence(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
if (FilterByOrigin(ctx, XLogRecGetOrigin(r)))
return;
- tupledata = XLogRecGetData(r);
- datalen = XLogRecGetDataLen(r);
- tuplelen = datalen - SizeOfHeapHeader - sizeof(xl_seq_rec);
-
- /* extract the WAL record, with "created" flag */
- xlrec = (xl_seq_rec *) XLogRecGetData(r);
-
- /* XXX how could we have sequence change without data? */
- if(!datalen || !tupledata)
- return;
-
- tuplebuf = ReorderBufferGetTupleBuf(ctx->reorder, tuplelen);
- DecodeSeqTuple(tupledata, datalen, tuplebuf);
-
- /*
- * Should we handle the sequence increment as transactional or not?
- *
- * If the sequence was created in a still-running transaction, treat
- * it as transactional and queue the increments. Otherwise it needs
- * to be treated as non-transactional, in which case we send it to
- * the plugin right away.
- */
- transactional = ReorderBufferSequenceIsTransactional(ctx->reorder,
- target_node,
- xlrec->created);
-
/* Skip the change if already processed (per the snapshot). */
if (transactional &&
!SnapBuildProcessChange(builder, xid, buf->origptr))
return;
else if (!transactional &&
(SnapBuildCurrentState(builder) != SNAPBUILD_CONSISTENT ||
- SnapBuildXactNeedsSkip(builder, buf->origptr)))
+ SnapBuildXactNeedsSkip(builder, buf->origptr)))
return;
/* Queue the increment (or send immediately if not transactional). */
snapshot = SnapBuildGetOrBuildSnapshot(builder, xid);
ReorderBufferQueueSequence(ctx->reorder, xid, snapshot, buf->endptr,
- origin_id, target_node, transactional,
- xlrec->created, tuplebuf);
+ origin_id, xlrec->reloid, target_node,
+ xlrec->last, xlrec->log_cnt, xlrec->is_called);
}
diff --git a/src/backend/replication/logical/message.c b/src/backend/replication/logical/message.c
index 93bd372421..3f0ac57fc6 100644
--- a/src/backend/replication/logical/message.c
+++ b/src/backend/replication/logical/message.c
@@ -81,7 +81,8 @@ logicalmsg_redo(XLogReaderState *record)
{
uint8 info = XLogRecGetInfo(record) & ~XLR_INFO_MASK;
- if (info != XLOG_LOGICAL_MESSAGE)
+ if (info != XLOG_LOGICAL_MESSAGE &&
+ info != XLOG_LOGICAL_SEQUENCE)
elog(PANIC, "logicalmsg_redo: unknown op code %u", info);
/* This is only interesting for logical decoding, see decode.c. */
diff --git a/src/backend/replication/logical/reorderbuffer.c b/src/backend/replication/logical/reorderbuffer.c
index 434926459f..1d9a39ac4b 100644
--- a/src/backend/replication/logical/reorderbuffer.c
+++ b/src/backend/replication/logical/reorderbuffer.c
@@ -569,17 +569,11 @@ ReorderBufferReturnChange(ReorderBuffer *rb, ReorderBufferChange *change,
change->data.truncate.relids = NULL;
}
break;
- case REORDER_BUFFER_CHANGE_SEQUENCE:
- if (change->data.sequence.tuple)
- {
- ReorderBufferReturnTupleBuf(rb, change->data.sequence.tuple);
- change->data.sequence.tuple = NULL;
- }
- break;
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM:
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_ABORT:
case REORDER_BUFFER_CHANGE_INTERNAL_COMMAND_ID:
case REORDER_BUFFER_CHANGE_INTERNAL_TUPLECID:
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
break;
}
@@ -973,9 +967,10 @@ ReorderBufferSequenceCleanup(ReorderBuffer *rb, TransactionId xid)
void
ReorderBufferQueueSequence(ReorderBuffer *rb, TransactionId xid,
Snapshot snapshot, XLogRecPtr lsn, RepOriginId origin_id,
- RelFileNode rnode, bool transactional, bool created,
- ReorderBufferTupleBuf *tuplebuf)
+ Oid reloid, RelFileNode rnode, int64 last, int64 log_cnt, bool is_called)
{
+ bool transactional = TransactionIdIsValid(xid);
+
/*
* Change needs to be handled as transactional, because the sequence was
* created in a transaction that is still running. In that case all the
@@ -992,38 +987,6 @@ ReorderBufferQueueSequence(ReorderBuffer *rb, TransactionId xid,
MemoryContext oldcontext;
ReorderBufferChange *change;
- /* lookup sequence by relfilenode */
- ReorderBufferSequenceEnt *ent;
- bool found;
-
- /* transactional changes require a transaction */
- Assert(xid != InvalidTransactionId);
-
- /* search the lookup table (we ignore the return value, found is enough) */
- ent = hash_search(rb->sequences,
- (void *) &rnode,
- created ? HASH_ENTER : HASH_FIND,
- &found);
-
- /*
- * If this is the "create" increment, we must not have found any
- * pre-existing entry in the hash table (i.e. there must not be
- * any conflicting sequence).
- */
- Assert(!(created && found));
-
- /* But we must have either created or found an existing entry. */
- Assert(created || found);
-
- /*
- * When creating the sequence, remember the XID of the transaction
- * that created id.
- */
- if (created)
- ent->xid = xid;
-
- /* XXX Maybe check that we're still in the same top-level xact? */
-
/* OK, allocate and queue the change */
oldcontext = MemoryContextSwitchTo(rb->context);
@@ -1034,38 +997,26 @@ ReorderBufferQueueSequence(ReorderBuffer *rb, TransactionId xid,
memcpy(&change->data.sequence.relnode, &rnode, sizeof(RelFileNode));
- change->data.sequence.created = created;
- change->data.sequence.tuple = tuplebuf;
+ change->data.sequence.reloid = reloid;
+ change->data.sequence.last = last;
+ change->data.sequence.log_cnt = log_cnt;
+ change->data.sequence.is_called = is_called;
/* add it to the same subxact that created the sequence */
- ReorderBufferQueueChange(rb, ent->xid, lsn, change, false);
+ ReorderBufferQueueChange(rb, xid, lsn, change, false);
MemoryContextSwitchTo(oldcontext);
}
else
{
/*
- * This increment is for a sequence that was not created in any
- * running transaction, so we treat it as non-transactional and
- * just send it to the output plugin directly.
- */
+ * This increment is for a sequence that was not created in any
+ * running transaction, so we treat it as non-transactional and
+ * just send it to the output plugin directly.
+ */
ReorderBufferTXN *txn = NULL;
volatile Snapshot snapshot_now = snapshot;
- bool using_subtxn;
-
-#ifdef USE_ASSERT_CHECKING
- /* All "creates" have to be handled as transactional. */
- Assert(!created);
-
- /* Make sure the sequence is not in the hash table. */
- {
- bool found;
- hash_search(rb->sequences,
- (void *) &rnode,
- HASH_FIND, &found);
- Assert(!found);
- }
-#endif
+ bool using_subtxn;
if (xid != InvalidTransactionId)
txn = ReorderBufferTXNByXid(rb, xid, true, NULL, lsn, true);
@@ -1074,40 +1025,35 @@ ReorderBufferQueueSequence(ReorderBuffer *rb, TransactionId xid,
SetupHistoricSnapshot(snapshot_now, NULL);
/*
- * Decoding needs access to syscaches et al., which in turn use
- * heavyweight locks and such. Thus we need to have enough state around to
- * keep track of those. The easiest way is to simply use a transaction
- * internally. That also allows us to easily enforce that nothing writes
- * to the database by checking for xid assignments.
- *
- * When we're called via the SQL SRF there's already a transaction
- * started, so start an explicit subtransaction there.
- */
+ * Decoding needs access to syscaches et al., which in turn use
+ * heavyweight locks and such. Thus we need to have enough state around to
+ * keep track of those. The easiest way is to simply use a transaction
+ * internally. That also allows us to easily enforce that nothing writes
+ * to the database by checking for xid assignments.
+ *
+ * When we're called via the SQL SRF there's already a transaction
+ * started, so start an explicit subtransaction there.
+ */
using_subtxn = IsTransactionOrTransactionBlock();
PG_TRY();
{
Relation relation;
- HeapTuple tuple;
- bool isnull;
- int64 last_value, log_cnt;
- bool is_called;
- Oid reloid;
+// Oid reloid;
if (using_subtxn)
BeginInternalSubTransaction("sequence");
else
StartTransactionCommand();
-
+/*
reloid = RelidByRelfilenode(rnode.spcNode, rnode.relNode);
if (reloid == InvalidOid)
elog(ERROR, "could not map filenode \"%s\" to relation OID",
- relpathperm(rnode,
- MAIN_FORKNUM));
-
+ relpathperm(rnode,
+ MAIN_FORKNUM));
+*/
relation = RelationIdGetRelation(reloid);
- tuple = &tuplebuf->tuple;
/*
* Extract the internal sequence values, describing the state.
@@ -1115,12 +1061,8 @@ ReorderBufferQueueSequence(ReorderBuffer *rb, TransactionId xid,
* XXX Seems a bit strange to access it directly. Maybe there's
* a better / more correct way?
*/
- last_value = heap_getattr(tuple, 1, RelationGetDescr(relation), &isnull);
- log_cnt = heap_getattr(tuple, 2, RelationGetDescr(relation), &isnull);
- is_called = heap_getattr(tuple, 3, RelationGetDescr(relation), &isnull);
-
- rb->sequence(rb, txn, lsn, relation, transactional, created,
- last_value, log_cnt, is_called);
+ rb->sequence(rb, txn, lsn, relation, transactional, false,
+ last, log_cnt, is_called);
RelationClose(relation);
@@ -2248,31 +2190,27 @@ ReorderBufferApplySequence(ReorderBuffer *rb, ReorderBufferTXN *txn,
Relation relation, ReorderBufferChange *change,
bool streaming)
{
- HeapTuple tuple;
- bool isnull;
int64 last_value, log_cnt;
bool is_called;
- tuple = &change->data.sequence.tuple->tuple;
-
/*
* Extract the internal sequence values, describing the state.
*
* XXX Seems a bit strange to access it directly. Maybe there's
* a better / more correct way?
*/
- last_value = heap_getattr(tuple, 1, RelationGetDescr(relation), &isnull);
- log_cnt = heap_getattr(tuple, 2, RelationGetDescr(relation), &isnull);
- is_called = heap_getattr(tuple, 3, RelationGetDescr(relation), &isnull);
+ last_value = change->data.sequence.last;
+ log_cnt = change->data.sequence.log_cnt;
+ is_called = change->data.sequence.is_called;
/* Only ever called from ReorderBufferApplySequence, so transational. */
if (streaming)
rb->stream_sequence(rb, txn, change->lsn, relation, true,
- change->data.sequence.created,
+ false, /* FIXME */
last_value, log_cnt, is_called);
else
rb->sequence(rb, txn, change->lsn, relation, true,
- change->data.sequence.created,
+ false, /* FIXME */
last_value, log_cnt, is_called);
}
@@ -2722,15 +2660,17 @@ ReorderBufferProcessTXN(ReorderBuffer *rb, ReorderBufferTXN *txn,
case REORDER_BUFFER_CHANGE_SEQUENCE:
Assert(snapshot_now);
+ /*
reloid = RelidByRelfilenode(change->data.sequence.relnode.spcNode,
change->data.sequence.relnode.relNode);
if (reloid == InvalidOid)
- elog(ERROR, "could not map filenode \"%s\" to relation OID",
+ elog(ERROR, "zz could not map filenode \"%s\" to relation OID",
relpathperm(change->data.sequence.relnode,
MAIN_FORKNUM));
- relation = RelationIdGetRelation(reloid);
+ */
+ relation = RelationIdGetRelation(change->data.sequence.reloid);
if (!RelationIsValid(relation))
elog(ERROR, "could not open relation with OID %u (for filenode \"%s\")",
@@ -4127,45 +4067,13 @@ ReorderBufferSerializeChange(ReorderBuffer *rb, ReorderBufferTXN *txn,
memcpy(data, change->data.truncate.relids, size);
data += size;
- break;
- }
- case REORDER_BUFFER_CHANGE_SEQUENCE:
- {
- char *data;
- ReorderBufferTupleBuf *tup;
- Size len = 0;
-
- tup = change->data.sequence.tuple;
-
- if (tup)
- {
- sz += sizeof(HeapTupleData);
- len = tup->tuple.t_len;
- sz += len;
- }
-
- /* make sure we have enough space */
- ReorderBufferSerializeReserve(rb, sz);
-
- data = ((char *) rb->outbuf) + sizeof(ReorderBufferDiskChange);
- /* might have been reallocated above */
- ondisk = (ReorderBufferDiskChange *) rb->outbuf;
-
- if (len)
- {
- memcpy(data, &tup->tuple, sizeof(HeapTupleData));
- data += sizeof(HeapTupleData);
-
- memcpy(data, tup->tuple.t_data, len);
- data += len;
- }
-
break;
}
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM:
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_ABORT:
case REORDER_BUFFER_CHANGE_INTERNAL_COMMAND_ID:
case REORDER_BUFFER_CHANGE_INTERNAL_TUPLECID:
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
/* ReorderBufferChange contains everything important */
break;
}
@@ -4424,28 +4332,13 @@ ReorderBufferChangeSize(ReorderBufferChange *change)
{
sz += sizeof(Oid) * change->data.truncate.nrelids;
- break;
- }
- case REORDER_BUFFER_CHANGE_SEQUENCE:
- {
- ReorderBufferTupleBuf *tup;
- Size len = 0;
-
- tup = change->data.sequence.tuple;
-
- if (tup)
- {
- sz += sizeof(HeapTupleData);
- len = tup->tuple.t_len;
- sz += len;
- }
-
break;
}
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM:
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_ABORT:
case REORDER_BUFFER_CHANGE_INTERNAL_COMMAND_ID:
case REORDER_BUFFER_CHANGE_INTERNAL_TUPLECID:
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
/* ReorderBufferChange contains everything important */
break;
}
@@ -4742,33 +4635,11 @@ ReorderBufferRestoreChange(ReorderBuffer *rb, ReorderBufferTXN *txn,
break;
}
- case REORDER_BUFFER_CHANGE_SEQUENCE:
- if (change->data.sequence.tuple)
- {
- uint32 tuplelen = ((HeapTuple) data)->t_len;
-
- change->data.sequence.tuple =
- ReorderBufferGetTupleBuf(rb, tuplelen - SizeofHeapTupleHeader);
-
- /* restore ->tuple */
- memcpy(&change->data.sequence.tuple->tuple, data,
- sizeof(HeapTupleData));
- data += sizeof(HeapTupleData);
-
- /* reset t_data pointer into the new tuplebuf */
- change->data.sequence.tuple->tuple.t_data =
- ReorderBufferTupleBufData(change->data.sequence.tuple);
-
- /* restore tuple data itself */
- memcpy(change->data.sequence.tuple->tuple.t_data, data, tuplelen);
- data += tuplelen;
- }
- break;
-
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM:
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_ABORT:
case REORDER_BUFFER_CHANGE_INTERNAL_COMMAND_ID:
case REORDER_BUFFER_CHANGE_INTERNAL_TUPLECID:
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
break;
}
diff --git a/src/include/commands/sequence.h b/src/include/commands/sequence.h
index 5919fb90ee..ddb36b5b17 100644
--- a/src/include/commands/sequence.h
+++ b/src/include/commands/sequence.h
@@ -62,6 +62,8 @@ extern void DeleteSequenceTuple(Oid relid);
extern void ResetSequence(Oid seq_relid);
extern void ResetSequenceCaches(void);
+extern void AtEOXact_Sequences(bool isCommit);
+
extern void seq_redo(XLogReaderState *rptr);
extern void seq_desc(StringInfo buf, XLogReaderState *rptr);
extern const char *seq_identify(uint8 info);
diff --git a/src/include/replication/message.h b/src/include/replication/message.h
index d3fb324c81..a39e8e984d 100644
--- a/src/include/replication/message.h
+++ b/src/include/replication/message.h
@@ -27,13 +27,28 @@ typedef struct xl_logical_message
char message[FLEXIBLE_ARRAY_MEMBER];
} xl_logical_message;
+/*
+ * Generic logical decoding sequence wal record.
+ */
+typedef struct xl_logical_sequence
+{
+ RelFileNode node;
+ Oid reloid;
+ int64 last; /* last value emitted for sequence */
+ int64 log_cnt; /* last value cached for sequence */
+ bool is_called;
+} xl_logical_sequence;
+
#define SizeOfLogicalMessage (offsetof(xl_logical_message, message))
+#define SizeOfLogicalSequence (sizeof(xl_logical_sequence))
extern XLogRecPtr LogLogicalMessage(const char *prefix, const char *message,
size_t size, bool transactional);
/* RMGR API*/
#define XLOG_LOGICAL_MESSAGE 0x00
+#define XLOG_LOGICAL_SEQUENCE 0x10
+
void logicalmsg_redo(XLogReaderState *record);
void logicalmsg_desc(StringInfo buf, XLogReaderState *record);
const char *logicalmsg_identify(uint8 info);
diff --git a/src/include/replication/reorderbuffer.h b/src/include/replication/reorderbuffer.h
index 63e6ed037b..eb78d5dff4 100644
--- a/src/include/replication/reorderbuffer.h
+++ b/src/include/replication/reorderbuffer.h
@@ -164,8 +164,10 @@ typedef struct ReorderBufferChange
struct
{
RelFileNode relnode;
- bool created;
- ReorderBufferTupleBuf *tuple;
+ Oid reloid;
+ int64 last;
+ int64 log_cnt;
+ bool is_called;
} sequence;
} data;
@@ -672,8 +674,8 @@ void ReorderBufferQueueMessage(ReorderBuffer *, TransactionId, Snapshot snapsho
Size message_size, const char *message);
void ReorderBufferQueueSequence(ReorderBuffer *rb, TransactionId xid,
Snapshot snapshot, XLogRecPtr lsn, RepOriginId origin_id,
- RelFileNode rnode, bool transactional, bool created,
- ReorderBufferTupleBuf *tuplebuf);
+ Oid reloid, RelFileNode rnode,
+ int64 last, int64 log_cnt, bool is_called);
void ReorderBufferCommit(ReorderBuffer *, TransactionId,
XLogRecPtr commit_lsn, XLogRecPtr end_lsn,
TimestampTz commit_time, RepOriginId origin_id, XLogRecPtr origin_lsn);
--
2.31.1
[text/x-patch] 0001-Logical-decoding-of-sequences-20211121.patch (43.5K, ../../[email protected]/5-0001-Logical-decoding-of-sequences-20211121.patch)
download | inline diff:
From c05f9777d843e7176b02c519444ed23650d39eb6 Mon Sep 17 00:00:00 2001
From: Tomas Vondra <[email protected]>
Date: Fri, 24 Sep 2021 00:41:33 +0200
Subject: [PATCH 1/4] Logical decoding of sequences
This adds the infrastructure for logical decoding of sequences. Support
for built-in logical replication and test_decoding is added separately.
When decoding sequences, we differentiate between a sequences created in
a (running) transaction, and sequences created in other (already
committed) transactions. Changes for sequences created in the same
top-level transaction are treated as "transactional" i.e. just like any
other change from that transaction (and discarded in case of a
rollback). Changes for sequences created earlier are treated as not
transactional - are processed immediately, as if performed outside any
transaction (and thus not rolled back).
This mixed behavior is necessary - sequences are non-transactional (e.g.
ROLLBACK does not undo the sequence increments). But for new sequences,
we need to handle them in a transactional way, because if we ever get
some DDL support, the sequence won't exist until the transaction gets
applied. So we need to ensure the increments don't happen until the
sequence gets created.
To differentiate which sequences are "old" and which were created in a
still-running transaction, we track sequences created in running
transactions in a hash table. Each sequence is identified by it's
relfilenode, and at transaction commit we discard all sequences created
in that particular transaction. For each sequence we track the XID of
the (sub)transaction that created it, and we cleanup the sequences for
each subtransaction when it completes (commit/rollback).
We don't use the XID to check if it's the same top-level transaction.
It's enough to know it was created in an in-progress transaction, and we
know it must be the current one because otherwise it wouldn't see the
sequence object.
The main changes in this patch are:
1) ensure WAL-logging of all necessary info for sequence advances
We need to be able to associate the advance with a XID, but until now
sequence advance might have XID 0 if it was the first thing the
transaction did. So we ensure the transaction has XID.
Note: This is needed because of subxacts. A XID 0 might still have the
sequence created in a different subxact of the same top-level xact.
Then there's a "created" flag added to the xl_seq_rec, which is
necessary to differentiate WAL for created/reset sequences.
2) decoding / queueing of sequences into ReorderBuffer
This is mostly copy-paste of existing code for other decoded events.
3) tracking sequences created in in-progress transactions
We use a simple hash table, indexed by relfilenode.
---
src/backend/commands/sequence.c | 83 +++-
src/backend/replication/logical/decode.c | 131 +++++-
src/backend/replication/logical/logical.c | 89 ++++
.../replication/logical/reorderbuffer.c | 424 ++++++++++++++++++
src/include/commands/sequence.h | 1 +
src/include/replication/output_plugin.h | 29 ++
src/include/replication/reorderbuffer.h | 44 +-
7 files changed, 793 insertions(+), 8 deletions(-)
diff --git a/src/backend/commands/sequence.c b/src/backend/commands/sequence.c
index 72bfdc07a4..a98fcc2e97 100644
--- a/src/backend/commands/sequence.c
+++ b/src/backend/commands/sequence.c
@@ -94,7 +94,7 @@ static HTAB *seqhashtab = NULL; /* hash table for SeqTable items */
*/
static SeqTableData *last_used_seq = NULL;
-static void fill_seq_with_data(Relation rel, HeapTuple tuple);
+static void fill_seq_with_data(Relation rel, HeapTuple tuple, bool create);
static Relation lock_and_open_sequence(SeqTable seq);
static void create_seq_hashtable(void);
static void init_sequence(Oid relid, SeqTable *p_elm, Relation *p_rel);
@@ -222,7 +222,7 @@ DefineSequence(ParseState *pstate, CreateSeqStmt *seq)
/* now initialize the sequence's data */
tuple = heap_form_tuple(tupDesc, value, null);
- fill_seq_with_data(rel, tuple);
+ fill_seq_with_data(rel, tuple, true);
/* process OWNED BY if given */
if (owned_by)
@@ -327,7 +327,7 @@ ResetSequence(Oid seq_relid)
/*
* Insert the modified tuple into the new storage file.
*/
- fill_seq_with_data(seq_rel, tuple);
+ fill_seq_with_data(seq_rel, tuple, true);
/* Clear local cache so that we don't think we have cached numbers */
/* Note that we do not change the currval() state */
@@ -340,7 +340,7 @@ ResetSequence(Oid seq_relid)
* Initialize a sequence's relation with the specified tuple as content
*/
static void
-fill_seq_with_data(Relation rel, HeapTuple tuple)
+fill_seq_with_data(Relation rel, HeapTuple tuple, bool create)
{
Buffer buf;
Page page;
@@ -378,8 +378,31 @@ fill_seq_with_data(Relation rel, HeapTuple tuple)
/* check the comment above nextval_internal()'s equivalent call. */
if (RelationNeedsWAL(rel))
+ {
GetTopTransactionId();
+ /*
+ * Ensure we have a proper XID, which will be included in the XLOG
+ * record by XLogRecordAssemble. Otherwise the first nextval() in
+ * a subxact (without any preceding changes) would get XID 0,
+ * and it'd be impossible to decide which top xact it belongs to.
+ * It'd also trigger assert in DecodeSequence.
+ *
+ * XXX This might seem unnecessary, because if there's no XID the
+ * transaction couldn't have done anything important yet, e.g. it
+ * could not have created a sequence. But that's incorrect, as it
+ * ignores subtransactions. The current subtransaction might not
+ * have done anything yet (thus no XID), but an earlier one might
+ * have created the sequence.
+ *
+ * XXX Not sure if this is the best solution. Maybe do this only
+ * with wal_level=logical to minimize the overhead. OTOH advancing
+ * the sequence is likely followed by using the value(s) for some
+ * other activity, which assigns the XID.
+ */
+ GetCurrentTransactionId();
+ }
+
START_CRIT_SECTION();
MarkBufferDirty(buf);
@@ -399,6 +422,7 @@ fill_seq_with_data(Relation rel, HeapTuple tuple)
XLogRegisterBuffer(0, buf, REGBUF_WILL_INIT);
xlrec.node = rel->rd_node;
+ xlrec.created = create;
XLogRegisterData((char *) &xlrec, sizeof(xl_seq_rec));
XLogRegisterData((char *) tuple->t_data, tuple->t_len);
@@ -502,7 +526,7 @@ AlterSequence(ParseState *pstate, AlterSeqStmt *stmt)
/*
* Insert the modified tuple into the new storage file.
*/
- fill_seq_with_data(seqrel, newdatatuple);
+ fill_seq_with_data(seqrel, newdatatuple, true);
}
/* process OWNED BY if given */
@@ -766,8 +790,31 @@ nextval_internal(Oid relid, bool check_permissions)
* (Have to do that here, so we're outside the critical section)
*/
if (logit && RelationNeedsWAL(seqrel))
+ {
GetTopTransactionId();
+ /*
+ * Ensure we have a proper XID, which will be included in the XLOG
+ * record by XLogRecordAssemble. Otherwise the first nextval() in
+ * a subxact (without any preceding changes) would get XID 0,
+ * and it'd be impossible to decide which top xact it belongs to.
+ * It'd also trigger assert in DecodeSequence.
+ *
+ * XXX This might seem unnecessary, because if there's no XID the
+ * transaction couldn't have done anything important yet, e.g. it
+ * could not have created a sequence. But that's incorrect, as it
+ * ignores subtransactions. The current subtransaction might not
+ * have done anything yet (thus no XID), but an earlier one might
+ * have created the sequence.
+ *
+ * XXX Not sure if this is the best solution. Maybe do this only
+ * with wal_level=logical to minimize the overhead. OTOH advancing
+ * the sequence is likely followed by using the value(s) for some
+ * other activity, which assigns the XID.
+ */
+ GetCurrentTransactionId();
+ }
+
/* ready to change the on-disk (or really, in-buffer) tuple */
START_CRIT_SECTION();
@@ -803,6 +850,7 @@ nextval_internal(Oid relid, bool check_permissions)
seq->log_cnt = 0;
xlrec.node = seqrel->rd_node;
+ xlrec.created = false;
XLogRegisterData((char *) &xlrec, sizeof(xl_seq_rec));
XLogRegisterData((char *) seqdatatuple.t_data, seqdatatuple.t_len);
@@ -977,8 +1025,31 @@ do_setval(Oid relid, int64 next, bool iscalled)
/* check the comment above nextval_internal()'s equivalent call. */
if (RelationNeedsWAL(seqrel))
+ {
GetTopTransactionId();
+ /*
+ * Ensure we have a proper XID, which will be included in the XLOG
+ * record by XLogRecordAssemble. Otherwise the first nextval() in
+ * a subxact (without any preceding changes) would get XID 0,
+ * and it'd be impossible to decide which top xact it belongs to.
+ * It'd also trigger assert in DecodeSequence.
+ *
+ * XXX This might seem unnecessary, because if there's no XID the
+ * transaction couldn't have done anything important yet, e.g. it
+ * could not have created a sequence. But that's incorrect, as it
+ * ignores subtransactions. The current subtransaction might not
+ * have done anything yet (thus no XID), but an earlier one might
+ * have created the sequence.
+ *
+ * XXX Not sure if this is the best solution. Maybe do this only
+ * with wal_level=logical to minimize the overhead. OTOH advancing
+ * the sequence is likely followed by using the value(s) for some
+ * other activity, which assigns the XID.
+ */
+ GetCurrentTransactionId();
+ }
+
/* ready to change the on-disk (or really, in-buffer) tuple */
START_CRIT_SECTION();
@@ -999,6 +1070,8 @@ do_setval(Oid relid, int64 next, bool iscalled)
XLogRegisterBuffer(0, buf, REGBUF_WILL_INIT);
xlrec.node = seqrel->rd_node;
+ xlrec.created = false;
+
XLogRegisterData((char *) &xlrec, sizeof(xl_seq_rec));
XLogRegisterData((char *) seqdatatuple.t_data, seqdatatuple.t_len);
diff --git a/src/backend/replication/logical/decode.c b/src/backend/replication/logical/decode.c
index a2b69511b4..b61f2e264a 100644
--- a/src/backend/replication/logical/decode.c
+++ b/src/backend/replication/logical/decode.c
@@ -42,6 +42,7 @@
#include "replication/reorderbuffer.h"
#include "replication/snapbuild.h"
#include "storage/standby.h"
+#include "commands/sequence.h"
typedef struct XLogRecordBuffer
{
@@ -74,10 +75,11 @@ static void DecodeAbort(LogicalDecodingContext *ctx, XLogRecordBuffer *buf,
bool two_phase);
static void DecodePrepare(LogicalDecodingContext *ctx, XLogRecordBuffer *buf,
xl_xact_parsed_prepare *parsed);
-
+static void DecodeSequence(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
/* common function to decode tuples */
static void DecodeXLogTuple(char *data, Size len, ReorderBufferTupleBuf *tup);
+static void DecodeSeqTuple(char *data, Size len, ReorderBufferTupleBuf *tuple);
/* helper functions for decoding transactions */
static inline bool FilterPrepare(LogicalDecodingContext *ctx,
@@ -158,6 +160,10 @@ LogicalDecodingProcessRecord(LogicalDecodingContext *ctx, XLogReaderState *recor
DecodeLogicalMsgOp(ctx, &buf);
break;
+ case RM_SEQ_ID:
+ DecodeSequence(ctx, &buf);
+ break;
+
/*
* Rmgrs irrelevant for logical decoding; they describe stuff not
* represented in logical decoding. Add new rmgrs in rmgrlist.h's
@@ -173,7 +179,6 @@ LogicalDecodingProcessRecord(LogicalDecodingContext *ctx, XLogReaderState *recor
case RM_HASH_ID:
case RM_GIN_ID:
case RM_GIST_ID:
- case RM_SEQ_ID:
case RM_SPGIST_ID:
case RM_BRIN_ID:
case RM_COMMIT_TS_ID:
@@ -1313,3 +1318,125 @@ DecodeTXNNeedSkip(LogicalDecodingContext *ctx, XLogRecordBuffer *buf,
(txn_dbid != InvalidOid && txn_dbid != ctx->slot->data.database) ||
ctx->fast_forward || FilterByOrigin(ctx, origin_id));
}
+
+/*
+ * Decode Sequence Tuple
+ */
+static void
+DecodeSeqTuple(char *data, Size len, ReorderBufferTupleBuf *tuple)
+{
+ int datalen = len - sizeof(xl_seq_rec) - SizeofHeapTupleHeader;
+
+ Assert(datalen >= 0);
+
+ tuple->tuple.t_len = datalen + SizeofHeapTupleHeader;;
+
+ ItemPointerSetInvalid(&tuple->tuple.t_self);
+
+ tuple->tuple.t_tableOid = InvalidOid;
+
+ memcpy(((char *) tuple->tuple.t_data),
+ data + sizeof(xl_seq_rec),
+ SizeofHeapTupleHeader);
+
+ memcpy(((char *) tuple->tuple.t_data) + SizeofHeapTupleHeader,
+ data + sizeof(xl_seq_rec) + SizeofHeapTupleHeader,
+ datalen);
+}
+
+/*
+ * Handle sequence decode
+ *
+ * Decoding sequences is a bit tricky, because while most sequence actions
+ * are non-transactional (not subject to rollback), some need to be handled
+ * as transactional.
+ *
+ * By default, a sequence increment is non-transactional - we must not queue
+ * it in a transaction as other changes, because the transaction might get
+ * rolled back and we'd discard the increment. The downstream would not be
+ * notified about the increment, which is wrong.
+ *
+ * On the other hand, the sequence may be created in a transaction. In this
+ * case we *should* queue the change as other changes in the transaction,
+ * because we don't want to send the increments for unknown sequence to the
+ * plugin - it might get confused about which sequence it's related to etc.
+ */
+static void
+DecodeSequence(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
+{
+ SnapBuild *builder = ctx->snapshot_builder;
+ ReorderBufferTupleBuf *tuplebuf;
+ RelFileNode target_node;
+ XLogReaderState *r = buf->record;
+ char *tupledata = NULL;
+ Size tuplelen;
+ Size datalen = 0;
+ TransactionId xid = XLogRecGetXid(r);
+ uint8 info = XLogRecGetInfo(buf->record) & ~XLR_INFO_MASK;
+ xl_seq_rec *xlrec;
+ Snapshot snapshot;
+ RepOriginId origin_id = XLogRecGetOrigin(r);
+ bool transactional;
+
+ /* only decode changes flagged with XLOG_SEQ_LOG */
+ if (info != XLOG_SEQ_LOG)
+ elog(ERROR, "unexpected RM_SEQ_ID record type: %u", info);
+
+ /*
+ * If we don't have snapshot or we are just fast-forwarding, there is no
+ * point in decoding messages.
+ */
+ if (SnapBuildCurrentState(builder) < SNAPBUILD_FULL_SNAPSHOT ||
+ ctx->fast_forward)
+ return;
+
+ /* only interested in our database */
+ XLogRecGetBlockTag(r, 0, &target_node, NULL, NULL);
+ if (target_node.dbNode != ctx->slot->data.database)
+ return;
+
+ /* output plugin doesn't look for this origin, no need to queue */
+ if (FilterByOrigin(ctx, XLogRecGetOrigin(r)))
+ return;
+
+ tupledata = XLogRecGetData(r);
+ datalen = XLogRecGetDataLen(r);
+ tuplelen = datalen - SizeOfHeapHeader - sizeof(xl_seq_rec);
+
+ /* extract the WAL record, with "created" flag */
+ xlrec = (xl_seq_rec *) XLogRecGetData(r);
+
+ /* XXX how could we have sequence change without data? */
+ if(!datalen || !tupledata)
+ return;
+
+ tuplebuf = ReorderBufferGetTupleBuf(ctx->reorder, tuplelen);
+ DecodeSeqTuple(tupledata, datalen, tuplebuf);
+
+ /*
+ * Should we handle the sequence increment as transactional or not?
+ *
+ * If the sequence was created in a still-running transaction, treat
+ * it as transactional and queue the increments. Otherwise it needs
+ * to be treated as non-transactional, in which case we send it to
+ * the plugin right away.
+ */
+ transactional = ReorderBufferSequenceIsTransactional(ctx->reorder,
+ target_node,
+ xlrec->created);
+
+ /* Skip the change if already processed (per the snapshot). */
+ if (transactional &&
+ !SnapBuildProcessChange(builder, xid, buf->origptr))
+ return;
+ else if (!transactional &&
+ (SnapBuildCurrentState(builder) != SNAPBUILD_CONSISTENT ||
+ SnapBuildXactNeedsSkip(builder, buf->origptr)))
+ return;
+
+ /* Queue the increment (or send immediately if not transactional). */
+ snapshot = SnapBuildGetOrBuildSnapshot(builder, xid);
+ ReorderBufferQueueSequence(ctx->reorder, xid, snapshot, buf->endptr,
+ origin_id, target_node, transactional,
+ xlrec->created, tuplebuf);
+}
diff --git a/src/backend/replication/logical/logical.c b/src/backend/replication/logical/logical.c
index b7d9521576..811d9912a1 100644
--- a/src/backend/replication/logical/logical.c
+++ b/src/backend/replication/logical/logical.c
@@ -73,6 +73,10 @@ static void truncate_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
static void message_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
XLogRecPtr message_lsn, bool transactional,
const char *prefix, Size message_size, const char *message);
+static void sequence_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ bool transactional, bool created,
+ int64 last_value, int64 log_cnt, bool is_called);
/* streaming callbacks */
static void stream_start_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
@@ -90,6 +94,10 @@ static void stream_change_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn
static void stream_message_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
XLogRecPtr message_lsn, bool transactional,
const char *prefix, Size message_size, const char *message);
+static void stream_sequence_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ bool transactional, bool created,
+ int64 last_value, int64 log_cnt, bool is_called);
static void stream_truncate_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
int nrelations, Relation relations[], ReorderBufferChange *change);
@@ -217,6 +225,7 @@ StartupDecodingContext(List *output_plugin_options,
ctx->reorder->apply_truncate = truncate_cb_wrapper;
ctx->reorder->commit = commit_cb_wrapper;
ctx->reorder->message = message_cb_wrapper;
+ ctx->reorder->sequence = sequence_cb_wrapper;
/*
* To support streaming, we require start/stop/abort/commit/change
@@ -233,6 +242,7 @@ StartupDecodingContext(List *output_plugin_options,
(ctx->callbacks.stream_commit_cb != NULL) ||
(ctx->callbacks.stream_change_cb != NULL) ||
(ctx->callbacks.stream_message_cb != NULL) ||
+ (ctx->callbacks.stream_sequence_cb != NULL) ||
(ctx->callbacks.stream_truncate_cb != NULL);
/*
@@ -250,6 +260,7 @@ StartupDecodingContext(List *output_plugin_options,
ctx->reorder->stream_commit = stream_commit_cb_wrapper;
ctx->reorder->stream_change = stream_change_cb_wrapper;
ctx->reorder->stream_message = stream_message_cb_wrapper;
+ ctx->reorder->stream_sequence = stream_sequence_cb_wrapper;
ctx->reorder->stream_truncate = stream_truncate_cb_wrapper;
@@ -1202,6 +1213,43 @@ message_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
error_context_stack = errcallback.previous;
}
+static void
+sequence_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ bool transactional, bool created,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ LogicalDecodingContext *ctx = cache->private_data;
+ LogicalErrorCallbackState state;
+ ErrorContextCallback errcallback;
+
+ Assert(!ctx->fast_forward);
+
+ if (ctx->callbacks.sequence_cb == NULL)
+ return;
+
+ /* Push callback + info on the error context stack */
+ state.ctx = ctx;
+ state.callback_name = "sequence";
+ state.report_location = sequence_lsn;
+ errcallback.callback = output_plugin_error_callback;
+ errcallback.arg = (void *) &state;
+ errcallback.previous = error_context_stack;
+ error_context_stack = &errcallback;
+
+ /* set output state */
+ ctx->accept_writes = true;
+ ctx->write_xid = txn != NULL ? txn->xid : InvalidTransactionId;
+ ctx->write_location = sequence_lsn;
+
+ /* do the actual work: call callback */
+ ctx->callbacks.sequence_cb(ctx, txn, sequence_lsn, rel, transactional,
+ created, last_value, log_cnt, is_called);
+
+ /* Pop the error context stack */
+ error_context_stack = errcallback.previous;
+}
+
static void
stream_start_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
XLogRecPtr first_lsn)
@@ -1507,6 +1555,47 @@ stream_message_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
error_context_stack = errcallback.previous;
}
+static void
+stream_sequence_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ bool transactional, bool created,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ LogicalDecodingContext *ctx = cache->private_data;
+ LogicalErrorCallbackState state;
+ ErrorContextCallback errcallback;
+
+ Assert(!ctx->fast_forward);
+
+ /* We're only supposed to call this when streaming is supported. */
+ Assert(ctx->streaming);
+
+ /* this callback is optional */
+ if (ctx->callbacks.stream_sequence_cb == NULL)
+ return;
+
+ /* Push callback + info on the error context stack */
+ state.ctx = ctx;
+ state.callback_name = "stream_sequence";
+ state.report_location = sequence_lsn;
+ errcallback.callback = output_plugin_error_callback;
+ errcallback.arg = (void *) &state;
+ errcallback.previous = error_context_stack;
+ error_context_stack = &errcallback;
+
+ /* set output state */
+ ctx->accept_writes = true;
+ ctx->write_xid = txn != NULL ? txn->xid : InvalidTransactionId;
+ ctx->write_location = sequence_lsn;
+
+ /* do the actual work: call callback */
+ ctx->callbacks.sequence_cb(ctx, txn, sequence_lsn, rel, transactional,
+ created, last_value, log_cnt, is_called);
+
+ /* Pop the error context stack */
+ error_context_stack = errcallback.previous;
+}
+
static void
stream_truncate_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
int nrelations, Relation relations[],
diff --git a/src/backend/replication/logical/reorderbuffer.c b/src/backend/replication/logical/reorderbuffer.c
index 46e66608cf..434926459f 100644
--- a/src/backend/replication/logical/reorderbuffer.c
+++ b/src/backend/replication/logical/reorderbuffer.c
@@ -77,6 +77,35 @@
* a bit more memory to the oldest subtransactions, because it's likely
* they are the source for the next sequence of changes.
*
+ * When decoding sequences, we differentiate between a sequences created
+ * in a (running) transaction, and sequences created in other (already
+ * committed) transactions. Changes for sequences created in the same
+ * top-level transaction are treated as "transactional" i.e. just like
+ * any other change from that transaction (and discarded in case of a
+ * rollback). Changes for sequences created earlier are treated as not
+ * transactional - are processed immediately, as if performed outside
+ * any transaction (and thus not rolled back).
+ *
+ * This mixed behavior is necessary - sequences are non-transactional
+ * (e.g. ROLLBACK does not undo the sequence increments). But for new
+ * sequences, we need to handle them in a transactional way, because if
+ * we ever get some DDL support, the sequence won't exist until the
+ * transaction gets applied. So we need to ensure the increments don't
+ * happen until the sequence gets created.
+ *
+ * To differentiate which sequences are "old" and which were created
+ * in a still-running transaction, we track sequences created in running
+ * transactions in a hash table. Each sequence is identified by it's
+ * relfilenode, and at transaction commit we discard all sequences
+ * created in that particular transaction. For each sequence we track
+ * the XID of the (sub)transaction that created it, and we cleanup the
+ * sequences for each subtransaction when it completes (commit/rollback).
+ *
+ * We don't use the XID to check if it's the same top-level transaction.
+ * It's enough to know it was created in an in-progress transaction,
+ * and we know it must be the current one because otherwise it wouldn't
+ * see the sequence object.
+ *
* -------------------------------------------------------------------------
*/
#include "postgres.h"
@@ -116,6 +145,13 @@ typedef struct ReorderBufferTXNByIdEnt
ReorderBufferTXN *txn;
} ReorderBufferTXNByIdEnt;
+/* entry for hash table we use to track sequences created in running xacts */
+typedef struct ReorderBufferSequenceEnt
+{
+ RelFileNode rnode;
+ TransactionId xid;
+} ReorderBufferSequenceEnt;
+
/* data structures for (relfilenode, ctid) => (cmin, cmax) mapping */
typedef struct ReorderBufferTupleCidKey
{
@@ -339,6 +375,14 @@ ReorderBufferAllocate(void)
buffer->by_txn = hash_create("ReorderBufferByXid", 1000, &hash_ctl,
HASH_ELEM | HASH_BLOBS | HASH_CONTEXT);
+ /* hash table of sequences, mapping relfilenode to XID of transaction */
+ hash_ctl.keysize = sizeof(RelFileNode);
+ hash_ctl.entrysize = sizeof(ReorderBufferSequenceEnt);
+ hash_ctl.hcxt = buffer->context;
+
+ buffer->sequences = hash_create("ReorderBufferSequenceHash", 1000, &hash_ctl,
+ HASH_ELEM | HASH_BLOBS | HASH_CONTEXT);
+
buffer->by_txn_last_xid = InvalidTransactionId;
buffer->by_txn_last_txn = NULL;
@@ -525,6 +569,13 @@ ReorderBufferReturnChange(ReorderBuffer *rb, ReorderBufferChange *change,
change->data.truncate.relids = NULL;
}
break;
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
+ if (change->data.sequence.tuple)
+ {
+ ReorderBufferReturnTupleBuf(rb, change->data.sequence.tuple);
+ change->data.sequence.tuple = NULL;
+ }
+ break;
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM:
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_ABORT:
case REORDER_BUFFER_CHANGE_INTERNAL_COMMAND_ID:
@@ -859,6 +910,242 @@ ReorderBufferQueueMessage(ReorderBuffer *rb, TransactionId xid,
}
}
+/*
+ * Treat the sequence increment as transactional?
+ *
+ * The hash table tracks all sequences created in in-progress transactions,
+ * so we simply do a lookup (the sequence is identified by relfilende). If
+ * we find a match, the increment should be handled as transactional.
+ */
+bool
+ReorderBufferSequenceIsTransactional(ReorderBuffer *rb,
+ RelFileNode rnode, bool created)
+{
+ bool found = false;
+
+ if (created)
+ return true;
+
+ hash_search(rb->sequences,
+ (void *) &rnode,
+ HASH_FIND,
+ &found);
+
+ return found;
+}
+
+/*
+ * Cleanup sequences created in in-progress transactions.
+ *
+ * There's no way to search by XID, so we simply do a seqscan of all
+ * the entries in the hash table. Hopefully there are only a couple
+ * entries in most cases - people generally don't create many new
+ * sequences over and over.
+ */
+static void
+ReorderBufferSequenceCleanup(ReorderBuffer *rb, TransactionId xid)
+{
+ HASH_SEQ_STATUS scan_status;
+ ReorderBufferSequenceEnt *ent;
+
+ hash_seq_init(&scan_status, rb->sequences);
+ while ((ent = (ReorderBufferSequenceEnt *) hash_seq_search(&scan_status)) != NULL)
+ {
+ /* skip sequences not from this transaction */
+ if (ent->xid != xid)
+ continue;
+
+ (void) hash_search(rb->sequences,
+ (void *) &(ent->rnode),
+ HASH_REMOVE, NULL);
+ }
+}
+
+/*
+ * A transactional sequence increment is queued to be processed upon commit
+ * and a non-transactional increment gets processed immediately.
+ *
+ * A sequence update may be both transactional and non-transactional. When
+ * created in a running transaction, treat it as transactional and queue
+ * the change in it. Otherwise treat it as non-transactional, so that we
+ * don't forget the increment in case of a rollback.
+ */
+void
+ReorderBufferQueueSequence(ReorderBuffer *rb, TransactionId xid,
+ Snapshot snapshot, XLogRecPtr lsn, RepOriginId origin_id,
+ RelFileNode rnode, bool transactional, bool created,
+ ReorderBufferTupleBuf *tuplebuf)
+{
+ /*
+ * Change needs to be handled as transactional, because the sequence was
+ * created in a transaction that is still running. In that case all the
+ * changes need to be queued in that transaction, we must not send them
+ * to the downstream until the transaction commits.
+ *
+ * There's a bit of a trouble with subtransactions - we can't queue it
+ * into the subxact, because it might be rolled back and we'd lose the
+ * increment. We need to queue it into the same (sub)xact that created
+ * the sequence, which is why we track the XID in the hash table.
+ */
+ if (transactional)
+ {
+ MemoryContext oldcontext;
+ ReorderBufferChange *change;
+
+ /* lookup sequence by relfilenode */
+ ReorderBufferSequenceEnt *ent;
+ bool found;
+
+ /* transactional changes require a transaction */
+ Assert(xid != InvalidTransactionId);
+
+ /* search the lookup table (we ignore the return value, found is enough) */
+ ent = hash_search(rb->sequences,
+ (void *) &rnode,
+ created ? HASH_ENTER : HASH_FIND,
+ &found);
+
+ /*
+ * If this is the "create" increment, we must not have found any
+ * pre-existing entry in the hash table (i.e. there must not be
+ * any conflicting sequence).
+ */
+ Assert(!(created && found));
+
+ /* But we must have either created or found an existing entry. */
+ Assert(created || found);
+
+ /*
+ * When creating the sequence, remember the XID of the transaction
+ * that created id.
+ */
+ if (created)
+ ent->xid = xid;
+
+ /* XXX Maybe check that we're still in the same top-level xact? */
+
+ /* OK, allocate and queue the change */
+ oldcontext = MemoryContextSwitchTo(rb->context);
+
+ change = ReorderBufferGetChange(rb);
+
+ change->action = REORDER_BUFFER_CHANGE_SEQUENCE;
+ change->origin_id = origin_id;
+
+ memcpy(&change->data.sequence.relnode, &rnode, sizeof(RelFileNode));
+
+ change->data.sequence.created = created;
+ change->data.sequence.tuple = tuplebuf;
+
+ /* add it to the same subxact that created the sequence */
+ ReorderBufferQueueChange(rb, ent->xid, lsn, change, false);
+
+ MemoryContextSwitchTo(oldcontext);
+ }
+ else
+ {
+ /*
+ * This increment is for a sequence that was not created in any
+ * running transaction, so we treat it as non-transactional and
+ * just send it to the output plugin directly.
+ */
+ ReorderBufferTXN *txn = NULL;
+ volatile Snapshot snapshot_now = snapshot;
+ bool using_subtxn;
+
+#ifdef USE_ASSERT_CHECKING
+ /* All "creates" have to be handled as transactional. */
+ Assert(!created);
+
+ /* Make sure the sequence is not in the hash table. */
+ {
+ bool found;
+ hash_search(rb->sequences,
+ (void *) &rnode,
+ HASH_FIND, &found);
+ Assert(!found);
+ }
+#endif
+
+ if (xid != InvalidTransactionId)
+ txn = ReorderBufferTXNByXid(rb, xid, true, NULL, lsn, true);
+
+ /* setup snapshot to allow catalog access */
+ SetupHistoricSnapshot(snapshot_now, NULL);
+
+ /*
+ * Decoding needs access to syscaches et al., which in turn use
+ * heavyweight locks and such. Thus we need to have enough state around to
+ * keep track of those. The easiest way is to simply use a transaction
+ * internally. That also allows us to easily enforce that nothing writes
+ * to the database by checking for xid assignments.
+ *
+ * When we're called via the SQL SRF there's already a transaction
+ * started, so start an explicit subtransaction there.
+ */
+ using_subtxn = IsTransactionOrTransactionBlock();
+
+ PG_TRY();
+ {
+ Relation relation;
+ HeapTuple tuple;
+ bool isnull;
+ int64 last_value, log_cnt;
+ bool is_called;
+ Oid reloid;
+
+ if (using_subtxn)
+ BeginInternalSubTransaction("sequence");
+ else
+ StartTransactionCommand();
+
+ reloid = RelidByRelfilenode(rnode.spcNode, rnode.relNode);
+
+ if (reloid == InvalidOid)
+ elog(ERROR, "could not map filenode \"%s\" to relation OID",
+ relpathperm(rnode,
+ MAIN_FORKNUM));
+
+ relation = RelationIdGetRelation(reloid);
+ tuple = &tuplebuf->tuple;
+
+ /*
+ * Extract the internal sequence values, describing the state.
+ *
+ * XXX Seems a bit strange to access it directly. Maybe there's
+ * a better / more correct way?
+ */
+ last_value = heap_getattr(tuple, 1, RelationGetDescr(relation), &isnull);
+ log_cnt = heap_getattr(tuple, 2, RelationGetDescr(relation), &isnull);
+ is_called = heap_getattr(tuple, 3, RelationGetDescr(relation), &isnull);
+
+ rb->sequence(rb, txn, lsn, relation, transactional, created,
+ last_value, log_cnt, is_called);
+
+ RelationClose(relation);
+
+ TeardownHistoricSnapshot(false);
+
+ AbortCurrentTransaction();
+
+ if (using_subtxn)
+ RollbackAndReleaseCurrentSubTransaction();
+ }
+ PG_CATCH();
+ {
+ TeardownHistoricSnapshot(true);
+
+ AbortCurrentTransaction();
+
+ if (using_subtxn)
+ RollbackAndReleaseCurrentSubTransaction();
+
+ PG_RE_THROW();
+ }
+ PG_END_TRY();
+ }
+}
+
/*
* AssertTXNLsnOrder
* Verify LSN ordering of transaction lists in the reorderbuffer
@@ -1535,6 +1822,9 @@ ReorderBufferCleanupTXN(ReorderBuffer *rb, ReorderBufferTXN *txn)
&found);
Assert(found);
+ /* Remove sequences created in this transaction (if any). */
+ ReorderBufferSequenceCleanup(rb, txn->xid);
+
/* remove entries spilled to disk */
if (rbtxn_is_serialized(txn))
ReorderBufferRestoreCleanup(rb, txn);
@@ -1950,6 +2240,42 @@ ReorderBufferApplyMessage(ReorderBuffer *rb, ReorderBufferTXN *txn,
change->data.msg.message);
}
+/*
+ * Helper function for ReorderBufferProcessTXN for applying sequences.
+ */
+static inline void
+ReorderBufferApplySequence(ReorderBuffer *rb, ReorderBufferTXN *txn,
+ Relation relation, ReorderBufferChange *change,
+ bool streaming)
+{
+ HeapTuple tuple;
+ bool isnull;
+ int64 last_value, log_cnt;
+ bool is_called;
+
+ tuple = &change->data.sequence.tuple->tuple;
+
+ /*
+ * Extract the internal sequence values, describing the state.
+ *
+ * XXX Seems a bit strange to access it directly. Maybe there's
+ * a better / more correct way?
+ */
+ last_value = heap_getattr(tuple, 1, RelationGetDescr(relation), &isnull);
+ log_cnt = heap_getattr(tuple, 2, RelationGetDescr(relation), &isnull);
+ is_called = heap_getattr(tuple, 3, RelationGetDescr(relation), &isnull);
+
+ /* Only ever called from ReorderBufferApplySequence, so transational. */
+ if (streaming)
+ rb->stream_sequence(rb, txn, change->lsn, relation, true,
+ change->data.sequence.created,
+ last_value, log_cnt, is_called);
+ else
+ rb->sequence(rb, txn, change->lsn, relation, true,
+ change->data.sequence.created,
+ last_value, log_cnt, is_called);
+}
+
/*
* Function to store the command id and snapshot at the end of the current
* stream so that we can reuse the same while sending the next stream.
@@ -2392,6 +2718,31 @@ ReorderBufferProcessTXN(ReorderBuffer *rb, ReorderBufferTXN *txn,
case REORDER_BUFFER_CHANGE_INTERNAL_TUPLECID:
elog(ERROR, "tuplecid value in changequeue");
break;
+
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
+ Assert(snapshot_now);
+
+ reloid = RelidByRelfilenode(change->data.sequence.relnode.spcNode,
+ change->data.sequence.relnode.relNode);
+
+ if (reloid == InvalidOid)
+ elog(ERROR, "could not map filenode \"%s\" to relation OID",
+ relpathperm(change->data.sequence.relnode,
+ MAIN_FORKNUM));
+
+ relation = RelationIdGetRelation(reloid);
+
+ if (!RelationIsValid(relation))
+ elog(ERROR, "could not open relation with OID %u (for filenode \"%s\")",
+ reloid,
+ relpathperm(change->data.sequence.relnode,
+ MAIN_FORKNUM));
+
+ if (RelationIsLogicallyLogged(relation))
+ ReorderBufferApplySequence(rb, txn, relation, change, streaming);
+
+ RelationClose(relation);
+ break;
}
}
@@ -3776,6 +4127,39 @@ ReorderBufferSerializeChange(ReorderBuffer *rb, ReorderBufferTXN *txn,
memcpy(data, change->data.truncate.relids, size);
data += size;
+ break;
+ }
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
+ {
+ char *data;
+ ReorderBufferTupleBuf *tup;
+ Size len = 0;
+
+ tup = change->data.sequence.tuple;
+
+ if (tup)
+ {
+ sz += sizeof(HeapTupleData);
+ len = tup->tuple.t_len;
+ sz += len;
+ }
+
+ /* make sure we have enough space */
+ ReorderBufferSerializeReserve(rb, sz);
+
+ data = ((char *) rb->outbuf) + sizeof(ReorderBufferDiskChange);
+ /* might have been reallocated above */
+ ondisk = (ReorderBufferDiskChange *) rb->outbuf;
+
+ if (len)
+ {
+ memcpy(data, &tup->tuple, sizeof(HeapTupleData));
+ data += sizeof(HeapTupleData);
+
+ memcpy(data, tup->tuple.t_data, len);
+ data += len;
+ }
+
break;
}
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM:
@@ -4040,6 +4424,22 @@ ReorderBufferChangeSize(ReorderBufferChange *change)
{
sz += sizeof(Oid) * change->data.truncate.nrelids;
+ break;
+ }
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
+ {
+ ReorderBufferTupleBuf *tup;
+ Size len = 0;
+
+ tup = change->data.sequence.tuple;
+
+ if (tup)
+ {
+ sz += sizeof(HeapTupleData);
+ len = tup->tuple.t_len;
+ sz += len;
+ }
+
break;
}
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM:
@@ -4341,6 +4741,30 @@ ReorderBufferRestoreChange(ReorderBuffer *rb, ReorderBufferTXN *txn,
break;
}
+
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
+ if (change->data.sequence.tuple)
+ {
+ uint32 tuplelen = ((HeapTuple) data)->t_len;
+
+ change->data.sequence.tuple =
+ ReorderBufferGetTupleBuf(rb, tuplelen - SizeofHeapTupleHeader);
+
+ /* restore ->tuple */
+ memcpy(&change->data.sequence.tuple->tuple, data,
+ sizeof(HeapTupleData));
+ data += sizeof(HeapTupleData);
+
+ /* reset t_data pointer into the new tuplebuf */
+ change->data.sequence.tuple->tuple.t_data =
+ ReorderBufferTupleBufData(change->data.sequence.tuple);
+
+ /* restore tuple data itself */
+ memcpy(change->data.sequence.tuple->tuple.t_data, data, tuplelen);
+ data += tuplelen;
+ }
+ break;
+
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM:
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_ABORT:
case REORDER_BUFFER_CHANGE_INTERNAL_COMMAND_ID:
diff --git a/src/include/commands/sequence.h b/src/include/commands/sequence.h
index 40544dd4c7..5919fb90ee 100644
--- a/src/include/commands/sequence.h
+++ b/src/include/commands/sequence.h
@@ -48,6 +48,7 @@ typedef FormData_pg_sequence_data *Form_pg_sequence_data;
typedef struct xl_seq_rec
{
RelFileNode node;
+ bool created; /* is this a CREATE SEQUENCE */
/* SEQUENCE TUPLE DATA FOLLOWS AT THE END */
} xl_seq_rec;
diff --git a/src/include/replication/output_plugin.h b/src/include/replication/output_plugin.h
index 810495ed0e..57bc13f11c 100644
--- a/src/include/replication/output_plugin.h
+++ b/src/include/replication/output_plugin.h
@@ -88,6 +88,19 @@ typedef void (*LogicalDecodeMessageCB) (struct LogicalDecodingContext *ctx,
Size message_size,
const char *message);
+/*
+ * Called for the generic logical decoding sequences.
+ */
+typedef void (*LogicalDecodeSequenceCB) (struct LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional,
+ bool created,
+ int64 last_value,
+ int64 log_cnt,
+ bool is_called);
+
/*
* Filter changes by origin.
*/
@@ -199,6 +212,20 @@ typedef void (*LogicalDecodeStreamMessageCB) (struct LogicalDecodingContext *ctx
Size message_size,
const char *message);
+/*
+ * Called for the streaming generic logical decoding sequences from in-progress
+ * transactions.
+ */
+typedef void (*LogicalDecodeStreamSequenceCB) (struct LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional,
+ bool created,
+ int64 last_value,
+ int64 log_cnt,
+ bool is_called);
+
/*
* Callback for streaming truncates from in-progress transactions.
*/
@@ -219,6 +246,7 @@ typedef struct OutputPluginCallbacks
LogicalDecodeTruncateCB truncate_cb;
LogicalDecodeCommitCB commit_cb;
LogicalDecodeMessageCB message_cb;
+ LogicalDecodeSequenceCB sequence_cb;
LogicalDecodeFilterByOriginCB filter_by_origin_cb;
LogicalDecodeShutdownCB shutdown_cb;
@@ -237,6 +265,7 @@ typedef struct OutputPluginCallbacks
LogicalDecodeStreamCommitCB stream_commit_cb;
LogicalDecodeStreamChangeCB stream_change_cb;
LogicalDecodeStreamMessageCB stream_message_cb;
+ LogicalDecodeStreamSequenceCB stream_sequence_cb;
LogicalDecodeStreamTruncateCB stream_truncate_cb;
} OutputPluginCallbacks;
diff --git a/src/include/replication/reorderbuffer.h b/src/include/replication/reorderbuffer.h
index 5b40ff75f7..63e6ed037b 100644
--- a/src/include/replication/reorderbuffer.h
+++ b/src/include/replication/reorderbuffer.h
@@ -64,7 +64,8 @@ enum ReorderBufferChangeType
REORDER_BUFFER_CHANGE_INTERNAL_SPEC_INSERT,
REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM,
REORDER_BUFFER_CHANGE_INTERNAL_SPEC_ABORT,
- REORDER_BUFFER_CHANGE_TRUNCATE
+ REORDER_BUFFER_CHANGE_TRUNCATE,
+ REORDER_BUFFER_CHANGE_SEQUENCE
};
/* forward declaration */
@@ -158,6 +159,14 @@ typedef struct ReorderBufferChange
uint32 ninvalidations; /* Number of messages */
SharedInvalidationMessage *invalidations; /* invalidation message */
} inval;
+
+ /* Context data for Sequence changes */
+ struct
+ {
+ RelFileNode relnode;
+ bool created;
+ ReorderBufferTupleBuf *tuple;
+ } sequence;
} data;
/*
@@ -430,6 +439,15 @@ typedef void (*ReorderBufferMessageCB) (ReorderBuffer *rb,
const char *prefix, Size sz,
const char *message);
+/* sequence callback signature */
+typedef void (*ReorderBufferSequenceCB) (ReorderBuffer *rb,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional, bool created,
+ int64 last_value, int64 log_cnt,
+ bool is_called);
+
/* begin prepare callback signature */
typedef void (*ReorderBufferBeginPrepareCB) (ReorderBuffer *rb,
ReorderBufferTXN *txn);
@@ -496,6 +514,15 @@ typedef void (*ReorderBufferStreamMessageCB) (
const char *prefix, Size sz,
const char *message);
+/* stream sequence callback signature */
+typedef void (*ReorderBufferStreamSequenceCB) (ReorderBuffer *rb,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional, bool created,
+ int64 last_value, int64 log_cnt,
+ bool is_called);
+
/* stream truncate callback signature */
typedef void (*ReorderBufferStreamTruncateCB) (
ReorderBuffer *rb,
@@ -511,6 +538,12 @@ struct ReorderBuffer
*/
HTAB *by_txn;
+ /*
+ * relfilenode => XID lookup table for sequences created in a transaction
+ * (also includes altered sequences, which assigns new relfilenode)
+ */
+ HTAB *sequences;
+
/*
* Transactions that could be a toplevel xact, ordered by LSN of the first
* record bearing that xid.
@@ -541,6 +574,7 @@ struct ReorderBuffer
ReorderBufferApplyTruncateCB apply_truncate;
ReorderBufferCommitCB commit;
ReorderBufferMessageCB message;
+ ReorderBufferSequenceCB sequence;
/*
* Callbacks to be called when streaming a transaction at prepare time.
@@ -560,6 +594,7 @@ struct ReorderBuffer
ReorderBufferStreamCommitCB stream_commit;
ReorderBufferStreamChangeCB stream_change;
ReorderBufferStreamMessageCB stream_message;
+ ReorderBufferStreamSequenceCB stream_sequence;
ReorderBufferStreamTruncateCB stream_truncate;
/*
@@ -635,6 +670,10 @@ void ReorderBufferQueueChange(ReorderBuffer *, TransactionId,
void ReorderBufferQueueMessage(ReorderBuffer *, TransactionId, Snapshot snapshot, XLogRecPtr lsn,
bool transactional, const char *prefix,
Size message_size, const char *message);
+void ReorderBufferQueueSequence(ReorderBuffer *rb, TransactionId xid,
+ Snapshot snapshot, XLogRecPtr lsn, RepOriginId origin_id,
+ RelFileNode rnode, bool transactional, bool created,
+ ReorderBufferTupleBuf *tuplebuf);
void ReorderBufferCommit(ReorderBuffer *, TransactionId,
XLogRecPtr commit_lsn, XLogRecPtr end_lsn,
TimestampTz commit_time, RepOriginId origin_id, XLogRecPtr origin_lsn);
@@ -682,4 +721,7 @@ void ReorderBufferSetRestartPoint(ReorderBuffer *, XLogRecPtr ptr);
void StartupReorderBuffer(void);
+bool ReorderBufferSequenceIsTransactional(ReorderBuffer *rb,
+ RelFileNode rnode, bool created);
+
#endif
--
2.31.1
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2021-11-22 15:44 Peter Eisentraut <[email protected]>
parent: Tomas Vondra <[email protected]>
1 sibling, 1 reply; 127+ messages in thread
From: Peter Eisentraut @ 2021-11-22 15:44 UTC (permalink / raw)
To: Tomas Vondra <[email protected]>; Hannu Krosing <[email protected]>; +Cc: vignesh C <[email protected]>; PostgreSQL Hackers <[email protected]>; Petr Jelinek <[email protected]>
On 22.11.21 01:47, Tomas Vondra wrote:
> So I think just decoding the sequence tuples is a better solution - for
> large transactions (consuming many values from the sequence) it may be
> more expensive (i.e. send more records to replica). But I doubt that
> matters too much - it's likely negligible compared to other data for
> large transactions.
I agree that the original approach is better. It was worth trying out
this alternative, but it seems quite complicated. I note that a lot of
additional code had to be added around several areas of the code,
whereas the original patch really just touched the logical decoding
code, as it should. This doesn't prevent anyone from attempting to
optimize things somehow in the future, but for now let's move forward
with the simple approach.
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2021-11-22 16:01 Petr Jelinek <[email protected]>
parent: Peter Eisentraut <[email protected]>
0 siblings, 0 replies; 127+ messages in thread
From: Petr Jelinek @ 2021-11-22 16:01 UTC (permalink / raw)
To: Peter Eisentraut <[email protected]>; +Cc: Tomas Vondra <[email protected]>; Hannu Krosing <[email protected]>; vignesh C <[email protected]>; PostgreSQL Hackers <[email protected]>
> On 22. 11. 2021, at 16:44, Peter Eisentraut <[email protected]> wrote:
>
> On 22.11.21 01:47, Tomas Vondra wrote:
>> So I think just decoding the sequence tuples is a better solution - for large transactions (consuming many values from the sequence) it may be more expensive (i.e. send more records to replica). But I doubt that matters too much - it's likely negligible compared to other data for large transactions.
>
> I agree that the original approach is better. It was worth trying out this alternative, but it seems quite complicated. I note that a lot of additional code had to be added around several areas of the code, whereas the original patch really just touched the logical decoding code, as it should. This doesn't prevent anyone from attempting to optimize things somehow in the future, but for now let's move forward with the simple approach.
+1
--
Petr Jelinek
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2021-11-23 01:01 Andres Freund <[email protected]>
parent: Hannu Krosing <[email protected]>
1 sibling, 1 reply; 127+ messages in thread
From: Andres Freund @ 2021-11-23 01:01 UTC (permalink / raw)
To: Hannu Krosing <[email protected]>; +Cc: Tomas Vondra <[email protected]>; Peter Eisentraut <[email protected]>; vignesh C <[email protected]>; PostgreSQL Hackers <[email protected]>; Petr Jelinek <[email protected]>
Hi,
On 2021-09-25 22:05:43 +0200, Hannu Krosing wrote:
> If our aim is just to make sure that all user-visible data in
> *transactional* tables is consistent with sequence state then one
> very much simplified approach to this could be to track the results of
> nextval() calls in a transaction at COMMIT put the latest sequence
> value in WAL (or just track the sequences affected and put the latest
> sequence state in WAL at commit which needs extra read of sequence but
> protects against race conditions with parallel transactions which get
> rolled back later)
I think this is a bad idea. It's architecturally more complicated and prevents
use cases because sequence values aren't guaranteed to be as new as on the
original system. You'd need to track all sequence use somehow *even if there
is no relevant WAL generated* in a transaction. There's simply no evidence of
sequence use in a transaction if that transaction uses a previously logged
sequence value.
Greetings,
Andres Freund
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2021-11-23 13:36 Tomas Vondra <[email protected]>
parent: Andres Freund <[email protected]>
0 siblings, 0 replies; 127+ messages in thread
From: Tomas Vondra @ 2021-11-23 13:36 UTC (permalink / raw)
To: Andres Freund <[email protected]>; Hannu Krosing <[email protected]>; +Cc: Peter Eisentraut <[email protected]>; vignesh C <[email protected]>; PostgreSQL Hackers <[email protected]>; Petr Jelinek <[email protected]>
On 11/23/21 02:01, Andres Freund wrote:
> Hi,
>
> On 2021-09-25 22:05:43 +0200, Hannu Krosing wrote:
>> If our aim is just to make sure that all user-visible data in
>> *transactional* tables is consistent with sequence state then one
>> very much simplified approach to this could be to track the results of
>> nextval() calls in a transaction at COMMIT put the latest sequence
>> value in WAL (or just track the sequences affected and put the latest
>> sequence state in WAL at commit which needs extra read of sequence but
>> protects against race conditions with parallel transactions which get
>> rolled back later)
>
> I think this is a bad idea. It's architecturally more complicated and prevents
> use cases because sequence values aren't guaranteed to be as new as on the
> original system. You'd need to track all sequence use somehow *even if there
> is no relevant WAL generated* in a transaction. There's simply no evidence of
> sequence use in a transaction if that transaction uses a previously logged
> sequence value.
>
Not quite. We already have a cache of all sequences used by a session
(see seqhashtab in sequence.c), and it's not that hard to extend it to
per-transaction tracking. That's what the last two versions do, mostly.
But there are various issues with that approach, described in my last
message(s).
regards
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2021-12-07 14:38 Peter Eisentraut <[email protected]>
parent: Tomas Vondra <[email protected]>
1 sibling, 1 reply; 127+ messages in thread
From: Peter Eisentraut @ 2021-12-07 14:38 UTC (permalink / raw)
To: Tomas Vondra <[email protected]>; +Cc: PostgreSQL Hackers <[email protected]>; Petr Jelinek <[email protected]>
I have checked the 0001 and 0003 patches. (I think we have dismissed
the approach in 0002 for now. And let's talk about 0004 later.)
I have attached a few fixup patches, described further below.
# 0001
The argument "create" for fill_seq_with_data() is always true (and
patch 0002 removes it). So I'm not sure if it's needed. If it is, it
should be documented somewhere.
About the comment you added in nextval_internal(): It's a good
explanation, so I would leave it in. I also agree with the
conclusion of the comment that the current solution is reasonable. We
probably don't need the same comment again in fill_seq_with_data() and
again in do_setval(). Note that both of the latter functions already
point to nextval_interval() for other comments, so the same can be
relied upon here as well.
The order of the new fields sequence_cb and stream_sequence_cb is a
bit inconsistent compared to truncate_cb and stream_truncate_cb.
Maybe take another look to make the order more uniform.
Some documentation needs to be added to the Logical Decoding chapter.
I have attached a patch that I think catches all the places that need
to be updated, with some details left for you to fill in. ;-) The
ordering of the some of the documentation chunks reflects the order in
which the callbacks appear in the header files, which might not be
optimal; see above.
In reorderbuffer.c, you left a comment about how to access a sequence
tuple. There is an easier way, using Form_pg_sequence_data, which is
how sequence.c also does it. See attached patch.
# 0003
The tests added in 0003 don't pass for me. It appears that you might
have forgotten to update the expected files after you added some tests
or something. The cfbot shows the same. See attached patch for a
correction, but do check what your intent was.
As mentioned before, we probably don't need the skip-sequences option
in test_decoding.
From 2dd9a0e76b9496016e7abebebd0006ddee72d3c1 Mon Sep 17 00:00:00 2001
From: Peter Eisentraut <[email protected]>
Date: Tue, 7 Dec 2021 14:19:42 +0100
Subject: [PATCH 1/4] Whitespace fixes
---
contrib/test_decoding/expected/sequence.out | 6 +++---
contrib/test_decoding/sql/sequence.sql | 6 +++---
2 files changed, 6 insertions(+), 6 deletions(-)
diff --git a/contrib/test_decoding/expected/sequence.out b/contrib/test_decoding/expected/sequence.out
index 17c88990b1..ca55bcebc3 100644
--- a/contrib/test_decoding/expected/sequence.out
+++ b/contrib/test_decoding/expected/sequence.out
@@ -177,7 +177,7 @@ SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'inc
------
(0 rows)
--- rollback on table creation with serial column
+-- rollback on table creation with serial column
BEGIN;
CREATE TABLE test_table (a SERIAL, b INT);
INSERT INTO test_table (b) VALUES (100);
@@ -202,7 +202,7 @@ SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'inc
------
(0 rows)
--- rollback on table with serial column
+-- rollback on table with serial column
CREATE TABLE test_table (a SERIAL, b INT);
BEGIN;
INSERT INTO test_table (b) VALUES (100);
@@ -222,7 +222,7 @@ INSERT INTO test_table (b) VALUES (700);
INSERT INTO test_table (b) VALUES (800);
INSERT INTO test_table (b) VALUES (900);
ROLLBACK;
--- check table and sequence values after rollback
+-- check table and sequence values after rollback
SELECT * from test_table_a_seq;
last_value | log_cnt | is_called
------------+---------+-----------
diff --git a/contrib/test_decoding/sql/sequence.sql b/contrib/test_decoding/sql/sequence.sql
index d8a34738f3..21c4b79222 100644
--- a/contrib/test_decoding/sql/sequence.sql
+++ b/contrib/test_decoding/sql/sequence.sql
@@ -48,7 +48,7 @@ CREATE TABLE test_table (a INT);
ROLLBACK;
SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'skip-sequences', '0');
--- rollback on table creation with serial column
+-- rollback on table creation with serial column
BEGIN;
CREATE TABLE test_table (a SERIAL, b INT);
INSERT INTO test_table (b) VALUES (100);
@@ -65,7 +65,7 @@ CREATE TABLE test_table (a SERIAL, b INT);
ROLLBACK;
SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'skip-sequences', '0');
--- rollback on table with serial column
+-- rollback on table with serial column
CREATE TABLE test_table (a SERIAL, b INT);
BEGIN;
@@ -82,7 +82,7 @@ CREATE TABLE test_table (a SERIAL, b INT);
INSERT INTO test_table (b) VALUES (900);
ROLLBACK;
--- check table and sequence values after rollback
+-- check table and sequence values after rollback
SELECT * from test_table_a_seq;
SELECT nextval('test_table_a_seq');
--
2.34.1
From c5e9eb2792bd276e493aed582627e93b4c681374 Mon Sep 17 00:00:00 2001
From: Peter Eisentraut <[email protected]>
Date: Tue, 7 Dec 2021 15:29:34 +0100
Subject: [PATCH 2/4] Make tests pass
---
contrib/test_decoding/expected/sequence.out | 18 ++++++++++++------
1 file changed, 12 insertions(+), 6 deletions(-)
diff --git a/contrib/test_decoding/expected/sequence.out b/contrib/test_decoding/expected/sequence.out
index ca55bcebc3..d94e185fb9 100644
--- a/contrib/test_decoding/expected/sequence.out
+++ b/contrib/test_decoding/expected/sequence.out
@@ -260,13 +260,15 @@ ROLLBACK TO SAVEPOINT a;
DROP TABLE test_table;
COMMIT;
SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
- data
---------------------------------------------------------------
+ data
+-------------------------------------------------------------------------------------------------
BEGIN
+ sequence public.test_table_a_seq: transactional:1 created:1 last_value:1 log_cnt:0 is_called:0
+ sequence public.test_table_a_seq: transactional:1 created:0 last_value:33 log_cnt:0 is_called:1
table public.test_table: INSERT: a[integer]:1 b[integer]:100
table public.test_table: INSERT: a[integer]:2 b[integer]:200
COMMIT
-(4 rows)
+(6 rows)
-- savepoint test on table with serial column
BEGIN;
@@ -307,11 +309,15 @@ SELECT * FROM test_sequence;
DROP SEQUENCE test_sequence;
COMMIT;
SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
- data
---------
+ data
+------------------------------------------------------------------------------------------------
BEGIN
+ sequence public.test_sequence: transactional:1 created:1 last_value:1 log_cnt:0 is_called:0
+ sequence public.test_sequence: transactional:1 created:0 last_value:33 log_cnt:0 is_called:1
+ sequence public.test_sequence: transactional:1 created:0 last_value:3000 log_cnt:0 is_called:1
+ sequence public.test_sequence: transactional:1 created:0 last_value:3033 log_cnt:0 is_called:1
COMMIT
-(2 rows)
+(6 rows)
SELECT pg_drop_replication_slot('regression_slot');
pg_drop_replication_slot
--
2.34.1
From 0338de62cd955b98244b5b624a2bcb7af76465b0 Mon Sep 17 00:00:00 2001
From: Peter Eisentraut <[email protected]>
Date: Tue, 7 Dec 2021 15:29:51 +0100
Subject: [PATCH 3/4] Some documentation updates
---
doc/src/sgml/logicaldecoding.sgml | 54 +++++++++++++++++++++++++++++--
1 file changed, 51 insertions(+), 3 deletions(-)
diff --git a/doc/src/sgml/logicaldecoding.sgml b/doc/src/sgml/logicaldecoding.sgml
index b6353c7a12..eea4f3cd50 100644
--- a/doc/src/sgml/logicaldecoding.sgml
+++ b/doc/src/sgml/logicaldecoding.sgml
@@ -458,6 +458,7 @@ <title>Initialization Function</title>
LogicalDecodeTruncateCB truncate_cb;
LogicalDecodeCommitCB commit_cb;
LogicalDecodeMessageCB message_cb;
+ LogicalDecodeSequenceCB sequence_cb;
LogicalDecodeFilterByOriginCB filter_by_origin_cb;
LogicalDecodeShutdownCB shutdown_cb;
LogicalDecodeFilterPrepareCB filter_prepare_cb;
@@ -472,6 +473,7 @@ <title>Initialization Function</title>
LogicalDecodeStreamCommitCB stream_commit_cb;
LogicalDecodeStreamChangeCB stream_change_cb;
LogicalDecodeStreamMessageCB stream_message_cb;
+ LogicalDecodeStreamSequenceCB stream_sequence_cb;
LogicalDecodeStreamTruncateCB stream_truncate_cb;
} OutputPluginCallbacks;
@@ -481,9 +483,12 @@ <title>Initialization Function</title>
and <function>commit_cb</function> callbacks are required,
while <function>startup_cb</function>,
<function>filter_by_origin_cb</function>, <function>truncate_cb</function>,
+ <function>sequence_cb</function>,
and <function>shutdown_cb</function> are optional.
If <function>truncate_cb</function> is not set but a
<command>TRUNCATE</command> is to be decoded, the action will be ignored.
+ Similarly, if <function>sequence_cb</function> is not set and a sequence
+ change is to be decoded, the action will be ignored.
</para>
<para>
@@ -492,7 +497,7 @@ <title>Initialization Function</title>
<function>stream_stop_cb</function>, <function>stream_abort_cb</function>,
<function>stream_commit_cb</function>, <function>stream_change_cb</function>,
and <function>stream_prepare_cb</function>
- are required, while <function>stream_message_cb</function> and
+ are required, while <function>stream_message_cb</function>, <function>stream_sequence_cb</function>, and
<function>stream_truncate_cb</function> are optional.
</para>
@@ -808,6 +813,27 @@ <title>Generic Message Callback</title>
</para>
</sect3>
+ <sect3 id="logicaldecoding-output-plugin-sequence">
+ <title>Sequence Callback</title>
+
+ <para>
+ The <function>sequence_cb</function> callback is called for actions that
+ change a sequence.
+<programlisting>
+typedef void (*LogicalDecodeSequenceCB) (struct LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional,
+ bool created,
+ int64 last_value,
+ int64 log_cnt,
+ bool is_called);
+</programlisting>
+ TODO
+ </para>
+ </sect3>
+
<sect3 id="logicaldecoding-output-plugin-filter-prepare">
<title>Prepare Filter Callback</title>
@@ -1017,6 +1043,28 @@ <title>Stream Message Callback</title>
</para>
</sect3>
+ <sect3 id="logicaldecoding-output-plugin-stream-sequence">
+ <title>Stream Sequence Callback</title>
+ <para>
+ The <function>stream_sequence_cb</function> callback is called for
+ actions that change a sequence in a block of streamed changes
+ (demarcated by <function>stream_start_cb</function> and
+ <function>stream_stop_cb</function> calls).
+<programlisting>
+typedef void (*LogicalDecodeStreamSequenceCB) (struct LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional,
+ bool created,
+ int64 last_value,
+ int64 log_cnt,
+ bool is_called);
+</programlisting>
+ TODO
+ </para>
+ </sect3>
+
<sect3 id="logicaldecoding-output-plugin-stream-truncate">
<title>Stream Truncate Callback</title>
<para>
@@ -1197,8 +1245,8 @@ <title>Streaming of Large Transactions for Logical Decoding</title>
in-progress transactions. There are multiple required streaming callbacks
(<function>stream_start_cb</function>, <function>stream_stop_cb</function>,
<function>stream_abort_cb</function>, <function>stream_commit_cb</function>
- and <function>stream_change_cb</function>) and two optional callbacks
- (<function>stream_message_cb</function> and <function>stream_truncate_cb</function>).
+ and <function>stream_change_cb</function>) and multiple optional callbacks
+ (<function>stream_message_cb</function>, <function>stream_sequence_cb</function>, and <function>stream_truncate_cb</function>).
Also, if streaming of two-phase commands is to be supported, then additional
callbacks must be provided. (See <xref linkend="logicaldecoding-two-phase-commits"/>
for details).
--
2.34.1
From 9f29fac7e479c255fb247d201fe8691655c11fbf Mon Sep 17 00:00:00 2001
From: Peter Eisentraut <[email protected]>
Date: Tue, 7 Dec 2021 15:31:36 +0100
Subject: [PATCH 4/4] Simplify accessing sequence tuple data
---
.../replication/logical/reorderbuffer.c | 37 ++++---------------
1 file changed, 8 insertions(+), 29 deletions(-)
diff --git a/src/backend/replication/logical/reorderbuffer.c b/src/backend/replication/logical/reorderbuffer.c
index 434926459f..b4fa16bebe 100644
--- a/src/backend/replication/logical/reorderbuffer.c
+++ b/src/backend/replication/logical/reorderbuffer.c
@@ -120,6 +120,7 @@
#include "access/xact.h"
#include "access/xlog_internal.h"
#include "catalog/catalog.h"
+#include "commands/sequence.h"
#include "lib/binaryheap.h"
#include "miscadmin.h"
#include "pgstat.h"
@@ -1089,9 +1090,7 @@ ReorderBufferQueueSequence(ReorderBuffer *rb, TransactionId xid,
{
Relation relation;
HeapTuple tuple;
- bool isnull;
- int64 last_value, log_cnt;
- bool is_called;
+ Form_pg_sequence_data seq;
Oid reloid;
if (using_subtxn)
@@ -1108,19 +1107,10 @@ ReorderBufferQueueSequence(ReorderBuffer *rb, TransactionId xid,
relation = RelationIdGetRelation(reloid);
tuple = &tuplebuf->tuple;
-
- /*
- * Extract the internal sequence values, describing the state.
- *
- * XXX Seems a bit strange to access it directly. Maybe there's
- * a better / more correct way?
- */
- last_value = heap_getattr(tuple, 1, RelationGetDescr(relation), &isnull);
- log_cnt = heap_getattr(tuple, 2, RelationGetDescr(relation), &isnull);
- is_called = heap_getattr(tuple, 3, RelationGetDescr(relation), &isnull);
+ seq = (Form_pg_sequence_data) GETSTRUCT(tuple);
rb->sequence(rb, txn, lsn, relation, transactional, created,
- last_value, log_cnt, is_called);
+ seq->last_value, seq->log_cnt, seq->is_called);
RelationClose(relation);
@@ -2249,31 +2239,20 @@ ReorderBufferApplySequence(ReorderBuffer *rb, ReorderBufferTXN *txn,
bool streaming)
{
HeapTuple tuple;
- bool isnull;
- int64 last_value, log_cnt;
- bool is_called;
+ Form_pg_sequence_data seq;
tuple = &change->data.sequence.tuple->tuple;
-
- /*
- * Extract the internal sequence values, describing the state.
- *
- * XXX Seems a bit strange to access it directly. Maybe there's
- * a better / more correct way?
- */
- last_value = heap_getattr(tuple, 1, RelationGetDescr(relation), &isnull);
- log_cnt = heap_getattr(tuple, 2, RelationGetDescr(relation), &isnull);
- is_called = heap_getattr(tuple, 3, RelationGetDescr(relation), &isnull);
+ seq = (Form_pg_sequence_data) GETSTRUCT(tuple);
/* Only ever called from ReorderBufferApplySequence, so transational. */
if (streaming)
rb->stream_sequence(rb, txn, change->lsn, relation, true,
change->data.sequence.created,
- last_value, log_cnt, is_called);
+ seq->last_value, seq->log_cnt, seq->is_called);
else
rb->sequence(rb, txn, change->lsn, relation, true,
change->data.sequence.created,
- last_value, log_cnt, is_called);
+ seq->last_value, seq->log_cnt, seq->is_called);
}
/*
--
2.34.1
Attachments:
[text/plain] 0001-Whitespace-fixes.patch (2.8K, ../../[email protected]/2-0001-Whitespace-fixes.patch)
download | inline diff:
From 2dd9a0e76b9496016e7abebebd0006ddee72d3c1 Mon Sep 17 00:00:00 2001
From: Peter Eisentraut <[email protected]>
Date: Tue, 7 Dec 2021 14:19:42 +0100
Subject: [PATCH 1/4] Whitespace fixes
---
contrib/test_decoding/expected/sequence.out | 6 +++---
contrib/test_decoding/sql/sequence.sql | 6 +++---
2 files changed, 6 insertions(+), 6 deletions(-)
diff --git a/contrib/test_decoding/expected/sequence.out b/contrib/test_decoding/expected/sequence.out
index 17c88990b1..ca55bcebc3 100644
--- a/contrib/test_decoding/expected/sequence.out
+++ b/contrib/test_decoding/expected/sequence.out
@@ -177,7 +177,7 @@ SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'inc
------
(0 rows)
--- rollback on table creation with serial column
+-- rollback on table creation with serial column
BEGIN;
CREATE TABLE test_table (a SERIAL, b INT);
INSERT INTO test_table (b) VALUES (100);
@@ -202,7 +202,7 @@ SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'inc
------
(0 rows)
--- rollback on table with serial column
+-- rollback on table with serial column
CREATE TABLE test_table (a SERIAL, b INT);
BEGIN;
INSERT INTO test_table (b) VALUES (100);
@@ -222,7 +222,7 @@ INSERT INTO test_table (b) VALUES (700);
INSERT INTO test_table (b) VALUES (800);
INSERT INTO test_table (b) VALUES (900);
ROLLBACK;
--- check table and sequence values after rollback
+-- check table and sequence values after rollback
SELECT * from test_table_a_seq;
last_value | log_cnt | is_called
------------+---------+-----------
diff --git a/contrib/test_decoding/sql/sequence.sql b/contrib/test_decoding/sql/sequence.sql
index d8a34738f3..21c4b79222 100644
--- a/contrib/test_decoding/sql/sequence.sql
+++ b/contrib/test_decoding/sql/sequence.sql
@@ -48,7 +48,7 @@ CREATE TABLE test_table (a INT);
ROLLBACK;
SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'skip-sequences', '0');
--- rollback on table creation with serial column
+-- rollback on table creation with serial column
BEGIN;
CREATE TABLE test_table (a SERIAL, b INT);
INSERT INTO test_table (b) VALUES (100);
@@ -65,7 +65,7 @@ CREATE TABLE test_table (a SERIAL, b INT);
ROLLBACK;
SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'skip-sequences', '0');
--- rollback on table with serial column
+-- rollback on table with serial column
CREATE TABLE test_table (a SERIAL, b INT);
BEGIN;
@@ -82,7 +82,7 @@ CREATE TABLE test_table (a SERIAL, b INT);
INSERT INTO test_table (b) VALUES (900);
ROLLBACK;
--- check table and sequence values after rollback
+-- check table and sequence values after rollback
SELECT * from test_table_a_seq;
SELECT nextval('test_table_a_seq');
--
2.34.1
[text/plain] 0002-Make-tests-pass.patch (2.4K, ../../[email protected]/3-0002-Make-tests-pass.patch)
download | inline diff:
From c5e9eb2792bd276e493aed582627e93b4c681374 Mon Sep 17 00:00:00 2001
From: Peter Eisentraut <[email protected]>
Date: Tue, 7 Dec 2021 15:29:34 +0100
Subject: [PATCH 2/4] Make tests pass
---
contrib/test_decoding/expected/sequence.out | 18 ++++++++++++------
1 file changed, 12 insertions(+), 6 deletions(-)
diff --git a/contrib/test_decoding/expected/sequence.out b/contrib/test_decoding/expected/sequence.out
index ca55bcebc3..d94e185fb9 100644
--- a/contrib/test_decoding/expected/sequence.out
+++ b/contrib/test_decoding/expected/sequence.out
@@ -260,13 +260,15 @@ ROLLBACK TO SAVEPOINT a;
DROP TABLE test_table;
COMMIT;
SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
- data
---------------------------------------------------------------
+ data
+-------------------------------------------------------------------------------------------------
BEGIN
+ sequence public.test_table_a_seq: transactional:1 created:1 last_value:1 log_cnt:0 is_called:0
+ sequence public.test_table_a_seq: transactional:1 created:0 last_value:33 log_cnt:0 is_called:1
table public.test_table: INSERT: a[integer]:1 b[integer]:100
table public.test_table: INSERT: a[integer]:2 b[integer]:200
COMMIT
-(4 rows)
+(6 rows)
-- savepoint test on table with serial column
BEGIN;
@@ -307,11 +309,15 @@ SELECT * FROM test_sequence;
DROP SEQUENCE test_sequence;
COMMIT;
SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
- data
---------
+ data
+------------------------------------------------------------------------------------------------
BEGIN
+ sequence public.test_sequence: transactional:1 created:1 last_value:1 log_cnt:0 is_called:0
+ sequence public.test_sequence: transactional:1 created:0 last_value:33 log_cnt:0 is_called:1
+ sequence public.test_sequence: transactional:1 created:0 last_value:3000 log_cnt:0 is_called:1
+ sequence public.test_sequence: transactional:1 created:0 last_value:3033 log_cnt:0 is_called:1
COMMIT
-(2 rows)
+(6 rows)
SELECT pg_drop_replication_slot('regression_slot');
pg_drop_replication_slot
--
2.34.1
[text/plain] 0003-Some-documentation-updates.patch (5.6K, ../../[email protected]/4-0003-Some-documentation-updates.patch)
download | inline diff:
From 0338de62cd955b98244b5b624a2bcb7af76465b0 Mon Sep 17 00:00:00 2001
From: Peter Eisentraut <[email protected]>
Date: Tue, 7 Dec 2021 15:29:51 +0100
Subject: [PATCH 3/4] Some documentation updates
---
doc/src/sgml/logicaldecoding.sgml | 54 +++++++++++++++++++++++++++++--
1 file changed, 51 insertions(+), 3 deletions(-)
diff --git a/doc/src/sgml/logicaldecoding.sgml b/doc/src/sgml/logicaldecoding.sgml
index b6353c7a12..eea4f3cd50 100644
--- a/doc/src/sgml/logicaldecoding.sgml
+++ b/doc/src/sgml/logicaldecoding.sgml
@@ -458,6 +458,7 @@ <title>Initialization Function</title>
LogicalDecodeTruncateCB truncate_cb;
LogicalDecodeCommitCB commit_cb;
LogicalDecodeMessageCB message_cb;
+ LogicalDecodeSequenceCB sequence_cb;
LogicalDecodeFilterByOriginCB filter_by_origin_cb;
LogicalDecodeShutdownCB shutdown_cb;
LogicalDecodeFilterPrepareCB filter_prepare_cb;
@@ -472,6 +473,7 @@ <title>Initialization Function</title>
LogicalDecodeStreamCommitCB stream_commit_cb;
LogicalDecodeStreamChangeCB stream_change_cb;
LogicalDecodeStreamMessageCB stream_message_cb;
+ LogicalDecodeStreamSequenceCB stream_sequence_cb;
LogicalDecodeStreamTruncateCB stream_truncate_cb;
} OutputPluginCallbacks;
@@ -481,9 +483,12 @@ <title>Initialization Function</title>
and <function>commit_cb</function> callbacks are required,
while <function>startup_cb</function>,
<function>filter_by_origin_cb</function>, <function>truncate_cb</function>,
+ <function>sequence_cb</function>,
and <function>shutdown_cb</function> are optional.
If <function>truncate_cb</function> is not set but a
<command>TRUNCATE</command> is to be decoded, the action will be ignored.
+ Similarly, if <function>sequence_cb</function> is not set and a sequence
+ change is to be decoded, the action will be ignored.
</para>
<para>
@@ -492,7 +497,7 @@ <title>Initialization Function</title>
<function>stream_stop_cb</function>, <function>stream_abort_cb</function>,
<function>stream_commit_cb</function>, <function>stream_change_cb</function>,
and <function>stream_prepare_cb</function>
- are required, while <function>stream_message_cb</function> and
+ are required, while <function>stream_message_cb</function>, <function>stream_sequence_cb</function>, and
<function>stream_truncate_cb</function> are optional.
</para>
@@ -808,6 +813,27 @@ <title>Generic Message Callback</title>
</para>
</sect3>
+ <sect3 id="logicaldecoding-output-plugin-sequence">
+ <title>Sequence Callback</title>
+
+ <para>
+ The <function>sequence_cb</function> callback is called for actions that
+ change a sequence.
+<programlisting>
+typedef void (*LogicalDecodeSequenceCB) (struct LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional,
+ bool created,
+ int64 last_value,
+ int64 log_cnt,
+ bool is_called);
+</programlisting>
+ TODO
+ </para>
+ </sect3>
+
<sect3 id="logicaldecoding-output-plugin-filter-prepare">
<title>Prepare Filter Callback</title>
@@ -1017,6 +1043,28 @@ <title>Stream Message Callback</title>
</para>
</sect3>
+ <sect3 id="logicaldecoding-output-plugin-stream-sequence">
+ <title>Stream Sequence Callback</title>
+ <para>
+ The <function>stream_sequence_cb</function> callback is called for
+ actions that change a sequence in a block of streamed changes
+ (demarcated by <function>stream_start_cb</function> and
+ <function>stream_stop_cb</function> calls).
+<programlisting>
+typedef void (*LogicalDecodeStreamSequenceCB) (struct LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional,
+ bool created,
+ int64 last_value,
+ int64 log_cnt,
+ bool is_called);
+</programlisting>
+ TODO
+ </para>
+ </sect3>
+
<sect3 id="logicaldecoding-output-plugin-stream-truncate">
<title>Stream Truncate Callback</title>
<para>
@@ -1197,8 +1245,8 @@ <title>Streaming of Large Transactions for Logical Decoding</title>
in-progress transactions. There are multiple required streaming callbacks
(<function>stream_start_cb</function>, <function>stream_stop_cb</function>,
<function>stream_abort_cb</function>, <function>stream_commit_cb</function>
- and <function>stream_change_cb</function>) and two optional callbacks
- (<function>stream_message_cb</function> and <function>stream_truncate_cb</function>).
+ and <function>stream_change_cb</function>) and multiple optional callbacks
+ (<function>stream_message_cb</function>, <function>stream_sequence_cb</function>, and <function>stream_truncate_cb</function>).
Also, if streaming of two-phase commands is to be supported, then additional
callbacks must be provided. (See <xref linkend="logicaldecoding-two-phase-commits"/>
for details).
--
2.34.1
[text/plain] 0004-Simplify-accessing-sequence-tuple-data.patch (3.1K, ../../[email protected]/5-0004-Simplify-accessing-sequence-tuple-data.patch)
download | inline diff:
From 9f29fac7e479c255fb247d201fe8691655c11fbf Mon Sep 17 00:00:00 2001
From: Peter Eisentraut <[email protected]>
Date: Tue, 7 Dec 2021 15:31:36 +0100
Subject: [PATCH 4/4] Simplify accessing sequence tuple data
---
.../replication/logical/reorderbuffer.c | 37 ++++---------------
1 file changed, 8 insertions(+), 29 deletions(-)
diff --git a/src/backend/replication/logical/reorderbuffer.c b/src/backend/replication/logical/reorderbuffer.c
index 434926459f..b4fa16bebe 100644
--- a/src/backend/replication/logical/reorderbuffer.c
+++ b/src/backend/replication/logical/reorderbuffer.c
@@ -120,6 +120,7 @@
#include "access/xact.h"
#include "access/xlog_internal.h"
#include "catalog/catalog.h"
+#include "commands/sequence.h"
#include "lib/binaryheap.h"
#include "miscadmin.h"
#include "pgstat.h"
@@ -1089,9 +1090,7 @@ ReorderBufferQueueSequence(ReorderBuffer *rb, TransactionId xid,
{
Relation relation;
HeapTuple tuple;
- bool isnull;
- int64 last_value, log_cnt;
- bool is_called;
+ Form_pg_sequence_data seq;
Oid reloid;
if (using_subtxn)
@@ -1108,19 +1107,10 @@ ReorderBufferQueueSequence(ReorderBuffer *rb, TransactionId xid,
relation = RelationIdGetRelation(reloid);
tuple = &tuplebuf->tuple;
-
- /*
- * Extract the internal sequence values, describing the state.
- *
- * XXX Seems a bit strange to access it directly. Maybe there's
- * a better / more correct way?
- */
- last_value = heap_getattr(tuple, 1, RelationGetDescr(relation), &isnull);
- log_cnt = heap_getattr(tuple, 2, RelationGetDescr(relation), &isnull);
- is_called = heap_getattr(tuple, 3, RelationGetDescr(relation), &isnull);
+ seq = (Form_pg_sequence_data) GETSTRUCT(tuple);
rb->sequence(rb, txn, lsn, relation, transactional, created,
- last_value, log_cnt, is_called);
+ seq->last_value, seq->log_cnt, seq->is_called);
RelationClose(relation);
@@ -2249,31 +2239,20 @@ ReorderBufferApplySequence(ReorderBuffer *rb, ReorderBufferTXN *txn,
bool streaming)
{
HeapTuple tuple;
- bool isnull;
- int64 last_value, log_cnt;
- bool is_called;
+ Form_pg_sequence_data seq;
tuple = &change->data.sequence.tuple->tuple;
-
- /*
- * Extract the internal sequence values, describing the state.
- *
- * XXX Seems a bit strange to access it directly. Maybe there's
- * a better / more correct way?
- */
- last_value = heap_getattr(tuple, 1, RelationGetDescr(relation), &isnull);
- log_cnt = heap_getattr(tuple, 2, RelationGetDescr(relation), &isnull);
- is_called = heap_getattr(tuple, 3, RelationGetDescr(relation), &isnull);
+ seq = (Form_pg_sequence_data) GETSTRUCT(tuple);
/* Only ever called from ReorderBufferApplySequence, so transational. */
if (streaming)
rb->stream_sequence(rb, txn, change->lsn, relation, true,
change->data.sequence.created,
- last_value, log_cnt, is_called);
+ seq->last_value, seq->log_cnt, seq->is_called);
else
rb->sequence(rb, txn, change->lsn, relation, true,
change->data.sequence.created,
- last_value, log_cnt, is_called);
+ seq->last_value, seq->log_cnt, seq->is_called);
}
/*
--
2.34.1
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2021-12-08 00:23 Tomas Vondra <[email protected]>
parent: Peter Eisentraut <[email protected]>
0 siblings, 1 reply; 127+ messages in thread
From: Tomas Vondra @ 2021-12-08 00:23 UTC (permalink / raw)
To: Peter Eisentraut <[email protected]>; +Cc: PostgreSQL Hackers <[email protected]>; Petr Jelinek <[email protected]>
On 12/7/21 15:38, Peter Eisentraut wrote:
> I have checked the 0001 and 0003 patches. (I think we have dismissed
> the approach in 0002 for now. And let's talk about 0004 later.)
>
Right, I think that's correct.
> I have attached a few fixup patches, described further below.
>
> # 0001
>
> The argument "create" for fill_seq_with_data() is always true (and
> patch 0002 removes it). So I'm not sure if it's needed. If it is, it
> should be documented somewhere.
>
Good point. I think it could be removed, but IIRC it'll be needed when
adding sequence decoding to the built-in replication, and that patch is
meant to be just an implementation of the API, without changing WAL.
OTOH I don't see it in the last version of that patch (in ResetSequence2
call) so maybe it's not really needed. I've kept it for now, but I'll
investigate.
> About the comment you added in nextval_internal(): It's a good
> explanation, so I would leave it in. I also agree with the
> conclusion of the comment that the current solution is reasonable. We
> probably don't need the same comment again in fill_seq_with_data() and
> again in do_setval(). Note that both of the latter functions already
> point to nextval_interval() for other comments, so the same can be
> relied upon here as well.
>
True. I moved it a bit in nextval_internal() and removed the other
copies. The existing references should be enough.
> The order of the new fields sequence_cb and stream_sequence_cb is a
> bit inconsistent compared to truncate_cb and stream_truncate_cb.
> Maybe take another look to make the order more uniform.
>
You mean in OutputPluginCallbacks? I'd actually argue it's the truncate
callbacks that are inconsistent - in the regular section truncate_cb is
right before commit_cb, while in the streaming section it's at the end.
Or what order do you think would be better? I'm fine with changing it.
> Some documentation needs to be added to the Logical Decoding chapter.
> I have attached a patch that I think catches all the places that need
> to be updated, with some details left for you to fill in. ;-) The
> ordering of the some of the documentation chunks reflects the order in
> which the callbacks appear in the header files, which might not be
> optimal; see above.
>
Thanks. I added a bit about the callbacks being optional and what the
parameters mean (only for sequence_cb, as the stream_ callbacks
generally don't copy that bit).
> In reorderbuffer.c, you left a comment about how to access a sequence
> tuple. There is an easier way, using Form_pg_sequence_data, which is
> how sequence.c also does it. See attached patch.
>
Yeah, that looks much nicer.
> # 0003
>
> The tests added in 0003 don't pass for me. It appears that you might
> have forgotten to update the expected files after you added some tests
> or something. The cfbot shows the same. See attached patch for a
> correction, but do check what your intent was.
>
Yeah. I suspect I removed the expected results due to the experimental
rework. I haven't noticed that because some of the tests for the
built-in replication are expected to fail.
Attached is an updated version of the first two parts (infrastructure
and test_decoding), with all your fixes merged.
regards
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
Attachments:
[text/x-patch] 0001-Logical-decoding-of-sequences-20211208.patch (47.2K, ../../[email protected]/2-0001-Logical-decoding-of-sequences-20211208.patch)
download | inline diff:
From 999514b9f802db9b888543532355e9c36646da30 Mon Sep 17 00:00:00 2001
From: Tomas Vondra <[email protected]>
Date: Fri, 24 Sep 2021 00:41:33 +0200
Subject: [PATCH 1/2] Logical decoding of sequences
This adds the infrastructure for logical decoding of sequences. Support
for built-in logical replication and test_decoding is added separately.
When decoding sequences, we differentiate between a sequences created in
a (running) transaction, and sequences created in other (already
committed) transactions. Changes for sequences created in the same
top-level transaction are treated as "transactional" i.e. just like any
other change from that transaction (and discarded in case of a
rollback). Changes for sequences created earlier are treated as not
transactional - are processed immediately, as if performed outside any
transaction (and thus not rolled back).
This mixed behavior is necessary - sequences are non-transactional (e.g.
ROLLBACK does not undo the sequence increments). But for new sequences,
we need to handle them in a transactional way, because if we ever get
some DDL support, the sequence won't exist until the transaction gets
applied. So we need to ensure the increments don't happen until the
sequence gets created.
To differentiate which sequences are "old" and which were created in a
still-running transaction, we track sequences created in running
transactions in a hash table. Each sequence is identified by it's
relfilenode, and at transaction commit we discard all sequences created
in that particular transaction. For each sequence we track the XID of
the (sub)transaction that created it, and we cleanup the sequences for
each subtransaction when it completes (commit/rollback).
We don't use the XID to check if it's the same top-level transaction.
It's enough to know it was created in an in-progress transaction, and we
know it must be the current one because otherwise it wouldn't see the
sequence object.
The main changes in this patch are:
1) ensure WAL-logging of all necessary info for sequence advances
We need to be able to associate the advance with a XID, but until now
sequence advance might have XID 0 if it was the first thing the
transaction did. So we ensure the transaction has XID.
Note: This is needed because of subxacts. A XID 0 might still have the
sequence created in a different subxact of the same top-level xact.
Then there's a "created" flag added to the xl_seq_rec, which is
necessary to differentiate WAL for created/reset sequences.
2) decoding / queueing of sequences into ReorderBuffer
This is mostly copy-paste of existing code for other decoded events.
3) tracking sequences created in in-progress transactions
We use a simple hash table, indexed by relfilenode.
---
doc/src/sgml/logicaldecoding.sgml | 65 ++-
src/backend/commands/sequence.c | 40 +-
src/backend/replication/logical/decode.c | 131 +++++-
src/backend/replication/logical/logical.c | 89 ++++
.../replication/logical/reorderbuffer.c | 403 ++++++++++++++++++
src/include/commands/sequence.h | 1 +
src/include/replication/output_plugin.h | 29 ++
src/include/replication/reorderbuffer.h | 44 +-
8 files changed, 791 insertions(+), 11 deletions(-)
diff --git a/doc/src/sgml/logicaldecoding.sgml b/doc/src/sgml/logicaldecoding.sgml
index b6353c7a12..4ede56678c 100644
--- a/doc/src/sgml/logicaldecoding.sgml
+++ b/doc/src/sgml/logicaldecoding.sgml
@@ -458,6 +458,7 @@ typedef struct OutputPluginCallbacks
LogicalDecodeTruncateCB truncate_cb;
LogicalDecodeCommitCB commit_cb;
LogicalDecodeMessageCB message_cb;
+ LogicalDecodeSequenceCB sequence_cb;
LogicalDecodeFilterByOriginCB filter_by_origin_cb;
LogicalDecodeShutdownCB shutdown_cb;
LogicalDecodeFilterPrepareCB filter_prepare_cb;
@@ -472,6 +473,7 @@ typedef struct OutputPluginCallbacks
LogicalDecodeStreamCommitCB stream_commit_cb;
LogicalDecodeStreamChangeCB stream_change_cb;
LogicalDecodeStreamMessageCB stream_message_cb;
+ LogicalDecodeStreamSequenceCB stream_sequence_cb;
LogicalDecodeStreamTruncateCB stream_truncate_cb;
} OutputPluginCallbacks;
@@ -481,9 +483,12 @@ typedef void (*LogicalOutputPluginInit) (struct OutputPluginCallbacks *cb);
and <function>commit_cb</function> callbacks are required,
while <function>startup_cb</function>,
<function>filter_by_origin_cb</function>, <function>truncate_cb</function>,
+ <function>sequence_cb</function>,
and <function>shutdown_cb</function> are optional.
If <function>truncate_cb</function> is not set but a
<command>TRUNCATE</command> is to be decoded, the action will be ignored.
+ Similarly, if <function>sequence_cb</function> is not set and a sequence
+ change is to be decoded, the action will be ignored.
</para>
<para>
@@ -492,7 +497,8 @@ typedef void (*LogicalOutputPluginInit) (struct OutputPluginCallbacks *cb);
<function>stream_stop_cb</function>, <function>stream_abort_cb</function>,
<function>stream_commit_cb</function>, <function>stream_change_cb</function>,
and <function>stream_prepare_cb</function>
- are required, while <function>stream_message_cb</function> and
+ are required, while <function>stream_message_cb</function>,
+ <function>stream_sequence_cb</function>, and
<function>stream_truncate_cb</function> are optional.
</para>
@@ -808,6 +814,37 @@ typedef void (*LogicalDecodeMessageCB) (struct LogicalDecodingContext *ctx,
</para>
</sect3>
+ <sect3 id="logicaldecoding-output-plugin-sequence">
+ <title>Sequence Callback</title>
+
+ <para>
+ The optional <function>sequence_cb</function> callback is called for
+ actions that update a sequence value.
+<programlisting>
+typedef void (*LogicalDecodeSequenceCB) (struct LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional,
+ bool created,
+ int64 last_value,
+ int64 log_cnt,
+ bool is_called);
+</programlisting>
+ The <parameter>txn</parameter> parameter contains meta information about
+ the transaction the sequence change is part of. Note however that it can
+ be NULL when the sequence change is non-transactional and the XID was not
+ assigned yet in the transaction which updated the sequence value.
+ The <parameter>sequence_lsn</parameter> has WAL location of the sequence
+ update. The <parameter>transactional</parameter> says if the sequence has
+ to be replayed as part of the transaction or directly.
+
+ The <parameter>created</parameter>, <parameter>last_value</parameter>,
+ <parameter>log_cnt</parameter> and <parameter>is_called</parameter> parameters
+ describe the sequence change.
+ </para>
+ </sect3>
+
<sect3 id="logicaldecoding-output-plugin-filter-prepare">
<title>Prepare Filter Callback</title>
@@ -1017,6 +1054,27 @@ typedef void (*LogicalDecodeStreamMessageCB) (struct LogicalDecodingContext *ctx
</para>
</sect3>
+ <sect3 id="logicaldecoding-output-plugin-stream-sequence">
+ <title>Stream Sequence Callback</title>
+ <para>
+ The optional <function>stream_sequence_cb</function> callback is called
+ for actions that change a sequence in a block of streamed changes
+ (demarcated by <function>stream_start_cb</function> and
+ <function>stream_stop_cb</function> calls).
+<programlisting>
+typedef void (*LogicalDecodeStreamSequenceCB) (struct LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional,
+ bool created,
+ int64 last_value,
+ int64 log_cnt,
+ bool is_called);
+</programlisting>
+ </para>
+ </sect3>
+
<sect3 id="logicaldecoding-output-plugin-stream-truncate">
<title>Stream Truncate Callback</title>
<para>
@@ -1197,8 +1255,9 @@ OutputPluginWrite(ctx, true);
in-progress transactions. There are multiple required streaming callbacks
(<function>stream_start_cb</function>, <function>stream_stop_cb</function>,
<function>stream_abort_cb</function>, <function>stream_commit_cb</function>
- and <function>stream_change_cb</function>) and two optional callbacks
- (<function>stream_message_cb</function> and <function>stream_truncate_cb</function>).
+ and <function>stream_change_cb</function>) and multiple optional callbacks
+ (<function>stream_message_cb</function>, <function>stream_sequence_cb</function>,
+ and <function>stream_truncate_cb</function>).
Also, if streaming of two-phase commands is to be supported, then additional
callbacks must be provided. (See <xref linkend="logicaldecoding-two-phase-commits"/>
for details).
diff --git a/src/backend/commands/sequence.c b/src/backend/commands/sequence.c
index 72bfdc07a4..c47c09a94e 100644
--- a/src/backend/commands/sequence.c
+++ b/src/backend/commands/sequence.c
@@ -94,7 +94,7 @@ static HTAB *seqhashtab = NULL; /* hash table for SeqTable items */
*/
static SeqTableData *last_used_seq = NULL;
-static void fill_seq_with_data(Relation rel, HeapTuple tuple);
+static void fill_seq_with_data(Relation rel, HeapTuple tuple, bool create);
static Relation lock_and_open_sequence(SeqTable seq);
static void create_seq_hashtable(void);
static void init_sequence(Oid relid, SeqTable *p_elm, Relation *p_rel);
@@ -222,7 +222,7 @@ DefineSequence(ParseState *pstate, CreateSeqStmt *seq)
/* now initialize the sequence's data */
tuple = heap_form_tuple(tupDesc, value, null);
- fill_seq_with_data(rel, tuple);
+ fill_seq_with_data(rel, tuple, true);
/* process OWNED BY if given */
if (owned_by)
@@ -327,7 +327,7 @@ ResetSequence(Oid seq_relid)
/*
* Insert the modified tuple into the new storage file.
*/
- fill_seq_with_data(seq_rel, tuple);
+ fill_seq_with_data(seq_rel, tuple, true);
/* Clear local cache so that we don't think we have cached numbers */
/* Note that we do not change the currval() state */
@@ -340,7 +340,7 @@ ResetSequence(Oid seq_relid)
* Initialize a sequence's relation with the specified tuple as content
*/
static void
-fill_seq_with_data(Relation rel, HeapTuple tuple)
+fill_seq_with_data(Relation rel, HeapTuple tuple, bool create)
{
Buffer buf;
Page page;
@@ -378,7 +378,10 @@ fill_seq_with_data(Relation rel, HeapTuple tuple)
/* check the comment above nextval_internal()'s equivalent call. */
if (RelationNeedsWAL(rel))
+ {
GetTopTransactionId();
+ GetCurrentTransactionId();
+ }
START_CRIT_SECTION();
@@ -399,6 +402,7 @@ fill_seq_with_data(Relation rel, HeapTuple tuple)
XLogRegisterBuffer(0, buf, REGBUF_WILL_INIT);
xlrec.node = rel->rd_node;
+ xlrec.created = create;
XLogRegisterData((char *) &xlrec, sizeof(xl_seq_rec));
XLogRegisterData((char *) tuple->t_data, tuple->t_len);
@@ -502,7 +506,7 @@ AlterSequence(ParseState *pstate, AlterSeqStmt *stmt)
/*
* Insert the modified tuple into the new storage file.
*/
- fill_seq_with_data(seqrel, newdatatuple);
+ fill_seq_with_data(seqrel, newdatatuple, true);
}
/* process OWNED BY if given */
@@ -764,9 +768,29 @@ nextval_internal(Oid relid, bool check_permissions)
* It's sufficient to ensure the toplevel transaction has an xid, no need
* to assign xids subxacts, that'll already trigger an appropriate wait.
* (Have to do that here, so we're outside the critical section)
+ *
+ * We have to ensure we have a proper XID, which will be included in
+ * the XLOG record by XLogRecordAssemble. Otherwise the first nextval()
+ * in a subxact (without any preceding changes) would get XID 0, and it
+ * would then be impossible to decide which top xact it belongs to.
+ * It'd also trigger assert in DecodeSequence.
+ *
+ * XXX This might seem unnecessary, because if there's no XID the xact
+ * couldn't have done anything important yet, e.g. it could not have
+ * created a sequence. But that's incorrect, as it ignores subxacts. The
+ * current subtransaction might not have done anything yet (thus no XID),
+ * but an earlier one might have created the sequence.
+ *
+ * XXX Not sure if this is the best solution. Maybe do this only with
+ * wal_level=logical to minimize the overhead. OTOH advancing the
+ * sequence is likely followed by using the value(s) for some other
+ * activity, which assigns the XID.
*/
if (logit && RelationNeedsWAL(seqrel))
+ {
GetTopTransactionId();
+ GetCurrentTransactionId();
+ }
/* ready to change the on-disk (or really, in-buffer) tuple */
START_CRIT_SECTION();
@@ -803,6 +827,7 @@ nextval_internal(Oid relid, bool check_permissions)
seq->log_cnt = 0;
xlrec.node = seqrel->rd_node;
+ xlrec.created = false;
XLogRegisterData((char *) &xlrec, sizeof(xl_seq_rec));
XLogRegisterData((char *) seqdatatuple.t_data, seqdatatuple.t_len);
@@ -977,7 +1002,10 @@ do_setval(Oid relid, int64 next, bool iscalled)
/* check the comment above nextval_internal()'s equivalent call. */
if (RelationNeedsWAL(seqrel))
+ {
GetTopTransactionId();
+ GetCurrentTransactionId();
+ }
/* ready to change the on-disk (or really, in-buffer) tuple */
START_CRIT_SECTION();
@@ -999,6 +1027,8 @@ do_setval(Oid relid, int64 next, bool iscalled)
XLogRegisterBuffer(0, buf, REGBUF_WILL_INIT);
xlrec.node = seqrel->rd_node;
+ xlrec.created = false;
+
XLogRegisterData((char *) &xlrec, sizeof(xl_seq_rec));
XLogRegisterData((char *) seqdatatuple.t_data, seqdatatuple.t_len);
diff --git a/src/backend/replication/logical/decode.c b/src/backend/replication/logical/decode.c
index a2b69511b4..b61f2e264a 100644
--- a/src/backend/replication/logical/decode.c
+++ b/src/backend/replication/logical/decode.c
@@ -42,6 +42,7 @@
#include "replication/reorderbuffer.h"
#include "replication/snapbuild.h"
#include "storage/standby.h"
+#include "commands/sequence.h"
typedef struct XLogRecordBuffer
{
@@ -74,10 +75,11 @@ static void DecodeAbort(LogicalDecodingContext *ctx, XLogRecordBuffer *buf,
bool two_phase);
static void DecodePrepare(LogicalDecodingContext *ctx, XLogRecordBuffer *buf,
xl_xact_parsed_prepare *parsed);
-
+static void DecodeSequence(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
/* common function to decode tuples */
static void DecodeXLogTuple(char *data, Size len, ReorderBufferTupleBuf *tup);
+static void DecodeSeqTuple(char *data, Size len, ReorderBufferTupleBuf *tuple);
/* helper functions for decoding transactions */
static inline bool FilterPrepare(LogicalDecodingContext *ctx,
@@ -158,6 +160,10 @@ LogicalDecodingProcessRecord(LogicalDecodingContext *ctx, XLogReaderState *recor
DecodeLogicalMsgOp(ctx, &buf);
break;
+ case RM_SEQ_ID:
+ DecodeSequence(ctx, &buf);
+ break;
+
/*
* Rmgrs irrelevant for logical decoding; they describe stuff not
* represented in logical decoding. Add new rmgrs in rmgrlist.h's
@@ -173,7 +179,6 @@ LogicalDecodingProcessRecord(LogicalDecodingContext *ctx, XLogReaderState *recor
case RM_HASH_ID:
case RM_GIN_ID:
case RM_GIST_ID:
- case RM_SEQ_ID:
case RM_SPGIST_ID:
case RM_BRIN_ID:
case RM_COMMIT_TS_ID:
@@ -1313,3 +1318,125 @@ DecodeTXNNeedSkip(LogicalDecodingContext *ctx, XLogRecordBuffer *buf,
(txn_dbid != InvalidOid && txn_dbid != ctx->slot->data.database) ||
ctx->fast_forward || FilterByOrigin(ctx, origin_id));
}
+
+/*
+ * Decode Sequence Tuple
+ */
+static void
+DecodeSeqTuple(char *data, Size len, ReorderBufferTupleBuf *tuple)
+{
+ int datalen = len - sizeof(xl_seq_rec) - SizeofHeapTupleHeader;
+
+ Assert(datalen >= 0);
+
+ tuple->tuple.t_len = datalen + SizeofHeapTupleHeader;;
+
+ ItemPointerSetInvalid(&tuple->tuple.t_self);
+
+ tuple->tuple.t_tableOid = InvalidOid;
+
+ memcpy(((char *) tuple->tuple.t_data),
+ data + sizeof(xl_seq_rec),
+ SizeofHeapTupleHeader);
+
+ memcpy(((char *) tuple->tuple.t_data) + SizeofHeapTupleHeader,
+ data + sizeof(xl_seq_rec) + SizeofHeapTupleHeader,
+ datalen);
+}
+
+/*
+ * Handle sequence decode
+ *
+ * Decoding sequences is a bit tricky, because while most sequence actions
+ * are non-transactional (not subject to rollback), some need to be handled
+ * as transactional.
+ *
+ * By default, a sequence increment is non-transactional - we must not queue
+ * it in a transaction as other changes, because the transaction might get
+ * rolled back and we'd discard the increment. The downstream would not be
+ * notified about the increment, which is wrong.
+ *
+ * On the other hand, the sequence may be created in a transaction. In this
+ * case we *should* queue the change as other changes in the transaction,
+ * because we don't want to send the increments for unknown sequence to the
+ * plugin - it might get confused about which sequence it's related to etc.
+ */
+static void
+DecodeSequence(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
+{
+ SnapBuild *builder = ctx->snapshot_builder;
+ ReorderBufferTupleBuf *tuplebuf;
+ RelFileNode target_node;
+ XLogReaderState *r = buf->record;
+ char *tupledata = NULL;
+ Size tuplelen;
+ Size datalen = 0;
+ TransactionId xid = XLogRecGetXid(r);
+ uint8 info = XLogRecGetInfo(buf->record) & ~XLR_INFO_MASK;
+ xl_seq_rec *xlrec;
+ Snapshot snapshot;
+ RepOriginId origin_id = XLogRecGetOrigin(r);
+ bool transactional;
+
+ /* only decode changes flagged with XLOG_SEQ_LOG */
+ if (info != XLOG_SEQ_LOG)
+ elog(ERROR, "unexpected RM_SEQ_ID record type: %u", info);
+
+ /*
+ * If we don't have snapshot or we are just fast-forwarding, there is no
+ * point in decoding messages.
+ */
+ if (SnapBuildCurrentState(builder) < SNAPBUILD_FULL_SNAPSHOT ||
+ ctx->fast_forward)
+ return;
+
+ /* only interested in our database */
+ XLogRecGetBlockTag(r, 0, &target_node, NULL, NULL);
+ if (target_node.dbNode != ctx->slot->data.database)
+ return;
+
+ /* output plugin doesn't look for this origin, no need to queue */
+ if (FilterByOrigin(ctx, XLogRecGetOrigin(r)))
+ return;
+
+ tupledata = XLogRecGetData(r);
+ datalen = XLogRecGetDataLen(r);
+ tuplelen = datalen - SizeOfHeapHeader - sizeof(xl_seq_rec);
+
+ /* extract the WAL record, with "created" flag */
+ xlrec = (xl_seq_rec *) XLogRecGetData(r);
+
+ /* XXX how could we have sequence change without data? */
+ if(!datalen || !tupledata)
+ return;
+
+ tuplebuf = ReorderBufferGetTupleBuf(ctx->reorder, tuplelen);
+ DecodeSeqTuple(tupledata, datalen, tuplebuf);
+
+ /*
+ * Should we handle the sequence increment as transactional or not?
+ *
+ * If the sequence was created in a still-running transaction, treat
+ * it as transactional and queue the increments. Otherwise it needs
+ * to be treated as non-transactional, in which case we send it to
+ * the plugin right away.
+ */
+ transactional = ReorderBufferSequenceIsTransactional(ctx->reorder,
+ target_node,
+ xlrec->created);
+
+ /* Skip the change if already processed (per the snapshot). */
+ if (transactional &&
+ !SnapBuildProcessChange(builder, xid, buf->origptr))
+ return;
+ else if (!transactional &&
+ (SnapBuildCurrentState(builder) != SNAPBUILD_CONSISTENT ||
+ SnapBuildXactNeedsSkip(builder, buf->origptr)))
+ return;
+
+ /* Queue the increment (or send immediately if not transactional). */
+ snapshot = SnapBuildGetOrBuildSnapshot(builder, xid);
+ ReorderBufferQueueSequence(ctx->reorder, xid, snapshot, buf->endptr,
+ origin_id, target_node, transactional,
+ xlrec->created, tuplebuf);
+}
diff --git a/src/backend/replication/logical/logical.c b/src/backend/replication/logical/logical.c
index 10cbdea124..582e8c2f7d 100644
--- a/src/backend/replication/logical/logical.c
+++ b/src/backend/replication/logical/logical.c
@@ -73,6 +73,10 @@ static void truncate_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
static void message_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
XLogRecPtr message_lsn, bool transactional,
const char *prefix, Size message_size, const char *message);
+static void sequence_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ bool transactional, bool created,
+ int64 last_value, int64 log_cnt, bool is_called);
/* streaming callbacks */
static void stream_start_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
@@ -90,6 +94,10 @@ static void stream_change_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn
static void stream_message_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
XLogRecPtr message_lsn, bool transactional,
const char *prefix, Size message_size, const char *message);
+static void stream_sequence_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ bool transactional, bool created,
+ int64 last_value, int64 log_cnt, bool is_called);
static void stream_truncate_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
int nrelations, Relation relations[], ReorderBufferChange *change);
@@ -218,6 +226,7 @@ StartupDecodingContext(List *output_plugin_options,
ctx->reorder->apply_truncate = truncate_cb_wrapper;
ctx->reorder->commit = commit_cb_wrapper;
ctx->reorder->message = message_cb_wrapper;
+ ctx->reorder->sequence = sequence_cb_wrapper;
/*
* To support streaming, we require start/stop/abort/commit/change
@@ -234,6 +243,7 @@ StartupDecodingContext(List *output_plugin_options,
(ctx->callbacks.stream_commit_cb != NULL) ||
(ctx->callbacks.stream_change_cb != NULL) ||
(ctx->callbacks.stream_message_cb != NULL) ||
+ (ctx->callbacks.stream_sequence_cb != NULL) ||
(ctx->callbacks.stream_truncate_cb != NULL);
/*
@@ -251,6 +261,7 @@ StartupDecodingContext(List *output_plugin_options,
ctx->reorder->stream_commit = stream_commit_cb_wrapper;
ctx->reorder->stream_change = stream_change_cb_wrapper;
ctx->reorder->stream_message = stream_message_cb_wrapper;
+ ctx->reorder->stream_sequence = stream_sequence_cb_wrapper;
ctx->reorder->stream_truncate = stream_truncate_cb_wrapper;
@@ -1203,6 +1214,43 @@ message_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
error_context_stack = errcallback.previous;
}
+static void
+sequence_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ bool transactional, bool created,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ LogicalDecodingContext *ctx = cache->private_data;
+ LogicalErrorCallbackState state;
+ ErrorContextCallback errcallback;
+
+ Assert(!ctx->fast_forward);
+
+ if (ctx->callbacks.sequence_cb == NULL)
+ return;
+
+ /* Push callback + info on the error context stack */
+ state.ctx = ctx;
+ state.callback_name = "sequence";
+ state.report_location = sequence_lsn;
+ errcallback.callback = output_plugin_error_callback;
+ errcallback.arg = (void *) &state;
+ errcallback.previous = error_context_stack;
+ error_context_stack = &errcallback;
+
+ /* set output state */
+ ctx->accept_writes = true;
+ ctx->write_xid = txn != NULL ? txn->xid : InvalidTransactionId;
+ ctx->write_location = sequence_lsn;
+
+ /* do the actual work: call callback */
+ ctx->callbacks.sequence_cb(ctx, txn, sequence_lsn, rel, transactional,
+ created, last_value, log_cnt, is_called);
+
+ /* Pop the error context stack */
+ error_context_stack = errcallback.previous;
+}
+
static void
stream_start_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
XLogRecPtr first_lsn)
@@ -1508,6 +1556,47 @@ stream_message_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
error_context_stack = errcallback.previous;
}
+static void
+stream_sequence_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ bool transactional, bool created,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ LogicalDecodingContext *ctx = cache->private_data;
+ LogicalErrorCallbackState state;
+ ErrorContextCallback errcallback;
+
+ Assert(!ctx->fast_forward);
+
+ /* We're only supposed to call this when streaming is supported. */
+ Assert(ctx->streaming);
+
+ /* this callback is optional */
+ if (ctx->callbacks.stream_sequence_cb == NULL)
+ return;
+
+ /* Push callback + info on the error context stack */
+ state.ctx = ctx;
+ state.callback_name = "stream_sequence";
+ state.report_location = sequence_lsn;
+ errcallback.callback = output_plugin_error_callback;
+ errcallback.arg = (void *) &state;
+ errcallback.previous = error_context_stack;
+ error_context_stack = &errcallback;
+
+ /* set output state */
+ ctx->accept_writes = true;
+ ctx->write_xid = txn != NULL ? txn->xid : InvalidTransactionId;
+ ctx->write_location = sequence_lsn;
+
+ /* do the actual work: call callback */
+ ctx->callbacks.sequence_cb(ctx, txn, sequence_lsn, rel, transactional,
+ created, last_value, log_cnt, is_called);
+
+ /* Pop the error context stack */
+ error_context_stack = errcallback.previous;
+}
+
static void
stream_truncate_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
int nrelations, Relation relations[],
diff --git a/src/backend/replication/logical/reorderbuffer.c b/src/backend/replication/logical/reorderbuffer.c
index 46e66608cf..b4fa16bebe 100644
--- a/src/backend/replication/logical/reorderbuffer.c
+++ b/src/backend/replication/logical/reorderbuffer.c
@@ -77,6 +77,35 @@
* a bit more memory to the oldest subtransactions, because it's likely
* they are the source for the next sequence of changes.
*
+ * When decoding sequences, we differentiate between a sequences created
+ * in a (running) transaction, and sequences created in other (already
+ * committed) transactions. Changes for sequences created in the same
+ * top-level transaction are treated as "transactional" i.e. just like
+ * any other change from that transaction (and discarded in case of a
+ * rollback). Changes for sequences created earlier are treated as not
+ * transactional - are processed immediately, as if performed outside
+ * any transaction (and thus not rolled back).
+ *
+ * This mixed behavior is necessary - sequences are non-transactional
+ * (e.g. ROLLBACK does not undo the sequence increments). But for new
+ * sequences, we need to handle them in a transactional way, because if
+ * we ever get some DDL support, the sequence won't exist until the
+ * transaction gets applied. So we need to ensure the increments don't
+ * happen until the sequence gets created.
+ *
+ * To differentiate which sequences are "old" and which were created
+ * in a still-running transaction, we track sequences created in running
+ * transactions in a hash table. Each sequence is identified by it's
+ * relfilenode, and at transaction commit we discard all sequences
+ * created in that particular transaction. For each sequence we track
+ * the XID of the (sub)transaction that created it, and we cleanup the
+ * sequences for each subtransaction when it completes (commit/rollback).
+ *
+ * We don't use the XID to check if it's the same top-level transaction.
+ * It's enough to know it was created in an in-progress transaction,
+ * and we know it must be the current one because otherwise it wouldn't
+ * see the sequence object.
+ *
* -------------------------------------------------------------------------
*/
#include "postgres.h"
@@ -91,6 +120,7 @@
#include "access/xact.h"
#include "access/xlog_internal.h"
#include "catalog/catalog.h"
+#include "commands/sequence.h"
#include "lib/binaryheap.h"
#include "miscadmin.h"
#include "pgstat.h"
@@ -116,6 +146,13 @@ typedef struct ReorderBufferTXNByIdEnt
ReorderBufferTXN *txn;
} ReorderBufferTXNByIdEnt;
+/* entry for hash table we use to track sequences created in running xacts */
+typedef struct ReorderBufferSequenceEnt
+{
+ RelFileNode rnode;
+ TransactionId xid;
+} ReorderBufferSequenceEnt;
+
/* data structures for (relfilenode, ctid) => (cmin, cmax) mapping */
typedef struct ReorderBufferTupleCidKey
{
@@ -339,6 +376,14 @@ ReorderBufferAllocate(void)
buffer->by_txn = hash_create("ReorderBufferByXid", 1000, &hash_ctl,
HASH_ELEM | HASH_BLOBS | HASH_CONTEXT);
+ /* hash table of sequences, mapping relfilenode to XID of transaction */
+ hash_ctl.keysize = sizeof(RelFileNode);
+ hash_ctl.entrysize = sizeof(ReorderBufferSequenceEnt);
+ hash_ctl.hcxt = buffer->context;
+
+ buffer->sequences = hash_create("ReorderBufferSequenceHash", 1000, &hash_ctl,
+ HASH_ELEM | HASH_BLOBS | HASH_CONTEXT);
+
buffer->by_txn_last_xid = InvalidTransactionId;
buffer->by_txn_last_txn = NULL;
@@ -525,6 +570,13 @@ ReorderBufferReturnChange(ReorderBuffer *rb, ReorderBufferChange *change,
change->data.truncate.relids = NULL;
}
break;
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
+ if (change->data.sequence.tuple)
+ {
+ ReorderBufferReturnTupleBuf(rb, change->data.sequence.tuple);
+ change->data.sequence.tuple = NULL;
+ }
+ break;
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM:
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_ABORT:
case REORDER_BUFFER_CHANGE_INTERNAL_COMMAND_ID:
@@ -859,6 +911,231 @@ ReorderBufferQueueMessage(ReorderBuffer *rb, TransactionId xid,
}
}
+/*
+ * Treat the sequence increment as transactional?
+ *
+ * The hash table tracks all sequences created in in-progress transactions,
+ * so we simply do a lookup (the sequence is identified by relfilende). If
+ * we find a match, the increment should be handled as transactional.
+ */
+bool
+ReorderBufferSequenceIsTransactional(ReorderBuffer *rb,
+ RelFileNode rnode, bool created)
+{
+ bool found = false;
+
+ if (created)
+ return true;
+
+ hash_search(rb->sequences,
+ (void *) &rnode,
+ HASH_FIND,
+ &found);
+
+ return found;
+}
+
+/*
+ * Cleanup sequences created in in-progress transactions.
+ *
+ * There's no way to search by XID, so we simply do a seqscan of all
+ * the entries in the hash table. Hopefully there are only a couple
+ * entries in most cases - people generally don't create many new
+ * sequences over and over.
+ */
+static void
+ReorderBufferSequenceCleanup(ReorderBuffer *rb, TransactionId xid)
+{
+ HASH_SEQ_STATUS scan_status;
+ ReorderBufferSequenceEnt *ent;
+
+ hash_seq_init(&scan_status, rb->sequences);
+ while ((ent = (ReorderBufferSequenceEnt *) hash_seq_search(&scan_status)) != NULL)
+ {
+ /* skip sequences not from this transaction */
+ if (ent->xid != xid)
+ continue;
+
+ (void) hash_search(rb->sequences,
+ (void *) &(ent->rnode),
+ HASH_REMOVE, NULL);
+ }
+}
+
+/*
+ * A transactional sequence increment is queued to be processed upon commit
+ * and a non-transactional increment gets processed immediately.
+ *
+ * A sequence update may be both transactional and non-transactional. When
+ * created in a running transaction, treat it as transactional and queue
+ * the change in it. Otherwise treat it as non-transactional, so that we
+ * don't forget the increment in case of a rollback.
+ */
+void
+ReorderBufferQueueSequence(ReorderBuffer *rb, TransactionId xid,
+ Snapshot snapshot, XLogRecPtr lsn, RepOriginId origin_id,
+ RelFileNode rnode, bool transactional, bool created,
+ ReorderBufferTupleBuf *tuplebuf)
+{
+ /*
+ * Change needs to be handled as transactional, because the sequence was
+ * created in a transaction that is still running. In that case all the
+ * changes need to be queued in that transaction, we must not send them
+ * to the downstream until the transaction commits.
+ *
+ * There's a bit of a trouble with subtransactions - we can't queue it
+ * into the subxact, because it might be rolled back and we'd lose the
+ * increment. We need to queue it into the same (sub)xact that created
+ * the sequence, which is why we track the XID in the hash table.
+ */
+ if (transactional)
+ {
+ MemoryContext oldcontext;
+ ReorderBufferChange *change;
+
+ /* lookup sequence by relfilenode */
+ ReorderBufferSequenceEnt *ent;
+ bool found;
+
+ /* transactional changes require a transaction */
+ Assert(xid != InvalidTransactionId);
+
+ /* search the lookup table (we ignore the return value, found is enough) */
+ ent = hash_search(rb->sequences,
+ (void *) &rnode,
+ created ? HASH_ENTER : HASH_FIND,
+ &found);
+
+ /*
+ * If this is the "create" increment, we must not have found any
+ * pre-existing entry in the hash table (i.e. there must not be
+ * any conflicting sequence).
+ */
+ Assert(!(created && found));
+
+ /* But we must have either created or found an existing entry. */
+ Assert(created || found);
+
+ /*
+ * When creating the sequence, remember the XID of the transaction
+ * that created id.
+ */
+ if (created)
+ ent->xid = xid;
+
+ /* XXX Maybe check that we're still in the same top-level xact? */
+
+ /* OK, allocate and queue the change */
+ oldcontext = MemoryContextSwitchTo(rb->context);
+
+ change = ReorderBufferGetChange(rb);
+
+ change->action = REORDER_BUFFER_CHANGE_SEQUENCE;
+ change->origin_id = origin_id;
+
+ memcpy(&change->data.sequence.relnode, &rnode, sizeof(RelFileNode));
+
+ change->data.sequence.created = created;
+ change->data.sequence.tuple = tuplebuf;
+
+ /* add it to the same subxact that created the sequence */
+ ReorderBufferQueueChange(rb, ent->xid, lsn, change, false);
+
+ MemoryContextSwitchTo(oldcontext);
+ }
+ else
+ {
+ /*
+ * This increment is for a sequence that was not created in any
+ * running transaction, so we treat it as non-transactional and
+ * just send it to the output plugin directly.
+ */
+ ReorderBufferTXN *txn = NULL;
+ volatile Snapshot snapshot_now = snapshot;
+ bool using_subtxn;
+
+#ifdef USE_ASSERT_CHECKING
+ /* All "creates" have to be handled as transactional. */
+ Assert(!created);
+
+ /* Make sure the sequence is not in the hash table. */
+ {
+ bool found;
+ hash_search(rb->sequences,
+ (void *) &rnode,
+ HASH_FIND, &found);
+ Assert(!found);
+ }
+#endif
+
+ if (xid != InvalidTransactionId)
+ txn = ReorderBufferTXNByXid(rb, xid, true, NULL, lsn, true);
+
+ /* setup snapshot to allow catalog access */
+ SetupHistoricSnapshot(snapshot_now, NULL);
+
+ /*
+ * Decoding needs access to syscaches et al., which in turn use
+ * heavyweight locks and such. Thus we need to have enough state around to
+ * keep track of those. The easiest way is to simply use a transaction
+ * internally. That also allows us to easily enforce that nothing writes
+ * to the database by checking for xid assignments.
+ *
+ * When we're called via the SQL SRF there's already a transaction
+ * started, so start an explicit subtransaction there.
+ */
+ using_subtxn = IsTransactionOrTransactionBlock();
+
+ PG_TRY();
+ {
+ Relation relation;
+ HeapTuple tuple;
+ Form_pg_sequence_data seq;
+ Oid reloid;
+
+ if (using_subtxn)
+ BeginInternalSubTransaction("sequence");
+ else
+ StartTransactionCommand();
+
+ reloid = RelidByRelfilenode(rnode.spcNode, rnode.relNode);
+
+ if (reloid == InvalidOid)
+ elog(ERROR, "could not map filenode \"%s\" to relation OID",
+ relpathperm(rnode,
+ MAIN_FORKNUM));
+
+ relation = RelationIdGetRelation(reloid);
+ tuple = &tuplebuf->tuple;
+ seq = (Form_pg_sequence_data) GETSTRUCT(tuple);
+
+ rb->sequence(rb, txn, lsn, relation, transactional, created,
+ seq->last_value, seq->log_cnt, seq->is_called);
+
+ RelationClose(relation);
+
+ TeardownHistoricSnapshot(false);
+
+ AbortCurrentTransaction();
+
+ if (using_subtxn)
+ RollbackAndReleaseCurrentSubTransaction();
+ }
+ PG_CATCH();
+ {
+ TeardownHistoricSnapshot(true);
+
+ AbortCurrentTransaction();
+
+ if (using_subtxn)
+ RollbackAndReleaseCurrentSubTransaction();
+
+ PG_RE_THROW();
+ }
+ PG_END_TRY();
+ }
+}
+
/*
* AssertTXNLsnOrder
* Verify LSN ordering of transaction lists in the reorderbuffer
@@ -1535,6 +1812,9 @@ ReorderBufferCleanupTXN(ReorderBuffer *rb, ReorderBufferTXN *txn)
&found);
Assert(found);
+ /* Remove sequences created in this transaction (if any). */
+ ReorderBufferSequenceCleanup(rb, txn->xid);
+
/* remove entries spilled to disk */
if (rbtxn_is_serialized(txn))
ReorderBufferRestoreCleanup(rb, txn);
@@ -1950,6 +2230,31 @@ ReorderBufferApplyMessage(ReorderBuffer *rb, ReorderBufferTXN *txn,
change->data.msg.message);
}
+/*
+ * Helper function for ReorderBufferProcessTXN for applying sequences.
+ */
+static inline void
+ReorderBufferApplySequence(ReorderBuffer *rb, ReorderBufferTXN *txn,
+ Relation relation, ReorderBufferChange *change,
+ bool streaming)
+{
+ HeapTuple tuple;
+ Form_pg_sequence_data seq;
+
+ tuple = &change->data.sequence.tuple->tuple;
+ seq = (Form_pg_sequence_data) GETSTRUCT(tuple);
+
+ /* Only ever called from ReorderBufferApplySequence, so transational. */
+ if (streaming)
+ rb->stream_sequence(rb, txn, change->lsn, relation, true,
+ change->data.sequence.created,
+ seq->last_value, seq->log_cnt, seq->is_called);
+ else
+ rb->sequence(rb, txn, change->lsn, relation, true,
+ change->data.sequence.created,
+ seq->last_value, seq->log_cnt, seq->is_called);
+}
+
/*
* Function to store the command id and snapshot at the end of the current
* stream so that we can reuse the same while sending the next stream.
@@ -2392,6 +2697,31 @@ ReorderBufferProcessTXN(ReorderBuffer *rb, ReorderBufferTXN *txn,
case REORDER_BUFFER_CHANGE_INTERNAL_TUPLECID:
elog(ERROR, "tuplecid value in changequeue");
break;
+
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
+ Assert(snapshot_now);
+
+ reloid = RelidByRelfilenode(change->data.sequence.relnode.spcNode,
+ change->data.sequence.relnode.relNode);
+
+ if (reloid == InvalidOid)
+ elog(ERROR, "could not map filenode \"%s\" to relation OID",
+ relpathperm(change->data.sequence.relnode,
+ MAIN_FORKNUM));
+
+ relation = RelationIdGetRelation(reloid);
+
+ if (!RelationIsValid(relation))
+ elog(ERROR, "could not open relation with OID %u (for filenode \"%s\")",
+ reloid,
+ relpathperm(change->data.sequence.relnode,
+ MAIN_FORKNUM));
+
+ if (RelationIsLogicallyLogged(relation))
+ ReorderBufferApplySequence(rb, txn, relation, change, streaming);
+
+ RelationClose(relation);
+ break;
}
}
@@ -3776,6 +4106,39 @@ ReorderBufferSerializeChange(ReorderBuffer *rb, ReorderBufferTXN *txn,
memcpy(data, change->data.truncate.relids, size);
data += size;
+ break;
+ }
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
+ {
+ char *data;
+ ReorderBufferTupleBuf *tup;
+ Size len = 0;
+
+ tup = change->data.sequence.tuple;
+
+ if (tup)
+ {
+ sz += sizeof(HeapTupleData);
+ len = tup->tuple.t_len;
+ sz += len;
+ }
+
+ /* make sure we have enough space */
+ ReorderBufferSerializeReserve(rb, sz);
+
+ data = ((char *) rb->outbuf) + sizeof(ReorderBufferDiskChange);
+ /* might have been reallocated above */
+ ondisk = (ReorderBufferDiskChange *) rb->outbuf;
+
+ if (len)
+ {
+ memcpy(data, &tup->tuple, sizeof(HeapTupleData));
+ data += sizeof(HeapTupleData);
+
+ memcpy(data, tup->tuple.t_data, len);
+ data += len;
+ }
+
break;
}
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM:
@@ -4040,6 +4403,22 @@ ReorderBufferChangeSize(ReorderBufferChange *change)
{
sz += sizeof(Oid) * change->data.truncate.nrelids;
+ break;
+ }
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
+ {
+ ReorderBufferTupleBuf *tup;
+ Size len = 0;
+
+ tup = change->data.sequence.tuple;
+
+ if (tup)
+ {
+ sz += sizeof(HeapTupleData);
+ len = tup->tuple.t_len;
+ sz += len;
+ }
+
break;
}
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM:
@@ -4341,6 +4720,30 @@ ReorderBufferRestoreChange(ReorderBuffer *rb, ReorderBufferTXN *txn,
break;
}
+
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
+ if (change->data.sequence.tuple)
+ {
+ uint32 tuplelen = ((HeapTuple) data)->t_len;
+
+ change->data.sequence.tuple =
+ ReorderBufferGetTupleBuf(rb, tuplelen - SizeofHeapTupleHeader);
+
+ /* restore ->tuple */
+ memcpy(&change->data.sequence.tuple->tuple, data,
+ sizeof(HeapTupleData));
+ data += sizeof(HeapTupleData);
+
+ /* reset t_data pointer into the new tuplebuf */
+ change->data.sequence.tuple->tuple.t_data =
+ ReorderBufferTupleBufData(change->data.sequence.tuple);
+
+ /* restore tuple data itself */
+ memcpy(change->data.sequence.tuple->tuple.t_data, data, tuplelen);
+ data += tuplelen;
+ }
+ break;
+
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM:
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_ABORT:
case REORDER_BUFFER_CHANGE_INTERNAL_COMMAND_ID:
diff --git a/src/include/commands/sequence.h b/src/include/commands/sequence.h
index 40544dd4c7..5919fb90ee 100644
--- a/src/include/commands/sequence.h
+++ b/src/include/commands/sequence.h
@@ -48,6 +48,7 @@ typedef FormData_pg_sequence_data *Form_pg_sequence_data;
typedef struct xl_seq_rec
{
RelFileNode node;
+ bool created; /* is this a CREATE SEQUENCE */
/* SEQUENCE TUPLE DATA FOLLOWS AT THE END */
} xl_seq_rec;
diff --git a/src/include/replication/output_plugin.h b/src/include/replication/output_plugin.h
index 810495ed0e..57bc13f11c 100644
--- a/src/include/replication/output_plugin.h
+++ b/src/include/replication/output_plugin.h
@@ -88,6 +88,19 @@ typedef void (*LogicalDecodeMessageCB) (struct LogicalDecodingContext *ctx,
Size message_size,
const char *message);
+/*
+ * Called for the generic logical decoding sequences.
+ */
+typedef void (*LogicalDecodeSequenceCB) (struct LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional,
+ bool created,
+ int64 last_value,
+ int64 log_cnt,
+ bool is_called);
+
/*
* Filter changes by origin.
*/
@@ -199,6 +212,20 @@ typedef void (*LogicalDecodeStreamMessageCB) (struct LogicalDecodingContext *ctx
Size message_size,
const char *message);
+/*
+ * Called for the streaming generic logical decoding sequences from in-progress
+ * transactions.
+ */
+typedef void (*LogicalDecodeStreamSequenceCB) (struct LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional,
+ bool created,
+ int64 last_value,
+ int64 log_cnt,
+ bool is_called);
+
/*
* Callback for streaming truncates from in-progress transactions.
*/
@@ -219,6 +246,7 @@ typedef struct OutputPluginCallbacks
LogicalDecodeTruncateCB truncate_cb;
LogicalDecodeCommitCB commit_cb;
LogicalDecodeMessageCB message_cb;
+ LogicalDecodeSequenceCB sequence_cb;
LogicalDecodeFilterByOriginCB filter_by_origin_cb;
LogicalDecodeShutdownCB shutdown_cb;
@@ -237,6 +265,7 @@ typedef struct OutputPluginCallbacks
LogicalDecodeStreamCommitCB stream_commit_cb;
LogicalDecodeStreamChangeCB stream_change_cb;
LogicalDecodeStreamMessageCB stream_message_cb;
+ LogicalDecodeStreamSequenceCB stream_sequence_cb;
LogicalDecodeStreamTruncateCB stream_truncate_cb;
} OutputPluginCallbacks;
diff --git a/src/include/replication/reorderbuffer.h b/src/include/replication/reorderbuffer.h
index 5b40ff75f7..63e6ed037b 100644
--- a/src/include/replication/reorderbuffer.h
+++ b/src/include/replication/reorderbuffer.h
@@ -64,7 +64,8 @@ enum ReorderBufferChangeType
REORDER_BUFFER_CHANGE_INTERNAL_SPEC_INSERT,
REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM,
REORDER_BUFFER_CHANGE_INTERNAL_SPEC_ABORT,
- REORDER_BUFFER_CHANGE_TRUNCATE
+ REORDER_BUFFER_CHANGE_TRUNCATE,
+ REORDER_BUFFER_CHANGE_SEQUENCE
};
/* forward declaration */
@@ -158,6 +159,14 @@ typedef struct ReorderBufferChange
uint32 ninvalidations; /* Number of messages */
SharedInvalidationMessage *invalidations; /* invalidation message */
} inval;
+
+ /* Context data for Sequence changes */
+ struct
+ {
+ RelFileNode relnode;
+ bool created;
+ ReorderBufferTupleBuf *tuple;
+ } sequence;
} data;
/*
@@ -430,6 +439,15 @@ typedef void (*ReorderBufferMessageCB) (ReorderBuffer *rb,
const char *prefix, Size sz,
const char *message);
+/* sequence callback signature */
+typedef void (*ReorderBufferSequenceCB) (ReorderBuffer *rb,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional, bool created,
+ int64 last_value, int64 log_cnt,
+ bool is_called);
+
/* begin prepare callback signature */
typedef void (*ReorderBufferBeginPrepareCB) (ReorderBuffer *rb,
ReorderBufferTXN *txn);
@@ -496,6 +514,15 @@ typedef void (*ReorderBufferStreamMessageCB) (
const char *prefix, Size sz,
const char *message);
+/* stream sequence callback signature */
+typedef void (*ReorderBufferStreamSequenceCB) (ReorderBuffer *rb,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional, bool created,
+ int64 last_value, int64 log_cnt,
+ bool is_called);
+
/* stream truncate callback signature */
typedef void (*ReorderBufferStreamTruncateCB) (
ReorderBuffer *rb,
@@ -511,6 +538,12 @@ struct ReorderBuffer
*/
HTAB *by_txn;
+ /*
+ * relfilenode => XID lookup table for sequences created in a transaction
+ * (also includes altered sequences, which assigns new relfilenode)
+ */
+ HTAB *sequences;
+
/*
* Transactions that could be a toplevel xact, ordered by LSN of the first
* record bearing that xid.
@@ -541,6 +574,7 @@ struct ReorderBuffer
ReorderBufferApplyTruncateCB apply_truncate;
ReorderBufferCommitCB commit;
ReorderBufferMessageCB message;
+ ReorderBufferSequenceCB sequence;
/*
* Callbacks to be called when streaming a transaction at prepare time.
@@ -560,6 +594,7 @@ struct ReorderBuffer
ReorderBufferStreamCommitCB stream_commit;
ReorderBufferStreamChangeCB stream_change;
ReorderBufferStreamMessageCB stream_message;
+ ReorderBufferStreamSequenceCB stream_sequence;
ReorderBufferStreamTruncateCB stream_truncate;
/*
@@ -635,6 +670,10 @@ void ReorderBufferQueueChange(ReorderBuffer *, TransactionId,
void ReorderBufferQueueMessage(ReorderBuffer *, TransactionId, Snapshot snapshot, XLogRecPtr lsn,
bool transactional, const char *prefix,
Size message_size, const char *message);
+void ReorderBufferQueueSequence(ReorderBuffer *rb, TransactionId xid,
+ Snapshot snapshot, XLogRecPtr lsn, RepOriginId origin_id,
+ RelFileNode rnode, bool transactional, bool created,
+ ReorderBufferTupleBuf *tuplebuf);
void ReorderBufferCommit(ReorderBuffer *, TransactionId,
XLogRecPtr commit_lsn, XLogRecPtr end_lsn,
TimestampTz commit_time, RepOriginId origin_id, XLogRecPtr origin_lsn);
@@ -682,4 +721,7 @@ void ReorderBufferSetRestartPoint(ReorderBuffer *, XLogRecPtr ptr);
void StartupReorderBuffer(void);
+bool ReorderBufferSequenceIsTransactional(ReorderBuffer *rb,
+ RelFileNode rnode, bool created);
+
#endif
--
2.31.1
[text/x-patch] 0002-Add-support-for-decoding-sequences-to-test_-20211208.patch (20.7K, ../../[email protected]/3-0002-Add-support-for-decoding-sequences-to-test_-20211208.patch)
download | inline diff:
From 936470a0cf0eaa56c90b32b23c25dc317a1d3a55 Mon Sep 17 00:00:00 2001
From: Tomas Vondra <[email protected]>
Date: Fri, 24 Sep 2021 00:42:04 +0200
Subject: [PATCH 2/2] Add support for decoding sequences to test_decoding
---
contrib/test_decoding/Makefile | 3 +-
contrib/test_decoding/expected/sequence.out | 327 ++++++++++++++++++++
contrib/test_decoding/sql/sequence.sql | 119 +++++++
contrib/test_decoding/test_decoding.c | 69 +++++
4 files changed, 517 insertions(+), 1 deletion(-)
create mode 100644 contrib/test_decoding/expected/sequence.out
create mode 100644 contrib/test_decoding/sql/sequence.sql
diff --git a/contrib/test_decoding/Makefile b/contrib/test_decoding/Makefile
index 9a31e0b879..56ddc3abae 100644
--- a/contrib/test_decoding/Makefile
+++ b/contrib/test_decoding/Makefile
@@ -5,7 +5,8 @@ PGFILEDESC = "test_decoding - example of a logical decoding output plugin"
REGRESS = ddl xact rewrite toast permissions decoding_in_xact \
decoding_into_rel binary prepared replorigin time messages \
- spill slot truncate stream stats twophase twophase_stream
+ spill slot truncate stream stats twophase twophase_stream \
+ sequence
ISOLATION = mxact delayed_startup ondisk_startup concurrent_ddl_dml \
oldest_xmin snapshot_transfer subxact_without_top concurrent_stream \
twophase_snapshot
diff --git a/contrib/test_decoding/expected/sequence.out b/contrib/test_decoding/expected/sequence.out
new file mode 100644
index 0000000000..d94e185fb9
--- /dev/null
+++ b/contrib/test_decoding/expected/sequence.out
@@ -0,0 +1,327 @@
+-- predictability
+SET synchronous_commit = on;
+SELECT 'init' FROM pg_create_logical_replication_slot('regression_slot', 'test_decoding');
+ ?column?
+----------
+ init
+(1 row)
+
+CREATE SEQUENCE test_sequence;
+-- test the sequence changes by several nextval() calls
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 1
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 2
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 4
+(1 row)
+
+-- test the sequence changes by several ALTER commands
+ALTER SEQUENCE test_sequence INCREMENT BY 10;
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 14
+(1 row)
+
+ALTER SEQUENCE test_sequence START WITH 3000;
+ALTER SEQUENCE test_sequence MAXVALUE 10000;
+ALTER SEQUENCE test_sequence RESTART WITH 4000;
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 4000
+(1 row)
+
+-- test the sequence changes by several setval() calls
+SELECT setval('test_sequence', 3500);
+ setval
+--------
+ 3500
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3510
+(1 row)
+
+SELECT setval('test_sequence', 3500, true);
+ setval
+--------
+ 3500
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3510
+(1 row)
+
+SELECT setval('test_sequence', 3500, false);
+ setval
+--------
+ 3500
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3500
+(1 row)
+
+-- show results and drop sequence
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+ data
+------------------------------------------------------------------------------------------------
+ BEGIN
+ sequence public.test_sequence: transactional:1 created:1 last_value:1 log_cnt:0 is_called:0
+ COMMIT
+ sequence public.test_sequence: transactional:0 created:0 last_value:33 log_cnt:0 is_called:1
+ BEGIN
+ sequence public.test_sequence: transactional:1 created:1 last_value:4 log_cnt:0 is_called:1
+ COMMIT
+ sequence public.test_sequence: transactional:0 created:0 last_value:334 log_cnt:0 is_called:1
+ BEGIN
+ sequence public.test_sequence: transactional:1 created:1 last_value:14 log_cnt:32 is_called:1
+ COMMIT
+ BEGIN
+ sequence public.test_sequence: transactional:1 created:1 last_value:14 log_cnt:0 is_called:1
+ COMMIT
+ BEGIN
+ sequence public.test_sequence: transactional:1 created:1 last_value:4000 log_cnt:0 is_called:0
+ COMMIT
+ sequence public.test_sequence: transactional:0 created:0 last_value:4320 log_cnt:0 is_called:1
+ sequence public.test_sequence: transactional:0 created:0 last_value:3500 log_cnt:0 is_called:1
+ sequence public.test_sequence: transactional:0 created:0 last_value:3830 log_cnt:0 is_called:1
+ sequence public.test_sequence: transactional:0 created:0 last_value:3500 log_cnt:0 is_called:1
+ sequence public.test_sequence: transactional:0 created:0 last_value:3830 log_cnt:0 is_called:1
+ sequence public.test_sequence: transactional:0 created:0 last_value:3500 log_cnt:0 is_called:0
+ sequence public.test_sequence: transactional:0 created:0 last_value:3820 log_cnt:0 is_called:1
+(24 rows)
+
+DROP SEQUENCE test_sequence;
+-- rollback on sequence creation and update
+BEGIN;
+CREATE SEQUENCE test_sequence;
+CREATE TABLE test_table (a INT);
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 1
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 2
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3
+(1 row)
+
+SELECT setval('test_sequence', 3000);
+ setval
+--------
+ 3000
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3001
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3002
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3003
+(1 row)
+
+ALTER SEQUENCE test_sequence RESTART WITH 6000;
+INSERT INTO test_table VALUES( (SELECT nextval('test_sequence')) );
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 6001
+(1 row)
+
+ROLLBACK;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'skip-sequences', '0');
+ data
+------
+(0 rows)
+
+-- rollback on table creation with serial column
+BEGIN;
+CREATE TABLE test_table (a SERIAL, b INT);
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+INSERT INTO test_table (b) VALUES (300);
+SELECT setval('test_table_a_seq', 3000);
+ setval
+--------
+ 3000
+(1 row)
+
+INSERT INTO test_table (b) VALUES (400);
+INSERT INTO test_table (b) VALUES (500);
+INSERT INTO test_table (b) VALUES (600);
+ALTER SEQUENCE test_table_a_seq RESTART WITH 6000;
+INSERT INTO test_table (b) VALUES (700);
+INSERT INTO test_table (b) VALUES (800);
+INSERT INTO test_table (b) VALUES (900);
+ROLLBACK;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'skip-sequences', '0');
+ data
+------
+(0 rows)
+
+-- rollback on table with serial column
+CREATE TABLE test_table (a SERIAL, b INT);
+BEGIN;
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+INSERT INTO test_table (b) VALUES (300);
+SELECT setval('test_table_a_seq', 3000);
+ setval
+--------
+ 3000
+(1 row)
+
+INSERT INTO test_table (b) VALUES (400);
+INSERT INTO test_table (b) VALUES (500);
+INSERT INTO test_table (b) VALUES (600);
+ALTER SEQUENCE test_table_a_seq RESTART WITH 6000;
+INSERT INTO test_table (b) VALUES (700);
+INSERT INTO test_table (b) VALUES (800);
+INSERT INTO test_table (b) VALUES (900);
+ROLLBACK;
+-- check table and sequence values after rollback
+SELECT * from test_table_a_seq;
+ last_value | log_cnt | is_called
+------------+---------+-----------
+ 3003 | 30 | t
+(1 row)
+
+SELECT nextval('test_table_a_seq');
+ nextval
+---------
+ 3004
+(1 row)
+
+DROP TABLE test_table;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+ data
+---------------------------------------------------------------------------------------------------
+ BEGIN
+ sequence public.test_table_a_seq: transactional:1 created:1 last_value:1 log_cnt:0 is_called:0
+ COMMIT
+ sequence public.test_table_a_seq: transactional:0 created:0 last_value:33 log_cnt:0 is_called:1
+ sequence public.test_table_a_seq: transactional:0 created:0 last_value:3000 log_cnt:0 is_called:1
+ sequence public.test_table_a_seq: transactional:0 created:0 last_value:3033 log_cnt:0 is_called:1
+ BEGIN
+ COMMIT
+(8 rows)
+
+-- savepoint test on table with serial column
+BEGIN;
+CREATE TABLE test_table (a SERIAL, b INT);
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+SAVEPOINT a;
+INSERT INTO test_table (b) VALUES (300);
+ROLLBACK TO SAVEPOINT a;
+DROP TABLE test_table;
+COMMIT;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+ data
+-------------------------------------------------------------------------------------------------
+ BEGIN
+ sequence public.test_table_a_seq: transactional:1 created:1 last_value:1 log_cnt:0 is_called:0
+ sequence public.test_table_a_seq: transactional:1 created:0 last_value:33 log_cnt:0 is_called:1
+ table public.test_table: INSERT: a[integer]:1 b[integer]:100
+ table public.test_table: INSERT: a[integer]:2 b[integer]:200
+ COMMIT
+(6 rows)
+
+-- savepoint test on table with serial column
+BEGIN;
+CREATE SEQUENCE test_sequence;
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 1
+(1 row)
+
+SELECT setval('test_sequence', 3000);
+ setval
+--------
+ 3000
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3001
+(1 row)
+
+SAVEPOINT a;
+ALTER SEQUENCE test_sequence START WITH 7000;
+SELECT setval('test_sequence', 5000);
+ setval
+--------
+ 5000
+(1 row)
+
+ROLLBACK TO SAVEPOINT a;
+SELECT * FROM test_sequence;
+ last_value | log_cnt | is_called
+------------+---------+-----------
+ 3001 | 32 | t
+(1 row)
+
+DROP SEQUENCE test_sequence;
+COMMIT;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+ data
+------------------------------------------------------------------------------------------------
+ BEGIN
+ sequence public.test_sequence: transactional:1 created:1 last_value:1 log_cnt:0 is_called:0
+ sequence public.test_sequence: transactional:1 created:0 last_value:33 log_cnt:0 is_called:1
+ sequence public.test_sequence: transactional:1 created:0 last_value:3000 log_cnt:0 is_called:1
+ sequence public.test_sequence: transactional:1 created:0 last_value:3033 log_cnt:0 is_called:1
+ COMMIT
+(6 rows)
+
+SELECT pg_drop_replication_slot('regression_slot');
+ pg_drop_replication_slot
+--------------------------
+
+(1 row)
+
diff --git a/contrib/test_decoding/sql/sequence.sql b/contrib/test_decoding/sql/sequence.sql
new file mode 100644
index 0000000000..21c4b79222
--- /dev/null
+++ b/contrib/test_decoding/sql/sequence.sql
@@ -0,0 +1,119 @@
+-- predictability
+SET synchronous_commit = on;
+SELECT 'init' FROM pg_create_logical_replication_slot('regression_slot', 'test_decoding');
+
+CREATE SEQUENCE test_sequence;
+
+-- test the sequence changes by several nextval() calls
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+
+-- test the sequence changes by several ALTER commands
+ALTER SEQUENCE test_sequence INCREMENT BY 10;
+SELECT nextval('test_sequence');
+
+ALTER SEQUENCE test_sequence START WITH 3000;
+ALTER SEQUENCE test_sequence MAXVALUE 10000;
+ALTER SEQUENCE test_sequence RESTART WITH 4000;
+SELECT nextval('test_sequence');
+
+-- test the sequence changes by several setval() calls
+SELECT setval('test_sequence', 3500);
+SELECT nextval('test_sequence');
+SELECT setval('test_sequence', 3500, true);
+SELECT nextval('test_sequence');
+SELECT setval('test_sequence', 3500, false);
+SELECT nextval('test_sequence');
+
+-- show results and drop sequence
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+DROP SEQUENCE test_sequence;
+
+-- rollback on sequence creation and update
+BEGIN;
+CREATE SEQUENCE test_sequence;
+CREATE TABLE test_table (a INT);
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT setval('test_sequence', 3000);
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+ALTER SEQUENCE test_sequence RESTART WITH 6000;
+INSERT INTO test_table VALUES( (SELECT nextval('test_sequence')) );
+SELECT nextval('test_sequence');
+ROLLBACK;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'skip-sequences', '0');
+
+-- rollback on table creation with serial column
+BEGIN;
+CREATE TABLE test_table (a SERIAL, b INT);
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+INSERT INTO test_table (b) VALUES (300);
+SELECT setval('test_table_a_seq', 3000);
+INSERT INTO test_table (b) VALUES (400);
+INSERT INTO test_table (b) VALUES (500);
+INSERT INTO test_table (b) VALUES (600);
+ALTER SEQUENCE test_table_a_seq RESTART WITH 6000;
+INSERT INTO test_table (b) VALUES (700);
+INSERT INTO test_table (b) VALUES (800);
+INSERT INTO test_table (b) VALUES (900);
+ROLLBACK;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'skip-sequences', '0');
+
+-- rollback on table with serial column
+CREATE TABLE test_table (a SERIAL, b INT);
+
+BEGIN;
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+INSERT INTO test_table (b) VALUES (300);
+SELECT setval('test_table_a_seq', 3000);
+INSERT INTO test_table (b) VALUES (400);
+INSERT INTO test_table (b) VALUES (500);
+INSERT INTO test_table (b) VALUES (600);
+ALTER SEQUENCE test_table_a_seq RESTART WITH 6000;
+INSERT INTO test_table (b) VALUES (700);
+INSERT INTO test_table (b) VALUES (800);
+INSERT INTO test_table (b) VALUES (900);
+ROLLBACK;
+
+-- check table and sequence values after rollback
+SELECT * from test_table_a_seq;
+SELECT nextval('test_table_a_seq');
+
+DROP TABLE test_table;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+
+-- savepoint test on table with serial column
+BEGIN;
+CREATE TABLE test_table (a SERIAL, b INT);
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+SAVEPOINT a;
+INSERT INTO test_table (b) VALUES (300);
+ROLLBACK TO SAVEPOINT a;
+DROP TABLE test_table;
+COMMIT;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+
+-- savepoint test on table with serial column
+BEGIN;
+CREATE SEQUENCE test_sequence;
+SELECT nextval('test_sequence');
+SELECT setval('test_sequence', 3000);
+SELECT nextval('test_sequence');
+SAVEPOINT a;
+ALTER SEQUENCE test_sequence START WITH 7000;
+SELECT setval('test_sequence', 5000);
+ROLLBACK TO SAVEPOINT a;
+SELECT * FROM test_sequence;
+DROP SEQUENCE test_sequence;
+COMMIT;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+
+SELECT pg_drop_replication_slot('regression_slot');
diff --git a/contrib/test_decoding/test_decoding.c b/contrib/test_decoding/test_decoding.c
index e5cd84e85e..45765d299e 100644
--- a/contrib/test_decoding/test_decoding.c
+++ b/contrib/test_decoding/test_decoding.c
@@ -35,6 +35,7 @@ typedef struct
bool include_timestamp;
bool skip_empty_xacts;
bool only_local;
+ bool skip_sequences;
} TestDecodingData;
/*
@@ -76,6 +77,10 @@ static void pg_decode_message(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn, XLogRecPtr message_lsn,
bool transactional, const char *prefix,
Size sz, const char *message);
+static void pg_decode_sequence(LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn, XLogRecPtr sequence_lsn,
+ Relation rel, bool transactional, bool created,
+ int64 last_value, int64 log_cnt, bool is_called);
static bool pg_decode_filter_prepare(LogicalDecodingContext *ctx,
TransactionId xid,
const char *gid);
@@ -116,6 +121,10 @@ static void pg_decode_stream_message(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn, XLogRecPtr message_lsn,
bool transactional, const char *prefix,
Size sz, const char *message);
+static void pg_decode_stream_sequence(LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn, XLogRecPtr sequence_lsn,
+ Relation rel, bool transactional, bool created,
+ int64 last_value, int64 log_cnt, bool is_called);
static void pg_decode_stream_truncate(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn,
int nrelations, Relation relations[],
@@ -141,6 +150,7 @@ _PG_output_plugin_init(OutputPluginCallbacks *cb)
cb->filter_by_origin_cb = pg_decode_filter;
cb->shutdown_cb = pg_decode_shutdown;
cb->message_cb = pg_decode_message;
+ cb->sequence_cb = pg_decode_sequence;
cb->filter_prepare_cb = pg_decode_filter_prepare;
cb->begin_prepare_cb = pg_decode_begin_prepare_txn;
cb->prepare_cb = pg_decode_prepare_txn;
@@ -153,6 +163,7 @@ _PG_output_plugin_init(OutputPluginCallbacks *cb)
cb->stream_commit_cb = pg_decode_stream_commit;
cb->stream_change_cb = pg_decode_stream_change;
cb->stream_message_cb = pg_decode_stream_message;
+ cb->stream_sequence_cb = pg_decode_stream_sequence;
cb->stream_truncate_cb = pg_decode_stream_truncate;
}
@@ -175,6 +186,9 @@ pg_decode_startup(LogicalDecodingContext *ctx, OutputPluginOptions *opt,
data->skip_empty_xacts = false;
data->only_local = false;
+ /* skip sequences by default for backwards compatibility */
+ data->skip_sequences = true;
+
ctx->output_plugin_private = data;
opt->output_type = OUTPUT_PLUGIN_TEXTUAL_OUTPUT;
@@ -265,6 +279,17 @@ pg_decode_startup(LogicalDecodingContext *ctx, OutputPluginOptions *opt,
errmsg("could not parse value \"%s\" for parameter \"%s\"",
strVal(elem->arg), elem->defname)));
}
+ else if (strcmp(elem->defname, "skip-sequences") == 0)
+ {
+
+ if (elem->arg == NULL)
+ continue; /* true by default */
+ else if (!parse_bool(strVal(elem->arg), &data->skip_sequences))
+ ereport(ERROR,
+ (errcode(ERRCODE_INVALID_PARAMETER_VALUE),
+ errmsg("could not parse value \"%s\" for parameter \"%s\"",
+ strVal(elem->arg), elem->defname)));
+ }
else
{
ereport(ERROR,
@@ -744,6 +769,28 @@ pg_decode_message(LogicalDecodingContext *ctx,
OutputPluginWrite(ctx, true);
}
+static void
+pg_decode_sequence(LogicalDecodingContext *ctx, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ bool transactional, bool created,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ TestDecodingData *data = ctx->output_plugin_private;
+
+ /* return if requested to skip_sequences */
+ if (data->skip_sequences)
+ return;
+
+ OutputPluginPrepareWrite(ctx, true);
+ appendStringInfoString(ctx->out, "sequence ");
+ appendStringInfoString(ctx->out,
+ quote_qualified_identifier(get_namespace_name(get_rel_namespace(RelationGetRelid(rel))),
+ RelationGetRelationName(rel)));
+ appendStringInfo(ctx->out, ": transactional:%d created:%d last_value:%zu log_cnt:%zu is_called:%d",
+ transactional, created, last_value, log_cnt, is_called);
+ OutputPluginWrite(ctx, true);
+}
+
static void
pg_decode_stream_start(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn)
@@ -943,6 +990,28 @@ pg_decode_stream_message(LogicalDecodingContext *ctx,
OutputPluginWrite(ctx, true);
}
+static void
+pg_decode_stream_sequence(LogicalDecodingContext *ctx, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ bool transactional, bool created,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ TestDecodingData *data = ctx->output_plugin_private;
+
+ /* return if requested to skip_sequences */
+ if (data->skip_sequences)
+ return;
+
+ OutputPluginPrepareWrite(ctx, true);
+ appendStringInfoString(ctx->out, "streaming sequence ");
+ appendStringInfoString(ctx->out,
+ quote_qualified_identifier(get_namespace_name(get_rel_namespace(RelationGetRelid(rel))),
+ RelationGetRelationName(rel)));
+ appendStringInfo(ctx->out, ": transactional:%d created:%d last_value:%zu log_cnt:%zu is_called:%d",
+ transactional, created, last_value, log_cnt, is_called);
+ OutputPluginWrite(ctx, true);
+}
+
/*
* In streaming mode, we don't display the detailed information of Truncate.
* See pg_decode_stream_change.
--
2.31.1
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2021-12-08 09:04 Peter Eisentraut <[email protected]>
parent: Tomas Vondra <[email protected]>
0 siblings, 1 reply; 127+ messages in thread
From: Peter Eisentraut @ 2021-12-08 09:04 UTC (permalink / raw)
To: Tomas Vondra <[email protected]>; +Cc: PostgreSQL Hackers <[email protected]>; Petr Jelinek <[email protected]>
On 08.12.21 01:23, Tomas Vondra wrote:
>> The argument "create" for fill_seq_with_data() is always true (and
>> patch 0002 removes it). So I'm not sure if it's needed. If it is, it
>> should be documented somewhere.
>>
>
> Good point. I think it could be removed, but IIRC it'll be needed when
> adding sequence decoding to the built-in replication, and that patch is
> meant to be just an implementation of the API, without changing WAL.
>
> OTOH I don't see it in the last version of that patch (in ResetSequence2
> call) so maybe it's not really needed. I've kept it for now, but I'll
> investigate.
Ok, please check. If it is needed, perhaps then we need a way for
test_decoding() to simulate it, for testing. But perhaps it's not needed.
>> The order of the new fields sequence_cb and stream_sequence_cb is a
>> bit inconsistent compared to truncate_cb and stream_truncate_cb.
>> Maybe take another look to make the order more uniform.
>>
>
> You mean in OutputPluginCallbacks? I'd actually argue it's the truncate
> callbacks that are inconsistent - in the regular section truncate_cb is
> right before commit_cb, while in the streaming section it's at the end.
Ok, that makes sense. Then leave yours.
When the question about fill_seq_with_data() is resolved, I have no more
comments on this part.
^ permalink raw reply [nested|flat] 127+ messages in thread
* [PATCH v10 6/6] Move removal of old logical rewrite mapping files to custodian.
@ 2021-12-13 06:07 Nathan Bossart <[email protected]>
0 siblings, 0 replies; 127+ messages in thread
From: Nathan Bossart @ 2021-12-13 06:07 UTC (permalink / raw)
If there are many such files to remove, checkpoints can take much
longer. To avoid this, move this work to the newly-introduced
custodian process.
Since the mapping files include 32-bit transaction IDs, there is a
risk of wraparound if the files are not cleaned up fast enough.
Removing these files in checkpoints offered decent wraparound
protection simply due to the relatively high frequency of
checkpointing. With this change, servers should still clean up
mappings files with decently high frequency, but in theory the
wraparound risk might worsen for some (e.g., if the custodian is
spending a lot of time on a different task). Given this is an
existing problem, this change makes no effort to handle the
wraparound risk, and it is left as a future exercise.
---
src/backend/access/heap/rewriteheap.c | 78 +++++++++++++++++++++++----
src/backend/postmaster/custodian.c | 43 +++++++++++++++
src/include/access/rewriteheap.h | 1 +
src/include/postmaster/custodian.h | 4 ++
4 files changed, 116 insertions(+), 10 deletions(-)
diff --git a/src/backend/access/heap/rewriteheap.c b/src/backend/access/heap/rewriteheap.c
index 2f08fbe8d3..a01edf8a1f 100644
--- a/src/backend/access/heap/rewriteheap.c
+++ b/src/backend/access/heap/rewriteheap.c
@@ -117,6 +117,7 @@
#include "lib/ilist.h"
#include "miscadmin.h"
#include "pgstat.h"
+#include "postmaster/custodian.h"
#include "replication/logical.h"
#include "replication/slot.h"
#include "storage/bufmgr.h"
@@ -124,6 +125,7 @@
#include "storage/procarray.h"
#include "storage/smgr.h"
#include "utils/memutils.h"
+#include "utils/pg_lsn.h"
#include "utils/rel.h"
/*
@@ -1183,7 +1185,8 @@ heap_xlog_logical_rewrite(XLogReaderState *r)
* Perform a checkpoint for logical rewrite mappings
*
* This serves two tasks:
- * 1) Remove all mappings not needed anymore based on the logical restart LSN
+ * 1) Alert the custodian to remove all mappings not needed anymore based on the
+ * logical restart LSN
* 2) Flush all remaining mappings to disk, so that replay after a checkpoint
* only has to deal with the parts of a mapping that have been written out
* after the checkpoint started.
@@ -1211,6 +1214,11 @@ CheckPointLogicalRewriteHeap(void)
if (cutoff != InvalidXLogRecPtr && redo < cutoff)
cutoff = redo;
+ /* let the custodian know what it can remove */
+ RequestCustodian(CUSTODIAN_REMOVE_REWRITE_MAPPINGS,
+ !IsUnderPostmaster,
+ LSNGetDatum(cutoff));
+
mappings_dir = AllocateDir("pg_logical/mappings");
while ((mapping_de = ReadDir(mappings_dir, "pg_logical/mappings")) != NULL)
{
@@ -1243,15 +1251,7 @@ CheckPointLogicalRewriteHeap(void)
lsn = ((uint64) hi) << 32 | lo;
- if (lsn < cutoff || cutoff == InvalidXLogRecPtr)
- {
- elog(DEBUG1, "removing logical rewrite file \"%s\"", path);
- if (unlink(path) < 0)
- ereport(ERROR,
- (errcode_for_file_access(),
- errmsg("could not remove file \"%s\": %m", path)));
- }
- else
+ if (lsn >= cutoff && cutoff != InvalidXLogRecPtr)
{
/* on some operating systems fsyncing a file requires O_RDWR */
int fd = OpenTransientFile(path, O_RDWR | PG_BINARY);
@@ -1289,3 +1289,61 @@ CheckPointLogicalRewriteHeap(void)
/* persist directory entries to disk */
fsync_fname("pg_logical/mappings", true);
}
+
+/*
+ * Remove all mappings not needed anymore based on the logical restart LSN saved
+ * by the checkpointer. We use this saved value instead of calling
+ * ReplicationSlotsComputeLogicalRestartLSN() so that we don't try to remove
+ * files that a concurrent call to CheckPointLogicalRewriteHeap() is trying to
+ * flush to disk.
+ */
+void
+RemoveOldLogicalRewriteMappings(void)
+{
+ XLogRecPtr cutoff;
+ DIR *mappings_dir;
+ struct dirent *mapping_de;
+ char path[MAXPGPATH + 20];
+
+ cutoff = CustodianGetLogicalRewriteCutoff();
+
+ mappings_dir = AllocateDir("pg_logical/mappings");
+ while ((mapping_de = ReadDir(mappings_dir, "pg_logical/mappings")) != NULL)
+ {
+ struct stat statbuf;
+ Oid dboid;
+ Oid relid;
+ XLogRecPtr lsn;
+ TransactionId rewrite_xid;
+ TransactionId create_xid;
+ uint32 hi,
+ lo;
+
+ if (strcmp(mapping_de->d_name, ".") == 0 ||
+ strcmp(mapping_de->d_name, "..") == 0)
+ continue;
+
+ snprintf(path, sizeof(path), "pg_logical/mappings/%s", mapping_de->d_name);
+ if (lstat(path, &statbuf) == 0 && !S_ISREG(statbuf.st_mode))
+ continue;
+
+ /* Skip over files that cannot be ours. */
+ if (strncmp(mapping_de->d_name, "map-", 4) != 0)
+ continue;
+
+ if (sscanf(mapping_de->d_name, LOGICAL_REWRITE_FORMAT,
+ &dboid, &relid, &hi, &lo, &rewrite_xid, &create_xid) != 6)
+ elog(ERROR, "could not parse filename \"%s\"", mapping_de->d_name);
+
+ lsn = ((uint64) hi) << 32 | lo;
+ if (lsn >= cutoff && cutoff != InvalidXLogRecPtr)
+ continue;
+
+ elog(DEBUG1, "removing logical rewrite file \"%s\"", path);
+ if (unlink(path) < 0)
+ ereport(ERROR,
+ (errcode_for_file_access(),
+ errmsg("could not remove file \"%s\": %m", path)));
+ }
+ FreeDir(mappings_dir);
+}
diff --git a/src/backend/postmaster/custodian.c b/src/backend/postmaster/custodian.c
index 855a756ca0..d4be19e5de 100644
--- a/src/backend/postmaster/custodian.c
+++ b/src/backend/postmaster/custodian.c
@@ -21,6 +21,7 @@
*/
#include "postgres.h"
+#include "access/rewriteheap.h"
#include "libpq/pqsignal.h"
#include "pgstat.h"
#include "postmaster/custodian.h"
@@ -33,11 +34,13 @@
#include "storage/procsignal.h"
#include "storage/smgr.h"
#include "utils/memutils.h"
+#include "utils/pg_lsn.h"
static void DoCustodianTasks(bool retry);
static CustodianTask CustodianGetNextTask(void);
static void CustodianEnqueueTask(CustodianTask task);
static const struct cust_task_funcs_entry *LookupCustodianFunctions(CustodianTask task);
+static void CustodianSetLogicalRewriteCutoff(Datum arg);
typedef struct
{
@@ -45,6 +48,8 @@ typedef struct
CustodianTask task_queue_elems[NUM_CUSTODIAN_TASKS];
int task_queue_head;
+
+ XLogRecPtr logical_rewrite_mappings_cutoff; /* can remove older mappings */
} CustodianShmemStruct;
static CustodianShmemStruct *CustodianShmem;
@@ -73,6 +78,7 @@ struct cust_task_funcs_entry
static const struct cust_task_funcs_entry cust_task_functions[] = {
{CUSTODIAN_REMOVE_TEMP_FILES, RemovePgTempFiles, NULL},
{CUSTODIAN_REMOVE_SERIALIZED_SNAPSHOTS, RemoveOldSerializedSnapshots, NULL},
+ {CUSTODIAN_REMOVE_REWRITE_MAPPINGS, RemoveOldLogicalRewriteMappings, CustodianSetLogicalRewriteCutoff},
{INVALID_CUSTODIAN_TASK, NULL, NULL} /* must be last */
};
@@ -384,3 +390,40 @@ LookupCustodianFunctions(CustodianTask task)
elog(ERROR, "could not lookup functions for custodian task %d", task);
pg_unreachable();
}
+
+/*
+ * Stores the provided cutoff LSN in the custodian's shared memory.
+ *
+ * It's okay if the cutoff LSN is updated before a previously set cutoff has
+ * been used for cleaning up files. If that happens, it just means that the
+ * next invocation of RemoveOldLogicalRewriteMappings() will use a more accurate
+ * cutoff.
+ */
+static void
+CustodianSetLogicalRewriteCutoff(Datum arg)
+{
+ SpinLockAcquire(&CustodianShmem->cust_lck);
+ CustodianShmem->logical_rewrite_mappings_cutoff = DatumGetLSN(arg);
+ SpinLockRelease(&CustodianShmem->cust_lck);
+
+ /* if pass-by-ref, free Datum memory */
+#ifndef USE_FLOAT8_BYVAL
+ pfree(DatumGetPointer(arg));
+#endif
+}
+
+/*
+ * Used by the custodian to determine which logical rewrite mapping files it can
+ * remove.
+ */
+XLogRecPtr
+CustodianGetLogicalRewriteCutoff(void)
+{
+ XLogRecPtr cutoff;
+
+ SpinLockAcquire(&CustodianShmem->cust_lck);
+ cutoff = CustodianShmem->logical_rewrite_mappings_cutoff;
+ SpinLockRelease(&CustodianShmem->cust_lck);
+
+ return cutoff;
+}
diff --git a/src/include/access/rewriteheap.h b/src/include/access/rewriteheap.h
index 353cbb2924..965372b5ff 100644
--- a/src/include/access/rewriteheap.h
+++ b/src/include/access/rewriteheap.h
@@ -53,5 +53,6 @@ typedef struct LogicalRewriteMappingData
*/
#define LOGICAL_REWRITE_FORMAT "map-%x-%x-%X_%X-%x-%x"
extern void CheckPointLogicalRewriteHeap(void);
+extern void RemoveOldLogicalRewriteMappings(void);
#endif /* REWRITE_HEAP_H */
diff --git a/src/include/postmaster/custodian.h b/src/include/postmaster/custodian.h
index 37334941cc..f177d55159 100644
--- a/src/include/postmaster/custodian.h
+++ b/src/include/postmaster/custodian.h
@@ -12,6 +12,8 @@
#ifndef _CUSTODIAN_H
#define _CUSTODIAN_H
+#include "access/xlogdefs.h"
+
/*
* If you add a new task here, be sure to add its corresponding function
* pointers to cust_task_functions in custodian.c.
@@ -20,6 +22,7 @@ typedef enum CustodianTask
{
CUSTODIAN_REMOVE_TEMP_FILES,
CUSTODIAN_REMOVE_SERIALIZED_SNAPSHOTS,
+ CUSTODIAN_REMOVE_REWRITE_MAPPINGS,
NUM_CUSTODIAN_TASKS, /* new tasks go above */
INVALID_CUSTODIAN_TASK
@@ -29,5 +32,6 @@ extern void CustodianMain(void) pg_attribute_noreturn();
extern Size CustodianShmemSize(void);
extern void CustodianShmemInit(void);
extern void RequestCustodian(CustodianTask task, bool immediate, Datum arg);
+extern XLogRecPtr CustodianGetLogicalRewriteCutoff(void);
#endif /* _CUSTODIAN_H */
--
2.25.1
--+QahgC5+KEYLbs62--
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2021-12-14 01:31 Tomas Vondra <[email protected]>
parent: Peter Eisentraut <[email protected]>
0 siblings, 3 replies; 127+ messages in thread
From: Tomas Vondra @ 2021-12-14 01:31 UTC (permalink / raw)
To: Peter Eisentraut <[email protected]>; +Cc: PostgreSQL Hackers <[email protected]>; Petr Jelinek <[email protected]>
Hi,
here's an updated version of the patches, dealing with almost all of the
issues (at least in the 0001 and 0002 parts). The main changes:
1) I've removed the 'created' flag from fill_seq_with_data, as
discussed. I don't think it's needed by any of the parts (not even 0003,
AFAICS). We still need it in xl_seq_rec, though.
2) GetCurrentTransactionId() added to sequence.c are called only with
wal_level=logical, to minimize the overhead.
There's still one remaining problem, that I already explained in [1].
The problem is that with this:
BEGIN;
SELECT nextval('s') FROM generate_series(1,100);
ROLLBACK;
The root cause is that pg_current_wal_lsn() uses the LogwrtResult.Write,
which is updated by XLogFlush() - but only in RecordTransactionCommit.
Which makes sense, because only the committed stuff is "visible".
But the non-transactional behavior of sequence decoding disagrees with
this, because now some of the changes from aborted transactions may be
replicated. Which means the wait_for_catchup() ends up not waiting for
the sequence change to be replicated. This is an issue for tests in
patch 0003, at least.
My concern is this actually affects other places waiting for things
getting replicated :-/
I have outlined some ideas how to deal with this in [1], but we got
sidetracked by exploring the alternative approach.
regards
[1]
https://www.postgresql.org/message-id/3d6df331-5532-6848-eb45-344b265e0238%40enterprisedb.com
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
Attachments:
[text/x-patch] 0003-Add-support-for-decoding-sequences-to-built-20211214.patch (76.9K, ../../[email protected]/2-0003-Add-support-for-decoding-sequences-to-built-20211214.patch)
download | inline diff:
From 27522e687bca7c067611343e4d22f57deeb9d03c Mon Sep 17 00:00:00 2001
From: Tomas Vondra <[email protected]>
Date: Tue, 14 Dec 2021 00:29:09 +0100
Subject: [PATCH 3/3] Add support for decoding sequences to built-in
replication
---
doc/src/sgml/catalogs.sgml | 71 ++++
doc/src/sgml/ref/alter_publication.sgml | 24 +-
doc/src/sgml/ref/alter_subscription.sgml | 4 +-
src/backend/catalog/pg_publication.c | 149 ++++++++-
src/backend/catalog/system_views.sql | 10 +
src/backend/commands/publicationcmds.c | 350 +++++++++++++++++++-
src/backend/commands/sequence.c | 79 +++++
src/backend/commands/subscriptioncmds.c | 272 +++++++++++++++
src/backend/executor/execReplication.c | 2 +-
src/backend/nodes/copyfuncs.c | 1 +
src/backend/nodes/equalfuncs.c | 1 +
src/backend/parser/gram.y | 32 ++
src/backend/replication/logical/proto.c | 52 +++
src/backend/replication/logical/tablesync.c | 118 ++++++-
src/backend/replication/logical/worker.c | 60 ++++
src/backend/replication/pgoutput/pgoutput.c | 85 ++++-
src/backend/utils/cache/relcache.c | 4 +-
src/bin/psql/tab-complete.c | 6 +-
src/include/catalog/pg_proc.dat | 5 +
src/include/catalog/pg_publication.h | 14 +
src/include/commands/sequence.h | 1 +
src/include/nodes/parsenodes.h | 6 +
src/include/replication/logicalproto.h | 19 ++
src/include/replication/pgoutput.h | 1 +
src/test/regress/expected/rules.out | 8 +
src/test/subscription/t/027_sequences.pl | 196 +++++++++++
26 files changed, 1538 insertions(+), 32 deletions(-)
create mode 100644 src/test/subscription/t/027_sequences.pl
diff --git a/doc/src/sgml/catalogs.sgml b/doc/src/sgml/catalogs.sgml
index 03e2537b07d..0dfeb36f03f 100644
--- a/doc/src/sgml/catalogs.sgml
+++ b/doc/src/sgml/catalogs.sgml
@@ -9514,6 +9514,11 @@ SCRAM-SHA-256$<replaceable><iteration count></replaceable>:<replaceable>&l
<entry>prepared transactions</entry>
</row>
+ <row>
+ <entry><link linkend="view-pg-publication-sequences"><structname>pg_publication_sequences</structname></link></entry>
+ <entry>publications and their associated sequences</entry>
+ </row>
+
<row>
<entry><link linkend="view-pg-publication-tables"><structname>pg_publication_tables</structname></link></entry>
<entry>publications and their associated tables</entry>
@@ -11349,6 +11354,72 @@ SELECT * FROM pg_locks pl LEFT JOIN pg_prepared_xacts ppx
</sect1>
+ <sect1 id="view-pg-publication-sequences">
+ <title><structname>pg_publication_sequences</structname></title>
+
+ <indexterm zone="view-pg-publication-sequences">
+ <primary>pg_publication_sequences</primary>
+ </indexterm>
+
+ <para>
+ The view <structname>pg_publication_sequences</structname> provides
+ information about the mapping between publications and the sequences they
+ contain. Unlike the underlying catalog
+ <link linkend="catalog-pg-publication-rel"><structname>pg_publication_rel</structname></link>,
+ this view expands
+ publications defined as <literal>FOR ALL SEQUENCES</literal>, so for such
+ publications there will be a row for each eligible sequence.
+ </para>
+
+ <table>
+ <title><structname>pg_publication_sequences</structname> Columns</title>
+ <tgroup cols="1">
+ <thead>
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ Column Type
+ </para>
+ <para>
+ Description
+ </para></entry>
+ </row>
+ </thead>
+
+ <tbody>
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ <structfield>pubname</structfield> <type>name</type>
+ (references <link linkend="catalog-pg-publication"><structname>pg_publication</structname></link>.<structfield>pubname</structfield>)
+ </para>
+ <para>
+ Name of publication
+ </para></entry>
+ </row>
+
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ <structfield>schemaname</structfield> <type>name</type>
+ (references <link linkend="catalog-pg-namespace"><structname>pg_namespace</structname></link>.<structfield>nspname</structfield>)
+ </para>
+ <para>
+ Name of schema containing sequence
+ </para></entry>
+ </row>
+
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ <structfield>sequencename</structfield> <type>name</type>
+ (references <link linkend="catalog-pg-class"><structname>pg_class</structname></link>.<structfield>relname</structfield>)
+ </para>
+ <para>
+ Name of sequence
+ </para></entry>
+ </row>
+ </tbody>
+ </tgroup>
+ </table>
+ </sect1>
+
<sect1 id="view-pg-publication-tables">
<title><structname>pg_publication_tables</structname></title>
diff --git a/doc/src/sgml/ref/alter_publication.sgml b/doc/src/sgml/ref/alter_publication.sgml
index bb4ef5e5e22..4d166ad3f9c 100644
--- a/doc/src/sgml/ref/alter_publication.sgml
+++ b/doc/src/sgml/ref/alter_publication.sgml
@@ -31,7 +31,9 @@ ALTER PUBLICATION <replaceable class="parameter">name</replaceable> RENAME TO <r
<phrase>where <replaceable class="parameter">publication_object</replaceable> is one of:</phrase>
TABLE [ ONLY ] <replaceable class="parameter">table_name</replaceable> [ * ] [, ... ]
+ SEQUENCE <replaceable class="parameter">sequence_name</replaceable> [ * ] [, ... ]
ALL TABLES IN SCHEMA { <replaceable class="parameter">schema_name</replaceable> | CURRENT_SCHEMA } [, ... ]
+ ALL SEQUENCE IN SCHEMA { <replaceable class="parameter">schema_name</replaceable> | CURRENT_SCHEMA } [, ... ]
</synopsis>
</refsynopsisdiv>
@@ -56,7 +58,18 @@ ALTER PUBLICATION <replaceable class="parameter">name</replaceable> RENAME TO <r
</para>
<para>
- The fourth variant of this command listed in the synopsis can change
+ The next three variants change which sequences are part of the publication.
+ The <literal>SET SEQUENCE</literal> clause will replace the list of sequences
+ in the publication with the specified one. The <literal>ADD SEQUENCE</literal>
+ and <literal>DROP SEQUENCE</literal> clauses will add and remove one or more
+ sequences from the publication. Note that adding sequences to a publication
+ that is already subscribed to will require a <literal>ALTER SUBSCRIPTION
+ ... REFRESH PUBLICATION</literal> action on the subscribing side in order
+ to become effective.
+ </para>
+
+ <para>
+ The seventh variant of this command listed in the synopsis can change
all of the publication properties specified in
<xref linkend="sql-createpublication"/>. Properties not mentioned in the
command retain their previous settings.
@@ -123,6 +136,15 @@ ALTER PUBLICATION <replaceable class="parameter">name</replaceable> RENAME TO <r
</listitem>
</varlistentry>
+ <varlistentry>
+ <term><replaceable class="parameter">sequence_name</replaceable></term>
+ <listitem>
+ <para>
+ Name of an existing sequence.
+ </para>
+ </listitem>
+ </varlistentry>
+
<varlistentry>
<term><literal>SET ( <replaceable class="parameter">publication_parameter</replaceable> [= <replaceable class="parameter">value</replaceable>] [, ... ] )</literal></term>
<listitem>
diff --git a/doc/src/sgml/ref/alter_subscription.sgml b/doc/src/sgml/ref/alter_subscription.sgml
index 0b027cc3462..8f28cf03f40 100644
--- a/doc/src/sgml/ref/alter_subscription.sgml
+++ b/doc/src/sgml/ref/alter_subscription.sgml
@@ -147,7 +147,7 @@ ALTER SUBSCRIPTION <replaceable class="parameter">name</replaceable> RENAME TO <
<listitem>
<para>
Fetch missing table information from publisher. This will start
- replication of tables that were added to the subscribed-to publications
+ replication of tables and sequences that were added to the subscribed-to publications
since <command>CREATE SUBSCRIPTION</command> or
the last invocation of <command>REFRESH PUBLICATION</command>.
</para>
@@ -164,7 +164,7 @@ ALTER SUBSCRIPTION <replaceable class="parameter">name</replaceable> RENAME TO <
Specifies whether to copy pre-existing data in the publications
that are being subscribed to when the replication starts.
The default is <literal>true</literal>. (Previously-subscribed
- tables are not copied.)
+ tables and sequences are not copied.)
</para>
</listitem>
</varlistentry>
diff --git a/src/backend/catalog/pg_publication.c b/src/backend/catalog/pg_publication.c
index 62f10bcbd28..6844ac61897 100644
--- a/src/backend/catalog/pg_publication.c
+++ b/src/backend/catalog/pg_publication.c
@@ -54,7 +54,8 @@ check_publication_add_relation(Relation targetrel)
{
/* Must be a regular or partitioned table */
if (RelationGetForm(targetrel)->relkind != RELKIND_RELATION &&
- RelationGetForm(targetrel)->relkind != RELKIND_PARTITIONED_TABLE)
+ RelationGetForm(targetrel)->relkind != RELKIND_PARTITIONED_TABLE &&
+ RelationGetForm(targetrel)->relkind != RELKIND_SEQUENCE)
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("cannot add relation \"%s\" to publication",
@@ -131,7 +132,8 @@ static bool
is_publishable_class(Oid relid, Form_pg_class reltuple)
{
return (reltuple->relkind == RELKIND_RELATION ||
- reltuple->relkind == RELKIND_PARTITIONED_TABLE) &&
+ reltuple->relkind == RELKIND_PARTITIONED_TABLE ||
+ reltuple->relkind == RELKIND_SEQUENCE) &&
!IsCatalogRelationOid(relid) &&
reltuple->relpersistence == RELPERSISTENCE_PERMANENT &&
relid >= FirstNormalObjectId;
@@ -502,6 +504,11 @@ GetPublicationRelations(Oid pubid, PublicationPartOpt pub_partopt)
Form_pg_publication_rel pubrel;
pubrel = (Form_pg_publication_rel) GETSTRUCT(tup);
+
+ /* skip sequences here */
+ if (get_rel_relkind(pubrel->prrelid) == RELKIND_SEQUENCE)
+ continue;
+
result = GetPubPartitionOptionRelations(result, pub_partopt,
pubrel->prrelid);
}
@@ -516,6 +523,49 @@ GetPublicationRelations(Oid pubid, PublicationPartOpt pub_partopt)
return result;
}
+/*
+ * Gets list of relation oids for a publication (sequences only).
+ *
+ * This should only be used for normal publications, the FOR ALL TABLES
+ * should use GetAllSequencesPublicationRelations().
+ */
+List *
+GetPublicationSequenceRelations(Oid pubid)
+{
+ List *result;
+ Relation pubrelsrel;
+ ScanKeyData scankey;
+ SysScanDesc scan;
+ HeapTuple tup;
+
+ /* Find all publications associated with the relation. */
+ pubrelsrel = table_open(PublicationRelRelationId, AccessShareLock);
+
+ ScanKeyInit(&scankey,
+ Anum_pg_publication_rel_prpubid,
+ BTEqualStrategyNumber, F_OIDEQ,
+ ObjectIdGetDatum(pubid));
+
+ scan = systable_beginscan(pubrelsrel, PublicationRelPrrelidPrpubidIndexId,
+ true, NULL, 1, &scankey);
+
+ result = NIL;
+ while (HeapTupleIsValid(tup = systable_getnext(scan)))
+ {
+ Form_pg_publication_rel pubrel;
+
+ pubrel = (Form_pg_publication_rel) GETSTRUCT(tup);
+
+ if (get_rel_relkind(pubrel->prrelid) == RELKIND_SEQUENCE)
+ result = lappend_oid(result, pubrel->prrelid);
+ }
+
+ systable_endscan(scan);
+ table_close(pubrelsrel, AccessShareLock);
+
+ return result;
+}
+
/*
* Gets list of publication oids for publications marked as FOR ALL TABLES.
*/
@@ -761,6 +811,46 @@ GetAllSchemaPublicationRelations(Oid pubid, PublicationPartOpt pub_partopt)
return result;
}
+/*
+ * Gets list of all relation published by FOR ALL TABLES publication(s).
+ *
+ * If the publication publishes partition changes via their respective root
+ * partitioned tables, we must exclude partitions in favor of including the
+ * root partitioned tables.
+ */
+List *
+GetAllSequencesPublicationRelations(void)
+{
+ Relation classRel;
+ ScanKeyData key[1];
+ TableScanDesc scan;
+ HeapTuple tuple;
+ List *result = NIL;
+
+ classRel = table_open(RelationRelationId, AccessShareLock);
+
+ ScanKeyInit(&key[0],
+ Anum_pg_class_relkind,
+ BTEqualStrategyNumber, F_CHAREQ,
+ CharGetDatum(RELKIND_SEQUENCE));
+
+ scan = table_beginscan_catalog(classRel, 1, key);
+
+ while ((tuple = heap_getnext(scan, ForwardScanDirection)) != NULL)
+ {
+ Form_pg_class relForm = (Form_pg_class) GETSTRUCT(tuple);
+ Oid relid = relForm->oid;
+
+ if (is_publishable_class(relid, relForm))
+ result = lappend_oid(result, relid);
+ }
+
+ table_endscan(scan);
+
+ table_close(classRel, AccessShareLock);
+ return result;
+}
+
/*
* Get publication using oid
*
@@ -783,10 +873,12 @@ GetPublication(Oid pubid)
pub->oid = pubid;
pub->name = pstrdup(NameStr(pubform->pubname));
pub->alltables = pubform->puballtables;
+ pub->allsequences = pubform->puballsequences;
pub->pubactions.pubinsert = pubform->pubinsert;
pub->pubactions.pubupdate = pubform->pubupdate;
pub->pubactions.pubdelete = pubform->pubdelete;
pub->pubactions.pubtruncate = pubform->pubtruncate;
+ pub->pubactions.pubsequence = pubform->pubsequence;
pub->pubviaroot = pubform->pubviaroot;
ReleaseSysCache(tup);
@@ -936,3 +1028,56 @@ pg_get_publication_tables(PG_FUNCTION_ARGS)
SRF_RETURN_DONE(funcctx);
}
+
+/*
+ * Returns Oids of sequences in a publication.
+ */
+Datum
+pg_get_publication_sequences(PG_FUNCTION_ARGS)
+{
+ FuncCallContext *funcctx;
+ char *pubname = text_to_cstring(PG_GETARG_TEXT_PP(0));
+ Publication *publication;
+ List *sequences;
+
+ /* stuff done only on the first call of the function */
+ if (SRF_IS_FIRSTCALL())
+ {
+ MemoryContext oldcontext;
+
+ /* create a function context for cross-call persistence */
+ funcctx = SRF_FIRSTCALL_INIT();
+
+ /* switch to memory context appropriate for multiple function calls */
+ oldcontext = MemoryContextSwitchTo(funcctx->multi_call_memory_ctx);
+
+ publication = GetPublicationByName(pubname, false);
+
+ /*
+ * Publications support partitioned tables, although all changes are
+ * replicated using leaf partition identity and schema, so we only
+ * need those.
+ */
+ if (publication->allsequences)
+ sequences = GetAllSequencesPublicationRelations();
+ else
+ sequences = GetPublicationSequenceRelations(publication->oid);
+
+ funcctx->user_fctx = (void *) sequences;
+
+ MemoryContextSwitchTo(oldcontext);
+ }
+
+ /* stuff done on every call of the function */
+ funcctx = SRF_PERCALL_SETUP();
+ sequences = (List *) funcctx->user_fctx;
+
+ if (funcctx->call_cntr < list_length(sequences))
+ {
+ Oid relid = list_nth_oid(sequences, funcctx->call_cntr);
+
+ SRF_RETURN_NEXT(funcctx, ObjectIdGetDatum(relid));
+ }
+
+ SRF_RETURN_DONE(funcctx);
+}
diff --git a/src/backend/catalog/system_views.sql b/src/backend/catalog/system_views.sql
index 61b515cdb85..0df91d7a5ad 100644
--- a/src/backend/catalog/system_views.sql
+++ b/src/backend/catalog/system_views.sql
@@ -372,6 +372,16 @@ CREATE VIEW pg_publication_tables AS
pg_class C JOIN pg_namespace N ON (N.oid = C.relnamespace)
WHERE C.oid = GPT.relid;
+CREATE VIEW pg_publication_sequences AS
+ SELECT
+ P.pubname AS pubname,
+ N.nspname AS schemaname,
+ C.relname AS sequencename
+ FROM pg_publication P,
+ LATERAL pg_get_publication_sequences(P.pubname) GPT,
+ pg_class C JOIN pg_namespace N ON (N.oid = C.relnamespace)
+ WHERE C.oid = GPT.relid;
+
CREATE VIEW pg_locks AS
SELECT * FROM pg_lock_status() AS L;
diff --git a/src/backend/commands/publicationcmds.c b/src/backend/commands/publicationcmds.c
index 404bb5d0c87..2c183d63f7d 100644
--- a/src/backend/commands/publicationcmds.c
+++ b/src/backend/commands/publicationcmds.c
@@ -16,6 +16,7 @@
#include "access/genam.h"
#include "access/htup_details.h"
+#include "access/relation.h"
#include "access/table.h"
#include "access/xact.h"
#include "catalog/catalog.h"
@@ -59,6 +60,12 @@ static void PublicationAddSchemas(Oid pubid, List *schemas, bool if_not_exists,
AlterPublicationStmt *stmt);
static void PublicationDropSchemas(Oid pubid, List *schemas, bool missing_ok);
+static List *OpenSequenceList(List *sequences);
+static void CloseSequenceList(List *rels);
+static void PublicationAddSequences(Oid pubid, List *rels, bool if_not_exists,
+ AlterPublicationStmt *stmt);
+static void PublicationDropSequences(Oid pubid, List *rels, bool missing_ok);
+
static void
parse_publication_options(ParseState *pstate,
List *options,
@@ -77,6 +84,7 @@ parse_publication_options(ParseState *pstate,
pubactions->pubupdate = true;
pubactions->pubdelete = true;
pubactions->pubtruncate = true;
+ pubactions->pubsequence = true;
*publish_via_partition_root = false;
/* Parse options */
@@ -101,6 +109,7 @@ parse_publication_options(ParseState *pstate,
pubactions->pubupdate = false;
pubactions->pubdelete = false;
pubactions->pubtruncate = false;
+ pubactions->pubsequence = false;
*publish_given = true;
publish = defGetString(defel);
@@ -123,6 +132,8 @@ parse_publication_options(ParseState *pstate,
pubactions->pubdelete = true;
else if (strcmp(publish_opt, "truncate") == 0)
pubactions->pubtruncate = true;
+ else if (strcmp(publish_opt, "sequence") == 0)
+ pubactions->pubsequence = true;
else
ereport(ERROR,
(errcode(ERRCODE_SYNTAX_ERROR),
@@ -149,7 +160,9 @@ parse_publication_options(ParseState *pstate,
*/
static void
ObjectsInPublicationToOids(List *pubobjspec_list, ParseState *pstate,
- List **rels, List **schemas)
+ List **tables, List **sequences,
+ List **tables_schemas, List **sequences_schemas,
+ List **schemas)
{
ListCell *cell;
PublicationObjSpec *pubobj;
@@ -167,12 +180,23 @@ ObjectsInPublicationToOids(List *pubobjspec_list, ParseState *pstate,
switch (pubobj->pubobjtype)
{
case PUBLICATIONOBJ_TABLE:
- *rels = lappend(*rels, pubobj->pubtable);
+ *tables = lappend(*tables, pubobj->pubtable);
+ break;
+ case PUBLICATIONOBJ_SEQUENCE:
+ *sequences = lappend(*sequences, pubobj->pubtable);
break;
case PUBLICATIONOBJ_TABLE_IN_SCHEMA:
schemaid = get_namespace_oid(pubobj->name, false);
/* Filter out duplicates if user specifies "sch1, sch1" */
+ *tables_schemas = list_append_unique_oid(*tables_schemas, schemaid);
+ *schemas = list_append_unique_oid(*schemas, schemaid);
+ break;
+ case PUBLICATIONOBJ_SEQUENCE_IN_SCHEMA:
+ schemaid = get_namespace_oid(pubobj->name, false);
+
+ /* Filter out duplicates if user specifies "sch1, sch1" */
+ *sequences_schemas = list_append_unique_oid(*sequences_schemas, schemaid);
*schemas = list_append_unique_oid(*schemas, schemaid);
break;
case PUBLICATIONOBJ_TABLE_IN_CUR_SCHEMA:
@@ -186,6 +210,21 @@ ObjectsInPublicationToOids(List *pubobjspec_list, ParseState *pstate,
list_free(search_path);
/* Filter out duplicates if user specifies "sch1, sch1" */
+ *tables_schemas = list_append_unique_oid(*tables_schemas, schemaid);
+ *schemas = list_append_unique_oid(*schemas, schemaid);
+ break;
+ case PUBLICATIONOBJ_SEQUENCE_IN_CUR_SCHEMA:
+ search_path = fetch_search_path(false);
+ if (search_path == NIL) /* nothing valid in search_path? */
+ ereport(ERROR,
+ errcode(ERRCODE_UNDEFINED_SCHEMA),
+ errmsg("no schema has been selected for CURRENT_SCHEMA"));
+
+ schemaid = linitial_oid(search_path);
+ list_free(search_path);
+
+ /* Filter out duplicates if user specifies "sch1, sch1" */
+ *sequences_schemas = list_append_unique_oid(*sequences_schemas, schemaid);
*schemas = list_append_unique_oid(*schemas, schemaid);
break;
default:
@@ -251,7 +290,10 @@ CreatePublication(ParseState *pstate, CreatePublicationStmt *stmt)
bool publish_via_partition_root_given;
bool publish_via_partition_root;
AclResult aclresult;
- List *relations = NIL;
+ List *tables = NIL;
+ List *sequences = NIL;
+ List *tables_schemaidlist = NIL;
+ List *sequences_schemaidlist = NIL;
List *schemaidlist = NIL;
/* must have CREATE privilege on database */
@@ -306,6 +348,8 @@ CreatePublication(ParseState *pstate, CreatePublicationStmt *stmt)
BoolGetDatum(pubactions.pubdelete);
values[Anum_pg_publication_pubtruncate - 1] =
BoolGetDatum(pubactions.pubtruncate);
+ values[Anum_pg_publication_pubsequence - 1] =
+ BoolGetDatum(pubactions.pubsequence);
values[Anum_pg_publication_pubviaroot - 1] =
BoolGetDatum(publish_via_partition_root);
@@ -330,26 +374,40 @@ CreatePublication(ParseState *pstate, CreatePublicationStmt *stmt)
}
else
{
- ObjectsInPublicationToOids(stmt->pubobjects, pstate, &relations,
+ ObjectsInPublicationToOids(stmt->pubobjects, pstate,
+ &tables, &sequences,
+ &tables_schemaidlist,
+ &sequences_schemaidlist,
&schemaidlist);
/* FOR ALL TABLES IN SCHEMA requires superuser */
- if (list_length(schemaidlist) > 0 && !superuser())
+ if (list_length(tables_schemaidlist) > 0 && !superuser())
ereport(ERROR,
errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
errmsg("must be superuser to create FOR ALL TABLES IN SCHEMA publication"));
- if (list_length(relations) > 0)
+ if (list_length(tables) > 0)
{
List *rels;
- rels = OpenTableList(relations);
- CheckObjSchemaNotAlreadyInPublication(rels, schemaidlist,
+ rels = OpenTableList(tables);
+ CheckObjSchemaNotAlreadyInPublication(rels, tables_schemaidlist,
PUBLICATIONOBJ_TABLE);
PublicationAddTables(puboid, rels, true, NULL);
CloseTableList(rels);
}
+ if (list_length(sequences) > 0)
+ {
+ List *rels;
+
+ rels = OpenSequenceList(sequences);
+ CheckObjSchemaNotAlreadyInPublication(rels, sequences_schemaidlist,
+ PUBLICATIONOBJ_SEQUENCE);
+ PublicationAddTables(puboid, rels, true, NULL);
+ CloseSequenceList(rels);
+ }
+
if (list_length(schemaidlist) > 0)
{
/*
@@ -651,12 +709,13 @@ AlterPublicationSchemas(AlterPublicationStmt *stmt,
*/
static void
CheckAlterPublication(AlterPublicationStmt *stmt, HeapTuple tup,
- List *tables, List *schemaidlist)
+ List *tables, List *tables_schemaidlist,
+ List *sequences, List *sequences_schemaidlist)
{
Form_pg_publication pubform = (Form_pg_publication) GETSTRUCT(tup);
if ((stmt->action == DEFELEM_ADD || stmt->action == DEFELEM_SET) &&
- schemaidlist && !superuser())
+ (tables_schemaidlist || sequences_schemaidlist) && !superuser())
ereport(ERROR,
(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
errmsg("must be superuser to add or set schemas")));
@@ -665,13 +724,24 @@ CheckAlterPublication(AlterPublicationStmt *stmt, HeapTuple tup,
* Check that user is allowed to manipulate the publication tables in
* schema
*/
- if (schemaidlist && pubform->puballtables)
+ if (tables_schemaidlist && pubform->puballtables)
ereport(ERROR,
(errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
errmsg("publication \"%s\" is defined as FOR ALL TABLES",
NameStr(pubform->pubname)),
errdetail("Tables from schema cannot be added to, dropped from, or set on FOR ALL TABLES publications.")));
+ /*
+ * Check that user is allowed to manipulate the publication sequences in
+ * schema
+ */
+ if (sequences_schemaidlist && pubform->puballsequences)
+ ereport(ERROR,
+ (errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
+ errmsg("publication \"%s\" is defined as FOR ALL SEQUENCES",
+ NameStr(pubform->pubname)),
+ errdetail("Sequences from schema cannot be added to, dropped from, or set on FOR ALL SEQUENCES publications.")));
+
/* Check that user is allowed to manipulate the publication tables. */
if (tables && pubform->puballtables)
ereport(ERROR,
@@ -679,6 +749,108 @@ CheckAlterPublication(AlterPublicationStmt *stmt, HeapTuple tup,
errmsg("publication \"%s\" is defined as FOR ALL TABLES",
NameStr(pubform->pubname)),
errdetail("Tables cannot be added to or dropped from FOR ALL TABLES publications.")));
+
+ /* Check that user is allowed to manipulate the publication tables. */
+ if (sequences && pubform->puballsequences)
+ ereport(ERROR,
+ (errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
+ errmsg("publication \"%s\" is defined as FOR ALL SEQUENCES",
+ NameStr(pubform->pubname)),
+ errdetail("Sequences cannot be added to or dropped from FOR ALL SEQUENCES publications.")));
+}
+
+/*
+ * Add or remove sequence to/from publication.
+ */
+static void
+AlterPublicationSequences(AlterPublicationStmt *stmt, HeapTuple tup,
+ List *sequences, List *schemaidlist)
+{
+ List *rels = NIL;
+ Form_pg_publication pubform = (Form_pg_publication) GETSTRUCT(tup);
+ Oid pubid = pubform->oid;
+
+ /*
+ * It is quite possible that for the SET case user has not specified any
+ * tables in which case we need to remove all the existing tables.
+ */
+ if (!sequences && stmt->action != DEFELEM_SET)
+ return;
+
+ rels = OpenSequenceList(sequences);
+
+ if (stmt->action == DEFELEM_ADD)
+ {
+ List *schemas = NIL;
+
+ /*
+ * Check if the relation is member of the existing schema in the
+ * publication or member of the schema list specified.
+ */
+ schemas = list_concat_copy(schemaidlist, GetPublicationSchemas(pubid));
+ CheckObjSchemaNotAlreadyInPublication(rels, schemas,
+ PUBLICATIONOBJ_SEQUENCE);
+ PublicationAddSequences(pubid, rels, false, stmt);
+ }
+ else if (stmt->action == DEFELEM_DROP)
+ PublicationDropSequences(pubid, rels, false);
+ else /* DEFELEM_SET */
+ {
+ List *oldrelids = GetPublicationRelations(pubid,
+ PUBLICATION_PART_ROOT);
+ List *delrels = NIL;
+ ListCell *oldlc;
+
+ CheckObjSchemaNotAlreadyInPublication(rels, schemaidlist,
+ PUBLICATIONOBJ_SEQUENCE);
+
+ /* Calculate which relations to drop. */
+ foreach(oldlc, oldrelids)
+ {
+ Oid oldrelid = lfirst_oid(oldlc);
+ ListCell *newlc;
+ bool found = false;
+
+ foreach(newlc, rels)
+ {
+ PublicationRelInfo *newpubrel;
+
+ newpubrel = (PublicationRelInfo *) lfirst(newlc);
+ if (RelationGetRelid(newpubrel->relation) == oldrelid)
+ {
+ found = true;
+ break;
+ }
+ }
+ /* Not yet in the list, open it and add to the list */
+ if (!found)
+ {
+ Relation oldrel;
+ PublicationRelInfo *pubrel;
+
+ /* Wrap relation into PublicationRelInfo */
+ oldrel = table_open(oldrelid, ShareUpdateExclusiveLock);
+
+ pubrel = palloc(sizeof(PublicationRelInfo));
+ pubrel->relation = oldrel;
+
+ delrels = lappend(delrels, pubrel);
+ }
+ }
+
+ /* And drop them. */
+ PublicationDropSequences(pubid, delrels, true);
+
+ /*
+ * Don't bother calculating the difference for adding, we'll catch and
+ * skip existing ones when doing catalog update.
+ */
+ PublicationAddSequences(pubid, rels, true, stmt);
+
+ CloseSequenceList(delrels);
+ }
+
+ CloseSequenceList(rels);
}
/*
@@ -716,13 +888,21 @@ AlterPublication(ParseState *pstate, AlterPublicationStmt *stmt)
AlterPublicationOptions(pstate, stmt, rel, tup);
else
{
- List *relations = NIL;
+ List *tables = NIL;
+ List *sequences = NIL;
+ List *tables_schemaidlist = NIL;
+ List *sequences_schemaidlist = NIL;
List *schemaidlist = NIL;
- ObjectsInPublicationToOids(stmt->pubobjects, pstate, &relations,
+ ObjectsInPublicationToOids(stmt->pubobjects, pstate,
+ &tables, &sequences,
+ &tables_schemaidlist,
+ &sequences_schemaidlist,
&schemaidlist);
- CheckAlterPublication(stmt, tup, relations, schemaidlist);
+ CheckAlterPublication(stmt, tup,
+ tables, tables_schemaidlist,
+ sequences, sequences_schemaidlist);
/*
* Lock the publication so nobody else can do anything with it. This
@@ -747,7 +927,9 @@ AlterPublication(ParseState *pstate, AlterPublicationStmt *stmt)
errmsg("publication \"%s\" does not exist",
stmt->pubname));
- AlterPublicationTables(stmt, tup, relations, schemaidlist);
+ AlterPublicationTables(stmt, tup, tables, tables_schemaidlist);
+ AlterPublicationSequences(stmt, tup, sequences, sequences_schemaidlist);
+
AlterPublicationSchemas(stmt, tup, schemaidlist);
}
@@ -1155,6 +1337,144 @@ PublicationDropSchemas(Oid pubid, List *schemas, bool missing_ok)
}
}
+/*
+ * Open relations specified by a PublicationTable list.
+ * In the returned list of PublicationRelInfo, tables are locked
+ * in ShareUpdateExclusiveLock mode in order to add them to a publication.
+ */
+static List *
+OpenSequenceList(List *sequences)
+{
+ List *relids = NIL;
+ List *rels = NIL;
+ ListCell *lc;
+
+ /*
+ * Open, share-lock, and check all the explicitly-specified relations
+ */
+ foreach(lc, sequences)
+ {
+ PublicationTable *s = lfirst_node(PublicationTable, lc);
+ Relation rel;
+ Oid myrelid;
+ PublicationRelInfo *pub_rel;
+
+ /* Allow query cancel in case this takes a long time */
+ CHECK_FOR_INTERRUPTS();
+
+ rel = table_openrv(s->relation, ShareUpdateExclusiveLock);
+ myrelid = RelationGetRelid(rel);
+
+ /*
+ * Filter out duplicates if user specifies "foo, foo".
+ *
+ * Note that this algorithm is known to not be very efficient (O(N^2))
+ * but given that it only works on list of tables given to us by user
+ * it's deemed acceptable.
+ */
+ if (list_member_oid(relids, myrelid))
+ {
+ table_close(rel, ShareUpdateExclusiveLock);
+ continue;
+ }
+
+ pub_rel = palloc(sizeof(PublicationRelInfo));
+ pub_rel->relation = rel;
+ rels = lappend(rels, pub_rel);
+ relids = lappend_oid(relids, myrelid);
+ }
+
+ list_free(relids);
+
+ return rels;
+}
+
+/*
+ * Close all relations in the list.
+ */
+static void
+CloseSequenceList(List *rels)
+{
+ ListCell *lc;
+
+ foreach(lc, rels)
+ {
+ PublicationRelInfo *pub_rel;
+
+ pub_rel = (PublicationRelInfo *) lfirst(lc);
+ table_close(pub_rel->relation, NoLock);
+ }
+}
+
+/*
+ * Add listed tables to the publication.
+ */
+static void
+PublicationAddSequences(Oid pubid, List *rels, bool if_not_exists,
+ AlterPublicationStmt *stmt)
+{
+ ListCell *lc;
+
+ Assert(!stmt || !stmt->for_all_sequences);
+
+ foreach(lc, rels)
+ {
+ PublicationRelInfo *pub_rel = (PublicationRelInfo *) lfirst(lc);
+ Relation rel = pub_rel->relation;
+ ObjectAddress obj;
+
+ /* Must be owner of the sequence or superuser. */
+ if (!pg_class_ownercheck(RelationGetRelid(rel), GetUserId()))
+ aclcheck_error(ACLCHECK_NOT_OWNER, get_relkind_objtype(rel->rd_rel->relkind),
+ RelationGetRelationName(rel));
+
+ obj = publication_add_relation(pubid, pub_rel, if_not_exists);
+ if (stmt)
+ {
+ EventTriggerCollectSimpleCommand(obj, InvalidObjectAddress,
+ (Node *) stmt);
+
+ InvokeObjectPostCreateHook(PublicationRelRelationId,
+ obj.objectId, 0);
+ }
+ }
+}
+
+/*
+ * Remove listed sequences from the publication.
+ */
+static void
+PublicationDropSequences(Oid pubid, List *rels, bool missing_ok)
+{
+ ObjectAddress obj;
+ ListCell *lc;
+ Oid prid;
+
+ foreach(lc, rels)
+ {
+ Relation rel = (Relation) lfirst(lc);
+ Oid relid = RelationGetRelid(rel);
+
+ prid = GetSysCacheOid2(PUBLICATIONRELMAP, Anum_pg_publication_rel_oid,
+ ObjectIdGetDatum(relid),
+ ObjectIdGetDatum(pubid));
+ if (!OidIsValid(prid))
+ {
+ if (missing_ok)
+ continue;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_UNDEFINED_OBJECT),
+ errmsg("relation \"%s\" is not part of the publication",
+ RelationGetRelationName(rel))));
+ }
+
+ ObjectAddressSet(obj, PublicationRelRelationId, prid);
+ performDeletion(&obj, DROP_CASCADE, 0);
+ }
+}
+
+
/*
* Internal workhorse for changing a publication owner
*/
diff --git a/src/backend/commands/sequence.c b/src/backend/commands/sequence.c
index 626b5b252f4..b8888213f54 100644
--- a/src/backend/commands/sequence.c
+++ b/src/backend/commands/sequence.c
@@ -336,6 +336,85 @@ ResetSequence(Oid seq_relid)
relation_close(seq_rel, NoLock);
}
+/*
+ * Reset a sequence to its initial value.
+ *
+ * The change is made transactionally, so that on failure of the current
+ * transaction, the sequence will be restored to its previous state.
+ * We do that by creating a whole new relfilenode for the sequence; so this
+ * works much like the rewriting forms of ALTER TABLE.
+ *
+ * Caller is assumed to have acquired AccessExclusiveLock on the sequence,
+ * which must not be released until end of transaction. Caller is also
+ * responsible for permissions checking.
+ */
+void
+ResetSequence2(Oid seq_relid, int64 last_value, int64 log_cnt, bool is_called)
+{
+ Relation seq_rel;
+ SeqTable elm;
+ Form_pg_sequence_data seq;
+ Buffer buf;
+ HeapTupleData seqdatatuple;
+ HeapTuple tuple;
+
+ /*
+ * Read the old sequence. This does a bit more work than really
+ * necessary, but it's simple, and we do want to double-check that it's
+ * indeed a sequence.
+ */
+ init_sequence(seq_relid, &elm, &seq_rel);
+ (void) read_seq_tuple(seq_rel, &buf, &seqdatatuple);
+
+ /*
+ * Copy the existing sequence tuple.
+ */
+ tuple = heap_copytuple(&seqdatatuple);
+
+ /* Now we're done with the old page */
+ UnlockReleaseBuffer(buf);
+
+ /*
+ * Modify the copied tuple to execute the restart (compare the RESTART
+ * action in AlterSequence)
+ */
+ seq = (Form_pg_sequence_data) GETSTRUCT(tuple);
+ seq->last_value = last_value;
+ seq->is_called = is_called;
+ seq->log_cnt = log_cnt;
+
+ /*
+ * Create a new storage file for the sequence.
+ */
+ RelationSetNewRelfilenode(seq_rel, seq_rel->rd_rel->relpersistence);
+
+ /*
+ * Ensure sequence's relfrozenxid is at 0, since it won't contain any
+ * unfrozen XIDs. Same with relminmxid, since a sequence will never
+ * contain multixacts.
+ */
+ Assert(seq_rel->rd_rel->relfrozenxid == InvalidTransactionId);
+ Assert(seq_rel->rd_rel->relminmxid == InvalidMultiXactId);
+
+ /*
+ * Insert the modified tuple into the new storage file.
+ *
+ * XXX Maybe this should also use created=true, just like the other places
+ * calling fill_seq_with_data. That's probably needed for correct cascading
+ * replication.
+ *
+ * XXX That'd mean all fill_seq_with_data callers use created=true, making
+ * the parameter unnecessary.
+ */
+ fill_seq_with_data(seq_rel, tuple);
+
+ /* Clear local cache so that we don't think we have cached numbers */
+ /* Note that we do not change the currval() state */
+ elm->cached = elm->last;
+
+ relation_close(seq_rel, NoLock);
+}
+
/*
* Initialize a sequence's relation with the specified tuple as content
*/
diff --git a/src/backend/commands/subscriptioncmds.c b/src/backend/commands/subscriptioncmds.c
index 2b658080fee..fc6374c4450 100644
--- a/src/backend/commands/subscriptioncmds.c
+++ b/src/backend/commands/subscriptioncmds.c
@@ -85,6 +85,7 @@ typedef struct SubOpts
} SubOpts;
static List *fetch_table_list(WalReceiverConn *wrconn, List *publications);
+static List *fetch_sequence_list(WalReceiverConn *wrconn, List *publications);
static void check_duplicates_in_publist(List *publist, Datum *datums);
static List *merge_publications(List *oldpublist, List *newpublist, bool addpub, const char *subname);
static void ReportSlotConnectionError(List *rstates, Oid subid, char *slotname, char *err);
@@ -496,6 +497,7 @@ CreateSubscription(ParseState *pstate, CreateSubscriptionStmt *stmt,
char *err;
WalReceiverConn *wrconn;
List *tables;
+ List *sequences;
ListCell *lc;
char table_state;
@@ -534,6 +536,26 @@ CreateSubscription(ParseState *pstate, CreateSubscriptionStmt *stmt,
InvalidXLogRecPtr);
}
+ /*
+ * Get the sequence list from publisher and build local sequence
+ * status info.
+ */
+ sequences = fetch_sequence_list(wrconn, publications);
+ foreach(lc, sequences)
+ {
+ RangeVar *rv = (RangeVar *) lfirst(lc);
+ Oid relid;
+
+ relid = RangeVarGetRelid(rv, AccessShareLock, false);
+
+ /* Check for supported relkind. */
+ CheckSubscriptionRelkind(get_rel_relkind(relid),
+ rv->schemaname, rv->relname);
+
+ AddSubscriptionRelState(subid, relid, table_state,
+ InvalidXLogRecPtr);
+ }
+
/*
* If requested, create permanent slot for the subscription. We
* won't use the initial snapshot for anything, so no need to
@@ -706,6 +728,10 @@ AlterSubscription_refresh(Subscription *sub, bool copy_data)
{
Oid relid = subrel_local_oids[off];
+ /* XXX ignore sequences - maybe do this in GetSubscriptionRelations? */
+ if (get_rel_relkind(relid) == RELKIND_SEQUENCE)
+ continue;
+
if (!bsearch(&relid, pubrel_local_oids,
list_length(pubrel_names), sizeof(Oid), oid_cmp))
{
@@ -797,6 +823,183 @@ AlterSubscription_refresh(Subscription *sub, bool copy_data)
ReplicationSlotDropAtPubNode(wrconn, syncslotname, true);
}
}
+
+ /*
+ * XXX now do the same thing for sequences, maybe before the preceding
+ * block, or earlier?
+ */
+
+ /* Get the table list from publisher. */
+ pubrel_names = fetch_sequence_list(wrconn, sub->publications);
+
+ /* Get local table list. */
+ subrel_states = GetSubscriptionRelations(sub->oid);
+
+ /*
+ * Build qsorted array of local table oids for faster lookup. This can
+ * potentially contain all tables in the database so speed of lookup
+ * is important.
+ */
+ subrel_local_oids = palloc(list_length(subrel_states) * sizeof(Oid));
+ off = 0;
+ foreach(lc, subrel_states)
+ {
+ SubscriptionRelState *relstate = (SubscriptionRelState *) lfirst(lc);
+
+ subrel_local_oids[off++] = relstate->relid;
+ }
+ qsort(subrel_local_oids, list_length(subrel_states),
+ sizeof(Oid), oid_cmp);
+
+ /*
+ * Rels that we want to remove from subscription and drop any slots
+ * and origins corresponding to them.
+ */
+ sub_remove_rels = palloc(list_length(subrel_states) * sizeof(SubRemoveRels));
+
+ /*
+ * Walk over the remote tables and try to match them to locally known
+ * tables. If the table is not known locally create a new state for
+ * it.
+ *
+ * Also builds array of local oids of remote tables for the next step.
+ */
+ off = 0;
+ pubrel_local_oids = palloc(list_length(pubrel_names) * sizeof(Oid));
+
+ foreach(lc, pubrel_names)
+ {
+ RangeVar *rv = (RangeVar *) lfirst(lc);
+ Oid relid;
+
+ relid = RangeVarGetRelid(rv, AccessShareLock, false);
+
+ /* Check for supported relkind. */
+ CheckSubscriptionRelkind(get_rel_relkind(relid),
+ rv->schemaname, rv->relname);
+
+ pubrel_local_oids[off++] = relid;
+
+ if (!bsearch(&relid, subrel_local_oids,
+ list_length(subrel_states), sizeof(Oid), oid_cmp))
+ {
+ AddSubscriptionRelState(sub->oid, relid,
+ copy_data ? SUBREL_STATE_INIT : SUBREL_STATE_READY,
+ InvalidXLogRecPtr);
+ ereport(DEBUG1,
+ (errmsg_internal("table \"%s.%s\" added to subscription \"%s\"",
+ rv->schemaname, rv->relname, sub->name)));
+ }
+ }
+
+ /*
+ * Next remove state for tables we should not care about anymore using
+ * the data we collected above
+ */
+ qsort(pubrel_local_oids, list_length(pubrel_names),
+ sizeof(Oid), oid_cmp);
+
+ remove_rel_len = 0;
+ for (off = 0; off < list_length(subrel_states); off++)
+ {
+ Oid relid = subrel_local_oids[off];
+
+ /* XXX ignore non-sequences - maybe do this in GetSubscriptionRelations? */
+ if (get_rel_relkind(relid) != RELKIND_SEQUENCE)
+ continue;
+
+ if (!bsearch(&relid, pubrel_local_oids,
+ list_length(pubrel_names), sizeof(Oid), oid_cmp))
+ {
+ char state;
+ XLogRecPtr statelsn;
+
+ /*
+ * Lock pg_subscription_rel with AccessExclusiveLock to
+ * prevent any race conditions with the apply worker
+ * re-launching workers at the same time this code is trying
+ * to remove those tables.
+ *
+ * Even if new worker for this particular rel is restarted it
+ * won't be able to make any progress as we hold exclusive
+ * lock on subscription_rel till the transaction end. It will
+ * simply exit as there is no corresponding rel entry.
+ *
+ * This locking also ensures that the state of rels won't
+ * change till we are done with this refresh operation.
+ */
+ if (!rel)
+ rel = table_open(SubscriptionRelRelationId, AccessExclusiveLock);
+
+ /* Last known rel state. */
+ state = GetSubscriptionRelState(sub->oid, relid, &statelsn);
+
+ sub_remove_rels[remove_rel_len].relid = relid;
+ sub_remove_rels[remove_rel_len++].state = state;
+
+ RemoveSubscriptionRel(sub->oid, relid);
+
+ logicalrep_worker_stop(sub->oid, relid);
+
+ /*
+ * For READY state, we would have already dropped the
+ * tablesync origin.
+ */
+ if (state != SUBREL_STATE_READY)
+ {
+ char originname[NAMEDATALEN];
+
+ /*
+ * Drop the tablesync's origin tracking if exists.
+ *
+ * It is possible that the origin is not yet created for
+ * tablesync worker, this can happen for the states before
+ * SUBREL_STATE_FINISHEDCOPY. The apply worker can also
+ * concurrently try to drop the origin and by this time
+ * the origin might be already removed. For these reasons,
+ * passing missing_ok = true.
+ */
+ ReplicationOriginNameForTablesync(sub->oid, relid, originname,
+ sizeof(originname));
+ replorigin_drop_by_name(originname, true, false);
+ }
+
+ ereport(DEBUG1,
+ (errmsg_internal("table \"%s.%s\" removed from subscription \"%s\"",
+ get_namespace_name(get_rel_namespace(relid)),
+ get_rel_name(relid),
+ sub->name)));
+ }
+ }
+
+ /*
+ * Drop the tablesync slots associated with removed tables. This has
+ * to be at the end because otherwise if there is an error while doing
+ * the database operations we won't be able to rollback dropped slots.
+ */
+ for (off = 0; off < remove_rel_len; off++)
+ {
+ if (sub_remove_rels[off].state != SUBREL_STATE_READY &&
+ sub_remove_rels[off].state != SUBREL_STATE_SYNCDONE)
+ {
+ char syncslotname[NAMEDATALEN] = {0};
+
+ /*
+ * For READY/SYNCDONE states we know the tablesync slot has
+ * already been dropped by the tablesync worker.
+ *
+ * For other states, there is no certainty, maybe the slot
+ * does not exist yet. Also, if we fail after removing some of
+ * the slots, next time, it will again try to drop already
+ * dropped slots and fail. For these reasons, we allow
+ * missing_ok = true for the drop.
+ */
+ ReplicationSlotNameForTablesync(sub->oid, sub_remove_rels[off].relid,
+ syncslotname, sizeof(syncslotname));
+ ReplicationSlotDropAtPubNode(wrconn, syncslotname, true);
+ }
+ }
+
}
PG_FINALLY();
{
@@ -1614,6 +1817,75 @@ fetch_table_list(WalReceiverConn *wrconn, List *publications)
return tablelist;
}
+/*
+ * Get the list of sequences which belong to specified publications on the
+ * publisher connection.
+ */
+static List *
+fetch_sequence_list(WalReceiverConn *wrconn, List *publications)
+{
+ WalRcvExecResult *res;
+ StringInfoData cmd;
+ TupleTableSlot *slot;
+ Oid tableRow[2] = {TEXTOID, TEXTOID};
+ ListCell *lc;
+ bool first;
+ List *tablelist = NIL;
+
+ Assert(list_length(publications) > 0);
+
+ initStringInfo(&cmd);
+ appendStringInfoString(&cmd, "SELECT DISTINCT s.schemaname, s.sequencename\n"
+ " FROM pg_catalog.pg_publication_sequences s\n"
+ " WHERE s.pubname IN (");
+ first = true;
+ foreach(lc, publications)
+ {
+ char *pubname = strVal(lfirst(lc));
+
+ if (first)
+ first = false;
+ else
+ appendStringInfoString(&cmd, ", ");
+
+ appendStringInfoString(&cmd, quote_literal_cstr(pubname));
+ }
+ appendStringInfoChar(&cmd, ')');
+
+ res = walrcv_exec(wrconn, cmd.data, 2, tableRow);
+ pfree(cmd.data);
+
+ if (res->status != WALRCV_OK_TUPLES)
+ ereport(ERROR,
+ (errmsg("could not receive list of replicated tables from the publisher: %s",
+ res->err)));
+
+ /* Process tables. */
+ slot = MakeSingleTupleTableSlot(res->tupledesc, &TTSOpsMinimalTuple);
+ while (tuplestore_gettupleslot(res->tuplestore, true, false, slot))
+ {
+ char *nspname;
+ char *relname;
+ bool isnull;
+ RangeVar *rv;
+
+ nspname = TextDatumGetCString(slot_getattr(slot, 1, &isnull));
+ Assert(!isnull);
+ relname = TextDatumGetCString(slot_getattr(slot, 2, &isnull));
+ Assert(!isnull);
+
+ rv = makeRangeVar(nspname, relname, -1);
+ tablelist = lappend(tablelist, rv);
+
+ ExecClearTuple(slot);
+ }
+ ExecDropSingleTupleTableSlot(slot);
+
+ walrcv_clear_result(res);
+
+ return tablelist;
+}
+
/*
* This is to report the connection failure while dropping replication slots.
* Here, we report the WARNING for all tablesync slots so that user can drop
diff --git a/src/backend/executor/execReplication.c b/src/backend/executor/execReplication.c
index 574d7d27fd9..6f0d42d5519 100644
--- a/src/backend/executor/execReplication.c
+++ b/src/backend/executor/execReplication.c
@@ -608,7 +608,7 @@ void
CheckSubscriptionRelkind(char relkind, const char *nspname,
const char *relname)
{
- if (relkind != RELKIND_RELATION && relkind != RELKIND_PARTITIONED_TABLE)
+ if (relkind != RELKIND_RELATION && relkind != RELKIND_PARTITIONED_TABLE && relkind != RELKIND_SEQUENCE)
ereport(ERROR,
(errcode(ERRCODE_WRONG_OBJECT_TYPE),
errmsg("cannot use relation \"%s.%s\" as logical replication target",
diff --git a/src/backend/nodes/copyfuncs.c b/src/backend/nodes/copyfuncs.c
index df0b7478838..6a5718572f9 100644
--- a/src/backend/nodes/copyfuncs.c
+++ b/src/backend/nodes/copyfuncs.c
@@ -4846,6 +4846,7 @@ _copyCreatePublicationStmt(const CreatePublicationStmt *from)
COPY_NODE_FIELD(options);
COPY_NODE_FIELD(pubobjects);
COPY_SCALAR_FIELD(for_all_tables);
+ COPY_SCALAR_FIELD(for_all_sequences);
return newnode;
}
diff --git a/src/backend/nodes/equalfuncs.c b/src/backend/nodes/equalfuncs.c
index cb7ddd463cb..5dccc0d0adc 100644
--- a/src/backend/nodes/equalfuncs.c
+++ b/src/backend/nodes/equalfuncs.c
@@ -2336,6 +2336,7 @@ _equalAlterPublicationStmt(const AlterPublicationStmt *a,
COMPARE_NODE_FIELD(options);
COMPARE_NODE_FIELD(pubobjects);
COMPARE_SCALAR_FIELD(for_all_tables);
+ COMPARE_SCALAR_FIELD(for_all_sequences);
COMPARE_SCALAR_FIELD(action);
return true;
diff --git a/src/backend/parser/gram.y b/src/backend/parser/gram.y
index 3d4dd43e47b..f037c17985b 100644
--- a/src/backend/parser/gram.y
+++ b/src/backend/parser/gram.y
@@ -9762,6 +9762,26 @@ PublicationObjSpec:
$$->pubobjtype = PUBLICATIONOBJ_TABLE_IN_CUR_SCHEMA;
$$->location = @5;
}
+ | SEQUENCE relation_expr
+ {
+ $$ = makeNode(PublicationObjSpec);
+ $$->pubobjtype = PUBLICATIONOBJ_SEQUENCE;
+ $$->pubtable = makeNode(PublicationTable);
+ $$->pubtable->relation = $2;
+ }
+ | ALL SEQUENCE IN_P SCHEMA ColId
+ {
+ $$ = makeNode(PublicationObjSpec);
+ $$->pubobjtype = PUBLICATIONOBJ_SEQUENCE_IN_SCHEMA;
+ $$->name = $5;
+ $$->location = @5;
+ }
+ | ALL SEQUENCES IN_P SCHEMA CURRENT_SCHEMA
+ {
+ $$ = makeNode(PublicationObjSpec);
+ $$->pubobjtype = PUBLICATIONOBJ_SEQUENCE_IN_CUR_SCHEMA;
+ $$->location = @5;
+ }
| ColId
{
$$ = makeNode(PublicationObjSpec);
@@ -10097,6 +10117,12 @@ UnlistenStmt:
}
;
+/*
+ * FIXME
+ *
+ * opt_publication_for_sequences and publication_for_sequences should be
+ * copies for sequences
+ */
/*****************************************************************************
*
@@ -10105,6 +10131,12 @@ UnlistenStmt:
* BEGIN / COMMIT / ROLLBACK
* (also older versions END / ABORT)
*
+ * ALTER PUBLICATION name ADD SEQUENCE sequence [, sequence2]
+ *
+ * ALTER PUBLICATION name DROP SEQUENCE sequence [, sequence2]
+ *
+ * ALTER PUBLICATION name SET SEQUENCE sequence [, sequence2]
+ *
*****************************************************************************/
TransactionStmt:
diff --git a/src/backend/replication/logical/proto.c b/src/backend/replication/logical/proto.c
index 9f5bf4b639f..79c6f626388 100644
--- a/src/backend/replication/logical/proto.c
+++ b/src/backend/replication/logical/proto.c
@@ -647,6 +647,56 @@ logicalrep_write_message(StringInfo out, TransactionId xid, XLogRecPtr lsn,
pq_sendbytes(out, message, sz);
}
+/*
+ * Write SEQUENCE to stream
+ */
+void
+logicalrep_write_sequence(StringInfo out, Relation rel, TransactionId xid,
+ XLogRecPtr lsn, bool transactional,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ uint8 flags = 0;
+ char *relname;
+
+ pq_sendbyte(out, LOGICAL_REP_MSG_SEQUENCE);
+
+ /* transaction ID (if not valid, we're not streaming) */
+ if (TransactionIdIsValid(xid))
+ pq_sendint32(out, xid);
+
+ pq_sendint8(out, flags);
+ pq_sendint64(out, lsn);
+
+ logicalrep_write_namespace(out, RelationGetNamespace(rel));
+ relname = RelationGetRelationName(rel);
+ pq_sendstring(out, relname);
+
+ pq_sendint8(out, transactional);
+ pq_sendint64(out, last_value);
+ pq_sendint64(out, log_cnt);
+ pq_sendint8(out, is_called);
+}
+
+/*
+ * Read SEQUENCE from the stream.
+ */
+void
+logicalrep_read_sequence(StringInfo in, LogicalRepSequence *seqdata)
+{
+ /* XXX skipping flags and lsn */
+ pq_getmsgint(in, 1);
+ pq_getmsgint64(in);
+
+ /* Read relation name from stream */
+ seqdata->nspname = pstrdup(logicalrep_read_namespace(in));
+ seqdata->seqname = pstrdup(pq_getmsgstring(in));
+
+ seqdata->transactional = pq_getmsgint(in, 1);
+ seqdata->last_value = pq_getmsgint64(in);
+ seqdata->log_cnt = pq_getmsgint64(in);
+ seqdata->is_called = pq_getmsgint(in, 1);
+}
+
/*
* Write relation description to the output stream.
*/
@@ -1203,6 +1253,8 @@ logicalrep_message_type(LogicalRepMsgType action)
return "STREAM ABORT";
case LOGICAL_REP_MSG_STREAM_PREPARE:
return "STREAM PREPARE";
+ case LOGICAL_REP_MSG_SEQUENCE:
+ return "SEQUENCE";
}
elog(ERROR, "invalid logical replication message type \"%c\"", action);
diff --git a/src/backend/replication/logical/tablesync.c b/src/backend/replication/logical/tablesync.c
index f07983a43cb..24369a1522e 100644
--- a/src/backend/replication/logical/tablesync.c
+++ b/src/backend/replication/logical/tablesync.c
@@ -100,6 +100,7 @@
#include "catalog/pg_subscription_rel.h"
#include "catalog/pg_type.h"
#include "commands/copy.h"
+#include "commands/sequence.h"
#include "miscadmin.h"
#include "parser/parse_relation.h"
#include "pgstat.h"
@@ -357,6 +358,12 @@ process_syncing_tables_for_sync(XLogRecPtr current_lsn)
*
* If the synchronization position is reached (SYNCDONE), then the table can
* be marked as READY and is no longer tracked.
+ *
+ * XXX This needs to handle sequences too - after AlterSubscription_refresh
+ * starts caring about sequences, GetSubscriptionNotReadyRelations won't
+ * return just tables, and we'll have to sync them here. Not sure it's worth
+ * creating a new "sync" worker per sequence, maybe we should just sync them
+ * in the current process (it's pretty light-weight).
*/
static void
process_syncing_tables_for_apply(XLogRecPtr current_lsn)
@@ -871,6 +878,99 @@ copy_table(Relation rel)
logicalrep_rel_close(relmapentry, NoLock);
}
+
+
+/*
+ * FIXME add comment
+ */
+static void
+fetch_sequence_data(char *nspname, char *relname,
+ int64 *last_value, int64 *log_cnt, bool *is_called)
+{
+ WalRcvExecResult *res;
+ StringInfoData cmd;
+ TupleTableSlot *slot;
+ Oid tableRow[3] = {INT8OID, INT8OID, BOOLOID};
+
+ initStringInfo(&cmd);
+ appendStringInfo(&cmd, "SELECT last_value, log_cnt, is_called\n"
+ " FROM %s", quote_qualified_identifier(nspname, relname));
+
+ res = walrcv_exec(LogRepWorkerWalRcvConn, cmd.data, 3, tableRow);
+ pfree(cmd.data);
+
+ if (res->status != WALRCV_OK_TUPLES)
+ ereport(ERROR,
+ (errmsg("could not receive list of replicated tables from the publisher: %s",
+ res->err)));
+
+ /* Process the sequence. */
+ slot = MakeSingleTupleTableSlot(res->tupledesc, &TTSOpsMinimalTuple);
+ while (tuplestore_gettupleslot(res->tuplestore, true, false, slot))
+ {
+ bool isnull;
+
+ *last_value = DatumGetInt64(slot_getattr(slot, 1, &isnull));
+ Assert(!isnull);
+
+ *log_cnt = DatumGetInt64(slot_getattr(slot, 2, &isnull));
+ Assert(!isnull);
+
+ *is_called = DatumGetBool(slot_getattr(slot, 3, &isnull));
+ Assert(!isnull);
+
+ ExecClearTuple(slot);
+ }
+ ExecDropSingleTupleTableSlot(slot);
+
+ walrcv_clear_result(res);
+}
+
+/*
+ * Copy existing data of a sequence from publisher.
+ *
+ * Caller is responsible for locking the local relation.
+ */
+static void
+copy_sequence(Relation rel)
+{
+ LogicalRepRelMapEntry *relmapentry;
+ LogicalRepRelation lrel;
+ StringInfoData cmd;
+ int64 last_value = 0,
+ log_cnt = 0;
+ bool is_called = 0;
+
+ /* Get the publisher relation info. */
+ fetch_remote_table_info(get_namespace_name(RelationGetNamespace(rel)),
+ RelationGetRelationName(rel), &lrel);
+
+ /* Put the relation into relmap. */
+ logicalrep_relmap_update(&lrel);
+
+ /* Map the publisher relation to local one. */
+ relmapentry = logicalrep_rel_open(lrel.remoteid, NoLock);
+ Assert(rel == relmapentry->localrel);
+
+ /* Start copy on the publisher. */
+ initStringInfo(&cmd);
+
+ Assert(lrel.relkind == RELKIND_SEQUENCE);
+
+ fetch_sequence_data(lrel.nspname, lrel.relname, &last_value, &log_cnt, &is_called);
+
+ elog(WARNING, "sequence %s info last_value %ld log_cnt %ld is_called %d",
+ quote_qualified_identifier(lrel.nspname, lrel.relname),
+ last_value, log_cnt, is_called);
+
+ ResetSequence2(RelationGetRelid(rel), last_value, log_cnt, is_called);
+
+ logicalrep_rel_close(relmapentry, NoLock);
+}
+
+
+
+
/*
* Determine the tablesync slot name.
*
@@ -1106,10 +1206,20 @@ LogicalRepSyncTableStart(XLogRecPtr *origin_startpos)
originname)));
}
- /* Now do the initial data copy */
- PushActiveSnapshot(GetTransactionSnapshot());
- copy_table(rel);
- PopActiveSnapshot();
+ if (get_rel_relkind(RelationGetRelid(rel)) == RELKIND_SEQUENCE)
+ {
+ /* Now do the initial sequence copy */
+ PushActiveSnapshot(GetTransactionSnapshot());
+ copy_sequence(rel);
+ PopActiveSnapshot();
+ }
+ else
+ {
+ /* Now do the initial data copy */
+ PushActiveSnapshot(GetTransactionSnapshot());
+ copy_table(rel);
+ PopActiveSnapshot();
+ }
res = walrcv_exec(LogRepWorkerWalRcvConn, "COMMIT", 0, NULL);
if (res->status != WALRCV_OK_COMMAND)
diff --git a/src/backend/replication/logical/worker.c b/src/backend/replication/logical/worker.c
index 2e79302a48a..089d989dd73 100644
--- a/src/backend/replication/logical/worker.c
+++ b/src/backend/replication/logical/worker.c
@@ -144,6 +144,7 @@
#include "catalog/pg_tablespace.h"
#include "commands/tablecmds.h"
#include "commands/tablespace.h"
+#include "commands/sequence.h"
#include "commands/trigger.h"
#include "executor/executor.h"
#include "executor/execPartition.h"
@@ -1091,6 +1092,61 @@ apply_handle_origin(StringInfo s)
errmsg_internal("ORIGIN message sent out of order")));
}
+/*
+ * Handle SEQUENCE message.
+ */
+static void
+apply_handle_sequence(StringInfo s)
+{
+ LogicalRepSequence seq;
+ Oid relid;
+
+ if (handle_streamed_transaction(LOGICAL_REP_MSG_SEQUENCE, s))
+ return;
+
+ logicalrep_read_sequence(s, &seq);
+
+ /*
+ * Non-transactional sequence updates should not be part of a remote
+ * transaction. There should not be any running transaction.
+ */
+ Assert((!seq.transactional) || in_remote_transaction);
+ Assert(!(!seq.transactional && in_remote_transaction));
+ Assert(!(!seq.transactional && IsTransactionState()));
+
+ /*
+ * Make sure we're in a transaction (needed by ResetSequence2). For
+ * non-transactional updates we're guaranteed to start a new one,
+ * and we'll commit it at the end.
+ */
+ if (!IsTransactionState())
+ {
+ StartTransactionCommand();
+ maybe_reread_subscription();
+ }
+
+ relid = RangeVarGetRelid(makeRangeVar(seq.nspname,
+ seq.seqname, -1),
+ RowExclusiveLock, false);
+
+ /* lock the sequence in AccessExclusiveLock, as expected by ResetSequence2 */
+ elog(WARNING, "locking sequence %d in exclusive mode", relid);
+ LockRelationOid(relid, AccessExclusiveLock);
+
+ elog(WARNING, "applying sequence %s.%s transactional %d last_value %ld log_cnt %ld is_called %d",
+ seq.nspname, seq.seqname, seq.transactional, seq.last_value, seq.log_cnt, seq.is_called);
+
+ /* apply the sequence change */
+ ResetSequence2(relid, seq.last_value, seq.log_cnt, seq.is_called);
+
+ /*
+ * Commit the per-stream transaction (we only do this when not in
+ * remote transaction, i.e. for non-transactional sequence updates.
+ */
+ if (!in_remote_transaction)
+ CommitTransactionCommand();
+}
+
/*
* Handle STREAM START message.
*/
@@ -2374,6 +2430,10 @@ apply_dispatch(StringInfo s)
*/
break;
+ case LOGICAL_REP_MSG_SEQUENCE:
+ apply_handle_sequence(s);
+ return;
+
case LOGICAL_REP_MSG_STREAM_START:
apply_handle_stream_start(s);
break;
diff --git a/src/backend/replication/pgoutput/pgoutput.c b/src/backend/replication/pgoutput/pgoutput.c
index 6f6a203dea7..9ab8b1d4ec4 100644
--- a/src/backend/replication/pgoutput/pgoutput.c
+++ b/src/backend/replication/pgoutput/pgoutput.c
@@ -49,6 +49,10 @@ static void pgoutput_message(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn, XLogRecPtr message_lsn,
bool transactional, const char *prefix,
Size sz, const char *message);
+static void pgoutput_sequence(LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn, XLogRecPtr sequence_lsn,
+ Relation rel, bool transactional,
+ int64 last_value, int64 log_cnt, bool is_called);
static bool pgoutput_origin_filter(LogicalDecodingContext *ctx,
RepOriginId origin_id);
static void pgoutput_begin_prepare_txn(LogicalDecodingContext *ctx,
@@ -159,6 +163,7 @@ _PG_output_plugin_init(OutputPluginCallbacks *cb)
cb->change_cb = pgoutput_change;
cb->truncate_cb = pgoutput_truncate;
cb->message_cb = pgoutput_message;
+ cb->sequence_cb = pgoutput_sequence;
cb->commit_cb = pgoutput_commit_txn;
cb->begin_prepare_cb = pgoutput_begin_prepare_txn;
@@ -175,6 +180,7 @@ _PG_output_plugin_init(OutputPluginCallbacks *cb)
cb->stream_commit_cb = pgoutput_stream_commit;
cb->stream_change_cb = pgoutput_change;
cb->stream_message_cb = pgoutput_message;
+ cb->stream_sequence_cb = pgoutput_sequence;
cb->stream_truncate_cb = pgoutput_truncate;
/* transaction streaming - two-phase commit */
cb->stream_prepare_cb = pgoutput_stream_prepare_txn;
@@ -188,6 +194,7 @@ parse_output_parameters(List *options, PGOutputData *data)
bool publication_names_given = false;
bool binary_option_given = false;
bool messages_option_given = false;
+ bool sequences_option_given = false;
bool streaming_given = false;
bool two_phase_option_given = false;
@@ -195,6 +202,7 @@ parse_output_parameters(List *options, PGOutputData *data)
data->streaming = false;
data->messages = false;
data->two_phase = false;
+ data->sequences = true;
foreach(lc, options)
{
@@ -260,6 +268,16 @@ parse_output_parameters(List *options, PGOutputData *data)
data->messages = defGetBoolean(defel);
}
+ else if (strcmp(defel->defname, "sequences") == 0)
+ {
+ if (sequences_option_given)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("conflicting or redundant options")));
+ sequences_option_given = true;
+
+ data->sequences = defGetBoolean(defel);
+ }
else if (strcmp(defel->defname, "streaming") == 0)
{
if (streaming_given)
@@ -856,6 +874,51 @@ pgoutput_message(LogicalDecodingContext *ctx, ReorderBufferTXN *txn,
OutputPluginWrite(ctx, true);
}
+static void
+pgoutput_sequence(LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn, XLogRecPtr sequence_lsn,
+ Relation rel, bool transactional,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ PGOutputData *data = (PGOutputData *) ctx->output_plugin_private;
+ TransactionId xid = InvalidTransactionId;
+ RelationSyncEntry *relentry;
+
+ if (!data->sequences)
+ return;
+
+ if (!is_publishable_relation(rel))
+ return;
+
+ /*
+ * Remember the xid for the message in streaming mode. See
+ * pgoutput_change.
+ */
+ if (in_streaming)
+ xid = txn->xid;
+
+ relentry = get_rel_sync_entry(data, RelationGetRelid(rel));
+
+ /*
+ * First check the sequence filter.
+ *
+ * We handle just REORDER_BUFFER_CHANGE_SEQUENCE here.
+ */
+ if (!relentry->pubactions.pubsequence)
+ return;
+
+ OutputPluginPrepareWrite(ctx, true);
+ logicalrep_write_sequence(ctx->out,
+ rel,
+ xid,
+ sequence_lsn,
+ transactional,
+ last_value,
+ log_cnt,
+ is_called);
+ OutputPluginWrite(ctx, true);
+}
+
/*
* Currently we always forward.
*/
@@ -1140,7 +1203,8 @@ get_rel_sync_entry(PGOutputData *data, Oid relid)
entry->streamed_txns = NIL;
entry->replicate_valid = false;
entry->pubactions.pubinsert = entry->pubactions.pubupdate =
- entry->pubactions.pubdelete = entry->pubactions.pubtruncate = false;
+ entry->pubactions.pubdelete = entry->pubactions.pubtruncate =
+ entry->pubactions.pubsequence = false;
entry->publish_as_relid = InvalidOid;
entry->map = NULL; /* will be set by maybe_send_schema() if
* needed */
@@ -1160,6 +1224,7 @@ get_rel_sync_entry(PGOutputData *data, Oid relid)
List *schemaPubids = GetSchemaPublications(schemaId);
ListCell *lc;
Oid publish_as_relid = relid;
+ bool is_sequence = (get_rel_relkind(relid) == RELKIND_SEQUENCE);
/* Reload publications if needed before use. */
if (!publications_valid)
@@ -1183,12 +1248,23 @@ get_rel_sync_entry(PGOutputData *data, Oid relid)
Publication *pub = lfirst(lc);
bool publish = false;
- if (pub->alltables)
+ if (pub->alltables && (!is_sequence))
{
publish = true;
if (pub->pubviaroot && am_partition)
publish_as_relid = llast_oid(get_partition_ancestors(relid));
}
+ else if (pub->allsequences && is_sequence)
+ {
+ publish = true;
+ }
+
+ /* if a sequence, just cross-check the list of publications */
+ if (!publish && is_sequence)
+ {
+ if (list_member_oid(pubids, pub->oid))
+ publish = true;
+ }
if (!publish)
{
@@ -1243,10 +1319,12 @@ get_rel_sync_entry(PGOutputData *data, Oid relid)
entry->pubactions.pubupdate |= pub->pubactions.pubupdate;
entry->pubactions.pubdelete |= pub->pubactions.pubdelete;
entry->pubactions.pubtruncate |= pub->pubactions.pubtruncate;
+ entry->pubactions.pubsequence |= pub->pubactions.pubsequence;
}
if (entry->pubactions.pubinsert && entry->pubactions.pubupdate &&
- entry->pubactions.pubdelete && entry->pubactions.pubtruncate)
+ entry->pubactions.pubdelete && entry->pubactions.pubtruncate &&
+ entry->pubactions.pubsequence)
break;
}
@@ -1391,6 +1469,7 @@ rel_sync_cache_publication_cb(Datum arg, int cacheid, uint32 hashvalue)
entry->pubactions.pubupdate = false;
entry->pubactions.pubdelete = false;
entry->pubactions.pubtruncate = false;
+ entry->pubactions.pubsequence = false;
}
}
diff --git a/src/backend/utils/cache/relcache.c b/src/backend/utils/cache/relcache.c
index 105d8d4601c..1f8634b50a1 100644
--- a/src/backend/utils/cache/relcache.c
+++ b/src/backend/utils/cache/relcache.c
@@ -5585,6 +5585,7 @@ GetRelationPublicationActions(Relation relation)
pubactions->pubupdate |= pubform->pubupdate;
pubactions->pubdelete |= pubform->pubdelete;
pubactions->pubtruncate |= pubform->pubtruncate;
+ pubactions->pubsequence |= pubform->pubsequence;
ReleaseSysCache(tup);
@@ -5593,7 +5594,8 @@ GetRelationPublicationActions(Relation relation)
* other publications.
*/
if (pubactions->pubinsert && pubactions->pubupdate &&
- pubactions->pubdelete && pubactions->pubtruncate)
+ pubactions->pubdelete && pubactions->pubtruncate &&
+ pubactions->pubsequence)
break;
}
diff --git a/src/bin/psql/tab-complete.c b/src/bin/psql/tab-complete.c
index 2f412ca3db3..e30bf7b1b55 100644
--- a/src/bin/psql/tab-complete.c
+++ b/src/bin/psql/tab-complete.c
@@ -1647,13 +1647,13 @@ psql_completion(const char *text, int start, int end)
COMPLETE_WITH("ADD", "DROP", "OWNER TO", "RENAME TO", "SET");
/* ALTER PUBLICATION <name> ADD */
else if (Matches("ALTER", "PUBLICATION", MatchAny, "ADD"))
- COMPLETE_WITH("ALL TABLES IN SCHEMA", "TABLE");
+ COMPLETE_WITH("ALL TABLES IN SCHEMA", "TABLE|SEQUENCE");
/* ALTER PUBLICATION <name> DROP */
else if (Matches("ALTER", "PUBLICATION", MatchAny, "DROP"))
- COMPLETE_WITH("ALL TABLES IN SCHEMA", "TABLE");
+ COMPLETE_WITH("ALL TABLES IN SCHEMA", "TABLE|SEQUENCE");
/* ALTER PUBLICATION <name> SET */
else if (Matches("ALTER", "PUBLICATION", MatchAny, "SET"))
- COMPLETE_WITH("(", "ALL TABLES IN SCHEMA", "TABLE");
+ COMPLETE_WITH("(", "ALL TABLES IN SCHEMA", "TABLE|SEQUENCE");
else if (Matches("ALTER", "PUBLICATION", MatchAny, "ADD|DROP|SET", "ALL", "TABLES", "IN", "SCHEMA"))
COMPLETE_WITH_QUERY(Query_for_list_of_schemas
" AND nspname != 'pg_catalog' "
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 79d787cd26a..cd929f6e0e2 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -11531,6 +11531,11 @@
provolatile => 's', prorettype => 'oid', proargtypes => 'text',
proallargtypes => '{text,oid}', proargmodes => '{i,o}',
proargnames => '{pubname,relid}', prosrc => 'pg_get_publication_tables' },
+{ oid => '8000', descr => 'get OIDs of sequences in a publication',
+ proname => 'pg_get_publication_sequences', prorows => '1000', proretset => 't',
+ provolatile => 's', prorettype => 'oid', proargtypes => 'text',
+ proallargtypes => '{text,oid}', proargmodes => '{i,o}',
+ proargnames => '{pubname,relid}', prosrc => 'pg_get_publication_sequences' },
{ oid => '6121',
descr => 'returns whether a relation can be part of a publication',
proname => 'pg_relation_is_publishable', provolatile => 's',
diff --git a/src/include/catalog/pg_publication.h b/src/include/catalog/pg_publication.h
index 902f2f2f0da..b44bb0e5e06 100644
--- a/src/include/catalog/pg_publication.h
+++ b/src/include/catalog/pg_publication.h
@@ -40,6 +40,12 @@ CATALOG(pg_publication,6104,PublicationRelationId)
*/
bool puballtables;
+ /*
+ * indicates that this is special publication which should encompass all
+ * sequences in the database (except for the unlogged and temp ones)
+ */
+ bool puballsequences;
+
/* true if inserts are published */
bool pubinsert;
@@ -52,6 +58,9 @@ CATALOG(pg_publication,6104,PublicationRelationId)
/* true if truncates are published */
bool pubtruncate;
+ /* true if sequences are published */
+ bool pubsequence;
+
/* true if partition changes are published using root schema */
bool pubviaroot;
} FormData_pg_publication;
@@ -72,6 +81,7 @@ typedef struct PublicationActions
bool pubupdate;
bool pubdelete;
bool pubtruncate;
+ bool pubsequence;
} PublicationActions;
typedef struct Publication
@@ -79,6 +89,7 @@ typedef struct Publication
Oid oid;
char *name;
bool alltables;
+ bool allsequences;
bool pubviaroot;
PublicationActions pubactions;
} Publication;
@@ -121,6 +132,9 @@ extern List *GetPubPartitionOptionRelations(List *result,
PublicationPartOpt pub_partopt,
Oid relid);
+extern List *GetAllSequencesPublicationRelations(void);
+extern List *GetPublicationSequenceRelations(Oid pubid);
+
extern bool is_publishable_relation(Relation rel);
extern bool is_schema_publication(Oid pubid);
extern ObjectAddress publication_add_relation(Oid pubid, PublicationRelInfo *targetrel,
diff --git a/src/include/commands/sequence.h b/src/include/commands/sequence.h
index 5919fb90ee1..c28e8695cb8 100644
--- a/src/include/commands/sequence.h
+++ b/src/include/commands/sequence.h
@@ -60,6 +60,7 @@ extern ObjectAddress DefineSequence(ParseState *pstate, CreateSeqStmt *stmt);
extern ObjectAddress AlterSequence(ParseState *pstate, AlterSeqStmt *stmt);
extern void DeleteSequenceTuple(Oid relid);
extern void ResetSequence(Oid seq_relid);
+extern void ResetSequence2(Oid seq_relid, int64 last_value, int64 log_cnt, bool is_called);
extern void ResetSequenceCaches(void);
extern void seq_redo(XLogReaderState *rptr);
diff --git a/src/include/nodes/parsenodes.h b/src/include/nodes/parsenodes.h
index 4c5a8a39bf1..d32ba498a7c 100644
--- a/src/include/nodes/parsenodes.h
+++ b/src/include/nodes/parsenodes.h
@@ -3653,6 +3653,10 @@ typedef enum PublicationObjSpecType
PUBLICATIONOBJ_TABLE_IN_SCHEMA, /* Tables in schema type */
PUBLICATIONOBJ_TABLE_IN_CUR_SCHEMA, /* Get the first element from
* search_path */
+ PUBLICATIONOBJ_SEQUENCE, /* Sequence type */
+ PUBLICATIONOBJ_SEQUENCE_IN_SCHEMA, /* Sequences in schema type */
+ PUBLICATIONOBJ_SEQUENCE_IN_CUR_SCHEMA, /* Get the first element from
+ * search_path */
PUBLICATIONOBJ_CONTINUATION /* Continuation of previous type */
} PublicationObjSpecType;
@@ -3672,6 +3676,7 @@ typedef struct CreatePublicationStmt
List *options; /* List of DefElem nodes */
List *pubobjects; /* Optional list of publication objects */
bool for_all_tables; /* Special publication for all tables in db */
+ bool for_all_sequences; /* Special publication for all sequences in db */
} CreatePublicationStmt;
typedef struct AlterPublicationStmt
@@ -3688,6 +3693,7 @@ typedef struct AlterPublicationStmt
*/
List *pubobjects; /* Optional list of publication objects */
bool for_all_tables; /* Special publication for all tables in db */
+ bool for_all_sequences; /* Special publication for all sequences in db */
DefElemAction action; /* What action to perform with the
* tables/schemas */
} AlterPublicationStmt;
diff --git a/src/include/replication/logicalproto.h b/src/include/replication/logicalproto.h
index 83741dcf424..fc1782c125f 100644
--- a/src/include/replication/logicalproto.h
+++ b/src/include/replication/logicalproto.h
@@ -60,6 +60,7 @@ typedef enum LogicalRepMsgType
LOGICAL_REP_MSG_RELATION = 'R',
LOGICAL_REP_MSG_TYPE = 'Y',
LOGICAL_REP_MSG_MESSAGE = 'M',
+ LOGICAL_REP_MSG_SEQUENCE = 'X', /* FIXME change */
LOGICAL_REP_MSG_BEGIN_PREPARE = 'b',
LOGICAL_REP_MSG_PREPARE = 'P',
LOGICAL_REP_MSG_COMMIT_PREPARED = 'K',
@@ -117,6 +118,18 @@ typedef struct LogicalRepTyp
char *typname; /* name of the remote type */
} LogicalRepTyp;
+/* Sequence info */
+typedef struct LogicalRepSequence
+{
+ Oid remoteid; /* unique id of the remote sequence */
+ char *nspname; /* schema name of remote sequence */
+ char *seqname; /* name of the remote sequence */
+ bool transactional;
+ int64 last_value;
+ int64 log_cnt;
+ bool is_called;
+} LogicalRepSequence;
+
/* Transaction info */
typedef struct LogicalRepBeginData
{
@@ -227,6 +240,12 @@ extern List *logicalrep_read_truncate(StringInfo in,
bool *cascade, bool *restart_seqs);
extern void logicalrep_write_message(StringInfo out, TransactionId xid, XLogRecPtr lsn,
bool transactional, const char *prefix, Size sz, const char *message);
+extern void logicalrep_write_sequence(StringInfo out, Relation rel,
+ TransactionId xid, XLogRecPtr lsn,
+ bool transactional,
+ int64 last_value, int64 log_cnt,
+ bool is_called);
+extern void logicalrep_read_sequence(StringInfo in, LogicalRepSequence *seqdata);
extern void logicalrep_write_rel(StringInfo out, TransactionId xid,
Relation rel);
extern LogicalRepRelation *logicalrep_read_rel(StringInfo in);
diff --git a/src/include/replication/pgoutput.h b/src/include/replication/pgoutput.h
index 0dc460fb70a..97eb2a7ef15 100644
--- a/src/include/replication/pgoutput.h
+++ b/src/include/replication/pgoutput.h
@@ -28,6 +28,7 @@ typedef struct PGOutputData
bool streaming;
bool messages;
bool two_phase;
+ bool sequences;
} PGOutputData;
#endif /* PGOUTPUT_H */
diff --git a/src/test/regress/expected/rules.out b/src/test/regress/expected/rules.out
index b58b062b10d..6eb4f110038 100644
--- a/src/test/regress/expected/rules.out
+++ b/src/test/regress/expected/rules.out
@@ -1451,6 +1451,14 @@ pg_prepared_xacts| SELECT p.transaction,
FROM ((pg_prepared_xact() p(transaction, gid, prepared, ownerid, dbid)
LEFT JOIN pg_authid u ON ((p.ownerid = u.oid)))
LEFT JOIN pg_database d ON ((p.dbid = d.oid)));
+pg_publication_sequences| SELECT p.pubname,
+ n.nspname AS schemaname,
+ c.relname AS sequencename
+ FROM pg_publication p,
+ LATERAL pg_get_publication_sequences((p.pubname)::text) gpt(relid),
+ (pg_class c
+ JOIN pg_namespace n ON ((n.oid = c.relnamespace)))
+ WHERE (c.oid = gpt.relid);
pg_publication_tables| SELECT p.pubname,
n.nspname AS schemaname,
c.relname AS tablename
diff --git a/src/test/subscription/t/027_sequences.pl b/src/test/subscription/t/027_sequences.pl
new file mode 100644
index 00000000000..fd941619071
--- /dev/null
+++ b/src/test/subscription/t/027_sequences.pl
@@ -0,0 +1,196 @@
+
+# Copyright (c) 2021, PostgreSQL Global Development Group
+
+# This tests that sequences are replicated correctly by logical replication
+use strict;
+use warnings;
+use PostgreSQL::Test::Cluster;
+use PostgreSQL::Test::Utils;
+use Test::More tests => 6;
+
+# Initialize publisher node
+my $node_publisher = PostgreSQL::Test::Cluster->new('publisher');
+$node_publisher->init(allows_streaming => 'logical');
+$node_publisher->start;
+
+# Create subscriber node
+my $node_subscriber = PostgreSQL::Test::Cluster->new('subscriber');
+$node_subscriber->init(allows_streaming => 'logical');
+$node_subscriber->start;
+
+# Create some preexisting content on publisher
+my $ddl = qq(
+ CREATE SEQUENCE s;
+);
+
+# Setup structure on the publisher
+$node_publisher->safe_psql('postgres', $ddl);
+
+# Create some the same structure on subscriber, and an extra sequence that
+# we'll create on the publisher later
+$ddl = qq(
+ CREATE SEQUENCE s;
+ CREATE SEQUENCE s2;
+);
+
+$node_subscriber->safe_psql('postgres', $ddl);
+
+# Setup logical replication
+my $publisher_connstr = $node_publisher->connstr . ' dbname=postgres';
+$node_publisher->safe_psql('postgres',
+ "CREATE PUBLICATION seq_pub");
+
+$node_publisher->safe_psql('postgres',
+ "ALTER PUBLICATION seq_pub ADD SEQUENCE s");
+
+$node_subscriber->safe_psql('postgres',
+ "CREATE SUBSCRIPTION seq_sub CONNECTION '$publisher_connstr' PUBLICATION seq_pub WITH (slot_name = seq_sub_slot)"
+);
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Wait for initial sync to finish as well
+my $synced_query =
+ "SELECT count(1) = 0 FROM pg_subscription_rel WHERE srsubstate NOT IN ('s', 'r');";
+$node_subscriber->poll_query_until('postgres', $synced_query)
+ or die "Timed out while waiting for subscriber to synchronize data";
+
+# Insert initial test data
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ -- generate a number of values using the sequence
+ SELECT nextval('s') FROM generate_series(1,100);
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+my $result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s;
+));
+
+is( $result, '132|0|t',
+ 'check replicated sequence values on subscriber');
+
+
+# advance the sequence in a rolled-back transaction - should not be replicated
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ SELECT nextval('s') FROM generate_series(1,100);
+ ROLLBACK;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s;
+));
+
+is( $result, '231|0|t',
+ 'check replicated sequence values on subscriber');
+
+
+# create a new sequence and roll it back - should not be replicated, due to
+# the transactional behavior
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ CREATE SEQUENCE s2;
+ ALTER PUBLICATION seq_pub ADD SEQUENCE s2;
+ SELECT nextval('s2') FROM generate_series(1,100);
+ ROLLBACK;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s2;
+));
+
+is( $result, '1|0|f',
+ 'check replicated sequence values on subscriber');
+
+
+# create a new sequence, advance it in a rolled-back transaction, but commit
+# the create - the advance should be replicated nevertheless
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ CREATE SEQUENCE s2;
+ ALTER PUBLICATION seq_pub ADD SEQUENCE s2;
+ SAVEPOINT sp1;
+ SELECT nextval('s2') FROM generate_series(1,100);
+ ROLLBACK TO sp1;
+ COMMIT;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Wait for sync of the second sequence we just added to finish
+$synced_query =
+ "SELECT count(1) = 0 FROM pg_subscription_rel WHERE srsubstate NOT IN ('s', 'r');";
+$node_subscriber->poll_query_until('postgres', $synced_query)
+ or die "Timed out while waiting for subscriber to synchronize data";
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s2;
+));
+
+is( $result, '132|0|t',
+ 'check replicated sequence values on subscriber');
+
+
+# advance the new sequence in a transaction, and roll it back - in this case
+# it should not be replicated at commit
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ SELECT nextval('s2') FROM generate_series(1,100);
+ ROLLBACK;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s2;
+));
+
+is( $result, '132|0|t',
+ 'check replicated sequence values on subscriber');
+
+
+# advance the sequence in a subtransaction - the subtransaction gets rolled
+# back, but commit the main one - the changes should still be replicated
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ SAVEPOINT s1;
+ SELECT nextval('s2') FROM generate_series(1,100);
+ ROLLBACK TO s1;
+ COMMIT;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s2;
+));
+
+is( $result, '330|0|t',
+ 'check replicated sequence values on subscriber');
+
+
+$node_subscriber->stop('fast');
+$node_publisher->stop('fast');
--
2.31.1
[text/x-patch] 0002-Add-support-for-decoding-sequences-to-test_-20211214.patch (20.3K, ../../[email protected]/3-0002-Add-support-for-decoding-sequences-to-test_-20211214.patch)
download | inline diff:
From 71e95aa7ea701e305c2a95516b6b37c22e0b5e7b Mon Sep 17 00:00:00 2001
From: Tomas Vondra <[email protected]>
Date: Fri, 24 Sep 2021 00:42:04 +0200
Subject: [PATCH 2/3] Add support for decoding sequences to test_decoding
---
contrib/test_decoding/Makefile | 3 +-
contrib/test_decoding/expected/sequence.out | 327 ++++++++++++++++++++
contrib/test_decoding/sql/sequence.sql | 119 +++++++
contrib/test_decoding/test_decoding.c | 69 +++++
4 files changed, 517 insertions(+), 1 deletion(-)
create mode 100644 contrib/test_decoding/expected/sequence.out
create mode 100644 contrib/test_decoding/sql/sequence.sql
diff --git a/contrib/test_decoding/Makefile b/contrib/test_decoding/Makefile
index 9a31e0b8795..56ddc3abaeb 100644
--- a/contrib/test_decoding/Makefile
+++ b/contrib/test_decoding/Makefile
@@ -5,7 +5,8 @@ PGFILEDESC = "test_decoding - example of a logical decoding output plugin"
REGRESS = ddl xact rewrite toast permissions decoding_in_xact \
decoding_into_rel binary prepared replorigin time messages \
- spill slot truncate stream stats twophase twophase_stream
+ spill slot truncate stream stats twophase twophase_stream \
+ sequence
ISOLATION = mxact delayed_startup ondisk_startup concurrent_ddl_dml \
oldest_xmin snapshot_transfer subxact_without_top concurrent_stream \
twophase_snapshot
diff --git a/contrib/test_decoding/expected/sequence.out b/contrib/test_decoding/expected/sequence.out
new file mode 100644
index 00000000000..c3c2a18a35b
--- /dev/null
+++ b/contrib/test_decoding/expected/sequence.out
@@ -0,0 +1,327 @@
+-- predictability
+SET synchronous_commit = on;
+SELECT 'init' FROM pg_create_logical_replication_slot('regression_slot', 'test_decoding');
+ ?column?
+----------
+ init
+(1 row)
+
+CREATE SEQUENCE test_sequence;
+-- test the sequence changes by several nextval() calls
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 1
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 2
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 4
+(1 row)
+
+-- test the sequence changes by several ALTER commands
+ALTER SEQUENCE test_sequence INCREMENT BY 10;
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 14
+(1 row)
+
+ALTER SEQUENCE test_sequence START WITH 3000;
+ALTER SEQUENCE test_sequence MAXVALUE 10000;
+ALTER SEQUENCE test_sequence RESTART WITH 4000;
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 4000
+(1 row)
+
+-- test the sequence changes by several setval() calls
+SELECT setval('test_sequence', 3500);
+ setval
+--------
+ 3500
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3510
+(1 row)
+
+SELECT setval('test_sequence', 3500, true);
+ setval
+--------
+ 3500
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3510
+(1 row)
+
+SELECT setval('test_sequence', 3500, false);
+ setval
+--------
+ 3500
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3500
+(1 row)
+
+-- show results and drop sequence
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+ data
+--------------------------------------------------------------------------------------
+ BEGIN
+ sequence public.test_sequence: transactional:1 last_value:1 log_cnt:0 is_called:0
+ COMMIT
+ sequence public.test_sequence: transactional:0 last_value:33 log_cnt:0 is_called:1
+ BEGIN
+ sequence public.test_sequence: transactional:1 last_value:4 log_cnt:0 is_called:1
+ COMMIT
+ sequence public.test_sequence: transactional:0 last_value:334 log_cnt:0 is_called:1
+ BEGIN
+ sequence public.test_sequence: transactional:1 last_value:14 log_cnt:32 is_called:1
+ COMMIT
+ BEGIN
+ sequence public.test_sequence: transactional:1 last_value:14 log_cnt:0 is_called:1
+ COMMIT
+ BEGIN
+ sequence public.test_sequence: transactional:1 last_value:4000 log_cnt:0 is_called:0
+ COMMIT
+ sequence public.test_sequence: transactional:0 last_value:4320 log_cnt:0 is_called:1
+ sequence public.test_sequence: transactional:0 last_value:3500 log_cnt:0 is_called:1
+ sequence public.test_sequence: transactional:0 last_value:3830 log_cnt:0 is_called:1
+ sequence public.test_sequence: transactional:0 last_value:3500 log_cnt:0 is_called:1
+ sequence public.test_sequence: transactional:0 last_value:3830 log_cnt:0 is_called:1
+ sequence public.test_sequence: transactional:0 last_value:3500 log_cnt:0 is_called:0
+ sequence public.test_sequence: transactional:0 last_value:3820 log_cnt:0 is_called:1
+(24 rows)
+
+DROP SEQUENCE test_sequence;
+-- rollback on sequence creation and update
+BEGIN;
+CREATE SEQUENCE test_sequence;
+CREATE TABLE test_table (a INT);
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 1
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 2
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3
+(1 row)
+
+SELECT setval('test_sequence', 3000);
+ setval
+--------
+ 3000
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3001
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3002
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3003
+(1 row)
+
+ALTER SEQUENCE test_sequence RESTART WITH 6000;
+INSERT INTO test_table VALUES( (SELECT nextval('test_sequence')) );
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 6001
+(1 row)
+
+ROLLBACK;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'skip-sequences', '0');
+ data
+------
+(0 rows)
+
+-- rollback on table creation with serial column
+BEGIN;
+CREATE TABLE test_table (a SERIAL, b INT);
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+INSERT INTO test_table (b) VALUES (300);
+SELECT setval('test_table_a_seq', 3000);
+ setval
+--------
+ 3000
+(1 row)
+
+INSERT INTO test_table (b) VALUES (400);
+INSERT INTO test_table (b) VALUES (500);
+INSERT INTO test_table (b) VALUES (600);
+ALTER SEQUENCE test_table_a_seq RESTART WITH 6000;
+INSERT INTO test_table (b) VALUES (700);
+INSERT INTO test_table (b) VALUES (800);
+INSERT INTO test_table (b) VALUES (900);
+ROLLBACK;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'skip-sequences', '0');
+ data
+------
+(0 rows)
+
+-- rollback on table with serial column
+CREATE TABLE test_table (a SERIAL, b INT);
+BEGIN;
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+INSERT INTO test_table (b) VALUES (300);
+SELECT setval('test_table_a_seq', 3000);
+ setval
+--------
+ 3000
+(1 row)
+
+INSERT INTO test_table (b) VALUES (400);
+INSERT INTO test_table (b) VALUES (500);
+INSERT INTO test_table (b) VALUES (600);
+ALTER SEQUENCE test_table_a_seq RESTART WITH 6000;
+INSERT INTO test_table (b) VALUES (700);
+INSERT INTO test_table (b) VALUES (800);
+INSERT INTO test_table (b) VALUES (900);
+ROLLBACK;
+-- check table and sequence values after rollback
+SELECT * from test_table_a_seq;
+ last_value | log_cnt | is_called
+------------+---------+-----------
+ 3003 | 30 | t
+(1 row)
+
+SELECT nextval('test_table_a_seq');
+ nextval
+---------
+ 3004
+(1 row)
+
+DROP TABLE test_table;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+ data
+-----------------------------------------------------------------------------------------
+ BEGIN
+ sequence public.test_table_a_seq: transactional:1 last_value:1 log_cnt:0 is_called:0
+ COMMIT
+ sequence public.test_table_a_seq: transactional:0 last_value:33 log_cnt:0 is_called:1
+ sequence public.test_table_a_seq: transactional:0 last_value:3000 log_cnt:0 is_called:1
+ sequence public.test_table_a_seq: transactional:0 last_value:3033 log_cnt:0 is_called:1
+ BEGIN
+ COMMIT
+(8 rows)
+
+-- savepoint test on table with serial column
+BEGIN;
+CREATE TABLE test_table (a SERIAL, b INT);
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+SAVEPOINT a;
+INSERT INTO test_table (b) VALUES (300);
+ROLLBACK TO SAVEPOINT a;
+DROP TABLE test_table;
+COMMIT;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+ data
+---------------------------------------------------------------------------------------
+ BEGIN
+ sequence public.test_table_a_seq: transactional:1 last_value:1 log_cnt:0 is_called:0
+ sequence public.test_table_a_seq: transactional:1 last_value:33 log_cnt:0 is_called:1
+ table public.test_table: INSERT: a[integer]:1 b[integer]:100
+ table public.test_table: INSERT: a[integer]:2 b[integer]:200
+ COMMIT
+(6 rows)
+
+-- savepoint test on table with serial column
+BEGIN;
+CREATE SEQUENCE test_sequence;
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 1
+(1 row)
+
+SELECT setval('test_sequence', 3000);
+ setval
+--------
+ 3000
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3001
+(1 row)
+
+SAVEPOINT a;
+ALTER SEQUENCE test_sequence START WITH 7000;
+SELECT setval('test_sequence', 5000);
+ setval
+--------
+ 5000
+(1 row)
+
+ROLLBACK TO SAVEPOINT a;
+SELECT * FROM test_sequence;
+ last_value | log_cnt | is_called
+------------+---------+-----------
+ 3001 | 32 | t
+(1 row)
+
+DROP SEQUENCE test_sequence;
+COMMIT;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+ data
+--------------------------------------------------------------------------------------
+ BEGIN
+ sequence public.test_sequence: transactional:1 last_value:1 log_cnt:0 is_called:0
+ sequence public.test_sequence: transactional:1 last_value:33 log_cnt:0 is_called:1
+ sequence public.test_sequence: transactional:1 last_value:3000 log_cnt:0 is_called:1
+ sequence public.test_sequence: transactional:1 last_value:3033 log_cnt:0 is_called:1
+ COMMIT
+(6 rows)
+
+SELECT pg_drop_replication_slot('regression_slot');
+ pg_drop_replication_slot
+--------------------------
+
+(1 row)
+
diff --git a/contrib/test_decoding/sql/sequence.sql b/contrib/test_decoding/sql/sequence.sql
new file mode 100644
index 00000000000..21c4b792224
--- /dev/null
+++ b/contrib/test_decoding/sql/sequence.sql
@@ -0,0 +1,119 @@
+-- predictability
+SET synchronous_commit = on;
+SELECT 'init' FROM pg_create_logical_replication_slot('regression_slot', 'test_decoding');
+
+CREATE SEQUENCE test_sequence;
+
+-- test the sequence changes by several nextval() calls
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+
+-- test the sequence changes by several ALTER commands
+ALTER SEQUENCE test_sequence INCREMENT BY 10;
+SELECT nextval('test_sequence');
+
+ALTER SEQUENCE test_sequence START WITH 3000;
+ALTER SEQUENCE test_sequence MAXVALUE 10000;
+ALTER SEQUENCE test_sequence RESTART WITH 4000;
+SELECT nextval('test_sequence');
+
+-- test the sequence changes by several setval() calls
+SELECT setval('test_sequence', 3500);
+SELECT nextval('test_sequence');
+SELECT setval('test_sequence', 3500, true);
+SELECT nextval('test_sequence');
+SELECT setval('test_sequence', 3500, false);
+SELECT nextval('test_sequence');
+
+-- show results and drop sequence
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+DROP SEQUENCE test_sequence;
+
+-- rollback on sequence creation and update
+BEGIN;
+CREATE SEQUENCE test_sequence;
+CREATE TABLE test_table (a INT);
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT setval('test_sequence', 3000);
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+ALTER SEQUENCE test_sequence RESTART WITH 6000;
+INSERT INTO test_table VALUES( (SELECT nextval('test_sequence')) );
+SELECT nextval('test_sequence');
+ROLLBACK;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'skip-sequences', '0');
+
+-- rollback on table creation with serial column
+BEGIN;
+CREATE TABLE test_table (a SERIAL, b INT);
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+INSERT INTO test_table (b) VALUES (300);
+SELECT setval('test_table_a_seq', 3000);
+INSERT INTO test_table (b) VALUES (400);
+INSERT INTO test_table (b) VALUES (500);
+INSERT INTO test_table (b) VALUES (600);
+ALTER SEQUENCE test_table_a_seq RESTART WITH 6000;
+INSERT INTO test_table (b) VALUES (700);
+INSERT INTO test_table (b) VALUES (800);
+INSERT INTO test_table (b) VALUES (900);
+ROLLBACK;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'skip-sequences', '0');
+
+-- rollback on table with serial column
+CREATE TABLE test_table (a SERIAL, b INT);
+
+BEGIN;
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+INSERT INTO test_table (b) VALUES (300);
+SELECT setval('test_table_a_seq', 3000);
+INSERT INTO test_table (b) VALUES (400);
+INSERT INTO test_table (b) VALUES (500);
+INSERT INTO test_table (b) VALUES (600);
+ALTER SEQUENCE test_table_a_seq RESTART WITH 6000;
+INSERT INTO test_table (b) VALUES (700);
+INSERT INTO test_table (b) VALUES (800);
+INSERT INTO test_table (b) VALUES (900);
+ROLLBACK;
+
+-- check table and sequence values after rollback
+SELECT * from test_table_a_seq;
+SELECT nextval('test_table_a_seq');
+
+DROP TABLE test_table;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+
+-- savepoint test on table with serial column
+BEGIN;
+CREATE TABLE test_table (a SERIAL, b INT);
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+SAVEPOINT a;
+INSERT INTO test_table (b) VALUES (300);
+ROLLBACK TO SAVEPOINT a;
+DROP TABLE test_table;
+COMMIT;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+
+-- savepoint test on table with serial column
+BEGIN;
+CREATE SEQUENCE test_sequence;
+SELECT nextval('test_sequence');
+SELECT setval('test_sequence', 3000);
+SELECT nextval('test_sequence');
+SAVEPOINT a;
+ALTER SEQUENCE test_sequence START WITH 7000;
+SELECT setval('test_sequence', 5000);
+ROLLBACK TO SAVEPOINT a;
+SELECT * FROM test_sequence;
+DROP SEQUENCE test_sequence;
+COMMIT;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+
+SELECT pg_drop_replication_slot('regression_slot');
diff --git a/contrib/test_decoding/test_decoding.c b/contrib/test_decoding/test_decoding.c
index e5cd84e85e4..bc3250b8f4a 100644
--- a/contrib/test_decoding/test_decoding.c
+++ b/contrib/test_decoding/test_decoding.c
@@ -35,6 +35,7 @@ typedef struct
bool include_timestamp;
bool skip_empty_xacts;
bool only_local;
+ bool skip_sequences;
} TestDecodingData;
/*
@@ -76,6 +77,10 @@ static void pg_decode_message(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn, XLogRecPtr message_lsn,
bool transactional, const char *prefix,
Size sz, const char *message);
+static void pg_decode_sequence(LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn, XLogRecPtr sequence_lsn,
+ Relation rel, bool transactional,
+ int64 last_value, int64 log_cnt, bool is_called);
static bool pg_decode_filter_prepare(LogicalDecodingContext *ctx,
TransactionId xid,
const char *gid);
@@ -116,6 +121,10 @@ static void pg_decode_stream_message(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn, XLogRecPtr message_lsn,
bool transactional, const char *prefix,
Size sz, const char *message);
+static void pg_decode_stream_sequence(LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn, XLogRecPtr sequence_lsn,
+ Relation rel, bool transactional,
+ int64 last_value, int64 log_cnt, bool is_called);
static void pg_decode_stream_truncate(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn,
int nrelations, Relation relations[],
@@ -141,6 +150,7 @@ _PG_output_plugin_init(OutputPluginCallbacks *cb)
cb->filter_by_origin_cb = pg_decode_filter;
cb->shutdown_cb = pg_decode_shutdown;
cb->message_cb = pg_decode_message;
+ cb->sequence_cb = pg_decode_sequence;
cb->filter_prepare_cb = pg_decode_filter_prepare;
cb->begin_prepare_cb = pg_decode_begin_prepare_txn;
cb->prepare_cb = pg_decode_prepare_txn;
@@ -153,6 +163,7 @@ _PG_output_plugin_init(OutputPluginCallbacks *cb)
cb->stream_commit_cb = pg_decode_stream_commit;
cb->stream_change_cb = pg_decode_stream_change;
cb->stream_message_cb = pg_decode_stream_message;
+ cb->stream_sequence_cb = pg_decode_stream_sequence;
cb->stream_truncate_cb = pg_decode_stream_truncate;
}
@@ -175,6 +186,9 @@ pg_decode_startup(LogicalDecodingContext *ctx, OutputPluginOptions *opt,
data->skip_empty_xacts = false;
data->only_local = false;
+ /* skip sequences by default for backwards compatibility */
+ data->skip_sequences = true;
+
ctx->output_plugin_private = data;
opt->output_type = OUTPUT_PLUGIN_TEXTUAL_OUTPUT;
@@ -265,6 +279,17 @@ pg_decode_startup(LogicalDecodingContext *ctx, OutputPluginOptions *opt,
errmsg("could not parse value \"%s\" for parameter \"%s\"",
strVal(elem->arg), elem->defname)));
}
+ else if (strcmp(elem->defname, "skip-sequences") == 0)
+ {
+
+ if (elem->arg == NULL)
+ continue; /* true by default */
+ else if (!parse_bool(strVal(elem->arg), &data->skip_sequences))
+ ereport(ERROR,
+ (errcode(ERRCODE_INVALID_PARAMETER_VALUE),
+ errmsg("could not parse value \"%s\" for parameter \"%s\"",
+ strVal(elem->arg), elem->defname)));
+ }
else
{
ereport(ERROR,
@@ -744,6 +769,28 @@ pg_decode_message(LogicalDecodingContext *ctx,
OutputPluginWrite(ctx, true);
}
+static void
+pg_decode_sequence(LogicalDecodingContext *ctx, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ bool transactional,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ TestDecodingData *data = ctx->output_plugin_private;
+
+ /* return if requested to skip_sequences */
+ if (data->skip_sequences)
+ return;
+
+ OutputPluginPrepareWrite(ctx, true);
+ appendStringInfoString(ctx->out, "sequence ");
+ appendStringInfoString(ctx->out,
+ quote_qualified_identifier(get_namespace_name(get_rel_namespace(RelationGetRelid(rel))),
+ RelationGetRelationName(rel)));
+ appendStringInfo(ctx->out, ": transactional:%d last_value:%zu log_cnt:%zu is_called:%d",
+ transactional, last_value, log_cnt, is_called);
+ OutputPluginWrite(ctx, true);
+}
+
static void
pg_decode_stream_start(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn)
@@ -943,6 +990,28 @@ pg_decode_stream_message(LogicalDecodingContext *ctx,
OutputPluginWrite(ctx, true);
}
+static void
+pg_decode_stream_sequence(LogicalDecodingContext *ctx, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ bool transactional,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ TestDecodingData *data = ctx->output_plugin_private;
+
+ /* return if requested to skip_sequences */
+ if (data->skip_sequences)
+ return;
+
+ OutputPluginPrepareWrite(ctx, true);
+ appendStringInfoString(ctx->out, "streaming sequence ");
+ appendStringInfoString(ctx->out,
+ quote_qualified_identifier(get_namespace_name(get_rel_namespace(RelationGetRelid(rel))),
+ RelationGetRelationName(rel)));
+ appendStringInfo(ctx->out, ": transactional:%d last_value:%zu log_cnt:%zu is_called:%d",
+ transactional, last_value, log_cnt, is_called);
+ OutputPluginWrite(ctx, true);
+}
+
/*
* In streaming mode, we don't display the detailed information of Truncate.
* See pg_decode_stream_change.
--
2.31.1
[text/x-patch] 0001-Logical-decoding-of-sequences-20211214.patch (45.1K, ../../[email protected]/4-0001-Logical-decoding-of-sequences-20211214.patch)
download | inline diff:
From 58bf0d2abb4fa45bdf67c271cbd43f6afb354342 Mon Sep 17 00:00:00 2001
From: Tomas Vondra <[email protected]>
Date: Fri, 24 Sep 2021 00:41:33 +0200
Subject: [PATCH 1/3] Logical decoding of sequences
This adds the infrastructure for logical decoding of sequences. Support
for built-in logical replication and test_decoding is added separately.
When decoding sequences, we differentiate between a sequences created in
a (running) transaction, and sequences created in other (already
committed) transactions. Changes for sequences created in the same
top-level transaction are treated as "transactional" i.e. just like any
other change from that transaction (and discarded in case of a
rollback). Changes for sequences created earlier are treated as not
transactional - are processed immediately, as if performed outside any
transaction (and thus not rolled back).
This mixed behavior is necessary - sequences are non-transactional (e.g.
ROLLBACK does not undo the sequence increments). But for new sequences,
we need to handle them in a transactional way, because if we ever get
some DDL support, the sequence won't exist until the transaction gets
applied. So we need to ensure the increments don't happen until the
sequence gets created.
To differentiate which sequences are "old" and which were created in a
still-running transaction, we track sequences created in running
transactions in a hash table. Each sequence is identified by it's
relfilenode, and at transaction commit we discard all sequences created
in that particular transaction. For each sequence we track the XID of
the (sub)transaction that created it, and we cleanup the sequences for
each subtransaction when it completes (commit/rollback).
We don't use the XID to check if it's the same top-level transaction.
It's enough to know it was created in an in-progress transaction, and we
know it must be the current one because otherwise it wouldn't see the
sequence object.
The main changes in this patch are:
1) ensure WAL-logging of all necessary info for sequence advances
We need to be able to associate the advance with a XID, but until now
sequence advance might have XID 0 if it was the first thing the
transaction did. So we ensure the transaction has XID.
Note: This is needed because of subxacts. A XID 0 might still have the
sequence created in a different subxact of the same top-level xact.
Then there's a "created" flag added to the xl_seq_rec, which is
necessary to differentiate WAL for created/reset sequences.
2) decoding / queueing of sequences into ReorderBuffer
This is mostly copy-paste of existing code for other decoded events.
3) tracking sequences created in in-progress transactions
We use a simple hash table, indexed by relfilenode.
---
doc/src/sgml/logicaldecoding.sgml | 62 ++-
src/backend/commands/sequence.c | 32 ++
src/backend/replication/logical/decode.c | 131 +++++-
src/backend/replication/logical/logical.c | 88 ++++
.../replication/logical/reorderbuffer.c | 400 ++++++++++++++++++
src/include/commands/sequence.h | 1 +
src/include/replication/output_plugin.h | 27 ++
src/include/replication/reorderbuffer.h | 43 +-
8 files changed, 778 insertions(+), 6 deletions(-)
diff --git a/doc/src/sgml/logicaldecoding.sgml b/doc/src/sgml/logicaldecoding.sgml
index b6353c7a125..60a2b7d1ccb 100644
--- a/doc/src/sgml/logicaldecoding.sgml
+++ b/doc/src/sgml/logicaldecoding.sgml
@@ -458,6 +458,7 @@ typedef struct OutputPluginCallbacks
LogicalDecodeTruncateCB truncate_cb;
LogicalDecodeCommitCB commit_cb;
LogicalDecodeMessageCB message_cb;
+ LogicalDecodeSequenceCB sequence_cb;
LogicalDecodeFilterByOriginCB filter_by_origin_cb;
LogicalDecodeShutdownCB shutdown_cb;
LogicalDecodeFilterPrepareCB filter_prepare_cb;
@@ -472,6 +473,7 @@ typedef struct OutputPluginCallbacks
LogicalDecodeStreamCommitCB stream_commit_cb;
LogicalDecodeStreamChangeCB stream_change_cb;
LogicalDecodeStreamMessageCB stream_message_cb;
+ LogicalDecodeStreamSequenceCB stream_sequence_cb;
LogicalDecodeStreamTruncateCB stream_truncate_cb;
} OutputPluginCallbacks;
@@ -481,9 +483,12 @@ typedef void (*LogicalOutputPluginInit) (struct OutputPluginCallbacks *cb);
and <function>commit_cb</function> callbacks are required,
while <function>startup_cb</function>,
<function>filter_by_origin_cb</function>, <function>truncate_cb</function>,
+ <function>sequence_cb</function>,
and <function>shutdown_cb</function> are optional.
If <function>truncate_cb</function> is not set but a
<command>TRUNCATE</command> is to be decoded, the action will be ignored.
+ Similarly, if <function>sequence_cb</function> is not set and a sequence
+ change is to be decoded, the action will be ignored.
</para>
<para>
@@ -492,7 +497,8 @@ typedef void (*LogicalOutputPluginInit) (struct OutputPluginCallbacks *cb);
<function>stream_stop_cb</function>, <function>stream_abort_cb</function>,
<function>stream_commit_cb</function>, <function>stream_change_cb</function>,
and <function>stream_prepare_cb</function>
- are required, while <function>stream_message_cb</function> and
+ are required, while <function>stream_message_cb</function>,
+ <function>stream_sequence_cb</function>, and
<function>stream_truncate_cb</function> are optional.
</para>
@@ -808,6 +814,35 @@ typedef void (*LogicalDecodeMessageCB) (struct LogicalDecodingContext *ctx,
</para>
</sect3>
+ <sect3 id="logicaldecoding-output-plugin-sequence">
+ <title>Sequence Callback</title>
+
+ <para>
+ The optional <function>sequence_cb</function> callback is called for
+ actions that update a sequence value.
+<programlisting>
+typedef void (*LogicalDecodeSequenceCB) (struct LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional,
+ int64 last_value,
+ int64 log_cnt,
+ bool is_called);
+</programlisting>
+ The <parameter>txn</parameter> parameter contains meta information about
+ the transaction the sequence change is part of. Note however that it can
+ be NULL when the sequence change is non-transactional and the XID was not
+ assigned yet in the transaction which updated the sequence value.
+ The <parameter>sequence_lsn</parameter> has WAL location of the sequence
+ update. The <parameter>transactional</parameter> says if the sequence has
+ to be replayed as part of the transaction or directly.
+
+ The <parameter>last_value</parameter>, <parameter>log_cnt</parameter> and
+ <parameter>is_called</parameter> parameters describe the sequence change.
+ </para>
+ </sect3>
+
<sect3 id="logicaldecoding-output-plugin-filter-prepare">
<title>Prepare Filter Callback</title>
@@ -1017,6 +1052,26 @@ typedef void (*LogicalDecodeStreamMessageCB) (struct LogicalDecodingContext *ctx
</para>
</sect3>
+ <sect3 id="logicaldecoding-output-plugin-stream-sequence">
+ <title>Stream Sequence Callback</title>
+ <para>
+ The optional <function>stream_sequence_cb</function> callback is called
+ for actions that change a sequence in a block of streamed changes
+ (demarcated by <function>stream_start_cb</function> and
+ <function>stream_stop_cb</function> calls).
+<programlisting>
+typedef void (*LogicalDecodeStreamSequenceCB) (struct LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional,
+ int64 last_value,
+ int64 log_cnt,
+ bool is_called);
+</programlisting>
+ </para>
+ </sect3>
+
<sect3 id="logicaldecoding-output-plugin-stream-truncate">
<title>Stream Truncate Callback</title>
<para>
@@ -1197,8 +1252,9 @@ OutputPluginWrite(ctx, true);
in-progress transactions. There are multiple required streaming callbacks
(<function>stream_start_cb</function>, <function>stream_stop_cb</function>,
<function>stream_abort_cb</function>, <function>stream_commit_cb</function>
- and <function>stream_change_cb</function>) and two optional callbacks
- (<function>stream_message_cb</function> and <function>stream_truncate_cb</function>).
+ and <function>stream_change_cb</function>) and multiple optional callbacks
+ (<function>stream_message_cb</function>, <function>stream_sequence_cb</function>,
+ and <function>stream_truncate_cb</function>).
Also, if streaming of two-phase commands is to be supported, then additional
callbacks must be provided. (See <xref linkend="logicaldecoding-two-phase-commits"/>
for details).
diff --git a/src/backend/commands/sequence.c b/src/backend/commands/sequence.c
index 72bfdc07a49..626b5b252f4 100644
--- a/src/backend/commands/sequence.c
+++ b/src/backend/commands/sequence.c
@@ -378,8 +378,13 @@ fill_seq_with_data(Relation rel, HeapTuple tuple)
/* check the comment above nextval_internal()'s equivalent call. */
if (RelationNeedsWAL(rel))
+ {
GetTopTransactionId();
+ if (XLogLogicalInfoActive())
+ GetCurrentTransactionId();
+ }
+
START_CRIT_SECTION();
MarkBufferDirty(buf);
@@ -399,6 +404,7 @@ fill_seq_with_data(Relation rel, HeapTuple tuple)
XLogRegisterBuffer(0, buf, REGBUF_WILL_INIT);
xlrec.node = rel->rd_node;
+ xlrec.created = true;
XLogRegisterData((char *) &xlrec, sizeof(xl_seq_rec));
XLogRegisterData((char *) tuple->t_data, tuple->t_len);
@@ -764,10 +770,28 @@ nextval_internal(Oid relid, bool check_permissions)
* It's sufficient to ensure the toplevel transaction has an xid, no need
* to assign xids subxacts, that'll already trigger an appropriate wait.
* (Have to do that here, so we're outside the critical section)
+ *
+ * We have to ensure we have a proper XID, which will be included in
+ * the XLOG record by XLogRecordAssemble. Otherwise the first nextval()
+ * in a subxact (without any preceding changes) would get XID 0, and it
+ * would then be impossible to decide which top xact it belongs to.
+ * It'd also trigger assert in DecodeSequence. We only do that with
+ * wal_level=logical, though.
+ *
+ * XXX This might seem unnecessary, because if there's no XID the xact
+ * couldn't have done anything important yet, e.g. it could not have
+ * created a sequence. But that's incorrect, as it ignores subxacts. The
+ * current subtransaction might not have done anything yet (thus no XID),
+ * but an earlier one might have created the sequence.
*/
if (logit && RelationNeedsWAL(seqrel))
+ {
GetTopTransactionId();
+ if (XLogLogicalInfoActive())
+ GetCurrentTransactionId();
+ }
+
/* ready to change the on-disk (or really, in-buffer) tuple */
START_CRIT_SECTION();
@@ -803,6 +827,7 @@ nextval_internal(Oid relid, bool check_permissions)
seq->log_cnt = 0;
xlrec.node = seqrel->rd_node;
+ xlrec.created = false;
XLogRegisterData((char *) &xlrec, sizeof(xl_seq_rec));
XLogRegisterData((char *) seqdatatuple.t_data, seqdatatuple.t_len);
@@ -977,8 +1002,13 @@ do_setval(Oid relid, int64 next, bool iscalled)
/* check the comment above nextval_internal()'s equivalent call. */
if (RelationNeedsWAL(seqrel))
+ {
GetTopTransactionId();
+ if (XLogLogicalInfoActive())
+ GetCurrentTransactionId();
+ }
+
/* ready to change the on-disk (or really, in-buffer) tuple */
START_CRIT_SECTION();
@@ -999,6 +1029,8 @@ do_setval(Oid relid, int64 next, bool iscalled)
XLogRegisterBuffer(0, buf, REGBUF_WILL_INIT);
xlrec.node = seqrel->rd_node;
+ xlrec.created = false;
+
XLogRegisterData((char *) &xlrec, sizeof(xl_seq_rec));
XLogRegisterData((char *) seqdatatuple.t_data, seqdatatuple.t_len);
diff --git a/src/backend/replication/logical/decode.c b/src/backend/replication/logical/decode.c
index 59aed6cee6c..6965be53caa 100644
--- a/src/backend/replication/logical/decode.c
+++ b/src/backend/replication/logical/decode.c
@@ -42,6 +42,7 @@
#include "replication/reorderbuffer.h"
#include "replication/snapbuild.h"
#include "storage/standby.h"
+#include "commands/sequence.h"
typedef struct XLogRecordBuffer
{
@@ -74,10 +75,11 @@ static void DecodeAbort(LogicalDecodingContext *ctx, XLogRecordBuffer *buf,
bool two_phase);
static void DecodePrepare(LogicalDecodingContext *ctx, XLogRecordBuffer *buf,
xl_xact_parsed_prepare *parsed);
-
+static void DecodeSequence(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
/* common function to decode tuples */
static void DecodeXLogTuple(char *data, Size len, ReorderBufferTupleBuf *tup);
+static void DecodeSeqTuple(char *data, Size len, ReorderBufferTupleBuf *tuple);
/* helper functions for decoding transactions */
static inline bool FilterPrepare(LogicalDecodingContext *ctx,
@@ -158,6 +160,10 @@ LogicalDecodingProcessRecord(LogicalDecodingContext *ctx, XLogReaderState *recor
DecodeLogicalMsgOp(ctx, &buf);
break;
+ case RM_SEQ_ID:
+ DecodeSequence(ctx, &buf);
+ break;
+
/*
* Rmgrs irrelevant for logical decoding; they describe stuff not
* represented in logical decoding. Add new rmgrs in rmgrlist.h's
@@ -173,7 +179,6 @@ LogicalDecodingProcessRecord(LogicalDecodingContext *ctx, XLogReaderState *recor
case RM_HASH_ID:
case RM_GIN_ID:
case RM_GIST_ID:
- case RM_SEQ_ID:
case RM_SPGIST_ID:
case RM_BRIN_ID:
case RM_COMMIT_TS_ID:
@@ -1313,3 +1318,125 @@ DecodeTXNNeedSkip(LogicalDecodingContext *ctx, XLogRecordBuffer *buf,
(txn_dbid != InvalidOid && txn_dbid != ctx->slot->data.database) ||
ctx->fast_forward || FilterByOrigin(ctx, origin_id));
}
+
+/*
+ * Decode Sequence Tuple
+ */
+static void
+DecodeSeqTuple(char *data, Size len, ReorderBufferTupleBuf *tuple)
+{
+ int datalen = len - sizeof(xl_seq_rec) - SizeofHeapTupleHeader;
+
+ Assert(datalen >= 0);
+
+ tuple->tuple.t_len = datalen + SizeofHeapTupleHeader;;
+
+ ItemPointerSetInvalid(&tuple->tuple.t_self);
+
+ tuple->tuple.t_tableOid = InvalidOid;
+
+ memcpy(((char *) tuple->tuple.t_data),
+ data + sizeof(xl_seq_rec),
+ SizeofHeapTupleHeader);
+
+ memcpy(((char *) tuple->tuple.t_data) + SizeofHeapTupleHeader,
+ data + sizeof(xl_seq_rec) + SizeofHeapTupleHeader,
+ datalen);
+}
+
+/*
+ * Handle sequence decode
+ *
+ * Decoding sequences is a bit tricky, because while most sequence actions
+ * are non-transactional (not subject to rollback), some need to be handled
+ * as transactional.
+ *
+ * By default, a sequence increment is non-transactional - we must not queue
+ * it in a transaction as other changes, because the transaction might get
+ * rolled back and we'd discard the increment. The downstream would not be
+ * notified about the increment, which is wrong.
+ *
+ * On the other hand, the sequence may be created in a transaction. In this
+ * case we *should* queue the change as other changes in the transaction,
+ * because we don't want to send the increments for unknown sequence to the
+ * plugin - it might get confused about which sequence it's related to etc.
+ */
+static void
+DecodeSequence(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
+{
+ SnapBuild *builder = ctx->snapshot_builder;
+ ReorderBufferTupleBuf *tuplebuf;
+ RelFileNode target_node;
+ XLogReaderState *r = buf->record;
+ char *tupledata = NULL;
+ Size tuplelen;
+ Size datalen = 0;
+ TransactionId xid = XLogRecGetXid(r);
+ uint8 info = XLogRecGetInfo(buf->record) & ~XLR_INFO_MASK;
+ xl_seq_rec *xlrec;
+ Snapshot snapshot;
+ RepOriginId origin_id = XLogRecGetOrigin(r);
+ bool transactional;
+
+ /* only decode changes flagged with XLOG_SEQ_LOG */
+ if (info != XLOG_SEQ_LOG)
+ elog(ERROR, "unexpected RM_SEQ_ID record type: %u", info);
+
+ /*
+ * If we don't have snapshot or we are just fast-forwarding, there is no
+ * point in decoding messages.
+ */
+ if (SnapBuildCurrentState(builder) < SNAPBUILD_FULL_SNAPSHOT ||
+ ctx->fast_forward)
+ return;
+
+ /* only interested in our database */
+ XLogRecGetBlockTag(r, 0, &target_node, NULL, NULL);
+ if (target_node.dbNode != ctx->slot->data.database)
+ return;
+
+ /* output plugin doesn't look for this origin, no need to queue */
+ if (FilterByOrigin(ctx, XLogRecGetOrigin(r)))
+ return;
+
+ tupledata = XLogRecGetData(r);
+ datalen = XLogRecGetDataLen(r);
+ tuplelen = datalen - SizeOfHeapHeader - sizeof(xl_seq_rec);
+
+ /* extract the WAL record, with "created" flag */
+ xlrec = (xl_seq_rec *) XLogRecGetData(r);
+
+ /* XXX how could we have sequence change without data? */
+ if(!datalen || !tupledata)
+ return;
+
+ tuplebuf = ReorderBufferGetTupleBuf(ctx->reorder, tuplelen);
+ DecodeSeqTuple(tupledata, datalen, tuplebuf);
+
+ /*
+ * Should we handle the sequence increment as transactional or not?
+ *
+ * If the sequence was created in a still-running transaction, treat
+ * it as transactional and queue the increments. Otherwise it needs
+ * to be treated as non-transactional, in which case we send it to
+ * the plugin right away.
+ */
+ transactional = ReorderBufferSequenceIsTransactional(ctx->reorder,
+ target_node,
+ xlrec->created);
+
+ /* Skip the change if already processed (per the snapshot). */
+ if (transactional &&
+ !SnapBuildProcessChange(builder, xid, buf->origptr))
+ return;
+ else if (!transactional &&
+ (SnapBuildCurrentState(builder) != SNAPBUILD_CONSISTENT ||
+ SnapBuildXactNeedsSkip(builder, buf->origptr)))
+ return;
+
+ /* Queue the increment (or send immediately if not transactional). */
+ snapshot = SnapBuildGetOrBuildSnapshot(builder, xid);
+ ReorderBufferQueueSequence(ctx->reorder, xid, snapshot, buf->endptr,
+ origin_id, target_node, transactional,
+ xlrec->created, tuplebuf);
+}
diff --git a/src/backend/replication/logical/logical.c b/src/backend/replication/logical/logical.c
index 10cbdea124c..aab5e1226d4 100644
--- a/src/backend/replication/logical/logical.c
+++ b/src/backend/replication/logical/logical.c
@@ -73,6 +73,10 @@ static void truncate_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
static void message_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
XLogRecPtr message_lsn, bool transactional,
const char *prefix, Size message_size, const char *message);
+static void sequence_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ bool transactional,
+ int64 last_value, int64 log_cnt, bool is_called);
/* streaming callbacks */
static void stream_start_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
@@ -90,6 +94,10 @@ static void stream_change_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn
static void stream_message_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
XLogRecPtr message_lsn, bool transactional,
const char *prefix, Size message_size, const char *message);
+static void stream_sequence_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ bool transactional,
+ int64 last_value, int64 log_cnt, bool is_called);
static void stream_truncate_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
int nrelations, Relation relations[], ReorderBufferChange *change);
@@ -218,6 +226,7 @@ StartupDecodingContext(List *output_plugin_options,
ctx->reorder->apply_truncate = truncate_cb_wrapper;
ctx->reorder->commit = commit_cb_wrapper;
ctx->reorder->message = message_cb_wrapper;
+ ctx->reorder->sequence = sequence_cb_wrapper;
/*
* To support streaming, we require start/stop/abort/commit/change
@@ -234,6 +243,7 @@ StartupDecodingContext(List *output_plugin_options,
(ctx->callbacks.stream_commit_cb != NULL) ||
(ctx->callbacks.stream_change_cb != NULL) ||
(ctx->callbacks.stream_message_cb != NULL) ||
+ (ctx->callbacks.stream_sequence_cb != NULL) ||
(ctx->callbacks.stream_truncate_cb != NULL);
/*
@@ -251,6 +261,7 @@ StartupDecodingContext(List *output_plugin_options,
ctx->reorder->stream_commit = stream_commit_cb_wrapper;
ctx->reorder->stream_change = stream_change_cb_wrapper;
ctx->reorder->stream_message = stream_message_cb_wrapper;
+ ctx->reorder->stream_sequence = stream_sequence_cb_wrapper;
ctx->reorder->stream_truncate = stream_truncate_cb_wrapper;
@@ -1203,6 +1214,42 @@ message_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
error_context_stack = errcallback.previous;
}
+static void
+sequence_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel, bool transactional,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ LogicalDecodingContext *ctx = cache->private_data;
+ LogicalErrorCallbackState state;
+ ErrorContextCallback errcallback;
+
+ Assert(!ctx->fast_forward);
+
+ if (ctx->callbacks.sequence_cb == NULL)
+ return;
+
+ /* Push callback + info on the error context stack */
+ state.ctx = ctx;
+ state.callback_name = "sequence";
+ state.report_location = sequence_lsn;
+ errcallback.callback = output_plugin_error_callback;
+ errcallback.arg = (void *) &state;
+ errcallback.previous = error_context_stack;
+ error_context_stack = &errcallback;
+
+ /* set output state */
+ ctx->accept_writes = true;
+ ctx->write_xid = txn != NULL ? txn->xid : InvalidTransactionId;
+ ctx->write_location = sequence_lsn;
+
+ /* do the actual work: call callback */
+ ctx->callbacks.sequence_cb(ctx, txn, sequence_lsn, rel, transactional,
+ last_value, log_cnt, is_called);
+
+ /* Pop the error context stack */
+ error_context_stack = errcallback.previous;
+}
+
static void
stream_start_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
XLogRecPtr first_lsn)
@@ -1508,6 +1555,47 @@ stream_message_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
error_context_stack = errcallback.previous;
}
+static void
+stream_sequence_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ bool transactional,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ LogicalDecodingContext *ctx = cache->private_data;
+ LogicalErrorCallbackState state;
+ ErrorContextCallback errcallback;
+
+ Assert(!ctx->fast_forward);
+
+ /* We're only supposed to call this when streaming is supported. */
+ Assert(ctx->streaming);
+
+ /* this callback is optional */
+ if (ctx->callbacks.stream_sequence_cb == NULL)
+ return;
+
+ /* Push callback + info on the error context stack */
+ state.ctx = ctx;
+ state.callback_name = "stream_sequence";
+ state.report_location = sequence_lsn;
+ errcallback.callback = output_plugin_error_callback;
+ errcallback.arg = (void *) &state;
+ errcallback.previous = error_context_stack;
+ error_context_stack = &errcallback;
+
+ /* set output state */
+ ctx->accept_writes = true;
+ ctx->write_xid = txn != NULL ? txn->xid : InvalidTransactionId;
+ ctx->write_location = sequence_lsn;
+
+ /* do the actual work: call callback */
+ ctx->callbacks.sequence_cb(ctx, txn, sequence_lsn, rel, transactional,
+ last_value, log_cnt, is_called);
+
+ /* Pop the error context stack */
+ error_context_stack = errcallback.previous;
+}
+
static void
stream_truncate_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
int nrelations, Relation relations[],
diff --git a/src/backend/replication/logical/reorderbuffer.c b/src/backend/replication/logical/reorderbuffer.c
index 7aa5647a2c6..6b6d0d9b955 100644
--- a/src/backend/replication/logical/reorderbuffer.c
+++ b/src/backend/replication/logical/reorderbuffer.c
@@ -77,6 +77,35 @@
* a bit more memory to the oldest subtransactions, because it's likely
* they are the source for the next sequence of changes.
*
+ * When decoding sequences, we differentiate between a sequences created
+ * in a (running) transaction, and sequences created in other (already
+ * committed) transactions. Changes for sequences created in the same
+ * top-level transaction are treated as "transactional" i.e. just like
+ * any other change from that transaction (and discarded in case of a
+ * rollback). Changes for sequences created earlier are treated as not
+ * transactional - are processed immediately, as if performed outside
+ * any transaction (and thus not rolled back).
+ *
+ * This mixed behavior is necessary - sequences are non-transactional
+ * (e.g. ROLLBACK does not undo the sequence increments). But for new
+ * sequences, we need to handle them in a transactional way, because if
+ * we ever get some DDL support, the sequence won't exist until the
+ * transaction gets applied. So we need to ensure the increments don't
+ * happen until the sequence gets created.
+ *
+ * To differentiate which sequences are "old" and which were created
+ * in a still-running transaction, we track sequences created in running
+ * transactions in a hash table. Each sequence is identified by it's
+ * relfilenode, and at transaction commit we discard all sequences
+ * created in that particular transaction. For each sequence we track
+ * the XID of the (sub)transaction that created it, and we cleanup the
+ * sequences for each subtransaction when it completes (commit/rollback).
+ *
+ * We don't use the XID to check if it's the same top-level transaction.
+ * It's enough to know it was created in an in-progress transaction,
+ * and we know it must be the current one because otherwise it wouldn't
+ * see the sequence object.
+ *
* -------------------------------------------------------------------------
*/
#include "postgres.h"
@@ -91,6 +120,7 @@
#include "access/xact.h"
#include "access/xlog_internal.h"
#include "catalog/catalog.h"
+#include "commands/sequence.h"
#include "lib/binaryheap.h"
#include "miscadmin.h"
#include "pgstat.h"
@@ -116,6 +146,13 @@ typedef struct ReorderBufferTXNByIdEnt
ReorderBufferTXN *txn;
} ReorderBufferTXNByIdEnt;
+/* entry for hash table we use to track sequences created in running xacts */
+typedef struct ReorderBufferSequenceEnt
+{
+ RelFileNode rnode;
+ TransactionId xid;
+} ReorderBufferSequenceEnt;
+
/* data structures for (relfilenode, ctid) => (cmin, cmax) mapping */
typedef struct ReorderBufferTupleCidKey
{
@@ -339,6 +376,14 @@ ReorderBufferAllocate(void)
buffer->by_txn = hash_create("ReorderBufferByXid", 1000, &hash_ctl,
HASH_ELEM | HASH_BLOBS | HASH_CONTEXT);
+ /* hash table of sequences, mapping relfilenode to XID of transaction */
+ hash_ctl.keysize = sizeof(RelFileNode);
+ hash_ctl.entrysize = sizeof(ReorderBufferSequenceEnt);
+ hash_ctl.hcxt = buffer->context;
+
+ buffer->sequences = hash_create("ReorderBufferSequenceHash", 1000, &hash_ctl,
+ HASH_ELEM | HASH_BLOBS | HASH_CONTEXT);
+
buffer->by_txn_last_xid = InvalidTransactionId;
buffer->by_txn_last_txn = NULL;
@@ -525,6 +570,13 @@ ReorderBufferReturnChange(ReorderBuffer *rb, ReorderBufferChange *change,
change->data.truncate.relids = NULL;
}
break;
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
+ if (change->data.sequence.tuple)
+ {
+ ReorderBufferReturnTupleBuf(rb, change->data.sequence.tuple);
+ change->data.sequence.tuple = NULL;
+ }
+ break;
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM:
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_ABORT:
case REORDER_BUFFER_CHANGE_INTERNAL_COMMAND_ID:
@@ -859,6 +911,230 @@ ReorderBufferQueueMessage(ReorderBuffer *rb, TransactionId xid,
}
}
+/*
+ * Treat the sequence increment as transactional?
+ *
+ * The hash table tracks all sequences created in in-progress transactions,
+ * so we simply do a lookup (the sequence is identified by relfilende). If
+ * we find a match, the increment should be handled as transactional.
+ */
+bool
+ReorderBufferSequenceIsTransactional(ReorderBuffer *rb,
+ RelFileNode rnode, bool created)
+{
+ bool found = false;
+
+ if (created)
+ return true;
+
+ hash_search(rb->sequences,
+ (void *) &rnode,
+ HASH_FIND,
+ &found);
+
+ return found;
+}
+
+/*
+ * Cleanup sequences created in in-progress transactions.
+ *
+ * There's no way to search by XID, so we simply do a seqscan of all
+ * the entries in the hash table. Hopefully there are only a couple
+ * entries in most cases - people generally don't create many new
+ * sequences over and over.
+ */
+static void
+ReorderBufferSequenceCleanup(ReorderBuffer *rb, TransactionId xid)
+{
+ HASH_SEQ_STATUS scan_status;
+ ReorderBufferSequenceEnt *ent;
+
+ hash_seq_init(&scan_status, rb->sequences);
+ while ((ent = (ReorderBufferSequenceEnt *) hash_seq_search(&scan_status)) != NULL)
+ {
+ /* skip sequences not from this transaction */
+ if (ent->xid != xid)
+ continue;
+
+ (void) hash_search(rb->sequences,
+ (void *) &(ent->rnode),
+ HASH_REMOVE, NULL);
+ }
+}
+
+/*
+ * A transactional sequence increment is queued to be processed upon commit
+ * and a non-transactional increment gets processed immediately.
+ *
+ * A sequence update may be both transactional and non-transactional. When
+ * created in a running transaction, treat it as transactional and queue
+ * the change in it. Otherwise treat it as non-transactional, so that we
+ * don't forget the increment in case of a rollback.
+ */
+void
+ReorderBufferQueueSequence(ReorderBuffer *rb, TransactionId xid,
+ Snapshot snapshot, XLogRecPtr lsn, RepOriginId origin_id,
+ RelFileNode rnode, bool transactional, bool created,
+ ReorderBufferTupleBuf *tuplebuf)
+{
+ /*
+ * Change needs to be handled as transactional, because the sequence was
+ * created in a transaction that is still running. In that case all the
+ * changes need to be queued in that transaction, we must not send them
+ * to the downstream until the transaction commits.
+ *
+ * There's a bit of a trouble with subtransactions - we can't queue it
+ * into the subxact, because it might be rolled back and we'd lose the
+ * increment. We need to queue it into the same (sub)xact that created
+ * the sequence, which is why we track the XID in the hash table.
+ */
+ if (transactional)
+ {
+ MemoryContext oldcontext;
+ ReorderBufferChange *change;
+
+ /* lookup sequence by relfilenode */
+ ReorderBufferSequenceEnt *ent;
+ bool found;
+
+ /* transactional changes require a transaction */
+ Assert(xid != InvalidTransactionId);
+
+ /* search the lookup table (we ignore the return value, found is enough) */
+ ent = hash_search(rb->sequences,
+ (void *) &rnode,
+ created ? HASH_ENTER : HASH_FIND,
+ &found);
+
+ /*
+ * If this is the "create" increment, we must not have found any
+ * pre-existing entry in the hash table (i.e. there must not be
+ * any conflicting sequence).
+ */
+ Assert(!(created && found));
+
+ /* But we must have either created or found an existing entry. */
+ Assert(created || found);
+
+ /*
+ * When creating the sequence, remember the XID of the transaction
+ * that created id.
+ */
+ if (created)
+ ent->xid = xid;
+
+ /* XXX Maybe check that we're still in the same top-level xact? */
+
+ /* OK, allocate and queue the change */
+ oldcontext = MemoryContextSwitchTo(rb->context);
+
+ change = ReorderBufferGetChange(rb);
+
+ change->action = REORDER_BUFFER_CHANGE_SEQUENCE;
+ change->origin_id = origin_id;
+
+ memcpy(&change->data.sequence.relnode, &rnode, sizeof(RelFileNode));
+
+ change->data.sequence.tuple = tuplebuf;
+
+ /* add it to the same subxact that created the sequence */
+ ReorderBufferQueueChange(rb, ent->xid, lsn, change, false);
+
+ MemoryContextSwitchTo(oldcontext);
+ }
+ else
+ {
+ /*
+ * This increment is for a sequence that was not created in any
+ * running transaction, so we treat it as non-transactional and
+ * just send it to the output plugin directly.
+ */
+ ReorderBufferTXN *txn = NULL;
+ volatile Snapshot snapshot_now = snapshot;
+ bool using_subtxn;
+
+#ifdef USE_ASSERT_CHECKING
+ /* All "creates" have to be handled as transactional. */
+ Assert(!created);
+
+ /* Make sure the sequence is not in the hash table. */
+ {
+ bool found;
+ hash_search(rb->sequences,
+ (void *) &rnode,
+ HASH_FIND, &found);
+ Assert(!found);
+ }
+#endif
+
+ if (xid != InvalidTransactionId)
+ txn = ReorderBufferTXNByXid(rb, xid, true, NULL, lsn, true);
+
+ /* setup snapshot to allow catalog access */
+ SetupHistoricSnapshot(snapshot_now, NULL);
+
+ /*
+ * Decoding needs access to syscaches et al., which in turn use
+ * heavyweight locks and such. Thus we need to have enough state around to
+ * keep track of those. The easiest way is to simply use a transaction
+ * internally. That also allows us to easily enforce that nothing writes
+ * to the database by checking for xid assignments.
+ *
+ * When we're called via the SQL SRF there's already a transaction
+ * started, so start an explicit subtransaction there.
+ */
+ using_subtxn = IsTransactionOrTransactionBlock();
+
+ PG_TRY();
+ {
+ Relation relation;
+ HeapTuple tuple;
+ Form_pg_sequence_data seq;
+ Oid reloid;
+
+ if (using_subtxn)
+ BeginInternalSubTransaction("sequence");
+ else
+ StartTransactionCommand();
+
+ reloid = RelidByRelfilenode(rnode.spcNode, rnode.relNode);
+
+ if (reloid == InvalidOid)
+ elog(ERROR, "could not map filenode \"%s\" to relation OID",
+ relpathperm(rnode,
+ MAIN_FORKNUM));
+
+ relation = RelationIdGetRelation(reloid);
+ tuple = &tuplebuf->tuple;
+ seq = (Form_pg_sequence_data) GETSTRUCT(tuple);
+
+ rb->sequence(rb, txn, lsn, relation, transactional,
+ seq->last_value, seq->log_cnt, seq->is_called);
+
+ RelationClose(relation);
+
+ TeardownHistoricSnapshot(false);
+
+ AbortCurrentTransaction();
+
+ if (using_subtxn)
+ RollbackAndReleaseCurrentSubTransaction();
+ }
+ PG_CATCH();
+ {
+ TeardownHistoricSnapshot(true);
+
+ AbortCurrentTransaction();
+
+ if (using_subtxn)
+ RollbackAndReleaseCurrentSubTransaction();
+
+ PG_RE_THROW();
+ }
+ PG_END_TRY();
+ }
+}
+
/*
* AssertTXNLsnOrder
* Verify LSN ordering of transaction lists in the reorderbuffer
@@ -1535,6 +1811,9 @@ ReorderBufferCleanupTXN(ReorderBuffer *rb, ReorderBufferTXN *txn)
&found);
Assert(found);
+ /* Remove sequences created in this transaction (if any). */
+ ReorderBufferSequenceCleanup(rb, txn->xid);
+
/* remove entries spilled to disk */
if (rbtxn_is_serialized(txn))
ReorderBufferRestoreCleanup(rb, txn);
@@ -1950,6 +2229,29 @@ ReorderBufferApplyMessage(ReorderBuffer *rb, ReorderBufferTXN *txn,
change->data.msg.message);
}
+/*
+ * Helper function for ReorderBufferProcessTXN for applying sequences.
+ */
+static inline void
+ReorderBufferApplySequence(ReorderBuffer *rb, ReorderBufferTXN *txn,
+ Relation relation, ReorderBufferChange *change,
+ bool streaming)
+{
+ HeapTuple tuple;
+ Form_pg_sequence_data seq;
+
+ tuple = &change->data.sequence.tuple->tuple;
+ seq = (Form_pg_sequence_data) GETSTRUCT(tuple);
+
+ /* Only ever called from ReorderBufferApplySequence, so transational. */
+ if (streaming)
+ rb->stream_sequence(rb, txn, change->lsn, relation, true,
+ seq->last_value, seq->log_cnt, seq->is_called);
+ else
+ rb->sequence(rb, txn, change->lsn, relation, true,
+ seq->last_value, seq->log_cnt, seq->is_called);
+}
+
/*
* Function to store the command id and snapshot at the end of the current
* stream so that we can reuse the same while sending the next stream.
@@ -2392,6 +2694,31 @@ ReorderBufferProcessTXN(ReorderBuffer *rb, ReorderBufferTXN *txn,
case REORDER_BUFFER_CHANGE_INTERNAL_TUPLECID:
elog(ERROR, "tuplecid value in changequeue");
break;
+
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
+ Assert(snapshot_now);
+
+ reloid = RelidByRelfilenode(change->data.sequence.relnode.spcNode,
+ change->data.sequence.relnode.relNode);
+
+ if (reloid == InvalidOid)
+ elog(ERROR, "could not map filenode \"%s\" to relation OID",
+ relpathperm(change->data.sequence.relnode,
+ MAIN_FORKNUM));
+
+ relation = RelationIdGetRelation(reloid);
+
+ if (!RelationIsValid(relation))
+ elog(ERROR, "could not open relation with OID %u (for filenode \"%s\")",
+ reloid,
+ relpathperm(change->data.sequence.relnode,
+ MAIN_FORKNUM));
+
+ if (RelationIsLogicallyLogged(relation))
+ ReorderBufferApplySequence(rb, txn, relation, change, streaming);
+
+ RelationClose(relation);
+ break;
}
}
@@ -3776,6 +4103,39 @@ ReorderBufferSerializeChange(ReorderBuffer *rb, ReorderBufferTXN *txn,
memcpy(data, change->data.truncate.relids, size);
data += size;
+ break;
+ }
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
+ {
+ char *data;
+ ReorderBufferTupleBuf *tup;
+ Size len = 0;
+
+ tup = change->data.sequence.tuple;
+
+ if (tup)
+ {
+ sz += sizeof(HeapTupleData);
+ len = tup->tuple.t_len;
+ sz += len;
+ }
+
+ /* make sure we have enough space */
+ ReorderBufferSerializeReserve(rb, sz);
+
+ data = ((char *) rb->outbuf) + sizeof(ReorderBufferDiskChange);
+ /* might have been reallocated above */
+ ondisk = (ReorderBufferDiskChange *) rb->outbuf;
+
+ if (len)
+ {
+ memcpy(data, &tup->tuple, sizeof(HeapTupleData));
+ data += sizeof(HeapTupleData);
+
+ memcpy(data, tup->tuple.t_data, len);
+ data += len;
+ }
+
break;
}
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM:
@@ -4040,6 +4400,22 @@ ReorderBufferChangeSize(ReorderBufferChange *change)
{
sz += sizeof(Oid) * change->data.truncate.nrelids;
+ break;
+ }
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
+ {
+ ReorderBufferTupleBuf *tup;
+ Size len = 0;
+
+ tup = change->data.sequence.tuple;
+
+ if (tup)
+ {
+ sz += sizeof(HeapTupleData);
+ len = tup->tuple.t_len;
+ sz += len;
+ }
+
break;
}
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM:
@@ -4341,6 +4717,30 @@ ReorderBufferRestoreChange(ReorderBuffer *rb, ReorderBufferTXN *txn,
break;
}
+
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
+ if (change->data.sequence.tuple)
+ {
+ uint32 tuplelen = ((HeapTuple) data)->t_len;
+
+ change->data.sequence.tuple =
+ ReorderBufferGetTupleBuf(rb, tuplelen - SizeofHeapTupleHeader);
+
+ /* restore ->tuple */
+ memcpy(&change->data.sequence.tuple->tuple, data,
+ sizeof(HeapTupleData));
+ data += sizeof(HeapTupleData);
+
+ /* reset t_data pointer into the new tuplebuf */
+ change->data.sequence.tuple->tuple.t_data =
+ ReorderBufferTupleBufData(change->data.sequence.tuple);
+
+ /* restore tuple data itself */
+ memcpy(change->data.sequence.tuple->tuple.t_data, data, tuplelen);
+ data += tuplelen;
+ }
+ break;
+
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM:
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_ABORT:
case REORDER_BUFFER_CHANGE_INTERNAL_COMMAND_ID:
diff --git a/src/include/commands/sequence.h b/src/include/commands/sequence.h
index 40544dd4c70..5919fb90ee1 100644
--- a/src/include/commands/sequence.h
+++ b/src/include/commands/sequence.h
@@ -48,6 +48,7 @@ typedef FormData_pg_sequence_data *Form_pg_sequence_data;
typedef struct xl_seq_rec
{
RelFileNode node;
+ bool created; /* is this a CREATE SEQUENCE */
/* SEQUENCE TUPLE DATA FOLLOWS AT THE END */
} xl_seq_rec;
diff --git a/src/include/replication/output_plugin.h b/src/include/replication/output_plugin.h
index 810495ed0e4..43b8a63ee4f 100644
--- a/src/include/replication/output_plugin.h
+++ b/src/include/replication/output_plugin.h
@@ -88,6 +88,18 @@ typedef void (*LogicalDecodeMessageCB) (struct LogicalDecodingContext *ctx,
Size message_size,
const char *message);
+/*
+ * Called for the generic logical decoding sequences.
+ */
+typedef void (*LogicalDecodeSequenceCB) (struct LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional,
+ int64 last_value,
+ int64 log_cnt,
+ bool is_called);
+
/*
* Filter changes by origin.
*/
@@ -199,6 +211,19 @@ typedef void (*LogicalDecodeStreamMessageCB) (struct LogicalDecodingContext *ctx
Size message_size,
const char *message);
+/*
+ * Called for the streaming generic logical decoding sequences from in-progress
+ * transactions.
+ */
+typedef void (*LogicalDecodeStreamSequenceCB) (struct LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional,
+ int64 last_value,
+ int64 log_cnt,
+ bool is_called);
+
/*
* Callback for streaming truncates from in-progress transactions.
*/
@@ -219,6 +244,7 @@ typedef struct OutputPluginCallbacks
LogicalDecodeTruncateCB truncate_cb;
LogicalDecodeCommitCB commit_cb;
LogicalDecodeMessageCB message_cb;
+ LogicalDecodeSequenceCB sequence_cb;
LogicalDecodeFilterByOriginCB filter_by_origin_cb;
LogicalDecodeShutdownCB shutdown_cb;
@@ -237,6 +263,7 @@ typedef struct OutputPluginCallbacks
LogicalDecodeStreamCommitCB stream_commit_cb;
LogicalDecodeStreamChangeCB stream_change_cb;
LogicalDecodeStreamMessageCB stream_message_cb;
+ LogicalDecodeStreamSequenceCB stream_sequence_cb;
LogicalDecodeStreamTruncateCB stream_truncate_cb;
} OutputPluginCallbacks;
diff --git a/src/include/replication/reorderbuffer.h b/src/include/replication/reorderbuffer.h
index 5b40ff75f79..ab469fb2109 100644
--- a/src/include/replication/reorderbuffer.h
+++ b/src/include/replication/reorderbuffer.h
@@ -64,7 +64,8 @@ enum ReorderBufferChangeType
REORDER_BUFFER_CHANGE_INTERNAL_SPEC_INSERT,
REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM,
REORDER_BUFFER_CHANGE_INTERNAL_SPEC_ABORT,
- REORDER_BUFFER_CHANGE_TRUNCATE
+ REORDER_BUFFER_CHANGE_TRUNCATE,
+ REORDER_BUFFER_CHANGE_SEQUENCE
};
/* forward declaration */
@@ -158,6 +159,13 @@ typedef struct ReorderBufferChange
uint32 ninvalidations; /* Number of messages */
SharedInvalidationMessage *invalidations; /* invalidation message */
} inval;
+
+ /* Context data for Sequence changes */
+ struct
+ {
+ RelFileNode relnode;
+ ReorderBufferTupleBuf *tuple;
+ } sequence;
} data;
/*
@@ -430,6 +438,15 @@ typedef void (*ReorderBufferMessageCB) (ReorderBuffer *rb,
const char *prefix, Size sz,
const char *message);
+/* sequence callback signature */
+typedef void (*ReorderBufferSequenceCB) (ReorderBuffer *rb,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional,
+ int64 last_value, int64 log_cnt,
+ bool is_called);
+
/* begin prepare callback signature */
typedef void (*ReorderBufferBeginPrepareCB) (ReorderBuffer *rb,
ReorderBufferTXN *txn);
@@ -496,6 +513,15 @@ typedef void (*ReorderBufferStreamMessageCB) (
const char *prefix, Size sz,
const char *message);
+/* stream sequence callback signature */
+typedef void (*ReorderBufferStreamSequenceCB) (ReorderBuffer *rb,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional,
+ int64 last_value, int64 log_cnt,
+ bool is_called);
+
/* stream truncate callback signature */
typedef void (*ReorderBufferStreamTruncateCB) (
ReorderBuffer *rb,
@@ -511,6 +537,12 @@ struct ReorderBuffer
*/
HTAB *by_txn;
+ /*
+ * relfilenode => XID lookup table for sequences created in a transaction
+ * (also includes altered sequences, which assigns new relfilenode)
+ */
+ HTAB *sequences;
+
/*
* Transactions that could be a toplevel xact, ordered by LSN of the first
* record bearing that xid.
@@ -541,6 +573,7 @@ struct ReorderBuffer
ReorderBufferApplyTruncateCB apply_truncate;
ReorderBufferCommitCB commit;
ReorderBufferMessageCB message;
+ ReorderBufferSequenceCB sequence;
/*
* Callbacks to be called when streaming a transaction at prepare time.
@@ -560,6 +593,7 @@ struct ReorderBuffer
ReorderBufferStreamCommitCB stream_commit;
ReorderBufferStreamChangeCB stream_change;
ReorderBufferStreamMessageCB stream_message;
+ ReorderBufferStreamSequenceCB stream_sequence;
ReorderBufferStreamTruncateCB stream_truncate;
/*
@@ -635,6 +669,10 @@ void ReorderBufferQueueChange(ReorderBuffer *, TransactionId,
void ReorderBufferQueueMessage(ReorderBuffer *, TransactionId, Snapshot snapshot, XLogRecPtr lsn,
bool transactional, const char *prefix,
Size message_size, const char *message);
+void ReorderBufferQueueSequence(ReorderBuffer *rb, TransactionId xid,
+ Snapshot snapshot, XLogRecPtr lsn, RepOriginId origin_id,
+ RelFileNode rnode, bool transactional, bool created,
+ ReorderBufferTupleBuf *tuplebuf);
void ReorderBufferCommit(ReorderBuffer *, TransactionId,
XLogRecPtr commit_lsn, XLogRecPtr end_lsn,
TimestampTz commit_time, RepOriginId origin_id, XLogRecPtr origin_lsn);
@@ -682,4 +720,7 @@ void ReorderBufferSetRestartPoint(ReorderBuffer *, XLogRecPtr ptr);
void StartupReorderBuffer(void);
+bool ReorderBufferSequenceIsTransactional(ReorderBuffer *rb,
+ RelFileNode rnode, bool created);
+
#endif
--
2.31.1
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2021-12-15 13:20 Amit Kapila <[email protected]>
parent: Tomas Vondra <[email protected]>
2 siblings, 1 reply; 127+ messages in thread
From: Amit Kapila @ 2021-12-15 13:20 UTC (permalink / raw)
To: Tomas Vondra <[email protected]>; +Cc: Peter Eisentraut <[email protected]>; PostgreSQL Hackers <[email protected]>; Petr Jelinek <[email protected]>
On Tue, Dec 14, 2021 at 7:02 AM Tomas Vondra
<[email protected]> wrote:
>
> Hi,
>
> here's an updated version of the patches, dealing with almost all of the
> issues (at least in the 0001 and 0002 parts). The main changes:
>
> 1) I've removed the 'created' flag from fill_seq_with_data, as
> discussed. I don't think it's needed by any of the parts (not even 0003,
> AFAICS). We still need it in xl_seq_rec, though.
>
> 2) GetCurrentTransactionId() added to sequence.c are called only with
> wal_level=logical, to minimize the overhead.
>
>
> There's still one remaining problem, that I already explained in [1].
> The problem is that with this:
>
> BEGIN;
> SELECT nextval('s') FROM generate_series(1,100);
> ROLLBACK;
>
>
> The root cause is that pg_current_wal_lsn() uses the LogwrtResult.Write,
> which is updated by XLogFlush() - but only in RecordTransactionCommit.
> Which makes sense, because only the committed stuff is "visible".
>
> But the non-transactional behavior of sequence decoding disagrees with
> this, because now some of the changes from aborted transactions may be
> replicated. Which means the wait_for_catchup() ends up not waiting for
> the sequence change to be replicated. This is an issue for tests in
> patch 0003, at least.
>
> My concern is this actually affects other places waiting for things
> getting replicated :-/
>
By any chance, will this impact synchronous replication as well which
waits for commits to be replicated?
How is this patch dealing with prepared transaction case where at
prepare we will send transactional changes and then later if rollback
prepared happens then the publisher will rollback changes related to
new relfilenode but subscriber would have already replayed the updated
seqval change which won't be rolled back?
--
With Regards,
Amit Kapila.
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2021-12-15 13:44 Peter Eisentraut <[email protected]>
parent: Tomas Vondra <[email protected]>
2 siblings, 0 replies; 127+ messages in thread
From: Peter Eisentraut @ 2021-12-15 13:44 UTC (permalink / raw)
To: Tomas Vondra <[email protected]>; +Cc: PostgreSQL Hackers <[email protected]>; Petr Jelinek <[email protected]>
On 14.12.21 02:31, Tomas Vondra wrote:
> There's still one remaining problem, that I already explained in [1].
> The problem is that with this:
>
> BEGIN;
> SELECT nextval('s') FROM generate_series(1,100);
> ROLLBACK;
>
>
> The root cause is that pg_current_wal_lsn() uses the LogwrtResult.Write,
> which is updated by XLogFlush() - but only in RecordTransactionCommit.
> Which makes sense, because only the committed stuff is "visible".
>
> But the non-transactional behavior of sequence decoding disagrees with
> this, because now some of the changes from aborted transactions may be
> replicated. Which means the wait_for_catchup() ends up not waiting for
> the sequence change to be replicated.
I can't think of a reason why this might be a problem.
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2021-12-15 13:51 Tomas Vondra <[email protected]>
parent: Amit Kapila <[email protected]>
0 siblings, 3 replies; 127+ messages in thread
From: Tomas Vondra @ 2021-12-15 13:51 UTC (permalink / raw)
To: Amit Kapila <[email protected]>; +Cc: Peter Eisentraut <[email protected]>; PostgreSQL Hackers <[email protected]>; Petr Jelinek <[email protected]>
On 12/15/21 14:20, Amit Kapila wrote:
> On Tue, Dec 14, 2021 at 7:02 AM Tomas Vondra
> <[email protected]> wrote:
>>
>> Hi,
>>
>> here's an updated version of the patches, dealing with almost all of the
>> issues (at least in the 0001 and 0002 parts). The main changes:
>>
>> 1) I've removed the 'created' flag from fill_seq_with_data, as
>> discussed. I don't think it's needed by any of the parts (not even 0003,
>> AFAICS). We still need it in xl_seq_rec, though.
>>
>> 2) GetCurrentTransactionId() added to sequence.c are called only with
>> wal_level=logical, to minimize the overhead.
>>
>>
>> There's still one remaining problem, that I already explained in [1].
>> The problem is that with this:
>>
>> BEGIN;
>> SELECT nextval('s') FROM generate_series(1,100);
>> ROLLBACK;
>>
>>
>> The root cause is that pg_current_wal_lsn() uses the LogwrtResult.Write,
>> which is updated by XLogFlush() - but only in RecordTransactionCommit.
>> Which makes sense, because only the committed stuff is "visible".
>>
>> But the non-transactional behavior of sequence decoding disagrees with
>> this, because now some of the changes from aborted transactions may be
>> replicated. Which means the wait_for_catchup() ends up not waiting for
>> the sequence change to be replicated. This is an issue for tests in
>> patch 0003, at least.
>>
>> My concern is this actually affects other places waiting for things
>> getting replicated :-/
>>
>
> By any chance, will this impact synchronous replication as well which
> waits for commits to be replicated?
>
Physical or logical replication? Physical is certainly not replicated.
For logical replication, it's more complicated.
> How is this patch dealing with prepared transaction case where at
> prepare we will send transactional changes and then later if rollback
> prepared happens then the publisher will rollback changes related to
> new relfilenode but subscriber would have already replayed the updated
> seqval change which won't be rolled back?
>
I'm not sure what exact scenario you are describing, but in general we
don't rollback sequence changes anyway, so this should not cause any
divergence between the publisher and subscriber.
Or are you talking about something like ALTER SEQUENCE? I think that
should trigger the transactional behavior for the new relfilenode, so
the subscriber won't see that changes.
regards
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2021-12-15 13:58 Tomas Vondra <[email protected]>
parent: Tomas Vondra <[email protected]>
2 siblings, 0 replies; 127+ messages in thread
From: Tomas Vondra @ 2021-12-15 13:58 UTC (permalink / raw)
To: Amit Kapila <[email protected]>; +Cc: Peter Eisentraut <[email protected]>; PostgreSQL Hackers <[email protected]>; Petr Jelinek <[email protected]>
On 12/15/21 14:51, Tomas Vondra wrote:
> On 12/15/21 14:20, Amit Kapila wrote:
>> On Tue, Dec 14, 2021 at 7:02 AM Tomas Vondra
>> <[email protected]> wrote:
>>>
>>> Hi,
>>>
>>> here's an updated version of the patches, dealing with almost all of the
>>> issues (at least in the 0001 and 0002 parts). The main changes:
>>>
>>> 1) I've removed the 'created' flag from fill_seq_with_data, as
>>> discussed. I don't think it's needed by any of the parts (not even 0003,
>>> AFAICS). We still need it in xl_seq_rec, though.
>>>
>>> 2) GetCurrentTransactionId() added to sequence.c are called only with
>>> wal_level=logical, to minimize the overhead.
>>>
>>>
>>> There's still one remaining problem, that I already explained in [1].
>>> The problem is that with this:
>>>
>>> BEGIN;
>>> SELECT nextval('s') FROM generate_series(1,100);
>>> ROLLBACK;
>>>
>>>
>>> The root cause is that pg_current_wal_lsn() uses the LogwrtResult.Write,
>>> which is updated by XLogFlush() - but only in RecordTransactionCommit.
>>> Which makes sense, because only the committed stuff is "visible".
>>>
>>> But the non-transactional behavior of sequence decoding disagrees with
>>> this, because now some of the changes from aborted transactions may be
>>> replicated. Which means the wait_for_catchup() ends up not waiting for
>>> the sequence change to be replicated. This is an issue for tests in
>>> patch 0003, at least.
>>>
>>> My concern is this actually affects other places waiting for things
>>> getting replicated :-/
>>>
>>
>> By any chance, will this impact synchronous replication as well which
>> waits for commits to be replicated?
>>
>
> Physical or logical replication? Physical is certainly not replicated.
>
> For logical replication, it's more complicated.
>
Apologies, sent too early ... I think it's more complicated for logical
sync replication, because of a scenario like this:
BEGIN;
SELECT nextval('s') FROM generate_series(1,100); <-- writes WAL
ROLLBACK;
SELECT nextval('s');
The first transaction advances the sequence enough to generate a WAL,
which we do every 32 values. But it's rolled back, so it does not update
LogwrtResult.Write, because that happens only at commit.
And then the nextval() generates a value from the sequence without
generating WAL, so it doesn't update the LSN either (IIRC). That'd mean
a sync replication may not wait for this change to reach the subscriber.
regards
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2021-12-15 16:42 Alvaro Herrera <[email protected]>
parent: Tomas Vondra <[email protected]>
2 siblings, 1 reply; 127+ messages in thread
From: Alvaro Herrera @ 2021-12-15 16:42 UTC (permalink / raw)
To: Tomas Vondra <[email protected]>; +Cc: Peter Eisentraut <[email protected]>; PostgreSQL Hackers <[email protected]>; Petr Jelinek <[email protected]>
Looking at 0003,
On 2021-Dec-14, Tomas Vondra wrote:
> diff --git a/doc/src/sgml/ref/alter_publication.sgml b/doc/src/sgml/ref/alter_publication.sgml
> index bb4ef5e5e22..4d166ad3f9c 100644
> --- a/doc/src/sgml/ref/alter_publication.sgml
> +++ b/doc/src/sgml/ref/alter_publication.sgml
> @@ -31,7 +31,9 @@ ALTER PUBLICATION <replaceable class="parameter">name</replaceable> RENAME TO <r
> <phrase>where <replaceable class="parameter">publication_object</replaceable> is one of:</phrase>
>
> TABLE [ ONLY ] <replaceable class="parameter">table_name</replaceable> [ * ] [, ... ]
> + SEQUENCE <replaceable class="parameter">sequence_name</replaceable> [ * ] [, ... ]
> ALL TABLES IN SCHEMA { <replaceable class="parameter">schema_name</replaceable> | CURRENT_SCHEMA } [, ... ]
> + ALL SEQUENCE IN SCHEMA { <replaceable class="parameter">schema_name</replaceable> | CURRENT_SCHEMA } [, ... ]
Note that this says ALL SEQUENCE; I think it should be ALL SEQUENCES.
> diff --git a/src/backend/parser/gram.y b/src/backend/parser/gram.y
> index 3d4dd43e47b..f037c17985b 100644
> --- a/src/backend/parser/gram.y
> +++ b/src/backend/parser/gram.y
> @@ -9762,6 +9762,26 @@ PublicationObjSpec:
...
> + | ALL SEQUENCE IN_P SCHEMA ColId
> + {
> + $$ = makeNode(PublicationObjSpec);
> + $$->pubobjtype = PUBLICATIONOBJ_SEQUENCE_IN_SCHEMA;
> + $$->name = $5;
> + $$->location = @5;
> + }
> + | ALL SEQUENCES IN_P SCHEMA CURRENT_SCHEMA
> + {
> + $$ = makeNode(PublicationObjSpec);
> + $$->pubobjtype = PUBLICATIONOBJ_SEQUENCE_IN_CUR_SCHEMA;
> + $$->location = @5;
> + }
And here you have ALL SEQUENCE in one spot and ALL SEQUENCES in the
other.
BTW I think these enum values should use the plural too,
PUBLICATIONOBJ_SEQUENCES_IN_CUR_SCHEMA (not SEQUENCE). I suppose you
copied from PUBLICATIONOBJ_TABLE_IN_CUR_SCHEMA, but that too seems to be
a mistake: should be PUBLICATIONOBJ_TABLES_IN_CUR_SCHEMA.
> @@ -10097,6 +10117,12 @@ UnlistenStmt:
> }
> ;
>
> +/*
> + * FIXME
> + *
> + * opt_publication_for_sequences and publication_for_sequences should be
> + * copies for sequences
> + */
Not sure if this FIXME is relevant or should just be removed.
> @@ -10105,6 +10131,12 @@ UnlistenStmt:
> * BEGIN / COMMIT / ROLLBACK
> * (also older versions END / ABORT)
> *
> + * ALTER PUBLICATION name ADD SEQUENCE sequence [, sequence2]
> + *
> + * ALTER PUBLICATION name DROP SEQUENCE sequence [, sequence2]
> + *
> + * ALTER PUBLICATION name SET SEQUENCE sequence [, sequence2]
> + *
This comment addition seems misplaced?
> diff --git a/src/bin/psql/tab-complete.c b/src/bin/psql/tab-complete.c
> index 2f412ca3db3..e30bf7b1b55 100644
> --- a/src/bin/psql/tab-complete.c
> +++ b/src/bin/psql/tab-complete.c
> @@ -1647,13 +1647,13 @@ psql_completion(const char *text, int start, int end)
> COMPLETE_WITH("ADD", "DROP", "OWNER TO", "RENAME TO", "SET");
> /* ALTER PUBLICATION <name> ADD */
> else if (Matches("ALTER", "PUBLICATION", MatchAny, "ADD"))
> - COMPLETE_WITH("ALL TABLES IN SCHEMA", "TABLE");
> + COMPLETE_WITH("ALL TABLES IN SCHEMA", "TABLE|SEQUENCE");
> /* ALTER PUBLICATION <name> DROP */
> else if (Matches("ALTER", "PUBLICATION", MatchAny, "DROP"))
> - COMPLETE_WITH("ALL TABLES IN SCHEMA", "TABLE");
> + COMPLETE_WITH("ALL TABLES IN SCHEMA", "TABLE|SEQUENCE");
> /* ALTER PUBLICATION <name> SET */
> else if (Matches("ALTER", "PUBLICATION", MatchAny, "SET"))
> - COMPLETE_WITH("(", "ALL TABLES IN SCHEMA", "TABLE");
> + COMPLETE_WITH("(", "ALL TABLES IN SCHEMA", "TABLE|SEQUENCE");
I think you should also add "ALL SEQUENCES IN SCHEMA" to these lists.
> else if (Matches("ALTER", "PUBLICATION", MatchAny, "ADD|DROP|SET", "ALL", "TABLES", "IN", "SCHEMA"))
... and perhaps make this "ALL", "TABLES|SEQUENCES", "IN", "SCHEMA".
--
Álvaro Herrera PostgreSQL Developer — https://www.EnterpriseDB.com/
"Nunca confiaré en un traidor. Ni siquiera si el traidor lo he creado yo"
(Barón Vladimir Harkonnen)
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2021-12-15 16:56 Tomas Vondra <[email protected]>
parent: Alvaro Herrera <[email protected]>
0 siblings, 0 replies; 127+ messages in thread
From: Tomas Vondra @ 2021-12-15 16:56 UTC (permalink / raw)
To: Alvaro Herrera <[email protected]>; +Cc: Peter Eisentraut <[email protected]>; PostgreSQL Hackers <[email protected]>; Petr Jelinek <[email protected]>
On 12/15/21 17:42, Alvaro Herrera wrote:
> Looking at 0003,
>
> On 2021-Dec-14, Tomas Vondra wrote:
>
>> diff --git a/doc/src/sgml/ref/alter_publication.sgml b/doc/src/sgml/ref/alter_publication.sgml
>> index bb4ef5e5e22..4d166ad3f9c 100644
>> --- a/doc/src/sgml/ref/alter_publication.sgml
>> +++ b/doc/src/sgml/ref/alter_publication.sgml
>> @@ -31,7 +31,9 @@ ALTER PUBLICATION <replaceable class="parameter">name</replaceable> RENAME TO <r
>> <phrase>where <replaceable class="parameter">publication_object</replaceable> is one of:</phrase>
>>
>> TABLE [ ONLY ] <replaceable class="parameter">table_name</replaceable> [ * ] [, ... ]
>> + SEQUENCE <replaceable class="parameter">sequence_name</replaceable> [ * ] [, ... ]
>> ALL TABLES IN SCHEMA { <replaceable class="parameter">schema_name</replaceable> | CURRENT_SCHEMA } [, ... ]
>> + ALL SEQUENCE IN SCHEMA { <replaceable class="parameter">schema_name</replaceable> | CURRENT_SCHEMA } [, ... ]
>
> Note that this says ALL SEQUENCE; I think it should be ALL SEQUENCES.
>
>> diff --git a/src/backend/parser/gram.y b/src/backend/parser/gram.y
>> index 3d4dd43e47b..f037c17985b 100644
>> --- a/src/backend/parser/gram.y
>> +++ b/src/backend/parser/gram.y
>> @@ -9762,6 +9762,26 @@ PublicationObjSpec:
> ...
>> + | ALL SEQUENCE IN_P SCHEMA ColId
>> + {
>> + $$ = makeNode(PublicationObjSpec);
>> + $$->pubobjtype = PUBLICATIONOBJ_SEQUENCE_IN_SCHEMA;
>> + $$->name = $5;
>> + $$->location = @5;
>> + }
>> + | ALL SEQUENCES IN_P SCHEMA CURRENT_SCHEMA
>> + {
>> + $$ = makeNode(PublicationObjSpec);
>> + $$->pubobjtype = PUBLICATIONOBJ_SEQUENCE_IN_CUR_SCHEMA;
>> + $$->location = @5;
>> + }
>
> And here you have ALL SEQUENCE in one spot and ALL SEQUENCES in the
> other.
>
> BTW I think these enum values should use the plural too,
> PUBLICATIONOBJ_SEQUENCES_IN_CUR_SCHEMA (not SEQUENCE). I suppose you
> copied from PUBLICATIONOBJ_TABLE_IN_CUR_SCHEMA, but that too seems to be
> a mistake: should be PUBLICATIONOBJ_TABLES_IN_CUR_SCHEMA.
>
>
>> @@ -10097,6 +10117,12 @@ UnlistenStmt:
>> }
>> ;
>>
>> +/*
>> + * FIXME
>> + *
>> + * opt_publication_for_sequences and publication_for_sequences should be
>> + * copies for sequences
>> + */
>
> Not sure if this FIXME is relevant or should just be removed.
>
>> @@ -10105,6 +10131,12 @@ UnlistenStmt:
>> * BEGIN / COMMIT / ROLLBACK
>> * (also older versions END / ABORT)
>> *
>> + * ALTER PUBLICATION name ADD SEQUENCE sequence [, sequence2]
>> + *
>> + * ALTER PUBLICATION name DROP SEQUENCE sequence [, sequence2]
>> + *
>> + * ALTER PUBLICATION name SET SEQUENCE sequence [, sequence2]
>> + *
>
> This comment addition seems misplaced?
>
>> diff --git a/src/bin/psql/tab-complete.c b/src/bin/psql/tab-complete.c
>> index 2f412ca3db3..e30bf7b1b55 100644
>> --- a/src/bin/psql/tab-complete.c
>> +++ b/src/bin/psql/tab-complete.c
>> @@ -1647,13 +1647,13 @@ psql_completion(const char *text, int start, int end)
>> COMPLETE_WITH("ADD", "DROP", "OWNER TO", "RENAME TO", "SET");
>> /* ALTER PUBLICATION <name> ADD */
>> else if (Matches("ALTER", "PUBLICATION", MatchAny, "ADD"))
>> - COMPLETE_WITH("ALL TABLES IN SCHEMA", "TABLE");
>> + COMPLETE_WITH("ALL TABLES IN SCHEMA", "TABLE|SEQUENCE");
>> /* ALTER PUBLICATION <name> DROP */
>> else if (Matches("ALTER", "PUBLICATION", MatchAny, "DROP"))
>> - COMPLETE_WITH("ALL TABLES IN SCHEMA", "TABLE");
>> + COMPLETE_WITH("ALL TABLES IN SCHEMA", "TABLE|SEQUENCE");
>> /* ALTER PUBLICATION <name> SET */
>> else if (Matches("ALTER", "PUBLICATION", MatchAny, "SET"))
>> - COMPLETE_WITH("(", "ALL TABLES IN SCHEMA", "TABLE");
>> + COMPLETE_WITH("(", "ALL TABLES IN SCHEMA", "TABLE|SEQUENCE");
>
> I think you should also add "ALL SEQUENCES IN SCHEMA" to these lists.
>
>
>> else if (Matches("ALTER", "PUBLICATION", MatchAny, "ADD|DROP|SET", "ALL", "TABLES", "IN", "SCHEMA"))
>
> ... and perhaps make this "ALL", "TABLES|SEQUENCES", "IN", "SCHEMA".
>
Thanks for the review. I'm aware 0003 is still incomplete and subject to
change - it's certainly not meant for commit yet. The current 0003 patch
is sufficient for testing the infrastructure, but we need to figure out
how to make it easier to use, what to do with implicit sequences and
similar things. Peter had some ideas in [1].
[1]
https://www.postgresql.org/message-id/359bf6d0-413d-292a-4305-e99eeafead39%40enterprisedb.com
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2021-12-16 11:59 Amit Kapila <[email protected]>
parent: Tomas Vondra <[email protected]>
2 siblings, 1 reply; 127+ messages in thread
From: Amit Kapila @ 2021-12-16 11:59 UTC (permalink / raw)
To: Tomas Vondra <[email protected]>; +Cc: Peter Eisentraut <[email protected]>; PostgreSQL Hackers <[email protected]>; Petr Jelinek <[email protected]>
On Wed, Dec 15, 2021 at 7:21 PM Tomas Vondra
<[email protected]> wrote:
>
> On 12/15/21 14:20, Amit Kapila wrote:
> > On Tue, Dec 14, 2021 at 7:02 AM Tomas Vondra
> > <[email protected]> wrote:
> >>
> >> Hi,
> >>
> >> here's an updated version of the patches, dealing with almost all of the
> >> issues (at least in the 0001 and 0002 parts). The main changes:
> >>
> >> 1) I've removed the 'created' flag from fill_seq_with_data, as
> >> discussed. I don't think it's needed by any of the parts (not even 0003,
> >> AFAICS). We still need it in xl_seq_rec, though.
> >>
> >> 2) GetCurrentTransactionId() added to sequence.c are called only with
> >> wal_level=logical, to minimize the overhead.
> >>
> >>
> >> There's still one remaining problem, that I already explained in [1].
> >> The problem is that with this:
> >>
> >> BEGIN;
> >> SELECT nextval('s') FROM generate_series(1,100);
> >> ROLLBACK;
> >>
> >>
> >> The root cause is that pg_current_wal_lsn() uses the LogwrtResult.Write,
> >> which is updated by XLogFlush() - but only in RecordTransactionCommit.
> >> Which makes sense, because only the committed stuff is "visible".
> >>
> >> But the non-transactional behavior of sequence decoding disagrees with
> >> this, because now some of the changes from aborted transactions may be
> >> replicated. Which means the wait_for_catchup() ends up not waiting for
> >> the sequence change to be replicated. This is an issue for tests in
> >> patch 0003, at least.
> >>
> >> My concern is this actually affects other places waiting for things
> >> getting replicated :-/
> >>
> >
> > By any chance, will this impact synchronous replication as well which
> > waits for commits to be replicated?
> >
>
> Physical or logical replication?
>
logical replication.
> Physical is certainly not replicated.
>
> For logical replication, it's more complicated.
>
> > How is this patch dealing with prepared transaction case where at
> > prepare we will send transactional changes and then later if rollback
> > prepared happens then the publisher will rollback changes related to
> > new relfilenode but subscriber would have already replayed the updated
> > seqval change which won't be rolled back?
> >
>
> I'm not sure what exact scenario you are describing, but in general we
> don't rollback sequence changes anyway, so this should not cause any
> divergence between the publisher and subscriber.
>
> Or are you talking about something like ALTER SEQUENCE? I think that
> should trigger the transactional behavior for the new relfilenode, so
> the subscriber won't see that changes.
>
Yeah, something like Alter Sequence and nextval combination. I see
that it will be handled because of the transactional behavior for the
new relfilenode as for applying each sequence change, a new
relfilenode is created. I think on apply side, the patch always
creates a new relfilenode irrespective of whether the sequence message
is transactional or not. Do we need to do that for the
non-transactional messages as well?
--
With Regards,
Amit Kapila.
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2021-12-16 14:54 Tomas Vondra <[email protected]>
parent: Amit Kapila <[email protected]>
0 siblings, 0 replies; 127+ messages in thread
From: Tomas Vondra @ 2021-12-16 14:54 UTC (permalink / raw)
To: Amit Kapila <[email protected]>; +Cc: Peter Eisentraut <[email protected]>; PostgreSQL Hackers <[email protected]>; Petr Jelinek <[email protected]>
On 12/16/21 12:59, Amit Kapila wrote:
> On Wed, Dec 15, 2021 at 7:21 PM Tomas Vondra
> <[email protected]> wrote:
>>
>> On 12/15/21 14:20, Amit Kapila wrote:
>>> On Tue, Dec 14, 2021 at 7:02 AM Tomas Vondra
>>> <[email protected]> wrote:
>>>>
>>>> Hi,
>>>>
>>>> here's an updated version of the patches, dealing with almost all of the
>>>> issues (at least in the 0001 and 0002 parts). The main changes:
>>>>
>>>> 1) I've removed the 'created' flag from fill_seq_with_data, as
>>>> discussed. I don't think it's needed by any of the parts (not even 0003,
>>>> AFAICS). We still need it in xl_seq_rec, though.
>>>>
>>>> 2) GetCurrentTransactionId() added to sequence.c are called only with
>>>> wal_level=logical, to minimize the overhead.
>>>>
>>>>
>>>> There's still one remaining problem, that I already explained in [1].
>>>> The problem is that with this:
>>>>
>>>> BEGIN;
>>>> SELECT nextval('s') FROM generate_series(1,100);
>>>> ROLLBACK;
>>>>
>>>>
>>>> The root cause is that pg_current_wal_lsn() uses the LogwrtResult.Write,
>>>> which is updated by XLogFlush() - but only in RecordTransactionCommit.
>>>> Which makes sense, because only the committed stuff is "visible".
>>>>
>>>> But the non-transactional behavior of sequence decoding disagrees with
>>>> this, because now some of the changes from aborted transactions may be
>>>> replicated. Which means the wait_for_catchup() ends up not waiting for
>>>> the sequence change to be replicated. This is an issue for tests in
>>>> patch 0003, at least.
>>>>
>>>> My concern is this actually affects other places waiting for things
>>>> getting replicated :-/
>>>>
>>>
>>> By any chance, will this impact synchronous replication as well which
>>> waits for commits to be replicated?
>>>
>>
>> Physical or logical replication?
>>
>
> logical replication.
>
>> Physical is certainly not replicated.
>>
>> For logical replication, it's more complicated.
>>
>>> How is this patch dealing with prepared transaction case where at
>>> prepare we will send transactional changes and then later if rollback
>>> prepared happens then the publisher will rollback changes related to
>>> new relfilenode but subscriber would have already replayed the updated
>>> seqval change which won't be rolled back?
>>>
>>
>> I'm not sure what exact scenario you are describing, but in general we
>> don't rollback sequence changes anyway, so this should not cause any
>> divergence between the publisher and subscriber.
>>
>> Or are you talking about something like ALTER SEQUENCE? I think that
>> should trigger the transactional behavior for the new relfilenode, so
>> the subscriber won't see that changes.
>>
>
> Yeah, something like Alter Sequence and nextval combination. I see
> that it will be handled because of the transactional behavior for the
> new relfilenode as for applying each sequence change, a new
> relfilenode is created.
Right.
> I think on apply side, the patch always creates a new relfilenode
> irrespective of whether the sequence message is transactional or not.
> Do we need to do that for the non-transactional messages as well?
>
Good question. I don't think that's necessary, I'll see if we can simply
update the tuple (mostly just like redo).
regards
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2021-12-18 01:56 Tomas Vondra <[email protected]>
parent: Tomas Vondra <[email protected]>
2 siblings, 1 reply; 127+ messages in thread
From: Tomas Vondra @ 2021-12-18 01:56 UTC (permalink / raw)
To: Amit Kapila <[email protected]>; +Cc: Peter Eisentraut <[email protected]>; PostgreSQL Hackers <[email protected]>; Petr Jelinek <[email protected]>
On 12/15/21 14:51, Tomas Vondra wrote:
> On 12/15/21 14:20, Amit Kapila wrote:
>> On Tue, Dec 14, 2021 at 7:02 AM Tomas Vondra
>> <[email protected]> wrote:
>>>
>>> Hi,
>>>
>>> here's an updated version of the patches, dealing with almost all of the
>>> issues (at least in the 0001 and 0002 parts). The main changes:
>>>
>>> 1) I've removed the 'created' flag from fill_seq_with_data, as
>>> discussed. I don't think it's needed by any of the parts (not even 0003,
>>> AFAICS). We still need it in xl_seq_rec, though.
>>>
>>> 2) GetCurrentTransactionId() added to sequence.c are called only with
>>> wal_level=logical, to minimize the overhead.
>>>
>>>
>>> There's still one remaining problem, that I already explained in [1].
>>> The problem is that with this:
>>>
>>> BEGIN;
>>> SELECT nextval('s') FROM generate_series(1,100);
>>> ROLLBACK;
>>>
>>>
>>> The root cause is that pg_current_wal_lsn() uses the LogwrtResult.Write,
>>> which is updated by XLogFlush() - but only in RecordTransactionCommit.
>>> Which makes sense, because only the committed stuff is "visible".
>>>
>>> But the non-transactional behavior of sequence decoding disagrees with
>>> this, because now some of the changes from aborted transactions may be
>>> replicated. Which means the wait_for_catchup() ends up not waiting for
>>> the sequence change to be replicated. This is an issue for tests in
>>> patch 0003, at least.
>>>
>>> My concern is this actually affects other places waiting for things
>>> getting replicated :-/
>>>
>>
>> By any chance, will this impact synchronous replication as well which
>> waits for commits to be replicated?
>>
>
> Physical or logical replication? Physical is certainly not replicated.
>
Actually, I take that back. It does affect physical (sync) replication
just as well, and I think it might be considered a data loss issue. I
started a new thread to discuss that, so that it's not buried in this
thread about logical replication.
regards
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2021-12-22 15:40 Tomas Vondra <[email protected]>
parent: Tomas Vondra <[email protected]>
0 siblings, 0 replies; 127+ messages in thread
From: Tomas Vondra @ 2021-12-22 15:40 UTC (permalink / raw)
To: Amit Kapila <[email protected]>; +Cc: Peter Eisentraut <[email protected]>; PostgreSQL Hackers <[email protected]>; Petr Jelinek <[email protected]>
Hi,
Here's an updated version of the patch series. The primary change is
tweaking the WAL-logging of sequences modified per [1]. This changes
test output in test_decoding and built-in replication patches, and to
make it clearer I left the changes in separate patches.
Assuming the WAL logging changes are acceptable, that resolves the data
loss issue.
I'm wondering what to do about changes with is_called=false, i.e.
changes generated by ALTER SEQUENCE etc. The current patch does decode
them and passes them to the output plugin, but I'm starting to think
that may not be the right behavior - if we haven't generated any data
from the sequence, there's no point in replicating that, I think.
regards
[1]
https://www.postgresql.org/message-id/712cad46-a9c8-1389-aef8-faf0203c9be9%40enterprisedb.com
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
Attachments:
[text/x-patch] 0001-WAL-log-individual-sequence-fetches-20211222.patch (1.9K, ../../[email protected]/2-0001-WAL-log-individual-sequence-fetches-20211222.patch)
download | inline diff:
From eeaa7cb36c69af048f0321e4883864ebe2542429 Mon Sep 17 00:00:00 2001
From: Tomas Vondra <[email protected]>
Date: Wed, 22 Dec 2021 03:18:46 +0100
Subject: [PATCH 1/6] WAL-log individual sequence fetches
---
src/backend/commands/sequence.c | 18 ++++++++++++++++--
1 file changed, 16 insertions(+), 2 deletions(-)
diff --git a/src/backend/commands/sequence.c b/src/backend/commands/sequence.c
index 72bfdc07a4..0f309d0a4e 100644
--- a/src/backend/commands/sequence.c
+++ b/src/backend/commands/sequence.c
@@ -52,6 +52,9 @@
* We don't want to log each fetching of a value from a sequence,
* so we pre-log a few fetches in advance. In the event of
* crash we can lose (skip over) as many values as we pre-logged.
+ *
+ * We only pre-log fetches in wal_level=minimal. For higher levels we
+ * WAL-log every individual sequence increment, as if this was 0.
*/
#define SEQ_LOG_VALS 32
@@ -666,11 +669,18 @@ nextval_internal(Oid relid, bool check_permissions)
* WAL record to be written anyway, else replay starting from the
* checkpoint would fail to advance the sequence past the logged values.
* In this case we may as well fetch extra values.
+ *
+ * We only pre-log fetches in wal_level=minimal. For higher levels we
+ * WAL-log every individual sequence increment.
*/
if (log < fetch || !seq->is_called)
{
/* forced log to satisfy local demand for values */
- fetch = log = fetch + SEQ_LOG_VALS;
+ if (XLogIsNeeded())
+ fetch = log = fetch;
+ else
+ fetch = log = fetch + SEQ_LOG_VALS;
+
logit = true;
}
else
@@ -680,7 +690,11 @@ nextval_internal(Oid relid, bool check_permissions)
if (PageGetLSN(page) <= redoptr)
{
/* last update of seq was before checkpoint */
- fetch = log = fetch + SEQ_LOG_VALS;
+ if (XLogIsNeeded())
+ fetch = log = fetch;
+ else
+ fetch = log = fetch + SEQ_LOG_VALS;
+
logit = true;
}
}
--
2.31.1
[text/x-patch] 0002-Logical-decoding-of-sequences-20211222.patch (45.1K, ../../[email protected]/3-0002-Logical-decoding-of-sequences-20211222.patch)
download | inline diff:
From b5da8cbba4a10db4ff4b7cdc7ed6555bc88d6232 Mon Sep 17 00:00:00 2001
From: Tomas Vondra <[email protected]>
Date: Fri, 24 Sep 2021 00:41:33 +0200
Subject: [PATCH 2/6] Logical decoding of sequences
This adds the infrastructure for logical decoding of sequences. Support
for built-in logical replication and test_decoding is added separately.
When decoding sequences, we differentiate between a sequences created in
a (running) transaction, and sequences created in other (already
committed) transactions. Changes for sequences created in the same
top-level transaction are treated as "transactional" i.e. just like any
other change from that transaction (and discarded in case of a
rollback). Changes for sequences created earlier are treated as not
transactional - are processed immediately, as if performed outside any
transaction (and thus not rolled back).
This mixed behavior is necessary - sequences are non-transactional (e.g.
ROLLBACK does not undo the sequence increments). But for new sequences,
we need to handle them in a transactional way, because if we ever get
some DDL support, the sequence won't exist until the transaction gets
applied. So we need to ensure the increments don't happen until the
sequence gets created.
To differentiate which sequences are "old" and which were created in a
still-running transaction, we track sequences created in running
transactions in a hash table. Each sequence is identified by it's
relfilenode, and at transaction commit we discard all sequences created
in that particular transaction. For each sequence we track the XID of
the (sub)transaction that created it, and we cleanup the sequences for
each subtransaction when it completes (commit/rollback).
We don't use the XID to check if it's the same top-level transaction.
It's enough to know it was created in an in-progress transaction, and we
know it must be the current one because otherwise it wouldn't see the
sequence object.
The main changes in this patch are:
1) ensure WAL-logging of all necessary info for sequence advances
We need to be able to associate the advance with a XID, but until now
sequence advance might have XID 0 if it was the first thing the
transaction did. So we ensure the transaction has XID.
Note: This is needed because of subxacts. A XID 0 might still have the
sequence created in a different subxact of the same top-level xact.
Then there's a "created" flag added to the xl_seq_rec, which is
necessary to differentiate WAL for created/reset sequences.
2) decoding / queueing of sequences into ReorderBuffer
This is mostly copy-paste of existing code for other decoded events.
3) tracking sequences created in in-progress transactions
We use a simple hash table, indexed by relfilenode.
---
doc/src/sgml/logicaldecoding.sgml | 62 ++-
src/backend/commands/sequence.c | 32 ++
src/backend/replication/logical/decode.c | 131 +++++-
src/backend/replication/logical/logical.c | 88 ++++
.../replication/logical/reorderbuffer.c | 400 ++++++++++++++++++
src/include/commands/sequence.h | 1 +
src/include/replication/output_plugin.h | 27 ++
src/include/replication/reorderbuffer.h | 43 +-
8 files changed, 778 insertions(+), 6 deletions(-)
diff --git a/doc/src/sgml/logicaldecoding.sgml b/doc/src/sgml/logicaldecoding.sgml
index b6353c7a12..60a2b7d1cc 100644
--- a/doc/src/sgml/logicaldecoding.sgml
+++ b/doc/src/sgml/logicaldecoding.sgml
@@ -458,6 +458,7 @@ typedef struct OutputPluginCallbacks
LogicalDecodeTruncateCB truncate_cb;
LogicalDecodeCommitCB commit_cb;
LogicalDecodeMessageCB message_cb;
+ LogicalDecodeSequenceCB sequence_cb;
LogicalDecodeFilterByOriginCB filter_by_origin_cb;
LogicalDecodeShutdownCB shutdown_cb;
LogicalDecodeFilterPrepareCB filter_prepare_cb;
@@ -472,6 +473,7 @@ typedef struct OutputPluginCallbacks
LogicalDecodeStreamCommitCB stream_commit_cb;
LogicalDecodeStreamChangeCB stream_change_cb;
LogicalDecodeStreamMessageCB stream_message_cb;
+ LogicalDecodeStreamSequenceCB stream_sequence_cb;
LogicalDecodeStreamTruncateCB stream_truncate_cb;
} OutputPluginCallbacks;
@@ -481,9 +483,12 @@ typedef void (*LogicalOutputPluginInit) (struct OutputPluginCallbacks *cb);
and <function>commit_cb</function> callbacks are required,
while <function>startup_cb</function>,
<function>filter_by_origin_cb</function>, <function>truncate_cb</function>,
+ <function>sequence_cb</function>,
and <function>shutdown_cb</function> are optional.
If <function>truncate_cb</function> is not set but a
<command>TRUNCATE</command> is to be decoded, the action will be ignored.
+ Similarly, if <function>sequence_cb</function> is not set and a sequence
+ change is to be decoded, the action will be ignored.
</para>
<para>
@@ -492,7 +497,8 @@ typedef void (*LogicalOutputPluginInit) (struct OutputPluginCallbacks *cb);
<function>stream_stop_cb</function>, <function>stream_abort_cb</function>,
<function>stream_commit_cb</function>, <function>stream_change_cb</function>,
and <function>stream_prepare_cb</function>
- are required, while <function>stream_message_cb</function> and
+ are required, while <function>stream_message_cb</function>,
+ <function>stream_sequence_cb</function>, and
<function>stream_truncate_cb</function> are optional.
</para>
@@ -808,6 +814,35 @@ typedef void (*LogicalDecodeMessageCB) (struct LogicalDecodingContext *ctx,
</para>
</sect3>
+ <sect3 id="logicaldecoding-output-plugin-sequence">
+ <title>Sequence Callback</title>
+
+ <para>
+ The optional <function>sequence_cb</function> callback is called for
+ actions that update a sequence value.
+<programlisting>
+typedef void (*LogicalDecodeSequenceCB) (struct LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional,
+ int64 last_value,
+ int64 log_cnt,
+ bool is_called);
+</programlisting>
+ The <parameter>txn</parameter> parameter contains meta information about
+ the transaction the sequence change is part of. Note however that it can
+ be NULL when the sequence change is non-transactional and the XID was not
+ assigned yet in the transaction which updated the sequence value.
+ The <parameter>sequence_lsn</parameter> has WAL location of the sequence
+ update. The <parameter>transactional</parameter> says if the sequence has
+ to be replayed as part of the transaction or directly.
+
+ The <parameter>last_value</parameter>, <parameter>log_cnt</parameter> and
+ <parameter>is_called</parameter> parameters describe the sequence change.
+ </para>
+ </sect3>
+
<sect3 id="logicaldecoding-output-plugin-filter-prepare">
<title>Prepare Filter Callback</title>
@@ -1017,6 +1052,26 @@ typedef void (*LogicalDecodeStreamMessageCB) (struct LogicalDecodingContext *ctx
</para>
</sect3>
+ <sect3 id="logicaldecoding-output-plugin-stream-sequence">
+ <title>Stream Sequence Callback</title>
+ <para>
+ The optional <function>stream_sequence_cb</function> callback is called
+ for actions that change a sequence in a block of streamed changes
+ (demarcated by <function>stream_start_cb</function> and
+ <function>stream_stop_cb</function> calls).
+<programlisting>
+typedef void (*LogicalDecodeStreamSequenceCB) (struct LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional,
+ int64 last_value,
+ int64 log_cnt,
+ bool is_called);
+</programlisting>
+ </para>
+ </sect3>
+
<sect3 id="logicaldecoding-output-plugin-stream-truncate">
<title>Stream Truncate Callback</title>
<para>
@@ -1197,8 +1252,9 @@ OutputPluginWrite(ctx, true);
in-progress transactions. There are multiple required streaming callbacks
(<function>stream_start_cb</function>, <function>stream_stop_cb</function>,
<function>stream_abort_cb</function>, <function>stream_commit_cb</function>
- and <function>stream_change_cb</function>) and two optional callbacks
- (<function>stream_message_cb</function> and <function>stream_truncate_cb</function>).
+ and <function>stream_change_cb</function>) and multiple optional callbacks
+ (<function>stream_message_cb</function>, <function>stream_sequence_cb</function>,
+ and <function>stream_truncate_cb</function>).
Also, if streaming of two-phase commands is to be supported, then additional
callbacks must be provided. (See <xref linkend="logicaldecoding-two-phase-commits"/>
for details).
diff --git a/src/backend/commands/sequence.c b/src/backend/commands/sequence.c
index 0f309d0a4e..a5435910a2 100644
--- a/src/backend/commands/sequence.c
+++ b/src/backend/commands/sequence.c
@@ -381,8 +381,13 @@ fill_seq_with_data(Relation rel, HeapTuple tuple)
/* check the comment above nextval_internal()'s equivalent call. */
if (RelationNeedsWAL(rel))
+ {
GetTopTransactionId();
+ if (XLogLogicalInfoActive())
+ GetCurrentTransactionId();
+ }
+
START_CRIT_SECTION();
MarkBufferDirty(buf);
@@ -402,6 +407,7 @@ fill_seq_with_data(Relation rel, HeapTuple tuple)
XLogRegisterBuffer(0, buf, REGBUF_WILL_INIT);
xlrec.node = rel->rd_node;
+ xlrec.created = true;
XLogRegisterData((char *) &xlrec, sizeof(xl_seq_rec));
XLogRegisterData((char *) tuple->t_data, tuple->t_len);
@@ -778,10 +784,28 @@ nextval_internal(Oid relid, bool check_permissions)
* It's sufficient to ensure the toplevel transaction has an xid, no need
* to assign xids subxacts, that'll already trigger an appropriate wait.
* (Have to do that here, so we're outside the critical section)
+ *
+ * We have to ensure we have a proper XID, which will be included in
+ * the XLOG record by XLogRecordAssemble. Otherwise the first nextval()
+ * in a subxact (without any preceding changes) would get XID 0, and it
+ * would then be impossible to decide which top xact it belongs to.
+ * It'd also trigger assert in DecodeSequence. We only do that with
+ * wal_level=logical, though.
+ *
+ * XXX This might seem unnecessary, because if there's no XID the xact
+ * couldn't have done anything important yet, e.g. it could not have
+ * created a sequence. But that's incorrect, as it ignores subxacts. The
+ * current subtransaction might not have done anything yet (thus no XID),
+ * but an earlier one might have created the sequence.
*/
if (logit && RelationNeedsWAL(seqrel))
+ {
GetTopTransactionId();
+ if (XLogLogicalInfoActive())
+ GetCurrentTransactionId();
+ }
+
/* ready to change the on-disk (or really, in-buffer) tuple */
START_CRIT_SECTION();
@@ -817,6 +841,7 @@ nextval_internal(Oid relid, bool check_permissions)
seq->log_cnt = 0;
xlrec.node = seqrel->rd_node;
+ xlrec.created = false;
XLogRegisterData((char *) &xlrec, sizeof(xl_seq_rec));
XLogRegisterData((char *) seqdatatuple.t_data, seqdatatuple.t_len);
@@ -991,8 +1016,13 @@ do_setval(Oid relid, int64 next, bool iscalled)
/* check the comment above nextval_internal()'s equivalent call. */
if (RelationNeedsWAL(seqrel))
+ {
GetTopTransactionId();
+ if (XLogLogicalInfoActive())
+ GetCurrentTransactionId();
+ }
+
/* ready to change the on-disk (or really, in-buffer) tuple */
START_CRIT_SECTION();
@@ -1013,6 +1043,8 @@ do_setval(Oid relid, int64 next, bool iscalled)
XLogRegisterBuffer(0, buf, REGBUF_WILL_INIT);
xlrec.node = seqrel->rd_node;
+ xlrec.created = false;
+
XLogRegisterData((char *) &xlrec, sizeof(xl_seq_rec));
XLogRegisterData((char *) seqdatatuple.t_data, seqdatatuple.t_len);
diff --git a/src/backend/replication/logical/decode.c b/src/backend/replication/logical/decode.c
index 59aed6cee6..6965be53ca 100644
--- a/src/backend/replication/logical/decode.c
+++ b/src/backend/replication/logical/decode.c
@@ -42,6 +42,7 @@
#include "replication/reorderbuffer.h"
#include "replication/snapbuild.h"
#include "storage/standby.h"
+#include "commands/sequence.h"
typedef struct XLogRecordBuffer
{
@@ -74,10 +75,11 @@ static void DecodeAbort(LogicalDecodingContext *ctx, XLogRecordBuffer *buf,
bool two_phase);
static void DecodePrepare(LogicalDecodingContext *ctx, XLogRecordBuffer *buf,
xl_xact_parsed_prepare *parsed);
-
+static void DecodeSequence(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
/* common function to decode tuples */
static void DecodeXLogTuple(char *data, Size len, ReorderBufferTupleBuf *tup);
+static void DecodeSeqTuple(char *data, Size len, ReorderBufferTupleBuf *tuple);
/* helper functions for decoding transactions */
static inline bool FilterPrepare(LogicalDecodingContext *ctx,
@@ -158,6 +160,10 @@ LogicalDecodingProcessRecord(LogicalDecodingContext *ctx, XLogReaderState *recor
DecodeLogicalMsgOp(ctx, &buf);
break;
+ case RM_SEQ_ID:
+ DecodeSequence(ctx, &buf);
+ break;
+
/*
* Rmgrs irrelevant for logical decoding; they describe stuff not
* represented in logical decoding. Add new rmgrs in rmgrlist.h's
@@ -173,7 +179,6 @@ LogicalDecodingProcessRecord(LogicalDecodingContext *ctx, XLogReaderState *recor
case RM_HASH_ID:
case RM_GIN_ID:
case RM_GIST_ID:
- case RM_SEQ_ID:
case RM_SPGIST_ID:
case RM_BRIN_ID:
case RM_COMMIT_TS_ID:
@@ -1313,3 +1318,125 @@ DecodeTXNNeedSkip(LogicalDecodingContext *ctx, XLogRecordBuffer *buf,
(txn_dbid != InvalidOid && txn_dbid != ctx->slot->data.database) ||
ctx->fast_forward || FilterByOrigin(ctx, origin_id));
}
+
+/*
+ * Decode Sequence Tuple
+ */
+static void
+DecodeSeqTuple(char *data, Size len, ReorderBufferTupleBuf *tuple)
+{
+ int datalen = len - sizeof(xl_seq_rec) - SizeofHeapTupleHeader;
+
+ Assert(datalen >= 0);
+
+ tuple->tuple.t_len = datalen + SizeofHeapTupleHeader;;
+
+ ItemPointerSetInvalid(&tuple->tuple.t_self);
+
+ tuple->tuple.t_tableOid = InvalidOid;
+
+ memcpy(((char *) tuple->tuple.t_data),
+ data + sizeof(xl_seq_rec),
+ SizeofHeapTupleHeader);
+
+ memcpy(((char *) tuple->tuple.t_data) + SizeofHeapTupleHeader,
+ data + sizeof(xl_seq_rec) + SizeofHeapTupleHeader,
+ datalen);
+}
+
+/*
+ * Handle sequence decode
+ *
+ * Decoding sequences is a bit tricky, because while most sequence actions
+ * are non-transactional (not subject to rollback), some need to be handled
+ * as transactional.
+ *
+ * By default, a sequence increment is non-transactional - we must not queue
+ * it in a transaction as other changes, because the transaction might get
+ * rolled back and we'd discard the increment. The downstream would not be
+ * notified about the increment, which is wrong.
+ *
+ * On the other hand, the sequence may be created in a transaction. In this
+ * case we *should* queue the change as other changes in the transaction,
+ * because we don't want to send the increments for unknown sequence to the
+ * plugin - it might get confused about which sequence it's related to etc.
+ */
+static void
+DecodeSequence(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
+{
+ SnapBuild *builder = ctx->snapshot_builder;
+ ReorderBufferTupleBuf *tuplebuf;
+ RelFileNode target_node;
+ XLogReaderState *r = buf->record;
+ char *tupledata = NULL;
+ Size tuplelen;
+ Size datalen = 0;
+ TransactionId xid = XLogRecGetXid(r);
+ uint8 info = XLogRecGetInfo(buf->record) & ~XLR_INFO_MASK;
+ xl_seq_rec *xlrec;
+ Snapshot snapshot;
+ RepOriginId origin_id = XLogRecGetOrigin(r);
+ bool transactional;
+
+ /* only decode changes flagged with XLOG_SEQ_LOG */
+ if (info != XLOG_SEQ_LOG)
+ elog(ERROR, "unexpected RM_SEQ_ID record type: %u", info);
+
+ /*
+ * If we don't have snapshot or we are just fast-forwarding, there is no
+ * point in decoding messages.
+ */
+ if (SnapBuildCurrentState(builder) < SNAPBUILD_FULL_SNAPSHOT ||
+ ctx->fast_forward)
+ return;
+
+ /* only interested in our database */
+ XLogRecGetBlockTag(r, 0, &target_node, NULL, NULL);
+ if (target_node.dbNode != ctx->slot->data.database)
+ return;
+
+ /* output plugin doesn't look for this origin, no need to queue */
+ if (FilterByOrigin(ctx, XLogRecGetOrigin(r)))
+ return;
+
+ tupledata = XLogRecGetData(r);
+ datalen = XLogRecGetDataLen(r);
+ tuplelen = datalen - SizeOfHeapHeader - sizeof(xl_seq_rec);
+
+ /* extract the WAL record, with "created" flag */
+ xlrec = (xl_seq_rec *) XLogRecGetData(r);
+
+ /* XXX how could we have sequence change without data? */
+ if(!datalen || !tupledata)
+ return;
+
+ tuplebuf = ReorderBufferGetTupleBuf(ctx->reorder, tuplelen);
+ DecodeSeqTuple(tupledata, datalen, tuplebuf);
+
+ /*
+ * Should we handle the sequence increment as transactional or not?
+ *
+ * If the sequence was created in a still-running transaction, treat
+ * it as transactional and queue the increments. Otherwise it needs
+ * to be treated as non-transactional, in which case we send it to
+ * the plugin right away.
+ */
+ transactional = ReorderBufferSequenceIsTransactional(ctx->reorder,
+ target_node,
+ xlrec->created);
+
+ /* Skip the change if already processed (per the snapshot). */
+ if (transactional &&
+ !SnapBuildProcessChange(builder, xid, buf->origptr))
+ return;
+ else if (!transactional &&
+ (SnapBuildCurrentState(builder) != SNAPBUILD_CONSISTENT ||
+ SnapBuildXactNeedsSkip(builder, buf->origptr)))
+ return;
+
+ /* Queue the increment (or send immediately if not transactional). */
+ snapshot = SnapBuildGetOrBuildSnapshot(builder, xid);
+ ReorderBufferQueueSequence(ctx->reorder, xid, snapshot, buf->endptr,
+ origin_id, target_node, transactional,
+ xlrec->created, tuplebuf);
+}
diff --git a/src/backend/replication/logical/logical.c b/src/backend/replication/logical/logical.c
index 10cbdea124..aab5e1226d 100644
--- a/src/backend/replication/logical/logical.c
+++ b/src/backend/replication/logical/logical.c
@@ -73,6 +73,10 @@ static void truncate_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
static void message_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
XLogRecPtr message_lsn, bool transactional,
const char *prefix, Size message_size, const char *message);
+static void sequence_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ bool transactional,
+ int64 last_value, int64 log_cnt, bool is_called);
/* streaming callbacks */
static void stream_start_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
@@ -90,6 +94,10 @@ static void stream_change_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn
static void stream_message_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
XLogRecPtr message_lsn, bool transactional,
const char *prefix, Size message_size, const char *message);
+static void stream_sequence_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ bool transactional,
+ int64 last_value, int64 log_cnt, bool is_called);
static void stream_truncate_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
int nrelations, Relation relations[], ReorderBufferChange *change);
@@ -218,6 +226,7 @@ StartupDecodingContext(List *output_plugin_options,
ctx->reorder->apply_truncate = truncate_cb_wrapper;
ctx->reorder->commit = commit_cb_wrapper;
ctx->reorder->message = message_cb_wrapper;
+ ctx->reorder->sequence = sequence_cb_wrapper;
/*
* To support streaming, we require start/stop/abort/commit/change
@@ -234,6 +243,7 @@ StartupDecodingContext(List *output_plugin_options,
(ctx->callbacks.stream_commit_cb != NULL) ||
(ctx->callbacks.stream_change_cb != NULL) ||
(ctx->callbacks.stream_message_cb != NULL) ||
+ (ctx->callbacks.stream_sequence_cb != NULL) ||
(ctx->callbacks.stream_truncate_cb != NULL);
/*
@@ -251,6 +261,7 @@ StartupDecodingContext(List *output_plugin_options,
ctx->reorder->stream_commit = stream_commit_cb_wrapper;
ctx->reorder->stream_change = stream_change_cb_wrapper;
ctx->reorder->stream_message = stream_message_cb_wrapper;
+ ctx->reorder->stream_sequence = stream_sequence_cb_wrapper;
ctx->reorder->stream_truncate = stream_truncate_cb_wrapper;
@@ -1203,6 +1214,42 @@ message_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
error_context_stack = errcallback.previous;
}
+static void
+sequence_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel, bool transactional,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ LogicalDecodingContext *ctx = cache->private_data;
+ LogicalErrorCallbackState state;
+ ErrorContextCallback errcallback;
+
+ Assert(!ctx->fast_forward);
+
+ if (ctx->callbacks.sequence_cb == NULL)
+ return;
+
+ /* Push callback + info on the error context stack */
+ state.ctx = ctx;
+ state.callback_name = "sequence";
+ state.report_location = sequence_lsn;
+ errcallback.callback = output_plugin_error_callback;
+ errcallback.arg = (void *) &state;
+ errcallback.previous = error_context_stack;
+ error_context_stack = &errcallback;
+
+ /* set output state */
+ ctx->accept_writes = true;
+ ctx->write_xid = txn != NULL ? txn->xid : InvalidTransactionId;
+ ctx->write_location = sequence_lsn;
+
+ /* do the actual work: call callback */
+ ctx->callbacks.sequence_cb(ctx, txn, sequence_lsn, rel, transactional,
+ last_value, log_cnt, is_called);
+
+ /* Pop the error context stack */
+ error_context_stack = errcallback.previous;
+}
+
static void
stream_start_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
XLogRecPtr first_lsn)
@@ -1508,6 +1555,47 @@ stream_message_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
error_context_stack = errcallback.previous;
}
+static void
+stream_sequence_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ bool transactional,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ LogicalDecodingContext *ctx = cache->private_data;
+ LogicalErrorCallbackState state;
+ ErrorContextCallback errcallback;
+
+ Assert(!ctx->fast_forward);
+
+ /* We're only supposed to call this when streaming is supported. */
+ Assert(ctx->streaming);
+
+ /* this callback is optional */
+ if (ctx->callbacks.stream_sequence_cb == NULL)
+ return;
+
+ /* Push callback + info on the error context stack */
+ state.ctx = ctx;
+ state.callback_name = "stream_sequence";
+ state.report_location = sequence_lsn;
+ errcallback.callback = output_plugin_error_callback;
+ errcallback.arg = (void *) &state;
+ errcallback.previous = error_context_stack;
+ error_context_stack = &errcallback;
+
+ /* set output state */
+ ctx->accept_writes = true;
+ ctx->write_xid = txn != NULL ? txn->xid : InvalidTransactionId;
+ ctx->write_location = sequence_lsn;
+
+ /* do the actual work: call callback */
+ ctx->callbacks.sequence_cb(ctx, txn, sequence_lsn, rel, transactional,
+ last_value, log_cnt, is_called);
+
+ /* Pop the error context stack */
+ error_context_stack = errcallback.previous;
+}
+
static void
stream_truncate_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
int nrelations, Relation relations[],
diff --git a/src/backend/replication/logical/reorderbuffer.c b/src/backend/replication/logical/reorderbuffer.c
index 7aa5647a2c..6b6d0d9b95 100644
--- a/src/backend/replication/logical/reorderbuffer.c
+++ b/src/backend/replication/logical/reorderbuffer.c
@@ -77,6 +77,35 @@
* a bit more memory to the oldest subtransactions, because it's likely
* they are the source for the next sequence of changes.
*
+ * When decoding sequences, we differentiate between a sequences created
+ * in a (running) transaction, and sequences created in other (already
+ * committed) transactions. Changes for sequences created in the same
+ * top-level transaction are treated as "transactional" i.e. just like
+ * any other change from that transaction (and discarded in case of a
+ * rollback). Changes for sequences created earlier are treated as not
+ * transactional - are processed immediately, as if performed outside
+ * any transaction (and thus not rolled back).
+ *
+ * This mixed behavior is necessary - sequences are non-transactional
+ * (e.g. ROLLBACK does not undo the sequence increments). But for new
+ * sequences, we need to handle them in a transactional way, because if
+ * we ever get some DDL support, the sequence won't exist until the
+ * transaction gets applied. So we need to ensure the increments don't
+ * happen until the sequence gets created.
+ *
+ * To differentiate which sequences are "old" and which were created
+ * in a still-running transaction, we track sequences created in running
+ * transactions in a hash table. Each sequence is identified by it's
+ * relfilenode, and at transaction commit we discard all sequences
+ * created in that particular transaction. For each sequence we track
+ * the XID of the (sub)transaction that created it, and we cleanup the
+ * sequences for each subtransaction when it completes (commit/rollback).
+ *
+ * We don't use the XID to check if it's the same top-level transaction.
+ * It's enough to know it was created in an in-progress transaction,
+ * and we know it must be the current one because otherwise it wouldn't
+ * see the sequence object.
+ *
* -------------------------------------------------------------------------
*/
#include "postgres.h"
@@ -91,6 +120,7 @@
#include "access/xact.h"
#include "access/xlog_internal.h"
#include "catalog/catalog.h"
+#include "commands/sequence.h"
#include "lib/binaryheap.h"
#include "miscadmin.h"
#include "pgstat.h"
@@ -116,6 +146,13 @@ typedef struct ReorderBufferTXNByIdEnt
ReorderBufferTXN *txn;
} ReorderBufferTXNByIdEnt;
+/* entry for hash table we use to track sequences created in running xacts */
+typedef struct ReorderBufferSequenceEnt
+{
+ RelFileNode rnode;
+ TransactionId xid;
+} ReorderBufferSequenceEnt;
+
/* data structures for (relfilenode, ctid) => (cmin, cmax) mapping */
typedef struct ReorderBufferTupleCidKey
{
@@ -339,6 +376,14 @@ ReorderBufferAllocate(void)
buffer->by_txn = hash_create("ReorderBufferByXid", 1000, &hash_ctl,
HASH_ELEM | HASH_BLOBS | HASH_CONTEXT);
+ /* hash table of sequences, mapping relfilenode to XID of transaction */
+ hash_ctl.keysize = sizeof(RelFileNode);
+ hash_ctl.entrysize = sizeof(ReorderBufferSequenceEnt);
+ hash_ctl.hcxt = buffer->context;
+
+ buffer->sequences = hash_create("ReorderBufferSequenceHash", 1000, &hash_ctl,
+ HASH_ELEM | HASH_BLOBS | HASH_CONTEXT);
+
buffer->by_txn_last_xid = InvalidTransactionId;
buffer->by_txn_last_txn = NULL;
@@ -525,6 +570,13 @@ ReorderBufferReturnChange(ReorderBuffer *rb, ReorderBufferChange *change,
change->data.truncate.relids = NULL;
}
break;
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
+ if (change->data.sequence.tuple)
+ {
+ ReorderBufferReturnTupleBuf(rb, change->data.sequence.tuple);
+ change->data.sequence.tuple = NULL;
+ }
+ break;
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM:
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_ABORT:
case REORDER_BUFFER_CHANGE_INTERNAL_COMMAND_ID:
@@ -859,6 +911,230 @@ ReorderBufferQueueMessage(ReorderBuffer *rb, TransactionId xid,
}
}
+/*
+ * Treat the sequence increment as transactional?
+ *
+ * The hash table tracks all sequences created in in-progress transactions,
+ * so we simply do a lookup (the sequence is identified by relfilende). If
+ * we find a match, the increment should be handled as transactional.
+ */
+bool
+ReorderBufferSequenceIsTransactional(ReorderBuffer *rb,
+ RelFileNode rnode, bool created)
+{
+ bool found = false;
+
+ if (created)
+ return true;
+
+ hash_search(rb->sequences,
+ (void *) &rnode,
+ HASH_FIND,
+ &found);
+
+ return found;
+}
+
+/*
+ * Cleanup sequences created in in-progress transactions.
+ *
+ * There's no way to search by XID, so we simply do a seqscan of all
+ * the entries in the hash table. Hopefully there are only a couple
+ * entries in most cases - people generally don't create many new
+ * sequences over and over.
+ */
+static void
+ReorderBufferSequenceCleanup(ReorderBuffer *rb, TransactionId xid)
+{
+ HASH_SEQ_STATUS scan_status;
+ ReorderBufferSequenceEnt *ent;
+
+ hash_seq_init(&scan_status, rb->sequences);
+ while ((ent = (ReorderBufferSequenceEnt *) hash_seq_search(&scan_status)) != NULL)
+ {
+ /* skip sequences not from this transaction */
+ if (ent->xid != xid)
+ continue;
+
+ (void) hash_search(rb->sequences,
+ (void *) &(ent->rnode),
+ HASH_REMOVE, NULL);
+ }
+}
+
+/*
+ * A transactional sequence increment is queued to be processed upon commit
+ * and a non-transactional increment gets processed immediately.
+ *
+ * A sequence update may be both transactional and non-transactional. When
+ * created in a running transaction, treat it as transactional and queue
+ * the change in it. Otherwise treat it as non-transactional, so that we
+ * don't forget the increment in case of a rollback.
+ */
+void
+ReorderBufferQueueSequence(ReorderBuffer *rb, TransactionId xid,
+ Snapshot snapshot, XLogRecPtr lsn, RepOriginId origin_id,
+ RelFileNode rnode, bool transactional, bool created,
+ ReorderBufferTupleBuf *tuplebuf)
+{
+ /*
+ * Change needs to be handled as transactional, because the sequence was
+ * created in a transaction that is still running. In that case all the
+ * changes need to be queued in that transaction, we must not send them
+ * to the downstream until the transaction commits.
+ *
+ * There's a bit of a trouble with subtransactions - we can't queue it
+ * into the subxact, because it might be rolled back and we'd lose the
+ * increment. We need to queue it into the same (sub)xact that created
+ * the sequence, which is why we track the XID in the hash table.
+ */
+ if (transactional)
+ {
+ MemoryContext oldcontext;
+ ReorderBufferChange *change;
+
+ /* lookup sequence by relfilenode */
+ ReorderBufferSequenceEnt *ent;
+ bool found;
+
+ /* transactional changes require a transaction */
+ Assert(xid != InvalidTransactionId);
+
+ /* search the lookup table (we ignore the return value, found is enough) */
+ ent = hash_search(rb->sequences,
+ (void *) &rnode,
+ created ? HASH_ENTER : HASH_FIND,
+ &found);
+
+ /*
+ * If this is the "create" increment, we must not have found any
+ * pre-existing entry in the hash table (i.e. there must not be
+ * any conflicting sequence).
+ */
+ Assert(!(created && found));
+
+ /* But we must have either created or found an existing entry. */
+ Assert(created || found);
+
+ /*
+ * When creating the sequence, remember the XID of the transaction
+ * that created id.
+ */
+ if (created)
+ ent->xid = xid;
+
+ /* XXX Maybe check that we're still in the same top-level xact? */
+
+ /* OK, allocate and queue the change */
+ oldcontext = MemoryContextSwitchTo(rb->context);
+
+ change = ReorderBufferGetChange(rb);
+
+ change->action = REORDER_BUFFER_CHANGE_SEQUENCE;
+ change->origin_id = origin_id;
+
+ memcpy(&change->data.sequence.relnode, &rnode, sizeof(RelFileNode));
+
+ change->data.sequence.tuple = tuplebuf;
+
+ /* add it to the same subxact that created the sequence */
+ ReorderBufferQueueChange(rb, ent->xid, lsn, change, false);
+
+ MemoryContextSwitchTo(oldcontext);
+ }
+ else
+ {
+ /*
+ * This increment is for a sequence that was not created in any
+ * running transaction, so we treat it as non-transactional and
+ * just send it to the output plugin directly.
+ */
+ ReorderBufferTXN *txn = NULL;
+ volatile Snapshot snapshot_now = snapshot;
+ bool using_subtxn;
+
+#ifdef USE_ASSERT_CHECKING
+ /* All "creates" have to be handled as transactional. */
+ Assert(!created);
+
+ /* Make sure the sequence is not in the hash table. */
+ {
+ bool found;
+ hash_search(rb->sequences,
+ (void *) &rnode,
+ HASH_FIND, &found);
+ Assert(!found);
+ }
+#endif
+
+ if (xid != InvalidTransactionId)
+ txn = ReorderBufferTXNByXid(rb, xid, true, NULL, lsn, true);
+
+ /* setup snapshot to allow catalog access */
+ SetupHistoricSnapshot(snapshot_now, NULL);
+
+ /*
+ * Decoding needs access to syscaches et al., which in turn use
+ * heavyweight locks and such. Thus we need to have enough state around to
+ * keep track of those. The easiest way is to simply use a transaction
+ * internally. That also allows us to easily enforce that nothing writes
+ * to the database by checking for xid assignments.
+ *
+ * When we're called via the SQL SRF there's already a transaction
+ * started, so start an explicit subtransaction there.
+ */
+ using_subtxn = IsTransactionOrTransactionBlock();
+
+ PG_TRY();
+ {
+ Relation relation;
+ HeapTuple tuple;
+ Form_pg_sequence_data seq;
+ Oid reloid;
+
+ if (using_subtxn)
+ BeginInternalSubTransaction("sequence");
+ else
+ StartTransactionCommand();
+
+ reloid = RelidByRelfilenode(rnode.spcNode, rnode.relNode);
+
+ if (reloid == InvalidOid)
+ elog(ERROR, "could not map filenode \"%s\" to relation OID",
+ relpathperm(rnode,
+ MAIN_FORKNUM));
+
+ relation = RelationIdGetRelation(reloid);
+ tuple = &tuplebuf->tuple;
+ seq = (Form_pg_sequence_data) GETSTRUCT(tuple);
+
+ rb->sequence(rb, txn, lsn, relation, transactional,
+ seq->last_value, seq->log_cnt, seq->is_called);
+
+ RelationClose(relation);
+
+ TeardownHistoricSnapshot(false);
+
+ AbortCurrentTransaction();
+
+ if (using_subtxn)
+ RollbackAndReleaseCurrentSubTransaction();
+ }
+ PG_CATCH();
+ {
+ TeardownHistoricSnapshot(true);
+
+ AbortCurrentTransaction();
+
+ if (using_subtxn)
+ RollbackAndReleaseCurrentSubTransaction();
+
+ PG_RE_THROW();
+ }
+ PG_END_TRY();
+ }
+}
+
/*
* AssertTXNLsnOrder
* Verify LSN ordering of transaction lists in the reorderbuffer
@@ -1535,6 +1811,9 @@ ReorderBufferCleanupTXN(ReorderBuffer *rb, ReorderBufferTXN *txn)
&found);
Assert(found);
+ /* Remove sequences created in this transaction (if any). */
+ ReorderBufferSequenceCleanup(rb, txn->xid);
+
/* remove entries spilled to disk */
if (rbtxn_is_serialized(txn))
ReorderBufferRestoreCleanup(rb, txn);
@@ -1950,6 +2229,29 @@ ReorderBufferApplyMessage(ReorderBuffer *rb, ReorderBufferTXN *txn,
change->data.msg.message);
}
+/*
+ * Helper function for ReorderBufferProcessTXN for applying sequences.
+ */
+static inline void
+ReorderBufferApplySequence(ReorderBuffer *rb, ReorderBufferTXN *txn,
+ Relation relation, ReorderBufferChange *change,
+ bool streaming)
+{
+ HeapTuple tuple;
+ Form_pg_sequence_data seq;
+
+ tuple = &change->data.sequence.tuple->tuple;
+ seq = (Form_pg_sequence_data) GETSTRUCT(tuple);
+
+ /* Only ever called from ReorderBufferApplySequence, so transational. */
+ if (streaming)
+ rb->stream_sequence(rb, txn, change->lsn, relation, true,
+ seq->last_value, seq->log_cnt, seq->is_called);
+ else
+ rb->sequence(rb, txn, change->lsn, relation, true,
+ seq->last_value, seq->log_cnt, seq->is_called);
+}
+
/*
* Function to store the command id and snapshot at the end of the current
* stream so that we can reuse the same while sending the next stream.
@@ -2392,6 +2694,31 @@ ReorderBufferProcessTXN(ReorderBuffer *rb, ReorderBufferTXN *txn,
case REORDER_BUFFER_CHANGE_INTERNAL_TUPLECID:
elog(ERROR, "tuplecid value in changequeue");
break;
+
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
+ Assert(snapshot_now);
+
+ reloid = RelidByRelfilenode(change->data.sequence.relnode.spcNode,
+ change->data.sequence.relnode.relNode);
+
+ if (reloid == InvalidOid)
+ elog(ERROR, "could not map filenode \"%s\" to relation OID",
+ relpathperm(change->data.sequence.relnode,
+ MAIN_FORKNUM));
+
+ relation = RelationIdGetRelation(reloid);
+
+ if (!RelationIsValid(relation))
+ elog(ERROR, "could not open relation with OID %u (for filenode \"%s\")",
+ reloid,
+ relpathperm(change->data.sequence.relnode,
+ MAIN_FORKNUM));
+
+ if (RelationIsLogicallyLogged(relation))
+ ReorderBufferApplySequence(rb, txn, relation, change, streaming);
+
+ RelationClose(relation);
+ break;
}
}
@@ -3776,6 +4103,39 @@ ReorderBufferSerializeChange(ReorderBuffer *rb, ReorderBufferTXN *txn,
memcpy(data, change->data.truncate.relids, size);
data += size;
+ break;
+ }
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
+ {
+ char *data;
+ ReorderBufferTupleBuf *tup;
+ Size len = 0;
+
+ tup = change->data.sequence.tuple;
+
+ if (tup)
+ {
+ sz += sizeof(HeapTupleData);
+ len = tup->tuple.t_len;
+ sz += len;
+ }
+
+ /* make sure we have enough space */
+ ReorderBufferSerializeReserve(rb, sz);
+
+ data = ((char *) rb->outbuf) + sizeof(ReorderBufferDiskChange);
+ /* might have been reallocated above */
+ ondisk = (ReorderBufferDiskChange *) rb->outbuf;
+
+ if (len)
+ {
+ memcpy(data, &tup->tuple, sizeof(HeapTupleData));
+ data += sizeof(HeapTupleData);
+
+ memcpy(data, tup->tuple.t_data, len);
+ data += len;
+ }
+
break;
}
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM:
@@ -4040,6 +4400,22 @@ ReorderBufferChangeSize(ReorderBufferChange *change)
{
sz += sizeof(Oid) * change->data.truncate.nrelids;
+ break;
+ }
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
+ {
+ ReorderBufferTupleBuf *tup;
+ Size len = 0;
+
+ tup = change->data.sequence.tuple;
+
+ if (tup)
+ {
+ sz += sizeof(HeapTupleData);
+ len = tup->tuple.t_len;
+ sz += len;
+ }
+
break;
}
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM:
@@ -4341,6 +4717,30 @@ ReorderBufferRestoreChange(ReorderBuffer *rb, ReorderBufferTXN *txn,
break;
}
+
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
+ if (change->data.sequence.tuple)
+ {
+ uint32 tuplelen = ((HeapTuple) data)->t_len;
+
+ change->data.sequence.tuple =
+ ReorderBufferGetTupleBuf(rb, tuplelen - SizeofHeapTupleHeader);
+
+ /* restore ->tuple */
+ memcpy(&change->data.sequence.tuple->tuple, data,
+ sizeof(HeapTupleData));
+ data += sizeof(HeapTupleData);
+
+ /* reset t_data pointer into the new tuplebuf */
+ change->data.sequence.tuple->tuple.t_data =
+ ReorderBufferTupleBufData(change->data.sequence.tuple);
+
+ /* restore tuple data itself */
+ memcpy(change->data.sequence.tuple->tuple.t_data, data, tuplelen);
+ data += tuplelen;
+ }
+ break;
+
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM:
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_ABORT:
case REORDER_BUFFER_CHANGE_INTERNAL_COMMAND_ID:
diff --git a/src/include/commands/sequence.h b/src/include/commands/sequence.h
index 40544dd4c7..5919fb90ee 100644
--- a/src/include/commands/sequence.h
+++ b/src/include/commands/sequence.h
@@ -48,6 +48,7 @@ typedef FormData_pg_sequence_data *Form_pg_sequence_data;
typedef struct xl_seq_rec
{
RelFileNode node;
+ bool created; /* is this a CREATE SEQUENCE */
/* SEQUENCE TUPLE DATA FOLLOWS AT THE END */
} xl_seq_rec;
diff --git a/src/include/replication/output_plugin.h b/src/include/replication/output_plugin.h
index 810495ed0e..43b8a63ee4 100644
--- a/src/include/replication/output_plugin.h
+++ b/src/include/replication/output_plugin.h
@@ -88,6 +88,18 @@ typedef void (*LogicalDecodeMessageCB) (struct LogicalDecodingContext *ctx,
Size message_size,
const char *message);
+/*
+ * Called for the generic logical decoding sequences.
+ */
+typedef void (*LogicalDecodeSequenceCB) (struct LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional,
+ int64 last_value,
+ int64 log_cnt,
+ bool is_called);
+
/*
* Filter changes by origin.
*/
@@ -199,6 +211,19 @@ typedef void (*LogicalDecodeStreamMessageCB) (struct LogicalDecodingContext *ctx
Size message_size,
const char *message);
+/*
+ * Called for the streaming generic logical decoding sequences from in-progress
+ * transactions.
+ */
+typedef void (*LogicalDecodeStreamSequenceCB) (struct LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional,
+ int64 last_value,
+ int64 log_cnt,
+ bool is_called);
+
/*
* Callback for streaming truncates from in-progress transactions.
*/
@@ -219,6 +244,7 @@ typedef struct OutputPluginCallbacks
LogicalDecodeTruncateCB truncate_cb;
LogicalDecodeCommitCB commit_cb;
LogicalDecodeMessageCB message_cb;
+ LogicalDecodeSequenceCB sequence_cb;
LogicalDecodeFilterByOriginCB filter_by_origin_cb;
LogicalDecodeShutdownCB shutdown_cb;
@@ -237,6 +263,7 @@ typedef struct OutputPluginCallbacks
LogicalDecodeStreamCommitCB stream_commit_cb;
LogicalDecodeStreamChangeCB stream_change_cb;
LogicalDecodeStreamMessageCB stream_message_cb;
+ LogicalDecodeStreamSequenceCB stream_sequence_cb;
LogicalDecodeStreamTruncateCB stream_truncate_cb;
} OutputPluginCallbacks;
diff --git a/src/include/replication/reorderbuffer.h b/src/include/replication/reorderbuffer.h
index 5b40ff75f7..ab469fb210 100644
--- a/src/include/replication/reorderbuffer.h
+++ b/src/include/replication/reorderbuffer.h
@@ -64,7 +64,8 @@ enum ReorderBufferChangeType
REORDER_BUFFER_CHANGE_INTERNAL_SPEC_INSERT,
REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM,
REORDER_BUFFER_CHANGE_INTERNAL_SPEC_ABORT,
- REORDER_BUFFER_CHANGE_TRUNCATE
+ REORDER_BUFFER_CHANGE_TRUNCATE,
+ REORDER_BUFFER_CHANGE_SEQUENCE
};
/* forward declaration */
@@ -158,6 +159,13 @@ typedef struct ReorderBufferChange
uint32 ninvalidations; /* Number of messages */
SharedInvalidationMessage *invalidations; /* invalidation message */
} inval;
+
+ /* Context data for Sequence changes */
+ struct
+ {
+ RelFileNode relnode;
+ ReorderBufferTupleBuf *tuple;
+ } sequence;
} data;
/*
@@ -430,6 +438,15 @@ typedef void (*ReorderBufferMessageCB) (ReorderBuffer *rb,
const char *prefix, Size sz,
const char *message);
+/* sequence callback signature */
+typedef void (*ReorderBufferSequenceCB) (ReorderBuffer *rb,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional,
+ int64 last_value, int64 log_cnt,
+ bool is_called);
+
/* begin prepare callback signature */
typedef void (*ReorderBufferBeginPrepareCB) (ReorderBuffer *rb,
ReorderBufferTXN *txn);
@@ -496,6 +513,15 @@ typedef void (*ReorderBufferStreamMessageCB) (
const char *prefix, Size sz,
const char *message);
+/* stream sequence callback signature */
+typedef void (*ReorderBufferStreamSequenceCB) (ReorderBuffer *rb,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional,
+ int64 last_value, int64 log_cnt,
+ bool is_called);
+
/* stream truncate callback signature */
typedef void (*ReorderBufferStreamTruncateCB) (
ReorderBuffer *rb,
@@ -511,6 +537,12 @@ struct ReorderBuffer
*/
HTAB *by_txn;
+ /*
+ * relfilenode => XID lookup table for sequences created in a transaction
+ * (also includes altered sequences, which assigns new relfilenode)
+ */
+ HTAB *sequences;
+
/*
* Transactions that could be a toplevel xact, ordered by LSN of the first
* record bearing that xid.
@@ -541,6 +573,7 @@ struct ReorderBuffer
ReorderBufferApplyTruncateCB apply_truncate;
ReorderBufferCommitCB commit;
ReorderBufferMessageCB message;
+ ReorderBufferSequenceCB sequence;
/*
* Callbacks to be called when streaming a transaction at prepare time.
@@ -560,6 +593,7 @@ struct ReorderBuffer
ReorderBufferStreamCommitCB stream_commit;
ReorderBufferStreamChangeCB stream_change;
ReorderBufferStreamMessageCB stream_message;
+ ReorderBufferStreamSequenceCB stream_sequence;
ReorderBufferStreamTruncateCB stream_truncate;
/*
@@ -635,6 +669,10 @@ void ReorderBufferQueueChange(ReorderBuffer *, TransactionId,
void ReorderBufferQueueMessage(ReorderBuffer *, TransactionId, Snapshot snapshot, XLogRecPtr lsn,
bool transactional, const char *prefix,
Size message_size, const char *message);
+void ReorderBufferQueueSequence(ReorderBuffer *rb, TransactionId xid,
+ Snapshot snapshot, XLogRecPtr lsn, RepOriginId origin_id,
+ RelFileNode rnode, bool transactional, bool created,
+ ReorderBufferTupleBuf *tuplebuf);
void ReorderBufferCommit(ReorderBuffer *, TransactionId,
XLogRecPtr commit_lsn, XLogRecPtr end_lsn,
TimestampTz commit_time, RepOriginId origin_id, XLogRecPtr origin_lsn);
@@ -682,4 +720,7 @@ void ReorderBufferSetRestartPoint(ReorderBuffer *, XLogRecPtr ptr);
void StartupReorderBuffer(void);
+bool ReorderBufferSequenceIsTransactional(ReorderBuffer *rb,
+ RelFileNode rnode, bool created);
+
#endif
--
2.31.1
[text/x-patch] 0003-Add-support-for-decoding-sequences-to-test_-20211222.patch (20.3K, ../../[email protected]/4-0003-Add-support-for-decoding-sequences-to-test_-20211222.patch)
download | inline diff:
From 4b249e65b55d5e9fb7b08135ead68bfa3c58034c Mon Sep 17 00:00:00 2001
From: Tomas Vondra <[email protected]>
Date: Fri, 24 Sep 2021 00:42:04 +0200
Subject: [PATCH 3/6] Add support for decoding sequences to test_decoding
---
contrib/test_decoding/Makefile | 3 +-
contrib/test_decoding/expected/sequence.out | 327 ++++++++++++++++++++
contrib/test_decoding/sql/sequence.sql | 119 +++++++
contrib/test_decoding/test_decoding.c | 69 +++++
4 files changed, 517 insertions(+), 1 deletion(-)
create mode 100644 contrib/test_decoding/expected/sequence.out
create mode 100644 contrib/test_decoding/sql/sequence.sql
diff --git a/contrib/test_decoding/Makefile b/contrib/test_decoding/Makefile
index 9a31e0b879..56ddc3abae 100644
--- a/contrib/test_decoding/Makefile
+++ b/contrib/test_decoding/Makefile
@@ -5,7 +5,8 @@ PGFILEDESC = "test_decoding - example of a logical decoding output plugin"
REGRESS = ddl xact rewrite toast permissions decoding_in_xact \
decoding_into_rel binary prepared replorigin time messages \
- spill slot truncate stream stats twophase twophase_stream
+ spill slot truncate stream stats twophase twophase_stream \
+ sequence
ISOLATION = mxact delayed_startup ondisk_startup concurrent_ddl_dml \
oldest_xmin snapshot_transfer subxact_without_top concurrent_stream \
twophase_snapshot
diff --git a/contrib/test_decoding/expected/sequence.out b/contrib/test_decoding/expected/sequence.out
new file mode 100644
index 0000000000..c3c2a18a35
--- /dev/null
+++ b/contrib/test_decoding/expected/sequence.out
@@ -0,0 +1,327 @@
+-- predictability
+SET synchronous_commit = on;
+SELECT 'init' FROM pg_create_logical_replication_slot('regression_slot', 'test_decoding');
+ ?column?
+----------
+ init
+(1 row)
+
+CREATE SEQUENCE test_sequence;
+-- test the sequence changes by several nextval() calls
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 1
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 2
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 4
+(1 row)
+
+-- test the sequence changes by several ALTER commands
+ALTER SEQUENCE test_sequence INCREMENT BY 10;
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 14
+(1 row)
+
+ALTER SEQUENCE test_sequence START WITH 3000;
+ALTER SEQUENCE test_sequence MAXVALUE 10000;
+ALTER SEQUENCE test_sequence RESTART WITH 4000;
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 4000
+(1 row)
+
+-- test the sequence changes by several setval() calls
+SELECT setval('test_sequence', 3500);
+ setval
+--------
+ 3500
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3510
+(1 row)
+
+SELECT setval('test_sequence', 3500, true);
+ setval
+--------
+ 3500
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3510
+(1 row)
+
+SELECT setval('test_sequence', 3500, false);
+ setval
+--------
+ 3500
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3500
+(1 row)
+
+-- show results and drop sequence
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+ data
+--------------------------------------------------------------------------------------
+ BEGIN
+ sequence public.test_sequence: transactional:1 last_value:1 log_cnt:0 is_called:0
+ COMMIT
+ sequence public.test_sequence: transactional:0 last_value:33 log_cnt:0 is_called:1
+ BEGIN
+ sequence public.test_sequence: transactional:1 last_value:4 log_cnt:0 is_called:1
+ COMMIT
+ sequence public.test_sequence: transactional:0 last_value:334 log_cnt:0 is_called:1
+ BEGIN
+ sequence public.test_sequence: transactional:1 last_value:14 log_cnt:32 is_called:1
+ COMMIT
+ BEGIN
+ sequence public.test_sequence: transactional:1 last_value:14 log_cnt:0 is_called:1
+ COMMIT
+ BEGIN
+ sequence public.test_sequence: transactional:1 last_value:4000 log_cnt:0 is_called:0
+ COMMIT
+ sequence public.test_sequence: transactional:0 last_value:4320 log_cnt:0 is_called:1
+ sequence public.test_sequence: transactional:0 last_value:3500 log_cnt:0 is_called:1
+ sequence public.test_sequence: transactional:0 last_value:3830 log_cnt:0 is_called:1
+ sequence public.test_sequence: transactional:0 last_value:3500 log_cnt:0 is_called:1
+ sequence public.test_sequence: transactional:0 last_value:3830 log_cnt:0 is_called:1
+ sequence public.test_sequence: transactional:0 last_value:3500 log_cnt:0 is_called:0
+ sequence public.test_sequence: transactional:0 last_value:3820 log_cnt:0 is_called:1
+(24 rows)
+
+DROP SEQUENCE test_sequence;
+-- rollback on sequence creation and update
+BEGIN;
+CREATE SEQUENCE test_sequence;
+CREATE TABLE test_table (a INT);
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 1
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 2
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3
+(1 row)
+
+SELECT setval('test_sequence', 3000);
+ setval
+--------
+ 3000
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3001
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3002
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3003
+(1 row)
+
+ALTER SEQUENCE test_sequence RESTART WITH 6000;
+INSERT INTO test_table VALUES( (SELECT nextval('test_sequence')) );
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 6001
+(1 row)
+
+ROLLBACK;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'skip-sequences', '0');
+ data
+------
+(0 rows)
+
+-- rollback on table creation with serial column
+BEGIN;
+CREATE TABLE test_table (a SERIAL, b INT);
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+INSERT INTO test_table (b) VALUES (300);
+SELECT setval('test_table_a_seq', 3000);
+ setval
+--------
+ 3000
+(1 row)
+
+INSERT INTO test_table (b) VALUES (400);
+INSERT INTO test_table (b) VALUES (500);
+INSERT INTO test_table (b) VALUES (600);
+ALTER SEQUENCE test_table_a_seq RESTART WITH 6000;
+INSERT INTO test_table (b) VALUES (700);
+INSERT INTO test_table (b) VALUES (800);
+INSERT INTO test_table (b) VALUES (900);
+ROLLBACK;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'skip-sequences', '0');
+ data
+------
+(0 rows)
+
+-- rollback on table with serial column
+CREATE TABLE test_table (a SERIAL, b INT);
+BEGIN;
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+INSERT INTO test_table (b) VALUES (300);
+SELECT setval('test_table_a_seq', 3000);
+ setval
+--------
+ 3000
+(1 row)
+
+INSERT INTO test_table (b) VALUES (400);
+INSERT INTO test_table (b) VALUES (500);
+INSERT INTO test_table (b) VALUES (600);
+ALTER SEQUENCE test_table_a_seq RESTART WITH 6000;
+INSERT INTO test_table (b) VALUES (700);
+INSERT INTO test_table (b) VALUES (800);
+INSERT INTO test_table (b) VALUES (900);
+ROLLBACK;
+-- check table and sequence values after rollback
+SELECT * from test_table_a_seq;
+ last_value | log_cnt | is_called
+------------+---------+-----------
+ 3003 | 30 | t
+(1 row)
+
+SELECT nextval('test_table_a_seq');
+ nextval
+---------
+ 3004
+(1 row)
+
+DROP TABLE test_table;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+ data
+-----------------------------------------------------------------------------------------
+ BEGIN
+ sequence public.test_table_a_seq: transactional:1 last_value:1 log_cnt:0 is_called:0
+ COMMIT
+ sequence public.test_table_a_seq: transactional:0 last_value:33 log_cnt:0 is_called:1
+ sequence public.test_table_a_seq: transactional:0 last_value:3000 log_cnt:0 is_called:1
+ sequence public.test_table_a_seq: transactional:0 last_value:3033 log_cnt:0 is_called:1
+ BEGIN
+ COMMIT
+(8 rows)
+
+-- savepoint test on table with serial column
+BEGIN;
+CREATE TABLE test_table (a SERIAL, b INT);
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+SAVEPOINT a;
+INSERT INTO test_table (b) VALUES (300);
+ROLLBACK TO SAVEPOINT a;
+DROP TABLE test_table;
+COMMIT;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+ data
+---------------------------------------------------------------------------------------
+ BEGIN
+ sequence public.test_table_a_seq: transactional:1 last_value:1 log_cnt:0 is_called:0
+ sequence public.test_table_a_seq: transactional:1 last_value:33 log_cnt:0 is_called:1
+ table public.test_table: INSERT: a[integer]:1 b[integer]:100
+ table public.test_table: INSERT: a[integer]:2 b[integer]:200
+ COMMIT
+(6 rows)
+
+-- savepoint test on table with serial column
+BEGIN;
+CREATE SEQUENCE test_sequence;
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 1
+(1 row)
+
+SELECT setval('test_sequence', 3000);
+ setval
+--------
+ 3000
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3001
+(1 row)
+
+SAVEPOINT a;
+ALTER SEQUENCE test_sequence START WITH 7000;
+SELECT setval('test_sequence', 5000);
+ setval
+--------
+ 5000
+(1 row)
+
+ROLLBACK TO SAVEPOINT a;
+SELECT * FROM test_sequence;
+ last_value | log_cnt | is_called
+------------+---------+-----------
+ 3001 | 32 | t
+(1 row)
+
+DROP SEQUENCE test_sequence;
+COMMIT;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+ data
+--------------------------------------------------------------------------------------
+ BEGIN
+ sequence public.test_sequence: transactional:1 last_value:1 log_cnt:0 is_called:0
+ sequence public.test_sequence: transactional:1 last_value:33 log_cnt:0 is_called:1
+ sequence public.test_sequence: transactional:1 last_value:3000 log_cnt:0 is_called:1
+ sequence public.test_sequence: transactional:1 last_value:3033 log_cnt:0 is_called:1
+ COMMIT
+(6 rows)
+
+SELECT pg_drop_replication_slot('regression_slot');
+ pg_drop_replication_slot
+--------------------------
+
+(1 row)
+
diff --git a/contrib/test_decoding/sql/sequence.sql b/contrib/test_decoding/sql/sequence.sql
new file mode 100644
index 0000000000..21c4b79222
--- /dev/null
+++ b/contrib/test_decoding/sql/sequence.sql
@@ -0,0 +1,119 @@
+-- predictability
+SET synchronous_commit = on;
+SELECT 'init' FROM pg_create_logical_replication_slot('regression_slot', 'test_decoding');
+
+CREATE SEQUENCE test_sequence;
+
+-- test the sequence changes by several nextval() calls
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+
+-- test the sequence changes by several ALTER commands
+ALTER SEQUENCE test_sequence INCREMENT BY 10;
+SELECT nextval('test_sequence');
+
+ALTER SEQUENCE test_sequence START WITH 3000;
+ALTER SEQUENCE test_sequence MAXVALUE 10000;
+ALTER SEQUENCE test_sequence RESTART WITH 4000;
+SELECT nextval('test_sequence');
+
+-- test the sequence changes by several setval() calls
+SELECT setval('test_sequence', 3500);
+SELECT nextval('test_sequence');
+SELECT setval('test_sequence', 3500, true);
+SELECT nextval('test_sequence');
+SELECT setval('test_sequence', 3500, false);
+SELECT nextval('test_sequence');
+
+-- show results and drop sequence
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+DROP SEQUENCE test_sequence;
+
+-- rollback on sequence creation and update
+BEGIN;
+CREATE SEQUENCE test_sequence;
+CREATE TABLE test_table (a INT);
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT setval('test_sequence', 3000);
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+ALTER SEQUENCE test_sequence RESTART WITH 6000;
+INSERT INTO test_table VALUES( (SELECT nextval('test_sequence')) );
+SELECT nextval('test_sequence');
+ROLLBACK;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'skip-sequences', '0');
+
+-- rollback on table creation with serial column
+BEGIN;
+CREATE TABLE test_table (a SERIAL, b INT);
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+INSERT INTO test_table (b) VALUES (300);
+SELECT setval('test_table_a_seq', 3000);
+INSERT INTO test_table (b) VALUES (400);
+INSERT INTO test_table (b) VALUES (500);
+INSERT INTO test_table (b) VALUES (600);
+ALTER SEQUENCE test_table_a_seq RESTART WITH 6000;
+INSERT INTO test_table (b) VALUES (700);
+INSERT INTO test_table (b) VALUES (800);
+INSERT INTO test_table (b) VALUES (900);
+ROLLBACK;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'skip-sequences', '0');
+
+-- rollback on table with serial column
+CREATE TABLE test_table (a SERIAL, b INT);
+
+BEGIN;
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+INSERT INTO test_table (b) VALUES (300);
+SELECT setval('test_table_a_seq', 3000);
+INSERT INTO test_table (b) VALUES (400);
+INSERT INTO test_table (b) VALUES (500);
+INSERT INTO test_table (b) VALUES (600);
+ALTER SEQUENCE test_table_a_seq RESTART WITH 6000;
+INSERT INTO test_table (b) VALUES (700);
+INSERT INTO test_table (b) VALUES (800);
+INSERT INTO test_table (b) VALUES (900);
+ROLLBACK;
+
+-- check table and sequence values after rollback
+SELECT * from test_table_a_seq;
+SELECT nextval('test_table_a_seq');
+
+DROP TABLE test_table;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+
+-- savepoint test on table with serial column
+BEGIN;
+CREATE TABLE test_table (a SERIAL, b INT);
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+SAVEPOINT a;
+INSERT INTO test_table (b) VALUES (300);
+ROLLBACK TO SAVEPOINT a;
+DROP TABLE test_table;
+COMMIT;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+
+-- savepoint test on table with serial column
+BEGIN;
+CREATE SEQUENCE test_sequence;
+SELECT nextval('test_sequence');
+SELECT setval('test_sequence', 3000);
+SELECT nextval('test_sequence');
+SAVEPOINT a;
+ALTER SEQUENCE test_sequence START WITH 7000;
+SELECT setval('test_sequence', 5000);
+ROLLBACK TO SAVEPOINT a;
+SELECT * FROM test_sequence;
+DROP SEQUENCE test_sequence;
+COMMIT;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
+
+SELECT pg_drop_replication_slot('regression_slot');
diff --git a/contrib/test_decoding/test_decoding.c b/contrib/test_decoding/test_decoding.c
index e5cd84e85e..bc3250b8f4 100644
--- a/contrib/test_decoding/test_decoding.c
+++ b/contrib/test_decoding/test_decoding.c
@@ -35,6 +35,7 @@ typedef struct
bool include_timestamp;
bool skip_empty_xacts;
bool only_local;
+ bool skip_sequences;
} TestDecodingData;
/*
@@ -76,6 +77,10 @@ static void pg_decode_message(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn, XLogRecPtr message_lsn,
bool transactional, const char *prefix,
Size sz, const char *message);
+static void pg_decode_sequence(LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn, XLogRecPtr sequence_lsn,
+ Relation rel, bool transactional,
+ int64 last_value, int64 log_cnt, bool is_called);
static bool pg_decode_filter_prepare(LogicalDecodingContext *ctx,
TransactionId xid,
const char *gid);
@@ -116,6 +121,10 @@ static void pg_decode_stream_message(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn, XLogRecPtr message_lsn,
bool transactional, const char *prefix,
Size sz, const char *message);
+static void pg_decode_stream_sequence(LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn, XLogRecPtr sequence_lsn,
+ Relation rel, bool transactional,
+ int64 last_value, int64 log_cnt, bool is_called);
static void pg_decode_stream_truncate(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn,
int nrelations, Relation relations[],
@@ -141,6 +150,7 @@ _PG_output_plugin_init(OutputPluginCallbacks *cb)
cb->filter_by_origin_cb = pg_decode_filter;
cb->shutdown_cb = pg_decode_shutdown;
cb->message_cb = pg_decode_message;
+ cb->sequence_cb = pg_decode_sequence;
cb->filter_prepare_cb = pg_decode_filter_prepare;
cb->begin_prepare_cb = pg_decode_begin_prepare_txn;
cb->prepare_cb = pg_decode_prepare_txn;
@@ -153,6 +163,7 @@ _PG_output_plugin_init(OutputPluginCallbacks *cb)
cb->stream_commit_cb = pg_decode_stream_commit;
cb->stream_change_cb = pg_decode_stream_change;
cb->stream_message_cb = pg_decode_stream_message;
+ cb->stream_sequence_cb = pg_decode_stream_sequence;
cb->stream_truncate_cb = pg_decode_stream_truncate;
}
@@ -175,6 +186,9 @@ pg_decode_startup(LogicalDecodingContext *ctx, OutputPluginOptions *opt,
data->skip_empty_xacts = false;
data->only_local = false;
+ /* skip sequences by default for backwards compatibility */
+ data->skip_sequences = true;
+
ctx->output_plugin_private = data;
opt->output_type = OUTPUT_PLUGIN_TEXTUAL_OUTPUT;
@@ -265,6 +279,17 @@ pg_decode_startup(LogicalDecodingContext *ctx, OutputPluginOptions *opt,
errmsg("could not parse value \"%s\" for parameter \"%s\"",
strVal(elem->arg), elem->defname)));
}
+ else if (strcmp(elem->defname, "skip-sequences") == 0)
+ {
+
+ if (elem->arg == NULL)
+ continue; /* true by default */
+ else if (!parse_bool(strVal(elem->arg), &data->skip_sequences))
+ ereport(ERROR,
+ (errcode(ERRCODE_INVALID_PARAMETER_VALUE),
+ errmsg("could not parse value \"%s\" for parameter \"%s\"",
+ strVal(elem->arg), elem->defname)));
+ }
else
{
ereport(ERROR,
@@ -744,6 +769,28 @@ pg_decode_message(LogicalDecodingContext *ctx,
OutputPluginWrite(ctx, true);
}
+static void
+pg_decode_sequence(LogicalDecodingContext *ctx, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ bool transactional,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ TestDecodingData *data = ctx->output_plugin_private;
+
+ /* return if requested to skip_sequences */
+ if (data->skip_sequences)
+ return;
+
+ OutputPluginPrepareWrite(ctx, true);
+ appendStringInfoString(ctx->out, "sequence ");
+ appendStringInfoString(ctx->out,
+ quote_qualified_identifier(get_namespace_name(get_rel_namespace(RelationGetRelid(rel))),
+ RelationGetRelationName(rel)));
+ appendStringInfo(ctx->out, ": transactional:%d last_value:%zu log_cnt:%zu is_called:%d",
+ transactional, last_value, log_cnt, is_called);
+ OutputPluginWrite(ctx, true);
+}
+
static void
pg_decode_stream_start(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn)
@@ -943,6 +990,28 @@ pg_decode_stream_message(LogicalDecodingContext *ctx,
OutputPluginWrite(ctx, true);
}
+static void
+pg_decode_stream_sequence(LogicalDecodingContext *ctx, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ bool transactional,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ TestDecodingData *data = ctx->output_plugin_private;
+
+ /* return if requested to skip_sequences */
+ if (data->skip_sequences)
+ return;
+
+ OutputPluginPrepareWrite(ctx, true);
+ appendStringInfoString(ctx->out, "streaming sequence ");
+ appendStringInfoString(ctx->out,
+ quote_qualified_identifier(get_namespace_name(get_rel_namespace(RelationGetRelid(rel))),
+ RelationGetRelationName(rel)));
+ appendStringInfo(ctx->out, ": transactional:%d last_value:%zu log_cnt:%zu is_called:%d",
+ transactional, last_value, log_cnt, is_called);
+ OutputPluginWrite(ctx, true);
+}
+
/*
* In streaming mode, we don't display the detailed information of Truncate.
* See pg_decode_stream_change.
--
2.31.1
[text/x-patch] 0004-tweak-test_decoding-tests-20211222.patch (6.5K, ../../[email protected]/5-0004-tweak-test_decoding-tests-20211222.patch)
download | inline diff:
From ee0f4ffe25d6ced57cd2822616ceb131407aeb5d Mon Sep 17 00:00:00 2001
From: Tomas Vondra <[email protected]>
Date: Wed, 22 Dec 2021 03:53:10 +0100
Subject: [PATCH 4/6] tweak test_decoding tests
---
contrib/test_decoding/expected/sequence.out | 48 +++++++++++++--------
1 file changed, 29 insertions(+), 19 deletions(-)
diff --git a/contrib/test_decoding/expected/sequence.out b/contrib/test_decoding/expected/sequence.out
index c3c2a18a35..971373e083 100644
--- a/contrib/test_decoding/expected/sequence.out
+++ b/contrib/test_decoding/expected/sequence.out
@@ -93,13 +93,16 @@ SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'inc
BEGIN
sequence public.test_sequence: transactional:1 last_value:1 log_cnt:0 is_called:0
COMMIT
- sequence public.test_sequence: transactional:0 last_value:33 log_cnt:0 is_called:1
+ sequence public.test_sequence: transactional:0 last_value:1 log_cnt:0 is_called:1
+ sequence public.test_sequence: transactional:0 last_value:2 log_cnt:0 is_called:1
+ sequence public.test_sequence: transactional:0 last_value:3 log_cnt:0 is_called:1
+ sequence public.test_sequence: transactional:0 last_value:4 log_cnt:0 is_called:1
BEGIN
sequence public.test_sequence: transactional:1 last_value:4 log_cnt:0 is_called:1
COMMIT
- sequence public.test_sequence: transactional:0 last_value:334 log_cnt:0 is_called:1
+ sequence public.test_sequence: transactional:0 last_value:14 log_cnt:0 is_called:1
BEGIN
- sequence public.test_sequence: transactional:1 last_value:14 log_cnt:32 is_called:1
+ sequence public.test_sequence: transactional:1 last_value:14 log_cnt:0 is_called:1
COMMIT
BEGIN
sequence public.test_sequence: transactional:1 last_value:14 log_cnt:0 is_called:1
@@ -107,14 +110,14 @@ SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'inc
BEGIN
sequence public.test_sequence: transactional:1 last_value:4000 log_cnt:0 is_called:0
COMMIT
- sequence public.test_sequence: transactional:0 last_value:4320 log_cnt:0 is_called:1
+ sequence public.test_sequence: transactional:0 last_value:4000 log_cnt:0 is_called:1
sequence public.test_sequence: transactional:0 last_value:3500 log_cnt:0 is_called:1
- sequence public.test_sequence: transactional:0 last_value:3830 log_cnt:0 is_called:1
+ sequence public.test_sequence: transactional:0 last_value:3510 log_cnt:0 is_called:1
sequence public.test_sequence: transactional:0 last_value:3500 log_cnt:0 is_called:1
- sequence public.test_sequence: transactional:0 last_value:3830 log_cnt:0 is_called:1
+ sequence public.test_sequence: transactional:0 last_value:3510 log_cnt:0 is_called:1
sequence public.test_sequence: transactional:0 last_value:3500 log_cnt:0 is_called:0
- sequence public.test_sequence: transactional:0 last_value:3820 log_cnt:0 is_called:1
-(24 rows)
+ sequence public.test_sequence: transactional:0 last_value:3500 log_cnt:0 is_called:1
+(27 rows)
DROP SEQUENCE test_sequence;
-- rollback on sequence creation and update
@@ -226,7 +229,7 @@ ROLLBACK;
SELECT * from test_table_a_seq;
last_value | log_cnt | is_called
------------+---------+-----------
- 3003 | 30 | t
+ 3003 | 0 | t
(1 row)
SELECT nextval('test_table_a_seq');
@@ -242,12 +245,17 @@ SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'inc
BEGIN
sequence public.test_table_a_seq: transactional:1 last_value:1 log_cnt:0 is_called:0
COMMIT
- sequence public.test_table_a_seq: transactional:0 last_value:33 log_cnt:0 is_called:1
+ sequence public.test_table_a_seq: transactional:0 last_value:1 log_cnt:0 is_called:1
+ sequence public.test_table_a_seq: transactional:0 last_value:2 log_cnt:0 is_called:1
+ sequence public.test_table_a_seq: transactional:0 last_value:3 log_cnt:0 is_called:1
sequence public.test_table_a_seq: transactional:0 last_value:3000 log_cnt:0 is_called:1
- sequence public.test_table_a_seq: transactional:0 last_value:3033 log_cnt:0 is_called:1
+ sequence public.test_table_a_seq: transactional:0 last_value:3001 log_cnt:0 is_called:1
+ sequence public.test_table_a_seq: transactional:0 last_value:3002 log_cnt:0 is_called:1
+ sequence public.test_table_a_seq: transactional:0 last_value:3003 log_cnt:0 is_called:1
+ sequence public.test_table_a_seq: transactional:0 last_value:3004 log_cnt:0 is_called:1
BEGIN
COMMIT
-(8 rows)
+(13 rows)
-- savepoint test on table with serial column
BEGIN;
@@ -260,15 +268,17 @@ ROLLBACK TO SAVEPOINT a;
DROP TABLE test_table;
COMMIT;
SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '0', 'skip-sequences', '0');
- data
----------------------------------------------------------------------------------------
+ data
+--------------------------------------------------------------------------------------
BEGIN
sequence public.test_table_a_seq: transactional:1 last_value:1 log_cnt:0 is_called:0
- sequence public.test_table_a_seq: transactional:1 last_value:33 log_cnt:0 is_called:1
+ sequence public.test_table_a_seq: transactional:1 last_value:1 log_cnt:0 is_called:1
table public.test_table: INSERT: a[integer]:1 b[integer]:100
+ sequence public.test_table_a_seq: transactional:1 last_value:2 log_cnt:0 is_called:1
table public.test_table: INSERT: a[integer]:2 b[integer]:200
+ sequence public.test_table_a_seq: transactional:1 last_value:3 log_cnt:0 is_called:1
COMMIT
-(6 rows)
+(8 rows)
-- savepoint test on table with serial column
BEGIN;
@@ -303,7 +313,7 @@ ROLLBACK TO SAVEPOINT a;
SELECT * FROM test_sequence;
last_value | log_cnt | is_called
------------+---------+-----------
- 3001 | 32 | t
+ 3001 | 0 | t
(1 row)
DROP SEQUENCE test_sequence;
@@ -313,9 +323,9 @@ SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'inc
--------------------------------------------------------------------------------------
BEGIN
sequence public.test_sequence: transactional:1 last_value:1 log_cnt:0 is_called:0
- sequence public.test_sequence: transactional:1 last_value:33 log_cnt:0 is_called:1
+ sequence public.test_sequence: transactional:1 last_value:1 log_cnt:0 is_called:1
sequence public.test_sequence: transactional:1 last_value:3000 log_cnt:0 is_called:1
- sequence public.test_sequence: transactional:1 last_value:3033 log_cnt:0 is_called:1
+ sequence public.test_sequence: transactional:1 last_value:3001 log_cnt:0 is_called:1
COMMIT
(6 rows)
--
2.31.1
[text/x-patch] 0005-Add-support-for-decoding-sequences-to-built-20211222.patch (76.8K, ../../[email protected]/6-0005-Add-support-for-decoding-sequences-to-built-20211222.patch)
download | inline diff:
From 4eab8773976137234e9fec46b0bc01b5fdde7e79 Mon Sep 17 00:00:00 2001
From: Tomas Vondra <[email protected]>
Date: Tue, 14 Dec 2021 00:29:09 +0100
Subject: [PATCH 5/6] Add support for decoding sequences to built-in
replication
---
doc/src/sgml/catalogs.sgml | 71 ++++
doc/src/sgml/ref/alter_publication.sgml | 24 +-
doc/src/sgml/ref/alter_subscription.sgml | 4 +-
src/backend/catalog/pg_publication.c | 149 ++++++++-
src/backend/catalog/system_views.sql | 10 +
src/backend/commands/publicationcmds.c | 350 +++++++++++++++++++-
src/backend/commands/sequence.c | 79 +++++
src/backend/commands/subscriptioncmds.c | 272 +++++++++++++++
src/backend/executor/execReplication.c | 2 +-
src/backend/nodes/copyfuncs.c | 1 +
src/backend/nodes/equalfuncs.c | 1 +
src/backend/parser/gram.y | 32 ++
src/backend/replication/logical/proto.c | 52 +++
src/backend/replication/logical/tablesync.c | 118 ++++++-
src/backend/replication/logical/worker.c | 60 ++++
src/backend/replication/pgoutput/pgoutput.c | 85 ++++-
src/backend/utils/cache/relcache.c | 4 +-
src/bin/psql/tab-complete.c | 6 +-
src/include/catalog/pg_proc.dat | 5 +
src/include/catalog/pg_publication.h | 14 +
src/include/commands/sequence.h | 1 +
src/include/nodes/parsenodes.h | 6 +
src/include/replication/logicalproto.h | 19 ++
src/include/replication/pgoutput.h | 1 +
src/test/regress/expected/rules.out | 8 +
src/test/subscription/t/027_sequences.pl | 196 +++++++++++
26 files changed, 1538 insertions(+), 32 deletions(-)
create mode 100644 src/test/subscription/t/027_sequences.pl
diff --git a/doc/src/sgml/catalogs.sgml b/doc/src/sgml/catalogs.sgml
index 03e2537b07..0dfeb36f03 100644
--- a/doc/src/sgml/catalogs.sgml
+++ b/doc/src/sgml/catalogs.sgml
@@ -9514,6 +9514,11 @@ SCRAM-SHA-256$<replaceable><iteration count></replaceable>:<replaceable>&l
<entry>prepared transactions</entry>
</row>
+ <row>
+ <entry><link linkend="view-pg-publication-sequences"><structname>pg_publication_sequences</structname></link></entry>
+ <entry>publications and their associated sequences</entry>
+ </row>
+
<row>
<entry><link linkend="view-pg-publication-tables"><structname>pg_publication_tables</structname></link></entry>
<entry>publications and their associated tables</entry>
@@ -11349,6 +11354,72 @@ SELECT * FROM pg_locks pl LEFT JOIN pg_prepared_xacts ppx
</sect1>
+ <sect1 id="view-pg-publication-sequences">
+ <title><structname>pg_publication_sequences</structname></title>
+
+ <indexterm zone="view-pg-publication-sequences">
+ <primary>pg_publication_sequences</primary>
+ </indexterm>
+
+ <para>
+ The view <structname>pg_publication_sequences</structname> provides
+ information about the mapping between publications and the sequences they
+ contain. Unlike the underlying catalog
+ <link linkend="catalog-pg-publication-rel"><structname>pg_publication_rel</structname></link>,
+ this view expands
+ publications defined as <literal>FOR ALL SEQUENCES</literal>, so for such
+ publications there will be a row for each eligible sequence.
+ </para>
+
+ <table>
+ <title><structname>pg_publication_sequences</structname> Columns</title>
+ <tgroup cols="1">
+ <thead>
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ Column Type
+ </para>
+ <para>
+ Description
+ </para></entry>
+ </row>
+ </thead>
+
+ <tbody>
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ <structfield>pubname</structfield> <type>name</type>
+ (references <link linkend="catalog-pg-publication"><structname>pg_publication</structname></link>.<structfield>pubname</structfield>)
+ </para>
+ <para>
+ Name of publication
+ </para></entry>
+ </row>
+
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ <structfield>schemaname</structfield> <type>name</type>
+ (references <link linkend="catalog-pg-namespace"><structname>pg_namespace</structname></link>.<structfield>nspname</structfield>)
+ </para>
+ <para>
+ Name of schema containing sequence
+ </para></entry>
+ </row>
+
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ <structfield>sequencename</structfield> <type>name</type>
+ (references <link linkend="catalog-pg-class"><structname>pg_class</structname></link>.<structfield>relname</structfield>)
+ </para>
+ <para>
+ Name of sequence
+ </para></entry>
+ </row>
+ </tbody>
+ </tgroup>
+ </table>
+ </sect1>
+
<sect1 id="view-pg-publication-tables">
<title><structname>pg_publication_tables</structname></title>
diff --git a/doc/src/sgml/ref/alter_publication.sgml b/doc/src/sgml/ref/alter_publication.sgml
index bb4ef5e5e2..4d166ad3f9 100644
--- a/doc/src/sgml/ref/alter_publication.sgml
+++ b/doc/src/sgml/ref/alter_publication.sgml
@@ -31,7 +31,9 @@ ALTER PUBLICATION <replaceable class="parameter">name</replaceable> RENAME TO <r
<phrase>where <replaceable class="parameter">publication_object</replaceable> is one of:</phrase>
TABLE [ ONLY ] <replaceable class="parameter">table_name</replaceable> [ * ] [, ... ]
+ SEQUENCE <replaceable class="parameter">sequence_name</replaceable> [ * ] [, ... ]
ALL TABLES IN SCHEMA { <replaceable class="parameter">schema_name</replaceable> | CURRENT_SCHEMA } [, ... ]
+ ALL SEQUENCE IN SCHEMA { <replaceable class="parameter">schema_name</replaceable> | CURRENT_SCHEMA } [, ... ]
</synopsis>
</refsynopsisdiv>
@@ -56,7 +58,18 @@ ALTER PUBLICATION <replaceable class="parameter">name</replaceable> RENAME TO <r
</para>
<para>
- The fourth variant of this command listed in the synopsis can change
+ The next three variants change which sequences are part of the publication.
+ The <literal>SET SEQUENCE</literal> clause will replace the list of sequences
+ in the publication with the specified one. The <literal>ADD SEQUENCE</literal>
+ and <literal>DROP SEQUENCE</literal> clauses will add and remove one or more
+ sequences from the publication. Note that adding sequences to a publication
+ that is already subscribed to will require a <literal>ALTER SUBSCRIPTION
+ ... REFRESH PUBLICATION</literal> action on the subscribing side in order
+ to become effective.
+ </para>
+
+ <para>
+ The seventh variant of this command listed in the synopsis can change
all of the publication properties specified in
<xref linkend="sql-createpublication"/>. Properties not mentioned in the
command retain their previous settings.
@@ -123,6 +136,15 @@ ALTER PUBLICATION <replaceable class="parameter">name</replaceable> RENAME TO <r
</listitem>
</varlistentry>
+ <varlistentry>
+ <term><replaceable class="parameter">sequence_name</replaceable></term>
+ <listitem>
+ <para>
+ Name of an existing sequence.
+ </para>
+ </listitem>
+ </varlistentry>
+
<varlistentry>
<term><literal>SET ( <replaceable class="parameter">publication_parameter</replaceable> [= <replaceable class="parameter">value</replaceable>] [, ... ] )</literal></term>
<listitem>
diff --git a/doc/src/sgml/ref/alter_subscription.sgml b/doc/src/sgml/ref/alter_subscription.sgml
index 0b027cc346..8f28cf03f4 100644
--- a/doc/src/sgml/ref/alter_subscription.sgml
+++ b/doc/src/sgml/ref/alter_subscription.sgml
@@ -147,7 +147,7 @@ ALTER SUBSCRIPTION <replaceable class="parameter">name</replaceable> RENAME TO <
<listitem>
<para>
Fetch missing table information from publisher. This will start
- replication of tables that were added to the subscribed-to publications
+ replication of tables and sequences that were added to the subscribed-to publications
since <command>CREATE SUBSCRIPTION</command> or
the last invocation of <command>REFRESH PUBLICATION</command>.
</para>
@@ -164,7 +164,7 @@ ALTER SUBSCRIPTION <replaceable class="parameter">name</replaceable> RENAME TO <
Specifies whether to copy pre-existing data in the publications
that are being subscribed to when the replication starts.
The default is <literal>true</literal>. (Previously-subscribed
- tables are not copied.)
+ tables and sequences are not copied.)
</para>
</listitem>
</varlistentry>
diff --git a/src/backend/catalog/pg_publication.c b/src/backend/catalog/pg_publication.c
index 62f10bcbd2..6844ac6189 100644
--- a/src/backend/catalog/pg_publication.c
+++ b/src/backend/catalog/pg_publication.c
@@ -54,7 +54,8 @@ check_publication_add_relation(Relation targetrel)
{
/* Must be a regular or partitioned table */
if (RelationGetForm(targetrel)->relkind != RELKIND_RELATION &&
- RelationGetForm(targetrel)->relkind != RELKIND_PARTITIONED_TABLE)
+ RelationGetForm(targetrel)->relkind != RELKIND_PARTITIONED_TABLE &&
+ RelationGetForm(targetrel)->relkind != RELKIND_SEQUENCE)
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("cannot add relation \"%s\" to publication",
@@ -131,7 +132,8 @@ static bool
is_publishable_class(Oid relid, Form_pg_class reltuple)
{
return (reltuple->relkind == RELKIND_RELATION ||
- reltuple->relkind == RELKIND_PARTITIONED_TABLE) &&
+ reltuple->relkind == RELKIND_PARTITIONED_TABLE ||
+ reltuple->relkind == RELKIND_SEQUENCE) &&
!IsCatalogRelationOid(relid) &&
reltuple->relpersistence == RELPERSISTENCE_PERMANENT &&
relid >= FirstNormalObjectId;
@@ -502,6 +504,11 @@ GetPublicationRelations(Oid pubid, PublicationPartOpt pub_partopt)
Form_pg_publication_rel pubrel;
pubrel = (Form_pg_publication_rel) GETSTRUCT(tup);
+
+ /* skip sequences here */
+ if (get_rel_relkind(pubrel->prrelid) == RELKIND_SEQUENCE)
+ continue;
+
result = GetPubPartitionOptionRelations(result, pub_partopt,
pubrel->prrelid);
}
@@ -516,6 +523,49 @@ GetPublicationRelations(Oid pubid, PublicationPartOpt pub_partopt)
return result;
}
+/*
+ * Gets list of relation oids for a publication (sequences only).
+ *
+ * This should only be used for normal publications, the FOR ALL TABLES
+ * should use GetAllSequencesPublicationRelations().
+ */
+List *
+GetPublicationSequenceRelations(Oid pubid)
+{
+ List *result;
+ Relation pubrelsrel;
+ ScanKeyData scankey;
+ SysScanDesc scan;
+ HeapTuple tup;
+
+ /* Find all publications associated with the relation. */
+ pubrelsrel = table_open(PublicationRelRelationId, AccessShareLock);
+
+ ScanKeyInit(&scankey,
+ Anum_pg_publication_rel_prpubid,
+ BTEqualStrategyNumber, F_OIDEQ,
+ ObjectIdGetDatum(pubid));
+
+ scan = systable_beginscan(pubrelsrel, PublicationRelPrrelidPrpubidIndexId,
+ true, NULL, 1, &scankey);
+
+ result = NIL;
+ while (HeapTupleIsValid(tup = systable_getnext(scan)))
+ {
+ Form_pg_publication_rel pubrel;
+
+ pubrel = (Form_pg_publication_rel) GETSTRUCT(tup);
+
+ if (get_rel_relkind(pubrel->prrelid) == RELKIND_SEQUENCE)
+ result = lappend_oid(result, pubrel->prrelid);
+ }
+
+ systable_endscan(scan);
+ table_close(pubrelsrel, AccessShareLock);
+
+ return result;
+}
+
/*
* Gets list of publication oids for publications marked as FOR ALL TABLES.
*/
@@ -761,6 +811,46 @@ GetAllSchemaPublicationRelations(Oid pubid, PublicationPartOpt pub_partopt)
return result;
}
+/*
+ * Gets list of all relation published by FOR ALL TABLES publication(s).
+ *
+ * If the publication publishes partition changes via their respective root
+ * partitioned tables, we must exclude partitions in favor of including the
+ * root partitioned tables.
+ */
+List *
+GetAllSequencesPublicationRelations(void)
+{
+ Relation classRel;
+ ScanKeyData key[1];
+ TableScanDesc scan;
+ HeapTuple tuple;
+ List *result = NIL;
+
+ classRel = table_open(RelationRelationId, AccessShareLock);
+
+ ScanKeyInit(&key[0],
+ Anum_pg_class_relkind,
+ BTEqualStrategyNumber, F_CHAREQ,
+ CharGetDatum(RELKIND_SEQUENCE));
+
+ scan = table_beginscan_catalog(classRel, 1, key);
+
+ while ((tuple = heap_getnext(scan, ForwardScanDirection)) != NULL)
+ {
+ Form_pg_class relForm = (Form_pg_class) GETSTRUCT(tuple);
+ Oid relid = relForm->oid;
+
+ if (is_publishable_class(relid, relForm))
+ result = lappend_oid(result, relid);
+ }
+
+ table_endscan(scan);
+
+ table_close(classRel, AccessShareLock);
+ return result;
+}
+
/*
* Get publication using oid
*
@@ -783,10 +873,12 @@ GetPublication(Oid pubid)
pub->oid = pubid;
pub->name = pstrdup(NameStr(pubform->pubname));
pub->alltables = pubform->puballtables;
+ pub->allsequences = pubform->puballsequences;
pub->pubactions.pubinsert = pubform->pubinsert;
pub->pubactions.pubupdate = pubform->pubupdate;
pub->pubactions.pubdelete = pubform->pubdelete;
pub->pubactions.pubtruncate = pubform->pubtruncate;
+ pub->pubactions.pubsequence = pubform->pubsequence;
pub->pubviaroot = pubform->pubviaroot;
ReleaseSysCache(tup);
@@ -936,3 +1028,56 @@ pg_get_publication_tables(PG_FUNCTION_ARGS)
SRF_RETURN_DONE(funcctx);
}
+
+/*
+ * Returns Oids of sequences in a publication.
+ */
+Datum
+pg_get_publication_sequences(PG_FUNCTION_ARGS)
+{
+ FuncCallContext *funcctx;
+ char *pubname = text_to_cstring(PG_GETARG_TEXT_PP(0));
+ Publication *publication;
+ List *sequences;
+
+ /* stuff done only on the first call of the function */
+ if (SRF_IS_FIRSTCALL())
+ {
+ MemoryContext oldcontext;
+
+ /* create a function context for cross-call persistence */
+ funcctx = SRF_FIRSTCALL_INIT();
+
+ /* switch to memory context appropriate for multiple function calls */
+ oldcontext = MemoryContextSwitchTo(funcctx->multi_call_memory_ctx);
+
+ publication = GetPublicationByName(pubname, false);
+
+ /*
+ * Publications support partitioned tables, although all changes are
+ * replicated using leaf partition identity and schema, so we only
+ * need those.
+ */
+ if (publication->allsequences)
+ sequences = GetAllSequencesPublicationRelations();
+ else
+ sequences = GetPublicationSequenceRelations(publication->oid);
+
+ funcctx->user_fctx = (void *) sequences;
+
+ MemoryContextSwitchTo(oldcontext);
+ }
+
+ /* stuff done on every call of the function */
+ funcctx = SRF_PERCALL_SETUP();
+ sequences = (List *) funcctx->user_fctx;
+
+ if (funcctx->call_cntr < list_length(sequences))
+ {
+ Oid relid = list_nth_oid(sequences, funcctx->call_cntr);
+
+ SRF_RETURN_NEXT(funcctx, ObjectIdGetDatum(relid));
+ }
+
+ SRF_RETURN_DONE(funcctx);
+}
diff --git a/src/backend/catalog/system_views.sql b/src/backend/catalog/system_views.sql
index 61b515cdb8..0df91d7a5a 100644
--- a/src/backend/catalog/system_views.sql
+++ b/src/backend/catalog/system_views.sql
@@ -372,6 +372,16 @@ CREATE VIEW pg_publication_tables AS
pg_class C JOIN pg_namespace N ON (N.oid = C.relnamespace)
WHERE C.oid = GPT.relid;
+CREATE VIEW pg_publication_sequences AS
+ SELECT
+ P.pubname AS pubname,
+ N.nspname AS schemaname,
+ C.relname AS sequencename
+ FROM pg_publication P,
+ LATERAL pg_get_publication_sequences(P.pubname) GPT,
+ pg_class C JOIN pg_namespace N ON (N.oid = C.relnamespace)
+ WHERE C.oid = GPT.relid;
+
CREATE VIEW pg_locks AS
SELECT * FROM pg_lock_status() AS L;
diff --git a/src/backend/commands/publicationcmds.c b/src/backend/commands/publicationcmds.c
index 404bb5d0c8..2c183d63f7 100644
--- a/src/backend/commands/publicationcmds.c
+++ b/src/backend/commands/publicationcmds.c
@@ -16,6 +16,7 @@
#include "access/genam.h"
#include "access/htup_details.h"
+#include "access/relation.h"
#include "access/table.h"
#include "access/xact.h"
#include "catalog/catalog.h"
@@ -59,6 +60,12 @@ static void PublicationAddSchemas(Oid pubid, List *schemas, bool if_not_exists,
AlterPublicationStmt *stmt);
static void PublicationDropSchemas(Oid pubid, List *schemas, bool missing_ok);
+static List *OpenSequenceList(List *sequences);
+static void CloseSequenceList(List *rels);
+static void PublicationAddSequences(Oid pubid, List *rels, bool if_not_exists,
+ AlterPublicationStmt *stmt);
+static void PublicationDropSequences(Oid pubid, List *rels, bool missing_ok);
+
static void
parse_publication_options(ParseState *pstate,
List *options,
@@ -77,6 +84,7 @@ parse_publication_options(ParseState *pstate,
pubactions->pubupdate = true;
pubactions->pubdelete = true;
pubactions->pubtruncate = true;
+ pubactions->pubsequence = true;
*publish_via_partition_root = false;
/* Parse options */
@@ -101,6 +109,7 @@ parse_publication_options(ParseState *pstate,
pubactions->pubupdate = false;
pubactions->pubdelete = false;
pubactions->pubtruncate = false;
+ pubactions->pubsequence = false;
*publish_given = true;
publish = defGetString(defel);
@@ -123,6 +132,8 @@ parse_publication_options(ParseState *pstate,
pubactions->pubdelete = true;
else if (strcmp(publish_opt, "truncate") == 0)
pubactions->pubtruncate = true;
+ else if (strcmp(publish_opt, "sequence") == 0)
+ pubactions->pubsequence = true;
else
ereport(ERROR,
(errcode(ERRCODE_SYNTAX_ERROR),
@@ -149,7 +160,9 @@ parse_publication_options(ParseState *pstate,
*/
static void
ObjectsInPublicationToOids(List *pubobjspec_list, ParseState *pstate,
- List **rels, List **schemas)
+ List **tables, List **sequences,
+ List **tables_schemas, List **sequences_schemas,
+ List **schemas)
{
ListCell *cell;
PublicationObjSpec *pubobj;
@@ -167,12 +180,23 @@ ObjectsInPublicationToOids(List *pubobjspec_list, ParseState *pstate,
switch (pubobj->pubobjtype)
{
case PUBLICATIONOBJ_TABLE:
- *rels = lappend(*rels, pubobj->pubtable);
+ *tables = lappend(*tables, pubobj->pubtable);
+ break;
+ case PUBLICATIONOBJ_SEQUENCE:
+ *sequences = lappend(*sequences, pubobj->pubtable);
break;
case PUBLICATIONOBJ_TABLE_IN_SCHEMA:
schemaid = get_namespace_oid(pubobj->name, false);
/* Filter out duplicates if user specifies "sch1, sch1" */
+ *tables_schemas = list_append_unique_oid(*tables_schemas, schemaid);
+ *schemas = list_append_unique_oid(*schemas, schemaid);
+ break;
+ case PUBLICATIONOBJ_SEQUENCE_IN_SCHEMA:
+ schemaid = get_namespace_oid(pubobj->name, false);
+
+ /* Filter out duplicates if user specifies "sch1, sch1" */
+ *sequences_schemas = list_append_unique_oid(*sequences_schemas, schemaid);
*schemas = list_append_unique_oid(*schemas, schemaid);
break;
case PUBLICATIONOBJ_TABLE_IN_CUR_SCHEMA:
@@ -186,6 +210,21 @@ ObjectsInPublicationToOids(List *pubobjspec_list, ParseState *pstate,
list_free(search_path);
/* Filter out duplicates if user specifies "sch1, sch1" */
+ *tables_schemas = list_append_unique_oid(*tables_schemas, schemaid);
+ *schemas = list_append_unique_oid(*schemas, schemaid);
+ break;
+ case PUBLICATIONOBJ_SEQUENCE_IN_CUR_SCHEMA:
+ search_path = fetch_search_path(false);
+ if (search_path == NIL) /* nothing valid in search_path? */
+ ereport(ERROR,
+ errcode(ERRCODE_UNDEFINED_SCHEMA),
+ errmsg("no schema has been selected for CURRENT_SCHEMA"));
+
+ schemaid = linitial_oid(search_path);
+ list_free(search_path);
+
+ /* Filter out duplicates if user specifies "sch1, sch1" */
+ *sequences_schemas = list_append_unique_oid(*sequences_schemas, schemaid);
*schemas = list_append_unique_oid(*schemas, schemaid);
break;
default:
@@ -251,7 +290,10 @@ CreatePublication(ParseState *pstate, CreatePublicationStmt *stmt)
bool publish_via_partition_root_given;
bool publish_via_partition_root;
AclResult aclresult;
- List *relations = NIL;
+ List *tables = NIL;
+ List *sequences = NIL;
+ List *tables_schemaidlist = NIL;
+ List *sequences_schemaidlist = NIL;
List *schemaidlist = NIL;
/* must have CREATE privilege on database */
@@ -306,6 +348,8 @@ CreatePublication(ParseState *pstate, CreatePublicationStmt *stmt)
BoolGetDatum(pubactions.pubdelete);
values[Anum_pg_publication_pubtruncate - 1] =
BoolGetDatum(pubactions.pubtruncate);
+ values[Anum_pg_publication_pubsequence - 1] =
+ BoolGetDatum(pubactions.pubsequence);
values[Anum_pg_publication_pubviaroot - 1] =
BoolGetDatum(publish_via_partition_root);
@@ -330,26 +374,40 @@ CreatePublication(ParseState *pstate, CreatePublicationStmt *stmt)
}
else
{
- ObjectsInPublicationToOids(stmt->pubobjects, pstate, &relations,
+ ObjectsInPublicationToOids(stmt->pubobjects, pstate,
+ &tables, &sequences,
+ &tables_schemaidlist,
+ &sequences_schemaidlist,
&schemaidlist);
/* FOR ALL TABLES IN SCHEMA requires superuser */
- if (list_length(schemaidlist) > 0 && !superuser())
+ if (list_length(tables_schemaidlist) > 0 && !superuser())
ereport(ERROR,
errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
errmsg("must be superuser to create FOR ALL TABLES IN SCHEMA publication"));
- if (list_length(relations) > 0)
+ if (list_length(tables) > 0)
{
List *rels;
- rels = OpenTableList(relations);
- CheckObjSchemaNotAlreadyInPublication(rels, schemaidlist,
+ rels = OpenTableList(tables);
+ CheckObjSchemaNotAlreadyInPublication(rels, tables_schemaidlist,
PUBLICATIONOBJ_TABLE);
PublicationAddTables(puboid, rels, true, NULL);
CloseTableList(rels);
}
+ if (list_length(sequences) > 0)
+ {
+ List *rels;
+
+ rels = OpenSequenceList(sequences);
+ CheckObjSchemaNotAlreadyInPublication(rels, sequences_schemaidlist,
+ PUBLICATIONOBJ_SEQUENCE);
+ PublicationAddTables(puboid, rels, true, NULL);
+ CloseSequenceList(rels);
+ }
+
if (list_length(schemaidlist) > 0)
{
/*
@@ -651,12 +709,13 @@ AlterPublicationSchemas(AlterPublicationStmt *stmt,
*/
static void
CheckAlterPublication(AlterPublicationStmt *stmt, HeapTuple tup,
- List *tables, List *schemaidlist)
+ List *tables, List *tables_schemaidlist,
+ List *sequences, List *sequences_schemaidlist)
{
Form_pg_publication pubform = (Form_pg_publication) GETSTRUCT(tup);
if ((stmt->action == DEFELEM_ADD || stmt->action == DEFELEM_SET) &&
- schemaidlist && !superuser())
+ (tables_schemaidlist || sequences_schemaidlist) && !superuser())
ereport(ERROR,
(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
errmsg("must be superuser to add or set schemas")));
@@ -665,13 +724,24 @@ CheckAlterPublication(AlterPublicationStmt *stmt, HeapTuple tup,
* Check that user is allowed to manipulate the publication tables in
* schema
*/
- if (schemaidlist && pubform->puballtables)
+ if (tables_schemaidlist && pubform->puballtables)
ereport(ERROR,
(errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
errmsg("publication \"%s\" is defined as FOR ALL TABLES",
NameStr(pubform->pubname)),
errdetail("Tables from schema cannot be added to, dropped from, or set on FOR ALL TABLES publications.")));
+ /*
+ * Check that user is allowed to manipulate the publication sequences in
+ * schema
+ */
+ if (sequences_schemaidlist && pubform->puballsequences)
+ ereport(ERROR,
+ (errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
+ errmsg("publication \"%s\" is defined as FOR ALL SEQUENCES",
+ NameStr(pubform->pubname)),
+ errdetail("Sequences from schema cannot be added to, dropped from, or set on FOR ALL SEQUENCES publications.")));
+
/* Check that user is allowed to manipulate the publication tables. */
if (tables && pubform->puballtables)
ereport(ERROR,
@@ -679,6 +749,108 @@ CheckAlterPublication(AlterPublicationStmt *stmt, HeapTuple tup,
errmsg("publication \"%s\" is defined as FOR ALL TABLES",
NameStr(pubform->pubname)),
errdetail("Tables cannot be added to or dropped from FOR ALL TABLES publications.")));
+
+ /* Check that user is allowed to manipulate the publication tables. */
+ if (sequences && pubform->puballsequences)
+ ereport(ERROR,
+ (errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
+ errmsg("publication \"%s\" is defined as FOR ALL SEQUENCES",
+ NameStr(pubform->pubname)),
+ errdetail("Sequences cannot be added to or dropped from FOR ALL SEQUENCES publications.")));
+}
+
+/*
+ * Add or remove sequence to/from publication.
+ */
+static void
+AlterPublicationSequences(AlterPublicationStmt *stmt, HeapTuple tup,
+ List *sequences, List *schemaidlist)
+{
+ List *rels = NIL;
+ Form_pg_publication pubform = (Form_pg_publication) GETSTRUCT(tup);
+ Oid pubid = pubform->oid;
+
+ /*
+ * It is quite possible that for the SET case user has not specified any
+ * tables in which case we need to remove all the existing tables.
+ */
+ if (!sequences && stmt->action != DEFELEM_SET)
+ return;
+
+ rels = OpenSequenceList(sequences);
+
+ if (stmt->action == DEFELEM_ADD)
+ {
+ List *schemas = NIL;
+
+ /*
+ * Check if the relation is member of the existing schema in the
+ * publication or member of the schema list specified.
+ */
+ schemas = list_concat_copy(schemaidlist, GetPublicationSchemas(pubid));
+ CheckObjSchemaNotAlreadyInPublication(rels, schemas,
+ PUBLICATIONOBJ_SEQUENCE);
+ PublicationAddSequences(pubid, rels, false, stmt);
+ }
+ else if (stmt->action == DEFELEM_DROP)
+ PublicationDropSequences(pubid, rels, false);
+ else /* DEFELEM_SET */
+ {
+ List *oldrelids = GetPublicationRelations(pubid,
+ PUBLICATION_PART_ROOT);
+ List *delrels = NIL;
+ ListCell *oldlc;
+
+ CheckObjSchemaNotAlreadyInPublication(rels, schemaidlist,
+ PUBLICATIONOBJ_SEQUENCE);
+
+ /* Calculate which relations to drop. */
+ foreach(oldlc, oldrelids)
+ {
+ Oid oldrelid = lfirst_oid(oldlc);
+ ListCell *newlc;
+ bool found = false;
+
+ foreach(newlc, rels)
+ {
+ PublicationRelInfo *newpubrel;
+
+ newpubrel = (PublicationRelInfo *) lfirst(newlc);
+ if (RelationGetRelid(newpubrel->relation) == oldrelid)
+ {
+ found = true;
+ break;
+ }
+ }
+ /* Not yet in the list, open it and add to the list */
+ if (!found)
+ {
+ Relation oldrel;
+ PublicationRelInfo *pubrel;
+
+ /* Wrap relation into PublicationRelInfo */
+ oldrel = table_open(oldrelid, ShareUpdateExclusiveLock);
+
+ pubrel = palloc(sizeof(PublicationRelInfo));
+ pubrel->relation = oldrel;
+
+ delrels = lappend(delrels, pubrel);
+ }
+ }
+
+ /* And drop them. */
+ PublicationDropSequences(pubid, delrels, true);
+
+ /*
+ * Don't bother calculating the difference for adding, we'll catch and
+ * skip existing ones when doing catalog update.
+ */
+ PublicationAddSequences(pubid, rels, true, stmt);
+
+ CloseSequenceList(delrels);
+ }
+
+ CloseSequenceList(rels);
}
/*
@@ -716,13 +888,21 @@ AlterPublication(ParseState *pstate, AlterPublicationStmt *stmt)
AlterPublicationOptions(pstate, stmt, rel, tup);
else
{
- List *relations = NIL;
+ List *tables = NIL;
+ List *sequences = NIL;
+ List *tables_schemaidlist = NIL;
+ List *sequences_schemaidlist = NIL;
List *schemaidlist = NIL;
- ObjectsInPublicationToOids(stmt->pubobjects, pstate, &relations,
+ ObjectsInPublicationToOids(stmt->pubobjects, pstate,
+ &tables, &sequences,
+ &tables_schemaidlist,
+ &sequences_schemaidlist,
&schemaidlist);
- CheckAlterPublication(stmt, tup, relations, schemaidlist);
+ CheckAlterPublication(stmt, tup,
+ tables, tables_schemaidlist,
+ sequences, sequences_schemaidlist);
/*
* Lock the publication so nobody else can do anything with it. This
@@ -747,7 +927,9 @@ AlterPublication(ParseState *pstate, AlterPublicationStmt *stmt)
errmsg("publication \"%s\" does not exist",
stmt->pubname));
- AlterPublicationTables(stmt, tup, relations, schemaidlist);
+ AlterPublicationTables(stmt, tup, tables, tables_schemaidlist);
+ AlterPublicationSequences(stmt, tup, sequences, sequences_schemaidlist);
+
AlterPublicationSchemas(stmt, tup, schemaidlist);
}
@@ -1155,6 +1337,144 @@ PublicationDropSchemas(Oid pubid, List *schemas, bool missing_ok)
}
}
+/*
+ * Open relations specified by a PublicationTable list.
+ * In the returned list of PublicationRelInfo, tables are locked
+ * in ShareUpdateExclusiveLock mode in order to add them to a publication.
+ */
+static List *
+OpenSequenceList(List *sequences)
+{
+ List *relids = NIL;
+ List *rels = NIL;
+ ListCell *lc;
+
+ /*
+ * Open, share-lock, and check all the explicitly-specified relations
+ */
+ foreach(lc, sequences)
+ {
+ PublicationTable *s = lfirst_node(PublicationTable, lc);
+ Relation rel;
+ Oid myrelid;
+ PublicationRelInfo *pub_rel;
+
+ /* Allow query cancel in case this takes a long time */
+ CHECK_FOR_INTERRUPTS();
+
+ rel = table_openrv(s->relation, ShareUpdateExclusiveLock);
+ myrelid = RelationGetRelid(rel);
+
+ /*
+ * Filter out duplicates if user specifies "foo, foo".
+ *
+ * Note that this algorithm is known to not be very efficient (O(N^2))
+ * but given that it only works on list of tables given to us by user
+ * it's deemed acceptable.
+ */
+ if (list_member_oid(relids, myrelid))
+ {
+ table_close(rel, ShareUpdateExclusiveLock);
+ continue;
+ }
+
+ pub_rel = palloc(sizeof(PublicationRelInfo));
+ pub_rel->relation = rel;
+ rels = lappend(rels, pub_rel);
+ relids = lappend_oid(relids, myrelid);
+ }
+
+ list_free(relids);
+
+ return rels;
+}
+
+/*
+ * Close all relations in the list.
+ */
+static void
+CloseSequenceList(List *rels)
+{
+ ListCell *lc;
+
+ foreach(lc, rels)
+ {
+ PublicationRelInfo *pub_rel;
+
+ pub_rel = (PublicationRelInfo *) lfirst(lc);
+ table_close(pub_rel->relation, NoLock);
+ }
+}
+
+/*
+ * Add listed tables to the publication.
+ */
+static void
+PublicationAddSequences(Oid pubid, List *rels, bool if_not_exists,
+ AlterPublicationStmt *stmt)
+{
+ ListCell *lc;
+
+ Assert(!stmt || !stmt->for_all_sequences);
+
+ foreach(lc, rels)
+ {
+ PublicationRelInfo *pub_rel = (PublicationRelInfo *) lfirst(lc);
+ Relation rel = pub_rel->relation;
+ ObjectAddress obj;
+
+ /* Must be owner of the sequence or superuser. */
+ if (!pg_class_ownercheck(RelationGetRelid(rel), GetUserId()))
+ aclcheck_error(ACLCHECK_NOT_OWNER, get_relkind_objtype(rel->rd_rel->relkind),
+ RelationGetRelationName(rel));
+
+ obj = publication_add_relation(pubid, pub_rel, if_not_exists);
+ if (stmt)
+ {
+ EventTriggerCollectSimpleCommand(obj, InvalidObjectAddress,
+ (Node *) stmt);
+
+ InvokeObjectPostCreateHook(PublicationRelRelationId,
+ obj.objectId, 0);
+ }
+ }
+}
+
+/*
+ * Remove listed sequences from the publication.
+ */
+static void
+PublicationDropSequences(Oid pubid, List *rels, bool missing_ok)
+{
+ ObjectAddress obj;
+ ListCell *lc;
+ Oid prid;
+
+ foreach(lc, rels)
+ {
+ Relation rel = (Relation) lfirst(lc);
+ Oid relid = RelationGetRelid(rel);
+
+ prid = GetSysCacheOid2(PUBLICATIONRELMAP, Anum_pg_publication_rel_oid,
+ ObjectIdGetDatum(relid),
+ ObjectIdGetDatum(pubid));
+ if (!OidIsValid(prid))
+ {
+ if (missing_ok)
+ continue;
+
+ ereport(ERROR,
+ (errcode(ERRCODE_UNDEFINED_OBJECT),
+ errmsg("relation \"%s\" is not part of the publication",
+ RelationGetRelationName(rel))));
+ }
+
+ ObjectAddressSet(obj, PublicationRelRelationId, prid);
+ performDeletion(&obj, DROP_CASCADE, 0);
+ }
+}
+
+
/*
* Internal workhorse for changing a publication owner
*/
diff --git a/src/backend/commands/sequence.c b/src/backend/commands/sequence.c
index a5435910a2..9a689e6305 100644
--- a/src/backend/commands/sequence.c
+++ b/src/backend/commands/sequence.c
@@ -339,6 +339,85 @@ ResetSequence(Oid seq_relid)
relation_close(seq_rel, NoLock);
}
+/*
+ * Reset a sequence to its initial value.
+ *
+ * The change is made transactionally, so that on failure of the current
+ * transaction, the sequence will be restored to its previous state.
+ * We do that by creating a whole new relfilenode for the sequence; so this
+ * works much like the rewriting forms of ALTER TABLE.
+ *
+ * Caller is assumed to have acquired AccessExclusiveLock on the sequence,
+ * which must not be released until end of transaction. Caller is also
+ * responsible for permissions checking.
+ */
+void
+ResetSequence2(Oid seq_relid, int64 last_value, int64 log_cnt, bool is_called)
+{
+ Relation seq_rel;
+ SeqTable elm;
+ Form_pg_sequence_data seq;
+ Buffer buf;
+ HeapTupleData seqdatatuple;
+ HeapTuple tuple;
+
+ /*
+ * Read the old sequence. This does a bit more work than really
+ * necessary, but it's simple, and we do want to double-check that it's
+ * indeed a sequence.
+ */
+ init_sequence(seq_relid, &elm, &seq_rel);
+ (void) read_seq_tuple(seq_rel, &buf, &seqdatatuple);
+
+ /*
+ * Copy the existing sequence tuple.
+ */
+ tuple = heap_copytuple(&seqdatatuple);
+
+ /* Now we're done with the old page */
+ UnlockReleaseBuffer(buf);
+
+ /*
+ * Modify the copied tuple to execute the restart (compare the RESTART
+ * action in AlterSequence)
+ */
+ seq = (Form_pg_sequence_data) GETSTRUCT(tuple);
+ seq->last_value = last_value;
+ seq->is_called = is_called;
+ seq->log_cnt = log_cnt;
+
+ /*
+ * Create a new storage file for the sequence.
+ */
+ RelationSetNewRelfilenode(seq_rel, seq_rel->rd_rel->relpersistence);
+
+ /*
+ * Ensure sequence's relfrozenxid is at 0, since it won't contain any
+ * unfrozen XIDs. Same with relminmxid, since a sequence will never
+ * contain multixacts.
+ */
+ Assert(seq_rel->rd_rel->relfrozenxid == InvalidTransactionId);
+ Assert(seq_rel->rd_rel->relminmxid == InvalidMultiXactId);
+
+ /*
+ * Insert the modified tuple into the new storage file.
+ *
+ * XXX Maybe this should also use created=true, just like the other places
+ * calling fill_seq_with_data. That's probably needed for correct cascading
+ * replication.
+ *
+ * XXX That'd mean all fill_seq_with_data callers use created=true, making
+ * the parameter unnecessary.
+ */
+ fill_seq_with_data(seq_rel, tuple);
+
+ /* Clear local cache so that we don't think we have cached numbers */
+ /* Note that we do not change the currval() state */
+ elm->cached = elm->last;
+
+ relation_close(seq_rel, NoLock);
+}
+
/*
* Initialize a sequence's relation with the specified tuple as content
*/
diff --git a/src/backend/commands/subscriptioncmds.c b/src/backend/commands/subscriptioncmds.c
index 2b658080fe..fc6374c445 100644
--- a/src/backend/commands/subscriptioncmds.c
+++ b/src/backend/commands/subscriptioncmds.c
@@ -85,6 +85,7 @@ typedef struct SubOpts
} SubOpts;
static List *fetch_table_list(WalReceiverConn *wrconn, List *publications);
+static List *fetch_sequence_list(WalReceiverConn *wrconn, List *publications);
static void check_duplicates_in_publist(List *publist, Datum *datums);
static List *merge_publications(List *oldpublist, List *newpublist, bool addpub, const char *subname);
static void ReportSlotConnectionError(List *rstates, Oid subid, char *slotname, char *err);
@@ -496,6 +497,7 @@ CreateSubscription(ParseState *pstate, CreateSubscriptionStmt *stmt,
char *err;
WalReceiverConn *wrconn;
List *tables;
+ List *sequences;
ListCell *lc;
char table_state;
@@ -534,6 +536,26 @@ CreateSubscription(ParseState *pstate, CreateSubscriptionStmt *stmt,
InvalidXLogRecPtr);
}
+ /*
+ * Get the sequence list from publisher and build local sequence
+ * status info.
+ */
+ sequences = fetch_sequence_list(wrconn, publications);
+ foreach(lc, sequences)
+ {
+ RangeVar *rv = (RangeVar *) lfirst(lc);
+ Oid relid;
+
+ relid = RangeVarGetRelid(rv, AccessShareLock, false);
+
+ /* Check for supported relkind. */
+ CheckSubscriptionRelkind(get_rel_relkind(relid),
+ rv->schemaname, rv->relname);
+
+ AddSubscriptionRelState(subid, relid, table_state,
+ InvalidXLogRecPtr);
+ }
+
/*
* If requested, create permanent slot for the subscription. We
* won't use the initial snapshot for anything, so no need to
@@ -706,6 +728,10 @@ AlterSubscription_refresh(Subscription *sub, bool copy_data)
{
Oid relid = subrel_local_oids[off];
+ /* XXX ignore sequences - maybe do this in GetSubscriptionRelations? */
+ if (get_rel_relkind(relid) == RELKIND_SEQUENCE)
+ continue;
+
if (!bsearch(&relid, pubrel_local_oids,
list_length(pubrel_names), sizeof(Oid), oid_cmp))
{
@@ -797,6 +823,183 @@ AlterSubscription_refresh(Subscription *sub, bool copy_data)
ReplicationSlotDropAtPubNode(wrconn, syncslotname, true);
}
}
+
+ /*
+ * XXX now do the same thing for sequences, maybe before the preceding
+ * block, or earlier?
+ */
+
+ /* Get the table list from publisher. */
+ pubrel_names = fetch_sequence_list(wrconn, sub->publications);
+
+ /* Get local table list. */
+ subrel_states = GetSubscriptionRelations(sub->oid);
+
+ /*
+ * Build qsorted array of local table oids for faster lookup. This can
+ * potentially contain all tables in the database so speed of lookup
+ * is important.
+ */
+ subrel_local_oids = palloc(list_length(subrel_states) * sizeof(Oid));
+ off = 0;
+ foreach(lc, subrel_states)
+ {
+ SubscriptionRelState *relstate = (SubscriptionRelState *) lfirst(lc);
+
+ subrel_local_oids[off++] = relstate->relid;
+ }
+ qsort(subrel_local_oids, list_length(subrel_states),
+ sizeof(Oid), oid_cmp);
+
+ /*
+ * Rels that we want to remove from subscription and drop any slots
+ * and origins corresponding to them.
+ */
+ sub_remove_rels = palloc(list_length(subrel_states) * sizeof(SubRemoveRels));
+
+ /*
+ * Walk over the remote tables and try to match them to locally known
+ * tables. If the table is not known locally create a new state for
+ * it.
+ *
+ * Also builds array of local oids of remote tables for the next step.
+ */
+ off = 0;
+ pubrel_local_oids = palloc(list_length(pubrel_names) * sizeof(Oid));
+
+ foreach(lc, pubrel_names)
+ {
+ RangeVar *rv = (RangeVar *) lfirst(lc);
+ Oid relid;
+
+ relid = RangeVarGetRelid(rv, AccessShareLock, false);
+
+ /* Check for supported relkind. */
+ CheckSubscriptionRelkind(get_rel_relkind(relid),
+ rv->schemaname, rv->relname);
+
+ pubrel_local_oids[off++] = relid;
+
+ if (!bsearch(&relid, subrel_local_oids,
+ list_length(subrel_states), sizeof(Oid), oid_cmp))
+ {
+ AddSubscriptionRelState(sub->oid, relid,
+ copy_data ? SUBREL_STATE_INIT : SUBREL_STATE_READY,
+ InvalidXLogRecPtr);
+ ereport(DEBUG1,
+ (errmsg_internal("table \"%s.%s\" added to subscription \"%s\"",
+ rv->schemaname, rv->relname, sub->name)));
+ }
+ }
+
+ /*
+ * Next remove state for tables we should not care about anymore using
+ * the data we collected above
+ */
+ qsort(pubrel_local_oids, list_length(pubrel_names),
+ sizeof(Oid), oid_cmp);
+
+ remove_rel_len = 0;
+ for (off = 0; off < list_length(subrel_states); off++)
+ {
+ Oid relid = subrel_local_oids[off];
+
+ /* XXX ignore non-sequences - maybe do this in GetSubscriptionRelations? */
+ if (get_rel_relkind(relid) != RELKIND_SEQUENCE)
+ continue;
+
+ if (!bsearch(&relid, pubrel_local_oids,
+ list_length(pubrel_names), sizeof(Oid), oid_cmp))
+ {
+ char state;
+ XLogRecPtr statelsn;
+
+ /*
+ * Lock pg_subscription_rel with AccessExclusiveLock to
+ * prevent any race conditions with the apply worker
+ * re-launching workers at the same time this code is trying
+ * to remove those tables.
+ *
+ * Even if new worker for this particular rel is restarted it
+ * won't be able to make any progress as we hold exclusive
+ * lock on subscription_rel till the transaction end. It will
+ * simply exit as there is no corresponding rel entry.
+ *
+ * This locking also ensures that the state of rels won't
+ * change till we are done with this refresh operation.
+ */
+ if (!rel)
+ rel = table_open(SubscriptionRelRelationId, AccessExclusiveLock);
+
+ /* Last known rel state. */
+ state = GetSubscriptionRelState(sub->oid, relid, &statelsn);
+
+ sub_remove_rels[remove_rel_len].relid = relid;
+ sub_remove_rels[remove_rel_len++].state = state;
+
+ RemoveSubscriptionRel(sub->oid, relid);
+
+ logicalrep_worker_stop(sub->oid, relid);
+
+ /*
+ * For READY state, we would have already dropped the
+ * tablesync origin.
+ */
+ if (state != SUBREL_STATE_READY)
+ {
+ char originname[NAMEDATALEN];
+
+ /*
+ * Drop the tablesync's origin tracking if exists.
+ *
+ * It is possible that the origin is not yet created for
+ * tablesync worker, this can happen for the states before
+ * SUBREL_STATE_FINISHEDCOPY. The apply worker can also
+ * concurrently try to drop the origin and by this time
+ * the origin might be already removed. For these reasons,
+ * passing missing_ok = true.
+ */
+ ReplicationOriginNameForTablesync(sub->oid, relid, originname,
+ sizeof(originname));
+ replorigin_drop_by_name(originname, true, false);
+ }
+
+ ereport(DEBUG1,
+ (errmsg_internal("table \"%s.%s\" removed from subscription \"%s\"",
+ get_namespace_name(get_rel_namespace(relid)),
+ get_rel_name(relid),
+ sub->name)));
+ }
+ }
+
+ /*
+ * Drop the tablesync slots associated with removed tables. This has
+ * to be at the end because otherwise if there is an error while doing
+ * the database operations we won't be able to rollback dropped slots.
+ */
+ for (off = 0; off < remove_rel_len; off++)
+ {
+ if (sub_remove_rels[off].state != SUBREL_STATE_READY &&
+ sub_remove_rels[off].state != SUBREL_STATE_SYNCDONE)
+ {
+ char syncslotname[NAMEDATALEN] = {0};
+
+ /*
+ * For READY/SYNCDONE states we know the tablesync slot has
+ * already been dropped by the tablesync worker.
+ *
+ * For other states, there is no certainty, maybe the slot
+ * does not exist yet. Also, if we fail after removing some of
+ * the slots, next time, it will again try to drop already
+ * dropped slots and fail. For these reasons, we allow
+ * missing_ok = true for the drop.
+ */
+ ReplicationSlotNameForTablesync(sub->oid, sub_remove_rels[off].relid,
+ syncslotname, sizeof(syncslotname));
+ ReplicationSlotDropAtPubNode(wrconn, syncslotname, true);
+ }
+ }
+
}
PG_FINALLY();
{
@@ -1614,6 +1817,75 @@ fetch_table_list(WalReceiverConn *wrconn, List *publications)
return tablelist;
}
+/*
+ * Get the list of sequences which belong to specified publications on the
+ * publisher connection.
+ */
+static List *
+fetch_sequence_list(WalReceiverConn *wrconn, List *publications)
+{
+ WalRcvExecResult *res;
+ StringInfoData cmd;
+ TupleTableSlot *slot;
+ Oid tableRow[2] = {TEXTOID, TEXTOID};
+ ListCell *lc;
+ bool first;
+ List *tablelist = NIL;
+
+ Assert(list_length(publications) > 0);
+
+ initStringInfo(&cmd);
+ appendStringInfoString(&cmd, "SELECT DISTINCT s.schemaname, s.sequencename\n"
+ " FROM pg_catalog.pg_publication_sequences s\n"
+ " WHERE s.pubname IN (");
+ first = true;
+ foreach(lc, publications)
+ {
+ char *pubname = strVal(lfirst(lc));
+
+ if (first)
+ first = false;
+ else
+ appendStringInfoString(&cmd, ", ");
+
+ appendStringInfoString(&cmd, quote_literal_cstr(pubname));
+ }
+ appendStringInfoChar(&cmd, ')');
+
+ res = walrcv_exec(wrconn, cmd.data, 2, tableRow);
+ pfree(cmd.data);
+
+ if (res->status != WALRCV_OK_TUPLES)
+ ereport(ERROR,
+ (errmsg("could not receive list of replicated tables from the publisher: %s",
+ res->err)));
+
+ /* Process tables. */
+ slot = MakeSingleTupleTableSlot(res->tupledesc, &TTSOpsMinimalTuple);
+ while (tuplestore_gettupleslot(res->tuplestore, true, false, slot))
+ {
+ char *nspname;
+ char *relname;
+ bool isnull;
+ RangeVar *rv;
+
+ nspname = TextDatumGetCString(slot_getattr(slot, 1, &isnull));
+ Assert(!isnull);
+ relname = TextDatumGetCString(slot_getattr(slot, 2, &isnull));
+ Assert(!isnull);
+
+ rv = makeRangeVar(nspname, relname, -1);
+ tablelist = lappend(tablelist, rv);
+
+ ExecClearTuple(slot);
+ }
+ ExecDropSingleTupleTableSlot(slot);
+
+ walrcv_clear_result(res);
+
+ return tablelist;
+}
+
/*
* This is to report the connection failure while dropping replication slots.
* Here, we report the WARNING for all tablesync slots so that user can drop
diff --git a/src/backend/executor/execReplication.c b/src/backend/executor/execReplication.c
index 574d7d27fd..6f0d42d551 100644
--- a/src/backend/executor/execReplication.c
+++ b/src/backend/executor/execReplication.c
@@ -608,7 +608,7 @@ void
CheckSubscriptionRelkind(char relkind, const char *nspname,
const char *relname)
{
- if (relkind != RELKIND_RELATION && relkind != RELKIND_PARTITIONED_TABLE)
+ if (relkind != RELKIND_RELATION && relkind != RELKIND_PARTITIONED_TABLE && relkind != RELKIND_SEQUENCE)
ereport(ERROR,
(errcode(ERRCODE_WRONG_OBJECT_TYPE),
errmsg("cannot use relation \"%s.%s\" as logical replication target",
diff --git a/src/backend/nodes/copyfuncs.c b/src/backend/nodes/copyfuncs.c
index df0b747883..6a5718572f 100644
--- a/src/backend/nodes/copyfuncs.c
+++ b/src/backend/nodes/copyfuncs.c
@@ -4846,6 +4846,7 @@ _copyCreatePublicationStmt(const CreatePublicationStmt *from)
COPY_NODE_FIELD(options);
COPY_NODE_FIELD(pubobjects);
COPY_SCALAR_FIELD(for_all_tables);
+ COPY_SCALAR_FIELD(for_all_sequences);
return newnode;
}
diff --git a/src/backend/nodes/equalfuncs.c b/src/backend/nodes/equalfuncs.c
index cb7ddd463c..5dccc0d0ad 100644
--- a/src/backend/nodes/equalfuncs.c
+++ b/src/backend/nodes/equalfuncs.c
@@ -2336,6 +2336,7 @@ _equalAlterPublicationStmt(const AlterPublicationStmt *a,
COMPARE_NODE_FIELD(options);
COMPARE_NODE_FIELD(pubobjects);
COMPARE_SCALAR_FIELD(for_all_tables);
+ COMPARE_SCALAR_FIELD(for_all_sequences);
COMPARE_SCALAR_FIELD(action);
return true;
diff --git a/src/backend/parser/gram.y b/src/backend/parser/gram.y
index 3d4dd43e47..f037c17985 100644
--- a/src/backend/parser/gram.y
+++ b/src/backend/parser/gram.y
@@ -9762,6 +9762,26 @@ PublicationObjSpec:
$$->pubobjtype = PUBLICATIONOBJ_TABLE_IN_CUR_SCHEMA;
$$->location = @5;
}
+ | SEQUENCE relation_expr
+ {
+ $$ = makeNode(PublicationObjSpec);
+ $$->pubobjtype = PUBLICATIONOBJ_SEQUENCE;
+ $$->pubtable = makeNode(PublicationTable);
+ $$->pubtable->relation = $2;
+ }
+ | ALL SEQUENCE IN_P SCHEMA ColId
+ {
+ $$ = makeNode(PublicationObjSpec);
+ $$->pubobjtype = PUBLICATIONOBJ_SEQUENCE_IN_SCHEMA;
+ $$->name = $5;
+ $$->location = @5;
+ }
+ | ALL SEQUENCES IN_P SCHEMA CURRENT_SCHEMA
+ {
+ $$ = makeNode(PublicationObjSpec);
+ $$->pubobjtype = PUBLICATIONOBJ_SEQUENCE_IN_CUR_SCHEMA;
+ $$->location = @5;
+ }
| ColId
{
$$ = makeNode(PublicationObjSpec);
@@ -10097,6 +10117,12 @@ UnlistenStmt:
}
;
+/*
+ * FIXME
+ *
+ * opt_publication_for_sequences and publication_for_sequences should be
+ * copies for sequences
+ */
/*****************************************************************************
*
@@ -10105,6 +10131,12 @@ UnlistenStmt:
* BEGIN / COMMIT / ROLLBACK
* (also older versions END / ABORT)
*
+ * ALTER PUBLICATION name ADD SEQUENCE sequence [, sequence2]
+ *
+ * ALTER PUBLICATION name DROP SEQUENCE sequence [, sequence2]
+ *
+ * ALTER PUBLICATION name SET SEQUENCE sequence [, sequence2]
+ *
*****************************************************************************/
TransactionStmt:
diff --git a/src/backend/replication/logical/proto.c b/src/backend/replication/logical/proto.c
index 9f5bf4b639..79c6f62638 100644
--- a/src/backend/replication/logical/proto.c
+++ b/src/backend/replication/logical/proto.c
@@ -647,6 +647,56 @@ logicalrep_write_message(StringInfo out, TransactionId xid, XLogRecPtr lsn,
pq_sendbytes(out, message, sz);
}
+/*
+ * Write SEQUENCE to stream
+ */
+void
+logicalrep_write_sequence(StringInfo out, Relation rel, TransactionId xid,
+ XLogRecPtr lsn, bool transactional,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ uint8 flags = 0;
+ char *relname;
+
+ pq_sendbyte(out, LOGICAL_REP_MSG_SEQUENCE);
+
+ /* transaction ID (if not valid, we're not streaming) */
+ if (TransactionIdIsValid(xid))
+ pq_sendint32(out, xid);
+
+ pq_sendint8(out, flags);
+ pq_sendint64(out, lsn);
+
+ logicalrep_write_namespace(out, RelationGetNamespace(rel));
+ relname = RelationGetRelationName(rel);
+ pq_sendstring(out, relname);
+
+ pq_sendint8(out, transactional);
+ pq_sendint64(out, last_value);
+ pq_sendint64(out, log_cnt);
+ pq_sendint8(out, is_called);
+}
+
+/*
+ * Read SEQUENCE from the stream.
+ */
+void
+logicalrep_read_sequence(StringInfo in, LogicalRepSequence *seqdata)
+{
+ /* XXX skipping flags and lsn */
+ pq_getmsgint(in, 1);
+ pq_getmsgint64(in);
+
+ /* Read relation name from stream */
+ seqdata->nspname = pstrdup(logicalrep_read_namespace(in));
+ seqdata->seqname = pstrdup(pq_getmsgstring(in));
+
+ seqdata->transactional = pq_getmsgint(in, 1);
+ seqdata->last_value = pq_getmsgint64(in);
+ seqdata->log_cnt = pq_getmsgint64(in);
+ seqdata->is_called = pq_getmsgint(in, 1);
+}
+
/*
* Write relation description to the output stream.
*/
@@ -1203,6 +1253,8 @@ logicalrep_message_type(LogicalRepMsgType action)
return "STREAM ABORT";
case LOGICAL_REP_MSG_STREAM_PREPARE:
return "STREAM PREPARE";
+ case LOGICAL_REP_MSG_SEQUENCE:
+ return "SEQUENCE";
}
elog(ERROR, "invalid logical replication message type \"%c\"", action);
diff --git a/src/backend/replication/logical/tablesync.c b/src/backend/replication/logical/tablesync.c
index f07983a43c..24369a1522 100644
--- a/src/backend/replication/logical/tablesync.c
+++ b/src/backend/replication/logical/tablesync.c
@@ -100,6 +100,7 @@
#include "catalog/pg_subscription_rel.h"
#include "catalog/pg_type.h"
#include "commands/copy.h"
+#include "commands/sequence.h"
#include "miscadmin.h"
#include "parser/parse_relation.h"
#include "pgstat.h"
@@ -357,6 +358,12 @@ process_syncing_tables_for_sync(XLogRecPtr current_lsn)
*
* If the synchronization position is reached (SYNCDONE), then the table can
* be marked as READY and is no longer tracked.
+ *
+ * XXX This needs to handle sequences too - after AlterSubscription_refresh
+ * starts caring about sequences, GetSubscriptionNotReadyRelations won't
+ * return just tables, and we'll have to sync them here. Not sure it's worth
+ * creating a new "sync" worker per sequence, maybe we should just sync them
+ * in the current process (it's pretty light-weight).
*/
static void
process_syncing_tables_for_apply(XLogRecPtr current_lsn)
@@ -871,6 +878,99 @@ copy_table(Relation rel)
logicalrep_rel_close(relmapentry, NoLock);
}
+
+
+/*
+ * FIXME add comment
+ */
+static void
+fetch_sequence_data(char *nspname, char *relname,
+ int64 *last_value, int64 *log_cnt, bool *is_called)
+{
+ WalRcvExecResult *res;
+ StringInfoData cmd;
+ TupleTableSlot *slot;
+ Oid tableRow[3] = {INT8OID, INT8OID, BOOLOID};
+
+ initStringInfo(&cmd);
+ appendStringInfo(&cmd, "SELECT last_value, log_cnt, is_called\n"
+ " FROM %s", quote_qualified_identifier(nspname, relname));
+
+ res = walrcv_exec(LogRepWorkerWalRcvConn, cmd.data, 3, tableRow);
+ pfree(cmd.data);
+
+ if (res->status != WALRCV_OK_TUPLES)
+ ereport(ERROR,
+ (errmsg("could not receive list of replicated tables from the publisher: %s",
+ res->err)));
+
+ /* Process the sequence. */
+ slot = MakeSingleTupleTableSlot(res->tupledesc, &TTSOpsMinimalTuple);
+ while (tuplestore_gettupleslot(res->tuplestore, true, false, slot))
+ {
+ bool isnull;
+
+ *last_value = DatumGetInt64(slot_getattr(slot, 1, &isnull));
+ Assert(!isnull);
+
+ *log_cnt = DatumGetInt64(slot_getattr(slot, 2, &isnull));
+ Assert(!isnull);
+
+ *is_called = DatumGetBool(slot_getattr(slot, 3, &isnull));
+ Assert(!isnull);
+
+ ExecClearTuple(slot);
+ }
+ ExecDropSingleTupleTableSlot(slot);
+
+ walrcv_clear_result(res);
+}
+
+/*
+ * Copy existing data of a sequence from publisher.
+ *
+ * Caller is responsible for locking the local relation.
+ */
+static void
+copy_sequence(Relation rel)
+{
+ LogicalRepRelMapEntry *relmapentry;
+ LogicalRepRelation lrel;
+ StringInfoData cmd;
+ int64 last_value = 0,
+ log_cnt = 0;
+ bool is_called = 0;
+
+ /* Get the publisher relation info. */
+ fetch_remote_table_info(get_namespace_name(RelationGetNamespace(rel)),
+ RelationGetRelationName(rel), &lrel);
+
+ /* Put the relation into relmap. */
+ logicalrep_relmap_update(&lrel);
+
+ /* Map the publisher relation to local one. */
+ relmapentry = logicalrep_rel_open(lrel.remoteid, NoLock);
+ Assert(rel == relmapentry->localrel);
+
+ /* Start copy on the publisher. */
+ initStringInfo(&cmd);
+
+ Assert(lrel.relkind == RELKIND_SEQUENCE);
+
+ fetch_sequence_data(lrel.nspname, lrel.relname, &last_value, &log_cnt, &is_called);
+
+ elog(WARNING, "sequence %s info last_value %ld log_cnt %ld is_called %d",
+ quote_qualified_identifier(lrel.nspname, lrel.relname),
+ last_value, log_cnt, is_called);
+
+ ResetSequence2(RelationGetRelid(rel), last_value, log_cnt, is_called);
+
+ logicalrep_rel_close(relmapentry, NoLock);
+}
+
+
+
+
/*
* Determine the tablesync slot name.
*
@@ -1106,10 +1206,20 @@ LogicalRepSyncTableStart(XLogRecPtr *origin_startpos)
originname)));
}
- /* Now do the initial data copy */
- PushActiveSnapshot(GetTransactionSnapshot());
- copy_table(rel);
- PopActiveSnapshot();
+ if (get_rel_relkind(RelationGetRelid(rel)) == RELKIND_SEQUENCE)
+ {
+ /* Now do the initial sequence copy */
+ PushActiveSnapshot(GetTransactionSnapshot());
+ copy_sequence(rel);
+ PopActiveSnapshot();
+ }
+ else
+ {
+ /* Now do the initial data copy */
+ PushActiveSnapshot(GetTransactionSnapshot());
+ copy_table(rel);
+ PopActiveSnapshot();
+ }
res = walrcv_exec(LogRepWorkerWalRcvConn, "COMMIT", 0, NULL);
if (res->status != WALRCV_OK_COMMAND)
diff --git a/src/backend/replication/logical/worker.c b/src/backend/replication/logical/worker.c
index 2e79302a48..089d989dd7 100644
--- a/src/backend/replication/logical/worker.c
+++ b/src/backend/replication/logical/worker.c
@@ -144,6 +144,7 @@
#include "catalog/pg_tablespace.h"
#include "commands/tablecmds.h"
#include "commands/tablespace.h"
+#include "commands/sequence.h"
#include "commands/trigger.h"
#include "executor/executor.h"
#include "executor/execPartition.h"
@@ -1091,6 +1092,61 @@ apply_handle_origin(StringInfo s)
errmsg_internal("ORIGIN message sent out of order")));
}
+/*
+ * Handle SEQUENCE message.
+ */
+static void
+apply_handle_sequence(StringInfo s)
+{
+ LogicalRepSequence seq;
+ Oid relid;
+
+ if (handle_streamed_transaction(LOGICAL_REP_MSG_SEQUENCE, s))
+ return;
+
+ logicalrep_read_sequence(s, &seq);
+
+ /*
+ * Non-transactional sequence updates should not be part of a remote
+ * transaction. There should not be any running transaction.
+ */
+ Assert((!seq.transactional) || in_remote_transaction);
+ Assert(!(!seq.transactional && in_remote_transaction));
+ Assert(!(!seq.transactional && IsTransactionState()));
+
+ /*
+ * Make sure we're in a transaction (needed by ResetSequence2). For
+ * non-transactional updates we're guaranteed to start a new one,
+ * and we'll commit it at the end.
+ */
+ if (!IsTransactionState())
+ {
+ StartTransactionCommand();
+ maybe_reread_subscription();
+ }
+
+ relid = RangeVarGetRelid(makeRangeVar(seq.nspname,
+ seq.seqname, -1),
+ RowExclusiveLock, false);
+
+ /* lock the sequence in AccessExclusiveLock, as expected by ResetSequence2 */
+ elog(WARNING, "locking sequence %d in exclusive mode", relid);
+ LockRelationOid(relid, AccessExclusiveLock);
+
+ elog(WARNING, "applying sequence %s.%s transactional %d last_value %ld log_cnt %ld is_called %d",
+ seq.nspname, seq.seqname, seq.transactional, seq.last_value, seq.log_cnt, seq.is_called);
+
+ /* apply the sequence change */
+ ResetSequence2(relid, seq.last_value, seq.log_cnt, seq.is_called);
+
+ /*
+ * Commit the per-stream transaction (we only do this when not in
+ * remote transaction, i.e. for non-transactional sequence updates.
+ */
+ if (!in_remote_transaction)
+ CommitTransactionCommand();
+}
+
/*
* Handle STREAM START message.
*/
@@ -2374,6 +2430,10 @@ apply_dispatch(StringInfo s)
*/
break;
+ case LOGICAL_REP_MSG_SEQUENCE:
+ apply_handle_sequence(s);
+ return;
+
case LOGICAL_REP_MSG_STREAM_START:
apply_handle_stream_start(s);
break;
diff --git a/src/backend/replication/pgoutput/pgoutput.c b/src/backend/replication/pgoutput/pgoutput.c
index 6f6a203dea..9ab8b1d4ec 100644
--- a/src/backend/replication/pgoutput/pgoutput.c
+++ b/src/backend/replication/pgoutput/pgoutput.c
@@ -49,6 +49,10 @@ static void pgoutput_message(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn, XLogRecPtr message_lsn,
bool transactional, const char *prefix,
Size sz, const char *message);
+static void pgoutput_sequence(LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn, XLogRecPtr sequence_lsn,
+ Relation rel, bool transactional,
+ int64 last_value, int64 log_cnt, bool is_called);
static bool pgoutput_origin_filter(LogicalDecodingContext *ctx,
RepOriginId origin_id);
static void pgoutput_begin_prepare_txn(LogicalDecodingContext *ctx,
@@ -159,6 +163,7 @@ _PG_output_plugin_init(OutputPluginCallbacks *cb)
cb->change_cb = pgoutput_change;
cb->truncate_cb = pgoutput_truncate;
cb->message_cb = pgoutput_message;
+ cb->sequence_cb = pgoutput_sequence;
cb->commit_cb = pgoutput_commit_txn;
cb->begin_prepare_cb = pgoutput_begin_prepare_txn;
@@ -175,6 +180,7 @@ _PG_output_plugin_init(OutputPluginCallbacks *cb)
cb->stream_commit_cb = pgoutput_stream_commit;
cb->stream_change_cb = pgoutput_change;
cb->stream_message_cb = pgoutput_message;
+ cb->stream_sequence_cb = pgoutput_sequence;
cb->stream_truncate_cb = pgoutput_truncate;
/* transaction streaming - two-phase commit */
cb->stream_prepare_cb = pgoutput_stream_prepare_txn;
@@ -188,6 +194,7 @@ parse_output_parameters(List *options, PGOutputData *data)
bool publication_names_given = false;
bool binary_option_given = false;
bool messages_option_given = false;
+ bool sequences_option_given = false;
bool streaming_given = false;
bool two_phase_option_given = false;
@@ -195,6 +202,7 @@ parse_output_parameters(List *options, PGOutputData *data)
data->streaming = false;
data->messages = false;
data->two_phase = false;
+ data->sequences = true;
foreach(lc, options)
{
@@ -260,6 +268,16 @@ parse_output_parameters(List *options, PGOutputData *data)
data->messages = defGetBoolean(defel);
}
+ else if (strcmp(defel->defname, "sequences") == 0)
+ {
+ if (sequences_option_given)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("conflicting or redundant options")));
+ sequences_option_given = true;
+
+ data->sequences = defGetBoolean(defel);
+ }
else if (strcmp(defel->defname, "streaming") == 0)
{
if (streaming_given)
@@ -856,6 +874,51 @@ pgoutput_message(LogicalDecodingContext *ctx, ReorderBufferTXN *txn,
OutputPluginWrite(ctx, true);
}
+static void
+pgoutput_sequence(LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn, XLogRecPtr sequence_lsn,
+ Relation rel, bool transactional,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ PGOutputData *data = (PGOutputData *) ctx->output_plugin_private;
+ TransactionId xid = InvalidTransactionId;
+ RelationSyncEntry *relentry;
+
+ if (!data->sequences)
+ return;
+
+ if (!is_publishable_relation(rel))
+ return;
+
+ /*
+ * Remember the xid for the message in streaming mode. See
+ * pgoutput_change.
+ */
+ if (in_streaming)
+ xid = txn->xid;
+
+ relentry = get_rel_sync_entry(data, RelationGetRelid(rel));
+
+ /*
+ * First check the sequence filter.
+ *
+ * We handle just REORDER_BUFFER_CHANGE_SEQUENCE here.
+ */
+ if (!relentry->pubactions.pubsequence)
+ return;
+
+ OutputPluginPrepareWrite(ctx, true);
+ logicalrep_write_sequence(ctx->out,
+ rel,
+ xid,
+ sequence_lsn,
+ transactional,
+ last_value,
+ log_cnt,
+ is_called);
+ OutputPluginWrite(ctx, true);
+}
+
/*
* Currently we always forward.
*/
@@ -1140,7 +1203,8 @@ get_rel_sync_entry(PGOutputData *data, Oid relid)
entry->streamed_txns = NIL;
entry->replicate_valid = false;
entry->pubactions.pubinsert = entry->pubactions.pubupdate =
- entry->pubactions.pubdelete = entry->pubactions.pubtruncate = false;
+ entry->pubactions.pubdelete = entry->pubactions.pubtruncate =
+ entry->pubactions.pubsequence = false;
entry->publish_as_relid = InvalidOid;
entry->map = NULL; /* will be set by maybe_send_schema() if
* needed */
@@ -1160,6 +1224,7 @@ get_rel_sync_entry(PGOutputData *data, Oid relid)
List *schemaPubids = GetSchemaPublications(schemaId);
ListCell *lc;
Oid publish_as_relid = relid;
+ bool is_sequence = (get_rel_relkind(relid) == RELKIND_SEQUENCE);
/* Reload publications if needed before use. */
if (!publications_valid)
@@ -1183,12 +1248,23 @@ get_rel_sync_entry(PGOutputData *data, Oid relid)
Publication *pub = lfirst(lc);
bool publish = false;
- if (pub->alltables)
+ if (pub->alltables && (!is_sequence))
{
publish = true;
if (pub->pubviaroot && am_partition)
publish_as_relid = llast_oid(get_partition_ancestors(relid));
}
+ else if (pub->allsequences && is_sequence)
+ {
+ publish = true;
+ }
+
+ /* if a sequence, just cross-check the list of publications */
+ if (!publish && is_sequence)
+ {
+ if (list_member_oid(pubids, pub->oid))
+ publish = true;
+ }
if (!publish)
{
@@ -1243,10 +1319,12 @@ get_rel_sync_entry(PGOutputData *data, Oid relid)
entry->pubactions.pubupdate |= pub->pubactions.pubupdate;
entry->pubactions.pubdelete |= pub->pubactions.pubdelete;
entry->pubactions.pubtruncate |= pub->pubactions.pubtruncate;
+ entry->pubactions.pubsequence |= pub->pubactions.pubsequence;
}
if (entry->pubactions.pubinsert && entry->pubactions.pubupdate &&
- entry->pubactions.pubdelete && entry->pubactions.pubtruncate)
+ entry->pubactions.pubdelete && entry->pubactions.pubtruncate &&
+ entry->pubactions.pubsequence)
break;
}
@@ -1391,6 +1469,7 @@ rel_sync_cache_publication_cb(Datum arg, int cacheid, uint32 hashvalue)
entry->pubactions.pubupdate = false;
entry->pubactions.pubdelete = false;
entry->pubactions.pubtruncate = false;
+ entry->pubactions.pubsequence = false;
}
}
diff --git a/src/backend/utils/cache/relcache.c b/src/backend/utils/cache/relcache.c
index 105d8d4601..1f8634b50a 100644
--- a/src/backend/utils/cache/relcache.c
+++ b/src/backend/utils/cache/relcache.c
@@ -5585,6 +5585,7 @@ GetRelationPublicationActions(Relation relation)
pubactions->pubupdate |= pubform->pubupdate;
pubactions->pubdelete |= pubform->pubdelete;
pubactions->pubtruncate |= pubform->pubtruncate;
+ pubactions->pubsequence |= pubform->pubsequence;
ReleaseSysCache(tup);
@@ -5593,7 +5594,8 @@ GetRelationPublicationActions(Relation relation)
* other publications.
*/
if (pubactions->pubinsert && pubactions->pubupdate &&
- pubactions->pubdelete && pubactions->pubtruncate)
+ pubactions->pubdelete && pubactions->pubtruncate &&
+ pubactions->pubsequence)
break;
}
diff --git a/src/bin/psql/tab-complete.c b/src/bin/psql/tab-complete.c
index cf30239f6d..c12e14e93e 100644
--- a/src/bin/psql/tab-complete.c
+++ b/src/bin/psql/tab-complete.c
@@ -1647,13 +1647,13 @@ psql_completion(const char *text, int start, int end)
COMPLETE_WITH("ADD", "DROP", "OWNER TO", "RENAME TO", "SET");
/* ALTER PUBLICATION <name> ADD */
else if (Matches("ALTER", "PUBLICATION", MatchAny, "ADD"))
- COMPLETE_WITH("ALL TABLES IN SCHEMA", "TABLE");
+ COMPLETE_WITH("ALL TABLES IN SCHEMA", "TABLE|SEQUENCE");
/* ALTER PUBLICATION <name> DROP */
else if (Matches("ALTER", "PUBLICATION", MatchAny, "DROP"))
- COMPLETE_WITH("ALL TABLES IN SCHEMA", "TABLE");
+ COMPLETE_WITH("ALL TABLES IN SCHEMA", "TABLE|SEQUENCE");
/* ALTER PUBLICATION <name> SET */
else if (Matches("ALTER", "PUBLICATION", MatchAny, "SET"))
- COMPLETE_WITH("(", "ALL TABLES IN SCHEMA", "TABLE");
+ COMPLETE_WITH("(", "ALL TABLES IN SCHEMA", "TABLE|SEQUENCE");
else if (Matches("ALTER", "PUBLICATION", MatchAny, "ADD|DROP|SET", "ALL", "TABLES", "IN", "SCHEMA"))
COMPLETE_WITH_QUERY(Query_for_list_of_schemas
" AND nspname != 'pg_catalog' "
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 4d992dc224..349ffe1b51 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -11513,6 +11513,11 @@
provolatile => 's', prorettype => 'oid', proargtypes => 'text',
proallargtypes => '{text,oid}', proargmodes => '{i,o}',
proargnames => '{pubname,relid}', prosrc => 'pg_get_publication_tables' },
+{ oid => '8000', descr => 'get OIDs of sequences in a publication',
+ proname => 'pg_get_publication_sequences', prorows => '1000', proretset => 't',
+ provolatile => 's', prorettype => 'oid', proargtypes => 'text',
+ proallargtypes => '{text,oid}', proargmodes => '{i,o}',
+ proargnames => '{pubname,relid}', prosrc => 'pg_get_publication_sequences' },
{ oid => '6121',
descr => 'returns whether a relation can be part of a publication',
proname => 'pg_relation_is_publishable', provolatile => 's',
diff --git a/src/include/catalog/pg_publication.h b/src/include/catalog/pg_publication.h
index 902f2f2f0d..b44bb0e5e0 100644
--- a/src/include/catalog/pg_publication.h
+++ b/src/include/catalog/pg_publication.h
@@ -40,6 +40,12 @@ CATALOG(pg_publication,6104,PublicationRelationId)
*/
bool puballtables;
+ /*
+ * indicates that this is special publication which should encompass all
+ * sequences in the database (except for the unlogged and temp ones)
+ */
+ bool puballsequences;
+
/* true if inserts are published */
bool pubinsert;
@@ -52,6 +58,9 @@ CATALOG(pg_publication,6104,PublicationRelationId)
/* true if truncates are published */
bool pubtruncate;
+ /* true if sequences are published */
+ bool pubsequence;
+
/* true if partition changes are published using root schema */
bool pubviaroot;
} FormData_pg_publication;
@@ -72,6 +81,7 @@ typedef struct PublicationActions
bool pubupdate;
bool pubdelete;
bool pubtruncate;
+ bool pubsequence;
} PublicationActions;
typedef struct Publication
@@ -79,6 +89,7 @@ typedef struct Publication
Oid oid;
char *name;
bool alltables;
+ bool allsequences;
bool pubviaroot;
PublicationActions pubactions;
} Publication;
@@ -121,6 +132,9 @@ extern List *GetPubPartitionOptionRelations(List *result,
PublicationPartOpt pub_partopt,
Oid relid);
+extern List *GetAllSequencesPublicationRelations(void);
+extern List *GetPublicationSequenceRelations(Oid pubid);
+
extern bool is_publishable_relation(Relation rel);
extern bool is_schema_publication(Oid pubid);
extern ObjectAddress publication_add_relation(Oid pubid, PublicationRelInfo *targetrel,
diff --git a/src/include/commands/sequence.h b/src/include/commands/sequence.h
index 5919fb90ee..c28e8695cb 100644
--- a/src/include/commands/sequence.h
+++ b/src/include/commands/sequence.h
@@ -60,6 +60,7 @@ extern ObjectAddress DefineSequence(ParseState *pstate, CreateSeqStmt *stmt);
extern ObjectAddress AlterSequence(ParseState *pstate, AlterSeqStmt *stmt);
extern void DeleteSequenceTuple(Oid relid);
extern void ResetSequence(Oid seq_relid);
+extern void ResetSequence2(Oid seq_relid, int64 last_value, int64 log_cnt, bool is_called);
extern void ResetSequenceCaches(void);
extern void seq_redo(XLogReaderState *rptr);
diff --git a/src/include/nodes/parsenodes.h b/src/include/nodes/parsenodes.h
index 4c5a8a39bf..d32ba498a7 100644
--- a/src/include/nodes/parsenodes.h
+++ b/src/include/nodes/parsenodes.h
@@ -3653,6 +3653,10 @@ typedef enum PublicationObjSpecType
PUBLICATIONOBJ_TABLE_IN_SCHEMA, /* Tables in schema type */
PUBLICATIONOBJ_TABLE_IN_CUR_SCHEMA, /* Get the first element from
* search_path */
+ PUBLICATIONOBJ_SEQUENCE, /* Sequence type */
+ PUBLICATIONOBJ_SEQUENCE_IN_SCHEMA, /* Sequences in schema type */
+ PUBLICATIONOBJ_SEQUENCE_IN_CUR_SCHEMA, /* Get the first element from
+ * search_path */
PUBLICATIONOBJ_CONTINUATION /* Continuation of previous type */
} PublicationObjSpecType;
@@ -3672,6 +3676,7 @@ typedef struct CreatePublicationStmt
List *options; /* List of DefElem nodes */
List *pubobjects; /* Optional list of publication objects */
bool for_all_tables; /* Special publication for all tables in db */
+ bool for_all_sequences; /* Special publication for all sequences in db */
} CreatePublicationStmt;
typedef struct AlterPublicationStmt
@@ -3688,6 +3693,7 @@ typedef struct AlterPublicationStmt
*/
List *pubobjects; /* Optional list of publication objects */
bool for_all_tables; /* Special publication for all tables in db */
+ bool for_all_sequences; /* Special publication for all sequences in db */
DefElemAction action; /* What action to perform with the
* tables/schemas */
} AlterPublicationStmt;
diff --git a/src/include/replication/logicalproto.h b/src/include/replication/logicalproto.h
index 83741dcf42..fc1782c125 100644
--- a/src/include/replication/logicalproto.h
+++ b/src/include/replication/logicalproto.h
@@ -60,6 +60,7 @@ typedef enum LogicalRepMsgType
LOGICAL_REP_MSG_RELATION = 'R',
LOGICAL_REP_MSG_TYPE = 'Y',
LOGICAL_REP_MSG_MESSAGE = 'M',
+ LOGICAL_REP_MSG_SEQUENCE = 'X', /* FIXME change */
LOGICAL_REP_MSG_BEGIN_PREPARE = 'b',
LOGICAL_REP_MSG_PREPARE = 'P',
LOGICAL_REP_MSG_COMMIT_PREPARED = 'K',
@@ -117,6 +118,18 @@ typedef struct LogicalRepTyp
char *typname; /* name of the remote type */
} LogicalRepTyp;
+/* Sequence info */
+typedef struct LogicalRepSequence
+{
+ Oid remoteid; /* unique id of the remote sequence */
+ char *nspname; /* schema name of remote sequence */
+ char *seqname; /* name of the remote sequence */
+ bool transactional;
+ int64 last_value;
+ int64 log_cnt;
+ bool is_called;
+} LogicalRepSequence;
+
/* Transaction info */
typedef struct LogicalRepBeginData
{
@@ -227,6 +240,12 @@ extern List *logicalrep_read_truncate(StringInfo in,
bool *cascade, bool *restart_seqs);
extern void logicalrep_write_message(StringInfo out, TransactionId xid, XLogRecPtr lsn,
bool transactional, const char *prefix, Size sz, const char *message);
+extern void logicalrep_write_sequence(StringInfo out, Relation rel,
+ TransactionId xid, XLogRecPtr lsn,
+ bool transactional,
+ int64 last_value, int64 log_cnt,
+ bool is_called);
+extern void logicalrep_read_sequence(StringInfo in, LogicalRepSequence *seqdata);
extern void logicalrep_write_rel(StringInfo out, TransactionId xid,
Relation rel);
extern LogicalRepRelation *logicalrep_read_rel(StringInfo in);
diff --git a/src/include/replication/pgoutput.h b/src/include/replication/pgoutput.h
index 0dc460fb70..97eb2a7ef1 100644
--- a/src/include/replication/pgoutput.h
+++ b/src/include/replication/pgoutput.h
@@ -28,6 +28,7 @@ typedef struct PGOutputData
bool streaming;
bool messages;
bool two_phase;
+ bool sequences;
} PGOutputData;
#endif /* PGOUTPUT_H */
diff --git a/src/test/regress/expected/rules.out b/src/test/regress/expected/rules.out
index b58b062b10..6eb4f11003 100644
--- a/src/test/regress/expected/rules.out
+++ b/src/test/regress/expected/rules.out
@@ -1451,6 +1451,14 @@ pg_prepared_xacts| SELECT p.transaction,
FROM ((pg_prepared_xact() p(transaction, gid, prepared, ownerid, dbid)
LEFT JOIN pg_authid u ON ((p.ownerid = u.oid)))
LEFT JOIN pg_database d ON ((p.dbid = d.oid)));
+pg_publication_sequences| SELECT p.pubname,
+ n.nspname AS schemaname,
+ c.relname AS sequencename
+ FROM pg_publication p,
+ LATERAL pg_get_publication_sequences((p.pubname)::text) gpt(relid),
+ (pg_class c
+ JOIN pg_namespace n ON ((n.oid = c.relnamespace)))
+ WHERE (c.oid = gpt.relid);
pg_publication_tables| SELECT p.pubname,
n.nspname AS schemaname,
c.relname AS tablename
diff --git a/src/test/subscription/t/027_sequences.pl b/src/test/subscription/t/027_sequences.pl
new file mode 100644
index 0000000000..fd94161907
--- /dev/null
+++ b/src/test/subscription/t/027_sequences.pl
@@ -0,0 +1,196 @@
+
+# Copyright (c) 2021, PostgreSQL Global Development Group
+
+# This tests that sequences are replicated correctly by logical replication
+use strict;
+use warnings;
+use PostgreSQL::Test::Cluster;
+use PostgreSQL::Test::Utils;
+use Test::More tests => 6;
+
+# Initialize publisher node
+my $node_publisher = PostgreSQL::Test::Cluster->new('publisher');
+$node_publisher->init(allows_streaming => 'logical');
+$node_publisher->start;
+
+# Create subscriber node
+my $node_subscriber = PostgreSQL::Test::Cluster->new('subscriber');
+$node_subscriber->init(allows_streaming => 'logical');
+$node_subscriber->start;
+
+# Create some preexisting content on publisher
+my $ddl = qq(
+ CREATE SEQUENCE s;
+);
+
+# Setup structure on the publisher
+$node_publisher->safe_psql('postgres', $ddl);
+
+# Create some the same structure on subscriber, and an extra sequence that
+# we'll create on the publisher later
+$ddl = qq(
+ CREATE SEQUENCE s;
+ CREATE SEQUENCE s2;
+);
+
+$node_subscriber->safe_psql('postgres', $ddl);
+
+# Setup logical replication
+my $publisher_connstr = $node_publisher->connstr . ' dbname=postgres';
+$node_publisher->safe_psql('postgres',
+ "CREATE PUBLICATION seq_pub");
+
+$node_publisher->safe_psql('postgres',
+ "ALTER PUBLICATION seq_pub ADD SEQUENCE s");
+
+$node_subscriber->safe_psql('postgres',
+ "CREATE SUBSCRIPTION seq_sub CONNECTION '$publisher_connstr' PUBLICATION seq_pub WITH (slot_name = seq_sub_slot)"
+);
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Wait for initial sync to finish as well
+my $synced_query =
+ "SELECT count(1) = 0 FROM pg_subscription_rel WHERE srsubstate NOT IN ('s', 'r');";
+$node_subscriber->poll_query_until('postgres', $synced_query)
+ or die "Timed out while waiting for subscriber to synchronize data";
+
+# Insert initial test data
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ -- generate a number of values using the sequence
+ SELECT nextval('s') FROM generate_series(1,100);
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+my $result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s;
+));
+
+is( $result, '132|0|t',
+ 'check replicated sequence values on subscriber');
+
+
+# advance the sequence in a rolled-back transaction - should not be replicated
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ SELECT nextval('s') FROM generate_series(1,100);
+ ROLLBACK;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s;
+));
+
+is( $result, '231|0|t',
+ 'check replicated sequence values on subscriber');
+
+
+# create a new sequence and roll it back - should not be replicated, due to
+# the transactional behavior
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ CREATE SEQUENCE s2;
+ ALTER PUBLICATION seq_pub ADD SEQUENCE s2;
+ SELECT nextval('s2') FROM generate_series(1,100);
+ ROLLBACK;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s2;
+));
+
+is( $result, '1|0|f',
+ 'check replicated sequence values on subscriber');
+
+
+# create a new sequence, advance it in a rolled-back transaction, but commit
+# the create - the advance should be replicated nevertheless
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ CREATE SEQUENCE s2;
+ ALTER PUBLICATION seq_pub ADD SEQUENCE s2;
+ SAVEPOINT sp1;
+ SELECT nextval('s2') FROM generate_series(1,100);
+ ROLLBACK TO sp1;
+ COMMIT;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Wait for sync of the second sequence we just added to finish
+$synced_query =
+ "SELECT count(1) = 0 FROM pg_subscription_rel WHERE srsubstate NOT IN ('s', 'r');";
+$node_subscriber->poll_query_until('postgres', $synced_query)
+ or die "Timed out while waiting for subscriber to synchronize data";
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s2;
+));
+
+is( $result, '132|0|t',
+ 'check replicated sequence values on subscriber');
+
+
+# advance the new sequence in a transaction, and roll it back - in this case
+# it should not be replicated at commit
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ SELECT nextval('s2') FROM generate_series(1,100);
+ ROLLBACK;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s2;
+));
+
+is( $result, '132|0|t',
+ 'check replicated sequence values on subscriber');
+
+
+# advance the sequence in a subtransaction - the subtransaction gets rolled
+# back, but commit the main one - the changes should still be replicated
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ SAVEPOINT s1;
+ SELECT nextval('s2') FROM generate_series(1,100);
+ ROLLBACK TO s1;
+ COMMIT;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s2;
+));
+
+is( $result, '330|0|t',
+ 'check replicated sequence values on subscriber');
+
+
+$node_subscriber->stop('fast');
+$node_publisher->stop('fast');
--
2.31.1
[text/x-patch] 0006-update-TAP-tests-20211222.patch (1.5K, ../../[email protected]/7-0006-update-TAP-tests-20211222.patch)
download | inline diff:
From 82c8457acfc96f93dab92881d75a349921049c8a Mon Sep 17 00:00:00 2001
From: Tomas Vondra <[email protected]>
Date: Wed, 22 Dec 2021 03:19:04 +0100
Subject: [PATCH 6/6] update TAP tests
---
src/test/subscription/t/027_sequences.pl | 10 +++++-----
1 file changed, 5 insertions(+), 5 deletions(-)
diff --git a/src/test/subscription/t/027_sequences.pl b/src/test/subscription/t/027_sequences.pl
index fd94161907..b9c1757a64 100644
--- a/src/test/subscription/t/027_sequences.pl
+++ b/src/test/subscription/t/027_sequences.pl
@@ -70,7 +70,7 @@ my $result = $node_subscriber->safe_psql(
SELECT * FROM s;
));
-is( $result, '132|0|t',
+is( $result, '100|0|t',
'check replicated sequence values on subscriber');
@@ -90,7 +90,7 @@ $result = $node_subscriber->safe_psql(
SELECT * FROM s;
));
-is( $result, '231|0|t',
+is( $result, '200|0|t',
'check replicated sequence values on subscriber');
@@ -144,7 +144,7 @@ $result = $node_subscriber->safe_psql(
SELECT * FROM s2;
));
-is( $result, '132|0|t',
+is( $result, '100|0|t',
'check replicated sequence values on subscriber');
@@ -165,7 +165,7 @@ $result = $node_subscriber->safe_psql(
SELECT * FROM s2;
));
-is( $result, '132|0|t',
+is( $result, '200|0|t',
'check replicated sequence values on subscriber');
@@ -188,7 +188,7 @@ $result = $node_subscriber->safe_psql(
SELECT * FROM s2;
));
-is( $result, '330|0|t',
+is( $result, '300|0|t',
'check replicated sequence values on subscriber');
--
2.31.1
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2022-03-23 11:50 Amit Kapila <[email protected]>
parent: Tomas Vondra <[email protected]>
2 siblings, 1 reply; 127+ messages in thread
From: Amit Kapila @ 2022-03-23 11:50 UTC (permalink / raw)
To: Petr Jelinek <[email protected]>; +Cc: Tomas Vondra <[email protected]>; Peter Eisentraut <[email protected]>; PostgreSQL Hackers <[email protected]>
On Tue, Mar 22, 2022 at 5:41 PM Petr Jelinek
<[email protected]> wrote:
>
> > On 22. 3. 2022, at 13:09, Amit Kapila <[email protected]> wrote:
> >
> > On Mon, Mar 21, 2022 at 4:25 AM Tomas Vondra
> > <[email protected]> wrote:
> >>
> >> Hi,
> >>
> >> Attached is a rebased patch, addressing most of the remaining issues.
> >>
> >
> > It appears that on the apply side, the patch always creates a new
> > relfilenode irrespective of whether the sequence message is
> > transactional or not. Is it required to create a new relfilenode for
> > non-transactional messages? If not that could be costly?
> >
>
>
> That's a good catch, I think we should just write the page in the non-transactional case, no need to mess with relnodes.
>
What if the current node has also incremented from the existing
sequence? Basically, how will we deal with conflicts? It seems we will
overwrite the actions done on the existing node which means sequence
values can go back.
On looking a bit more closely, I think I see some more fundamental
problems here:
* Don't we need some syncing mechanism between apply worker and
sequence sync worker so that apply worker skips the sequence changes
till the sync worker is finished, otherwise, there is a risk of one
overriding the values of the other?
* Currently, the patch uses one sync worker per sequence. It seems to
be a waste of resources considering apart from one additional process,
we need origin/slot to sync each sequence.
* Don't we need explicit privilege checking before applying sequence
data as we do in commit a2ab9c06ea15fbcb2bfde570986a06b37f52bcca for
tables?
--
With Regards,
Amit Kapila.
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2022-03-23 12:46 Petr Jelinek <[email protected]>
parent: Amit Kapila <[email protected]>
0 siblings, 1 reply; 127+ messages in thread
From: Petr Jelinek @ 2022-03-23 12:46 UTC (permalink / raw)
To: Amit Kapila <[email protected]>; +Cc: Tomas Vondra <[email protected]>; Peter Eisentraut <[email protected]>; PostgreSQL Hackers <[email protected]>
> On 23. 3. 2022, at 12:50, Amit Kapila <[email protected]> wrote:
>
> On Tue, Mar 22, 2022 at 5:41 PM Petr Jelinek
> <[email protected] <mailto:[email protected]>> wrote:
>>
>>> On 22. 3. 2022, at 13:09, Amit Kapila <[email protected]> wrote:
>>>
>>> On Mon, Mar 21, 2022 at 4:25 AM Tomas Vondra
>>> <[email protected]> wrote:
>>>>
>>>> Hi,
>>>>
>>>> Attached is a rebased patch, addressing most of the remaining issues.
>>>>
>>>
>>> It appears that on the apply side, the patch always creates a new
>>> relfilenode irrespective of whether the sequence message is
>>> transactional or not. Is it required to create a new relfilenode for
>>> non-transactional messages? If not that could be costly?
>>>
>>
>>
>> That's a good catch, I think we should just write the page in the non-transactional case, no need to mess with relnodes.
>>
>
> What if the current node has also incremented from the existing
> sequence? Basically, how will we deal with conflicts? It seems we will
> overwrite the actions done on the existing node which means sequence
> values can go back.
>
I think this is perfectly acceptable behavior, we are replicating state from upstream, not reconciling state on downstream.
You can't really use the builtin sequences to implement distributed sequence via replication. If user wants to write to both nodes they should not replicate the sequence value and instead offset the sequence on each node so they produce different ranges, that's quite common approach. One day we might want revisit adding support for custom sequence AMs.
> * Currently, the patch uses one sync worker per sequence. It seems to
> be a waste of resources considering apart from one additional process,
> we need origin/slot to sync each sequence.
>
This is indeed wasteful but not something that I'd consider blocker for the patch personally.
--
Petr
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2022-03-23 22:30 Tomas Vondra <[email protected]>
parent: Petr Jelinek <[email protected]>
0 siblings, 2 replies; 127+ messages in thread
From: Tomas Vondra @ 2022-03-23 22:30 UTC (permalink / raw)
To: Petr Jelinek <[email protected]>; Amit Kapila <[email protected]>; +Cc: Peter Eisentraut <[email protected]>; PostgreSQL Hackers <[email protected]>
On 3/23/22 13:46, Petr Jelinek wrote:
>
>> On 23. 3. 2022, at 12:50, Amit Kapila <[email protected]
>> <mailto:[email protected]>> wrote:
>>
>> On Tue, Mar 22, 2022 at 5:41 PM Petr Jelinek
>> <[email protected] <mailto:[email protected]>>
>> wrote:
>>>
>>>> On 22. 3. 2022, at 13:09, Amit Kapila <[email protected]
>>>> <mailto:[email protected]>> wrote:
>>>>
>>>> On Mon, Mar 21, 2022 at 4:25 AM Tomas Vondra
>>>> <[email protected]
>>>> <mailto:[email protected]>> wrote:
>>>>>
>>>>> Hi,
>>>>>
>>>>> Attached is a rebased patch, addressing most of the remaining issues.
>>>>>
>>>>
>>>> It appears that on the apply side, the patch always creates a new
>>>> relfilenode irrespective of whether the sequence message is
>>>> transactional or not. Is it required to create a new relfilenode for
>>>> non-transactional messages? If not that could be costly?
>>>>
>>>
>>>
>>> That's a good catch, I think we should just write the page in the
>>> non-transactional case, no need to mess with relnodes.
>>>
>>
>> What if the current node has also incremented from the existing
>> sequence? Basically, how will we deal with conflicts? It seems we will
>> overwrite the actions done on the existing node which means sequence
>> values can go back.
>>
>
>
> I think this is perfectly acceptable behavior, we are replicating state
> from upstream, not reconciling state on downstream.
>
> You can't really use the builtin sequences to implement distributed
> sequence via replication. If user wants to write to both nodes they
> should not replicate the sequence value and instead offset the sequence
> on each node so they produce different ranges, that's quite common
> approach. One day we might want revisit adding support for custom
> sequence AMs.
>
Exactly. Moreover it's about the same behavior as if you update table
data on the subscriber, and then an UPDATE gets replicated and
overwrites the local change.
Attached is a patch fixing the relfilenode issue - now we only allocate
a new relfilenode for the transactional case, and an in-place update
similar to a setval() otherwise. And thanks for noticing this.
>
>> * Currently, the patch uses one sync worker per sequence. It seems to
>> be a waste of resources considering apart from one additional process,
>> we need origin/slot to sync each sequence.
>>
>
>
> This is indeed wasteful but not something that I'd consider blocker for
> the patch personally.
>
Right, and the same argument can be made for tablesync of tiny tables
(which a sequence essentially is). I'm sure there are ways to improve
this, but that can be done later.
regards
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
Attachments:
[text/x-patch] 0001-Add-support-for-decoding-sequences-to-built-20220323.patch (243.5K, ../../[email protected]/2-0001-Add-support-for-decoding-sequences-to-built-20220323.patch)
download | inline diff:
From a53fc6e4a0b132587a56f43750bf9f6cfc01ed6c Mon Sep 17 00:00:00 2001
From: Tomas Vondra <[email protected]>
Date: Sat, 19 Mar 2022 20:15:06 +0100
Subject: [PATCH] Add support for decoding sequences to built-in replication
---
doc/src/sgml/catalogs.sgml | 81 ++
doc/src/sgml/protocol.sgml | 119 ++
doc/src/sgml/ref/alter_publication.sgml | 24 +-
doc/src/sgml/ref/alter_subscription.sgml | 3 +-
doc/src/sgml/ref/create_publication.sgml | 45 +-
src/backend/catalog/objectaddress.c | 44 +-
src/backend/catalog/pg_publication.c | 321 +++++-
src/backend/catalog/system_views.sql | 10 +
src/backend/commands/publicationcmds.c | 426 +++++--
src/backend/commands/sequence.c | 154 +++
src/backend/commands/subscriptioncmds.c | 101 +-
src/backend/commands/tablecmds.c | 27 +-
src/backend/executor/execReplication.c | 4 +-
src/backend/nodes/copyfuncs.c | 4 +-
src/backend/nodes/equalfuncs.c | 4 +-
src/backend/parser/gram.y | 52 +-
src/backend/replication/logical/proto.c | 52 +
src/backend/replication/logical/tablesync.c | 111 +-
src/backend/replication/logical/worker.c | 56 +
src/backend/replication/pgoutput/pgoutput.c | 79 +-
src/backend/utils/cache/relcache.c | 28 +-
src/backend/utils/cache/syscache.c | 6 +-
src/bin/pg_dump/pg_dump.c | 65 +-
src/bin/pg_dump/pg_dump.h | 3 +
src/bin/pg_dump/t/002_pg_dump.pl | 40 +-
src/bin/psql/describe.c | 287 +++--
src/bin/psql/tab-complete.c | 12 +-
src/include/catalog/pg_proc.dat | 5 +
src/include/catalog/pg_publication.h | 26 +-
.../catalog/pg_publication_namespace.h | 10 +-
src/include/commands/sequence.h | 1 +
src/include/nodes/parsenodes.h | 8 +-
src/include/replication/logicalproto.h | 19 +
src/include/replication/pgoutput.h | 1 +
src/test/regress/expected/object_address.out | 10 +-
src/test/regress/expected/publication.out | 1009 ++++++++++++++---
src/test/regress/expected/rules.out | 8 +
src/test/regress/sql/object_address.sql | 5 +-
src/test/regress/sql/publication.sql | 226 +++-
src/test/subscription/t/030_sequences.pl | 202 ++++
40 files changed, 3231 insertions(+), 457 deletions(-)
create mode 100644 src/test/subscription/t/030_sequences.pl
diff --git a/doc/src/sgml/catalogs.sgml b/doc/src/sgml/catalogs.sgml
index 2a8cd026649..b8c954a5547 100644
--- a/doc/src/sgml/catalogs.sgml
+++ b/doc/src/sgml/catalogs.sgml
@@ -6281,6 +6281,16 @@ SCRAM-SHA-256$<replaceable><iteration count></replaceable>:<replaceable>&l
Reference to schema
</para></entry>
</row>
+
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ <structfield>pntype</structfield> <type>char</type>
+ Determines which object type is included from this schema.
+ </para>
+ <para>
+ Reference to schema
+ </para></entry>
+ </row>
</tbody>
</tgroup>
</table>
@@ -9598,6 +9608,11 @@ SCRAM-SHA-256$<replaceable><iteration count></replaceable>:<replaceable>&l
<entry>prepared transactions</entry>
</row>
+ <row>
+ <entry><link linkend="view-pg-publication-sequences"><structname>pg_publication_sequences</structname></link></entry>
+ <entry>publications and their associated sequences</entry>
+ </row>
+
<row>
<entry><link linkend="view-pg-publication-tables"><structname>pg_publication_tables</structname></link></entry>
<entry>publications and their associated tables</entry>
@@ -11433,6 +11448,72 @@ SELECT * FROM pg_locks pl LEFT JOIN pg_prepared_xacts ppx
</sect1>
+ <sect1 id="view-pg-publication-sequences">
+ <title><structname>pg_publication_sequences</structname></title>
+
+ <indexterm zone="view-pg-publication-sequences">
+ <primary>pg_publication_sequences</primary>
+ </indexterm>
+
+ <para>
+ The view <structname>pg_publication_sequences</structname> provides
+ information about the mapping between publications and the sequences they
+ contain. Unlike the underlying catalog
+ <link linkend="catalog-pg-publication-rel"><structname>pg_publication_rel</structname></link>,
+ this view expands
+ publications defined as <literal>FOR ALL SEQUENCES</literal>, so for such
+ publications there will be a row for each eligible sequence.
+ </para>
+
+ <table>
+ <title><structname>pg_publication_sequences</structname> Columns</title>
+ <tgroup cols="1">
+ <thead>
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ Column Type
+ </para>
+ <para>
+ Description
+ </para></entry>
+ </row>
+ </thead>
+
+ <tbody>
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ <structfield>pubname</structfield> <type>name</type>
+ (references <link linkend="catalog-pg-publication"><structname>pg_publication</structname></link>.<structfield>pubname</structfield>)
+ </para>
+ <para>
+ Name of publication
+ </para></entry>
+ </row>
+
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ <structfield>schemaname</structfield> <type>name</type>
+ (references <link linkend="catalog-pg-namespace"><structname>pg_namespace</structname></link>.<structfield>nspname</structfield>)
+ </para>
+ <para>
+ Name of schema containing sequence
+ </para></entry>
+ </row>
+
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ <structfield>sequencename</structfield> <type>name</type>
+ (references <link linkend="catalog-pg-class"><structname>pg_class</structname></link>.<structfield>relname</structfield>)
+ </para>
+ <para>
+ Name of sequence
+ </para></entry>
+ </row>
+ </tbody>
+ </tgroup>
+ </table>
+ </sect1>
+
<sect1 id="view-pg-publication-tables">
<title><structname>pg_publication_tables</structname></title>
diff --git a/doc/src/sgml/protocol.sgml b/doc/src/sgml/protocol.sgml
index 719b947ef4e..c61c310e176 100644
--- a/doc/src/sgml/protocol.sgml
+++ b/doc/src/sgml/protocol.sgml
@@ -7065,6 +7065,125 @@ Relation
</listitem>
</varlistentry>
+<varlistentry id="protocol-logicalrep-message-formats-Sequence">
+<term>
+Sequence
+</term>
+<listitem>
+<para>
+
+<variablelist>
+<varlistentry>
+<term>
+ Byte1('X')
+</term>
+<listitem>
+<para>
+ Identifies the message as a sequence message.
+</para>
+</listitem>
+</varlistentry>
+<varlistentry>
+<term>
+ Int32 (TransactionId)
+</term>
+<listitem>
+<para>
+ Xid of the transaction (only present for streamed transactions).
+ This field is available since protocol version 2.
+</para>
+</listitem>
+</varlistentry>
+<varlistentry>
+<term>
+ Int8(0)
+</term>
+<listitem>
+<para>
+ Flags; currently unused.
+</para>
+</listitem>
+</varlistentry>
+<varlistentry>
+<term>
+ Int64 (XLogRecPtr)
+</term>
+<listitem>
+<para>
+ The LSN of the sequence increment.
+</para>
+</listitem>
+</varlistentry>
+<varlistentry>
+<term>
+ String
+</term>
+<listitem>
+<para>
+ Namespace (empty string for <literal>pg_catalog</literal>).
+</para>
+</listitem>
+</varlistentry>
+<varlistentry>
+<term>
+ String
+</term>
+<listitem>
+<para>
+ Relation name.
+</para>
+</listitem>
+</varlistentry>
+
+<varlistentry>
+<term>
+ Int8
+</term>
+<listitem>
+<para>
+ 1 if the sequence update is transactions, 0 otherwise.
+</para>
+</listitem>
+</varlistentry>
+
+<varlistentry>
+<term>
+ Int64
+</term>
+<listitem>
+<para>
+ <structfield>last_value</structfield> value of the sequence.
+</para>
+</listitem>
+</varlistentry>
+
+<varlistentry>
+<term>
+ Int64
+</term>
+<listitem>
+<para>
+ <structfield>log_cnt</structfield> value of the sequence.
+</para>
+</listitem>
+</varlistentry>
+
+<varlistentry>
+<term>
+ Int8
+</term>
+<listitem>
+<para>
+ <structfield>is_called</structfield> value of the sequence.
+</para>
+</listitem>
+</varlistentry>
+
+</variablelist>
+</para>
+</listitem>
+</varlistentry>
+
<varlistentry id="protocol-logicalrep-message-formats-Type">
<term>
Type
diff --git a/doc/src/sgml/ref/alter_publication.sgml b/doc/src/sgml/ref/alter_publication.sgml
index 32b75f6c78e..5dacb732b69 100644
--- a/doc/src/sgml/ref/alter_publication.sgml
+++ b/doc/src/sgml/ref/alter_publication.sgml
@@ -31,7 +31,9 @@ ALTER PUBLICATION <replaceable class="parameter">name</replaceable> RENAME TO <r
<phrase>where <replaceable class="parameter">publication_object</replaceable> is one of:</phrase>
TABLE [ ONLY ] <replaceable class="parameter">table_name</replaceable> [ * ] [ WHERE ( <replaceable class="parameter">expression</replaceable> ) ] [, ... ]
+ SEQUENCE <replaceable class="parameter">sequence_name</replaceable> [, ... ]
ALL TABLES IN SCHEMA { <replaceable class="parameter">schema_name</replaceable> | CURRENT_SCHEMA } [, ... ]
+ ALL SEQUENCES IN SCHEMA { <replaceable class="parameter">schema_name</replaceable> | CURRENT_SCHEMA } [, ... ]
</synopsis>
</refsynopsisdiv>
@@ -58,7 +60,18 @@ ALTER PUBLICATION <replaceable class="parameter">name</replaceable> RENAME TO <r
</para>
<para>
- The fourth variant of this command listed in the synopsis can change
+ The next three variants change which sequences are part of the publication.
+ The <literal>SET SEQUENCE</literal> clause will replace the list of sequences
+ in the publication with the specified one. The <literal>ADD SEQUENCE</literal>
+ and <literal>DROP SEQUENCE</literal> clauses will add and remove one or more
+ sequences from the publication. Note that adding sequences to a publication
+ that is already subscribed to will require a <literal>ALTER SUBSCRIPTION
+ ... REFRESH PUBLICATION</literal> action on the subscribing side in order
+ to become effective.
+ </para>
+
+ <para>
+ The seventh variant of this command listed in the synopsis can change
all of the publication properties specified in
<xref linkend="sql-createpublication"/>. Properties not mentioned in the
command retain their previous settings.
@@ -122,6 +135,15 @@ ALTER PUBLICATION <replaceable class="parameter">name</replaceable> RENAME TO <r
</listitem>
</varlistentry>
+ <varlistentry>
+ <term><replaceable class="parameter">sequence_name</replaceable></term>
+ <listitem>
+ <para>
+ Name of an existing sequence.
+ </para>
+ </listitem>
+ </varlistentry>
+
<varlistentry>
<term><replaceable class="parameter">schema_name</replaceable></term>
<listitem>
diff --git a/doc/src/sgml/ref/alter_subscription.sgml b/doc/src/sgml/ref/alter_subscription.sgml
index ac2db249cbb..70c5c69be97 100644
--- a/doc/src/sgml/ref/alter_subscription.sgml
+++ b/doc/src/sgml/ref/alter_subscription.sgml
@@ -148,7 +148,7 @@ ALTER SUBSCRIPTION <replaceable class="parameter">name</replaceable> RENAME TO <
<listitem>
<para>
Fetch missing table information from publisher. This will start
- replication of tables that were added to the subscribed-to publications
+ replication of tables and sequences that were added to the subscribed-to publications
since <command>CREATE SUBSCRIPTION</command> or
the last invocation of <command>REFRESH PUBLICATION</command>.
</para>
@@ -169,6 +169,7 @@ ALTER SUBSCRIPTION <replaceable class="parameter">name</replaceable> RENAME TO <
<para>
Previously subscribed tables are not copied, even if a table's row
filter <literal>WHERE</literal> clause has since been modified.
+ Previously-subscribed sequences are not copied either.
</para>
</listitem>
</varlistentry>
diff --git a/doc/src/sgml/ref/create_publication.sgml b/doc/src/sgml/ref/create_publication.sgml
index 4979b9b646d..f355549966d 100644
--- a/doc/src/sgml/ref/create_publication.sgml
+++ b/doc/src/sgml/ref/create_publication.sgml
@@ -23,13 +23,16 @@ PostgreSQL documentation
<synopsis>
CREATE PUBLICATION <replaceable class="parameter">name</replaceable>
[ FOR ALL TABLES
+ | FOR ALL SEQUENCES
| FOR <replaceable class="parameter">publication_object</replaceable> [, ... ] ]
[ WITH ( <replaceable class="parameter">publication_parameter</replaceable> [= <replaceable class="parameter">value</replaceable>] [, ... ] ) ]
<phrase>where <replaceable class="parameter">publication_object</replaceable> is one of:</phrase>
TABLE [ ONLY ] <replaceable class="parameter">table_name</replaceable> [ * ] [ WHERE ( <replaceable class="parameter">expression</replaceable> ) ] [, ... ]
+ SEQUENCE <replaceable class="parameter">sequence_name</replaceable> [ * ] [, ... ]
ALL TABLES IN SCHEMA { <replaceable class="parameter">schema_name</replaceable> | CURRENT_SCHEMA } [, ... ]
+ ALL SEQUENCES IN SCHEMA { <replaceable class="parameter">schema_name</replaceable> | CURRENT_SCHEMA } [, ... ]
</synopsis>
</refsynopsisdiv>
@@ -107,27 +110,43 @@ CREATE PUBLICATION <replaceable class="parameter">name</replaceable>
</listitem>
</varlistentry>
+ <varlistentry>
+ <term><literal>FOR SEQUENCE</literal></term>
+ <listitem>
+ <para>
+ Specifies a list of sequences to add to the publication.
+ </para>
+
+ <para>
+ Specifying a sequence that is part of a schema specified by <literal>FOR
+ ALL SEQUENCES IN SCHEMA</literal> is not supported.
+ </para>
+ </listitem>
+ </varlistentry>
+
<varlistentry>
<term><literal>FOR ALL TABLES</literal></term>
+ <term><literal>FOR ALL SEQUENCES</literal></term>
<listitem>
<para>
- Marks the publication as one that replicates changes for all tables in
- the database, including tables created in the future.
+ Marks the publication as one that replicates changes for all tables/sequences in
+ the database, including tables/sequences created in the future.
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><literal>FOR ALL TABLES IN SCHEMA</literal></term>
+ <term><literal>FOR ALL SEQUENCES IN SCHEMA</literal></term>
<listitem>
<para>
- Marks the publication as one that replicates changes for all tables in
- the specified list of schemas, including tables created in the future.
+ Marks the publication as one that replicates changes for all sequences/tables in
+ the specified list of schemas, including sequences/tables created in the future.
</para>
<para>
- Specifying a schema along with a table which belongs to the specified
- schema using <literal>FOR TABLE</literal> is not supported.
+ Specifying a schema along with a sequence/table which belongs to the specified
+ schema using <literal>FOR SEQUENCE</literal>/<literal>FOR TABLE</literal> is not supported.
</para>
<para>
@@ -202,10 +221,9 @@ CREATE PUBLICATION <replaceable class="parameter">name</replaceable>
<title>Notes</title>
<para>
- If <literal>FOR TABLE</literal>, <literal>FOR ALL TABLES</literal> or
- <literal>FOR ALL TABLES IN SCHEMA</literal> are not specified, then the
- publication starts out with an empty set of tables. That is useful if
- tables or schemas are to be added later.
+ If <literal>FOR TABLE</literal>, <literal>FOR SEQUENCE</literal>, etc. is
+ not specified, then the publication starts out with an empty set of tables
+ and sequences. That is useful if objects are to be added later.
</para>
<para>
@@ -220,10 +238,9 @@ CREATE PUBLICATION <replaceable class="parameter">name</replaceable>
</para>
<para>
- To add a table to a publication, the invoking user must have ownership
- rights on the table. The <command>FOR ALL TABLES</command> and
- <command>FOR ALL TABLES IN SCHEMA</command> clauses require the invoking
- user to be a superuser.
+ To add a table or a sequence to a publication, the invoking user must
+ have ownership rights on the object. The <command>FOR ALL ...</command>
+ clauses require the invoking user to be a superuser.
</para>
<para>
diff --git a/src/backend/catalog/objectaddress.c b/src/backend/catalog/objectaddress.c
index d7ce063997a..3fd17ea64f0 100644
--- a/src/backend/catalog/objectaddress.c
+++ b/src/backend/catalog/objectaddress.c
@@ -1930,12 +1930,14 @@ get_object_address_publication_schema(List *object, bool missing_ok)
char *pubname;
char *schemaname;
Oid schemaid;
+ char *objtype;
ObjectAddressSet(address, PublicationNamespaceRelationId, InvalidOid);
/* Fetch schema name and publication name from input list */
schemaname = strVal(linitial(object));
pubname = strVal(lsecond(object));
+ objtype = strVal(lthird(object));
schemaid = get_namespace_oid(schemaname, missing_ok);
if (!OidIsValid(schemaid))
@@ -1948,10 +1950,12 @@ get_object_address_publication_schema(List *object, bool missing_ok)
/* Find the publication schema mapping in syscache */
address.objectId =
- GetSysCacheOid2(PUBLICATIONNAMESPACEMAP,
+ GetSysCacheOid3(PUBLICATIONNAMESPACEMAP,
Anum_pg_publication_namespace_oid,
ObjectIdGetDatum(schemaid),
- ObjectIdGetDatum(pub->oid));
+ ObjectIdGetDatum(pub->oid),
+ CharGetDatum(objtype[0]));
+
if (!OidIsValid(address.objectId) && !missing_ok)
ereport(ERROR,
(errcode(ERRCODE_UNDEFINED_OBJECT),
@@ -2232,7 +2236,6 @@ pg_get_object_address(PG_FUNCTION_ARGS)
case OBJECT_DOMCONSTRAINT:
case OBJECT_CAST:
case OBJECT_USER_MAPPING:
- case OBJECT_PUBLICATION_NAMESPACE:
case OBJECT_PUBLICATION_REL:
case OBJECT_DEFACL:
case OBJECT_TRANSFORM:
@@ -2257,6 +2260,7 @@ pg_get_object_address(PG_FUNCTION_ARGS)
/* fall through to check args length */
/* FALLTHROUGH */
case OBJECT_OPERATOR:
+ case OBJECT_PUBLICATION_NAMESPACE:
if (list_length(args) != 2)
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
@@ -2327,6 +2331,8 @@ pg_get_object_address(PG_FUNCTION_ARGS)
objnode = (Node *) list_make2(name, linitial(args));
break;
case OBJECT_PUBLICATION_NAMESPACE:
+ objnode = (Node *) list_make3(linitial(name), linitial(args), lsecond(args));
+ break;
case OBJECT_USER_MAPPING:
objnode = (Node *) list_make2(linitial(name), linitial(args));
break;
@@ -2881,11 +2887,12 @@ get_catalog_object_by_oid(Relation catalog, AttrNumber oidcol, Oid objectId)
*
* Get publication name and schema name from the object address into pubname and
* nspname. Both pubname and nspname are palloc'd strings which will be freed by
- * the caller.
+ * the caller. The last parameter specifies which object type is included from
+ * the schema.
*/
static bool
getPublicationSchemaInfo(const ObjectAddress *object, bool missing_ok,
- char **pubname, char **nspname)
+ char **pubname, char **nspname, char **objtype)
{
HeapTuple tup;
Form_pg_publication_namespace pnform;
@@ -2921,6 +2928,13 @@ getPublicationSchemaInfo(const ObjectAddress *object, bool missing_ok,
return false;
}
+ /*
+ * The type is always a single character, but we need to pass it as a string,
+ * so allocate two charaters and set the first one. The second one is \0.
+ */
+ *objtype = palloc0(2);
+ *objtype[0] = pnform->pntype;
+
ReleaseSysCache(tup);
return true;
}
@@ -3926,15 +3940,17 @@ getObjectDescription(const ObjectAddress *object, bool missing_ok)
{
char *pubname;
char *nspname;
+ char *objtype;
if (!getPublicationSchemaInfo(object, missing_ok,
- &pubname, &nspname))
+ &pubname, &nspname, &objtype))
break;
- appendStringInfo(&buffer, _("publication of schema %s in publication %s"),
- nspname, pubname);
+ appendStringInfo(&buffer, _("publication of schema %s in publication %s type %s"),
+ nspname, pubname, objtype);
pfree(pubname);
pfree(nspname);
+ pfree(objtype);
break;
}
@@ -5729,18 +5745,24 @@ getObjectIdentityParts(const ObjectAddress *object,
{
char *pubname;
char *nspname;
+ char *objtype;
if (!getPublicationSchemaInfo(object, missing_ok, &pubname,
- &nspname))
+ &nspname, &objtype))
break;
- appendStringInfo(&buffer, "%s in publication %s",
- nspname, pubname);
+ appendStringInfo(&buffer, "%s in publication %s type %s",
+ nspname, pubname, objtype);
if (objargs)
*objargs = list_make1(pubname);
else
pfree(pubname);
+ if (objargs)
+ *objargs = lappend(*objargs, objtype);
+ else
+ pfree(objtype);
+
if (objname)
*objname = list_make1(nspname);
else
diff --git a/src/backend/catalog/pg_publication.c b/src/backend/catalog/pg_publication.c
index 789b895db89..144d1ada9e2 100644
--- a/src/backend/catalog/pg_publication.c
+++ b/src/backend/catalog/pg_publication.c
@@ -54,7 +54,8 @@ check_publication_add_relation(Relation targetrel)
{
/* Must be a regular or partitioned table */
if (RelationGetForm(targetrel)->relkind != RELKIND_RELATION &&
- RelationGetForm(targetrel)->relkind != RELKIND_PARTITIONED_TABLE)
+ RelationGetForm(targetrel)->relkind != RELKIND_PARTITIONED_TABLE &&
+ RelationGetForm(targetrel)->relkind != RELKIND_SEQUENCE)
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("cannot add relation \"%s\" to publication",
@@ -131,7 +132,8 @@ static bool
is_publishable_class(Oid relid, Form_pg_class reltuple)
{
return (reltuple->relkind == RELKIND_RELATION ||
- reltuple->relkind == RELKIND_PARTITIONED_TABLE) &&
+ reltuple->relkind == RELKIND_PARTITIONED_TABLE ||
+ reltuple->relkind == RELKIND_SEQUENCE) &&
!IsCatalogRelationOid(relid) &&
reltuple->relpersistence == RELPERSISTENCE_PERMANENT &&
relid >= FirstNormalObjectId;
@@ -176,6 +178,47 @@ filter_partitions(List *relids)
return result;
}
+/* Check the character is a valid object type for schema publication. */
+static void
+AssertObjectTypeValid(char objectType)
+{
+#ifdef USE_ASSERT_CHECKING
+ Assert(objectType == PUB_OBJTYPE_SEQUENCE || objectType == PUB_OBJTYPE_TABLE);
+#endif
+}
+
+/*
+ * Determine object type given the object type set for a schema.
+ */
+char
+pub_get_object_type_for_relkind(char relkind)
+{
+ /* sequence maps directly to sequence relkind */
+ if (relkind == RELKIND_SEQUENCE)
+ return PUB_OBJTYPE_SEQUENCE;
+
+ /* for table, we match either regular or partitioned table */
+ if (relkind == RELKIND_RELATION ||
+ relkind == RELKIND_PARTITIONED_TABLE)
+ return PUB_OBJTYPE_TABLE;
+
+ return PUB_OBJTYPE_UNSUPPORTED;
+}
+
+/*
+ * Determine if publication object type matches the relkind.
+ *
+ * Returns true if the relation matches object type replicated by this schema,
+ * false otherwise.
+ */
+static bool
+pub_object_type_matches_relkind(char objectType, char relkind)
+{
+ AssertObjectTypeValid(objectType);
+
+ return (pub_get_object_type_for_relkind(relkind) == objectType);
+}
+
/*
* Another variant of this, taking a Relation.
*/
@@ -205,7 +248,7 @@ is_schema_publication(Oid pubid)
ObjectIdGetDatum(pubid));
scan = systable_beginscan(pubschsrel,
- PublicationNamespacePnnspidPnpubidIndexId,
+ PublicationNamespacePnnspidPnpubidPntypeIndexId,
true, NULL, 1, &scankey);
tup = systable_getnext(scan);
result = HeapTupleIsValid(tup);
@@ -313,7 +356,9 @@ GetTopMostAncestorInPublication(Oid puboid, List *ancestors, int *ancestor_level
}
else
{
- aschemaPubids = GetSchemaPublications(get_rel_namespace(ancestor));
+ /* we only search for ancestors of tables, so PUB_OBJTYPE_TABLE */
+ aschemaPubids = GetSchemaPublications(get_rel_namespace(ancestor),
+ PUB_OBJTYPE_TABLE);
if (list_member_oid(aschemaPubids, puboid))
{
topmost_relid = ancestor;
@@ -436,7 +481,7 @@ publication_add_relation(Oid pubid, PublicationRelInfo *pri,
* Insert new publication / schema mapping.
*/
ObjectAddress
-publication_add_schema(Oid pubid, Oid schemaid, bool if_not_exists)
+publication_add_schema(Oid pubid, Oid schemaid, char objectType, bool if_not_exists)
{
Relation rel;
HeapTuple tup;
@@ -448,6 +493,8 @@ publication_add_schema(Oid pubid, Oid schemaid, bool if_not_exists)
ObjectAddress myself,
referenced;
+ AssertObjectTypeValid(objectType);
+
rel = table_open(PublicationNamespaceRelationId, RowExclusiveLock);
/*
@@ -455,9 +502,10 @@ publication_add_schema(Oid pubid, Oid schemaid, bool if_not_exists)
* duplicates, it's here just to provide nicer error message in common
* case. The real protection is the unique key on the catalog.
*/
- if (SearchSysCacheExists2(PUBLICATIONNAMESPACEMAP,
+ if (SearchSysCacheExists3(PUBLICATIONNAMESPACEMAP,
ObjectIdGetDatum(schemaid),
- ObjectIdGetDatum(pubid)))
+ ObjectIdGetDatum(pubid),
+ CharGetDatum(objectType)))
{
table_close(rel, RowExclusiveLock);
@@ -483,6 +531,8 @@ publication_add_schema(Oid pubid, Oid schemaid, bool if_not_exists)
ObjectIdGetDatum(pubid);
values[Anum_pg_publication_namespace_pnnspid - 1] =
ObjectIdGetDatum(schemaid);
+ values[Anum_pg_publication_namespace_pntype - 1] =
+ CharGetDatum(objectType);
tup = heap_form_tuple(RelationGetDescr(rel), values, nulls);
@@ -508,7 +558,7 @@ publication_add_schema(Oid pubid, Oid schemaid, bool if_not_exists)
* publication_add_relation for why we need to consider all the
* partitions.
*/
- schemaRels = GetSchemaPublicationRelations(schemaid,
+ schemaRels = GetSchemaPublicationRelations(schemaid, objectType,
PUBLICATION_PART_ALL);
InvalidatePublicationRels(schemaRels);
@@ -542,11 +592,14 @@ GetRelationPublications(Oid relid)
/*
* Gets list of relation oids for a publication.
*
- * This should only be used FOR TABLE publications, the FOR ALL TABLES
- * should use GetAllTablesPublicationRelations().
+ * This should only be used FOR TABLE / FOR SEQUENCE publications, the FOR
+ * ALL TABLES / SEQUENCES should use GetAllTablesPublicationRelations()
+ * and GetAllSequencesPublicationRelations().
+ *
+ * XXX pub_partopt only matters for tables, not sequences.
*/
List *
-GetPublicationRelations(Oid pubid, PublicationPartOpt pub_partopt)
+GetPublicationRelations(Oid pubid, char objectType, PublicationPartOpt pub_partopt)
{
List *result;
Relation pubrelsrel;
@@ -554,6 +607,8 @@ GetPublicationRelations(Oid pubid, PublicationPartOpt pub_partopt)
SysScanDesc scan;
HeapTuple tup;
+ AssertObjectTypeValid(objectType);
+
/* Find all publications associated with the relation. */
pubrelsrel = table_open(PublicationRelRelationId, AccessShareLock);
@@ -568,11 +623,29 @@ GetPublicationRelations(Oid pubid, PublicationPartOpt pub_partopt)
result = NIL;
while (HeapTupleIsValid(tup = systable_getnext(scan)))
{
+ char relkind;
Form_pg_publication_rel pubrel;
pubrel = (Form_pg_publication_rel) GETSTRUCT(tup);
- result = GetPubPartitionOptionRelations(result, pub_partopt,
- pubrel->prrelid);
+ relkind = get_rel_relkind(pubrel->prrelid);
+
+ /*
+ * If the relkind does not match the requested object type, ignore the
+ * relation. For example we might be interested only in sequences, so
+ * we ignore tables.
+ */
+ if (!pub_object_type_matches_relkind(objectType, relkind))
+ continue;
+
+ /*
+ * We don't have partitioned sequences, so just add them to the list.
+ * Otherwise consider adding all child relations, if requested.
+ */
+ if (relkind == RELKIND_SEQUENCE)
+ result = lappend_oid(result, pubrel->prrelid);
+ else
+ result = GetPubPartitionOptionRelations(result, pub_partopt,
+ pubrel->prrelid);
}
systable_endscan(scan);
@@ -622,6 +695,43 @@ GetAllTablesPublications(void)
return result;
}
+/*
+ * Gets list of publication oids for publications marked as FOR ALL SEQUENCES.
+ */
+List *
+GetAllSequencesPublications(void)
+{
+ List *result;
+ Relation rel;
+ ScanKeyData scankey;
+ SysScanDesc scan;
+ HeapTuple tup;
+
+ /* Find all publications that are marked as for all sequences. */
+ rel = table_open(PublicationRelationId, AccessShareLock);
+
+ ScanKeyInit(&scankey,
+ Anum_pg_publication_puballsequences,
+ BTEqualStrategyNumber, F_BOOLEQ,
+ BoolGetDatum(true));
+
+ scan = systable_beginscan(rel, InvalidOid, false,
+ NULL, 1, &scankey);
+
+ result = NIL;
+ while (HeapTupleIsValid(tup = systable_getnext(scan)))
+ {
+ Oid oid = ((Form_pg_publication) GETSTRUCT(tup))->oid;
+
+ result = lappend_oid(result, oid);
+ }
+
+ systable_endscan(scan);
+ table_close(rel, AccessShareLock);
+
+ return result;
+}
+
/*
* Gets list of all relation published by FOR ALL TABLES publication(s).
*
@@ -688,28 +798,38 @@ GetAllTablesPublicationRelations(bool pubviaroot)
/*
* Gets the list of schema oids for a publication.
*
- * This should only be used FOR ALL TABLES IN SCHEMA publications.
+ * This should only be used FOR ALL TABLES IN SCHEMA and FOR ALL SEQUENCES
+ * publications.
+ *
+ * 'objectType' determines whether to get FOR TABLE or FOR SEQUENCES schemas
*/
List *
-GetPublicationSchemas(Oid pubid)
+GetPublicationSchemas(Oid pubid, char objectType)
{
List *result = NIL;
Relation pubschsrel;
- ScanKeyData scankey;
+ ScanKeyData scankey[2];
SysScanDesc scan;
HeapTuple tup;
+ AssertObjectTypeValid(objectType);
+
/* Find all schemas associated with the publication */
pubschsrel = table_open(PublicationNamespaceRelationId, AccessShareLock);
- ScanKeyInit(&scankey,
+ ScanKeyInit(&scankey[0],
Anum_pg_publication_namespace_pnpubid,
BTEqualStrategyNumber, F_OIDEQ,
ObjectIdGetDatum(pubid));
+ ScanKeyInit(&scankey[1],
+ Anum_pg_publication_namespace_pntype,
+ BTEqualStrategyNumber, F_CHAREQ,
+ CharGetDatum(objectType));
+
scan = systable_beginscan(pubschsrel,
- PublicationNamespacePnnspidPnpubidIndexId,
- true, NULL, 1, &scankey);
+ PublicationNamespacePnnspidPnpubidPntypeIndexId,
+ true, NULL, 2, scankey);
while (HeapTupleIsValid(tup = systable_getnext(scan)))
{
Form_pg_publication_namespace pubsch;
@@ -727,14 +847,26 @@ GetPublicationSchemas(Oid pubid)
/*
* Gets the list of publication oids associated with a specified schema.
+ *
+ * objectType specifies whether we're looking for schemas including tables or
+ * sequences.
+ *
+ * Note: relcache calls this for all object types, not just tables and sequences.
+ * Which is why we handle the PUB_OBJTYPE_UNSUPPORTED object type too.
*/
List *
-GetSchemaPublications(Oid schemaid)
+GetSchemaPublications(Oid schemaid, char objectType)
{
List *result = NIL;
CatCList *pubschlist;
int i;
+ /* unsupported object type */
+ if (objectType == PUB_OBJTYPE_UNSUPPORTED)
+ return result;
+
+ AssertObjectTypeValid(objectType);
+
/* Find all publications associated with the schema */
pubschlist = SearchSysCacheList1(PUBLICATIONNAMESPACEMAP,
ObjectIdGetDatum(schemaid));
@@ -742,6 +874,11 @@ GetSchemaPublications(Oid schemaid)
{
HeapTuple tup = &pubschlist->members[i]->tuple;
Oid pubid = ((Form_pg_publication_namespace) GETSTRUCT(tup))->pnpubid;
+ char pntype = ((Form_pg_publication_namespace) GETSTRUCT(tup))->pntype;
+
+ /* Skip schemas publishing a different object type. */
+ if (pntype != objectType)
+ continue;
result = lappend_oid(result, pubid);
}
@@ -753,9 +890,13 @@ GetSchemaPublications(Oid schemaid)
/*
* Get the list of publishable relation oids for a specified schema.
+ *
+ * objectType specifies whether this is FOR ALL TABLES IN SCHEMA or FOR ALL
+ * SEQUENCES IN SCHEMA
*/
List *
-GetSchemaPublicationRelations(Oid schemaid, PublicationPartOpt pub_partopt)
+GetSchemaPublicationRelations(Oid schemaid, char objectType,
+ PublicationPartOpt pub_partopt)
{
Relation classRel;
ScanKeyData key[1];
@@ -764,6 +905,7 @@ GetSchemaPublicationRelations(Oid schemaid, PublicationPartOpt pub_partopt)
List *result = NIL;
Assert(OidIsValid(schemaid));
+ AssertObjectTypeValid(objectType);
classRel = table_open(RelationRelationId, AccessShareLock);
@@ -784,9 +926,16 @@ GetSchemaPublicationRelations(Oid schemaid, PublicationPartOpt pub_partopt)
continue;
relkind = get_rel_relkind(relid);
- if (relkind == RELKIND_RELATION)
- result = lappend_oid(result, relid);
- else if (relkind == RELKIND_PARTITIONED_TABLE)
+
+ /* Skip if the relkind does not match FOR ALL TABLES / SEQUENCES. */
+ if (!pub_object_type_matches_relkind(objectType, relkind))
+ continue;
+
+ /*
+ * If the object is a partitioned table, lookup all the child relations
+ * (if requested). Otherwise just add the object to the list.
+ */
+ if (relkind == RELKIND_PARTITIONED_TABLE)
{
List *partitionrels = NIL;
@@ -799,7 +948,11 @@ GetSchemaPublicationRelations(Oid schemaid, PublicationPartOpt pub_partopt)
pub_partopt,
relForm->oid);
result = list_concat_unique_oid(result, partitionrels);
+ continue;
}
+
+ /* non-partitioned tables and sequences */
+ result = lappend_oid(result, relid);
}
table_endscan(scan);
@@ -809,27 +962,67 @@ GetSchemaPublicationRelations(Oid schemaid, PublicationPartOpt pub_partopt)
/*
* Gets the list of all relations published by FOR ALL TABLES IN SCHEMA
- * publication.
+ * or FOR ALL SEQUENCES IN SCHEMA publication.
*/
List *
-GetAllSchemaPublicationRelations(Oid pubid, PublicationPartOpt pub_partopt)
+GetAllSchemaPublicationRelations(Oid pubid, char objectType,
+ PublicationPartOpt pub_partopt)
{
List *result = NIL;
- List *pubschemalist = GetPublicationSchemas(pubid);
+ List *pubschemalist = GetPublicationSchemas(pubid, objectType);
ListCell *cell;
+ AssertObjectTypeValid(objectType);
+
foreach(cell, pubschemalist)
{
Oid schemaid = lfirst_oid(cell);
List *schemaRels = NIL;
- schemaRels = GetSchemaPublicationRelations(schemaid, pub_partopt);
+ schemaRels = GetSchemaPublicationRelations(schemaid, objectType,
+ pub_partopt);
result = list_concat(result, schemaRels);
}
return result;
}
+/*
+ * Gets list of all relation published by FOR ALL SEQUENCES publication(s).
+ */
+List *
+GetAllSequencesPublicationRelations(void)
+{
+ Relation classRel;
+ ScanKeyData key[1];
+ TableScanDesc scan;
+ HeapTuple tuple;
+ List *result = NIL;
+
+ classRel = table_open(RelationRelationId, AccessShareLock);
+
+ ScanKeyInit(&key[0],
+ Anum_pg_class_relkind,
+ BTEqualStrategyNumber, F_CHAREQ,
+ CharGetDatum(RELKIND_SEQUENCE));
+
+ scan = table_beginscan_catalog(classRel, 1, key);
+
+ while ((tuple = heap_getnext(scan, ForwardScanDirection)) != NULL)
+ {
+ Form_pg_class relForm = (Form_pg_class) GETSTRUCT(tuple);
+ Oid relid = relForm->oid;
+
+ if (is_publishable_class(relid, relForm))
+ result = lappend_oid(result, relid);
+ }
+
+ table_endscan(scan);
+
+ table_close(classRel, AccessShareLock);
+ return result;
+}
+
/*
* Get publication using oid
*
@@ -852,10 +1045,12 @@ GetPublication(Oid pubid)
pub->oid = pubid;
pub->name = pstrdup(NameStr(pubform->pubname));
pub->alltables = pubform->puballtables;
+ pub->allsequences = pubform->puballsequences;
pub->pubactions.pubinsert = pubform->pubinsert;
pub->pubactions.pubupdate = pubform->pubupdate;
pub->pubactions.pubdelete = pubform->pubdelete;
pub->pubactions.pubtruncate = pubform->pubtruncate;
+ pub->pubactions.pubsequence = pubform->pubsequence;
pub->pubviaroot = pubform->pubviaroot;
ReleaseSysCache(tup);
@@ -966,10 +1161,12 @@ pg_get_publication_tables(PG_FUNCTION_ARGS)
*schemarelids;
relids = GetPublicationRelations(publication->oid,
+ PUB_OBJTYPE_TABLE,
publication->pubviaroot ?
PUBLICATION_PART_ROOT :
PUBLICATION_PART_LEAF);
schemarelids = GetAllSchemaPublicationRelations(publication->oid,
+ PUB_OBJTYPE_TABLE,
publication->pubviaroot ?
PUBLICATION_PART_ROOT :
PUBLICATION_PART_LEAF);
@@ -1005,3 +1202,71 @@ pg_get_publication_tables(PG_FUNCTION_ARGS)
SRF_RETURN_DONE(funcctx);
}
+
+/*
+ * Returns Oids of sequences in a publication.
+ */
+Datum
+pg_get_publication_sequences(PG_FUNCTION_ARGS)
+{
+ FuncCallContext *funcctx;
+ char *pubname = text_to_cstring(PG_GETARG_TEXT_PP(0));
+ Publication *publication;
+ List *sequences;
+
+ /* stuff done only on the first call of the function */
+ if (SRF_IS_FIRSTCALL())
+ {
+ MemoryContext oldcontext;
+
+ /* create a function context for cross-call persistence */
+ funcctx = SRF_FIRSTCALL_INIT();
+
+ /* switch to memory context appropriate for multiple function calls */
+ oldcontext = MemoryContextSwitchTo(funcctx->multi_call_memory_ctx);
+
+ publication = GetPublicationByName(pubname, false);
+
+ /*
+ * Publications support partitioned tables, although all changes are
+ * replicated using leaf partition identity and schema, so we only
+ * need those.
+ */
+ if (publication->allsequences)
+ sequences = GetAllSequencesPublicationRelations();
+ else
+ {
+ List *relids,
+ *schemarelids;
+
+ relids = GetPublicationRelations(publication->oid,
+ PUB_OBJTYPE_SEQUENCE,
+ publication->pubviaroot ?
+ PUBLICATION_PART_ROOT :
+ PUBLICATION_PART_LEAF);
+ schemarelids = GetAllSchemaPublicationRelations(publication->oid,
+ PUB_OBJTYPE_SEQUENCE,
+ publication->pubviaroot ?
+ PUBLICATION_PART_ROOT :
+ PUBLICATION_PART_LEAF);
+ sequences = list_concat_unique_oid(relids, schemarelids);
+ }
+
+ funcctx->user_fctx = (void *) sequences;
+
+ MemoryContextSwitchTo(oldcontext);
+ }
+
+ /* stuff done on every call of the function */
+ funcctx = SRF_PERCALL_SETUP();
+ sequences = (List *) funcctx->user_fctx;
+
+ if (funcctx->call_cntr < list_length(sequences))
+ {
+ Oid relid = list_nth_oid(sequences, funcctx->call_cntr);
+
+ SRF_RETURN_NEXT(funcctx, ObjectIdGetDatum(relid));
+ }
+
+ SRF_RETURN_DONE(funcctx);
+}
diff --git a/src/backend/catalog/system_views.sql b/src/backend/catalog/system_views.sql
index bd48ee7bd25..9ac8e9a2998 100644
--- a/src/backend/catalog/system_views.sql
+++ b/src/backend/catalog/system_views.sql
@@ -374,6 +374,16 @@ CREATE VIEW pg_publication_tables AS
pg_class C JOIN pg_namespace N ON (N.oid = C.relnamespace)
WHERE C.oid = GPT.relid;
+CREATE VIEW pg_publication_sequences AS
+ SELECT
+ P.pubname AS pubname,
+ N.nspname AS schemaname,
+ C.relname AS sequencename
+ FROM pg_publication P,
+ LATERAL pg_get_publication_sequences(P.pubname) GPS,
+ pg_class C JOIN pg_namespace N ON (N.oid = C.relnamespace)
+ WHERE C.oid = GPS.relid;
+
CREATE VIEW pg_locks AS
SELECT * FROM pg_lock_status() AS L;
diff --git a/src/backend/commands/publicationcmds.c b/src/backend/commands/publicationcmds.c
index 1aad2e769cb..92505f4699f 100644
--- a/src/backend/commands/publicationcmds.c
+++ b/src/backend/commands/publicationcmds.c
@@ -16,6 +16,7 @@
#include "access/genam.h"
#include "access/htup_details.h"
+#include "access/relation.h"
#include "access/table.h"
#include "access/xact.h"
#include "catalog/catalog.h"
@@ -67,15 +68,17 @@ typedef struct rf_context
} rf_context;
static List *OpenRelIdList(List *relids);
-static List *OpenTableList(List *tables);
-static void CloseTableList(List *rels);
+static List *OpenRelationList(List *rels, char objectType);
+static void CloseRelationList(List *rels);
static void LockSchemaList(List *schemalist);
-static void PublicationAddTables(Oid pubid, List *rels, bool if_not_exists,
+static void PublicationAddRelations(Oid pubid, List *rels, bool if_not_exists,
AlterPublicationStmt *stmt);
-static void PublicationDropTables(Oid pubid, List *rels, bool missing_ok);
-static void PublicationAddSchemas(Oid pubid, List *schemas, bool if_not_exists,
- AlterPublicationStmt *stmt);
-static void PublicationDropSchemas(Oid pubid, List *schemas, bool missing_ok);
+static void PublicationDropRelations(Oid pubid, List *rels, bool missing_ok);
+static void PublicationAddSchemas(Oid pubid, List *schemas, char objectType,
+ bool if_not_exists, AlterPublicationStmt *stmt);
+static void PublicationDropSchemas(Oid pubid, List *schemas, char objectType,
+ bool missing_ok);
+
static void
parse_publication_options(ParseState *pstate,
@@ -95,6 +98,7 @@ parse_publication_options(ParseState *pstate,
pubactions->pubupdate = true;
pubactions->pubdelete = true;
pubactions->pubtruncate = true;
+ pubactions->pubsequence = true;
*publish_via_partition_root = false;
/* Parse options */
@@ -119,6 +123,7 @@ parse_publication_options(ParseState *pstate,
pubactions->pubupdate = false;
pubactions->pubdelete = false;
pubactions->pubtruncate = false;
+ pubactions->pubsequence = false;
*publish_given = true;
publish = defGetString(defel);
@@ -141,6 +146,8 @@ parse_publication_options(ParseState *pstate,
pubactions->pubdelete = true;
else if (strcmp(publish_opt, "truncate") == 0)
pubactions->pubtruncate = true;
+ else if (strcmp(publish_opt, "sequence") == 0)
+ pubactions->pubsequence = true;
else
ereport(ERROR,
(errcode(ERRCODE_SYNTAX_ERROR),
@@ -167,7 +174,9 @@ parse_publication_options(ParseState *pstate,
*/
static void
ObjectsInPublicationToOids(List *pubobjspec_list, ParseState *pstate,
- List **rels, List **schemas)
+ List **tables, List **sequences,
+ List **tables_schemas, List **sequences_schemas,
+ List **schemas)
{
ListCell *cell;
PublicationObjSpec *pubobj;
@@ -185,12 +194,23 @@ ObjectsInPublicationToOids(List *pubobjspec_list, ParseState *pstate,
switch (pubobj->pubobjtype)
{
case PUBLICATIONOBJ_TABLE:
- *rels = lappend(*rels, pubobj->pubtable);
+ *tables = lappend(*tables, pubobj->pubtable);
+ break;
+ case PUBLICATIONOBJ_SEQUENCE:
+ *sequences = lappend(*sequences, pubobj->pubtable);
break;
case PUBLICATIONOBJ_TABLES_IN_SCHEMA:
schemaid = get_namespace_oid(pubobj->name, false);
/* Filter out duplicates if user specifies "sch1, sch1" */
+ *tables_schemas = list_append_unique_oid(*tables_schemas, schemaid);
+ *schemas = list_append_unique_oid(*schemas, schemaid);
+ break;
+ case PUBLICATIONOBJ_SEQUENCES_IN_SCHEMA:
+ schemaid = get_namespace_oid(pubobj->name, false);
+
+ /* Filter out duplicates if user specifies "sch1, sch1" */
+ *sequences_schemas = list_append_unique_oid(*sequences_schemas, schemaid);
*schemas = list_append_unique_oid(*schemas, schemaid);
break;
case PUBLICATIONOBJ_TABLES_IN_CUR_SCHEMA:
@@ -204,6 +224,21 @@ ObjectsInPublicationToOids(List *pubobjspec_list, ParseState *pstate,
list_free(search_path);
/* Filter out duplicates if user specifies "sch1, sch1" */
+ *tables_schemas = list_append_unique_oid(*tables_schemas, schemaid);
+ *schemas = list_append_unique_oid(*schemas, schemaid);
+ break;
+ case PUBLICATIONOBJ_SEQUENCES_IN_CUR_SCHEMA:
+ search_path = fetch_search_path(false);
+ if (search_path == NIL) /* nothing valid in search_path? */
+ ereport(ERROR,
+ errcode(ERRCODE_UNDEFINED_SCHEMA),
+ errmsg("no schema has been selected for CURRENT_SCHEMA"));
+
+ schemaid = linitial_oid(search_path);
+ list_free(search_path);
+
+ /* Filter out duplicates if user specifies "sch1, sch1" */
+ *sequences_schemas = list_append_unique_oid(*sequences_schemas, schemaid);
*schemas = list_append_unique_oid(*schemas, schemaid);
break;
default:
@@ -240,6 +275,14 @@ CheckObjSchemaNotAlreadyInPublication(List *rels, List *schemaidlist,
errdetail("Table \"%s\" in schema \"%s\" is already part of the publication, adding the same schema is not supported.",
RelationGetRelationName(rel),
get_namespace_name(relSchemaId)));
+ else if (checkobjtype == PUBLICATIONOBJ_SEQUENCES_IN_SCHEMA)
+ ereport(ERROR,
+ errcode(ERRCODE_INVALID_PARAMETER_VALUE),
+ errmsg("cannot add schema \"%s\" to publication",
+ get_namespace_name(relSchemaId)),
+ errdetail("SEquence \"%s\" in schema \"%s\" is already part of the publication, adding the same schema is not supported.",
+ RelationGetRelationName(rel),
+ get_namespace_name(relSchemaId)));
else if (checkobjtype == PUBLICATIONOBJ_TABLE)
ereport(ERROR,
errcode(ERRCODE_INVALID_PARAMETER_VALUE),
@@ -248,6 +291,14 @@ CheckObjSchemaNotAlreadyInPublication(List *rels, List *schemaidlist,
RelationGetRelationName(rel)),
errdetail("Table's schema \"%s\" is already part of the publication or part of the specified schema list.",
get_namespace_name(relSchemaId)));
+ else if (checkobjtype == PUBLICATIONOBJ_SEQUENCE)
+ ereport(ERROR,
+ errcode(ERRCODE_INVALID_PARAMETER_VALUE),
+ errmsg("cannot add relation \"%s.%s\" to publication",
+ get_namespace_name(relSchemaId),
+ RelationGetRelationName(rel)),
+ errdetail("Sequence's schema \"%s\" is already part of the publication or part of the specified schema list.",
+ get_namespace_name(relSchemaId)));
}
}
}
@@ -615,6 +666,7 @@ TransformPubWhereClauses(List *tables, const char *queryString,
ObjectAddress
CreatePublication(ParseState *pstate, CreatePublicationStmt *stmt)
{
+ ListCell *lc;
Relation rel;
ObjectAddress myself;
Oid puboid;
@@ -626,9 +678,31 @@ CreatePublication(ParseState *pstate, CreatePublicationStmt *stmt)
bool publish_via_partition_root_given;
bool publish_via_partition_root;
AclResult aclresult;
- List *relations = NIL;
+ List *tables = NIL;
+ List *sequences = NIL;
+ List *tables_schemaidlist = NIL;
+ List *sequences_schemaidlist = NIL;
List *schemaidlist = NIL;
+ bool for_all_tables = false;
+ bool for_all_sequences = false;
+
+ /*
+ * XXX Translate the list of object types (represented by strings) to bool
+ * flags. Not sure this is the best way, maybe it should be done in the
+ * grammar directly, somehow.
+ *
+ * XXX Should prevent duplicates, like FOR ALL TABLES, TABLES, TABLES?
+ */
+ foreach (lc, stmt->for_all_objects)
+ {
+ char *val = strVal(lfirst(lc));
+ if (strcmp(val, "tables") == 0)
+ for_all_tables = true;
+ else if (strcmp(val, "sequences") == 0)
+ for_all_sequences = true;
+ }
+
/* must have CREATE privilege on database */
aclresult = pg_database_aclcheck(MyDatabaseId, GetUserId(), ACL_CREATE);
if (aclresult != ACLCHECK_OK)
@@ -636,7 +710,7 @@ CreatePublication(ParseState *pstate, CreatePublicationStmt *stmt)
get_database_name(MyDatabaseId));
/* FOR ALL TABLES requires superuser */
- if (stmt->for_all_tables && !superuser())
+ if (for_all_tables && !superuser())
ereport(ERROR,
(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
errmsg("must be superuser to create FOR ALL TABLES publication")));
@@ -672,7 +746,9 @@ CreatePublication(ParseState *pstate, CreatePublicationStmt *stmt)
Anum_pg_publication_oid);
values[Anum_pg_publication_oid - 1] = ObjectIdGetDatum(puboid);
values[Anum_pg_publication_puballtables - 1] =
- BoolGetDatum(stmt->for_all_tables);
+ BoolGetDatum(for_all_tables);
+ values[Anum_pg_publication_puballsequences - 1] =
+ BoolGetDatum(for_all_sequences);
values[Anum_pg_publication_pubinsert - 1] =
BoolGetDatum(pubactions.pubinsert);
values[Anum_pg_publication_pubupdate - 1] =
@@ -681,6 +757,8 @@ CreatePublication(ParseState *pstate, CreatePublicationStmt *stmt)
BoolGetDatum(pubactions.pubdelete);
values[Anum_pg_publication_pubtruncate - 1] =
BoolGetDatum(pubactions.pubtruncate);
+ values[Anum_pg_publication_pubsequence - 1] =
+ BoolGetDatum(pubactions.pubsequence);
values[Anum_pg_publication_pubviaroot - 1] =
BoolGetDatum(publish_via_partition_root);
@@ -698,45 +776,84 @@ CreatePublication(ParseState *pstate, CreatePublicationStmt *stmt)
CommandCounterIncrement();
/* Associate objects with the publication. */
- if (stmt->for_all_tables)
+ if (for_all_tables || for_all_sequences)
{
/* Invalidate relcache so that publication info is rebuilt. */
CacheInvalidateRelcacheAll();
}
- else
+
+ if (!for_all_tables || !for_all_sequences)
{
- ObjectsInPublicationToOids(stmt->pubobjects, pstate, &relations,
+ ObjectsInPublicationToOids(stmt->pubobjects, pstate,
+ &tables, &sequences,
+ &tables_schemaidlist,
+ &sequences_schemaidlist,
&schemaidlist);
/* FOR ALL TABLES IN SCHEMA requires superuser */
- if (list_length(schemaidlist) > 0 && !superuser())
+ if (list_length(tables_schemaidlist) > 0 && !superuser())
ereport(ERROR,
errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
errmsg("must be superuser to create FOR ALL TABLES IN SCHEMA publication"));
- if (list_length(relations) > 0)
+ /* FOR ALL SEQUENCES IN SCHEMA requires superuser */
+ if (list_length(sequences_schemaidlist) > 0 && !superuser())
+ ereport(ERROR,
+ errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
+ errmsg("must be superuser to create FOR ALL SEQUENCES IN SCHEMA publication"));
+
+ /* tables added directly */
+ if (list_length(tables) > 0)
{
List *rels;
- rels = OpenTableList(relations);
- CheckObjSchemaNotAlreadyInPublication(rels, schemaidlist,
+ rels = OpenRelationList(tables, PUB_OBJTYPE_TABLE);
+ CheckObjSchemaNotAlreadyInPublication(rels, tables_schemaidlist,
PUBLICATIONOBJ_TABLE);
TransformPubWhereClauses(rels, pstate->p_sourcetext,
publish_via_partition_root);
- PublicationAddTables(puboid, rels, true, NULL);
- CloseTableList(rels);
+ PublicationAddRelations(puboid, rels, true, NULL);
+ CloseRelationList(rels);
}
- if (list_length(schemaidlist) > 0)
+ /* sequences added directly */
+ if (list_length(sequences) > 0)
+ {
+ List *rels;
+
+ rels = OpenRelationList(sequences, PUB_OBJTYPE_SEQUENCE);
+ CheckObjSchemaNotAlreadyInPublication(rels, sequences_schemaidlist,
+ PUBLICATIONOBJ_SEQUENCE);
+ PublicationAddRelations(puboid, rels, true, NULL);
+ CloseRelationList(rels);
+ }
+
+ /* tables added through a schema */
+ if (list_length(tables_schemaidlist) > 0)
{
/*
* Schema lock is held until the publication is created to prevent
* concurrent schema deletion.
*/
- LockSchemaList(schemaidlist);
- PublicationAddSchemas(puboid, schemaidlist, true, NULL);
+ LockSchemaList(tables_schemaidlist);
+ PublicationAddSchemas(puboid,
+ tables_schemaidlist, PUB_OBJTYPE_TABLE,
+ true, NULL);
+ }
+
+ /* sequences added through a schema */
+ if (list_length(sequences_schemaidlist) > 0)
+ {
+ /*
+ * Schema lock is held until the publication is created to prevent
+ * concurrent schema deletion.
+ */
+ LockSchemaList(sequences_schemaidlist);
+ PublicationAddSchemas(puboid,
+ sequences_schemaidlist, PUB_OBJTYPE_SEQUENCE,
+ true, NULL);
}
}
@@ -799,6 +916,7 @@ AlterPublicationOptions(ParseState *pstate, AlterPublicationStmt *stmt,
AccessShareLock);
root_relids = GetPublicationRelations(pubform->oid,
+ PUB_OBJTYPE_TABLE,
PUBLICATION_PART_ROOT);
foreach(lc, root_relids)
@@ -857,6 +975,9 @@ AlterPublicationOptions(ParseState *pstate, AlterPublicationStmt *stmt,
values[Anum_pg_publication_pubtruncate - 1] = BoolGetDatum(pubactions.pubtruncate);
replaces[Anum_pg_publication_pubtruncate - 1] = true;
+
+ values[Anum_pg_publication_pubsequence - 1] = BoolGetDatum(pubactions.pubsequence);
+ replaces[Anum_pg_publication_pubsequence - 1] = true;
}
if (publish_via_partition_root_given)
@@ -876,7 +997,7 @@ AlterPublicationOptions(ParseState *pstate, AlterPublicationStmt *stmt,
pubform = (Form_pg_publication) GETSTRUCT(tup);
/* Invalidate the relcache. */
- if (pubform->puballtables)
+ if (pubform->puballtables || pubform->puballsequences)
{
CacheInvalidateRelcacheAll();
}
@@ -892,6 +1013,7 @@ AlterPublicationOptions(ParseState *pstate, AlterPublicationStmt *stmt,
*/
if (root_relids == NIL)
relids = GetPublicationRelations(pubform->oid,
+ PUB_OBJTYPE_TABLE,
PUBLICATION_PART_ALL);
else
{
@@ -905,7 +1027,20 @@ AlterPublicationOptions(ParseState *pstate, AlterPublicationStmt *stmt,
lfirst_oid(lc));
}
+ /* tables */
+ schemarelids = GetAllSchemaPublicationRelations(pubform->oid,
+ PUB_OBJTYPE_TABLE,
+ PUBLICATION_PART_ALL);
+ relids = list_concat_unique_oid(relids, schemarelids);
+
+ /* sequences */
+ relids = list_concat_unique_oid(relids,
+ GetPublicationRelations(pubform->oid,
+ PUB_OBJTYPE_SEQUENCE,
+ PUBLICATION_PART_ALL));
+
schemarelids = GetAllSchemaPublicationRelations(pubform->oid,
+ PUB_OBJTYPE_SEQUENCE,
PUBLICATION_PART_ALL);
relids = list_concat_unique_oid(relids, schemarelids);
@@ -960,7 +1095,7 @@ AlterPublicationTables(AlterPublicationStmt *stmt, HeapTuple tup,
if (!tables && stmt->action != AP_SetObjects)
return;
- rels = OpenTableList(tables);
+ rels = OpenRelationList(tables, PUB_OBJTYPE_TABLE);
if (stmt->action == AP_AddObjects)
{
@@ -970,19 +1105,22 @@ AlterPublicationTables(AlterPublicationStmt *stmt, HeapTuple tup,
* Check if the relation is member of the existing schema in the
* publication or member of the schema list specified.
*/
- schemas = list_concat_copy(schemaidlist, GetPublicationSchemas(pubid));
+ schemas = list_concat_copy(schemaidlist,
+ GetPublicationSchemas(pubid,
+ PUB_OBJTYPE_TABLE));
CheckObjSchemaNotAlreadyInPublication(rels, schemas,
PUBLICATIONOBJ_TABLE);
TransformPubWhereClauses(rels, queryString, pubform->pubviaroot);
- PublicationAddTables(pubid, rels, false, stmt);
+ PublicationAddRelations(pubid, rels, false, stmt);
}
else if (stmt->action == AP_DropObjects)
- PublicationDropTables(pubid, rels, false);
+ PublicationDropRelations(pubid, rels, false);
else /* AP_SetObjects */
{
List *oldrelids = GetPublicationRelations(pubid,
+ PUB_OBJTYPE_TABLE,
PUBLICATION_PART_ROOT);
List *delrels = NIL;
ListCell *oldlc;
@@ -1064,18 +1202,18 @@ AlterPublicationTables(AlterPublicationStmt *stmt, HeapTuple tup,
}
/* And drop them. */
- PublicationDropTables(pubid, delrels, true);
+ PublicationDropRelations(pubid, delrels, true);
/*
* Don't bother calculating the difference for adding, we'll catch and
* skip existing ones when doing catalog update.
*/
- PublicationAddTables(pubid, rels, true, stmt);
+ PublicationAddRelations(pubid, rels, true, stmt);
- CloseTableList(delrels);
+ CloseRelationList(delrels);
}
- CloseTableList(rels);
+ CloseRelationList(rels);
}
/*
@@ -1085,7 +1223,8 @@ AlterPublicationTables(AlterPublicationStmt *stmt, HeapTuple tup,
*/
static void
AlterPublicationSchemas(AlterPublicationStmt *stmt,
- HeapTuple tup, List *schemaidlist)
+ HeapTuple tup, List *schemaidlist,
+ char objectType)
{
Form_pg_publication pubform = (Form_pg_publication) GETSTRUCT(tup);
@@ -1107,20 +1246,20 @@ AlterPublicationSchemas(AlterPublicationStmt *stmt,
List *rels;
List *reloids;
- reloids = GetPublicationRelations(pubform->oid, PUBLICATION_PART_ROOT);
+ reloids = GetPublicationRelations(pubform->oid, objectType, PUBLICATION_PART_ROOT);
rels = OpenRelIdList(reloids);
CheckObjSchemaNotAlreadyInPublication(rels, schemaidlist,
PUBLICATIONOBJ_TABLES_IN_SCHEMA);
- CloseTableList(rels);
- PublicationAddSchemas(pubform->oid, schemaidlist, false, stmt);
+ CloseRelationList(rels);
+ PublicationAddSchemas(pubform->oid, schemaidlist, objectType, false, stmt);
}
else if (stmt->action == AP_DropObjects)
- PublicationDropSchemas(pubform->oid, schemaidlist, false);
+ PublicationDropSchemas(pubform->oid, schemaidlist, objectType, false);
else /* AP_SetObjects */
{
- List *oldschemaids = GetPublicationSchemas(pubform->oid);
+ List *oldschemaids = GetPublicationSchemas(pubform->oid, objectType);
List *delschemas = NIL;
/* Identify which schemas should be dropped */
@@ -1133,13 +1272,13 @@ AlterPublicationSchemas(AlterPublicationStmt *stmt,
LockSchemaList(delschemas);
/* And drop them */
- PublicationDropSchemas(pubform->oid, delschemas, true);
+ PublicationDropSchemas(pubform->oid, delschemas, objectType, true);
/*
* Don't bother calculating the difference for adding, we'll catch and
* skip existing ones when doing catalog update.
*/
- PublicationAddSchemas(pubform->oid, schemaidlist, true, stmt);
+ PublicationAddSchemas(pubform->oid, schemaidlist, objectType, true, stmt);
}
}
@@ -1149,12 +1288,13 @@ AlterPublicationSchemas(AlterPublicationStmt *stmt,
*/
static void
CheckAlterPublication(AlterPublicationStmt *stmt, HeapTuple tup,
- List *tables, List *schemaidlist)
+ List *tables, List *tables_schemaidlist,
+ List *sequences, List *sequences_schemaidlist)
{
Form_pg_publication pubform = (Form_pg_publication) GETSTRUCT(tup);
if ((stmt->action == AP_AddObjects || stmt->action == AP_SetObjects) &&
- schemaidlist && !superuser())
+ (tables_schemaidlist || sequences_schemaidlist) && !superuser())
ereport(ERROR,
(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
errmsg("must be superuser to add or set schemas")));
@@ -1163,13 +1303,24 @@ CheckAlterPublication(AlterPublicationStmt *stmt, HeapTuple tup,
* Check that user is allowed to manipulate the publication tables in
* schema
*/
- if (schemaidlist && pubform->puballtables)
+ if (tables_schemaidlist && pubform->puballtables)
ereport(ERROR,
(errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
errmsg("publication \"%s\" is defined as FOR ALL TABLES",
NameStr(pubform->pubname)),
errdetail("Tables from schema cannot be added to, dropped from, or set on FOR ALL TABLES publications.")));
+ /*
+ * Check that user is allowed to manipulate the publication sequences in
+ * schema
+ */
+ if (sequences_schemaidlist && pubform->puballsequences)
+ ereport(ERROR,
+ (errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
+ errmsg("publication \"%s\" is defined as FOR ALL SEQUENCES",
+ NameStr(pubform->pubname)),
+ errdetail("Sequences from schema cannot be added to, dropped from, or set on FOR ALL SEQUENCES publications.")));
+
/* Check that user is allowed to manipulate the publication tables. */
if (tables && pubform->puballtables)
ereport(ERROR,
@@ -1177,6 +1328,107 @@ CheckAlterPublication(AlterPublicationStmt *stmt, HeapTuple tup,
errmsg("publication \"%s\" is defined as FOR ALL TABLES",
NameStr(pubform->pubname)),
errdetail("Tables cannot be added to or dropped from FOR ALL TABLES publications.")));
+
+ /* Check that user is allowed to manipulate the publication tables. */
+ if (sequences && pubform->puballsequences)
+ ereport(ERROR,
+ (errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
+ errmsg("publication \"%s\" is defined as FOR ALL SEQUENCES",
+ NameStr(pubform->pubname)),
+ errdetail("Sequences cannot be added to or dropped from FOR ALL SEQUENCES publications.")));
+}
+
+/*
+ * Add or remove sequence to/from publication.
+ */
+static void
+AlterPublicationSequences(AlterPublicationStmt *stmt, HeapTuple tup,
+ List *sequences, List *schemaidlist)
+{
+ List *rels = NIL;
+ Form_pg_publication pubform = (Form_pg_publication) GETSTRUCT(tup);
+ Oid pubid = pubform->oid;
+
+ /*
+ * It is quite possible that for the SET case user has not specified any
+ * tables in which case we need to remove all the existing tables.
+ */
+ if (!sequences && stmt->action != AP_SetObjects)
+ return;
+
+ rels = OpenRelationList(sequences, PUB_OBJTYPE_SEQUENCE);
+
+ if (stmt->action == AP_AddObjects)
+ {
+ List *schemas = NIL;
+
+ /*
+ * Check if the relation is member of the existing schema in the
+ * publication or member of the schema list specified.
+ */
+ schemas = list_concat_copy(schemaidlist,
+ GetPublicationSchemas(pubid,
+ PUB_OBJTYPE_SEQUENCE));
+ CheckObjSchemaNotAlreadyInPublication(rels, schemas,
+ PUBLICATIONOBJ_SEQUENCE);
+ PublicationAddRelations(pubid, rels, false, stmt);
+ }
+ else if (stmt->action == AP_DropObjects)
+ PublicationDropRelations(pubid, rels, false);
+ else /* DEFELEM_SET */
+ {
+ List *oldrelids = GetPublicationRelations(pubid,
+ PUB_OBJTYPE_SEQUENCE,
+ PUBLICATION_PART_ROOT);
+ List *delrels = NIL;
+ ListCell *oldlc;
+
+ CheckObjSchemaNotAlreadyInPublication(rels, schemaidlist,
+ PUBLICATIONOBJ_SEQUENCE);
+
+ /* Calculate which relations to drop. */
+ foreach(oldlc, oldrelids)
+ {
+ Oid oldrelid = lfirst_oid(oldlc);
+ ListCell *newlc;
+ PublicationRelInfo *oldrel;
+ bool found = false;
+
+ foreach(newlc, rels)
+ {
+ PublicationRelInfo *newpubrel;
+
+ newpubrel = (PublicationRelInfo *) lfirst(newlc);
+ if (RelationGetRelid(newpubrel->relation) == oldrelid)
+ {
+ found = true;
+ break;
+ }
+ }
+ /* Not yet in the list, open it and add to the list */
+ if (!found)
+ {
+ oldrel = palloc(sizeof(PublicationRelInfo));
+ oldrel->whereClause = NULL;
+ oldrel->relation = table_open(oldrelid,
+ ShareUpdateExclusiveLock);
+ delrels = lappend(delrels, oldrel);
+ }
+ }
+
+ /* And drop them. */
+ PublicationDropRelations(pubid, delrels, true);
+
+ /*
+ * Don't bother calculating the difference for adding, we'll catch and
+ * skip existing ones when doing catalog update.
+ */
+ PublicationAddRelations(pubid, rels, true, stmt);
+
+ CloseRelationList(delrels);
+ }
+
+ CloseRelationList(rels);
}
/*
@@ -1214,14 +1466,22 @@ AlterPublication(ParseState *pstate, AlterPublicationStmt *stmt)
AlterPublicationOptions(pstate, stmt, rel, tup);
else
{
- List *relations = NIL;
+ List *tables = NIL;
+ List *sequences = NIL;
+ List *tables_schemaidlist = NIL;
+ List *sequences_schemaidlist = NIL;
List *schemaidlist = NIL;
Oid pubid = pubform->oid;
- ObjectsInPublicationToOids(stmt->pubobjects, pstate, &relations,
+ ObjectsInPublicationToOids(stmt->pubobjects, pstate,
+ &tables, &sequences,
+ &tables_schemaidlist,
+ &sequences_schemaidlist,
&schemaidlist);
- CheckAlterPublication(stmt, tup, relations, schemaidlist);
+ CheckAlterPublication(stmt, tup,
+ tables, tables_schemaidlist,
+ sequences, sequences_schemaidlist);
heap_freetuple(tup);
@@ -1249,9 +1509,16 @@ AlterPublication(ParseState *pstate, AlterPublicationStmt *stmt)
errmsg("publication \"%s\" does not exist",
stmt->pubname));
- AlterPublicationTables(stmt, tup, relations, schemaidlist,
+ AlterPublicationTables(stmt, tup, tables, tables_schemaidlist,
pstate->p_sourcetext);
- AlterPublicationSchemas(stmt, tup, schemaidlist);
+
+ AlterPublicationSequences(stmt, tup, sequences, sequences_schemaidlist);
+
+ AlterPublicationSchemas(stmt, tup, tables_schemaidlist,
+ PUB_OBJTYPE_TABLE);
+
+ AlterPublicationSchemas(stmt, tup, sequences_schemaidlist,
+ PUB_OBJTYPE_SEQUENCE);
}
/* Cleanup. */
@@ -1319,7 +1586,7 @@ RemovePublicationById(Oid pubid)
pubform = (Form_pg_publication) GETSTRUCT(tup);
/* Invalidate relcache so that publication info is rebuilt. */
- if (pubform->puballtables)
+ if (pubform->puballtables || pubform->puballsequences)
CacheInvalidateRelcacheAll();
CatalogTupleDelete(rel, &tup->t_self);
@@ -1355,6 +1622,7 @@ RemovePublicationSchemaById(Oid psoid)
* partitions.
*/
schemaRels = GetSchemaPublicationRelations(pubsch->pnnspid,
+ pubsch->pntype,
PUBLICATION_PART_ALL);
InvalidatePublicationRels(schemaRels);
@@ -1397,29 +1665,45 @@ OpenRelIdList(List *relids)
* add them to a publication.
*/
static List *
-OpenTableList(List *tables)
+OpenRelationList(List *rels, char objectType)
{
List *relids = NIL;
- List *rels = NIL;
+ List *result = NIL;
ListCell *lc;
List *relids_with_rf = NIL;
/*
* Open, share-lock, and check all the explicitly-specified relations
*/
- foreach(lc, tables)
+ foreach(lc, rels)
{
PublicationTable *t = lfirst_node(PublicationTable, lc);
bool recurse = t->relation->inh;
Relation rel;
Oid myrelid;
PublicationRelInfo *pub_rel;
+ char myrelkind;
/* Allow query cancel in case this takes a long time */
CHECK_FOR_INTERRUPTS();
rel = table_openrv(t->relation, ShareUpdateExclusiveLock);
myrelid = RelationGetRelid(rel);
+ myrelkind = get_rel_relkind(myrelid);
+
+ /*
+ * Make sure the relkind matches the expected object type. This happens
+ * e.g. when adding a sequence using ADD TABLE or a table using ADD
+ * SEQUENCE).
+ *
+ * XXX We let through unsupported object types (views etc.). Those are
+ * caught later in check_publication_add_relation. A bit weird.
+ *
+ * FIXME Use a better error message.
+ */
+ if (pub_get_object_type_for_relkind(myrelkind) != PUB_OBJTYPE_UNSUPPORTED &&
+ pub_get_object_type_for_relkind(myrelkind) != objectType)
+ elog(ERROR, "object type mismatch");
/*
* Filter out duplicates if user specifies "foo, foo".
@@ -1444,7 +1728,7 @@ OpenTableList(List *tables)
pub_rel = palloc(sizeof(PublicationRelInfo));
pub_rel->relation = rel;
pub_rel->whereClause = t->whereClause;
- rels = lappend(rels, pub_rel);
+ result = lappend(result, pub_rel);
relids = lappend_oid(relids, myrelid);
if (t->whereClause)
@@ -1498,7 +1782,7 @@ OpenTableList(List *tables)
pub_rel->relation = rel;
/* child inherits WHERE clause from parent */
pub_rel->whereClause = t->whereClause;
- rels = lappend(rels, pub_rel);
+ result = lappend(result, pub_rel);
relids = lappend_oid(relids, childrelid);
if (t->whereClause)
@@ -1510,14 +1794,14 @@ OpenTableList(List *tables)
list_free(relids);
list_free(relids_with_rf);
- return rels;
+ return result;
}
/*
* Close all relations in the list.
*/
static void
-CloseTableList(List *rels)
+CloseRelationList(List *rels)
{
ListCell *lc;
@@ -1565,12 +1849,12 @@ LockSchemaList(List *schemalist)
* Add listed tables to the publication.
*/
static void
-PublicationAddTables(Oid pubid, List *rels, bool if_not_exists,
+PublicationAddRelations(Oid pubid, List *rels, bool if_not_exists,
AlterPublicationStmt *stmt)
{
ListCell *lc;
- Assert(!stmt || !stmt->for_all_tables);
+ Assert(!stmt || !stmt->for_all_objects);
foreach(lc, rels)
{
@@ -1599,7 +1883,7 @@ PublicationAddTables(Oid pubid, List *rels, bool if_not_exists,
* Remove listed tables from the publication.
*/
static void
-PublicationDropTables(Oid pubid, List *rels, bool missing_ok)
+PublicationDropRelations(Oid pubid, List *rels, bool missing_ok)
{
ObjectAddress obj;
ListCell *lc;
@@ -1639,19 +1923,19 @@ PublicationDropTables(Oid pubid, List *rels, bool missing_ok)
* Add listed schemas to the publication.
*/
static void
-PublicationAddSchemas(Oid pubid, List *schemas, bool if_not_exists,
- AlterPublicationStmt *stmt)
+PublicationAddSchemas(Oid pubid, List *schemas, char objectType,
+ bool if_not_exists, AlterPublicationStmt *stmt)
{
ListCell *lc;
- Assert(!stmt || !stmt->for_all_tables);
+ Assert(!stmt || !stmt->for_all_objects);
foreach(lc, schemas)
{
Oid schemaid = lfirst_oid(lc);
ObjectAddress obj;
- obj = publication_add_schema(pubid, schemaid, if_not_exists);
+ obj = publication_add_schema(pubid, schemaid, objectType, if_not_exists);
if (stmt)
{
EventTriggerCollectSimpleCommand(obj, InvalidObjectAddress,
@@ -1667,7 +1951,7 @@ PublicationAddSchemas(Oid pubid, List *schemas, bool if_not_exists,
* Remove listed schemas from the publication.
*/
static void
-PublicationDropSchemas(Oid pubid, List *schemas, bool missing_ok)
+PublicationDropSchemas(Oid pubid, List *schemas, char objectType, bool missing_ok)
{
ObjectAddress obj;
ListCell *lc;
@@ -1677,10 +1961,11 @@ PublicationDropSchemas(Oid pubid, List *schemas, bool missing_ok)
{
Oid schemaid = lfirst_oid(lc);
- psid = GetSysCacheOid2(PUBLICATIONNAMESPACEMAP,
+ psid = GetSysCacheOid3(PUBLICATIONNAMESPACEMAP,
Anum_pg_publication_namespace_oid,
ObjectIdGetDatum(schemaid),
- ObjectIdGetDatum(pubid));
+ ObjectIdGetDatum(pubid),
+ CharGetDatum(objectType));
if (!OidIsValid(psid))
{
if (missing_ok)
@@ -1735,6 +2020,13 @@ AlterPublicationOwner_internal(Relation rel, HeapTuple tup, Oid newOwnerId)
NameStr(form->pubname)),
errhint("The owner of a FOR ALL TABLES publication must be a superuser.")));
+ if (form->puballsequences && !superuser_arg(newOwnerId))
+ ereport(ERROR,
+ (errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
+ errmsg("permission denied to change owner of publication \"%s\"",
+ NameStr(form->pubname)),
+ errhint("The owner of a FOR ALL SEQUENCES publication must be a superuser.")));
+
if (!superuser_arg(newOwnerId) && is_schema_publication(form->oid))
ereport(ERROR,
(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
diff --git a/src/backend/commands/sequence.c b/src/backend/commands/sequence.c
index c13cada3bf1..717bb0b2aa9 100644
--- a/src/backend/commands/sequence.c
+++ b/src/backend/commands/sequence.c
@@ -336,6 +336,160 @@ ResetSequence(Oid seq_relid)
relation_close(seq_rel, NoLock);
}
+/*
+ * Update the sequence state by modifying the existing sequence data row.
+ *
+ * This keeps the same relfilenode, so the behavior is non-transactional.
+ */
+static void
+SetSequence_non_transactional(Oid seqrelid, int64 last_value, int64 log_cnt, bool is_called)
+{
+ SeqTable elm;
+ Relation seqrel;
+ Buffer buf;
+ HeapTupleData seqdatatuple;
+ Form_pg_sequence_data seq;
+
+ /* open and lock sequence */
+ init_sequence(seqrelid, &elm, &seqrel);
+
+ /* lock page' buffer and read tuple */
+ seq = read_seq_tuple(seqrel, &buf, &seqdatatuple);
+
+ /* check the comment above nextval_internal()'s equivalent call. */
+ if (RelationNeedsWAL(seqrel))
+ {
+ GetTopTransactionId();
+
+ if (XLogLogicalInfoActive())
+ GetCurrentTransactionId();
+ }
+
+ /* ready to change the on-disk (or really, in-buffer) tuple */
+ START_CRIT_SECTION();
+
+ seq->last_value = last_value;
+ seq->is_called = is_called;
+ seq->log_cnt = log_cnt;
+
+ MarkBufferDirty(buf);
+
+ /* XLOG stuff */
+ if (RelationNeedsWAL(seqrel))
+ {
+ xl_seq_rec xlrec;
+ XLogRecPtr recptr;
+ Page page = BufferGetPage(buf);
+
+ XLogBeginInsert();
+ XLogRegisterBuffer(0, buf, REGBUF_WILL_INIT);
+
+ xlrec.node = seqrel->rd_node;
+ xlrec.created = false;
+
+ XLogRegisterData((char *) &xlrec, sizeof(xl_seq_rec));
+ XLogRegisterData((char *) seqdatatuple.t_data, seqdatatuple.t_len);
+
+ recptr = XLogInsert(RM_SEQ_ID, XLOG_SEQ_LOG);
+
+ PageSetLSN(page, recptr);
+ }
+
+ END_CRIT_SECTION();
+
+ UnlockReleaseBuffer(buf);
+
+ /* Clear local cache so that we don't think we have cached numbers */
+ /* Note that we do not change the currval() state */
+ elm->cached = elm->last;
+
+ relation_close(seqrel, NoLock);
+}
+
+/*
+ * Update the sequence state by creating a new relfilenode.
+ *
+ * This creates a new relfilenode, to allow transactional behavior.
+ */
+static void
+SetSequence_transactional(Oid seq_relid, int64 last_value, int64 log_cnt, bool is_called)
+{
+ SeqTable elm;
+ Relation seqrel;
+ Buffer buf;
+ HeapTupleData seqdatatuple;
+ Form_pg_sequence_data seq;
+ HeapTuple tuple;
+
+ /* open and lock sequence */
+ init_sequence(seq_relid, &elm, &seqrel);
+
+ /* lock page' buffer and read tuple */
+ seq = read_seq_tuple(seqrel, &buf, &seqdatatuple);
+
+ /* Copy the existing sequence tuple. */
+ tuple = heap_copytuple(&seqdatatuple);
+
+ /* Now we're done with the old page */
+ UnlockReleaseBuffer(buf);
+
+ /*
+ * Modify the copied tuple to update the sequence state (similar to what
+ * ResetSequence does).
+ */
+ seq = (Form_pg_sequence_data) GETSTRUCT(tuple);
+ seq->last_value = last_value;
+ seq->is_called = is_called;
+ seq->log_cnt = log_cnt;
+
+ /*
+ * Create a new storage file for the sequence - this is needed for the
+ * transactional behavior.
+ */
+ RelationSetNewRelfilenode(seqrel, seqrel->rd_rel->relpersistence);
+
+ /*
+ * Ensure sequence's relfrozenxid is at 0, since it won't contain any
+ * unfrozen XIDs. Same with relminmxid, since a sequence will never
+ * contain multixacts.
+ */
+ Assert(seqrel->rd_rel->relfrozenxid == InvalidTransactionId);
+ Assert(seqrel->rd_rel->relminmxid == InvalidMultiXactId);
+
+ /*
+ * Insert the modified tuple into the new storage file. This does all the
+ * necessary WAL-logging etc.
+ */
+ fill_seq_with_data(seqrel, tuple);
+
+ /* Clear local cache so that we don't think we have cached numbers */
+ /* Note that we do not change the currval() state */
+ elm->cached = elm->last;
+
+ relation_close(seqrel, NoLock);
+}
+
+/*
+ * Set a sequence to a specified internal state.
+ *
+ * The change is made transactionally, so that on failure of the current
+ * transaction, the sequence will be restored to its previous state.
+ * We do that by creating a whole new relfilenode for the sequence; so this
+ * works much like the rewriting forms of ALTER TABLE.
+ *
+ * Caller is assumed to have acquired AccessExclusiveLock on the sequence,
+ * which must not be released until end of transaction. Caller is also
+ * responsible for permissions checking.
+ */
+void
+SetSequence(Oid seq_relid, bool transactional, int64 last_value, int64 log_cnt, bool is_called)
+{
+ if (transactional)
+ SetSequence_transactional(seq_relid, last_value, log_cnt, is_called);
+ else
+ SetSequence_non_transactional(seq_relid, last_value, log_cnt, is_called);
+}
+
/*
* Initialize a sequence's relation with the specified tuple as content
*/
diff --git a/src/backend/commands/subscriptioncmds.c b/src/backend/commands/subscriptioncmds.c
index e16f04626de..14ee492a9ba 100644
--- a/src/backend/commands/subscriptioncmds.c
+++ b/src/backend/commands/subscriptioncmds.c
@@ -90,6 +90,7 @@ typedef struct SubOpts
} SubOpts;
static List *fetch_table_list(WalReceiverConn *wrconn, List *publications);
+static List *fetch_sequence_list(WalReceiverConn *wrconn, List *publications);
static void check_duplicates_in_publist(List *publist, Datum *datums);
static List *merge_publications(List *oldpublist, List *newpublist, bool addpub, const char *subname);
static void ReportSlotConnectionError(List *rstates, Oid subid, char *slotname, char *err);
@@ -541,9 +542,9 @@ CreateSubscription(ParseState *pstate, CreateSubscriptionStmt *stmt,
{
char *err;
WalReceiverConn *wrconn;
- List *tables;
+ List *relations;
ListCell *lc;
- char table_state;
+ char sync_state;
/* Try to connect to the publisher. */
wrconn = walrcv_connect(conninfo, true, stmt->subname, &err);
@@ -558,14 +559,17 @@ CreateSubscription(ParseState *pstate, CreateSubscriptionStmt *stmt,
* Set sync state based on if we were asked to do data copy or
* not.
*/
- table_state = opts.copy_data ? SUBREL_STATE_INIT : SUBREL_STATE_READY;
+ sync_state = opts.copy_data ? SUBREL_STATE_INIT : SUBREL_STATE_READY;
/*
- * Get the table list from publisher and build local table status
- * info.
+ * Get the table and sequence list from publisher and build
+ * local relation sync status info.
*/
- tables = fetch_table_list(wrconn, publications);
- foreach(lc, tables)
+ relations = fetch_table_list(wrconn, publications);
+ relations = list_concat(relations,
+ fetch_sequence_list(wrconn, publications));
+
+ foreach(lc, relations)
{
RangeVar *rv = (RangeVar *) lfirst(lc);
Oid relid;
@@ -576,7 +580,7 @@ CreateSubscription(ParseState *pstate, CreateSubscriptionStmt *stmt,
CheckSubscriptionRelkind(get_rel_relkind(relid),
rv->schemaname, rv->relname);
- AddSubscriptionRelState(subid, relid, table_state,
+ AddSubscriptionRelState(subid, relid, sync_state,
InvalidXLogRecPtr);
}
@@ -602,12 +606,12 @@ CreateSubscription(ParseState *pstate, CreateSubscriptionStmt *stmt,
*
* Note that if tables were specified but copy_data is false
* then it is safe to enable two_phase up-front because those
- * tables are already initially in READY state. When the
- * subscription has no tables, we leave the twophase state as
- * PENDING, to allow ALTER SUBSCRIPTION ... REFRESH
+ * relations are already initially in READY state. When the
+ * subscription has no relations, we leave the twophase state
+ * as PENDING, to allow ALTER SUBSCRIPTION ... REFRESH
* PUBLICATION to work.
*/
- if (opts.twophase && !opts.copy_data && tables != NIL)
+ if (opts.twophase && !opts.copy_data && relations != NIL)
twophase_enabled = true;
walrcv_create_slot(wrconn, opts.slot_name, false, twophase_enabled,
@@ -677,8 +681,10 @@ AlterSubscription_refresh(Subscription *sub, bool copy_data)
PG_TRY();
{
- /* Get the table list from publisher. */
+ /* Get the list of relations from publisher. */
pubrel_names = fetch_table_list(wrconn, sub->publications);
+ pubrel_names = list_concat(pubrel_names,
+ fetch_sequence_list(wrconn, sub->publications));
/* Get local table list. */
subrel_states = GetSubscriptionRelations(sub->oid);
@@ -1712,6 +1718,75 @@ fetch_table_list(WalReceiverConn *wrconn, List *publications)
return tablelist;
}
+/*
+ * Get the list of sequences which belong to specified publications on the
+ * publisher connection.
+ */
+static List *
+fetch_sequence_list(WalReceiverConn *wrconn, List *publications)
+{
+ WalRcvExecResult *res;
+ StringInfoData cmd;
+ TupleTableSlot *slot;
+ Oid tableRow[2] = {TEXTOID, TEXTOID};
+ ListCell *lc;
+ bool first;
+ List *tablelist = NIL;
+
+ Assert(list_length(publications) > 0);
+
+ initStringInfo(&cmd);
+ appendStringInfoString(&cmd, "SELECT DISTINCT s.schemaname, s.sequencename\n"
+ " FROM pg_catalog.pg_publication_sequences s\n"
+ " WHERE s.pubname IN (");
+ first = true;
+ foreach(lc, publications)
+ {
+ char *pubname = strVal(lfirst(lc));
+
+ if (first)
+ first = false;
+ else
+ appendStringInfoString(&cmd, ", ");
+
+ appendStringInfoString(&cmd, quote_literal_cstr(pubname));
+ }
+ appendStringInfoChar(&cmd, ')');
+
+ res = walrcv_exec(wrconn, cmd.data, 2, tableRow);
+ pfree(cmd.data);
+
+ if (res->status != WALRCV_OK_TUPLES)
+ ereport(ERROR,
+ (errmsg("could not receive list of replicated sequences from the publisher: %s",
+ res->err)));
+
+ /* Process tables. */
+ slot = MakeSingleTupleTableSlot(res->tupledesc, &TTSOpsMinimalTuple);
+ while (tuplestore_gettupleslot(res->tuplestore, true, false, slot))
+ {
+ char *nspname;
+ char *relname;
+ bool isnull;
+ RangeVar *rv;
+
+ nspname = TextDatumGetCString(slot_getattr(slot, 1, &isnull));
+ Assert(!isnull);
+ relname = TextDatumGetCString(slot_getattr(slot, 2, &isnull));
+ Assert(!isnull);
+
+ rv = makeRangeVar(nspname, relname, -1);
+ tablelist = lappend(tablelist, rv);
+
+ ExecClearTuple(slot);
+ }
+ ExecDropSingleTupleTableSlot(slot);
+
+ walrcv_clear_result(res);
+
+ return tablelist;
+}
+
/*
* This is to report the connection failure while dropping replication slots.
* Here, we report the WARNING for all tablesync slots so that user can drop
diff --git a/src/backend/commands/tablecmds.c b/src/backend/commands/tablecmds.c
index 80faae985e9..38bf572e1f6 100644
--- a/src/backend/commands/tablecmds.c
+++ b/src/backend/commands/tablecmds.c
@@ -42,6 +42,7 @@
#include "catalog/pg_inherits.h"
#include "catalog/pg_namespace.h"
#include "catalog/pg_opclass.h"
+#include "catalog/pg_publication_namespace.h"
#include "catalog/pg_statistic_ext.h"
#include "catalog/pg_tablespace.h"
#include "catalog/pg_trigger.h"
@@ -16381,11 +16382,14 @@ AlterTableNamespace(AlterObjectSchemaStmt *stmt, Oid *oldschema)
* Check that setting the relation to a different schema won't result in a
* publication having both a schema and the same schema's table, as this
* is not supported.
+ *
+ * XXX The blocks are almost exactly the same, so maybe combine them into
+ * a single block.
*/
if (stmt->objectType == OBJECT_TABLE)
{
ListCell *lc;
- List *schemaPubids = GetSchemaPublications(nspOid);
+ List *schemaPubids = GetSchemaPublications(nspOid, PUB_OBJTYPE_TABLE);
List *relPubids = GetRelationPublications(RelationGetRelid(rel));
foreach(lc, relPubids)
@@ -16403,6 +16407,27 @@ AlterTableNamespace(AlterObjectSchemaStmt *stmt, Oid *oldschema)
get_publication_name(pubid, false)));
}
}
+ else if (stmt->objectType == OBJECT_SEQUENCE)
+ {
+ ListCell *lc;
+ List *schemaPubids = GetSchemaPublications(nspOid, PUB_OBJTYPE_SEQUENCE);
+ List *relPubids = GetRelationPublications(RelationGetRelid(rel));
+
+ foreach(lc, relPubids)
+ {
+ Oid pubid = lfirst_oid(lc);
+
+ if (list_member_oid(schemaPubids, pubid))
+ ereport(ERROR,
+ errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
+ errmsg("cannot move sequence \"%s\" to schema \"%s\"",
+ RelationGetRelationName(rel), stmt->newschema),
+ errdetail("The schema \"%s\" and same schema's sequence \"%s\" cannot be part of the same publication \"%s\".",
+ stmt->newschema,
+ RelationGetRelationName(rel),
+ get_publication_name(pubid, false)));
+ }
+ }
/* common checks on switching namespaces */
CheckSetNamespace(oldNspOid, nspOid);
diff --git a/src/backend/executor/execReplication.c b/src/backend/executor/execReplication.c
index 13328141e23..0df7cf58747 100644
--- a/src/backend/executor/execReplication.c
+++ b/src/backend/executor/execReplication.c
@@ -636,7 +636,9 @@ void
CheckSubscriptionRelkind(char relkind, const char *nspname,
const char *relname)
{
- if (relkind != RELKIND_RELATION && relkind != RELKIND_PARTITIONED_TABLE)
+ if (relkind != RELKIND_RELATION &&
+ relkind != RELKIND_PARTITIONED_TABLE &&
+ relkind != RELKIND_SEQUENCE)
ereport(ERROR,
(errcode(ERRCODE_WRONG_OBJECT_TYPE),
errmsg("cannot use relation \"%s.%s\" as logical replication target",
diff --git a/src/backend/nodes/copyfuncs.c b/src/backend/nodes/copyfuncs.c
index d4f8455a2bd..55f720a88f4 100644
--- a/src/backend/nodes/copyfuncs.c
+++ b/src/backend/nodes/copyfuncs.c
@@ -4862,7 +4862,7 @@ _copyCreatePublicationStmt(const CreatePublicationStmt *from)
COPY_STRING_FIELD(pubname);
COPY_NODE_FIELD(options);
COPY_NODE_FIELD(pubobjects);
- COPY_SCALAR_FIELD(for_all_tables);
+ COPY_NODE_FIELD(for_all_objects);
return newnode;
}
@@ -4875,7 +4875,7 @@ _copyAlterPublicationStmt(const AlterPublicationStmt *from)
COPY_STRING_FIELD(pubname);
COPY_NODE_FIELD(options);
COPY_NODE_FIELD(pubobjects);
- COPY_SCALAR_FIELD(for_all_tables);
+ COPY_NODE_FIELD(for_all_objects);
COPY_SCALAR_FIELD(action);
return newnode;
diff --git a/src/backend/nodes/equalfuncs.c b/src/backend/nodes/equalfuncs.c
index f1002afe7a0..82562eb9b87 100644
--- a/src/backend/nodes/equalfuncs.c
+++ b/src/backend/nodes/equalfuncs.c
@@ -2333,7 +2333,7 @@ _equalCreatePublicationStmt(const CreatePublicationStmt *a,
COMPARE_STRING_FIELD(pubname);
COMPARE_NODE_FIELD(options);
COMPARE_NODE_FIELD(pubobjects);
- COMPARE_SCALAR_FIELD(for_all_tables);
+ COMPARE_NODE_FIELD(for_all_objects);
return true;
}
@@ -2345,7 +2345,7 @@ _equalAlterPublicationStmt(const AlterPublicationStmt *a,
COMPARE_STRING_FIELD(pubname);
COMPARE_NODE_FIELD(options);
COMPARE_NODE_FIELD(pubobjects);
- COMPARE_SCALAR_FIELD(for_all_tables);
+ COMPARE_NODE_FIELD(for_all_objects);
COMPARE_SCALAR_FIELD(action);
return true;
diff --git a/src/backend/parser/gram.y b/src/backend/parser/gram.y
index 0036c2f9e2d..e327bc735fb 100644
--- a/src/backend/parser/gram.y
+++ b/src/backend/parser/gram.y
@@ -446,7 +446,7 @@ static Node *makeRecursiveViewSelect(char *relname, List *aliases, Node *query);
transform_element_list transform_type_list
TriggerTransitions TriggerReferencing
vacuum_relation_list opt_vacuum_relation_list
- drop_option_list pub_obj_list
+ drop_option_list pub_obj_list pub_obj_type_list
%type <node> opt_routine_body
%type <groupclause> group_clause
@@ -575,6 +575,7 @@ static Node *makeRecursiveViewSelect(char *relname, List *aliases, Node *query);
%type <node> var_value zone_value
%type <rolespec> auth_ident RoleSpec opt_granted_by
%type <publicationobjectspec> PublicationObjSpec
+%type <node> pub_obj_type
%type <keyword> unreserved_keyword type_func_name_keyword
%type <keyword> col_name_keyword reserved_keyword
@@ -9701,12 +9702,9 @@ AlterOwnerStmt: ALTER AGGREGATE aggregate_with_argtypes OWNER TO RoleSpec
*
* CREATE PUBLICATION FOR ALL TABLES [WITH options]
*
- * CREATE PUBLICATION FOR pub_obj [, ...] [WITH options]
- *
- * pub_obj is one of:
+ * CREATE PUBLICATION FOR ALL SEQUENCES [WITH options]
*
- * TABLE table [, ...]
- * ALL TABLES IN SCHEMA schema [, ...]
+ * CREATE PUBLICATION FOR pub_obj [, ...] [WITH options]
*
*****************************************************************************/
@@ -9718,12 +9716,12 @@ CreatePublicationStmt:
n->options = $4;
$$ = (Node *)n;
}
- | CREATE PUBLICATION name FOR ALL TABLES opt_definition
+ | CREATE PUBLICATION name FOR ALL pub_obj_type_list opt_definition
{
CreatePublicationStmt *n = makeNode(CreatePublicationStmt);
n->pubname = $3;
n->options = $7;
- n->for_all_tables = true;
+ n->for_all_objects = $6;
$$ = (Node *)n;
}
| CREATE PUBLICATION name FOR pub_obj_list opt_definition
@@ -9772,6 +9770,26 @@ PublicationObjSpec:
$$->pubobjtype = PUBLICATIONOBJ_TABLES_IN_CUR_SCHEMA;
$$->location = @5;
}
+ | SEQUENCE relation_expr
+ {
+ $$ = makeNode(PublicationObjSpec);
+ $$->pubobjtype = PUBLICATIONOBJ_SEQUENCE;
+ $$->pubtable = makeNode(PublicationTable);
+ $$->pubtable->relation = $2;
+ }
+ | ALL SEQUENCES IN_P SCHEMA ColId
+ {
+ $$ = makeNode(PublicationObjSpec);
+ $$->pubobjtype = PUBLICATIONOBJ_SEQUENCES_IN_SCHEMA;
+ $$->name = $5;
+ $$->location = @5;
+ }
+ | ALL SEQUENCES IN_P SCHEMA CURRENT_SCHEMA
+ {
+ $$ = makeNode(PublicationObjSpec);
+ $$->pubobjtype = PUBLICATIONOBJ_SEQUENCES_IN_CUR_SCHEMA;
+ $$->location = @5;
+ }
| ColId OptWhereClause
{
$$ = makeNode(PublicationObjSpec);
@@ -9826,6 +9844,19 @@ pub_obj_list: PublicationObjSpec
{ $$ = lappend($1, $3); }
;
+pub_obj_type: TABLES
+ { $$ = (Node *) makeString("tables"); }
+ | SEQUENCES
+ { $$ = (Node *) makeString("sequences"); }
+ ;
+
+pub_obj_type_list: pub_obj_type
+ { $$ = list_make1($1); }
+ | pub_obj_type_list ',' pub_obj_type
+ { $$ = lappend($1, $3); }
+ ;
+
+
/*****************************************************************************
*
* ALTER PUBLICATION name SET ( options )
@@ -9836,11 +9867,6 @@ pub_obj_list: PublicationObjSpec
*
* ALTER PUBLICATION name SET pub_obj [, ...]
*
- * pub_obj is one of:
- *
- * TABLE table_name [, ...]
- * ALL TABLES IN SCHEMA schema_name [, ...]
- *
*****************************************************************************/
AlterPublicationStmt:
diff --git a/src/backend/replication/logical/proto.c b/src/backend/replication/logical/proto.c
index c9b0eeefd7e..3dbe85d61a2 100644
--- a/src/backend/replication/logical/proto.c
+++ b/src/backend/replication/logical/proto.c
@@ -648,6 +648,56 @@ logicalrep_write_message(StringInfo out, TransactionId xid, XLogRecPtr lsn,
pq_sendbytes(out, message, sz);
}
+/*
+ * Write SEQUENCE to stream
+ */
+void
+logicalrep_write_sequence(StringInfo out, Relation rel, TransactionId xid,
+ XLogRecPtr lsn, bool transactional,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ uint8 flags = 0;
+ char *relname;
+
+ pq_sendbyte(out, LOGICAL_REP_MSG_SEQUENCE);
+
+ /* transaction ID (if not valid, we're not streaming) */
+ if (TransactionIdIsValid(xid))
+ pq_sendint32(out, xid);
+
+ pq_sendint8(out, flags);
+ pq_sendint64(out, lsn);
+
+ logicalrep_write_namespace(out, RelationGetNamespace(rel));
+ relname = RelationGetRelationName(rel);
+ pq_sendstring(out, relname);
+
+ pq_sendint8(out, transactional);
+ pq_sendint64(out, last_value);
+ pq_sendint64(out, log_cnt);
+ pq_sendint8(out, is_called);
+}
+
+/*
+ * Read SEQUENCE from the stream.
+ */
+void
+logicalrep_read_sequence(StringInfo in, LogicalRepSequence *seqdata)
+{
+ /* XXX skipping flags and lsn */
+ pq_getmsgint(in, 1);
+ pq_getmsgint64(in);
+
+ /* Read relation name from stream */
+ seqdata->nspname = pstrdup(logicalrep_read_namespace(in));
+ seqdata->seqname = pstrdup(pq_getmsgstring(in));
+
+ seqdata->transactional = pq_getmsgint(in, 1);
+ seqdata->last_value = pq_getmsgint64(in);
+ seqdata->log_cnt = pq_getmsgint64(in);
+ seqdata->is_called = pq_getmsgint(in, 1);
+}
+
/*
* Write relation description to the output stream.
*/
@@ -1203,6 +1253,8 @@ logicalrep_message_type(LogicalRepMsgType action)
return "STREAM ABORT";
case LOGICAL_REP_MSG_STREAM_PREPARE:
return "STREAM PREPARE";
+ case LOGICAL_REP_MSG_SEQUENCE:
+ return "SEQUENCE";
}
elog(ERROR, "invalid logical replication message type \"%c\"", action);
diff --git a/src/backend/replication/logical/tablesync.c b/src/backend/replication/logical/tablesync.c
index 1659964571c..36ecc5d74c4 100644
--- a/src/backend/replication/logical/tablesync.c
+++ b/src/backend/replication/logical/tablesync.c
@@ -100,6 +100,7 @@
#include "catalog/pg_subscription_rel.h"
#include "catalog/pg_type.h"
#include "commands/copy.h"
+#include "commands/sequence.h"
#include "miscadmin.h"
#include "parser/parse_relation.h"
#include "pgstat.h"
@@ -359,6 +360,8 @@ process_syncing_tables_for_sync(XLogRecPtr current_lsn)
*
* If the synchronization position is reached (SYNCDONE), then the table can
* be marked as READY and is no longer tracked.
+ *
+ * XXX Maybe rename this to "_relations_" to cover both tables and sequences.
*/
static void
process_syncing_tables_for_apply(XLogRecPtr current_lsn)
@@ -999,6 +1002,95 @@ copy_table(Relation rel)
logicalrep_rel_close(relmapentry, NoLock);
}
+/*
+ * Fetch sequence data (current state) from the remote node.
+ */
+static void
+fetch_sequence_data(char *nspname, char *relname,
+ int64 *last_value, int64 *log_cnt, bool *is_called)
+{
+ WalRcvExecResult *res;
+ StringInfoData cmd;
+ TupleTableSlot *slot;
+ Oid tableRow[3] = {INT8OID, INT8OID, BOOLOID};
+
+ initStringInfo(&cmd);
+ appendStringInfo(&cmd, "SELECT last_value, log_cnt, is_called\n"
+ " FROM %s", quote_qualified_identifier(nspname, relname));
+
+ res = walrcv_exec(LogRepWorkerWalRcvConn, cmd.data, 3, tableRow);
+ pfree(cmd.data);
+
+ if (res->status != WALRCV_OK_TUPLES)
+ ereport(ERROR,
+ (errmsg("could not receive list of replicated tables from the publisher: %s",
+ res->err)));
+
+ /* Process the sequence. */
+ slot = MakeSingleTupleTableSlot(res->tupledesc, &TTSOpsMinimalTuple);
+ while (tuplestore_gettupleslot(res->tuplestore, true, false, slot))
+ {
+ bool isnull;
+
+ *last_value = DatumGetInt64(slot_getattr(slot, 1, &isnull));
+ Assert(!isnull);
+
+ *log_cnt = DatumGetInt64(slot_getattr(slot, 2, &isnull));
+ Assert(!isnull);
+
+ *is_called = DatumGetBool(slot_getattr(slot, 3, &isnull));
+ Assert(!isnull);
+
+ ExecClearTuple(slot);
+ }
+ ExecDropSingleTupleTableSlot(slot);
+
+ walrcv_clear_result(res);
+}
+
+/*
+ * Copy existing data of a sequence from publisher.
+ *
+ * Caller is responsible for locking the local relation.
+ */
+static void
+copy_sequence(Relation rel)
+{
+ LogicalRepRelMapEntry *relmapentry;
+ LogicalRepRelation lrel;
+ List *qual = NIL;
+ StringInfoData cmd;
+ int64 last_value = 0,
+ log_cnt = 0;
+ bool is_called = 0;
+
+ /* Get the publisher relation info. */
+ fetch_remote_table_info(get_namespace_name(RelationGetNamespace(rel)),
+ RelationGetRelationName(rel), &lrel, &qual);
+
+ /* sequences don't have row filters */
+ Assert(!qual);
+
+ /* Put the relation into relmap. */
+ logicalrep_relmap_update(&lrel);
+
+ /* Map the publisher relation to local one. */
+ relmapentry = logicalrep_rel_open(lrel.remoteid, NoLock);
+ Assert(rel == relmapentry->localrel);
+
+ /* Start copy on the publisher. */
+ initStringInfo(&cmd);
+
+ Assert(lrel.relkind == RELKIND_SEQUENCE);
+
+ fetch_sequence_data(lrel.nspname, lrel.relname, &last_value, &log_cnt, &is_called);
+
+ /* tablesync sets the sequences in non-transactional way */
+ SetSequence(RelationGetRelid(rel), false, last_value, log_cnt, is_called);
+
+ logicalrep_rel_close(relmapentry, NoLock);
+}
+
/*
* Determine the tablesync slot name.
*
@@ -1260,10 +1352,21 @@ LogicalRepSyncTableStart(XLogRecPtr *origin_startpos)
originname)));
}
- /* Now do the initial data copy */
- PushActiveSnapshot(GetTransactionSnapshot());
- copy_table(rel);
- PopActiveSnapshot();
+ /* Do the right action depending on the relation kind. */
+ if (get_rel_relkind(RelationGetRelid(rel)) == RELKIND_SEQUENCE)
+ {
+ /* Now do the initial sequence copy */
+ PushActiveSnapshot(GetTransactionSnapshot());
+ copy_sequence(rel);
+ PopActiveSnapshot();
+ }
+ else
+ {
+ /* Now do the initial data copy */
+ PushActiveSnapshot(GetTransactionSnapshot());
+ copy_table(rel);
+ PopActiveSnapshot();
+ }
res = walrcv_exec(LogRepWorkerWalRcvConn, "COMMIT", 0, NULL);
if (res->status != WALRCV_OK_COMMAND)
diff --git a/src/backend/replication/logical/worker.c b/src/backend/replication/logical/worker.c
index 82dcffc2db8..f3868b3e1f8 100644
--- a/src/backend/replication/logical/worker.c
+++ b/src/backend/replication/logical/worker.c
@@ -143,6 +143,7 @@
#include "catalog/pg_subscription.h"
#include "catalog/pg_subscription_rel.h"
#include "catalog/pg_tablespace.h"
+#include "commands/sequence.h"
#include "commands/tablecmds.h"
#include "commands/tablespace.h"
#include "commands/trigger.h"
@@ -1143,6 +1144,57 @@ apply_handle_origin(StringInfo s)
errmsg_internal("ORIGIN message sent out of order")));
}
+/*
+ * Handle SEQUENCE message.
+ */
+static void
+apply_handle_sequence(StringInfo s)
+{
+ LogicalRepSequence seq;
+ Oid relid;
+
+ if (handle_streamed_transaction(LOGICAL_REP_MSG_SEQUENCE, s))
+ return;
+
+ logicalrep_read_sequence(s, &seq);
+
+ /*
+ * Non-transactional sequence updates should not be part of a remote
+ * transaction. There should not be any running transaction.
+ */
+ Assert((!seq.transactional) || in_remote_transaction);
+ Assert(!(!seq.transactional && in_remote_transaction));
+ Assert(!(!seq.transactional && IsTransactionState()));
+
+ /*
+ * Make sure we're in a transaction (needed by SetSequence). For
+ * non-transactional updates we're guaranteed to start a new one,
+ * and we'll commit it at the end.
+ */
+ if (!IsTransactionState())
+ {
+ StartTransactionCommand();
+ maybe_reread_subscription();
+ }
+
+ relid = RangeVarGetRelid(makeRangeVar(seq.nspname,
+ seq.seqname, -1),
+ RowExclusiveLock, false);
+
+ /* lock the sequence in AccessExclusiveLock, as expected by SetSequence */
+ LockRelationOid(relid, AccessExclusiveLock);
+
+ /* apply the sequence change */
+ SetSequence(relid, seq.transactional, seq.last_value, seq.log_cnt, seq.is_called);
+
+ /*
+ * Commit the per-stream transaction (we only do this when not in
+ * remote transaction, i.e. for non-transactional sequence updates.
+ */
+ if (!in_remote_transaction)
+ CommitTransactionCommand();
+}
+
/*
* Handle STREAM START message.
*/
@@ -2511,6 +2563,10 @@ apply_dispatch(StringInfo s)
*/
break;
+ case LOGICAL_REP_MSG_SEQUENCE:
+ apply_handle_sequence(s);
+ return;
+
case LOGICAL_REP_MSG_STREAM_START:
apply_handle_stream_start(s);
break;
diff --git a/src/backend/replication/pgoutput/pgoutput.c b/src/backend/replication/pgoutput/pgoutput.c
index 5fddab3a3d4..4cdc698cbb3 100644
--- a/src/backend/replication/pgoutput/pgoutput.c
+++ b/src/backend/replication/pgoutput/pgoutput.c
@@ -15,6 +15,7 @@
#include "access/tupconvert.h"
#include "catalog/partition.h"
#include "catalog/pg_publication.h"
+#include "catalog/pg_publication_namespace.h"
#include "catalog/pg_publication_rel.h"
#include "commands/defrem.h"
#include "executor/executor.h"
@@ -53,6 +54,10 @@ static void pgoutput_message(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn, XLogRecPtr message_lsn,
bool transactional, const char *prefix,
Size sz, const char *message);
+static void pgoutput_sequence(LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn, XLogRecPtr sequence_lsn,
+ Relation relation, bool transactional,
+ int64 last_value, int64 log_cnt, bool is_called);
static bool pgoutput_origin_filter(LogicalDecodingContext *ctx,
RepOriginId origin_id);
static void pgoutput_begin_prepare_txn(LogicalDecodingContext *ctx,
@@ -208,6 +213,7 @@ _PG_output_plugin_init(OutputPluginCallbacks *cb)
cb->change_cb = pgoutput_change;
cb->truncate_cb = pgoutput_truncate;
cb->message_cb = pgoutput_message;
+ cb->sequence_cb = pgoutput_sequence;
cb->commit_cb = pgoutput_commit_txn;
cb->begin_prepare_cb = pgoutput_begin_prepare_txn;
@@ -224,6 +230,7 @@ _PG_output_plugin_init(OutputPluginCallbacks *cb)
cb->stream_commit_cb = pgoutput_stream_commit;
cb->stream_change_cb = pgoutput_change;
cb->stream_message_cb = pgoutput_message;
+ cb->stream_sequence_cb = pgoutput_sequence;
cb->stream_truncate_cb = pgoutput_truncate;
/* transaction streaming - two-phase commit */
cb->stream_prepare_cb = pgoutput_stream_prepare_txn;
@@ -237,6 +244,7 @@ parse_output_parameters(List *options, PGOutputData *data)
bool publication_names_given = false;
bool binary_option_given = false;
bool messages_option_given = false;
+ bool sequences_option_given = false;
bool streaming_given = false;
bool two_phase_option_given = false;
@@ -244,6 +252,7 @@ parse_output_parameters(List *options, PGOutputData *data)
data->streaming = false;
data->messages = false;
data->two_phase = false;
+ data->sequences = true;
foreach(lc, options)
{
@@ -312,6 +321,16 @@ parse_output_parameters(List *options, PGOutputData *data)
data->messages = defGetBoolean(defel);
}
+ else if (strcmp(defel->defname, "sequences") == 0)
+ {
+ if (sequences_option_given)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("conflicting or redundant options")));
+ sequences_option_given = true;
+
+ data->sequences = defGetBoolean(defel);
+ }
else if (strcmp(defel->defname, "streaming") == 0)
{
if (streaming_given)
@@ -1440,6 +1459,51 @@ pgoutput_message(LogicalDecodingContext *ctx, ReorderBufferTXN *txn,
OutputPluginWrite(ctx, true);
}
+static void
+pgoutput_sequence(LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn, XLogRecPtr sequence_lsn,
+ Relation relation, bool transactional,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ PGOutputData *data = (PGOutputData *) ctx->output_plugin_private;
+ TransactionId xid = InvalidTransactionId;
+ RelationSyncEntry *relentry;
+
+ if (!data->sequences)
+ return;
+
+ if (!is_publishable_relation(relation))
+ return;
+
+ /*
+ * Remember the xid for the message in streaming mode. See
+ * pgoutput_change.
+ */
+ if (in_streaming)
+ xid = txn->xid;
+
+ relentry = get_rel_sync_entry(data, relation);
+
+ /*
+ * First check the sequence filter.
+ *
+ * We handle just REORDER_BUFFER_CHANGE_SEQUENCE here.
+ */
+ if (!relentry->pubactions.pubsequence)
+ return;
+
+ OutputPluginPrepareWrite(ctx, true);
+ logicalrep_write_sequence(ctx->out,
+ relation,
+ xid,
+ sequence_lsn,
+ transactional,
+ last_value,
+ log_cnt,
+ is_called);
+ OutputPluginWrite(ctx, true);
+}
+
/*
* Currently we always forward.
*/
@@ -1725,7 +1789,8 @@ get_rel_sync_entry(PGOutputData *data, Relation relation)
entry->schema_sent = false;
entry->streamed_txns = NIL;
entry->pubactions.pubinsert = entry->pubactions.pubupdate =
- entry->pubactions.pubdelete = entry->pubactions.pubtruncate = false;
+ entry->pubactions.pubdelete = entry->pubactions.pubtruncate =
+ entry->pubactions.pubsequence = false;
entry->new_slot = NULL;
entry->old_slot = NULL;
memset(entry->exprstate, 0, sizeof(entry->exprstate));
@@ -1739,13 +1804,13 @@ get_rel_sync_entry(PGOutputData *data, Relation relation)
{
Oid schemaId = get_rel_namespace(relid);
List *pubids = GetRelationPublications(relid);
-
+ char objectType = pub_get_object_type_for_relkind(get_rel_relkind(relid));
/*
* We don't acquire a lock on the namespace system table as we build
* the cache entry using a historic snapshot and all the later changes
* are absorbed while decoding WAL.
*/
- List *schemaPubids = GetSchemaPublications(schemaId);
+ List *schemaPubids = GetSchemaPublications(schemaId, objectType);
ListCell *lc;
Oid publish_as_relid = relid;
int publish_ancestor_level = 0;
@@ -1780,6 +1845,7 @@ get_rel_sync_entry(PGOutputData *data, Relation relation)
entry->pubactions.pubupdate = false;
entry->pubactions.pubdelete = false;
entry->pubactions.pubtruncate = false;
+ entry->pubactions.pubsequence = false;
/*
* Tuple slots cleanups. (Will be rebuilt later if needed).
@@ -1826,9 +1892,11 @@ get_rel_sync_entry(PGOutputData *data, Relation relation)
/*
* If this is a FOR ALL TABLES publication, pick the partition root
- * and set the ancestor level accordingly.
+ * and set the ancestor level accordingly. If this is a FOR ALL
+ * SEQUENCES publication, we publish it too but we don't need to
+ * pick the partition root etc.
*/
- if (pub->alltables)
+ if (pub->alltables || pub->allsequences)
{
publish = true;
if (pub->pubviaroot && am_partition)
@@ -1889,6 +1957,7 @@ get_rel_sync_entry(PGOutputData *data, Relation relation)
entry->pubactions.pubupdate |= pub->pubactions.pubupdate;
entry->pubactions.pubdelete |= pub->pubactions.pubdelete;
entry->pubactions.pubtruncate |= pub->pubactions.pubtruncate;
+ entry->pubactions.pubsequence |= pub->pubactions.pubsequence;
/*
* We want to publish the changes as the top-most ancestor
diff --git a/src/backend/utils/cache/relcache.c b/src/backend/utils/cache/relcache.c
index fbd11883e17..9391467960b 100644
--- a/src/backend/utils/cache/relcache.c
+++ b/src/backend/utils/cache/relcache.c
@@ -56,6 +56,7 @@
#include "catalog/pg_opclass.h"
#include "catalog/pg_proc.h"
#include "catalog/pg_publication.h"
+#include "catalog/pg_publication_namespace.h"
#include "catalog/pg_rewrite.h"
#include "catalog/pg_shseclabel.h"
#include "catalog/pg_statistic_ext.h"
@@ -5562,6 +5563,8 @@ RelationBuildPublicationDesc(Relation relation, PublicationDesc *pubdesc)
Oid schemaid;
List *ancestors = NIL;
Oid relid = RelationGetRelid(relation);
+ char relkind = relation->rd_rel->relkind;
+ char objType;
/*
* If not publishable, it publishes no actions. (pgoutput_change() will
@@ -5588,8 +5591,15 @@ RelationBuildPublicationDesc(Relation relation, PublicationDesc *pubdesc)
/* Fetch the publication membership info. */
puboids = GetRelationPublications(relid);
schemaid = RelationGetNamespace(relation);
- puboids = list_concat_unique_oid(puboids, GetSchemaPublications(schemaid));
+ objType = pub_get_object_type_for_relkind(relkind);
+ puboids = list_concat_unique_oid(puboids,
+ GetSchemaPublications(schemaid, objType));
+
+ /*
+ * If this is a partion (and thus a table), lookup all ancestors and track
+ * all publications them too.
+ */
if (relation->rd_rel->relispartition)
{
/* Add publications that the ancestors are in too. */
@@ -5601,12 +5611,23 @@ RelationBuildPublicationDesc(Relation relation, PublicationDesc *pubdesc)
puboids = list_concat_unique_oid(puboids,
GetRelationPublications(ancestor));
+
+ /* include all publications publishing schema of all ancestors */
schemaid = get_rel_namespace(ancestor);
puboids = list_concat_unique_oid(puboids,
- GetSchemaPublications(schemaid));
+ GetSchemaPublications(schemaid,
+ PUB_OBJTYPE_TABLE));
}
}
- puboids = list_concat_unique_oid(puboids, GetAllTablesPublications());
+
+ /*
+ * Consider also FOR ALL TABLES and FOR ALL SEQUENCES publications,
+ * depending on the relkind of the relation.
+ */
+ if (relation->rd_rel->relkind == RELKIND_SEQUENCE)
+ puboids = list_concat_unique_oid(puboids, GetAllSequencesPublications());
+ else
+ puboids = list_concat_unique_oid(puboids, GetAllTablesPublications());
foreach(lc, puboids)
{
@@ -5625,6 +5646,7 @@ RelationBuildPublicationDesc(Relation relation, PublicationDesc *pubdesc)
pubdesc->pubactions.pubupdate |= pubform->pubupdate;
pubdesc->pubactions.pubdelete |= pubform->pubdelete;
pubdesc->pubactions.pubtruncate |= pubform->pubtruncate;
+ pubdesc->pubactions.pubsequence |= pubform->pubsequence;
/*
* Check if all columns referenced in the filter expression are part of
diff --git a/src/backend/utils/cache/syscache.c b/src/backend/utils/cache/syscache.c
index f4e7819f1e2..a675877d195 100644
--- a/src/backend/utils/cache/syscache.c
+++ b/src/backend/utils/cache/syscache.c
@@ -630,12 +630,12 @@ static const struct cachedesc cacheinfo[] = {
64
},
{PublicationNamespaceRelationId, /* PUBLICATIONNAMESPACEMAP */
- PublicationNamespacePnnspidPnpubidIndexId,
- 2,
+ PublicationNamespacePnnspidPnpubidPntypeIndexId,
+ 3,
{
Anum_pg_publication_namespace_pnnspid,
Anum_pg_publication_namespace_pnpubid,
- 0,
+ Anum_pg_publication_namespace_pntype,
0
},
64
diff --git a/src/bin/pg_dump/pg_dump.c b/src/bin/pg_dump/pg_dump.c
index e5816c4ccea..00629f08362 100644
--- a/src/bin/pg_dump/pg_dump.c
+++ b/src/bin/pg_dump/pg_dump.c
@@ -3814,10 +3814,12 @@ getPublications(Archive *fout, int *numPublications)
int i_pubname;
int i_pubowner;
int i_puballtables;
+ int i_puballsequences;
int i_pubinsert;
int i_pubupdate;
int i_pubdelete;
int i_pubtruncate;
+ int i_pubsequence;
int i_pubviaroot;
int i,
ntups;
@@ -3833,23 +3835,29 @@ getPublications(Archive *fout, int *numPublications)
resetPQExpBuffer(query);
/* Get the publications. */
- if (fout->remoteVersion >= 130000)
+ if (fout->remoteVersion >= 150000)
+ appendPQExpBuffer(query,
+ "SELECT p.tableoid, p.oid, p.pubname, "
+ "p.pubowner, "
+ "p.puballtables, p.puballsequences, p.pubinsert, p.pubupdate, p.pubdelete, p.pubtruncate, p.pubsequence, p.pubviaroot "
+ "FROM pg_publication p");
+ else if (fout->remoteVersion >= 130000)
appendPQExpBuffer(query,
"SELECT p.tableoid, p.oid, p.pubname, "
"p.pubowner, "
- "p.puballtables, p.pubinsert, p.pubupdate, p.pubdelete, p.pubtruncate, p.pubviaroot "
+ "p.puballtables, false AS puballsequences, p.pubinsert, p.pubupdate, p.pubdelete, p.pubtruncate, false AS pubsequence, p.pubviaroot "
"FROM pg_publication p");
else if (fout->remoteVersion >= 110000)
appendPQExpBuffer(query,
"SELECT p.tableoid, p.oid, p.pubname, "
"p.pubowner, "
- "p.puballtables, p.pubinsert, p.pubupdate, p.pubdelete, p.pubtruncate, false AS pubviaroot "
+ "p.puballtables, false AS puballsequences, p.pubinsert, p.pubupdate, p.pubdelete, p.pubtruncate, false AS pubsequence, false AS pubviaroot "
"FROM pg_publication p");
else
appendPQExpBuffer(query,
"SELECT p.tableoid, p.oid, p.pubname, "
"p.pubowner, "
- "p.puballtables, p.pubinsert, p.pubupdate, p.pubdelete, false AS pubtruncate, false AS pubviaroot "
+ "p.puballtables, false AS puballsequences, p.pubinsert, p.pubupdate, p.pubdelete, false AS pubtruncate, false AS pubsequence, false AS pubviaroot "
"FROM pg_publication p");
res = ExecuteSqlQuery(fout, query->data, PGRES_TUPLES_OK);
@@ -3861,10 +3869,12 @@ getPublications(Archive *fout, int *numPublications)
i_pubname = PQfnumber(res, "pubname");
i_pubowner = PQfnumber(res, "pubowner");
i_puballtables = PQfnumber(res, "puballtables");
+ i_puballsequences = PQfnumber(res, "puballsequences");
i_pubinsert = PQfnumber(res, "pubinsert");
i_pubupdate = PQfnumber(res, "pubupdate");
i_pubdelete = PQfnumber(res, "pubdelete");
i_pubtruncate = PQfnumber(res, "pubtruncate");
+ i_pubsequence = PQfnumber(res, "pubsequence");
i_pubviaroot = PQfnumber(res, "pubviaroot");
pubinfo = pg_malloc(ntups * sizeof(PublicationInfo));
@@ -3880,6 +3890,8 @@ getPublications(Archive *fout, int *numPublications)
pubinfo[i].rolname = getRoleName(PQgetvalue(res, i, i_pubowner));
pubinfo[i].puballtables =
(strcmp(PQgetvalue(res, i, i_puballtables), "t") == 0);
+ pubinfo[i].puballsequences =
+ (strcmp(PQgetvalue(res, i, i_puballsequences), "t") == 0);
pubinfo[i].pubinsert =
(strcmp(PQgetvalue(res, i, i_pubinsert), "t") == 0);
pubinfo[i].pubupdate =
@@ -3888,6 +3900,8 @@ getPublications(Archive *fout, int *numPublications)
(strcmp(PQgetvalue(res, i, i_pubdelete), "t") == 0);
pubinfo[i].pubtruncate =
(strcmp(PQgetvalue(res, i, i_pubtruncate), "t") == 0);
+ pubinfo[i].pubsequence =
+ (strcmp(PQgetvalue(res, i, i_pubsequence), "t") == 0);
pubinfo[i].pubviaroot =
(strcmp(PQgetvalue(res, i, i_pubviaroot), "t") == 0);
@@ -3933,6 +3947,9 @@ dumpPublication(Archive *fout, const PublicationInfo *pubinfo)
if (pubinfo->puballtables)
appendPQExpBufferStr(query, " FOR ALL TABLES");
+ if (pubinfo->puballsequences)
+ appendPQExpBufferStr(query, " FOR ALL SEQUENCES");
+
appendPQExpBufferStr(query, " WITH (publish = '");
if (pubinfo->pubinsert)
{
@@ -3967,6 +3984,15 @@ dumpPublication(Archive *fout, const PublicationInfo *pubinfo)
first = false;
}
+ if (pubinfo->pubsequence)
+ {
+ if (!first)
+ appendPQExpBufferStr(query, ", ");
+
+ appendPQExpBufferStr(query, "sequence");
+ first = false;
+ }
+
appendPQExpBufferStr(query, "'");
if (pubinfo->pubviaroot)
@@ -4013,6 +4039,7 @@ getPublicationNamespaces(Archive *fout)
int i_oid;
int i_pnpubid;
int i_pnnspid;
+ int i_pntype;
int i,
j,
ntups;
@@ -4024,7 +4051,7 @@ getPublicationNamespaces(Archive *fout)
/* Collect all publication membership info. */
appendPQExpBufferStr(query,
- "SELECT tableoid, oid, pnpubid, pnnspid "
+ "SELECT tableoid, oid, pnpubid, pnnspid, pntype "
"FROM pg_catalog.pg_publication_namespace");
res = ExecuteSqlQuery(fout, query->data, PGRES_TUPLES_OK);
@@ -4034,6 +4061,7 @@ getPublicationNamespaces(Archive *fout)
i_oid = PQfnumber(res, "oid");
i_pnpubid = PQfnumber(res, "pnpubid");
i_pnnspid = PQfnumber(res, "pnnspid");
+ i_pntype = PQfnumber(res, "pntype");
/* this allocation may be more than we need */
pubsinfo = pg_malloc(ntups * sizeof(PublicationSchemaInfo));
@@ -4043,6 +4071,7 @@ getPublicationNamespaces(Archive *fout)
{
Oid pnpubid = atooid(PQgetvalue(res, i, i_pnpubid));
Oid pnnspid = atooid(PQgetvalue(res, i, i_pnnspid));
+ char pntype = PQgetvalue(res, i, i_pntype)[0];
PublicationInfo *pubinfo;
NamespaceInfo *nspinfo;
@@ -4074,6 +4103,7 @@ getPublicationNamespaces(Archive *fout)
pubsinfo[j].dobj.name = nspinfo->dobj.name;
pubsinfo[j].publication = pubinfo;
pubsinfo[j].pubschema = nspinfo;
+ pubsinfo[j].pubtype = pntype;
/* Decide whether we want to dump it */
selectDumpablePublicationObject(&(pubsinfo[j].dobj), fout);
@@ -4207,7 +4237,11 @@ dumpPublicationNamespace(Archive *fout, const PublicationSchemaInfo *pubsinfo)
query = createPQExpBuffer();
appendPQExpBuffer(query, "ALTER PUBLICATION %s ", fmtId(pubinfo->dobj.name));
- appendPQExpBuffer(query, "ADD ALL TABLES IN SCHEMA %s;\n", fmtId(schemainfo->dobj.name));
+
+ if (pubsinfo->pubtype == 't')
+ appendPQExpBuffer(query, "ADD ALL TABLES IN SCHEMA %s;\n", fmtId(schemainfo->dobj.name));
+ else
+ appendPQExpBuffer(query, "ADD ALL SEQUENCES IN SCHEMA %s;\n", fmtId(schemainfo->dobj.name));
/*
* There is no point in creating drop query as the drop is done by schema
@@ -4240,6 +4274,7 @@ dumpPublicationTable(Archive *fout, const PublicationRelInfo *pubrinfo)
TableInfo *tbinfo = pubrinfo->pubtable;
PQExpBuffer query;
char *tag;
+ char *description;
/* Do nothing in data-only dump */
if (dopt->dataOnly)
@@ -4249,10 +4284,22 @@ dumpPublicationTable(Archive *fout, const PublicationRelInfo *pubrinfo)
query = createPQExpBuffer();
- appendPQExpBuffer(query, "ALTER PUBLICATION %s ADD TABLE ONLY",
- fmtId(pubinfo->dobj.name));
+ if (tbinfo->relkind == RELKIND_SEQUENCE)
+ {
+ appendPQExpBuffer(query, "ALTER PUBLICATION %s ADD SEQUENCE",
+ fmtId(pubinfo->dobj.name));
+ description = "PUBLICATION SEQUENCE";
+ }
+ else
+ {
+ appendPQExpBuffer(query, "ALTER PUBLICATION %s ADD TABLE ONLY",
+ fmtId(pubinfo->dobj.name));
+ description = "PUBLICATION TABLE";
+ }
+
appendPQExpBuffer(query, " %s",
fmtQualifiedDumpable(tbinfo));
+
if (pubrinfo->pubrelqual)
{
/*
@@ -4275,7 +4322,7 @@ dumpPublicationTable(Archive *fout, const PublicationRelInfo *pubrinfo)
ARCHIVE_OPTS(.tag = tag,
.namespace = tbinfo->dobj.namespace->dobj.name,
.owner = pubinfo->rolname,
- .description = "PUBLICATION TABLE",
+ .description = description,
.section = SECTION_POST_DATA,
.createStmt = query->data));
diff --git a/src/bin/pg_dump/pg_dump.h b/src/bin/pg_dump/pg_dump.h
index 772dc0cf7a2..893725d121f 100644
--- a/src/bin/pg_dump/pg_dump.h
+++ b/src/bin/pg_dump/pg_dump.h
@@ -615,10 +615,12 @@ typedef struct _PublicationInfo
DumpableObject dobj;
const char *rolname;
bool puballtables;
+ bool puballsequences;
bool pubinsert;
bool pubupdate;
bool pubdelete;
bool pubtruncate;
+ bool pubsequence;
bool pubviaroot;
} PublicationInfo;
@@ -643,6 +645,7 @@ typedef struct _PublicationSchemaInfo
DumpableObject dobj;
PublicationInfo *publication;
NamespaceInfo *pubschema;
+ char pubtype;
} PublicationSchemaInfo;
/*
diff --git a/src/bin/pg_dump/t/002_pg_dump.pl b/src/bin/pg_dump/t/002_pg_dump.pl
index fd1052e5db8..19f908f6006 100644
--- a/src/bin/pg_dump/t/002_pg_dump.pl
+++ b/src/bin/pg_dump/t/002_pg_dump.pl
@@ -2358,7 +2358,7 @@ my %tests = (
create_order => 50,
create_sql => 'CREATE PUBLICATION pub1;',
regexp => qr/^
- \QCREATE PUBLICATION pub1 WITH (publish = 'insert, update, delete, truncate');\E
+ \QCREATE PUBLICATION pub1 WITH (publish = 'insert, update, delete, truncate, sequence');\E
/xm,
like => { %full_runs, section_post_data => 1, },
},
@@ -2378,16 +2378,27 @@ my %tests = (
create_order => 50,
create_sql => 'CREATE PUBLICATION pub3;',
regexp => qr/^
- \QCREATE PUBLICATION pub3 WITH (publish = 'insert, update, delete, truncate');\E
+ \QCREATE PUBLICATION pub3 WITH (publish = 'insert, update, delete, truncate, sequence');\E
/xm,
like => { %full_runs, section_post_data => 1, },
},
'CREATE PUBLICATION pub4' => {
create_order => 50,
- create_sql => 'CREATE PUBLICATION pub4;',
+ create_sql => 'CREATE PUBLICATION pub4
+ FOR ALL SEQUENCES
+ WITH (publish = \'\');',
regexp => qr/^
- \QCREATE PUBLICATION pub4 WITH (publish = 'insert, update, delete, truncate');\E
+ \QCREATE PUBLICATION pub4 FOR ALL SEQUENCES WITH (publish = '');\E
+ /xm,
+ like => { %full_runs, section_post_data => 1, },
+ },
+
+ 'CREATE PUBLICATION pub5' => {
+ create_order => 50,
+ create_sql => 'CREATE PUBLICATION pub5;',
+ regexp => qr/^
+ \QCREATE PUBLICATION pub5 WITH (publish = 'insert, update, delete, truncate, sequence');\E
/xm,
like => { %full_runs, section_post_data => 1, },
},
@@ -2474,6 +2485,27 @@ my %tests = (
unlike => { exclude_dump_test_schema => 1, },
},
+ 'ALTER PUBLICATION pub3 ADD ALL SEQUENCES IN SCHEMA dump_test' => {
+ create_order => 51,
+ create_sql =>
+ 'ALTER PUBLICATION pub3 ADD ALL SEQUENCES IN SCHEMA dump_test;',
+ regexp => qr/^
+ \QALTER PUBLICATION pub3 ADD ALL SEQUENCES IN SCHEMA dump_test;\E
+ /xm,
+ like => { %full_runs, section_post_data => 1, },
+ unlike => { exclude_dump_test_schema => 1, },
+ },
+
+ 'ALTER PUBLICATION pub3 ADD ALL SEQUENCES IN SCHEMA public' => {
+ create_order => 52,
+ create_sql =>
+ 'ALTER PUBLICATION pub3 ADD ALL SEQUENCES IN SCHEMA public;',
+ regexp => qr/^
+ \QALTER PUBLICATION pub3 ADD ALL SEQUENCES IN SCHEMA public;\E
+ /xm,
+ like => { %full_runs, section_post_data => 1, },
+ },
+
'CREATE SCHEMA public' => {
regexp => qr/^CREATE SCHEMA public;/m,
diff --git a/src/bin/psql/describe.c b/src/bin/psql/describe.c
index 714097cad1b..b8a532a45f7 100644
--- a/src/bin/psql/describe.c
+++ b/src/bin/psql/describe.c
@@ -1633,28 +1633,19 @@ describeOneTableDetails(const char *schemaname,
if (tableinfo.relkind == RELKIND_SEQUENCE)
{
PGresult *result = NULL;
- printQueryOpt myopt = pset.popt;
- char *footers[2] = {NULL, NULL};
if (pset.sversion >= 100000)
{
printfPQExpBuffer(&buf,
- "SELECT pg_catalog.format_type(seqtypid, NULL) AS \"%s\",\n"
- " seqstart AS \"%s\",\n"
- " seqmin AS \"%s\",\n"
- " seqmax AS \"%s\",\n"
- " seqincrement AS \"%s\",\n"
- " CASE WHEN seqcycle THEN '%s' ELSE '%s' END AS \"%s\",\n"
- " seqcache AS \"%s\"\n",
- gettext_noop("Type"),
- gettext_noop("Start"),
- gettext_noop("Minimum"),
- gettext_noop("Maximum"),
- gettext_noop("Increment"),
+ "SELECT pg_catalog.format_type(seqtypid, NULL),\n"
+ " seqstart,\n"
+ " seqmin,\n"
+ " seqmax,\n"
+ " seqincrement,\n"
+ " CASE WHEN seqcycle THEN '%s' ELSE '%s' END,\n"
+ " seqcache\n",
gettext_noop("yes"),
- gettext_noop("no"),
- gettext_noop("Cycles?"),
- gettext_noop("Cache"));
+ gettext_noop("no"));
appendPQExpBuffer(&buf,
"FROM pg_catalog.pg_sequence\n"
"WHERE seqrelid = '%s';",
@@ -1663,22 +1654,15 @@ describeOneTableDetails(const char *schemaname,
else
{
printfPQExpBuffer(&buf,
- "SELECT 'bigint' AS \"%s\",\n"
- " start_value AS \"%s\",\n"
- " min_value AS \"%s\",\n"
- " max_value AS \"%s\",\n"
- " increment_by AS \"%s\",\n"
- " CASE WHEN is_cycled THEN '%s' ELSE '%s' END AS \"%s\",\n"
- " cache_value AS \"%s\"\n",
- gettext_noop("Type"),
- gettext_noop("Start"),
- gettext_noop("Minimum"),
- gettext_noop("Maximum"),
- gettext_noop("Increment"),
+ "SELECT 'bigint',\n"
+ " start_value,\n"
+ " min_value,\n"
+ " max_value,\n"
+ " increment_by,\n"
+ " CASE WHEN is_cycled THEN '%s' ELSE '%s' END,\n"
+ " cache_value\n",
gettext_noop("yes"),
- gettext_noop("no"),
- gettext_noop("Cycles?"),
- gettext_noop("Cache"));
+ gettext_noop("no"));
appendPQExpBuffer(&buf, "FROM %s", fmtId(schemaname));
/* must be separate because fmtId isn't reentrant */
appendPQExpBuffer(&buf, ".%s;", fmtId(relationname));
@@ -1688,6 +1672,51 @@ describeOneTableDetails(const char *schemaname,
if (!res)
goto error_return;
+ numrows = PQntuples(res);
+
+ /* XXX reset to use expanded output for sequences (maybe we should
+ * keep this disabled, just like for tables?) */
+ myopt.expanded = pset.popt.topt.expanded;
+
+ printTableInit(&cont, &myopt, title.data, 7, numrows);
+ printTableInitialized = true;
+
+ printfPQExpBuffer(&title, _("Sequence \"%s.%s\""),
+ schemaname, relationname);
+
+ printTableAddHeader(&cont, gettext_noop("Type"), true, 'l');
+ printTableAddHeader(&cont, gettext_noop("Start"), true, 'r');
+ printTableAddHeader(&cont, gettext_noop("Minimum"), true, 'r');
+ printTableAddHeader(&cont, gettext_noop("Maximum"), true, 'r');
+ printTableAddHeader(&cont, gettext_noop("Increment"), true, 'r');
+ printTableAddHeader(&cont, gettext_noop("Cycles?"), true, 'l');
+ printTableAddHeader(&cont, gettext_noop("Cache"), true, 'r');
+
+ /* Generate table cells to be printed */
+ for (i = 0; i < numrows; i++)
+ {
+ /* Type */
+ printTableAddCell(&cont, PQgetvalue(res, i, 0), false, false);
+
+ /* Start */
+ printTableAddCell(&cont, PQgetvalue(res, i, 1), false, false);
+
+ /* Minimum */
+ printTableAddCell(&cont, PQgetvalue(res, i, 2), false, false);
+
+ /* Maximum */
+ printTableAddCell(&cont, PQgetvalue(res, i, 3), false, false);
+
+ /* Increment */
+ printTableAddCell(&cont, PQgetvalue(res, i, 4), false, false);
+
+ /* Cycles? */
+ printTableAddCell(&cont, PQgetvalue(res, i, 5), false, false);
+
+ /* Cache */
+ printTableAddCell(&cont, PQgetvalue(res, i, 6), false, false);
+ }
+
/* Footer information about a sequence */
/* Get the column that owns this sequence */
@@ -1721,29 +1750,63 @@ describeOneTableDetails(const char *schemaname,
switch (PQgetvalue(result, 0, 1)[0])
{
case 'a':
- footers[0] = psprintf(_("Owned by: %s"),
- PQgetvalue(result, 0, 0));
+ printTableAddFooter(&cont,
+ psprintf(_("Owned by: %s"),
+ PQgetvalue(result, 0, 0)));
break;
case 'i':
- footers[0] = psprintf(_("Sequence for identity column: %s"),
- PQgetvalue(result, 0, 0));
+ printTableAddFooter(&cont,
+ psprintf(_("Sequence for identity column: %s"),
+ PQgetvalue(result, 0, 0)));
break;
}
}
PQclear(result);
- printfPQExpBuffer(&title, _("Sequence \"%s.%s\""),
- schemaname, relationname);
+ /* print any publications */
+ if (pset.sversion >= 150000)
+ {
+ int tuples = 0;
- myopt.footers = footers;
- myopt.topt.default_footer = false;
- myopt.title = title.data;
- myopt.translate_header = true;
+ printfPQExpBuffer(&buf,
+ "SELECT pubname\n"
+ "FROM pg_catalog.pg_publication p\n"
+ " JOIN pg_catalog.pg_publication_namespace pn ON p.oid = pn.pnpubid\n"
+ " JOIN pg_catalog.pg_class pc ON pc.relnamespace = pn.pnnspid\n"
+ "WHERE pc.oid ='%s' and pn.pntype = 's' and pg_catalog.pg_relation_is_publishable('%s')\n"
+ "UNION\n"
+ "SELECT pubname\n"
+ "FROM pg_catalog.pg_publication p\n"
+ " JOIN pg_catalog.pg_publication_rel pr ON p.oid = pr.prpubid\n"
+ "WHERE pr.prrelid = '%s'\n"
+ "UNION\n"
+ "SELECT pubname\n"
+ "FROM pg_catalog.pg_publication p\n"
+ "WHERE p.puballsequences AND pg_catalog.pg_relation_is_publishable('%s')\n"
+ "ORDER BY 1;",
+ oid, oid, oid, oid);
- printQuery(res, &myopt, pset.queryFout, false, pset.logfile);
+ result = PSQLexec(buf.data);
+ if (!result)
+ goto error_return;
+ else
+ tuples = PQntuples(result);
+
+ if (tuples > 0)
+ printTableAddFooter(&cont, _("Publications:"));
+
+ /* Might be an empty set - that's ok */
+ for (i = 0; i < tuples; i++)
+ {
+ printfPQExpBuffer(&buf, " \"%s\"",
+ PQgetvalue(result, i, 0));
+
+ printTableAddFooter(&cont, buf.data);
+ }
+ PQclear(result);
+ }
- if (footers[0])
- free(footers[0]);
+ printTable(&cont, pset.queryFout, false, pset.logfile);
retval = true;
goto error_return; /* not an error, just return early */
@@ -1970,6 +2033,11 @@ describeOneTableDetails(const char *schemaname,
for (i = 0; i < cols; i++)
printTableAddHeader(&cont, headers[i], true, 'l');
+ res = PSQLexec(buf.data);
+ if (!res)
+ goto error_return;
+ numrows = PQntuples(res);
+
/* Generate table cells to be printed */
for (i = 0; i < numrows; i++)
{
@@ -2895,7 +2963,7 @@ describeOneTableDetails(const char *schemaname,
"FROM pg_catalog.pg_publication p\n"
" JOIN pg_catalog.pg_publication_namespace pn ON p.oid = pn.pnpubid\n"
" JOIN pg_catalog.pg_class pc ON pc.relnamespace = pn.pnnspid\n"
- "WHERE pc.oid ='%s' and pg_catalog.pg_relation_is_publishable('%s')\n"
+ "WHERE pc.oid ='%s' and pn.pntype = 't' and pg_catalog.pg_relation_is_publishable('%s')\n"
"UNION\n"
"SELECT pubname\n"
" , pg_get_expr(pr.prqual, c.oid)\n"
@@ -4785,7 +4853,7 @@ listSchemas(const char *pattern, bool verbose, bool showSystem)
int i;
printfPQExpBuffer(&buf,
- "SELECT pubname \n"
+ "SELECT pubname, (CASE WHEN pntype = 't' THEN 'tables' ELSE 'sequences' END) AS pubtype\n"
"FROM pg_catalog.pg_publication p\n"
" JOIN pg_catalog.pg_publication_namespace pn ON p.oid = pn.pnpubid\n"
" JOIN pg_catalog.pg_namespace n ON n.oid = pn.pnnspid \n"
@@ -4814,8 +4882,9 @@ listSchemas(const char *pattern, bool verbose, bool showSystem)
/* Might be an empty set - that's ok */
for (i = 0; i < pub_schema_tuples; i++)
{
- printfPQExpBuffer(&buf, " \"%s\"",
- PQgetvalue(result, i, 0));
+ printfPQExpBuffer(&buf, " \"%s\" (%s)",
+ PQgetvalue(result, i, 0),
+ PQgetvalue(result, i, 1));
footers[i + 1] = pg_strdup(buf.data);
}
@@ -5820,7 +5889,7 @@ listPublications(const char *pattern)
PQExpBufferData buf;
PGresult *res;
printQueryOpt myopt = pset.popt;
- static const bool translate_columns[] = {false, false, false, false, false, false, false, false};
+ static const bool translate_columns[] = {false, false, false, false, false, false, false, false, false, false};
if (pset.sversion < 100000)
{
@@ -5834,23 +5903,45 @@ listPublications(const char *pattern)
initPQExpBuffer(&buf);
- printfPQExpBuffer(&buf,
- "SELECT pubname AS \"%s\",\n"
- " pg_catalog.pg_get_userbyid(pubowner) AS \"%s\",\n"
- " puballtables AS \"%s\",\n"
- " pubinsert AS \"%s\",\n"
- " pubupdate AS \"%s\",\n"
- " pubdelete AS \"%s\"",
- gettext_noop("Name"),
- gettext_noop("Owner"),
- gettext_noop("All tables"),
- gettext_noop("Inserts"),
- gettext_noop("Updates"),
- gettext_noop("Deletes"));
+ if (pset.sversion >= 150000)
+ printfPQExpBuffer(&buf,
+ "SELECT pubname AS \"%s\",\n"
+ " pg_catalog.pg_get_userbyid(pubowner) AS \"%s\",\n"
+ " puballtables AS \"%s\",\n"
+ " puballsequences AS \"%s\",\n"
+ " pubinsert AS \"%s\",\n"
+ " pubupdate AS \"%s\",\n"
+ " pubdelete AS \"%s\"",
+ gettext_noop("Name"),
+ gettext_noop("Owner"),
+ gettext_noop("All tables"),
+ gettext_noop("All sequences"),
+ gettext_noop("Inserts"),
+ gettext_noop("Updates"),
+ gettext_noop("Deletes"));
+ else
+ printfPQExpBuffer(&buf,
+ "SELECT pubname AS \"%s\",\n"
+ " pg_catalog.pg_get_userbyid(pubowner) AS \"%s\",\n"
+ " puballtables AS \"%s\",\n"
+ " pubinsert AS \"%s\",\n"
+ " pubupdate AS \"%s\",\n"
+ " pubdelete AS \"%s\"",
+ gettext_noop("Name"),
+ gettext_noop("Owner"),
+ gettext_noop("All tables"),
+ gettext_noop("Inserts"),
+ gettext_noop("Updates"),
+ gettext_noop("Deletes"));
+
if (pset.sversion >= 110000)
appendPQExpBuffer(&buf,
",\n pubtruncate AS \"%s\"",
gettext_noop("Truncates"));
+ if (pset.sversion >= 150000)
+ appendPQExpBuffer(&buf,
+ ",\n pubsequence AS \"%s\"",
+ gettext_noop("Sequences"));
if (pset.sversion >= 130000)
appendPQExpBuffer(&buf,
",\n pubviaroot AS \"%s\"",
@@ -5936,6 +6027,7 @@ describePublications(const char *pattern)
PGresult *res;
bool has_pubtruncate;
bool has_pubviaroot;
+ bool has_pubsequence;
PQExpBufferData title;
printTableContent cont;
@@ -5952,6 +6044,7 @@ describePublications(const char *pattern)
has_pubtruncate = (pset.sversion >= 110000);
has_pubviaroot = (pset.sversion >= 130000);
+ has_pubsequence = (pset.sversion >= 150000);
initPQExpBuffer(&buf);
@@ -5959,12 +6052,17 @@ describePublications(const char *pattern)
"SELECT oid, pubname,\n"
" pg_catalog.pg_get_userbyid(pubowner) AS owner,\n"
" puballtables, pubinsert, pubupdate, pubdelete");
+
if (has_pubtruncate)
appendPQExpBufferStr(&buf,
", pubtruncate");
if (has_pubviaroot)
appendPQExpBufferStr(&buf,
", pubviaroot");
+ if (has_pubsequence)
+ appendPQExpBufferStr(&buf,
+ ", puballsequences, pubsequence");
+
appendPQExpBufferStr(&buf,
"\nFROM pg_catalog.pg_publication\n");
@@ -6005,6 +6103,7 @@ describePublications(const char *pattern)
char *pubid = PQgetvalue(res, i, 0);
char *pubname = PQgetvalue(res, i, 1);
bool puballtables = strcmp(PQgetvalue(res, i, 3), "t") == 0;
+ bool puballsequences = strcmp(PQgetvalue(res, i, 9), "t") == 0;
printTableOpt myopt = pset.popt.topt;
if (has_pubtruncate)
@@ -6012,29 +6111,43 @@ describePublications(const char *pattern)
if (has_pubviaroot)
ncols++;
+ /* sequences have two extra columns (puballsequences, pubsequences) */
+ if (has_pubsequence)
+ ncols += 2;
+
initPQExpBuffer(&title);
printfPQExpBuffer(&title, _("Publication %s"), pubname);
printTableInit(&cont, &myopt, title.data, ncols, nrows);
printTableAddHeader(&cont, gettext_noop("Owner"), true, align);
printTableAddHeader(&cont, gettext_noop("All tables"), true, align);
+ if (has_pubsequence)
+ printTableAddHeader(&cont, gettext_noop("All sequences"), true, align);
printTableAddHeader(&cont, gettext_noop("Inserts"), true, align);
printTableAddHeader(&cont, gettext_noop("Updates"), true, align);
printTableAddHeader(&cont, gettext_noop("Deletes"), true, align);
if (has_pubtruncate)
printTableAddHeader(&cont, gettext_noop("Truncates"), true, align);
+ if (has_pubsequence)
+ printTableAddHeader(&cont, gettext_noop("Sequences"), true, align);
if (has_pubviaroot)
printTableAddHeader(&cont, gettext_noop("Via root"), true, align);
- printTableAddCell(&cont, PQgetvalue(res, i, 2), false, false);
- printTableAddCell(&cont, PQgetvalue(res, i, 3), false, false);
- printTableAddCell(&cont, PQgetvalue(res, i, 4), false, false);
- printTableAddCell(&cont, PQgetvalue(res, i, 5), false, false);
- printTableAddCell(&cont, PQgetvalue(res, i, 6), false, false);
+ printTableAddCell(&cont, PQgetvalue(res, i, 2), false, false); /* owner */
+ printTableAddCell(&cont, PQgetvalue(res, i, 3), false, false); /* all tables */
+
+ if (has_pubsequence)
+ printTableAddCell(&cont, PQgetvalue(res, i, 9), false, false); /* all sequences */
+
+ printTableAddCell(&cont, PQgetvalue(res, i, 4), false, false); /* insert */
+ printTableAddCell(&cont, PQgetvalue(res, i, 5), false, false); /* update */
+ printTableAddCell(&cont, PQgetvalue(res, i, 6), false, false); /* delete */
if (has_pubtruncate)
- printTableAddCell(&cont, PQgetvalue(res, i, 7), false, false);
+ printTableAddCell(&cont, PQgetvalue(res, i, 7), false, false); /* truncate */
+ if (has_pubsequence)
+ printTableAddCell(&cont, PQgetvalue(res, i, 10), false, false); /* sequence */
if (has_pubviaroot)
- printTableAddCell(&cont, PQgetvalue(res, i, 8), false, false);
+ printTableAddCell(&cont, PQgetvalue(res, i, 8), false, false); /* via root */
if (!puballtables)
{
@@ -6054,6 +6167,7 @@ describePublications(const char *pattern)
"WHERE c.relnamespace = n.oid\n"
" AND c.oid = pr.prrelid\n"
" AND pr.prpubid = '%s'\n"
+ " AND c.relkind != 'S'\n" /* exclude sequences */
"ORDER BY 1,2", pubid);
if (!addFooterToPublicationDesc(&buf, "Tables:", false, &cont))
goto error_return;
@@ -6065,7 +6179,7 @@ describePublications(const char *pattern)
"SELECT n.nspname\n"
"FROM pg_catalog.pg_namespace n\n"
" JOIN pg_catalog.pg_publication_namespace pn ON n.oid = pn.pnnspid\n"
- "WHERE pn.pnpubid = '%s'\n"
+ "WHERE pn.pnpubid = '%s' AND pn.pntype = 't'\n"
"ORDER BY 1", pubid);
if (!addFooterToPublicationDesc(&buf, "Tables from schemas:",
true, &cont))
@@ -6073,6 +6187,37 @@ describePublications(const char *pattern)
}
}
+ if (!puballsequences)
+ {
+ /* Get the tables for the specified publication */
+ printfPQExpBuffer(&buf,
+ "SELECT n.nspname, c.relname, NULL\n"
+ "FROM pg_catalog.pg_class c,\n"
+ " pg_catalog.pg_namespace n,\n"
+ " pg_catalog.pg_publication_rel pr\n"
+ "WHERE c.relnamespace = n.oid\n"
+ " AND c.oid = pr.prrelid\n"
+ " AND pr.prpubid = '%s'\n"
+ " AND c.relkind = 'S'\n" /* only sequences */
+ "ORDER BY 1,2", pubid);
+ if (!addFooterToPublicationDesc(&buf, "Sequences:", false, &cont))
+ goto error_return;
+
+ if (pset.sversion >= 150000)
+ {
+ /* Get the schemas for the specified publication */
+ printfPQExpBuffer(&buf,
+ "SELECT n.nspname\n"
+ "FROM pg_catalog.pg_namespace n\n"
+ " JOIN pg_catalog.pg_publication_namespace pn ON n.oid = pn.pnnspid\n"
+ "WHERE pn.pnpubid = '%s' AND pn.pntype = 's'\n"
+ "ORDER BY 1", pubid);
+ if (!addFooterToPublicationDesc(&buf, "Sequences from schemas:",
+ true, &cont))
+ goto error_return;
+ }
+ }
+
printTable(&cont, pset.queryFout, false, pset.logfile);
printTableCleanup(&cont);
diff --git a/src/bin/psql/tab-complete.c b/src/bin/psql/tab-complete.c
index 5c064595a97..e59bd8302d6 100644
--- a/src/bin/psql/tab-complete.c
+++ b/src/bin/psql/tab-complete.c
@@ -1815,11 +1815,15 @@ psql_completion(const char *text, int start, int end)
COMPLETE_WITH("ADD", "DROP", "OWNER TO", "RENAME TO", "SET");
/* ALTER PUBLICATION <name> ADD */
else if (Matches("ALTER", "PUBLICATION", MatchAny, "ADD"))
- COMPLETE_WITH("ALL TABLES IN SCHEMA", "TABLE");
+ COMPLETE_WITH("ALL TABLES IN SCHEMA", "ALL SEQUENCES IN SCHEMA", "TABLE", "SEQUENCE");
else if (Matches("ALTER", "PUBLICATION", MatchAny, "ADD|SET", "TABLE") ||
(HeadMatches("ALTER", "PUBLICATION", MatchAny, "ADD|SET", "TABLE") &&
ends_with(prev_wd, ',')))
COMPLETE_WITH_SCHEMA_QUERY(Query_for_list_of_tables);
+ else if (Matches("ALTER", "PUBLICATION", MatchAny, "ADD|SET", "SEQUENCE") ||
+ (HeadMatches("ALTER", "PUBLICATION", MatchAny, "ADD|SET", "SEQUENCE") &&
+ ends_with(prev_wd, ',')))
+ COMPLETE_WITH_SCHEMA_QUERY(Query_for_list_of_sequences);
/*
* "ALTER PUBLICATION <name> SET TABLE <name> WHERE (" - complete with
* table attributes
@@ -1838,11 +1842,11 @@ psql_completion(const char *text, int start, int end)
COMPLETE_WITH(",");
/* ALTER PUBLICATION <name> DROP */
else if (Matches("ALTER", "PUBLICATION", MatchAny, "DROP"))
- COMPLETE_WITH("ALL TABLES IN SCHEMA", "TABLE");
+ COMPLETE_WITH("ALL TABLES IN SCHEMA", "ALL SEQUENCES IN SCHEMA", "TABLE", "SEQUENCE");
/* ALTER PUBLICATION <name> SET */
else if (Matches("ALTER", "PUBLICATION", MatchAny, "SET"))
- COMPLETE_WITH("(", "ALL TABLES IN SCHEMA", "TABLE");
- else if (Matches("ALTER", "PUBLICATION", MatchAny, "ADD|DROP|SET", "ALL", "TABLES", "IN", "SCHEMA"))
+ COMPLETE_WITH("(", "ALL TABLES IN SCHEMA", "ALL SEQUENCES IN SCHEMA", "TABLE", "SEQUENCE");
+ else if (Matches("ALTER", "PUBLICATION", MatchAny, "ADD|DROP|SET", "ALL", "TABLES|SEQUENCES", "IN", "SCHEMA"))
COMPLETE_WITH_QUERY_PLUS(Query_for_list_of_schemas
" AND nspname NOT LIKE E'pg\\\\_%%'",
"CURRENT_SCHEMA");
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index d8e8715ed1c..699bd0aa3e3 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -11540,6 +11540,11 @@
provolatile => 's', prorettype => 'oid', proargtypes => 'text',
proallargtypes => '{text,oid}', proargmodes => '{i,o}',
proargnames => '{pubname,relid}', prosrc => 'pg_get_publication_tables' },
+{ oid => '8000', descr => 'get OIDs of sequences in a publication',
+ proname => 'pg_get_publication_sequences', prorows => '1000', proretset => 't',
+ provolatile => 's', prorettype => 'oid', proargtypes => 'text',
+ proallargtypes => '{text,oid}', proargmodes => '{i,o}',
+ proargnames => '{pubname,relid}', prosrc => 'pg_get_publication_sequences' },
{ oid => '6121',
descr => 'returns whether a relation can be part of a publication',
proname => 'pg_relation_is_publishable', provolatile => 's',
diff --git a/src/include/catalog/pg_publication.h b/src/include/catalog/pg_publication.h
index fe773cf9b7d..97f26208e1d 100644
--- a/src/include/catalog/pg_publication.h
+++ b/src/include/catalog/pg_publication.h
@@ -40,6 +40,12 @@ CATALOG(pg_publication,6104,PublicationRelationId)
*/
bool puballtables;
+ /*
+ * indicates that this is special publication which should encompass all
+ * sequences in the database (except for the unlogged and temp ones)
+ */
+ bool puballsequences;
+
/* true if inserts are published */
bool pubinsert;
@@ -52,6 +58,9 @@ CATALOG(pg_publication,6104,PublicationRelationId)
/* true if truncates are published */
bool pubtruncate;
+ /* true if sequences are published */
+ bool pubsequence;
+
/* true if partition changes are published using root schema */
bool pubviaroot;
} FormData_pg_publication;
@@ -72,6 +81,7 @@ typedef struct PublicationActions
bool pubupdate;
bool pubdelete;
bool pubtruncate;
+ bool pubsequence;
} PublicationActions;
typedef struct PublicationDesc
@@ -92,6 +102,7 @@ typedef struct Publication
Oid oid;
char *name;
bool alltables;
+ bool allsequences;
bool pubviaroot;
PublicationActions pubactions;
} Publication;
@@ -122,14 +133,15 @@ typedef enum PublicationPartOpt
PUBLICATION_PART_ALL,
} PublicationPartOpt;
-extern List *GetPublicationRelations(Oid pubid, PublicationPartOpt pub_partopt);
+extern List *GetPublicationRelations(Oid pubid, char objectType,
+ PublicationPartOpt pub_partopt);
extern List *GetAllTablesPublications(void);
extern List *GetAllTablesPublicationRelations(bool pubviaroot);
-extern List *GetPublicationSchemas(Oid pubid);
-extern List *GetSchemaPublications(Oid schemaid);
-extern List *GetSchemaPublicationRelations(Oid schemaid,
+extern List *GetPublicationSchemas(Oid pubid, char objectType);
+extern List *GetSchemaPublications(Oid schemaid, char objectType);
+extern List *GetSchemaPublicationRelations(Oid schemaid, char objectType,
PublicationPartOpt pub_partopt);
-extern List *GetAllSchemaPublicationRelations(Oid puboid,
+extern List *GetAllSchemaPublicationRelations(Oid puboid, char objectType,
PublicationPartOpt pub_partopt);
extern List *GetPubPartitionOptionRelations(List *result,
PublicationPartOpt pub_partopt,
@@ -137,11 +149,15 @@ extern List *GetPubPartitionOptionRelations(List *result,
extern Oid GetTopMostAncestorInPublication(Oid puboid, List *ancestors,
int *ancestor_level);
+extern List *GetAllSequencesPublications(void);
+extern List *GetAllSequencesPublicationRelations(void);
+
extern bool is_publishable_relation(Relation rel);
extern bool is_schema_publication(Oid pubid);
extern ObjectAddress publication_add_relation(Oid pubid, PublicationRelInfo *pri,
bool if_not_exists);
extern ObjectAddress publication_add_schema(Oid pubid, Oid schemaid,
+ char objectType,
bool if_not_exists);
extern Oid get_publication_oid(const char *pubname, bool missing_ok);
diff --git a/src/include/catalog/pg_publication_namespace.h b/src/include/catalog/pg_publication_namespace.h
index e4306da02e7..7340a1ec646 100644
--- a/src/include/catalog/pg_publication_namespace.h
+++ b/src/include/catalog/pg_publication_namespace.h
@@ -32,6 +32,7 @@ CATALOG(pg_publication_namespace,8901,PublicationNamespaceRelationId)
Oid oid; /* oid */
Oid pnpubid BKI_LOOKUP(pg_publication); /* Oid of the publication */
Oid pnnspid BKI_LOOKUP(pg_namespace); /* Oid of the schema */
+ char pntype; /* object type to include */
} FormData_pg_publication_namespace;
/* ----------------
@@ -42,6 +43,13 @@ CATALOG(pg_publication_namespace,8901,PublicationNamespaceRelationId)
typedef FormData_pg_publication_namespace *Form_pg_publication_namespace;
DECLARE_UNIQUE_INDEX_PKEY(pg_publication_namespace_oid_index, 8902, PublicationNamespaceObjectIndexId, on pg_publication_namespace using btree(oid oid_ops));
-DECLARE_UNIQUE_INDEX(pg_publication_namespace_pnnspid_pnpubid_index, 8903, PublicationNamespacePnnspidPnpubidIndexId, on pg_publication_namespace using btree(pnnspid oid_ops, pnpubid oid_ops));
+DECLARE_UNIQUE_INDEX(pg_publication_namespace_pnnspid_pnpubid_pntype_index, 8903, PublicationNamespacePnnspidPnpubidPntypeIndexId, on pg_publication_namespace using btree(pnnspid oid_ops, pnpubid oid_ops, pntype char_ops));
+
+/* object type to include from a schema, maps to relkind */
+#define PUB_OBJTYPE_TABLE 't' /* table (regular or partitioned) */
+#define PUB_OBJTYPE_SEQUENCE 's' /* sequence object */
+#define PUB_OBJTYPE_UNSUPPORTED 'u' /* used for non-replicated types */
+
+extern char pub_get_object_type_for_relkind(char relkind);
#endif /* PG_PUBLICATION_NAMESPACE_H */
diff --git a/src/include/commands/sequence.h b/src/include/commands/sequence.h
index 9fecc41954e..5bab90db8e0 100644
--- a/src/include/commands/sequence.h
+++ b/src/include/commands/sequence.h
@@ -60,6 +60,7 @@ extern ObjectAddress DefineSequence(ParseState *pstate, CreateSeqStmt *stmt);
extern ObjectAddress AlterSequence(ParseState *pstate, AlterSeqStmt *stmt);
extern void DeleteSequenceTuple(Oid relid);
extern void ResetSequence(Oid seq_relid);
+extern void SetSequence(Oid seq_relid, bool transactional, int64 last_value, int64 log_cnt, bool is_called);
extern void ResetSequenceCaches(void);
extern void seq_redo(XLogReaderState *rptr);
diff --git a/src/include/nodes/parsenodes.h b/src/include/nodes/parsenodes.h
index 2f618cb2292..cb1fcc0ee31 100644
--- a/src/include/nodes/parsenodes.h
+++ b/src/include/nodes/parsenodes.h
@@ -3665,6 +3665,10 @@ typedef enum PublicationObjSpecType
PUBLICATIONOBJ_TABLES_IN_SCHEMA, /* All tables in schema */
PUBLICATIONOBJ_TABLES_IN_CUR_SCHEMA, /* All tables in first element of
* search_path */
+ PUBLICATIONOBJ_SEQUENCE, /* Sequence type */
+ PUBLICATIONOBJ_SEQUENCES_IN_SCHEMA, /* Sequences in schema type */
+ PUBLICATIONOBJ_SEQUENCES_IN_CUR_SCHEMA, /* Get the first element of
+ * search_path */
PUBLICATIONOBJ_CONTINUATION /* Continuation of previous type */
} PublicationObjSpecType;
@@ -3683,7 +3687,7 @@ typedef struct CreatePublicationStmt
char *pubname; /* Name of the publication */
List *options; /* List of DefElem nodes */
List *pubobjects; /* Optional list of publication objects */
- bool for_all_tables; /* Special publication for all tables in db */
+ List *for_all_objects; /* Special publication for all objects in db */
} CreatePublicationStmt;
typedef enum AlterPublicationAction
@@ -3706,7 +3710,7 @@ typedef struct AlterPublicationStmt
* objects.
*/
List *pubobjects; /* Optional list of publication objects */
- bool for_all_tables; /* Special publication for all tables in db */
+ List *for_all_objects; /* Special publication for all objects in db */
AlterPublicationAction action; /* What action to perform with the given
* objects */
} AlterPublicationStmt;
diff --git a/src/include/replication/logicalproto.h b/src/include/replication/logicalproto.h
index 4d2c881644a..415757d8a2d 100644
--- a/src/include/replication/logicalproto.h
+++ b/src/include/replication/logicalproto.h
@@ -61,6 +61,7 @@ typedef enum LogicalRepMsgType
LOGICAL_REP_MSG_RELATION = 'R',
LOGICAL_REP_MSG_TYPE = 'Y',
LOGICAL_REP_MSG_MESSAGE = 'M',
+ LOGICAL_REP_MSG_SEQUENCE = 'X', /* FIXME change */
LOGICAL_REP_MSG_BEGIN_PREPARE = 'b',
LOGICAL_REP_MSG_PREPARE = 'P',
LOGICAL_REP_MSG_COMMIT_PREPARED = 'K',
@@ -118,6 +119,18 @@ typedef struct LogicalRepTyp
char *typname; /* name of the remote type */
} LogicalRepTyp;
+/* Sequence info */
+typedef struct LogicalRepSequence
+{
+ Oid remoteid; /* unique id of the remote sequence */
+ char *nspname; /* schema name of remote sequence */
+ char *seqname; /* name of the remote sequence */
+ bool transactional;
+ int64 last_value;
+ int64 log_cnt;
+ bool is_called;
+} LogicalRepSequence;
+
/* Transaction info */
typedef struct LogicalRepBeginData
{
@@ -230,6 +243,12 @@ extern List *logicalrep_read_truncate(StringInfo in,
bool *cascade, bool *restart_seqs);
extern void logicalrep_write_message(StringInfo out, TransactionId xid, XLogRecPtr lsn,
bool transactional, const char *prefix, Size sz, const char *message);
+extern void logicalrep_write_sequence(StringInfo out, Relation rel,
+ TransactionId xid, XLogRecPtr lsn,
+ bool transactional,
+ int64 last_value, int64 log_cnt,
+ bool is_called);
+extern void logicalrep_read_sequence(StringInfo in, LogicalRepSequence *seqdata);
extern void logicalrep_write_rel(StringInfo out, TransactionId xid,
Relation rel);
extern LogicalRepRelation *logicalrep_read_rel(StringInfo in);
diff --git a/src/include/replication/pgoutput.h b/src/include/replication/pgoutput.h
index eafedd610a5..f4e9f35d09d 100644
--- a/src/include/replication/pgoutput.h
+++ b/src/include/replication/pgoutput.h
@@ -29,6 +29,7 @@ typedef struct PGOutputData
bool streaming;
bool messages;
bool two_phase;
+ bool sequences;
} PGOutputData;
#endif /* PGOUTPUT_H */
diff --git a/src/test/regress/expected/object_address.out b/src/test/regress/expected/object_address.out
index 4117fc27c9a..c95d44b3db9 100644
--- a/src/test/regress/expected/object_address.out
+++ b/src/test/regress/expected/object_address.out
@@ -46,6 +46,7 @@ CREATE TRANSFORM FOR int LANGUAGE SQL (
SET client_min_messages = 'ERROR';
CREATE PUBLICATION addr_pub FOR TABLE addr_nsp.gentable;
CREATE PUBLICATION addr_pub_schema FOR ALL TABLES IN SCHEMA addr_nsp;
+CREATE PUBLICATION addr_pub_schema2 FOR ALL SEQUENCES IN SCHEMA addr_nsp;
RESET client_min_messages;
CREATE SUBSCRIPTION regress_addr_sub CONNECTION '' PUBLICATION bar WITH (connect = false, slot_name = NONE);
WARNING: tables were not subscribed, you will have to run ALTER SUBSCRIPTION ... REFRESH PUBLICATION to subscribe the tables
@@ -428,7 +429,8 @@ WITH objects (type, name, args) AS (VALUES
('transform', '{int}', '{sql}'),
('access method', '{btree}', '{}'),
('publication', '{addr_pub}', '{}'),
- ('publication namespace', '{addr_nsp}', '{addr_pub_schema}'),
+ ('publication namespace', '{addr_nsp}', '{addr_pub_schema,t}'),
+ ('publication namespace', '{addr_nsp}', '{addr_pub_schema2,s}'),
('publication relation', '{addr_nsp, gentable}', '{addr_pub}'),
('subscription', '{regress_addr_sub}', '{}'),
('statistics object', '{addr_nsp, gentable_stat}', '{}')
@@ -492,8 +494,9 @@ SELECT (pg_identify_object(addr1.classid, addr1.objid, addr1.objsubid)).*,
subscription | | regress_addr_sub | regress_addr_sub | t
publication | | addr_pub | addr_pub | t
publication relation | | | addr_nsp.gentable in publication addr_pub | t
- publication namespace | | | addr_nsp in publication addr_pub_schema | t
-(50 rows)
+ publication namespace | | | addr_nsp in publication addr_pub_schema type t | t
+ publication namespace | | | addr_nsp in publication addr_pub_schema2 type s | t
+(51 rows)
---
--- Cleanup resources
@@ -506,6 +509,7 @@ drop cascades to server integer
drop cascades to user mapping for regress_addr_user on server integer
DROP PUBLICATION addr_pub;
DROP PUBLICATION addr_pub_schema;
+DROP PUBLICATION addr_pub_schema2;
DROP SUBSCRIPTION regress_addr_sub;
DROP SCHEMA addr_nsp CASCADE;
NOTICE: drop cascades to 14 other objects
diff --git a/src/test/regress/expected/publication.out b/src/test/regress/expected/publication.out
index 4e191c120ac..58214ee742e 100644
--- a/src/test/regress/expected/publication.out
+++ b/src/test/regress/expected/publication.out
@@ -30,20 +30,20 @@ ERROR: conflicting or redundant options
LINE 1: ...ub_xxx WITH (publish_via_partition_root = 'true', publish_vi...
^
\dRp
- List of publications
- Name | Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------+--------------------------+------------+---------+---------+---------+-----------+----------
- testpib_ins_trunct | regress_publication_user | f | t | f | f | f | f
- testpub_default | regress_publication_user | f | f | t | f | f | f
+ List of publications
+ Name | Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ testpib_ins_trunct | regress_publication_user | f | f | t | f | f | f | f | f
+ testpub_default | regress_publication_user | f | f | f | t | f | f | f | f
(2 rows)
-ALTER PUBLICATION testpub_default SET (publish = 'insert, update, delete');
+ALTER PUBLICATION testpub_default SET (publish = 'insert, update, delete, sequence');
\dRp
- List of publications
- Name | Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------+--------------------------+------------+---------+---------+---------+-----------+----------
- testpib_ins_trunct | regress_publication_user | f | t | f | f | f | f
- testpub_default | regress_publication_user | f | t | t | t | f | f
+ List of publications
+ Name | Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ testpib_ins_trunct | regress_publication_user | f | f | t | f | f | f | f | f
+ testpub_default | regress_publication_user | f | f | t | t | t | f | t | f
(2 rows)
--- adding tables
@@ -61,6 +61,9 @@ CREATE TABLE testpub_tbl2 (id serial primary key, data text);
ALTER PUBLICATION testpub_foralltables ADD TABLE testpub_tbl2;
ERROR: publication "testpub_foralltables" is defined as FOR ALL TABLES
DETAIL: Tables cannot be added to or dropped from FOR ALL TABLES publications.
+-- fail - can't add a table using ADD SEQUENCE command
+ALTER PUBLICATION testpub_foralltables ADD SEQUENCE testpub_tbl2;
+ERROR: object type mismatch
-- fail - can't drop from all tables publication
ALTER PUBLICATION testpub_foralltables DROP TABLE testpub_tbl2;
ERROR: publication "testpub_foralltables" is defined as FOR ALL TABLES
@@ -87,10 +90,10 @@ RESET client_min_messages;
-- should be able to add schema to 'FOR TABLE' publication
ALTER PUBLICATION testpub_fortable ADD ALL TABLES IN SCHEMA pub_test;
\dRp+ testpub_fortable
- Publication testpub_fortable
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_fortable
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"public.testpub_tbl1"
Tables from schemas:
@@ -99,20 +102,20 @@ Tables from schemas:
-- should be able to drop schema from 'FOR TABLE' publication
ALTER PUBLICATION testpub_fortable DROP ALL TABLES IN SCHEMA pub_test;
\dRp+ testpub_fortable
- Publication testpub_fortable
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_fortable
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"public.testpub_tbl1"
-- should be able to set schema to 'FOR TABLE' publication
ALTER PUBLICATION testpub_fortable SET ALL TABLES IN SCHEMA pub_test;
\dRp+ testpub_fortable
- Publication testpub_fortable
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_fortable
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test"
@@ -134,10 +137,10 @@ ERROR: relation "testpub_nopk" is not part of the publication
-- should be able to set table to schema publication
ALTER PUBLICATION testpub_forschema SET TABLE pub_test.testpub_nopk;
\dRp+ testpub_forschema
- Publication testpub_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"pub_test.testpub_nopk"
@@ -159,10 +162,10 @@ Publications:
"testpub_foralltables"
\dRp+ testpub_foralltables
- Publication testpub_foralltables
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | t | t | t | f | f | f
+ Publication testpub_foralltables
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | t | f | t | t | f | f | f | f
(1 row)
DROP TABLE testpub_tbl2;
@@ -174,24 +177,527 @@ CREATE PUBLICATION testpub3 FOR TABLE testpub_tbl3;
CREATE PUBLICATION testpub4 FOR TABLE ONLY testpub_tbl3;
RESET client_min_messages;
\dRp+ testpub3
- Publication testpub3
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub3
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"public.testpub_tbl3"
"public.testpub_tbl3a"
\dRp+ testpub4
- Publication testpub4
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub4
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"public.testpub_tbl3"
DROP TABLE testpub_tbl3, testpub_tbl3a;
DROP PUBLICATION testpub3, testpub4;
+--- adding sequences
+CREATE SEQUENCE testpub_seq0;
+CREATE SEQUENCE pub_test.testpub_seq1;
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_forallsequences FOR ALL SEQUENCES WITH (publish = 'sequence');
+RESET client_min_messages;
+ALTER PUBLICATION testpub_forallsequences SET (publish = 'insert, sequence');
+CREATE SEQUENCE testpub_seq2;
+-- fail - can't add to for all sequences publication
+ALTER PUBLICATION testpub_forallsequences ADD SEQUENCE testpub_seq2;
+ERROR: publication "testpub_forallsequences" is defined as FOR ALL SEQUENCES
+DETAIL: Sequences cannot be added to or dropped from FOR ALL SEQUENCES publications.
+-- fail - can't drop from all sequences publication
+ALTER PUBLICATION testpub_forallsequences DROP SEQUENCE testpub_seq2;
+ERROR: publication "testpub_forallsequences" is defined as FOR ALL SEQUENCES
+DETAIL: Sequences cannot be added to or dropped from FOR ALL SEQUENCES publications.
+-- fail - can't add to for all sequences publication
+ALTER PUBLICATION testpub_forallsequences SET SEQUENCE pub_test.testpub_seq1;
+ERROR: publication "testpub_forallsequences" is defined as FOR ALL SEQUENCES
+DETAIL: Sequences cannot be added to or dropped from FOR ALL SEQUENCES publications.
+-- fail - can't add schema to 'FOR ALL SEQUENCES' publication
+ALTER PUBLICATION testpub_forallsequences ADD ALL SEQUENCES IN SCHEMA pub_test;
+ERROR: publication "testpub_forallsequences" is defined as FOR ALL SEQUENCES
+DETAIL: Sequences from schema cannot be added to, dropped from, or set on FOR ALL SEQUENCES publications.
+-- fail - can't drop schema from 'FOR ALL SEQUENCES' publication
+ALTER PUBLICATION testpub_forallsequences DROP ALL SEQUENCES IN SCHEMA pub_test;
+ERROR: publication "testpub_forallsequences" is defined as FOR ALL SEQUENCES
+DETAIL: Sequences from schema cannot be added to, dropped from, or set on FOR ALL SEQUENCES publications.
+-- fail - can't set schema to 'FOR ALL SEQUENCES' publication
+ALTER PUBLICATION testpub_forallsequences SET ALL SEQUENCES IN SCHEMA pub_test;
+ERROR: publication "testpub_forallsequences" is defined as FOR ALL SEQUENCES
+DETAIL: Sequences from schema cannot be added to, dropped from, or set on FOR ALL SEQUENCES publications.
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_forsequence FOR SEQUENCE testpub_seq0;
+RESET client_min_messages;
+-- should be able to add schema to 'FOR SEQUENCE' publication
+ALTER PUBLICATION testpub_forsequence ADD ALL SEQUENCES IN SCHEMA pub_test;
+\dRp+ testpub_forsequence
+ Publication testpub_forsequence
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Sequences:
+ "public.testpub_seq0"
+Sequences from schemas:
+ "pub_test"
+
+-- fail - can't add sequence from the schema we already added
+ALTER PUBLICATION testpub_forsequence ADD SEQUENCE pub_test.testpub_seq1;
+ERROR: cannot add relation "pub_test.testpub_seq1" to publication
+DETAIL: Sequence's schema "pub_test" is already part of the publication or part of the specified schema list.
+-- fail - can't add sequence using ADD TABLE command
+ALTER PUBLICATION testpub_forsequence ADD TABLE pub_test.testpub_seq1;
+ERROR: object type mismatch
+-- should be able to drop schema from 'FOR SEQUENCE' publication
+ALTER PUBLICATION testpub_forsequence DROP ALL SEQUENCES IN SCHEMA pub_test;
+\dRp+ testpub_forsequence
+ Publication testpub_forsequence
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Sequences:
+ "public.testpub_seq0"
+
+-- should be able to set schema to 'FOR SEQUENCE' publication
+ALTER PUBLICATION testpub_forsequence SET ALL SEQUENCES IN SCHEMA pub_test;
+\dRp+ testpub_forsequence
+ Publication testpub_forsequence
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Sequences from schemas:
+ "pub_test"
+
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_forschema FOR ALL SEQUENCES IN SCHEMA pub_test;
+RESET client_min_messages;
+-- fail - can't create publication with schema and sequence of the same schema
+CREATE PUBLICATION testpub_for_seq_schema FOR ALL SEQUENCES IN SCHEMA pub_test, SEQUENCE pub_test.testpub_seq1;
+ERROR: cannot add relation "pub_test.testpub_seq1" to publication
+DETAIL: Sequence's schema "pub_test" is already part of the publication or part of the specified schema list.
+-- fail - can't add a sequence of the same schema to the schema publication
+ALTER PUBLICATION testpub_forschema ADD SEQUENCE pub_test.testpub_seq1;
+ERROR: cannot add relation "pub_test.testpub_seq1" to publication
+DETAIL: Sequence's schema "pub_test" is already part of the publication or part of the specified schema list.
+-- fail - can't drop a sequence from the schema publication which isn't in the
+-- publication
+ALTER PUBLICATION testpub_forschema DROP SEQUENCE pub_test.testpub_seq1;
+ERROR: relation "testpub_seq1" is not part of the publication
+-- should be able to set sequence to schema publication
+ALTER PUBLICATION testpub_forschema SET SEQUENCE pub_test.testpub_seq1;
+\dRp+ testpub_forschema
+ Publication testpub_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Sequences:
+ "pub_test.testpub_seq1"
+
+SELECT pubname, puballtables, puballsequences FROM pg_publication WHERE pubname = 'testpub_forallsequences';
+ pubname | puballtables | puballsequences
+-------------------------+--------------+-----------------
+ testpub_forallsequences | f | t
+(1 row)
+
+\d+ pub_test.testpub_seq1
+ Sequence "pub_test.testpub_seq1"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+Publications:
+ "testpub_forallsequences"
+ "testpub_forschema"
+ "testpub_forsequence"
+
+\dRp+ testpub_forallsequences
+ Publication testpub_forallsequences
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | t | t | f | f | f | t | f
+(1 row)
+
+DROP SEQUENCE testpub_seq0, pub_test.testpub_seq1, testpub_seq2;
+DROP PUBLICATION testpub_forallsequences, testpub_forsequence, testpub_forschema;
+-- Publication mixing tables and sequences
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_mix;
+RESET client_min_messages;
+CREATE SEQUENCE testpub_seq1;
+CREATE SEQUENCE pub_test.testpub_seq2;
+ALTER PUBLICATION testpub_mix ADD SEQUENCE testpub_seq1, TABLE testpub_tbl1;
+\dRp+ testpub_mix
+ Publication testpub_mix
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Tables:
+ "public.testpub_tbl1"
+Sequences:
+ "public.testpub_seq1"
+
+ALTER PUBLICATION testpub_mix ADD ALL SEQUENCES IN SCHEMA pub_test, ALL TABLES IN SCHEMA pub_test;
+\dRp+ testpub_mix
+ Publication testpub_mix
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Tables:
+ "public.testpub_tbl1"
+Tables from schemas:
+ "pub_test"
+Sequences:
+ "public.testpub_seq1"
+Sequences from schemas:
+ "pub_test"
+
+ALTER PUBLICATION testpub_mix DROP ALL SEQUENCES IN SCHEMA pub_test;
+\dRp+ testpub_mix
+ Publication testpub_mix
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Tables:
+ "public.testpub_tbl1"
+Tables from schemas:
+ "pub_test"
+Sequences:
+ "public.testpub_seq1"
+
+ALTER PUBLICATION testpub_mix DROP ALL TABLES IN SCHEMA pub_test;
+\dRp+ testpub_mix
+ Publication testpub_mix
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Tables:
+ "public.testpub_tbl1"
+Sequences:
+ "public.testpub_seq1"
+
+DROP PUBLICATION testpub_mix;
+DROP SEQUENCE testpub_seq1;
+DROP SEQUENCE pub_test.testpub_seq2;
+-- make sure we replicate only the correct relation type
+CREATE SCHEMA pub_test1;
+CREATE SEQUENCE pub_test1.test_seq1;
+CREATE TABLE pub_test1.test_tbl1 (a int primary key, b int);
+CREATE SCHEMA pub_test2;
+CREATE SEQUENCE pub_test2.test_seq2;
+CREATE TABLE pub_test2.test_tbl2 (a int primary key, b int);
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_schemas;
+RESET client_min_messages;
+-- add tables from one schema, sequences from the other
+ALTER PUBLICATION testpub_schemas ADD ALL TABLES IN SCHEMA pub_test2;
+ALTER PUBLICATION testpub_schemas ADD ALL SEQUENCES IN SCHEMA pub_test1;
+\dRp+ testpub_schemas
+ Publication testpub_schemas
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Tables from schemas:
+ "pub_test2"
+Sequences from schemas:
+ "pub_test1"
+
+\dn+ pub_test1
+ List of schemas
+ Name | Owner | Access privileges | Description
+-----------+--------------------------+-------------------+-------------
+ pub_test1 | regress_publication_user | |
+Publications:
+ "testpub_schemas" (sequences)
+
+\dn+ pub_test2
+ List of schemas
+ Name | Owner | Access privileges | Description
+-----------+--------------------------+-------------------+-------------
+ pub_test2 | regress_publication_user | |
+Publications:
+ "testpub_schemas" (tables)
+
+\d+ pub_test1.test_seq1;
+ Sequence "pub_test1.test_seq1"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test1.test_tbl1;
+ Table "pub_test1.test_tbl1"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl1_pkey" PRIMARY KEY, btree (a)
+
+\d+ pub_test2.test_seq2;
+ Sequence "pub_test2.test_seq2"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+
+\d+ pub_test2.test_tbl2;
+ Table "pub_test2.test_tbl2"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl2_pkey" PRIMARY KEY, btree (a)
+Publications:
+ "testpub_schemas"
+
+-- add the other object type from each schema
+ALTER PUBLICATION testpub_schemas ADD ALL TABLES IN SCHEMA pub_test1;
+ALTER PUBLICATION testpub_schemas ADD ALL SEQUENCES IN SCHEMA pub_test2;
+\dRp+ testpub_schemas
+ Publication testpub_schemas
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Tables from schemas:
+ "pub_test1"
+ "pub_test2"
+Sequences from schemas:
+ "pub_test1"
+ "pub_test2"
+
+\dn+ pub_test1
+ List of schemas
+ Name | Owner | Access privileges | Description
+-----------+--------------------------+-------------------+-------------
+ pub_test1 | regress_publication_user | |
+Publications:
+ "testpub_schemas" (sequences)
+ "testpub_schemas" (tables)
+
+\dn+ pub_test2
+ List of schemas
+ Name | Owner | Access privileges | Description
+-----------+--------------------------+-------------------+-------------
+ pub_test2 | regress_publication_user | |
+Publications:
+ "testpub_schemas" (tables)
+ "testpub_schemas" (sequences)
+
+\d+ pub_test1.test_seq1;
+ Sequence "pub_test1.test_seq1"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test1.test_tbl1;
+ Table "pub_test1.test_tbl1"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl1_pkey" PRIMARY KEY, btree (a)
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test2.test_seq2;
+ Sequence "pub_test2.test_seq2"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test2.test_tbl2;
+ Table "pub_test2.test_tbl2"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl2_pkey" PRIMARY KEY, btree (a)
+Publications:
+ "testpub_schemas"
+
+-- now drop the object type added first
+ALTER PUBLICATION testpub_schemas DROP ALL TABLES IN SCHEMA pub_test2;
+ALTER PUBLICATION testpub_schemas DROP ALL SEQUENCES IN SCHEMA pub_test1;
+\dRp+ testpub_schemas
+ Publication testpub_schemas
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Tables from schemas:
+ "pub_test1"
+Sequences from schemas:
+ "pub_test2"
+
+\dn+ pub_test1
+ List of schemas
+ Name | Owner | Access privileges | Description
+-----------+--------------------------+-------------------+-------------
+ pub_test1 | regress_publication_user | |
+Publications:
+ "testpub_schemas" (tables)
+
+\dn+ pub_test2
+ List of schemas
+ Name | Owner | Access privileges | Description
+-----------+--------------------------+-------------------+-------------
+ pub_test2 | regress_publication_user | |
+Publications:
+ "testpub_schemas" (sequences)
+
+\d+ pub_test1.test_seq1;
+ Sequence "pub_test1.test_seq1"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+
+\d+ pub_test1.test_tbl1;
+ Table "pub_test1.test_tbl1"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl1_pkey" PRIMARY KEY, btree (a)
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test2.test_seq2;
+ Sequence "pub_test2.test_seq2"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test2.test_tbl2;
+ Table "pub_test2.test_tbl2"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl2_pkey" PRIMARY KEY, btree (a)
+
+-- should fail (publication contains the whole schema)
+ALTER PUBLICATION testpub_schemas ADD TABLE pub_test1.test_tbl1;
+ERROR: cannot add relation "pub_test1.test_tbl1" to publication
+DETAIL: Table's schema "pub_test1" is already part of the publication or part of the specified schema list.
+ALTER PUBLICATION testpub_schemas ADD SEQUENCE pub_test2.test_seq2;
+ERROR: cannot add relation "pub_test2.test_seq2" to publication
+DETAIL: Sequence's schema "pub_test2" is already part of the publication or part of the specified schema list.
+-- should work (different schema)
+ALTER PUBLICATION testpub_schemas ADD TABLE pub_test2.test_tbl2;
+ALTER PUBLICATION testpub_schemas ADD SEQUENCE pub_test1.test_seq1;
+\dRp+ testpub_schemas
+ Publication testpub_schemas
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Tables:
+ "pub_test2.test_tbl2"
+Tables from schemas:
+ "pub_test1"
+Sequences:
+ "pub_test1.test_seq1"
+Sequences from schemas:
+ "pub_test2"
+
+\d+ pub_test1.test_seq1;
+ Sequence "pub_test1.test_seq1"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test1.test_tbl1;
+ Table "pub_test1.test_tbl1"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl1_pkey" PRIMARY KEY, btree (a)
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test2.test_seq2;
+ Sequence "pub_test2.test_seq2"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test2.test_tbl2;
+ Table "pub_test2.test_tbl2"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl2_pkey" PRIMARY KEY, btree (a)
+Publications:
+ "testpub_schemas"
+
+-- now drop the explicitly added objects again
+ALTER PUBLICATION testpub_schemas DROP TABLE pub_test2.test_tbl2;
+ALTER PUBLICATION testpub_schemas DROP SEQUENCE pub_test1.test_seq1;
+\dRp+ testpub_schemas
+ Publication testpub_schemas
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Tables from schemas:
+ "pub_test1"
+Sequences from schemas:
+ "pub_test2"
+
+\d+ pub_test1.test_seq1;
+ Sequence "pub_test1.test_seq1"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+
+\d+ pub_test1.test_tbl1;
+ Table "pub_test1.test_tbl1"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl1_pkey" PRIMARY KEY, btree (a)
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test2.test_seq2;
+ Sequence "pub_test2.test_seq2"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test2.test_tbl2;
+ Table "pub_test2.test_tbl2"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl2_pkey" PRIMARY KEY, btree (a)
+
+DROP PUBLICATION testpub_schemas;
+DROP TABLE pub_test1.test_tbl1, pub_test2.test_tbl2;
+DROP SEQUENCE pub_test1.test_seq1, pub_test2.test_seq2;
+DROP SCHEMA pub_test1, pub_test2;
-- Tests for partitioned tables
SET client_min_messages = 'ERROR';
CREATE PUBLICATION testpub_forparted;
@@ -207,10 +713,10 @@ UPDATE testpub_parted1 SET a = 1;
-- only parent is listed as being in publication, not the partition
ALTER PUBLICATION testpub_forparted ADD TABLE testpub_parted;
\dRp+ testpub_forparted
- Publication testpub_forparted
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_forparted
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"public.testpub_parted"
@@ -223,10 +729,10 @@ ALTER TABLE testpub_parted DETACH PARTITION testpub_parted1;
UPDATE testpub_parted1 SET a = 1;
ALTER PUBLICATION testpub_forparted SET (publish_via_partition_root = true);
\dRp+ testpub_forparted
- Publication testpub_forparted
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | t
+ Publication testpub_forparted
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | t
Tables:
"public.testpub_parted"
@@ -255,10 +761,10 @@ SET client_min_messages = 'ERROR';
CREATE PUBLICATION testpub5 FOR TABLE testpub_rf_tbl1, testpub_rf_tbl2 WHERE (c <> 'test' AND d < 5) WITH (publish = 'insert');
RESET client_min_messages;
\dRp+ testpub5
- Publication testpub5
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | f | f | f | f
+ Publication testpub5
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | f | f | f | f | f
Tables:
"public.testpub_rf_tbl1"
"public.testpub_rf_tbl2" WHERE ((c <> 'test'::text) AND (d < 5))
@@ -271,10 +777,10 @@ Tables:
ALTER PUBLICATION testpub5 ADD TABLE testpub_rf_tbl3 WHERE (e > 1000 AND e < 2000);
\dRp+ testpub5
- Publication testpub5
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | f | f | f | f
+ Publication testpub5
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | f | f | f | f | f
Tables:
"public.testpub_rf_tbl1"
"public.testpub_rf_tbl2" WHERE ((c <> 'test'::text) AND (d < 5))
@@ -290,10 +796,10 @@ Publications:
ALTER PUBLICATION testpub5 DROP TABLE testpub_rf_tbl2;
\dRp+ testpub5
- Publication testpub5
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | f | f | f | f
+ Publication testpub5
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | f | f | f | f | f
Tables:
"public.testpub_rf_tbl1"
"public.testpub_rf_tbl3" WHERE ((e > 1000) AND (e < 2000))
@@ -301,10 +807,10 @@ Tables:
-- remove testpub_rf_tbl1 and add testpub_rf_tbl3 again (another WHERE expression)
ALTER PUBLICATION testpub5 SET TABLE testpub_rf_tbl3 WHERE (e > 300 AND e < 500);
\dRp+ testpub5
- Publication testpub5
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | f | f | f | f
+ Publication testpub5
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | f | f | f | f | f
Tables:
"public.testpub_rf_tbl3" WHERE ((e > 300) AND (e < 500))
@@ -337,10 +843,10 @@ SET client_min_messages = 'ERROR';
CREATE PUBLICATION testpub_syntax1 FOR TABLE testpub_rf_tbl1, ONLY testpub_rf_tbl3 WHERE (e < 999) WITH (publish = 'insert');
RESET client_min_messages;
\dRp+ testpub_syntax1
- Publication testpub_syntax1
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | f | f | f | f
+ Publication testpub_syntax1
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | f | f | f | f | f
Tables:
"public.testpub_rf_tbl1"
"public.testpub_rf_tbl3" WHERE (e < 999)
@@ -350,10 +856,10 @@ SET client_min_messages = 'ERROR';
CREATE PUBLICATION testpub_syntax2 FOR TABLE testpub_rf_tbl1, testpub_rf_schema1.testpub_rf_tbl5 WHERE (h < 999) WITH (publish = 'insert');
RESET client_min_messages;
\dRp+ testpub_syntax2
- Publication testpub_syntax2
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | f | f | f | f
+ Publication testpub_syntax2
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | f | f | f | f | f
Tables:
"public.testpub_rf_tbl1"
"testpub_rf_schema1.testpub_rf_tbl5" WHERE (h < 999)
@@ -658,10 +1164,10 @@ ERROR: relation "testpub_tbl1" is already member of publication "testpub_fortbl
CREATE PUBLICATION testpub_fortbl FOR TABLE testpub_tbl1;
ERROR: publication "testpub_fortbl" already exists
\dRp+ testpub_fortbl
- Publication testpub_fortbl
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_fortbl
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"pub_test.testpub_nopk"
"public.testpub_tbl1"
@@ -699,10 +1205,10 @@ Publications:
"testpub_fortbl"
\dRp+ testpub_default
- Publication testpub_default
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | f | f
+ Publication testpub_default
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | f | t | f
Tables:
"pub_test.testpub_nopk"
"public.testpub_tbl1"
@@ -780,10 +1286,10 @@ REVOKE CREATE ON DATABASE regression FROM regress_publication_user2;
DROP TABLE testpub_parted;
DROP TABLE testpub_tbl1;
\dRp+ testpub_default
- Publication testpub_default
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | f | f
+ Publication testpub_default
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | f | t | f
(1 row)
-- fail - must be owner of publication
@@ -793,20 +1299,20 @@ ERROR: must be owner of publication testpub_default
RESET ROLE;
ALTER PUBLICATION testpub_default RENAME TO testpub_foo;
\dRp testpub_foo
- List of publications
- Name | Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
--------------+--------------------------+------------+---------+---------+---------+-----------+----------
- testpub_foo | regress_publication_user | f | t | t | t | f | f
+ List of publications
+ Name | Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+-------------+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ testpub_foo | regress_publication_user | f | f | t | t | t | f | t | f
(1 row)
-- rename back to keep the rest simple
ALTER PUBLICATION testpub_foo RENAME TO testpub_default;
ALTER PUBLICATION testpub_default OWNER TO regress_publication_user2;
\dRp testpub_default
- List of publications
- Name | Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
------------------+---------------------------+------------+---------+---------+---------+-----------+----------
- testpub_default | regress_publication_user2 | f | t | t | t | f | f
+ List of publications
+ Name | Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+-----------------+---------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ testpub_default | regress_publication_user2 | f | f | t | t | t | f | t | f
(1 row)
-- adding schemas and tables
@@ -822,19 +1328,19 @@ CREATE TABLE "CURRENT_SCHEMA"."CURRENT_SCHEMA"(id int);
SET client_min_messages = 'ERROR';
CREATE PUBLICATION testpub1_forschema FOR ALL TABLES IN SCHEMA pub_test1;
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
CREATE PUBLICATION testpub2_forschema FOR ALL TABLES IN SCHEMA pub_test1, pub_test2, pub_test3;
\dRp+ testpub2_forschema
- Publication testpub2_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub2_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
"pub_test2"
@@ -848,44 +1354,44 @@ CREATE PUBLICATION testpub6_forschema FOR ALL TABLES IN SCHEMA "CURRENT_SCHEMA",
CREATE PUBLICATION testpub_fortable FOR TABLE "CURRENT_SCHEMA"."CURRENT_SCHEMA";
RESET client_min_messages;
\dRp+ testpub3_forschema
- Publication testpub3_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub3_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"public"
\dRp+ testpub4_forschema
- Publication testpub4_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub4_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"CURRENT_SCHEMA"
\dRp+ testpub5_forschema
- Publication testpub5_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub5_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"CURRENT_SCHEMA"
"public"
\dRp+ testpub6_forschema
- Publication testpub6_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub6_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"CURRENT_SCHEMA"
"public"
\dRp+ testpub_fortable
- Publication testpub_fortable
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_fortable
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"CURRENT_SCHEMA.CURRENT_SCHEMA"
@@ -919,10 +1425,10 @@ ERROR: schema "testpub_view" does not exist
-- dropping the schema should reflect the change in publication
DROP SCHEMA pub_test3;
\dRp+ testpub2_forschema
- Publication testpub2_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub2_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
"pub_test2"
@@ -930,20 +1436,20 @@ Tables from schemas:
-- renaming the schema should reflect the change in publication
ALTER SCHEMA pub_test1 RENAME to pub_test1_renamed;
\dRp+ testpub2_forschema
- Publication testpub2_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub2_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1_renamed"
"pub_test2"
ALTER SCHEMA pub_test1_renamed RENAME to pub_test1;
\dRp+ testpub2_forschema
- Publication testpub2_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub2_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
"pub_test2"
@@ -951,10 +1457,10 @@ Tables from schemas:
-- alter publication add schema
ALTER PUBLICATION testpub1_forschema ADD ALL TABLES IN SCHEMA pub_test2;
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
"pub_test2"
@@ -963,10 +1469,10 @@ Tables from schemas:
ALTER PUBLICATION testpub1_forschema ADD ALL TABLES IN SCHEMA non_existent_schema;
ERROR: schema "non_existent_schema" does not exist
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
"pub_test2"
@@ -975,10 +1481,10 @@ Tables from schemas:
ALTER PUBLICATION testpub1_forschema ADD ALL TABLES IN SCHEMA pub_test1;
ERROR: schema "pub_test1" is already member of publication "testpub1_forschema"
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
"pub_test2"
@@ -986,10 +1492,10 @@ Tables from schemas:
-- alter publication drop schema
ALTER PUBLICATION testpub1_forschema DROP ALL TABLES IN SCHEMA pub_test2;
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
@@ -997,10 +1503,10 @@ Tables from schemas:
ALTER PUBLICATION testpub1_forschema DROP ALL TABLES IN SCHEMA pub_test2;
ERROR: tables from schema "pub_test2" are not part of the publication
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
@@ -1008,29 +1514,29 @@ Tables from schemas:
ALTER PUBLICATION testpub1_forschema DROP ALL TABLES IN SCHEMA non_existent_schema;
ERROR: schema "non_existent_schema" does not exist
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
-- drop all schemas
ALTER PUBLICATION testpub1_forschema DROP ALL TABLES IN SCHEMA pub_test1;
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
(1 row)
-- alter publication set multiple schema
ALTER PUBLICATION testpub1_forschema SET ALL TABLES IN SCHEMA pub_test1, pub_test2;
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
"pub_test2"
@@ -1039,10 +1545,10 @@ Tables from schemas:
ALTER PUBLICATION testpub1_forschema SET ALL TABLES IN SCHEMA non_existent_schema;
ERROR: schema "non_existent_schema" does not exist
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
"pub_test2"
@@ -1051,10 +1557,10 @@ Tables from schemas:
-- removing the duplicate schemas
ALTER PUBLICATION testpub1_forschema SET ALL TABLES IN SCHEMA pub_test1, pub_test1;
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
@@ -1124,18 +1630,18 @@ SET client_min_messages = 'ERROR';
CREATE PUBLICATION testpub3_forschema;
RESET client_min_messages;
\dRp+ testpub3_forschema
- Publication testpub3_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub3_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
(1 row)
ALTER PUBLICATION testpub3_forschema SET ALL TABLES IN SCHEMA pub_test1;
\dRp+ testpub3_forschema
- Publication testpub3_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub3_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
@@ -1145,20 +1651,20 @@ CREATE PUBLICATION testpub_forschema_fortable FOR ALL TABLES IN SCHEMA pub_test1
CREATE PUBLICATION testpub_fortable_forschema FOR TABLE pub_test2.tbl1, ALL TABLES IN SCHEMA pub_test1;
RESET client_min_messages;
\dRp+ testpub_forschema_fortable
- Publication testpub_forschema_fortable
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_forschema_fortable
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"pub_test2.tbl1"
Tables from schemas:
"pub_test1"
\dRp+ testpub_fortable_forschema
- Publication testpub_fortable_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_fortable_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"pub_test2.tbl1"
Tables from schemas:
@@ -1202,40 +1708,85 @@ CREATE SCHEMA sch1;
CREATE SCHEMA sch2;
CREATE TABLE sch1.tbl1 (a int) PARTITION BY RANGE(a);
CREATE TABLE sch2.tbl1_part1 PARTITION OF sch1.tbl1 FOR VALUES FROM (1) to (10);
+CREATE SEQUENCE sch1.seq1;
+CREATE SEQUENCE sch2.seq2;
-- Schema publication that does not include the schema that has the parent table
CREATE PUBLICATION pub FOR ALL TABLES IN SCHEMA sch2 WITH (PUBLISH_VIA_PARTITION_ROOT=1);
+ALTER PUBLICATION pub ADD ALL SEQUENCES IN SCHEMA sch2;
SELECT * FROM pg_publication_tables;
pubname | schemaname | tablename
---------+------------+------------
pub | sch2 | tbl1_part1
(1 row)
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch2 | seq2
+(1 row)
+
DROP PUBLICATION pub;
-- Table publication that does not include the parent table
CREATE PUBLICATION pub FOR TABLE sch2.tbl1_part1 WITH (PUBLISH_VIA_PARTITION_ROOT=1);
+ALTER PUBLICATION pub ADD SEQUENCE sch2.seq2;
SELECT * FROM pg_publication_tables;
pubname | schemaname | tablename
---------+------------+------------
pub | sch2 | tbl1_part1
(1 row)
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch2 | seq2
+(1 row)
+
-- Table publication that includes both the parent table and the child table
ALTER PUBLICATION pub ADD TABLE sch1.tbl1;
+ALTER PUBLICATION pub ADD SEQUENCE sch1.seq1;
SELECT * FROM pg_publication_tables;
pubname | schemaname | tablename
---------+------------+-----------
pub | sch1 | tbl1
(1 row)
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch1 | seq1
+ pub | sch2 | seq2
+(2 rows)
+
DROP PUBLICATION pub;
-- Schema publication that does not include the schema that has the parent table
CREATE PUBLICATION pub FOR ALL TABLES IN SCHEMA sch2 WITH (PUBLISH_VIA_PARTITION_ROOT=0);
+ALTER PUBLICATION pub ADD SEQUENCE sch1.seq1;
SELECT * FROM pg_publication_tables;
pubname | schemaname | tablename
---------+------------+------------
pub | sch2 | tbl1_part1
(1 row)
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch1 | seq1
+(1 row)
+
+DROP PUBLICATION pub;
+-- Sequence publication
+CREATE PUBLICATION pub FOR SEQUENCE sch2.seq2;
+SELECT * FROM pg_publication_tables;
+ pubname | schemaname | tablename
+---------+------------+-----------
+(0 rows)
+
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch2 | seq2
+(1 row)
+
DROP PUBLICATION pub;
-- Table publication that does not include the parent table
CREATE PUBLICATION pub FOR TABLE sch2.tbl1_part1 WITH (PUBLISH_VIA_PARTITION_ROOT=0);
@@ -1245,14 +1796,26 @@ SELECT * FROM pg_publication_tables;
pub | sch2 | tbl1_part1
(1 row)
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+(0 rows)
+
-- Table publication that includes both the parent table and the child table
ALTER PUBLICATION pub ADD TABLE sch1.tbl1;
+ALTER PUBLICATION pub ADD ALL SEQUENCES IN SCHEMA sch2;
SELECT * FROM pg_publication_tables;
pubname | schemaname | tablename
---------+------------+------------
pub | sch2 | tbl1_part1
(1 row)
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch2 | seq2
+(1 row)
+
DROP PUBLICATION pub;
DROP TABLE sch2.tbl1_part1;
DROP TABLE sch1.tbl1;
@@ -1268,9 +1831,81 @@ SELECT * FROM pg_publication_tables;
pub | sch1 | tbl1
(1 row)
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+(0 rows)
+
+DROP PUBLICATION pub;
+-- Schema publication
+CREATE PUBLICATION pub FOR SEQUENCE sch2.seq2;
+SELECT * FROM pg_publication_tables;
+ pubname | schemaname | tablename
+---------+------------+-----------
+(0 rows)
+
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch2 | seq2
+(1 row)
+
+DROP PUBLICATION pub;
+-- Sequence publication
+CREATE PUBLICATION pub FOR ALL SEQUENCES IN SCHEMA sch2;
+SELECT * FROM pg_publication_tables;
+ pubname | schemaname | tablename
+---------+------------+-----------
+(0 rows)
+
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch2 | seq2
+(1 row)
+
+ALTER PUBLICATION pub ADD SEQUENCE sch1.seq1;
+SELECT * FROM pg_publication_tables;
+ pubname | schemaname | tablename
+---------+------------+-----------
+(0 rows)
+
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch1 | seq1
+ pub | sch2 | seq2
+(2 rows)
+
+ALTER PUBLICATION pub DROP SEQUENCE sch1.seq1;
+SELECT * FROM pg_publication_tables;
+ pubname | schemaname | tablename
+---------+------------+-----------
+(0 rows)
+
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch2 | seq2
+(1 row)
+
+ALTER PUBLICATION pub ADD ALL SEQUENCES IN SCHEMA sch1;
+SELECT * FROM pg_publication_tables;
+ pubname | schemaname | tablename
+---------+------------+-----------
+(0 rows)
+
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch1 | seq1
+ pub | sch2 | seq2
+(2 rows)
+
RESET client_min_messages;
DROP PUBLICATION pub;
DROP TABLE sch1.tbl1;
+DROP SEQUENCE sch1.seq1, sch2.seq2;
DROP SCHEMA sch1 cascade;
DROP SCHEMA sch2 cascade;
RESET SESSION AUTHORIZATION;
diff --git a/src/test/regress/expected/rules.out b/src/test/regress/expected/rules.out
index 6cb6388880d..8a5b20e62cb 100644
--- a/src/test/regress/expected/rules.out
+++ b/src/test/regress/expected/rules.out
@@ -1429,6 +1429,14 @@ pg_prepared_xacts| SELECT p.transaction,
FROM ((pg_prepared_xact() p(transaction, gid, prepared, ownerid, dbid)
LEFT JOIN pg_authid u ON ((p.ownerid = u.oid)))
LEFT JOIN pg_database d ON ((p.dbid = d.oid)));
+pg_publication_sequences| SELECT p.pubname,
+ n.nspname AS schemaname,
+ c.relname AS sequencename
+ FROM pg_publication p,
+ LATERAL pg_get_publication_sequences((p.pubname)::text) gps(relid),
+ (pg_class c
+ JOIN pg_namespace n ON ((n.oid = c.relnamespace)))
+ WHERE (c.oid = gps.relid);
pg_publication_tables| SELECT p.pubname,
n.nspname AS schemaname,
c.relname AS tablename
diff --git a/src/test/regress/sql/object_address.sql b/src/test/regress/sql/object_address.sql
index acd0468a9d9..9d8323468d6 100644
--- a/src/test/regress/sql/object_address.sql
+++ b/src/test/regress/sql/object_address.sql
@@ -49,6 +49,7 @@ CREATE TRANSFORM FOR int LANGUAGE SQL (
SET client_min_messages = 'ERROR';
CREATE PUBLICATION addr_pub FOR TABLE addr_nsp.gentable;
CREATE PUBLICATION addr_pub_schema FOR ALL TABLES IN SCHEMA addr_nsp;
+CREATE PUBLICATION addr_pub_schema2 FOR ALL SEQUENCES IN SCHEMA addr_nsp;
RESET client_min_messages;
CREATE SUBSCRIPTION regress_addr_sub CONNECTION '' PUBLICATION bar WITH (connect = false, slot_name = NONE);
CREATE STATISTICS addr_nsp.gentable_stat ON a, b FROM addr_nsp.gentable;
@@ -198,7 +199,8 @@ WITH objects (type, name, args) AS (VALUES
('transform', '{int}', '{sql}'),
('access method', '{btree}', '{}'),
('publication', '{addr_pub}', '{}'),
- ('publication namespace', '{addr_nsp}', '{addr_pub_schema}'),
+ ('publication namespace', '{addr_nsp}', '{addr_pub_schema,t}'),
+ ('publication namespace', '{addr_nsp}', '{addr_pub_schema2,s}'),
('publication relation', '{addr_nsp, gentable}', '{addr_pub}'),
('subscription', '{regress_addr_sub}', '{}'),
('statistics object', '{addr_nsp, gentable_stat}', '{}')
@@ -218,6 +220,7 @@ SELECT (pg_identify_object(addr1.classid, addr1.objid, addr1.objsubid)).*,
DROP FOREIGN DATA WRAPPER addr_fdw CASCADE;
DROP PUBLICATION addr_pub;
DROP PUBLICATION addr_pub_schema;
+DROP PUBLICATION addr_pub_schema2;
DROP SUBSCRIPTION regress_addr_sub;
DROP SCHEMA addr_nsp CASCADE;
diff --git a/src/test/regress/sql/publication.sql b/src/test/regress/sql/publication.sql
index 5457c56b33f..c195e75c6f0 100644
--- a/src/test/regress/sql/publication.sql
+++ b/src/test/regress/sql/publication.sql
@@ -27,7 +27,7 @@ CREATE PUBLICATION testpub_xxx WITH (publish_via_partition_root = 'true', publis
\dRp
-ALTER PUBLICATION testpub_default SET (publish = 'insert, update, delete');
+ALTER PUBLICATION testpub_default SET (publish = 'insert, update, delete, sequence');
\dRp
@@ -46,6 +46,8 @@ ALTER PUBLICATION testpub_foralltables SET (publish = 'insert, update');
CREATE TABLE testpub_tbl2 (id serial primary key, data text);
-- fail - can't add to for all tables publication
ALTER PUBLICATION testpub_foralltables ADD TABLE testpub_tbl2;
+-- fail - can't add a table using ADD SEQUENCE command
+ALTER PUBLICATION testpub_foralltables ADD SEQUENCE testpub_tbl2;
-- fail - can't drop from all tables publication
ALTER PUBLICATION testpub_foralltables DROP TABLE testpub_tbl2;
-- fail - can't add to for all tables publication
@@ -104,6 +106,183 @@ RESET client_min_messages;
DROP TABLE testpub_tbl3, testpub_tbl3a;
DROP PUBLICATION testpub3, testpub4;
+--- adding sequences
+CREATE SEQUENCE testpub_seq0;
+CREATE SEQUENCE pub_test.testpub_seq1;
+
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_forallsequences FOR ALL SEQUENCES WITH (publish = 'sequence');
+RESET client_min_messages;
+ALTER PUBLICATION testpub_forallsequences SET (publish = 'insert, sequence');
+
+CREATE SEQUENCE testpub_seq2;
+-- fail - can't add to for all sequences publication
+ALTER PUBLICATION testpub_forallsequences ADD SEQUENCE testpub_seq2;
+-- fail - can't drop from all sequences publication
+ALTER PUBLICATION testpub_forallsequences DROP SEQUENCE testpub_seq2;
+-- fail - can't add to for all sequences publication
+ALTER PUBLICATION testpub_forallsequences SET SEQUENCE pub_test.testpub_seq1;
+
+-- fail - can't add schema to 'FOR ALL SEQUENCES' publication
+ALTER PUBLICATION testpub_forallsequences ADD ALL SEQUENCES IN SCHEMA pub_test;
+-- fail - can't drop schema from 'FOR ALL SEQUENCES' publication
+ALTER PUBLICATION testpub_forallsequences DROP ALL SEQUENCES IN SCHEMA pub_test;
+-- fail - can't set schema to 'FOR ALL SEQUENCES' publication
+ALTER PUBLICATION testpub_forallsequences SET ALL SEQUENCES IN SCHEMA pub_test;
+
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_forsequence FOR SEQUENCE testpub_seq0;
+RESET client_min_messages;
+-- should be able to add schema to 'FOR SEQUENCE' publication
+ALTER PUBLICATION testpub_forsequence ADD ALL SEQUENCES IN SCHEMA pub_test;
+\dRp+ testpub_forsequence
+-- fail - can't add sequence from the schema we already added
+ALTER PUBLICATION testpub_forsequence ADD SEQUENCE pub_test.testpub_seq1;
+-- fail - can't add sequence using ADD TABLE command
+ALTER PUBLICATION testpub_forsequence ADD TABLE pub_test.testpub_seq1;
+-- should be able to drop schema from 'FOR SEQUENCE' publication
+ALTER PUBLICATION testpub_forsequence DROP ALL SEQUENCES IN SCHEMA pub_test;
+\dRp+ testpub_forsequence
+-- should be able to set schema to 'FOR SEQUENCE' publication
+ALTER PUBLICATION testpub_forsequence SET ALL SEQUENCES IN SCHEMA pub_test;
+\dRp+ testpub_forsequence
+
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_forschema FOR ALL SEQUENCES IN SCHEMA pub_test;
+RESET client_min_messages;
+-- fail - can't create publication with schema and sequence of the same schema
+CREATE PUBLICATION testpub_for_seq_schema FOR ALL SEQUENCES IN SCHEMA pub_test, SEQUENCE pub_test.testpub_seq1;
+-- fail - can't add a sequence of the same schema to the schema publication
+ALTER PUBLICATION testpub_forschema ADD SEQUENCE pub_test.testpub_seq1;
+-- fail - can't drop a sequence from the schema publication which isn't in the
+-- publication
+ALTER PUBLICATION testpub_forschema DROP SEQUENCE pub_test.testpub_seq1;
+-- should be able to set sequence to schema publication
+ALTER PUBLICATION testpub_forschema SET SEQUENCE pub_test.testpub_seq1;
+\dRp+ testpub_forschema
+
+SELECT pubname, puballtables, puballsequences FROM pg_publication WHERE pubname = 'testpub_forallsequences';
+\d+ pub_test.testpub_seq1
+\dRp+ testpub_forallsequences
+DROP SEQUENCE testpub_seq0, pub_test.testpub_seq1, testpub_seq2;
+DROP PUBLICATION testpub_forallsequences, testpub_forsequence, testpub_forschema;
+
+-- Publication mixing tables and sequences
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_mix;
+RESET client_min_messages;
+
+CREATE SEQUENCE testpub_seq1;
+CREATE SEQUENCE pub_test.testpub_seq2;
+
+ALTER PUBLICATION testpub_mix ADD SEQUENCE testpub_seq1, TABLE testpub_tbl1;
+\dRp+ testpub_mix
+
+ALTER PUBLICATION testpub_mix ADD ALL SEQUENCES IN SCHEMA pub_test, ALL TABLES IN SCHEMA pub_test;
+\dRp+ testpub_mix
+
+ALTER PUBLICATION testpub_mix DROP ALL SEQUENCES IN SCHEMA pub_test;
+\dRp+ testpub_mix
+
+ALTER PUBLICATION testpub_mix DROP ALL TABLES IN SCHEMA pub_test;
+\dRp+ testpub_mix
+
+DROP PUBLICATION testpub_mix;
+DROP SEQUENCE testpub_seq1;
+DROP SEQUENCE pub_test.testpub_seq2;
+
+
+-- make sure we replicate only the correct relation type
+CREATE SCHEMA pub_test1;
+CREATE SEQUENCE pub_test1.test_seq1;
+CREATE TABLE pub_test1.test_tbl1 (a int primary key, b int);
+
+CREATE SCHEMA pub_test2;
+CREATE SEQUENCE pub_test2.test_seq2;
+CREATE TABLE pub_test2.test_tbl2 (a int primary key, b int);
+
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_schemas;
+RESET client_min_messages;
+
+-- add tables from one schema, sequences from the other
+ALTER PUBLICATION testpub_schemas ADD ALL TABLES IN SCHEMA pub_test2;
+ALTER PUBLICATION testpub_schemas ADD ALL SEQUENCES IN SCHEMA pub_test1;
+
+\dRp+ testpub_schemas
+
+\dn+ pub_test1
+\dn+ pub_test2
+
+\d+ pub_test1.test_seq1;
+\d+ pub_test1.test_tbl1;
+
+\d+ pub_test2.test_seq2;
+\d+ pub_test2.test_tbl2;
+
+-- add the other object type from each schema
+ALTER PUBLICATION testpub_schemas ADD ALL TABLES IN SCHEMA pub_test1;
+ALTER PUBLICATION testpub_schemas ADD ALL SEQUENCES IN SCHEMA pub_test2;
+
+\dRp+ testpub_schemas
+
+\dn+ pub_test1
+\dn+ pub_test2
+
+\d+ pub_test1.test_seq1;
+\d+ pub_test1.test_tbl1;
+
+\d+ pub_test2.test_seq2;
+\d+ pub_test2.test_tbl2;
+
+-- now drop the object type added first
+ALTER PUBLICATION testpub_schemas DROP ALL TABLES IN SCHEMA pub_test2;
+ALTER PUBLICATION testpub_schemas DROP ALL SEQUENCES IN SCHEMA pub_test1;
+
+\dRp+ testpub_schemas
+
+\dn+ pub_test1
+\dn+ pub_test2
+
+\d+ pub_test1.test_seq1;
+\d+ pub_test1.test_tbl1;
+
+\d+ pub_test2.test_seq2;
+\d+ pub_test2.test_tbl2;
+
+-- should fail (publication contains the whole schema)
+ALTER PUBLICATION testpub_schemas ADD TABLE pub_test1.test_tbl1;
+ALTER PUBLICATION testpub_schemas ADD SEQUENCE pub_test2.test_seq2;
+
+-- should work (different schema)
+ALTER PUBLICATION testpub_schemas ADD TABLE pub_test2.test_tbl2;
+ALTER PUBLICATION testpub_schemas ADD SEQUENCE pub_test1.test_seq1;
+
+\dRp+ testpub_schemas
+
+\d+ pub_test1.test_seq1;
+\d+ pub_test1.test_tbl1;
+
+\d+ pub_test2.test_seq2;
+\d+ pub_test2.test_tbl2;
+
+-- now drop the explicitly added objects again
+ALTER PUBLICATION testpub_schemas DROP TABLE pub_test2.test_tbl2;
+ALTER PUBLICATION testpub_schemas DROP SEQUENCE pub_test1.test_seq1;
+
+\dRp+ testpub_schemas
+
+\d+ pub_test1.test_seq1;
+\d+ pub_test1.test_tbl1;
+
+\d+ pub_test2.test_seq2;
+\d+ pub_test2.test_tbl2;
+
+DROP PUBLICATION testpub_schemas;
+DROP TABLE pub_test1.test_tbl1, pub_test2.test_tbl2;
+DROP SEQUENCE pub_test1.test_seq1, pub_test2.test_seq2;
+DROP SCHEMA pub_test1, pub_test2;
+
-- Tests for partitioned tables
SET client_min_messages = 'ERROR';
CREATE PUBLICATION testpub_forparted;
@@ -717,32 +896,51 @@ CREATE SCHEMA sch1;
CREATE SCHEMA sch2;
CREATE TABLE sch1.tbl1 (a int) PARTITION BY RANGE(a);
CREATE TABLE sch2.tbl1_part1 PARTITION OF sch1.tbl1 FOR VALUES FROM (1) to (10);
+CREATE SEQUENCE sch1.seq1;
+CREATE SEQUENCE sch2.seq2;
-- Schema publication that does not include the schema that has the parent table
CREATE PUBLICATION pub FOR ALL TABLES IN SCHEMA sch2 WITH (PUBLISH_VIA_PARTITION_ROOT=1);
+ALTER PUBLICATION pub ADD ALL SEQUENCES IN SCHEMA sch2;
SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
DROP PUBLICATION pub;
-- Table publication that does not include the parent table
CREATE PUBLICATION pub FOR TABLE sch2.tbl1_part1 WITH (PUBLISH_VIA_PARTITION_ROOT=1);
+ALTER PUBLICATION pub ADD SEQUENCE sch2.seq2;
SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
-- Table publication that includes both the parent table and the child table
ALTER PUBLICATION pub ADD TABLE sch1.tbl1;
+ALTER PUBLICATION pub ADD SEQUENCE sch1.seq1;
SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
DROP PUBLICATION pub;
-- Schema publication that does not include the schema that has the parent table
CREATE PUBLICATION pub FOR ALL TABLES IN SCHEMA sch2 WITH (PUBLISH_VIA_PARTITION_ROOT=0);
+ALTER PUBLICATION pub ADD SEQUENCE sch1.seq1;
+SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
+
+DROP PUBLICATION pub;
+-- Sequence publication
+CREATE PUBLICATION pub FOR SEQUENCE sch2.seq2;
SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
DROP PUBLICATION pub;
-- Table publication that does not include the parent table
CREATE PUBLICATION pub FOR TABLE sch2.tbl1_part1 WITH (PUBLISH_VIA_PARTITION_ROOT=0);
SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
-- Table publication that includes both the parent table and the child table
ALTER PUBLICATION pub ADD TABLE sch1.tbl1;
+ALTER PUBLICATION pub ADD ALL SEQUENCES IN SCHEMA sch2;
SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
DROP PUBLICATION pub;
DROP TABLE sch2.tbl1_part1;
@@ -755,10 +953,36 @@ CREATE TABLE sch1.tbl1_part3 (a int) PARTITION BY RANGE(a);
ALTER TABLE sch1.tbl1 ATTACH PARTITION sch1.tbl1_part3 FOR VALUES FROM (20) to (30);
CREATE PUBLICATION pub FOR ALL TABLES IN SCHEMA sch1 WITH (PUBLISH_VIA_PARTITION_ROOT=1);
SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
+
+DROP PUBLICATION pub;
+-- Schema publication
+CREATE PUBLICATION pub FOR SEQUENCE sch2.seq2;
+SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
+
+DROP PUBLICATION pub;
+-- Sequence publication
+CREATE PUBLICATION pub FOR ALL SEQUENCES IN SCHEMA sch2;
+SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
+
+ALTER PUBLICATION pub ADD SEQUENCE sch1.seq1;
+SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
+
+ALTER PUBLICATION pub DROP SEQUENCE sch1.seq1;
+SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
+
+ALTER PUBLICATION pub ADD ALL SEQUENCES IN SCHEMA sch1;
+SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
RESET client_min_messages;
DROP PUBLICATION pub;
DROP TABLE sch1.tbl1;
+DROP SEQUENCE sch1.seq1, sch2.seq2;
DROP SCHEMA sch1 cascade;
DROP SCHEMA sch2 cascade;
diff --git a/src/test/subscription/t/030_sequences.pl b/src/test/subscription/t/030_sequences.pl
new file mode 100644
index 00000000000..9ae3c03d7d1
--- /dev/null
+++ b/src/test/subscription/t/030_sequences.pl
@@ -0,0 +1,202 @@
+
+# Copyright (c) 2021, PostgreSQL Global Development Group
+
+# This tests that sequences are replicated correctly by logical replication
+use strict;
+use warnings;
+use PostgreSQL::Test::Cluster;
+use PostgreSQL::Test::Utils;
+use Test::More;
+
+# Initialize publisher node
+my $node_publisher = PostgreSQL::Test::Cluster->new('publisher');
+$node_publisher->init(allows_streaming => 'logical');
+$node_publisher->start;
+
+# Create subscriber node
+my $node_subscriber = PostgreSQL::Test::Cluster->new('subscriber');
+$node_subscriber->init(allows_streaming => 'logical');
+$node_subscriber->start;
+
+# Create some preexisting content on publisher
+my $ddl = qq(
+ CREATE TABLE seq_test (v BIGINT);
+ CREATE SEQUENCE s;
+);
+
+# Setup structure on the publisher
+$node_publisher->safe_psql('postgres', $ddl);
+
+# Create some the same structure on subscriber, and an extra sequence that
+# we'll create on the publisher later
+$ddl = qq(
+ CREATE TABLE seq_test (v BIGINT);
+ CREATE SEQUENCE s;
+ CREATE SEQUENCE s2;
+);
+
+$node_subscriber->safe_psql('postgres', $ddl);
+
+# Setup logical replication
+my $publisher_connstr = $node_publisher->connstr . ' dbname=postgres';
+$node_publisher->safe_psql('postgres',
+ "CREATE PUBLICATION seq_pub");
+
+$node_publisher->safe_psql('postgres',
+ "ALTER PUBLICATION seq_pub ADD SEQUENCE s");
+
+$node_subscriber->safe_psql('postgres',
+ "CREATE SUBSCRIPTION seq_sub CONNECTION '$publisher_connstr' PUBLICATION seq_pub"
+);
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Wait for initial sync to finish as well
+my $synced_query =
+ "SELECT count(1) = 0 FROM pg_subscription_rel WHERE srsubstate NOT IN ('s', 'r');";
+$node_subscriber->poll_query_until('postgres', $synced_query)
+ or die "Timed out while waiting for subscriber to synchronize data";
+
+# Insert initial test data
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ -- generate a number of values using the sequence
+ INSERT INTO seq_test SELECT nextval('s') FROM generate_series(1,100);
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+my $result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s;
+));
+
+is( $result, '132|0|t',
+ 'initial test data replicated');
+
+
+# advance the sequence in a rolled-back transaction - the rollback
+# does not wait for the replication, so we could see any intermediate state
+# so do something else after the test, to ensure we wait for everything
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ INSERT INTO seq_test SELECT nextval('s') FROM generate_series(1,100);
+ ROLLBACK;
+ INSERT INTO seq_test VALUES (-1);
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s;
+));
+
+is( $result, '231|0|t',
+ 'advance sequence in rolled-back transaction');
+
+
+# create a new sequence and roll it back - should not be replicated, due to
+# the transactional behavior
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ CREATE SEQUENCE s2;
+ ALTER PUBLICATION seq_pub ADD SEQUENCE s2;
+ INSERT INTO seq_test SELECT nextval('s2') FROM generate_series(1,100);
+ ROLLBACK;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s2;
+));
+
+is( $result, '1|0|f',
+ 'create new sequence and roll it back');
+
+
+# create a new sequence, advance it in a rolled-back transaction, but commit
+# the create - the advance should be replicated nevertheless
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ CREATE SEQUENCE s2;
+ ALTER PUBLICATION seq_pub ADD SEQUENCE s2;
+ SAVEPOINT sp1;
+ INSERT INTO seq_test SELECT nextval('s2') FROM generate_series(1,100);
+ ROLLBACK TO sp1;
+ COMMIT;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Wait for sync of the second sequence we just added to finish
+$synced_query =
+ "SELECT count(1) = 0 FROM pg_subscription_rel WHERE srsubstate NOT IN ('s', 'r');";
+$node_subscriber->poll_query_until('postgres', $synced_query)
+ or die "Timed out while waiting for subscriber to synchronize data";
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s2;
+));
+
+is( $result, '132|0|t',
+ 'create sequence, advance it in rolled-back transaction, but commit the create');
+
+
+# advance the new sequence in a transaction, and roll it back - the rollback
+# does not wait for the replication, so we could see any intermediate state
+# so do something else after the test, to ensure we wait for everything
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ INSERT INTO seq_test SELECT nextval('s2') FROM generate_series(1,100);
+ ROLLBACK;
+ INSERT INTO seq_test VALUES (-1);
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s2;
+));
+
+is( $result, '231|0|t',
+ 'advance the new sequence in a transaction and roll it back');
+
+
+# advance the sequence in a subtransaction - the subtransaction gets rolled
+# back, but commit the main one - the changes should still be replicated
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ SAVEPOINT s1;
+ INSERT INTO seq_test SELECT nextval('s2') FROM generate_series(1,100);
+ ROLLBACK TO s1;
+ COMMIT;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s2;
+));
+
+is( $result, '330|0|t',
+ 'advance sequence in a subtransaction');
+
+
+done_testing();
--
2.34.1
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2022-03-24 21:52 Greg Stark <[email protected]>
parent: Tomas Vondra <[email protected]>
1 sibling, 1 reply; 127+ messages in thread
From: Greg Stark @ 2022-03-24 21:52 UTC (permalink / raw)
To: Tomas Vondra <[email protected]>; +Cc: Petr Jelinek <[email protected]>; Amit Kapila <[email protected]>; Peter Eisentraut <[email protected]>; PostgreSQL Hackers <[email protected]>
Awesome to see this get committed, thanks Tomas.
Is there anything left or shall I update the CF entry to committed?
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2022-03-24 21:59 Tomas Vondra <[email protected]>
parent: Tomas Vondra <[email protected]>
1 sibling, 3 replies; 127+ messages in thread
From: Tomas Vondra @ 2022-03-24 21:59 UTC (permalink / raw)
To: Petr Jelinek <[email protected]>; Amit Kapila <[email protected]>; +Cc: Peter Eisentraut <[email protected]>; PostgreSQL Hackers <[email protected]>
Hi,
Pushed, after going through the patch once more, addressed the remaining
FIXMEs, corrected a couple places in the docs and comments, etc. Minor
tweaks, nothing important.
I've been thinking about the grammar a bit more after pushing, and I
realized that maybe it'd be better to handle the FOR ALL TABLES /
SEQUENCES clause as PublicationObjSpec, not as a separate/special case.
regards
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2022-03-24 22:01 Tomas Vondra <[email protected]>
parent: Greg Stark <[email protected]>
0 siblings, 0 replies; 127+ messages in thread
From: Tomas Vondra @ 2022-03-24 22:01 UTC (permalink / raw)
To: Greg Stark <[email protected]>; +Cc: Petr Jelinek <[email protected]>; Amit Kapila <[email protected]>; Peter Eisentraut <[email protected]>; PostgreSQL Hackers <[email protected]>
On 3/24/22 22:52, Greg Stark wrote:
> Awesome to see this get committed, thanks Tomas.
>
> Is there anything left or shall I update the CF entry to committed?
Yeah, let's mark it as committed. I was waiting for some feedback from
the buildfarm - there are some failures, but it seems unrelated.
regards
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2022-03-25 04:01 Amit Kapila <[email protected]>
parent: Tomas Vondra <[email protected]>
2 siblings, 1 reply; 127+ messages in thread
From: Amit Kapila @ 2022-03-25 04:01 UTC (permalink / raw)
To: Tomas Vondra <[email protected]>; +Cc: Petr Jelinek <[email protected]>; Peter Eisentraut <[email protected]>; PostgreSQL Hackers <[email protected]>
On Fri, Mar 25, 2022 at 3:29 AM Tomas Vondra
<[email protected]> wrote:
>
> Pushed.
>
Some of the comments given by me [1] don't seem to be addressed or
responded to. Let me try to say again for the ease of discussion:
* Don't we need some syncing mechanism between apply worker and
sequence sync worker so that apply worker skips the sequence changes
till the sync worker is finished, otherwise, there is a risk of one
overriding the values of the other? See how we take care of this for a
table in should_apply_changes_for_rel() and its callers. If we don't
do this for sequences for some reason then probably a comment
somewhere is required.
* Don't we need explicit privilege checking before applying sequence
data as we do in commit a2ab9c06ea15fbcb2bfde570986a06b37f52bcca for
tables?
Few new comments:
=================
1. A simple test like the below crashes for me:
postgres=# create sequence s1;
CREATE SEQUENCE
postgres=# create sequence s2;
CREATE SEQUENCE
postgres=# create publication pub1 for sequence s1, s2;
server closed the connection unexpectedly
This probably means the server terminated abnormally
before or while processing the request.
The connection to the server was lost. Attempting reset: Failed.
2. In apply_handle_sequence() do we need AccessExclusiveLock for
non-transactional case?
3. In apply_handle_sequence(), I think for transactional case, we need
to skip the operation, if the skip lsn is set. See how we skip in
apply_handle_insert() and similar functions.
[1] - https://www.postgresql.org/message-id/CAA4eK1Jn-DttQ%3D4Pdh9YCe1w%2BzGbgC%2B0uR0sfg2RtkjiAPmB9g%40ma...
--
With Regards,
Amit Kapila.
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2022-03-25 07:00 Masahiko Sawada <[email protected]>
parent: Tomas Vondra <[email protected]>
2 siblings, 1 reply; 127+ messages in thread
From: Masahiko Sawada @ 2022-03-25 07:00 UTC (permalink / raw)
To: Tomas Vondra <[email protected]>; +Cc: Petr Jelinek <[email protected]>; Amit Kapila <[email protected]>; Peter Eisentraut <[email protected]>; PostgreSQL Hackers <[email protected]>
On Fri, Mar 25, 2022 at 6:59 AM Tomas Vondra
<[email protected]> wrote:
>
> Hi,
>
> Pushed, after going through the patch once more, addressed the remaining
> FIXMEs, corrected a couple places in the docs and comments, etc. Minor
> tweaks, nothing important.
>
The commit updates tab-completion for ALTER PUBLICATION but seems not
to update for CREATE PUBLICATION. I've attached a patch for that.
Also, the commit add a new pgoutput option "sequences":
+ else if (strcmp(defel->defname, "sequences") == 0)
+ {
+ if (sequences_option_given)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("conflicting
or redundant options")));
+ sequences_option_given = true;
+
+ data->sequences = defGetBoolean(defel);
+ }
But as far as I read changes, there is no use of this option, and this
code is not tested. Can we remove it or is it for upcoming changes?
Regards,
--
Masahiko Sawada
EDB: https://www.enterprisedb.com/
Attachments:
[application/octet-stream] tab_completion.patch (3.2K, ../../CAD21AoDtKpdJcHOLjfPQ7TmpFqNB5__=Q_g1e8OBRrwT5LP-=g@mail.gmail.com/2-tab_completion.patch)
download | inline diff:
diff --git a/src/bin/psql/tab-complete.c b/src/bin/psql/tab-complete.c
index e59bd8302d..63bfdf11c6 100644
--- a/src/bin/psql/tab-complete.c
+++ b/src/bin/psql/tab-complete.c
@@ -2971,21 +2971,27 @@ psql_completion(const char *text, int start, int end)
/* CREATE PUBLICATION */
else if (Matches("CREATE", "PUBLICATION", MatchAny))
- COMPLETE_WITH("FOR TABLE", "FOR ALL TABLES", "FOR ALL TABLES IN SCHEMA", "WITH (");
+ COMPLETE_WITH("FOR TABLE", "FOR ALL TABLES", "FOR ALL TABLES IN SCHEMA",
+ "FOR SEQUENCE", "FOR ALL SEQUENCES", "FOR ALL SEQUENCES IN SCHEMA",
+ "WITH (");
else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR"))
- COMPLETE_WITH("TABLE", "ALL TABLES", "ALL TABLES IN SCHEMA");
+ COMPLETE_WITH("TABLE", "ALL TABLES", "ALL TABLES IN SCHEMA",
+ "SEQUENCE", "ALL SEQUENCES", "ALL SEQUENCES IN SCHEMA");
else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "ALL"))
- COMPLETE_WITH("TABLES", "TABLES IN SCHEMA");
- else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "ALL", "TABLES"))
+ COMPLETE_WITH("TABLES", "TABLES IN SCHEMA", "SEQUENCES", "SEQUENCES IN SCHEMA");
+ else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "ALL", "TABLES|SEQUENCES"))
COMPLETE_WITH("IN SCHEMA", "WITH (");
- else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "TABLE", MatchAny) && !ends_with(prev_wd, ','))
+ else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "TABLE|SEQUENCE", MatchAny) && !ends_with(prev_wd, ','))
COMPLETE_WITH("WHERE (", "WITH (");
/* Complete "CREATE PUBLICATION <name> FOR TABLE" with "<table>, ..." */
else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "TABLE"))
COMPLETE_WITH_SCHEMA_QUERY(Query_for_list_of_tables);
+ /* Complete "CREATE PUBLICATION <name> FOR SEQUENCE" with "<sequence>, ..." */
+ else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "SEQUENCE"))
+ COMPLETE_WITH_SCHEMA_QUERY(Query_for_list_of_sequences);
/*
- * "CREATE PUBLICATION <name> FOR TABLE <name> WHERE (" - complete with
+ * "CREATE PUBLICATION <name> FOR TABLE|SEQUENCE <name> WHERE (" - complete with
* table attributes
*/
else if (HeadMatches("CREATE", "PUBLICATION", MatchAny) && TailMatches("WHERE"))
@@ -2996,14 +3002,14 @@ psql_completion(const char *text, int start, int end)
COMPLETE_WITH(" WITH (");
/*
- * Complete "CREATE PUBLICATION <name> FOR ALL TABLES IN SCHEMA <schema>,
+ * Complete "CREATE PUBLICATION <name> FOR ALL TABLES|SEQUENCES IN SCHEMA <schema>,
* ..."
*/
- else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "ALL", "TABLES", "IN", "SCHEMA"))
+ else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "ALL", "TABLES|SEQUENCES", "IN", "SCHEMA"))
COMPLETE_WITH_QUERY_PLUS(Query_for_list_of_schemas
" AND nspname NOT LIKE E'pg\\\\_%%'",
"CURRENT_SCHEMA");
- else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "ALL", "TABLES", "IN", "SCHEMA", MatchAny) && (!ends_with(prev_wd, ',')))
+ else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "ALL", "TABLES|SEQUENCES", "IN", "SCHEMA", MatchAny) && (!ends_with(prev_wd, ',')))
COMPLETE_WITH("WITH (");
/* Complete "CREATE PUBLICATION <name> [...] WITH" */
else if (HeadMatches("CREATE", "PUBLICATION") && TailMatches("WITH", "("))
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2022-03-25 10:26 Tomas Vondra <[email protected]>
parent: Amit Kapila <[email protected]>
0 siblings, 1 reply; 127+ messages in thread
From: Tomas Vondra @ 2022-03-25 10:26 UTC (permalink / raw)
To: Amit Kapila <[email protected]>; +Cc: Petr Jelinek <[email protected]>; Peter Eisentraut <[email protected]>; PostgreSQL Hackers <[email protected]>
On 3/25/22 05:01, Amit Kapila wrote:
> On Fri, Mar 25, 2022 at 3:29 AM Tomas Vondra
> <[email protected]> wrote:
>>
>> Pushed.
>>
>
> Some of the comments given by me [1] don't seem to be addressed or
> responded to. Let me try to say again for the ease of discussion:
>
D'oh! I got distracted by Petr's response to that message, and missed
this part ...
> * Don't we need some syncing mechanism between apply worker and
> sequence sync worker so that apply worker skips the sequence changes
> till the sync worker is finished, otherwise, there is a risk of one
> overriding the values of the other? See how we take care of this for a
> table in should_apply_changes_for_rel() and its callers. If we don't
> do this for sequences for some reason then probably a comment
> somewhere is required.
>
How would that happen? If we're effectively setting the sequence as a
side effect of inserting the data, then why should we even replicate the
sequence? We'll have the problem later too, no?
> * Don't we need explicit privilege checking before applying sequence
> data as we do in commit a2ab9c06ea15fbcb2bfde570986a06b37f52bcca for
> tables?
>
So essentially something like TargetPrivilegesCheck in the worker? I
think you're probably right we need something like that.
> Few new comments:
> =================
> 1. A simple test like the below crashes for me:
> postgres=# create sequence s1;
> CREATE SEQUENCE
> postgres=# create sequence s2;
> CREATE SEQUENCE
> postgres=# create publication pub1 for sequence s1, s2;
> server closed the connection unexpectedly
> This probably means the server terminated abnormally
> before or while processing the request.
> The connection to the server was lost. Attempting reset: Failed.
>
Yeah, preprocess_pubobj_list seems to be a few bricks shy. I have a fix,
will push shortly.
> 2. In apply_handle_sequence() do we need AccessExclusiveLock for
> non-transactional case?
>
Good catch. This lock was inherited from ResetSequence, but now that the
transactional case works differently, we probably don't need it.
> 3. In apply_handle_sequence(), I think for transactional case, we need
> to skip the operation, if the skip lsn is set. See how we skip in
> apply_handle_insert() and similar functions.
>
Right.
Thanks for these reports!
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2022-03-25 10:34 Tomas Vondra <[email protected]>
parent: Masahiko Sawada <[email protected]>
0 siblings, 0 replies; 127+ messages in thread
From: Tomas Vondra @ 2022-03-25 10:34 UTC (permalink / raw)
To: Masahiko Sawada <[email protected]>; +Cc: Petr Jelinek <[email protected]>; Amit Kapila <[email protected]>; Peter Eisentraut <[email protected]>; PostgreSQL Hackers <[email protected]>
On 3/25/22 08:00, Masahiko Sawada wrote:
> On Fri, Mar 25, 2022 at 6:59 AM Tomas Vondra
> <[email protected]> wrote:
>>
>> Hi,
>>
>> Pushed, after going through the patch once more, addressed the remaining
>> FIXMEs, corrected a couple places in the docs and comments, etc. Minor
>> tweaks, nothing important.
>>
>
> The commit updates tab-completion for ALTER PUBLICATION but seems not
> to update for CREATE PUBLICATION. I've attached a patch for that.
>
Thanks. I'm pretty sure the patch did that, but it likely got lost in
one of the rebases due to a conflict. Too bad we don't have tests for
tab-complete. Will fix.
> Also, the commit add a new pgoutput option "sequences":
>
> + else if (strcmp(defel->defname, "sequences") == 0)
> + {
> + if (sequences_option_given)
> + ereport(ERROR,
> + (errcode(ERRCODE_SYNTAX_ERROR),
> + errmsg("conflicting
> or redundant options")));
> + sequences_option_given = true;
> +
> + data->sequences = defGetBoolean(defel);
> + }
>
> But as far as I read changes, there is no use of this option, and this
> code is not tested. Can we remove it or is it for upcoming changes?
>
pgoutput_sequence uses this
if (!data->sequences)
return;
This was inspired by what we do for logical messages, but maybe there's
an argument we don't need this, considering we have "sequence" action
and that a sequence has to be added to the publication. I don't think
there's any future patch relying on this (and it could add it back, if
needed).
regards
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2022-03-25 11:21 Amit Kapila <[email protected]>
parent: Tomas Vondra <[email protected]>
0 siblings, 2 replies; 127+ messages in thread
From: Amit Kapila @ 2022-03-25 11:21 UTC (permalink / raw)
To: Tomas Vondra <[email protected]>; +Cc: Petr Jelinek <[email protected]>; Peter Eisentraut <[email protected]>; PostgreSQL Hackers <[email protected]>
On Fri, Mar 25, 2022 at 3:56 PM Tomas Vondra
<[email protected]> wrote:
>
>
> On 3/25/22 05:01, Amit Kapila wrote:
> > On Fri, Mar 25, 2022 at 3:29 AM Tomas Vondra
> > <[email protected]> wrote:
> >>
> >> Pushed.
> >>
> >
> > Some of the comments given by me [1] don't seem to be addressed or
> > responded to. Let me try to say again for the ease of discussion:
> >
>
> D'oh! I got distracted by Petr's response to that message, and missed
> this part ...
>
> > * Don't we need some syncing mechanism between apply worker and
> > sequence sync worker so that apply worker skips the sequence changes
> > till the sync worker is finished, otherwise, there is a risk of one
> > overriding the values of the other? See how we take care of this for a
> > table in should_apply_changes_for_rel() and its callers. If we don't
> > do this for sequences for some reason then probably a comment
> > somewhere is required.
> >
>
> How would that happen? If we're effectively setting the sequence as a
> side effect of inserting the data, then why should we even replicate the
> sequence?
>
I was talking just about sequence values here, considering that some
sequence is just replicating based on nextval. I think the problem is
that apply worker might override what copy has done if an apply worker
is behind the sequence sync worker as both can run in parallel. Let me
try to take some theoretical example to explain this:
Assume, at LSN 10000, the value of sequence s1 is 10. Then by LSN
12000, the value of s1 becomes 20. Now, say copy decides to copy the
sequence value till LSN 12000 which means it will make the value as 20
on the subscriber, now, in parallel, apply worker can process LSN
10000 and make it again 10. Apply worker might end up redoing all
sequence operations along with some transactional ones where we
recreate the file. I am not sure what exact problem it can lead to but
I think we don't need to redo the work.
We'll have the problem later too, no?
>
> > * Don't we need explicit privilege checking before applying sequence
> > data as we do in commit a2ab9c06ea15fbcb2bfde570986a06b37f52bcca for
> > tables?
> >
>
> So essentially something like TargetPrivilegesCheck in the worker?
>
Right.
Few more comments:
==================
1.
@@ -636,7 +704,7 @@ CreatePublication(ParseState *pstate,
CreatePublicationStmt *stmt)
get_database_name(MyDatabaseId));
/* FOR ALL TABLES requires superuser */
- if (stmt->for_all_tables && !superuser())
+ if (for_all_tables && !superuser())
ereport(ERROR,
(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
errmsg("must be superuser to create FOR ALL TABLES
publication")));
Don't we need a similar check for 'for_all_schema' publications?
2.
+<varlistentry>
+<term>
+ Int8
+</term>
+<listitem>
+<para>
+ 1 if the sequence update is transactions, 0 otherwise.
Shall we say transactional instead of transactions?
3.
+/*
+ * Determine object type given the object type set for a schema.
+ */
+char
+pub_get_object_type_for_relkind(char relkind)
Shouldn't it be 'relation' instead of 'schema' at the end of the sentence?
4.
@@ -1739,13 +1804,13 @@ get_rel_sync_entry(PGOutputData *data,
Relation relation)
{
Oid schemaId = get_rel_namespace(relid);
List *pubids = GetRelationPublications(relid);
-
+ char objectType =
pub_get_object_type_for_relkind(get_rel_relkind(relid));
A few lines after this we are again getting relkind which is not a big
deal but OTOH there doesn't seem to be a need to fetch the same thing
twice from the cache.
5.
+
+ /* Check that user is allowed to manipulate the publication tables. */
+ if (sequences && pubform->puballsequences)
/tables/sequences
6.
+apply_handle_sequence(StringInfo s)
{
...
+
+ relid = RangeVarGetRelid(makeRangeVar(seq.nspname,
+ seq.seqname, -1),
+ RowExclusiveLock, false);
...
}
As here, we are using missing_ok, if the sequence doesn't exist, it
will give a message like: "ERROR: relation "public.s1" does not
exist" whereas for tables we give a slightly more clear message like:
"ERROR: logical replication target relation "public.t1" does not
exist". This is handled via logicalrep_rel_open().
--
With Regards,
Amit Kapila.
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2022-03-25 11:59 Tomas Vondra <[email protected]>
parent: Amit Kapila <[email protected]>
1 sibling, 1 reply; 127+ messages in thread
From: Tomas Vondra @ 2022-03-25 11:59 UTC (permalink / raw)
To: Amit Kapila <[email protected]>; +Cc: Petr Jelinek <[email protected]>; Peter Eisentraut <[email protected]>; PostgreSQL Hackers <[email protected]>
On 3/25/22 12:21, Amit Kapila wrote:
> On Fri, Mar 25, 2022 at 3:56 PM Tomas Vondra
> <[email protected]> wrote:
>>
>>
>> On 3/25/22 05:01, Amit Kapila wrote:
>>> On Fri, Mar 25, 2022 at 3:29 AM Tomas Vondra
>>> <[email protected]> wrote:
>>>>
>>>> Pushed.
>>>>
>>>
>>> Some of the comments given by me [1] don't seem to be addressed or
>>> responded to. Let me try to say again for the ease of discussion:
>>>
>>
>> D'oh! I got distracted by Petr's response to that message, and missed
>> this part ...
>>
>>> * Don't we need some syncing mechanism between apply worker and
>>> sequence sync worker so that apply worker skips the sequence changes
>>> till the sync worker is finished, otherwise, there is a risk of one
>>> overriding the values of the other? See how we take care of this for a
>>> table in should_apply_changes_for_rel() and its callers. If we don't
>>> do this for sequences for some reason then probably a comment
>>> somewhere is required.
>>>
>>
>> How would that happen? If we're effectively setting the sequence as a
>> side effect of inserting the data, then why should we even replicate the
>> sequence?
>>
>
> I was talking just about sequence values here, considering that some
> sequence is just replicating based on nextval. I think the problem is
> that apply worker might override what copy has done if an apply worker
> is behind the sequence sync worker as both can run in parallel. Let me
> try to take some theoretical example to explain this:
>
> Assume, at LSN 10000, the value of sequence s1 is 10. Then by LSN
> 12000, the value of s1 becomes 20. Now, say copy decides to copy the
> sequence value till LSN 12000 which means it will make the value as 20
> on the subscriber, now, in parallel, apply worker can process LSN
> 10000 and make it again 10. Apply worker might end up redoing all
> sequence operations along with some transactional ones where we
> recreate the file. I am not sure what exact problem it can lead to but
> I think we don't need to redo the work.
>
> We'll have the problem later too, no?
>
Ah, I was confused why this would be an issue for sequences and not for
plain tables, but now I realize apply_handle_sequence() is not called in
apply_handle_sequence. Yes, that's probably a thinko.
>>> * Don't we need explicit privilege checking before applying sequence
>>> data as we do in commit a2ab9c06ea15fbcb2bfde570986a06b37f52bcca for
>>> tables?
>>>
>>
>> So essentially something like TargetPrivilegesCheck in the worker?
>>
>
> Right.
>
OK, will do.
> Few more comments:
> ==================
> 1.
> @@ -636,7 +704,7 @@ CreatePublication(ParseState *pstate,
> CreatePublicationStmt *stmt)
> get_database_name(MyDatabaseId));
>
> /* FOR ALL TABLES requires superuser */
> - if (stmt->for_all_tables && !superuser())
> + if (for_all_tables && !superuser())
> ereport(ERROR,
> (errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
> errmsg("must be superuser to create FOR ALL TABLES
> publication")));
>
> Don't we need a similar check for 'for_all_schema' publications?
>
I think you mean "for_all_sequences", right?
> 2.
> +<varlistentry>
> +<term>
> + Int8
> +</term>
> +<listitem>
> +<para>
> + 1 if the sequence update is transactions, 0 otherwise.
>
> Shall we say transactional instead of transactions?
>
> 3.
> +/*
> + * Determine object type given the object type set for a schema.
> + */
> +char
> +pub_get_object_type_for_relkind(char relkind)
>
> Shouldn't it be 'relation' instead of 'schema' at the end of the sentence?
>
> 4.
> @@ -1739,13 +1804,13 @@ get_rel_sync_entry(PGOutputData *data,
> Relation relation)
> {
> Oid schemaId = get_rel_namespace(relid);
> List *pubids = GetRelationPublications(relid);
> -
> + char objectType =
> pub_get_object_type_for_relkind(get_rel_relkind(relid));
>
> A few lines after this we are again getting relkind which is not a big
> deal but OTOH there doesn't seem to be a need to fetch the same thing
> twice from the cache.
>
> 5.
> +
> + /* Check that user is allowed to manipulate the publication tables. */
> + if (sequences && pubform->puballsequences)
>
> /tables/sequences
>
> 6.
> +apply_handle_sequence(StringInfo s)
> {
> ...
> +
> + relid = RangeVarGetRelid(makeRangeVar(seq.nspname,
> + seq.seqname, -1),
> + RowExclusiveLock, false);
> ...
> }
>
> As here, we are using missing_ok, if the sequence doesn't exist, it
> will give a message like: "ERROR: relation "public.s1" does not
> exist" whereas for tables we give a slightly more clear message like:
> "ERROR: logical replication target relation "public.t1" does not
> exist". This is handled via logicalrep_rel_open().
>
Thanks, I'll look at rewording these comments and messages.
regards
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2022-03-25 14:34 vignesh C <[email protected]>
parent: Tomas Vondra <[email protected]>
2 siblings, 1 reply; 127+ messages in thread
From: vignesh C @ 2022-03-25 14:34 UTC (permalink / raw)
To: Tomas Vondra <[email protected]>; +Cc: Petr Jelinek <[email protected]>; Amit Kapila <[email protected]>; Peter Eisentraut <[email protected]>; PostgreSQL Hackers <[email protected]>
On Fri, Mar 25, 2022 at 3:29 AM Tomas Vondra
<[email protected]> wrote:
>
> Hi,
>
> Pushed, after going through the patch once more, addressed the remaining
> FIXMEs, corrected a couple places in the docs and comments, etc. Minor
> tweaks, nothing important.
While rebasing patch [1] I found a couple of comments:
static void
ObjectsInPublicationToOids(List *pubobjspec_list, ParseState *pstate,
- List **rels, List **schemas)
+ List **tables, List **sequences,
+ List **tables_schemas, List **sequences_schemas,
+ List **schemas)
{
ListCell *cell;
PublicationObjSpec *pubobj;
@@ -185,12 +194,23 @@ ObjectsInPublicationToOids(List
*pubobjspec_list, ParseState *pstate,
switch (pubobj->pubobjtype)
{
case PUBLICATIONOBJ_TABLE:
- *rels = lappend(*rels, pubobj->pubtable);
+ *tables = lappend(*tables, pubobj->pubtable);
+ break;
+ case PUBLICATIONOBJ_SEQUENCE:
+ *sequences = lappend(*sequences, pubobj->pubtable);
break;
case PUBLICATIONOBJ_TABLES_IN_SCHEMA:
schemaid = get_namespace_oid(pubobj->name, false);
/* Filter out duplicates if user specifies "sch1, sch1" */
+ *tables_schemas = list_append_unique_oid(*tables_schemas, schemaid);
+ *schemas = list_append_unique_oid(*schemas, schemaid);
+ break;
Now tables_schemas and sequence_schemas are being updated and used in
ObjectsInPublicationToOids, schema parameter is no longer being used
after processing in ObjectsInPublicationToOids, I felt we can remove
that parameter.
/* ALTER PUBLICATION <name> ADD */
else if (Matches("ALTER", "PUBLICATION", MatchAny, "ADD"))
- COMPLETE_WITH("ALL TABLES IN SCHEMA", "TABLE");
+ COMPLETE_WITH("ALL TABLES IN SCHEMA", "ALL SEQUENCES IN SCHEMA",
"TABLE", "SEQUENCE");
Tab completion of alter publication for ADD and DROP is the same, we
could combine it.
Attached a patch for the same.
Thoughts?
[1] - https://www.postgresql.org/message-id/CALDaNm3%3DJrucjhiiwsYQw5-PGtBHFONa6F7hhWCXMsGvh%3DtamA%40mail...
Regards,
Vignesh
Attachments:
[text/x-patch] 0001-Removed-unused-parameter-from-ObjectsInPublicationTo.patch (5.1K, ../../CALDaNm3q_9AnXPwQ8fOod2=OhsB4Bz9UYrxOJ1Jnq85pZWXhnQ@mail.gmail.com/2-0001-Removed-unused-parameter-from-ObjectsInPublicationTo.patch)
download | inline diff:
From dc707ba93494e3ba0cffd8ab6e1fb1b7a1c0e70e Mon Sep 17 00:00:00 2001
From: Vigneshwaran C <[email protected]>
Date: Fri, 25 Mar 2022 19:49:40 +0530
Subject: [PATCH] Removed unused parameter from ObjectsInPublicationToOids.
Removed unused parameter from ObjectsInPublicationToOids.
---
src/backend/commands/publicationcmds.c | 15 +++------------
src/bin/psql/tab-complete.c | 5 +----
2 files changed, 4 insertions(+), 16 deletions(-)
diff --git a/src/backend/commands/publicationcmds.c b/src/backend/commands/publicationcmds.c
index c6437799c5..89298d7857 100644
--- a/src/backend/commands/publicationcmds.c
+++ b/src/backend/commands/publicationcmds.c
@@ -175,8 +175,7 @@ parse_publication_options(ParseState *pstate,
static void
ObjectsInPublicationToOids(List *pubobjspec_list, ParseState *pstate,
List **tables, List **sequences,
- List **tables_schemas, List **sequences_schemas,
- List **schemas)
+ List **tables_schemas, List **sequences_schemas)
{
ListCell *cell;
PublicationObjSpec *pubobj;
@@ -204,14 +203,12 @@ ObjectsInPublicationToOids(List *pubobjspec_list, ParseState *pstate,
/* Filter out duplicates if user specifies "sch1, sch1" */
*tables_schemas = list_append_unique_oid(*tables_schemas, schemaid);
- *schemas = list_append_unique_oid(*schemas, schemaid);
break;
case PUBLICATIONOBJ_SEQUENCES_IN_SCHEMA:
schemaid = get_namespace_oid(pubobj->name, false);
/* Filter out duplicates if user specifies "sch1, sch1" */
*sequences_schemas = list_append_unique_oid(*sequences_schemas, schemaid);
- *schemas = list_append_unique_oid(*schemas, schemaid);
break;
case PUBLICATIONOBJ_TABLES_IN_CUR_SCHEMA:
search_path = fetch_search_path(false);
@@ -225,7 +222,6 @@ ObjectsInPublicationToOids(List *pubobjspec_list, ParseState *pstate,
/* Filter out duplicates if user specifies "sch1, sch1" */
*tables_schemas = list_append_unique_oid(*tables_schemas, schemaid);
- *schemas = list_append_unique_oid(*schemas, schemaid);
break;
case PUBLICATIONOBJ_SEQUENCES_IN_CUR_SCHEMA:
search_path = fetch_search_path(false);
@@ -239,7 +235,6 @@ ObjectsInPublicationToOids(List *pubobjspec_list, ParseState *pstate,
/* Filter out duplicates if user specifies "sch1, sch1" */
*sequences_schemas = list_append_unique_oid(*sequences_schemas, schemaid);
- *schemas = list_append_unique_oid(*schemas, schemaid);
break;
default:
/* shouldn't happen */
@@ -679,7 +674,6 @@ CreatePublication(ParseState *pstate, CreatePublicationStmt *stmt)
List *sequences = NIL;
List *tables_schemaidlist = NIL;
List *sequences_schemaidlist = NIL;
- List *schemaidlist = NIL;
bool for_all_tables = false;
bool for_all_sequences = false;
@@ -782,8 +776,7 @@ CreatePublication(ParseState *pstate, CreatePublicationStmt *stmt)
ObjectsInPublicationToOids(stmt->pubobjects, pstate,
&tables, &sequences,
&tables_schemaidlist,
- &sequences_schemaidlist,
- &schemaidlist);
+ &sequences_schemaidlist);
/* FOR ALL TABLES IN SCHEMA requires superuser */
if (list_length(tables_schemaidlist) > 0 && !superuser())
@@ -1462,14 +1455,12 @@ AlterPublication(ParseState *pstate, AlterPublicationStmt *stmt)
List *sequences = NIL;
List *tables_schemaidlist = NIL;
List *sequences_schemaidlist = NIL;
- List *schemaidlist = NIL;
Oid pubid = pubform->oid;
ObjectsInPublicationToOids(stmt->pubobjects, pstate,
&tables, &sequences,
&tables_schemaidlist,
- &sequences_schemaidlist,
- &schemaidlist);
+ &sequences_schemaidlist);
CheckAlterPublication(stmt, tup,
tables, tables_schemaidlist,
diff --git a/src/bin/psql/tab-complete.c b/src/bin/psql/tab-complete.c
index 63bfdf11c6..682d4fe18d 100644
--- a/src/bin/psql/tab-complete.c
+++ b/src/bin/psql/tab-complete.c
@@ -1814,7 +1814,7 @@ psql_completion(const char *text, int start, int end)
else if (Matches("ALTER", "PUBLICATION", MatchAny))
COMPLETE_WITH("ADD", "DROP", "OWNER TO", "RENAME TO", "SET");
/* ALTER PUBLICATION <name> ADD */
- else if (Matches("ALTER", "PUBLICATION", MatchAny, "ADD"))
+ else if (Matches("ALTER", "PUBLICATION", MatchAny, "ADD|DROP"))
COMPLETE_WITH("ALL TABLES IN SCHEMA", "ALL SEQUENCES IN SCHEMA", "TABLE", "SEQUENCE");
else if (Matches("ALTER", "PUBLICATION", MatchAny, "ADD|SET", "TABLE") ||
(HeadMatches("ALTER", "PUBLICATION", MatchAny, "ADD|SET", "TABLE") &&
@@ -1840,9 +1840,6 @@ psql_completion(const char *text, int start, int end)
COMPLETE_WITH(",", "WHERE (");
else if (HeadMatches("ALTER", "PUBLICATION", MatchAny, "ADD|SET", "TABLE"))
COMPLETE_WITH(",");
- /* ALTER PUBLICATION <name> DROP */
- else if (Matches("ALTER", "PUBLICATION", MatchAny, "DROP"))
- COMPLETE_WITH("ALL TABLES IN SCHEMA", "ALL SEQUENCES IN SCHEMA", "TABLE", "SEQUENCE");
/* ALTER PUBLICATION <name> SET */
else if (Matches("ALTER", "PUBLICATION", MatchAny, "SET"))
COMPLETE_WITH("(", "ALL TABLES IN SCHEMA", "ALL SEQUENCES IN SCHEMA", "TABLE", "SEQUENCE");
--
2.32.0
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2022-03-25 16:50 Tomas Vondra <[email protected]>
parent: Tomas Vondra <[email protected]>
0 siblings, 1 reply; 127+ messages in thread
From: Tomas Vondra @ 2022-03-25 16:50 UTC (permalink / raw)
To: Amit Kapila <[email protected]>; +Cc: Petr Jelinek <[email protected]>; Peter Eisentraut <[email protected]>; PostgreSQL Hackers <[email protected]>
On 3/25/22 12:59, Tomas Vondra wrote:
>
> On 3/25/22 12:21, Amit Kapila wrote:
>> On Fri, Mar 25, 2022 at 3:56 PM Tomas Vondra
>> <[email protected]> wrote:
>>>
>>>
>>> On 3/25/22 05:01, Amit Kapila wrote:
>>>> On Fri, Mar 25, 2022 at 3:29 AM Tomas Vondra
>>>> <[email protected]> wrote:
>>>>>
>>>>> Pushed.
>>>>>
>>>>
>>>> Some of the comments given by me [1] don't seem to be addressed or
>>>> responded to. Let me try to say again for the ease of discussion:
>>>>
>>>
>>> D'oh! I got distracted by Petr's response to that message, and missed
>>> this part ...
>>>
>>>> * Don't we need some syncing mechanism between apply worker and
>>>> sequence sync worker so that apply worker skips the sequence changes
>>>> till the sync worker is finished, otherwise, there is a risk of one
>>>> overriding the values of the other? See how we take care of this for a
>>>> table in should_apply_changes_for_rel() and its callers. If we don't
>>>> do this for sequences for some reason then probably a comment
>>>> somewhere is required.
>>>>
>>>
>>> How would that happen? If we're effectively setting the sequence as a
>>> side effect of inserting the data, then why should we even replicate the
>>> sequence?
>>>
>>
>> I was talking just about sequence values here, considering that some
>> sequence is just replicating based on nextval. I think the problem is
>> that apply worker might override what copy has done if an apply worker
>> is behind the sequence sync worker as both can run in parallel. Let me
>> try to take some theoretical example to explain this:
>>
>> Assume, at LSN 10000, the value of sequence s1 is 10. Then by LSN
>> 12000, the value of s1 becomes 20. Now, say copy decides to copy the
>> sequence value till LSN 12000 which means it will make the value as 20
>> on the subscriber, now, in parallel, apply worker can process LSN
>> 10000 and make it again 10. Apply worker might end up redoing all
>> sequence operations along with some transactional ones where we
>> recreate the file. I am not sure what exact problem it can lead to but
>> I think we don't need to redo the work.
>>
>> We'll have the problem later too, no?
>>
>
> Ah, I was confused why this would be an issue for sequences and not for
> plain tables, but now I realize apply_handle_sequence() is not called in
> apply_handle_sequence. Yes, that's probably a thinko.
>
Hmm, so fixing this might be a bit trickier than I expected.
Firstly, currently we only send nspname/relname in the sequence message,
not the remote OID or schema. The idea was that for sequences we don't
really need schema info, so this seemed OK.
But should_apply_changes_for_rel() needs LogicalRepRelMapEntry, and to
create/maintain that those records we need to send the schema.
Attached is a WIP patch does that.
Two places need more work, I think:
1) maybe_send_schema needs ReorderBufferChange, but we don't have that
for sequences, we only have TXN. I created a simple wrapper, but maybe
we should just tweak maybe_send_schema to use TXN.
2) The transaction handling in is a bit confusing. The non-transactional
increments won't have any explicit commit later, so we can't just rely
on begin_replication_step/end_replication_step. But I want to try
spending a bit more time on this.
But there's a more serious issue, I think. So far, we allowed this:
BEGIN;
CREATE SEQUENCE s2;
ALTER PUBLICATION p ADD SEQUENCE s2;
INSERT INTO seq_test SELECT nextval('s2') FROM generate_series(1,100);
COMMIT;
and the behavior was that we replicated the changes. But with the patch
applied, that no longer happens, because should_apply_changes_for_rel
says the change should not be applied.
And after thinking about this, I think that's correct - we can't apply
changes until ALTER SUBSCRIPTION ... REFRESH PUBLICATION gets executed,
and we can't do that until the transaction commits.
So I guess that's correct, and the current behavior is a bug.
For a while I was thinking that maybe this means we don't need the
transactional behavior at all, but I think we do - we have to handle
ALTER SEQUENCE cases that are transactional.
Does that make sense, Amit?
regards
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
Attachments:
[text/x-patch] sequences-tablesync-fix.patch (6.0K, ../../[email protected]/2-sequences-tablesync-fix.patch)
download | inline diff:
diff --git a/src/backend/replication/logical/proto.c b/src/backend/replication/logical/proto.c
index 3dbe85d61a..0ae0378191 100644
--- a/src/backend/replication/logical/proto.c
+++ b/src/backend/replication/logical/proto.c
@@ -657,7 +657,6 @@ logicalrep_write_sequence(StringInfo out, Relation rel, TransactionId xid,
int64 last_value, int64 log_cnt, bool is_called)
{
uint8 flags = 0;
- char *relname;
pq_sendbyte(out, LOGICAL_REP_MSG_SEQUENCE);
@@ -668,9 +667,8 @@ logicalrep_write_sequence(StringInfo out, Relation rel, TransactionId xid,
pq_sendint8(out, flags);
pq_sendint64(out, lsn);
- logicalrep_write_namespace(out, RelationGetNamespace(rel));
- relname = RelationGetRelationName(rel);
- pq_sendstring(out, relname);
+ /* OID ad sequence identifier */
+ pq_sendint32(out, RelationGetRelid(rel));
pq_sendint8(out, transactional);
pq_sendint64(out, last_value);
@@ -688,9 +686,8 @@ logicalrep_read_sequence(StringInfo in, LogicalRepSequence *seqdata)
pq_getmsgint(in, 1);
pq_getmsgint64(in);
- /* Read relation name from stream */
- seqdata->nspname = pstrdup(logicalrep_read_namespace(in));
- seqdata->seqname = pstrdup(pq_getmsgstring(in));
+ /* Read relation OID from stream */
+ seqdata->remoteid = pq_getmsgint(in, 4);
seqdata->transactional = pq_getmsgint(in, 1);
seqdata->last_value = pq_getmsgint64(in);
diff --git a/src/backend/replication/logical/worker.c b/src/backend/replication/logical/worker.c
index f3868b3e1f..d2acb08996 100644
--- a/src/backend/replication/logical/worker.c
+++ b/src/backend/replication/logical/worker.c
@@ -1151,7 +1151,8 @@ static void
apply_handle_sequence(StringInfo s)
{
LogicalRepSequence seq;
- Oid relid;
+ LogicalRepRelMapEntry *rel;
+ LOCKMODE lockmode;
if (handle_streamed_transaction(LOGICAL_REP_MSG_SEQUENCE, s))
return;
@@ -1167,9 +1168,8 @@ apply_handle_sequence(StringInfo s)
Assert(!(!seq.transactional && IsTransactionState()));
/*
- * Make sure we're in a transaction (needed by SetSequence). For
- * non-transactional updates we're guaranteed to start a new one,
- * and we'll commit it at the end.
+ * Make sure there's a transaction (the non-transactional case may can't
+ * have one yet).
*/
if (!IsTransactionState())
{
@@ -1177,15 +1177,38 @@ apply_handle_sequence(StringInfo s)
maybe_reread_subscription();
}
- relid = RangeVarGetRelid(makeRangeVar(seq.nspname,
- seq.seqname, -1),
- RowExclusiveLock, false);
+ /*
+ * Use the necessary lock mode. For transactional changes we need
+ * AccessExclusiveLock (just like ResetSequence), while for regular
+ * updates RowExclusiveLock is enough).
+ */
+ lockmode = (seq.transactional ? AccessExclusiveLock : RowExclusiveLock);
- /* lock the sequence in AccessExclusiveLock, as expected by SetSequence */
- LockRelationOid(relid, AccessExclusiveLock);
+ rel = logicalrep_rel_open(seq.remoteid, lockmode);
+
+ if (!should_apply_changes_for_rel(rel))
+ {
+ /*
+ * The relation can't become interesting in the middle of the
+ * transaction so it's safe to unlock it.
+ */
+ logicalrep_rel_close(rel, RowExclusiveLock);
+
+ /*
+ * In non-transactional case there won't be any other changes, so
+ * abort the whole transaction.
+ */
+ if (!seq.transactional)
+ AbortCurrentTransaction();
+
+ return;
+ }
/* apply the sequence change */
- SetSequence(relid, seq.transactional, seq.last_value, seq.log_cnt, seq.is_called);
+ SetSequence(RelationGetRelid(rel->localrel),
+ seq.transactional, seq.last_value, seq.log_cnt, seq.is_called);
+
+ logicalrep_rel_close(rel, NoLock);
/*
* Commit the per-stream transaction (we only do this when not in
diff --git a/src/backend/replication/pgoutput/pgoutput.c b/src/backend/replication/pgoutput/pgoutput.c
index 4cdc698cbb..366faf9564 100644
--- a/src/backend/replication/pgoutput/pgoutput.c
+++ b/src/backend/replication/pgoutput/pgoutput.c
@@ -568,9 +568,9 @@ pgoutput_rollback_prepared_txn(LogicalDecodingContext *ctx,
* done yet.
*/
static void
-maybe_send_schema(LogicalDecodingContext *ctx,
- ReorderBufferChange *change,
- Relation relation, RelationSyncEntry *relentry)
+maybe_send_schema_txn(LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn,
+ Relation relation, RelationSyncEntry *relentry)
{
bool schema_sent;
TransactionId xid = InvalidTransactionId;
@@ -585,10 +585,10 @@ maybe_send_schema(LogicalDecodingContext *ctx,
* the write methods will not include it.
*/
if (in_streaming)
- xid = change->txn->xid;
+ xid = txn->xid;
- if (change->txn->toptxn)
- topxid = change->txn->toptxn->xid;
+ if (txn->toptxn)
+ topxid = txn->toptxn->xid;
else
topxid = xid;
@@ -634,6 +634,14 @@ maybe_send_schema(LogicalDecodingContext *ctx,
relentry->schema_sent = true;
}
+static void
+maybe_send_schema(LogicalDecodingContext *ctx,
+ ReorderBufferChange *change,
+ Relation relation, RelationSyncEntry *relentry)
+{
+ maybe_send_schema_txn(ctx, change->txn, relation, relentry);
+}
+
/*
* Sends a relation
*/
@@ -1492,6 +1500,8 @@ pgoutput_sequence(LogicalDecodingContext *ctx,
if (!relentry->pubactions.pubsequence)
return;
+ maybe_send_schema_txn(ctx, txn, relation, relentry);
+
OutputPluginPrepareWrite(ctx, true);
logicalrep_write_sequence(ctx->out,
relation,
diff --git a/src/include/replication/logicalproto.h b/src/include/replication/logicalproto.h
index fb86ca022d..b7ab68d9a2 100644
--- a/src/include/replication/logicalproto.h
+++ b/src/include/replication/logicalproto.h
@@ -122,9 +122,10 @@ typedef struct LogicalRepTyp
/* Sequence info */
typedef struct LogicalRepSequence
{
- Oid remoteid; /* unique id of the remote sequence */
- char *nspname; /* schema name of remote sequence */
- char *seqname; /* name of the remote sequence */
+ // Oid remoteid; /* unique id of the remote sequence */
+ LogicalRepRelId remoteid; /* unique id of the sequence */
+ // char *nspname; /* schema name of remote sequence */
+ // char *seqname; /* name of the remote sequence */
bool transactional;
int64 last_value;
int64 log_cnt;
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2022-03-25 19:58 Tomas Vondra <[email protected]>
parent: vignesh C <[email protected]>
0 siblings, 0 replies; 127+ messages in thread
From: Tomas Vondra @ 2022-03-25 19:58 UTC (permalink / raw)
To: vignesh C <[email protected]>; +Cc: Petr Jelinek <[email protected]>; Amit Kapila <[email protected]>; Peter Eisentraut <[email protected]>; PostgreSQL Hackers <[email protected]>
On 3/25/22 15:34, vignesh C wrote:
> On Fri, Mar 25, 2022 at 3:29 AM Tomas Vondra
> <[email protected]> wrote:
>>
>> Hi,
>>
>> Pushed, after going through the patch once more, addressed the remaining
>> FIXMEs, corrected a couple places in the docs and comments, etc. Minor
>> tweaks, nothing important.
>
> While rebasing patch [1] I found a couple of comments:
> static void
> ObjectsInPublicationToOids(List *pubobjspec_list, ParseState *pstate,
> - List **rels, List **schemas)
> + List **tables, List **sequences,
> + List **tables_schemas, List **sequences_schemas,
> + List **schemas)
> {
> ListCell *cell;
> PublicationObjSpec *pubobj;
> @@ -185,12 +194,23 @@ ObjectsInPublicationToOids(List
> *pubobjspec_list, ParseState *pstate,
> switch (pubobj->pubobjtype)
> {
> case PUBLICATIONOBJ_TABLE:
> - *rels = lappend(*rels, pubobj->pubtable);
> + *tables = lappend(*tables, pubobj->pubtable);
> + break;
> + case PUBLICATIONOBJ_SEQUENCE:
> + *sequences = lappend(*sequences, pubobj->pubtable);
> break;
> case PUBLICATIONOBJ_TABLES_IN_SCHEMA:
> schemaid = get_namespace_oid(pubobj->name, false);
>
> /* Filter out duplicates if user specifies "sch1, sch1" */
> + *tables_schemas = list_append_unique_oid(*tables_schemas, schemaid);
> + *schemas = list_append_unique_oid(*schemas, schemaid);
> + break;
>
> Now tables_schemas and sequence_schemas are being updated and used in
> ObjectsInPublicationToOids, schema parameter is no longer being used
> after processing in ObjectsInPublicationToOids, I felt we can remove
> that parameter.
>
Thanks! That's a nice simplification, I'll get that pushed in a couple
minutes.
> /* ALTER PUBLICATION <name> ADD */
> else if (Matches("ALTER", "PUBLICATION", MatchAny, "ADD"))
> - COMPLETE_WITH("ALL TABLES IN SCHEMA", "TABLE");
> + COMPLETE_WITH("ALL TABLES IN SCHEMA", "ALL SEQUENCES IN SCHEMA",
> "TABLE", "SEQUENCE");
>
> Tab completion of alter publication for ADD and DROP is the same, we
> could combine it.
>
We could, but I find these combined rules harder to read, so I'll keep
the current tab-completion.
> Attached a patch for the same.
> Thoughts?
Thanks for taking a look! Appreciated.
>
> [1] - https://www.postgresql.org/message-id/CALDaNm3%3DJrucjhiiwsYQw5-PGtBHFONa6F7hhWCXMsGvh%3DtamA%40mail...
>
regars
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2022-03-25 20:10 Tomas Vondra <[email protected]>
parent: Amit Kapila <[email protected]>
1 sibling, 0 replies; 127+ messages in thread
From: Tomas Vondra @ 2022-03-25 20:10 UTC (permalink / raw)
To: Amit Kapila <[email protected]>; +Cc: Petr Jelinek <[email protected]>; Peter Eisentraut <[email protected]>; PostgreSQL Hackers <[email protected]>
Hi,
I've fixed most of the reported issues (or at least I think so), with
the exception of those in apply_handle_sequence function, i.e.:
1) properly coordinating with the tablesync worker
2) considering skip_lsn, skipping changes
3) missing privilege check, similar to TargetPrivilegesCheck
4) nicer error message if the sequence does not exist
The apply_handle_sequence stuff seems to be inter-related, so I plan to
deal with that in a single separate commit - the main part being the
tablesync coordination, per the fix I shared earlier today. But I need
time to think about that, I don't want to rush that.
Thanks for the feedback!
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2022-03-26 07:28 Amit Kapila <[email protected]>
parent: Tomas Vondra <[email protected]>
0 siblings, 1 reply; 127+ messages in thread
From: Amit Kapila @ 2022-03-26 07:28 UTC (permalink / raw)
To: Tomas Vondra <[email protected]>; +Cc: Petr Jelinek <[email protected]>; Peter Eisentraut <[email protected]>; PostgreSQL Hackers <[email protected]>
On Fri, Mar 25, 2022 at 10:20 PM Tomas Vondra
<[email protected]> wrote:
>
> Hmm, so fixing this might be a bit trickier than I expected.
>
> Firstly, currently we only send nspname/relname in the sequence message,
> not the remote OID or schema. The idea was that for sequences we don't
> really need schema info, so this seemed OK.
>
> But should_apply_changes_for_rel() needs LogicalRepRelMapEntry, and to
> create/maintain that those records we need to send the schema.
>
> Attached is a WIP patch does that.
>
> Two places need more work, I think:
>
> 1) maybe_send_schema needs ReorderBufferChange, but we don't have that
> for sequences, we only have TXN. I created a simple wrapper, but maybe
> we should just tweak maybe_send_schema to use TXN.
>
> 2) The transaction handling in is a bit confusing. The non-transactional
> increments won't have any explicit commit later, so we can't just rely
> on begin_replication_step/end_replication_step. But I want to try
> spending a bit more time on this.
>
I didn't understand what you want to say in point (2).
>
> But there's a more serious issue, I think. So far, we allowed this:
>
> BEGIN;
> CREATE SEQUENCE s2;
> ALTER PUBLICATION p ADD SEQUENCE s2;
> INSERT INTO seq_test SELECT nextval('s2') FROM generate_series(1,100);
> COMMIT;
>
> and the behavior was that we replicated the changes. But with the patch
> applied, that no longer happens, because should_apply_changes_for_rel
> says the change should not be applied.
>
> And after thinking about this, I think that's correct - we can't apply
> changes until ALTER SUBSCRIPTION ... REFRESH PUBLICATION gets executed,
> and we can't do that until the transaction commits.
>
> So I guess that's correct, and the current behavior is a bug.
>
Yes, I also think that is a bug.
> For a while I was thinking that maybe this means we don't need the
> transactional behavior at all, but I think we do - we have to handle
> ALTER SEQUENCE cases that are transactional.
>
I need some time to think about this. At all places, it is mentioned
as creating a sequence for transactional cases which at the very least
need some tweak.
--
With Regards,
Amit Kapila.
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2022-03-26 09:56 Tomas Vondra <[email protected]>
parent: Amit Kapila <[email protected]>
0 siblings, 1 reply; 127+ messages in thread
From: Tomas Vondra @ 2022-03-26 09:56 UTC (permalink / raw)
To: Amit Kapila <[email protected]>; +Cc: Petr Jelinek <[email protected]>; Peter Eisentraut <[email protected]>; PostgreSQL Hackers <[email protected]>
On 3/26/22 08:28, Amit Kapila wrote:
> On Fri, Mar 25, 2022 at 10:20 PM Tomas Vondra
> <[email protected]> wrote:
>>
>> Hmm, so fixing this might be a bit trickier than I expected.
>>
>> Firstly, currently we only send nspname/relname in the sequence message,
>> not the remote OID or schema. The idea was that for sequences we don't
>> really need schema info, so this seemed OK.
>>
>> But should_apply_changes_for_rel() needs LogicalRepRelMapEntry, and to
>> create/maintain that those records we need to send the schema.
>>
>> Attached is a WIP patch does that.
>>
>> Two places need more work, I think:
>>
>> 1) maybe_send_schema needs ReorderBufferChange, but we don't have that
>> for sequences, we only have TXN. I created a simple wrapper, but maybe
>> we should just tweak maybe_send_schema to use TXN.
>>
>> 2) The transaction handling in is a bit confusing. The non-transactional
>> increments won't have any explicit commit later, so we can't just rely
>> on begin_replication_step/end_replication_step. But I want to try
>> spending a bit more time on this.
>>
>
> I didn't understand what you want to say in point (2).
>
My point is that handle_apply_sequence() either needs to use the same
transaction handling as other apply methods, or start (and commit) a
separate transaction for the "transactional" case.
Which means we can't use the begin_replication_step/end_replication_step
and the current code seems a bit complex. And I'm not sure it's quite
correct. So this place needs more work.
>>
>> But there's a more serious issue, I think. So far, we allowed this:
>>
>> BEGIN;
>> CREATE SEQUENCE s2;
>> ALTER PUBLICATION p ADD SEQUENCE s2;
>> INSERT INTO seq_test SELECT nextval('s2') FROM generate_series(1,100);
>> COMMIT;
>>
>> and the behavior was that we replicated the changes. But with the patch
>> applied, that no longer happens, because should_apply_changes_for_rel
>> says the change should not be applied.
>>
>> And after thinking about this, I think that's correct - we can't apply
>> changes until ALTER SUBSCRIPTION ... REFRESH PUBLICATION gets executed,
>> and we can't do that until the transaction commits.
>>
>> So I guess that's correct, and the current behavior is a bug.
>>
>
> Yes, I also think that is a bug.
>
OK
>> For a while I was thinking that maybe this means we don't need the
>> transactional behavior at all, but I think we do - we have to handle
>> ALTER SEQUENCE cases that are transactional.
>>
>
> I need some time to think about this.
Understood.
> At all places, it is mentioned
> as creating a sequence for transactional cases which at the very least
> need some tweak.
>
Which places?
regards
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2022-03-28 05:29 Amit Kapila <[email protected]>
parent: Tomas Vondra <[email protected]>
0 siblings, 2 replies; 127+ messages in thread
From: Amit Kapila @ 2022-03-28 05:29 UTC (permalink / raw)
To: Tomas Vondra <[email protected]>; +Cc: Petr Jelinek <[email protected]>; Peter Eisentraut <[email protected]>; PostgreSQL Hackers <[email protected]>
On Sat, Mar 26, 2022 at 3:26 PM Tomas Vondra
<[email protected]> wrote:
>
> On 3/26/22 08:28, Amit Kapila wrote:
> >>
> >> 2) The transaction handling in is a bit confusing. The non-transactional
> >> increments won't have any explicit commit later, so we can't just rely
> >> on begin_replication_step/end_replication_step. But I want to try
> >> spending a bit more time on this.
> >>
> >
> > I didn't understand what you want to say in point (2).
> >
>
> My point is that handle_apply_sequence() either needs to use the same
> transaction handling as other apply methods, or start (and commit) a
> separate transaction for the "transactional" case.
>
> Which means we can't use the begin_replication_step/end_replication_step
>
We already call CommitTransactionCommand after end_replication_step at
a few places in that file so as there is no explicit commit in
non-transactional case, we can probably call CommitTransactionCommand
for it.
> and the current code seems a bit complex. And I'm not sure it's quite
> correct. So this place needs more work.
>
Agreed.
>
> >> For a while I was thinking that maybe this means we don't need the
> >> transactional behavior at all, but I think we do - we have to handle
> >> ALTER SEQUENCE cases that are transactional.
> >>
> >
> > I need some time to think about this.
>
While thinking about this, I think I see a problem with the
non-transactional handling of sequences. It seems that we will skip
sending non-transactional sequence change if it occurs before the
decoding has reached a consistent point but the surrounding commit
occurs after a consistent point is reached. In such cases, the
corresponding DMLs like inserts will be sent but sequence changes
won't be sent. For example (this scenario is based on
twophase_snapshot.spec),
Initial setup:
==============
Create table t1_seq(c1 int);
Create Sequence seq1;
Test Execution via multiple sessions (this test allows insert in
session-2 to happen before we reach a consistent point and commit
happens after a consistent point):
=======================================================================================================
Session-2:
Begin;
SELECT pg_current_xact_id();
Session-1:
SELECT 'init' FROM pg_create_logical_replication_slot('test_slot',
'test_decoding', false, true);
Session-3:
Begin;
SELECT pg_current_xact_id();
Session-2:
Commit;
Begin;
INSERT INTO t1_seq SELECT nextval('seq1') FROM generate_series(1,100);
Session-3:
Commit;
Session-2:
Commit 'foo'
Session-1:
SELECT data FROM pg_logical_slot_get_changes('test_slot', NULL, NULL,
'include-xids', 'false', 'skip-empty-xacts', '1');
data
----------------------------------------------
BEGIN
table public.t1_seq: INSERT: c1[integer]:1
table public.t1_seq: INSERT: c1[integer]:2
table public.t1_seq: INSERT: c1[integer]:3
table public.t1_seq: INSERT: c1[integer]:4
table public.t1_seq: INSERT: c1[integer]:5
table public.t1_seq: INSERT: c1[integer]:6
Now, if we normally try to decode such an insert, the result would be
something like:
data
------------------------------------------------------------------------------
sequence public.seq1: transactional:0 last_value: 33 log_cnt: 0 is_called:1
sequence public.seq1: transactional:0 last_value: 66 log_cnt: 0 is_called:1
sequence public.seq1: transactional:0 last_value: 99 log_cnt: 0 is_called:1
sequence public.seq1: transactional:0 last_value: 132 log_cnt: 0 is_called:1
BEGIN
table public.t1_seq: INSERT: c1[integer]:1
table public.t1_seq: INSERT: c1[integer]:2
table public.t1_seq: INSERT: c1[integer]:3
table public.t1_seq: INSERT: c1[integer]:4
table public.t1_seq: INSERT: c1[integer]:5
table public.t1_seq: INSERT: c1[integer]:6
This will create an inconsistent replica as sequence changes won't be
replicated. I thought about changing snapshot dealing of
non-transactional sequence changes similar to transactional ones but
that also won't work because it is only at commit we decide whether we
can send the changes.
For the transactional case, as we are considering the create sequence
operation as transactional, we would unnecessarily queue them even
though that is not required. Basically, they don't need to be
considered transactional and we can simply ignore such messages like
other DDLs. But for that probably we need to distinguish Alter/Create
case which may or may not be straightforward. Now, queuing them is
probably harmless unless it causes the transaction to spill/stream.
I still couldn't think completely about cases where a mix of
transactional and non-transactional changes occur in the same
transaction as I think it somewhat depends on what we want to do about
the above cases.
> > At all places, it is mentioned
> > as creating a sequence for transactional cases which at the very least
> > need some tweak.
> >
>
> Which places?
>
In comments like:
a. When decoding sequences, we differentiate between sequences created
in a (running) transaction and sequences created in other (already
committed) transactions.
b. ... But for new sequences, we need to handle them in a transactional way, ..
c. ... Change needs to be handled as transactional, because the
sequence was created in a transaction that is still running ...
It seems all these places indicate a scenario of creating a sequence
whereas we want to do transactional stuff mainly for Alter.
--
With Regards,
Amit Kapila.
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2022-05-25 20:42 Zheng Li <[email protected]>
parent: Amit Kapila <[email protected]>
1 sibling, 0 replies; 127+ messages in thread
From: Zheng Li @ 2022-05-25 20:42 UTC (permalink / raw)
To: Tomas Vondra <[email protected]>; +Cc: Amit Kapila <[email protected]>; Petr Jelinek <[email protected]>; Peter Eisentraut <[email protected]>; PostgreSQL Hackers <[email protected]>
> But of course, if we expect/require to have a perfect snapshot for that
> exact position in the transaction, this won't work. IMO the whole idea
> that we can have non-transactional bits in naturally transactional
> decoding seems a bit suspicious (at least in hindsight).
>
> No matter what we do for sequences, though, this still affects logical
> messages too. Not sure what to do there :-(
Hi, I spent some time trying to understand this problem while I was
evaluating its impact on the DDL replication in [1]. I think for DDL
we could always remove the
non-transactional bits since DDL will probably always be processed
transactionally.
I attempted to solve the problem for messages. Here is a potential
solution by keeping track of
the last decoded/acked non-transactional message/operation lsn and use
it to check if a non-transactional message record should be skipped
during decoding,
to do that I added new fields
ReplicationSlotPersistentData.non_xact_op_at,
XLogReaderState.NonXactOpRecPtr and
SnapBuild.start_decoding_nonxactop_at.
This is the end LSN of the last non-transactional message/operation
decoded/acked. I verified this approach solves the issue of
missing decoding of non-transactional messages under
concurrency/before the builder state reaches SNAPBUILD_CONSISTENT.
Once
the builder state reach SNAPBUILD_CONSISTENT, the new field
ReplicationSlotPersistentData.non_xact_op_at can be set
to ReplicationSlotPersistentData.confirmed_flush.
Similar to the sequence issue, here is the test case for logical messages:
Test concurrent execution in 3 sessions that allows pg_logical_emit_message in
session-2 to happen before we reach a consistent point and commit
happens after a consistent point:
Session-2:
Begin;
SELECT pg_current_xact_id();
Session-1:
SELECT 'init' FROM pg_create_logical_replication_slot('test_slot',
'test_decoding', false, true);
Session-3:
Begin;
SELECT pg_current_xact_id();
Session-2:
Commit;
Begin;
SELECT pg_logical_emit_message(true, 'test_decoding', 'msg1');
SELECT pg_logical_emit_message(false, 'test_decoding', 'msg2');
Session-3:
Commit;
Session-1: (at this point, the session will crash without the fix)
SELECT data FROM pg_logical_slot_get_changes('test_slot', NULL, NULL,
'force-binary', '0', 'skip-empty-xacts', '1');
data
---------------------------------------------------------------------
message: transactional: 0 prefix: test_decoding, sz: 4 content:msg1
Session-2:
Commit;
Session-1:
SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL,
NULL, 'force-binary', '0', 'skip-empty-xacts', '1');
data
---------------------------------------------------------------------
message: transactional: 1 prefix: test_decoding, sz: 4 content:msg2
I also tried the same approach on sequences (on a commit before the
revert of sequence replication) and it seems to be working but
I think it needs further testing.
Patch 0001-Intorduce-new-field-ReplicationSlotPersistentData.no.patch
applies on master which contains the fix for logical messages.
[1] https://www.postgresql.org/message-id/flat/CAAD30U+pVmfKwUKy8cbZOnUXyguJ-uBNejwD75Kyo=OjdQGJ9g@mail....
Thoughts?
With Regards,
Zheng
Attachments:
[application/octet-stream] 0001-Intorduce-new-field-ReplicationSlotPersistentData.no.patch (12.9K, ../../CAAD30U+FD+QprMqY8wPo2n+3eS7B8=f2XYNjrUC9Sq-OFp6zDA@mail.gmail.com/2-0001-Intorduce-new-field-ReplicationSlotPersistentData.no.patch)
download | inline diff:
From 34621aef44b38af447c670f94da9e703bdd495fa Mon Sep 17 00:00:00 2001
From: "Zheng (Zane) Li" <[email protected]>
Date: Fri, 20 May 2022 19:55:50 +0000
Subject: [PATCH] Intorduce new field
ReplicationSlotPersistentData.non_xact_op_at, XLogReaderState.NonXactOpRecPtr
and SnapBuild.start_decoding_nonxactop_at. This is the end lsn of last
non-transactional message/operation decoded. We introduce this new field in
an attempt to solve the issue of missing decoding of non-transactional
message/operation under concurrency/before build state reach
SNAPBUILD_CONSISTENT. Once the build state reach SNAPBUILD_CONSISTENT,
non_xact_op_at can be set to ReplicationSlotPersistentData.confirmed_flush.
---
src/backend/replication/logical/decode.c | 18 ++++++++++---
src/backend/replication/logical/logical.c | 25 +++++++++++++++---
.../replication/logical/logicalfuncs.c | 2 +-
src/backend/replication/logical/snapbuild.c | 26 +++++++++++++++++--
src/backend/replication/slotfuncs.c | 2 +-
src/backend/replication/walsender.c | 2 +-
src/include/access/xlogreader.h | 3 +++
src/include/replication/logical.h | 2 +-
src/include/replication/slot.h | 9 +++++++
src/include/replication/snapbuild.h | 5 +++-
10 files changed, 80 insertions(+), 14 deletions(-)
diff --git a/src/backend/replication/logical/decode.c b/src/backend/replication/logical/decode.c
index aa2427ba73..a0b700a662 100644
--- a/src/backend/replication/logical/decode.c
+++ b/src/backend/replication/logical/decode.c
@@ -589,12 +589,24 @@ logicalmsg_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
if (message->transactional &&
!SnapBuildProcessChange(builder, xid, buf->origptr))
return;
+
else if (!message->transactional &&
- (SnapBuildCurrentState(builder) != SNAPBUILD_CONSISTENT ||
- SnapBuildXactNeedsSkip(builder, buf->origptr)))
+ ((SnapBuildCurrentState(builder) < SNAPBUILD_FULL_SNAPSHOT) ||
+ /* Can't prcess non-transactional msg immediately if consumer functions are not provided */
+ ctx->prepare_write == NULL ||
+ SnapBuildNonXactOpNeedsSkip(builder, buf->origptr)))
return;
- snapshot = SnapBuildGetOrBuildSnapshot(builder, xid);
+ /*
+ * Build or get snapshot for non-transactional msg for immedieate processing.
+ * Transactional msg will be queued and processed upon decoding of the commit
+ * record.
+ */
+ if (!message->transactional)
+ {
+ snapshot = SnapBuildGetOrBuildSnapshot(builder, xid);
+ }
+
ReorderBufferQueueMessage(ctx->reorder, xid, snapshot, buf->endptr,
message->transactional,
message->message, /* first part of message is
diff --git a/src/backend/replication/logical/logical.c b/src/backend/replication/logical/logical.c
index 625a7f4273..302a61efb8 100644
--- a/src/backend/replication/logical/logical.c
+++ b/src/backend/replication/logical/logical.c
@@ -208,7 +208,8 @@ StartupDecodingContext(List *output_plugin_options,
ctx->reorder = ReorderBufferAllocate();
ctx->snapshot_builder =
AllocateSnapshotBuilder(ctx->reorder, xmin_horizon, start_lsn,
- need_full_snapshot, slot->data.two_phase_at);
+ need_full_snapshot, slot->data.two_phase_at,
+ slot->data.non_xact_op_at);
ctx->reorder->private_data = ctx;
@@ -1216,6 +1217,12 @@ message_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
ctx->callbacks.message_cb(ctx, txn, message_lsn, transactional, prefix,
message_size, message);
+ if (!transactional)
+ {
+ SnapBuildSetNonXactOpAt(ctx->snapshot_builder, message_lsn);
+ ctx->reader->NonXactOpRecPtr = message_lsn;
+ }
+
/* Pop the error context stack */
error_context_stack = errcallback.previous;
}
@@ -1531,6 +1538,12 @@ stream_message_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
ctx->callbacks.stream_message_cb(ctx, txn, message_lsn, transactional, prefix,
message_size, message);
+ if (!transactional)
+ {
+ SnapBuildSetNonXactOpAt(ctx->snapshot_builder, message_lsn);
+ ctx->reader->NonXactOpRecPtr = message_lsn;
+ }
+
/* Pop the error context stack */
error_context_stack = errcallback.previous;
}
@@ -1648,7 +1661,7 @@ LogicalIncreaseXminForSlot(XLogRecPtr current_lsn, TransactionId xmin)
/* candidate already valid with the current flush position, apply */
if (updated_xmin)
- LogicalConfirmReceivedLocation(slot->data.confirmed_flush);
+ LogicalConfirmReceivedLocation(slot->data.confirmed_flush, InvalidXLogRecPtr);
}
/*
@@ -1726,14 +1739,14 @@ LogicalIncreaseRestartDecodingForSlot(XLogRecPtr current_lsn, XLogRecPtr restart
/* candidates are already valid with the current flush position, apply */
if (updated_lsn)
- LogicalConfirmReceivedLocation(slot->data.confirmed_flush);
+ LogicalConfirmReceivedLocation(slot->data.confirmed_flush, InvalidXLogRecPtr);
}
/*
* Handle a consumer's confirmation having received all changes up to lsn.
*/
void
-LogicalConfirmReceivedLocation(XLogRecPtr lsn)
+LogicalConfirmReceivedLocation(XLogRecPtr lsn, XLogRecPtr lsn_nonxact)
{
Assert(lsn != InvalidXLogRecPtr);
@@ -1747,6 +1760,8 @@ LogicalConfirmReceivedLocation(XLogRecPtr lsn)
SpinLockAcquire(&MyReplicationSlot->mutex);
MyReplicationSlot->data.confirmed_flush = lsn;
+ if (lsn_nonxact != InvalidXLogRecPtr)
+ MyReplicationSlot->data.non_xact_op_at = lsn_nonxact;
/* if we're past the location required for bumping xmin, do so */
if (MyReplicationSlot->candidate_xmin_lsn != InvalidXLogRecPtr &&
@@ -1812,6 +1827,8 @@ LogicalConfirmReceivedLocation(XLogRecPtr lsn)
{
SpinLockAcquire(&MyReplicationSlot->mutex);
MyReplicationSlot->data.confirmed_flush = lsn;
+ if (lsn_nonxact != InvalidXLogRecPtr)
+ MyReplicationSlot->data.non_xact_op_at = lsn_nonxact;
SpinLockRelease(&MyReplicationSlot->mutex);
}
}
diff --git a/src/backend/replication/logical/logicalfuncs.c b/src/backend/replication/logical/logicalfuncs.c
index 6058d36e0d..5d579218d8 100644
--- a/src/backend/replication/logical/logicalfuncs.c
+++ b/src/backend/replication/logical/logicalfuncs.c
@@ -294,7 +294,7 @@ pg_logical_slot_get_changes_guts(FunctionCallInfo fcinfo, bool confirm, bool bin
*/
if (ctx->reader->EndRecPtr != InvalidXLogRecPtr && confirm)
{
- LogicalConfirmReceivedLocation(ctx->reader->EndRecPtr);
+ LogicalConfirmReceivedLocation(ctx->reader->EndRecPtr, ctx->reader->NonXactOpRecPtr);
/*
* If only the confirmed_flush_lsn has changed the slot won't get
diff --git a/src/backend/replication/logical/snapbuild.c b/src/backend/replication/logical/snapbuild.c
index 1119a12db9..d5d378836a 100644
--- a/src/backend/replication/logical/snapbuild.c
+++ b/src/backend/replication/logical/snapbuild.c
@@ -175,6 +175,8 @@ struct SnapBuild
*/
XLogRecPtr two_phase_at;
+ XLogRecPtr start_decoding_nonxactop_at;
+
/*
* Don't start decoding WAL until the "xl_running_xacts" information
* indicates there are no running xids with an xid smaller than this.
@@ -281,7 +283,8 @@ AllocateSnapshotBuilder(ReorderBuffer *reorder,
TransactionId xmin_horizon,
XLogRecPtr start_lsn,
bool need_full_snapshot,
- XLogRecPtr two_phase_at)
+ XLogRecPtr two_phase_at,
+ XLogRecPtr non_xact_op_at)
{
MemoryContext context;
MemoryContext oldcontext;
@@ -310,6 +313,7 @@ AllocateSnapshotBuilder(ReorderBuffer *reorder,
builder->start_decoding_at = start_lsn;
builder->building_full_snapshot = need_full_snapshot;
builder->two_phase_at = two_phase_at;
+ builder->start_decoding_nonxactop_at = non_xact_op_at+1;
MemoryContextSwitchTo(oldcontext);
@@ -387,6 +391,15 @@ SnapBuildSetTwoPhaseAt(SnapBuild *builder, XLogRecPtr ptr)
builder->two_phase_at = ptr;
}
+/*
+ * Set the LSN at which non-transactional op decoding is at.
+ */
+void
+SnapBuildSetNonXactOpAt(SnapBuild *builder, XLogRecPtr ptr)
+{
+ builder->start_decoding_nonxactop_at = ptr;
+}
+
/*
* Should the contents of transaction ending at 'ptr' be decoded?
*/
@@ -396,6 +409,15 @@ SnapBuildXactNeedsSkip(SnapBuild *builder, XLogRecPtr ptr)
return ptr < builder->start_decoding_at;
}
+/*
+ * Should the contents of a non-transactional op ending at 'ptr' be decoded?
+ */
+bool
+SnapBuildNonXactOpNeedsSkip(SnapBuild *builder, XLogRecPtr ptr)
+{
+ return ptr < builder->start_decoding_nonxactop_at;
+}
+
/*
* Increase refcount of a snapshot.
*
@@ -661,7 +683,7 @@ SnapBuildExportSnapshot(SnapBuild *builder)
Snapshot
SnapBuildGetOrBuildSnapshot(SnapBuild *builder, TransactionId xid)
{
- Assert(builder->state == SNAPBUILD_CONSISTENT);
+ Assert(builder->state >= SNAPBUILD_FULL_SNAPSHOT);
/* only build a new snapshot if we don't have a prebuilt one */
if (builder->snapshot == NULL)
diff --git a/src/backend/replication/slotfuncs.c b/src/backend/replication/slotfuncs.c
index ca945994ef..3bbbe74058 100644
--- a/src/backend/replication/slotfuncs.c
+++ b/src/backend/replication/slotfuncs.c
@@ -530,7 +530,7 @@ pg_logical_replication_slot_advance(XLogRecPtr moveto)
if (ctx->reader->EndRecPtr != InvalidXLogRecPtr)
{
- LogicalConfirmReceivedLocation(moveto);
+ LogicalConfirmReceivedLocation(moveto, moveto);
/*
* If only the confirmed_flush LSN has changed the slot won't get
diff --git a/src/backend/replication/walsender.c b/src/backend/replication/walsender.c
index e42671722a..b1373e52fd 100644
--- a/src/backend/replication/walsender.c
+++ b/src/backend/replication/walsender.c
@@ -2170,7 +2170,7 @@ ProcessStandbyReplyMessage(void)
if (MyReplicationSlot && flushPtr != InvalidXLogRecPtr)
{
if (SlotIsLogical(MyReplicationSlot))
- LogicalConfirmReceivedLocation(flushPtr);
+ LogicalConfirmReceivedLocation(flushPtr, flushPtr);
else
PhysicalConfirmReceivedLocation(flushPtr);
}
diff --git a/src/include/access/xlogreader.h b/src/include/access/xlogreader.h
index e73ea4a840..b197a6a354 100644
--- a/src/include/access/xlogreader.h
+++ b/src/include/access/xlogreader.h
@@ -206,6 +206,9 @@ struct XLogReaderState
XLogRecPtr ReadRecPtr; /* start of last record read */
XLogRecPtr EndRecPtr; /* end+1 of last record read */
+ /* end of last non-transactional op record read */
+ XLogRecPtr NonXactOpRecPtr;
+
/*
* Set at the end of recovery: the start point of a partial record at the
* end of WAL (InvalidXLogRecPtr if there wasn't one), and the start
diff --git a/src/include/replication/logical.h b/src/include/replication/logical.h
index edadacd589..c9714b51f0 100644
--- a/src/include/replication/logical.h
+++ b/src/include/replication/logical.h
@@ -136,7 +136,7 @@ extern void FreeDecodingContext(LogicalDecodingContext *ctx);
extern void LogicalIncreaseXminForSlot(XLogRecPtr lsn, TransactionId xmin);
extern void LogicalIncreaseRestartDecodingForSlot(XLogRecPtr current_lsn,
XLogRecPtr restart_lsn);
-extern void LogicalConfirmReceivedLocation(XLogRecPtr lsn);
+extern void LogicalConfirmReceivedLocation(XLogRecPtr lsn, XLogRecPtr lsn_nonxact);
extern bool filter_prepare_cb_wrapper(LogicalDecodingContext *ctx,
TransactionId xid, const char *gid);
diff --git a/src/include/replication/slot.h b/src/include/replication/slot.h
index 8c9f3321d5..941c6f6da3 100644
--- a/src/include/replication/slot.h
+++ b/src/include/replication/slot.h
@@ -83,6 +83,15 @@ typedef struct ReplicationSlotPersistentData
*/
XLogRecPtr confirmed_flush;
+ /*
+ * Oldest LSN that the client has acked receipt for a non-transactional
+ * operation such as a non-transactional message. This is introduced
+ * to slove missing non-transactional decoding when creating logical
+ * replication slot under concurrency. Under normal situation
+ * (SNAPBUILD_CONSISTENT) we can make non_xact_op_at = confirmed_flush
+ */
+ XLogRecPtr non_xact_op_at;
+
/*
* LSN at which we enabled two_phase commit for this slot or LSN at which
* we found a consistent point at the time of slot creation.
diff --git a/src/include/replication/snapbuild.h b/src/include/replication/snapbuild.h
index d179251aad..7b055dba7c 100644
--- a/src/include/replication/snapbuild.h
+++ b/src/include/replication/snapbuild.h
@@ -62,7 +62,8 @@ extern void CheckPointSnapBuild(void);
extern SnapBuild *AllocateSnapshotBuilder(struct ReorderBuffer *cache,
TransactionId xmin_horizon, XLogRecPtr start_lsn,
bool need_full_snapshot,
- XLogRecPtr two_phase_at);
+ XLogRecPtr two_phase_at,
+ XLogRecPtr non_xact_op_at);
extern void FreeSnapshotBuilder(SnapBuild *cache);
extern void SnapBuildSnapDecRefcount(Snapshot snap);
@@ -77,8 +78,10 @@ extern Snapshot SnapBuildGetOrBuildSnapshot(SnapBuild *builder,
TransactionId xid);
extern bool SnapBuildXactNeedsSkip(SnapBuild *snapstate, XLogRecPtr ptr);
+extern bool SnapBuildNonXactOpNeedsSkip(SnapBuild *builder, XLogRecPtr ptr);
extern XLogRecPtr SnapBuildGetTwoPhaseAt(SnapBuild *builder);
extern void SnapBuildSetTwoPhaseAt(SnapBuild *builder, XLogRecPtr ptr);
+extern void SnapBuildSetNonXactOpAt(SnapBuild *builder, XLogRecPtr ptr);
extern void SnapBuildCommitTxn(SnapBuild *builder, XLogRecPtr lsn,
TransactionId xid, int nsubxacts,
--
2.32.0
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2022-08-07 00:36 Noah Misch <[email protected]>
parent: Amit Kapila <[email protected]>
1 sibling, 1 reply; 127+ messages in thread
From: Noah Misch @ 2022-08-07 00:36 UTC (permalink / raw)
To: Tomas Vondra <[email protected]>; +Cc: Amit Kapila <[email protected]>; Petr Jelinek <[email protected]>; Peter Eisentraut <[email protected]>; PostgreSQL Hackers <[email protected]>
On Thu, Apr 07, 2022 at 08:34:50PM +0200, Tomas Vondra wrote:
> I've pushed a revert af all the commits related to this - decoding of
> sequences and test_decoding / built-in replication changes.
Two July buildfarm runs failed with PANIC during standby promotion:
https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=tern&dt=2022-07-19%2004%3A13%3A18
https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=tern&dt=2022-07-31%2011%3A33%3A13
The attached patch hacks things so an ordinary x86_64 GNU/Linux machine
reproduces this consistently. "git bisect" then traced the regression to the
above revert commit (2c7ea57e56ca5f668c32d4266e0a3e45b455bef5). The pg_ctl
test suite passes under this hack in all supported branches, and it passed on
v15 until that revert. Would you investigate?
The buildfarm animal uses keep_error_builds. From kept data directories, I
deduced these events:
- After the base backup, auto-analyze ran on the primary and wrote WAL.
- Standby streamed and wrote up to 0/301FFF.
- Standby received the promote signal. Terminated streaming. WAL page at 0/302000 remained all-zeros.
- Somehow, end-of-recovery became a PANIC.
Key portions from buildfarm logs:
=== good run standby2 log
2022-07-21 22:55:16.860 UTC [25034912:5] LOG: received promote request
2022-07-21 22:55:16.878 UTC [26804682:2] FATAL: terminating walreceiver process due to administrator command
2022-07-21 22:55:16.878 UTC [25034912:6] LOG: invalid record length at 0/3000060: wanted 24, got 0
2022-07-21 22:55:16.878 UTC [25034912:7] LOG: redo done at 0/3000028 system usage: CPU: user: 0.00 s, system: 0.00 s, elapsed: 0.42 s
2022-07-21 22:55:16.878 UTC [25034912:8] LOG: selected new timeline ID: 2
2022-07-21 22:55:17.004 UTC [25034912:9] LOG: archive recovery complete
2022-07-21 22:55:17.005 UTC [23724044:1] LOG: checkpoint starting: force
2022-07-21 22:55:17.008 UTC [14549364:4] LOG: database system is ready to accept connections
2022-07-21 22:55:17.093 UTC [23724044:2] LOG: checkpoint complete: wrote 3 buffers (2.3%); 0 WAL file(s) added, 0 removed, 1 recycled; write=0.019 s, sync=0.001 s, total=0.089 s; sync files=0, longest=0.000 s, average=0.000 s; distance=16384 kB, estimate=16384 kB
2022-07-21 22:55:17.143 UTC [27394418:1] [unknown] LOG: connection received: host=[local]
2022-07-21 22:55:17.144 UTC [27394418:2] [unknown] LOG: connection authorized: user=nm database=postgres application_name=003_promote.pl
2022-07-21 22:55:17.147 UTC [27394418:3] 003_promote.pl LOG: statement: SELECT pg_is_in_recovery()
2022-07-21 22:55:17.148 UTC [27394418:4] 003_promote.pl LOG: disconnection: session time: 0:00:00.005 user=nm database=postgres host=[local]
2022-07-21 22:55:58.301 UTC [14549364:5] LOG: received immediate shutdown request
2022-07-21 22:55:58.337 UTC [14549364:6] LOG: database system is shut down
=== failed run standby2 log, with my annotations
2022-07-19 05:28:22.136 UTC [7340406:5] LOG: received promote request
2022-07-19 05:28:22.163 UTC [8519860:2] FATAL: terminating walreceiver process due to administrator command
2022-07-19 05:28:22.166 UTC [7340406:6] LOG: invalid magic number 0000 in log segment 000000010000000000000003, offset 131072
New compared to the good run. XLOG_PAGE_MAGIC didn't match. This implies the WAL ended at a WAL page boundary.
2022-07-19 05:28:22.166 UTC [7340406:7] LOG: redo done at 0/301F168 system usage: CPU: user: 0.00 s, system: 0.00 s, elapsed: 0.18 s
2022-07-19 05:28:22.166 UTC [7340406:8] LOG: last completed transaction was at log time 2022-07-19 05:28:13.956716+00
New compared to the good run. The good run had no transactions to replay. The bad run replayed records from an auto-analyze.
2022-07-19 05:28:22.166 UTC [7340406:9] PANIC: invalid record length at 0/301F168: wanted 24, got 0
More WAL overall in bad run, due to auto-analyze. End of recovery wrongly considered a PANIC.
2022-07-19 05:28:22.583 UTC [8388800:4] LOG: startup process (PID 7340406) was terminated by signal 6: IOT/Abort trap
2022-07-19 05:28:22.584 UTC [8388800:5] LOG: terminating any other active server processes
2022-07-19 05:28:22.587 UTC [8388800:6] LOG: shutting down due to startup process failure
2022-07-19 05:28:22.627 UTC [8388800:7] LOG: database system is shut down
Let me know if I've left out details you want; I may be able to dig more out
of the buildfarm artifacts.
Author: Noah Misch <[email protected]>
Commit: Noah Misch <[email protected]>
diff --git a/src/backend/replication/walreceiver.c b/src/backend/replication/walreceiver.c
index 8604fd4..6232237 100644
--- a/src/backend/replication/walreceiver.c
+++ b/src/backend/replication/walreceiver.c
@@ -885,6 +885,12 @@ XLogWalRcvWrite(char *buf, Size nbytes, XLogRecPtr recptr, TimeLineID tli)
{
int startoff;
int byteswritten;
+ bool spin_after = false;
+ /* force writing to end at a certain WAL page boundary */
+ if (recptr + nbytes > 0x03022000) {
+ nbytes = 0x03022000 - recptr;
+ spin_after = true;
+ }
Assert(tli != 0);
@@ -943,6 +949,8 @@ XLogWalRcvWrite(char *buf, Size nbytes, XLogRecPtr recptr, TimeLineID tli)
LogstreamResult.Write = recptr;
}
+ while (spin_after) /* we forced writing to end; wait for termination */
+ ProcessWalRcvInterrupts();
/* Update shared-memory status */
pg_atomic_write_u64(&WalRcv->writtenUpto, LogstreamResult.Write);
diff --git a/src/bin/pg_ctl/t/003_promote.pl b/src/bin/pg_ctl/t/003_promote.pl
index 84d28f4..57ae317 100644
--- a/src/bin/pg_ctl/t/003_promote.pl
+++ b/src/bin/pg_ctl/t/003_promote.pl
@@ -29,6 +29,7 @@ command_fails_like(
[ 'pg_ctl', '-D', $node_primary->data_dir, 'promote' ],
qr/not in standby mode/,
'pg_ctl promote of primary instance fails');
+if (1) { ok(1) for 1 .. 3; } else { # skip tests this way to avoid merge conflicts
my $node_standby = PostgreSQL::Test::Cluster->new('standby');
$node_primary->backup('my_backup');
@@ -46,12 +47,17 @@ ok( $node_standby->poll_query_until(
'postgres', 'SELECT NOT pg_is_in_recovery()'),
'promoted standby is not in recovery');
+}
+my $node_standby;
+$node_primary->backup('my_backup');
# same again with default wait option
$node_standby = PostgreSQL::Test::Cluster->new('standby2');
$node_standby->init_from_backup($node_primary, 'my_backup',
has_streaming => 1);
$node_standby->start;
+$node_primary->safe_psql('postgres', 'ANALYZE');
+
is($node_standby->safe_psql('postgres', 'SELECT pg_is_in_recovery()'),
't', 'standby is in recovery');
Attachments:
[text/plain] promote-invalid-magic-panic-v0.patch (2.1K, ../../[email protected]/2-promote-invalid-magic-panic-v0.patch)
download | inline diff:
Author: Noah Misch <[email protected]>
Commit: Noah Misch <[email protected]>
diff --git a/src/backend/replication/walreceiver.c b/src/backend/replication/walreceiver.c
index 8604fd4..6232237 100644
--- a/src/backend/replication/walreceiver.c
+++ b/src/backend/replication/walreceiver.c
@@ -885,6 +885,12 @@ XLogWalRcvWrite(char *buf, Size nbytes, XLogRecPtr recptr, TimeLineID tli)
{
int startoff;
int byteswritten;
+ bool spin_after = false;
+ /* force writing to end at a certain WAL page boundary */
+ if (recptr + nbytes > 0x03022000) {
+ nbytes = 0x03022000 - recptr;
+ spin_after = true;
+ }
Assert(tli != 0);
@@ -943,6 +949,8 @@ XLogWalRcvWrite(char *buf, Size nbytes, XLogRecPtr recptr, TimeLineID tli)
LogstreamResult.Write = recptr;
}
+ while (spin_after) /* we forced writing to end; wait for termination */
+ ProcessWalRcvInterrupts();
/* Update shared-memory status */
pg_atomic_write_u64(&WalRcv->writtenUpto, LogstreamResult.Write);
diff --git a/src/bin/pg_ctl/t/003_promote.pl b/src/bin/pg_ctl/t/003_promote.pl
index 84d28f4..57ae317 100644
--- a/src/bin/pg_ctl/t/003_promote.pl
+++ b/src/bin/pg_ctl/t/003_promote.pl
@@ -29,6 +29,7 @@ command_fails_like(
[ 'pg_ctl', '-D', $node_primary->data_dir, 'promote' ],
qr/not in standby mode/,
'pg_ctl promote of primary instance fails');
+if (1) { ok(1) for 1 .. 3; } else { # skip tests this way to avoid merge conflicts
my $node_standby = PostgreSQL::Test::Cluster->new('standby');
$node_primary->backup('my_backup');
@@ -46,12 +47,17 @@ ok( $node_standby->poll_query_until(
'postgres', 'SELECT NOT pg_is_in_recovery()'),
'promoted standby is not in recovery');
+}
+my $node_standby;
+$node_primary->backup('my_backup');
# same again with default wait option
$node_standby = PostgreSQL::Test::Cluster->new('standby2');
$node_standby->init_from_backup($node_primary, 'my_backup',
has_streaming => 1);
$node_standby->start;
+$node_primary->safe_psql('postgres', 'ANALYZE');
+
is($node_standby->safe_psql('postgres', 'SELECT pg_is_in_recovery()'),
't', 'standby is in recovery');
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2022-08-07 13:18 Tomas Vondra <[email protected]>
parent: Noah Misch <[email protected]>
0 siblings, 1 reply; 127+ messages in thread
From: Tomas Vondra @ 2022-08-07 13:18 UTC (permalink / raw)
To: Noah Misch <[email protected]>; +Cc: Amit Kapila <[email protected]>; Petr Jelinek <[email protected]>; Peter Eisentraut <[email protected]>; PostgreSQL Hackers <[email protected]>; Thomas Munro <[email protected]>
On 8/7/22 02:36, Noah Misch wrote:
> On Thu, Apr 07, 2022 at 08:34:50PM +0200, Tomas Vondra wrote:
>> I've pushed a revert af all the commits related to this - decoding of
>> sequences and test_decoding / built-in replication changes.
>
> Two July buildfarm runs failed with PANIC during standby promotion:
> https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=tern&dt=2022-07-19%2004%3A13%3A18
> https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=tern&dt=2022-07-31%2011%3A33%3A13
>
> The attached patch hacks things so an ordinary x86_64 GNU/Linux machine
> reproduces this consistently. "git bisect" then traced the regression to the
> above revert commit (2c7ea57e56ca5f668c32d4266e0a3e45b455bef5). The pg_ctl
> test suite passes under this hack in all supported branches, and it passed on
> v15 until that revert. Would you investigate?
>
> The buildfarm animal uses keep_error_builds. From kept data directories, I
> deduced these events:
>
> - After the base backup, auto-analyze ran on the primary and wrote WAL.
> - Standby streamed and wrote up to 0/301FFF.
> - Standby received the promote signal. Terminated streaming. WAL page at 0/302000 remained all-zeros.
> - Somehow, end-of-recovery became a PANIC.
>
I think it'd be really bizarre if this was due to the revert, as that
simply undoes minor WAL changes (and none of this should affect what
happens at WAL page boundary etc.). It just restores WAL as it was
before 0da92dc, nothing particularly complicated. I did go through all
of the changes again and I haven't spotted anything particularly
suspicious, but I'll give it another try tomorrow.
However, I did try bisecting this using the attached patch, and that
does not suggest the issue is in the revert commit. It actually fails
all the way back to 5dc0418fab2, and it starts working on 9553b4115f1.
...
6392f2a0968 Try to silence "-Wmissing-braces" complaints in ...
=> 5dc0418fab2 Prefetch data referenced by the WAL, take II.
9553b4115f1 Fix warning introduced in 5c279a6d350.
...
This is merely 10 commits before the revert, and it seems way more
related to WAL. Also, adding this to the two nodes in 003_standby.pl
makes the issue go away, it seems:
$node_standby->append_conf('postgresql.conf',
qq(recovery_prefetch = off));
I'd bet it's about WAL prefetching, not the revert, and the bisect was a
bit incorrect, because the commits are close and the failures happen to
be rare. (Presumably you first did the bisect and then wrote the patch
that reproduces this, right?)
Adding Thomas Munro to the thread, he's the WAL prefetching expert ;-)
regards
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2022-08-07 19:12 Noah Misch <[email protected]>
parent: Tomas Vondra <[email protected]>
0 siblings, 1 reply; 127+ messages in thread
From: Noah Misch @ 2022-08-07 19:12 UTC (permalink / raw)
To: Tomas Vondra <[email protected]>; +Cc: Amit Kapila <[email protected]>; Petr Jelinek <[email protected]>; Peter Eisentraut <[email protected]>; PostgreSQL Hackers <[email protected]>; Thomas Munro <[email protected]>
On Sun, Aug 07, 2022 at 03:18:52PM +0200, Tomas Vondra wrote:
> On 8/7/22 02:36, Noah Misch wrote:
> > On Thu, Apr 07, 2022 at 08:34:50PM +0200, Tomas Vondra wrote:
> >> I've pushed a revert af all the commits related to this - decoding of
> >> sequences and test_decoding / built-in replication changes.
> >
> > Two July buildfarm runs failed with PANIC during standby promotion:
> > https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=tern&dt=2022-07-19%2004%3A13%3A18
> > https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=tern&dt=2022-07-31%2011%3A33%3A13
> >
> > The attached patch hacks things so an ordinary x86_64 GNU/Linux machine
> > reproduces this consistently. "git bisect" then traced the regression to the
> > above revert commit (2c7ea57e56ca5f668c32d4266e0a3e45b455bef5). The pg_ctl
> > test suite passes under this hack in all supported branches, and it passed on
> > v15 until that revert. Would you investigate?
> >
> > The buildfarm animal uses keep_error_builds. From kept data directories, I
> > deduced these events:
> >
> > - After the base backup, auto-analyze ran on the primary and wrote WAL.
> > - Standby streamed and wrote up to 0/301FFF.
> > - Standby received the promote signal. Terminated streaming. WAL page at 0/302000 remained all-zeros.
> > - Somehow, end-of-recovery became a PANIC.
>
> I think it'd be really bizarre if this was due to the revert, as that
> simply undoes minor WAL changes (and none of this should affect what
> happens at WAL page boundary etc.). It just restores WAL as it was
> before 0da92dc, nothing particularly complicated. I did go through all
> of the changes again and I haven't spotted anything particularly
> suspicious, but I'll give it another try tomorrow.
>
> However, I did try bisecting this using the attached patch, and that
> does not suggest the issue is in the revert commit. It actually fails
> all the way back to 5dc0418fab2, and it starts working on 9553b4115f1.
>
> ...
> 6392f2a0968 Try to silence "-Wmissing-braces" complaints in ...
> => 5dc0418fab2 Prefetch data referenced by the WAL, take II.
> 9553b4115f1 Fix warning introduced in 5c279a6d350.
> ...
>
> This is merely 10 commits before the revert, and it seems way more
> related to WAL. Also, adding this to the two nodes in 003_standby.pl
> makes the issue go away, it seems:
>
> $node_standby->append_conf('postgresql.conf',
> qq(recovery_prefetch = off));
>
> I'd bet it's about WAL prefetching, not the revert, and the bisect was a
> bit incorrect, because the commits are close and the failures happen to
> be rare. (Presumably you first did the bisect and then wrote the patch
> that reproduces this, right?)
No. I wrote the patch, then used the patch to drive the bisect. With ten
iterations, commit 2c7ea57 passes 0/10, while 2c7ea57^ passes 10/10. I've now
tried recovery_prefetch=off. With that, the test passes 10/10 at 2c7ea57.
Given your observation of a failure at 5dc0418fab2, I agree with your
conclusion. Whatever the role of 2c7ea57 in exposing the failure on my
machine, a root cause in WAL prefetching looks more likely.
> Adding Thomas Munro to the thread, he's the WAL prefetching expert ;-)
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2022-08-07 21:09 Thomas Munro <[email protected]>
parent: Noah Misch <[email protected]>
0 siblings, 1 reply; 127+ messages in thread
From: Thomas Munro @ 2022-08-07 21:09 UTC (permalink / raw)
To: Noah Misch <[email protected]>; +Cc: Tomas Vondra <[email protected]>; Amit Kapila <[email protected]>; Petr Jelinek <[email protected]>; Peter Eisentraut <[email protected]>; PostgreSQL Hackers <[email protected]>
On Mon, Aug 8, 2022 at 7:12 AM Noah Misch <[email protected]> wrote:
> On Sun, Aug 07, 2022 at 03:18:52PM +0200, Tomas Vondra wrote:
> > I'd bet it's about WAL prefetching, not the revert, and the bisect was a
> > bit incorrect, because the commits are close and the failures happen to
> > be rare. (Presumably you first did the bisect and then wrote the patch
> > that reproduces this, right?)
>
> No. I wrote the patch, then used the patch to drive the bisect. With ten
> iterations, commit 2c7ea57 passes 0/10, while 2c7ea57^ passes 10/10. I've now
> tried recovery_prefetch=off. With that, the test passes 10/10 at 2c7ea57.
> Given your observation of a failure at 5dc0418fab2, I agree with your
> conclusion. Whatever the role of 2c7ea57 in exposing the failure on my
> machine, a root cause in WAL prefetching looks more likely.
>
> > Adding Thomas Munro to the thread, he's the WAL prefetching expert ;-)
Thanks for the repro patch and bisection work. Looking...
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2022-08-08 06:15 Thomas Munro <[email protected]>
parent: Thomas Munro <[email protected]>
0 siblings, 2 replies; 127+ messages in thread
From: Thomas Munro @ 2022-08-08 06:15 UTC (permalink / raw)
To: Noah Misch <[email protected]>; +Cc: Tomas Vondra <[email protected]>; Amit Kapila <[email protected]>; Petr Jelinek <[email protected]>; Peter Eisentraut <[email protected]>; PostgreSQL Hackers <[email protected]>
On Mon, Aug 8, 2022 at 9:09 AM Thomas Munro <[email protected]> wrote:
> Thanks for the repro patch and bisection work. Looking...
I don't have the complete explanation yet, but it's something like
this. We hit the following branch in xlogrecovery.c...
if (StandbyMode &&
!XLogReaderValidatePageHeader(xlogreader,
targetPagePtr, readBuf))
{
/*
* Emit this error right now then retry this page
immediately. Use
* errmsg_internal() because the message was already translated.
*/
if (xlogreader->errormsg_buf[0])
ereport(emode_for_corrupt_record(emode,
xlogreader->EndRecPtr),
(errmsg_internal("%s",
xlogreader->errormsg_buf)));
/* reset any error XLogReaderValidatePageHeader()
might have set */
xlogreader->errormsg_buf[0] = '\0';
goto next_record_is_invalid;
}
... but, even though there was a (suppressed) error, nothing
invalidates the reader's page buffer. Normally,
XLogReadValidatePageHeader() failure or any other kind of error
encountered by xlogreader.c'd decoding logic would do that, but here
the read_page callback is directly calling the header validation.
Without prefetching, that doesn't seem to matter, but reading ahead
can cause us to have the problem page in our buffer at the wrong time,
and then not re-read it when we should. Or something like that.
The attached patch that simply moves the cache invalidation into
report_invalid_record(), so that it's reached by the above code and
everything else that reports an error, seems to fix the problem in
src/bin/pg_ctl/t/003_promote.pl with Noah's spanner-in-the-works patch
applied, and passes check-world without it. I need to look at this
some more, though, and figure out if it's the right fix.
Attachments:
[text/x-patch] 0001-WIP-Fix-XLogPageRead-cache-invalidation.patch (1.6K, ../../CA+hUKGLptJc_k=cEqGtyRb-2pyF++uUMNi-+VEhrUV6pmezJmg@mail.gmail.com/2-0001-WIP-Fix-XLogPageRead-cache-invalidation.patch)
download | inline diff:
From 1f4fc659441a805a50b77f4e7810f9a25b5e62ae Mon Sep 17 00:00:00 2001
From: Thomas Munro <[email protected]>
Date: Mon, 8 Aug 2022 14:09:32 +1200
Subject: [PATCH] WIP Fix XLogPageRead() cache invalidation.
When ReadPageInternal() reports an error, including due to failure of
XLogReaderValidatePageHeader(), the reader calls
XLogReaderInvalReaderState() to invalidate the contents of its buffer,
forcing a re-read next time.
This did not work if a header validation error was detected by the
read_page callback itself. XLogPageRead() calls
XLogValidatePageHeader() directly, and if that failed, there was no
cache invalidation.
Fix, by invalidating the cache directly in report_invalid_record(), so
that XLogValidatePageHeader()'s failure invalidates the cache even with
called directly by xlogrecovery.c.
XXX More research/testing needed
diff --git a/src/backend/access/transam/xlogreader.c b/src/backend/access/transam/xlogreader.c
index 06e91547dd..e6431722e7 100644
--- a/src/backend/access/transam/xlogreader.c
+++ b/src/backend/access/transam/xlogreader.c
@@ -81,6 +81,12 @@ report_invalid_record(XLogReaderState *state, const char *fmt,...)
va_end(args);
state->errormsg_deferred = true;
+
+ /*
+ * Invalidate the read state. We might read from a different source after
+ * failure.
+ */
+ XLogReaderInvalReadState(state);
}
/*
@@ -1066,11 +1072,6 @@ ReadPageInternal(XLogReaderState *state, XLogRecPtr pageptr, int reqLen)
return readLen;
err:
- if (state->errormsg_buf[0] != '\0')
- {
- state->errormsg_deferred = true;
- XLogReaderInvalReadState(state);
- }
return XLREAD_FAIL;
}
--
2.37.1
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2022-08-08 08:33 Kyotaro Horiguchi <[email protected]>
parent: Thomas Munro <[email protected]>
1 sibling, 1 reply; 127+ messages in thread
From: Kyotaro Horiguchi @ 2022-08-08 08:33 UTC (permalink / raw)
To: [email protected]; +Cc: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]
At Mon, 8 Aug 2022 18:15:46 +1200, Thomas Munro <[email protected]> wrote in
> On Mon, Aug 8, 2022 at 9:09 AM Thomas Munro <[email protected]> wrote:
> > Thanks for the repro patch and bisection work. Looking...
>
> I don't have the complete explanation yet, but it's something like
> this. We hit the following branch in xlogrecovery.c...
>
> if (StandbyMode &&
> !XLogReaderValidatePageHeader(xlogreader,
> targetPagePtr, readBuf))
> {
> /*
> * Emit this error right now then retry this page
> immediately. Use
> * errmsg_internal() because the message was already translated.
> */
> if (xlogreader->errormsg_buf[0])
> ereport(emode_for_corrupt_record(emode,
> xlogreader->EndRecPtr),
> (errmsg_internal("%s",
> xlogreader->errormsg_buf)));
>
> /* reset any error XLogReaderValidatePageHeader()
> might have set */
> xlogreader->errormsg_buf[0] = '\0';
> goto next_record_is_invalid;
> }
>
> ... but, even though there was a (suppressed) error, nothing
> invalidates the reader's page buffer. Normally,
> XLogReadValidatePageHeader() failure or any other kind of error
> encountered by xlogreader.c'd decoding logic would do that, but here
> the read_page callback is directly calling the header validation.
> Without prefetching, that doesn't seem to matter, but reading ahead
> can cause us to have the problem page in our buffer at the wrong time,
> and then not re-read it when we should. Or something like that.
>
> The attached patch that simply moves the cache invalidation into
> report_invalid_record(), so that it's reached by the above code and
> everything else that reports an error, seems to fix the problem in
> src/bin/pg_ctl/t/003_promote.pl with Noah's spanner-in-the-works patch
> applied, and passes check-world without it. I need to look at this
> some more, though, and figure out if it's the right fix.
If WaitForWALToBecomeAvailable returned by promotion, ReadPageInteral
misses the chance to inavlidate reader-state. That state is not an
error while in StandbyMode.
In the repro case, XLogPageRead returns XLREAD_WOULDBLOCK after the
first failure. This situation (of course) was not considered when
that code was introduced. If that function is going to return with
XLREAD_WOULDBLOCK while lastSourceFailed, it should be turned into
XLREAD_FAIL. So, the following also works.
diff --git a/src/backend/access/transam/xlogrecovery.c b/src/backend/access/transam/xlogrecovery.c
index 21088e78f6..9f242fe656 100644
--- a/src/backend/access/transam/xlogrecovery.c
+++ b/src/backend/access/transam/xlogrecovery.c
@@ -3220,7 +3220,9 @@ retry:
xlogreader->nonblocking))
{
case XLREAD_WOULDBLOCK:
- return XLREAD_WOULDBLOCK;
+ if (!lastSourceFailed)
+ return XLREAD_WOULDBLOCK;
+ /* Fall through. */
case XLREAD_FAIL:
if (readFile >= 0)
close(readFile);
regards.
--
Kyotaro Horiguchi
NTT Open Source Software Center
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2022-08-08 08:56 Kyotaro Horiguchi <[email protected]>
parent: Kyotaro Horiguchi <[email protected]>
0 siblings, 1 reply; 127+ messages in thread
From: Kyotaro Horiguchi @ 2022-08-08 08:56 UTC (permalink / raw)
To: [email protected]; +Cc: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]
At Mon, 08 Aug 2022 17:33:22 +0900 (JST), Kyotaro Horiguchi <[email protected]> wrote in
> If WaitForWALToBecomeAvailable returned by promotion, ReadPageInteral
> misses the chance to inavlidate reader-state. That state is not an
> error while in StandbyMode.
Mmm... Maybe I wanted to say: (Still I'm not sure the rewrite works..)
If WaitForWALToBecomeAvailable returned by promotion, ReadPageInteral
would miss the chance to invalidate reader-state. When XLogPageRead
is called in blocking mode while in StandbyMode (that is, the
traditional condition) , the function continues retrying until it
succeeds, or returns XLRAD_FAIL if promote is triggered. In other
words, it was not supposed to return non-failure while the header
validation is failing while in standby mode. But while in nonblocking
mode, the function can return non-failure with lastSourceFailed =
true, which seems wrong.
regards.
--
Kyotaro Horiguchi
NTT Open Source Software Center
^ permalink raw reply [nested|flat] 127+ messages in thread
* logical decoding and replication of sequences, take 2
@ 2022-08-18 21:10 Tomas Vondra <[email protected]>
0 siblings, 1 reply; 127+ messages in thread
From: Tomas Vondra @ 2022-08-18 21:10 UTC (permalink / raw)
To: PostgreSQL Hackers <[email protected]>
Hi,
Here's a rebased version of the patch adding logical decoding of
sequences. The previous attempt [1] ended up getting reverted, due to
running into issues with non-transactional nature of sequences when
decoding the existing WAL records. See [2] for details.
This patch uses a different approach, proposed by Hannu Krosing [3],
based on tracking sequences actually modified in each transaction, and
then WAL-logging the state at the end.
This does work, but I'm not very happy about WAL-logging all sequences
at the end. The "problem" is we have to re-read the current state of the
sequence from disk, because it might be concurrently updated by another
transaction.
Imagine two transactions, T1 and T2:
T1: BEGIN
T1: SELECT nextval('s') FROM generate_series(1,1000)
T2: BEGIN
T2: SELECT nextval('s') FROM generate_series(1,1000)
T2: COMMIT
T1: COMMIT
The expected outcome is that the sequence value is ~2000. We must not
blindly apply the changes from T2 by the increments in T1. So the patch
simply reads "current" state of the transaction at commit time. Which is
annoying, because it involves I/O, increases the commit duration, etc.
On the other hand, this is likely cheaper than the other approach based
on WAL-logging every sequence increment (that would have to be careful
about obsoleted increments too, when applying them transactionally).
I wonder if we might deal with this by simply WAL-logging LSN of the
last change for each sequence (in the given xact), which would allow
discarding the "obsolete" changes quite easily I think. nextval() would
simply look at LSN in the page header.
And maybe we could then use the LSN to read the increment from the WAL
during decoding, instead of having to read it and WAL-log it during
commit. Essentially, we'd run a local XLogReader. Of course, we'd have
to be careful about checkpoints, not sure what to do about that.
Another idea that just occurred to me is that if we end up having to
read the sequence state during commit, maybe we could at least optimize
it somehow. For example we might track LSN of the last logged state for
each sequence (in shared memory or something), and the other sessions
could just skip the WAL-log if their "local" LSN is <= than this LSN.
regards
[1]
https://www.postgresql.org/message-id/flat/[email protected]
[2]
https://www.postgresql.org/message-id/00708727-d856-1886-48e3-811296c7ba8c%40enterprisedb.com
[3]
https://www.postgresql.org/message-id/CAMT0RQQeDR51xs8zTa25YpfKB1B34nS-Q4hhsRPznVsjMB_P1w%40mail.gma...
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
Attachments:
[text/x-patch] decoding-sequences-tracking-20220818.patch (577.0K, ../../[email protected]/2-decoding-sequences-tracking-20220818.patch)
download | inline diff:
diff --git a/contrib/test_decoding/Makefile b/contrib/test_decoding/Makefile
index c7ce603706..05c0c5a2f8 100644
--- a/contrib/test_decoding/Makefile
+++ b/contrib/test_decoding/Makefile
@@ -5,7 +5,8 @@ PGFILEDESC = "test_decoding - example of a logical decoding output plugin"
REGRESS = ddl xact rewrite toast permissions decoding_in_xact \
decoding_into_rel binary prepared replorigin time messages \
- spill slot truncate stream stats twophase twophase_stream
+ spill slot truncate stream stats twophase twophase_stream \
+ sequence
ISOLATION = mxact delayed_startup ondisk_startup concurrent_ddl_dml \
oldest_xmin snapshot_transfer subxact_without_top concurrent_stream \
twophase_snapshot slot_creation_error catalog_change_snapshot
diff --git a/contrib/test_decoding/expected/ddl.out b/contrib/test_decoding/expected/ddl.out
index 9a28b5ddc5..1e37c8c897 100644
--- a/contrib/test_decoding/expected/ddl.out
+++ b/contrib/test_decoding/expected/ddl.out
@@ -40,7 +40,7 @@ SELECT 'init' FROM pg_create_physical_replication_slot('repl');
init
(1 row)
-SELECT data FROM pg_logical_slot_get_changes('repl', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('repl', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
ERROR: cannot use physical replication slot for logical decoding
SELECT pg_drop_replication_slot('repl');
pg_drop_replication_slot
@@ -89,7 +89,7 @@ COMMIT;
ALTER TABLE replication_example RENAME COLUMN text TO somenum;
INSERT INTO replication_example(somedata, somenum) VALUES (4, 1);
-- collect all changes
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
---------------------------------------------------------------------------------------------------------------------------
BEGIN
@@ -118,7 +118,7 @@ SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'inc
ALTER TABLE replication_example ALTER COLUMN somenum TYPE int4 USING (somenum::int4);
-- check that this doesn't produce any changes from the heap rewrite
-SELECT count(data) FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT count(data) FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
count
-------
0
@@ -134,7 +134,7 @@ INSERT INTO replication_example(somedata, somenum, zaphod2) VALUES (6, 3, 1);
INSERT INTO replication_example(somedata, somenum, zaphod1) VALUES (6, 4, 2);
COMMIT;
-- show changes
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
------------------------------------------------------------------------------------------------------------------------------------------
BEGIN
@@ -154,7 +154,7 @@ INSERT INTO replication_example(id, somedata, somenum) SELECT i, i, i FROM gener
ON CONFLICT (id) DO UPDATE SET somenum = excluded.somenum + 1;
COMMIT;
/* display results */
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
--------------------------------------------------------------------------------------------------------------------------------------------------
BEGIN
@@ -242,7 +242,7 @@ CREATE TABLE tr_unique(id2 serial unique NOT NULL, data int);
INSERT INTO tr_unique(data) VALUES(10);
ALTER TABLE tr_unique RENAME TO tr_pkey;
ALTER TABLE tr_pkey ADD COLUMN id serial primary key;
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-rewrites', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-rewrites', '1', 'include-sequences', '0');
data
-----------------------------------------------------------------------------
BEGIN
@@ -257,7 +257,7 @@ INSERT INTO tr_pkey(data) VALUES(1);
--show deletion with primary key
DELETE FROM tr_pkey;
/* display results */
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
----------------------------------------------------------------------------
BEGIN
@@ -288,7 +288,7 @@ INSERT INTO tr_oddlength VALUES('ab', 'foo');
COMMIT;
/* display results, but hide most of the output */
SELECT count(*), min(data), max(data)
-FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1')
+FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0')
GROUP BY substring(data, 1, 24)
ORDER BY 1,2;
count | min | max
@@ -310,7 +310,7 @@ DELETE FROM spoolme;
DROP TABLE spoolme;
COMMIT;
SELECT data
-FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1')
+FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0')
WHERE data ~ 'UPDATE';
data
-------------------------------------------------------------------------------------------------------------
@@ -324,7 +324,7 @@ INSERT INTO tr_etoomuch (id, data)
SELECT g.i, -g.i FROM generate_series(8000, 12000) g(i)
ON CONFLICT(id) DO UPDATE SET data = EXCLUDED.data;
SELECT substring(data, 1, 29), count(*)
-FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1') WITH ORDINALITY
+FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0') WITH ORDINALITY
GROUP BY 1
ORDER BY min(ordinality);
substring | count
@@ -355,7 +355,7 @@ RELEASE SAVEPOINT c;
INSERT INTO tr_sub(path) VALUES ('1-top-2-#1');
RELEASE SAVEPOINT b;
COMMIT;
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
----------------------------------------------------------------------
BEGIN
@@ -394,7 +394,7 @@ INSERT INTO tr_sub(path) VALUES ('2-top-1...--#3');
RELEASE SAVEPOINT subtop;
INSERT INTO tr_sub(path) VALUES ('2-top-#1');
COMMIT;
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
------------------------------------------------------------------------
BEGIN
@@ -415,7 +415,7 @@ INSERT INTO tr_sub(path) VALUES ('3-top-2-2-#1');
ROLLBACK TO SAVEPOINT b;
INSERT INTO tr_sub(path) VALUES ('3-top-2-#2');
COMMIT;
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
-----------------------------------------------------------------------
BEGIN
@@ -444,7 +444,7 @@ BEGIN;
SAVEPOINT a;
INSERT INTO tr_sub(path) VALUES ('5-top-1-#1');
COMMIT;
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
---------------------------------------------------------------------
BEGIN
@@ -465,7 +465,7 @@ ROLLBACK TO SAVEPOINT a;
ALTER TABLE tr_sub_ddl ALTER COLUMN data TYPE bigint;
INSERT INTO tr_sub_ddl VALUES(43);
COMMIT;
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
--------------------------------------------------
BEGIN
@@ -542,7 +542,7 @@ Options: user_catalog_table=false
INSERT INTO replication_metadata(relation, options)
VALUES ('zaphod', NULL);
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
------------------------------------------------------------------------------------------------------------------------------------
BEGIN
@@ -659,7 +659,7 @@ INSERT INTO toasttable(toasted_col2) SELECT repeat(string_agg(to_char(g.i, 'FM00
UPDATE toasttable
SET toasted_col1 = (SELECT string_agg(g.i::text, '') FROM generate_series(1, 2000) g(i))
WHERE id = 1;
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
BEGIN
@@ -814,7 +814,7 @@ UPDATE toasttable
WHERE id = 1;
-- make sure we decode correctly even if the toast table is gone
DROP TABLE toasttable;
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
BEGIN
@@ -826,7 +826,7 @@ SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'inc
(6 rows)
-- done, free logical replication slot
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
------
(0 rows)
diff --git a/contrib/test_decoding/expected/decoding_in_xact.out b/contrib/test_decoding/expected/decoding_in_xact.out
index b65253f463..0816c780fe 100644
--- a/contrib/test_decoding/expected/decoding_in_xact.out
+++ b/contrib/test_decoding/expected/decoding_in_xact.out
@@ -58,7 +58,7 @@ SELECT pg_current_xact_id() = '0';
-- don't show yet, haven't committed
INSERT INTO nobarf(data) VALUES('2');
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
-----------------------------------------------------------
BEGIN
@@ -68,7 +68,7 @@ SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'inc
COMMIT;
INSERT INTO nobarf(data) VALUES('3');
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
-----------------------------------------------------------
BEGIN
diff --git a/contrib/test_decoding/expected/decoding_into_rel.out b/contrib/test_decoding/expected/decoding_into_rel.out
index 8fd3390066..03966b8b1c 100644
--- a/contrib/test_decoding/expected/decoding_into_rel.out
+++ b/contrib/test_decoding/expected/decoding_into_rel.out
@@ -19,7 +19,7 @@ SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'inc
CREATE TABLE somechange(id serial primary key);
INSERT INTO somechange DEFAULT VALUES;
CREATE TABLE changeresult AS
- SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+ SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
SELECT * FROM changeresult;
data
------------------------------------------------
@@ -29,9 +29,9 @@ SELECT * FROM changeresult;
(3 rows)
INSERT INTO changeresult
- SELECT data FROM pg_logical_slot_peek_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+ SELECT data FROM pg_logical_slot_peek_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
INSERT INTO changeresult
- SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+ SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
SELECT * FROM changeresult;
data
--------------------------------------------------------------------------------------------------------------------------------------------------
@@ -63,7 +63,7 @@ DROP TABLE somechange;
CREATE FUNCTION slot_changes_wrapper(slot_name name) RETURNS SETOF TEXT AS $$
BEGIN
RETURN QUERY
- SELECT data FROM pg_logical_slot_peek_changes(slot_name, NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+ SELECT data FROM pg_logical_slot_peek_changes(slot_name, NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
END$$ LANGUAGE plpgsql;
SELECT * FROM slot_changes_wrapper('regression_slot');
slot_changes_wrapper
@@ -84,7 +84,7 @@ SELECT * FROM slot_changes_wrapper('regression_slot');
COMMIT
(14 rows)
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
BEGIN
diff --git a/contrib/test_decoding/expected/mxact.out b/contrib/test_decoding/expected/mxact.out
index 03ad3df099..c5bc26c804 100644
--- a/contrib/test_decoding/expected/mxact.out
+++ b/contrib/test_decoding/expected/mxact.out
@@ -7,7 +7,7 @@ step s0init: SELECT 'init' FROM pg_create_logical_replication_slot('isolation_sl
init
(1 row)
-step s0start: SELECT data FROM pg_logical_slot_get_changes('isolation_slot', NULL, NULL, 'include-xids', 'false');
+step s0start: SELECT data FROM pg_logical_slot_get_changes('isolation_slot', NULL, NULL, 'include-xids', 'false', 'include-sequences', 'false');
data
----
(0 rows)
@@ -27,7 +27,7 @@ t
(1 row)
step s0w: INSERT INTO do_write DEFAULT VALUES;
-step s0start: SELECT data FROM pg_logical_slot_get_changes('isolation_slot', NULL, NULL, 'include-xids', 'false');
+step s0start: SELECT data FROM pg_logical_slot_get_changes('isolation_slot', NULL, NULL, 'include-xids', 'false', 'include-sequences', 'false');
data
--------------------------------------------
BEGIN
@@ -50,7 +50,7 @@ step s0init: SELECT 'init' FROM pg_create_logical_replication_slot('isolation_sl
init
(1 row)
-step s0start: SELECT data FROM pg_logical_slot_get_changes('isolation_slot', NULL, NULL, 'include-xids', 'false');
+step s0start: SELECT data FROM pg_logical_slot_get_changes('isolation_slot', NULL, NULL, 'include-xids', 'false', 'include-sequences', 'false');
data
----
(0 rows)
@@ -71,7 +71,7 @@ t
step s0alter: ALTER TABLE do_write ADD column ts timestamptz;
step s0w: INSERT INTO do_write DEFAULT VALUES;
-step s0start: SELECT data FROM pg_logical_slot_get_changes('isolation_slot', NULL, NULL, 'include-xids', 'false');
+step s0start: SELECT data FROM pg_logical_slot_get_changes('isolation_slot', NULL, NULL, 'include-xids', 'false', 'include-sequences', 'false');
data
------------------------------------------------------------------------------
BEGIN
diff --git a/contrib/test_decoding/expected/ondisk_startup.out b/contrib/test_decoding/expected/ondisk_startup.out
index bc7ff07164..3d2fa4a92e 100644
--- a/contrib/test_decoding/expected/ondisk_startup.out
+++ b/contrib/test_decoding/expected/ondisk_startup.out
@@ -35,7 +35,7 @@ init
step s2c: COMMIT;
step s1insert: INSERT INTO do_write DEFAULT VALUES;
step s1checkpoint: CHECKPOINT;
-step s1start: SELECT data FROM pg_logical_slot_get_changes('isolation_slot', NULL, NULL, 'include-xids', 'false');
+step s1start: SELECT data FROM pg_logical_slot_get_changes('isolation_slot', NULL, NULL, 'include-xids', 'false', 'include-sequences', 'false');
data
--------------------------------------------------------------------
BEGIN
@@ -46,7 +46,7 @@ COMMIT
step s1insert: INSERT INTO do_write DEFAULT VALUES;
step s1alter: ALTER TABLE do_write ADD COLUMN addedbys1 int;
step s1insert: INSERT INTO do_write DEFAULT VALUES;
-step s1start: SELECT data FROM pg_logical_slot_get_changes('isolation_slot', NULL, NULL, 'include-xids', 'false');
+step s1start: SELECT data FROM pg_logical_slot_get_changes('isolation_slot', NULL, NULL, 'include-xids', 'false', 'include-sequences', 'false');
data
--------------------------------------------------------------------------------------------
BEGIN
diff --git a/contrib/test_decoding/expected/replorigin.out b/contrib/test_decoding/expected/replorigin.out
index 49ffaeea2d..a828a453b1 100644
--- a/contrib/test_decoding/expected/replorigin.out
+++ b/contrib/test_decoding/expected/replorigin.out
@@ -82,9 +82,9 @@ SELECT 'init' FROM pg_create_logical_replication_slot('regression_slot', 'test_d
-- origin tx
INSERT INTO origin_tbl(data) VALUES ('will be replicated and decoded and decoded again');
INSERT INTO target_tbl(data)
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
-- as is normal, the insert into target_tbl shows up
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
----------------------------------------------------------------------------------------------------------------------------------------------------------------------------
BEGIN
@@ -120,7 +120,7 @@ SELECT pg_replication_origin_xact_setup('0/aabbccdd', '2013-01-01 00:00');
(1 row)
INSERT INTO target_tbl(data)
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'only-local', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'only-local', '1', 'include-sequences', '0');
COMMIT;
-- check replication progress for the session is correct
SELECT pg_replication_origin_session_progress(false);
@@ -164,14 +164,14 @@ SELECT pg_replication_origin_progress('regress_test_decoding: regression_slot',
SELECT pg_replication_origin_session_reset();
ERROR: no replication origin is configured
-- and magically the replayed xact will be filtered!
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'only-local', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'only-local', '1', 'include-sequences', '0');
data
------
(0 rows)
--but new original changes still show up
INSERT INTO origin_tbl(data) VALUES ('will be replicated');
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'only-local', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'only-local', '1', 'include-sequences', '0');
data
--------------------------------------------------------------------------------
BEGIN
@@ -237,7 +237,7 @@ SELECT local_id, external_id,
1 | regress_test_decoding: regression_slot_no_lsn | f | t
(1 row)
-SELECT data FROM pg_logical_slot_get_changes('regression_slot_no_lsn', NULL, NULL, 'skip-empty-xacts', '1', 'include-xids', '0');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot_no_lsn', NULL, NULL, 'skip-empty-xacts', '1', 'include-xids', '0', 'include-sequences', '0');
data
-------------------------------------------------------------------------------------
BEGIN
diff --git a/contrib/test_decoding/expected/rewrite.out b/contrib/test_decoding/expected/rewrite.out
index b30999c436..5d15b192ed 100644
--- a/contrib/test_decoding/expected/rewrite.out
+++ b/contrib/test_decoding/expected/rewrite.out
@@ -64,7 +64,7 @@ SELECT 'init' FROM pg_create_logical_replication_slot('regression_slot', 'test_d
CREATE TABLE replication_example(id SERIAL PRIMARY KEY, somedata int, text varchar(120));
INSERT INTO replication_example(somedata) VALUES (1);
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
----------------------------------------------------------------------------------------------------------
BEGIN
@@ -115,7 +115,7 @@ INSERT INTO replication_example(somedata, testcolumn1, testcolumn3) VALUES (7, 5
COMMIT;
-- make old files go away
CHECKPOINT;
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
BEGIN
@@ -141,7 +141,7 @@ VACUUM FULL pg_proc; VACUUM FULL pg_description; VACUUM FULL pg_shdescription; V
INSERT INTO replication_example(somedata, testcolumn1, testcolumn3) VALUES (8, 6, 1);
VACUUM FULL pg_proc; VACUUM FULL pg_description; VACUUM FULL pg_shdescription; VACUUM FULL iamalargetable;
INSERT INTO replication_example(somedata, testcolumn1, testcolumn3) VALUES (9, 7, 1);
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
BEGIN
diff --git a/contrib/test_decoding/expected/sequence.out b/contrib/test_decoding/expected/sequence.out
new file mode 100644
index 0000000000..dc1bf67012
--- /dev/null
+++ b/contrib/test_decoding/expected/sequence.out
@@ -0,0 +1,325 @@
+-- predictability
+SET synchronous_commit = on;
+SELECT 'init' FROM pg_create_logical_replication_slot('regression_slot', 'test_decoding');
+ ?column?
+----------
+ init
+(1 row)
+
+CREATE SEQUENCE test_sequence;
+-- test the sequence changes by several nextval() calls
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 1
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 2
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 4
+(1 row)
+
+-- test the sequence changes by several ALTER commands
+ALTER SEQUENCE test_sequence INCREMENT BY 10;
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 14
+(1 row)
+
+ALTER SEQUENCE test_sequence START WITH 3000;
+ALTER SEQUENCE test_sequence MAXVALUE 10000;
+ALTER SEQUENCE test_sequence RESTART WITH 4000;
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 4000
+(1 row)
+
+-- test the sequence changes by several setval() calls
+SELECT setval('test_sequence', 3500);
+ setval
+--------
+ 3500
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3510
+(1 row)
+
+SELECT setval('test_sequence', 3500, true);
+ setval
+--------
+ 3500
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3510
+(1 row)
+
+SELECT setval('test_sequence', 3500, false);
+ setval
+--------
+ 3500
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3500
+(1 row)
+
+-- show results and drop sequence
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+ data
+----------------------------------------------------------------------------------------
+ BEGIN
+ sequence public.test_sequence: transactional:1 last_value: 1 log_cnt: 0 is_called:0
+ COMMIT
+ sequence public.test_sequence: transactional:0 last_value: 33 log_cnt: 0 is_called:1
+ BEGIN
+ sequence public.test_sequence: transactional:1 last_value: 4 log_cnt: 0 is_called:1
+ COMMIT
+ sequence public.test_sequence: transactional:0 last_value: 334 log_cnt: 0 is_called:1
+ BEGIN
+ sequence public.test_sequence: transactional:1 last_value: 14 log_cnt: 32 is_called:1
+ COMMIT
+ BEGIN
+ sequence public.test_sequence: transactional:1 last_value: 14 log_cnt: 0 is_called:1
+ COMMIT
+ BEGIN
+ sequence public.test_sequence: transactional:1 last_value: 4000 log_cnt: 0 is_called:0
+ COMMIT
+ sequence public.test_sequence: transactional:0 last_value: 4320 log_cnt: 0 is_called:1
+ sequence public.test_sequence: transactional:0 last_value: 3500 log_cnt: 0 is_called:1
+ sequence public.test_sequence: transactional:0 last_value: 3830 log_cnt: 0 is_called:1
+ sequence public.test_sequence: transactional:0 last_value: 3500 log_cnt: 0 is_called:1
+ sequence public.test_sequence: transactional:0 last_value: 3830 log_cnt: 0 is_called:1
+ sequence public.test_sequence: transactional:0 last_value: 3500 log_cnt: 0 is_called:0
+ sequence public.test_sequence: transactional:0 last_value: 3820 log_cnt: 0 is_called:1
+(24 rows)
+
+DROP SEQUENCE test_sequence;
+-- rollback on sequence creation and update
+BEGIN;
+CREATE SEQUENCE test_sequence;
+CREATE TABLE test_table (a INT);
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 1
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 2
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3
+(1 row)
+
+SELECT setval('test_sequence', 3000);
+ setval
+--------
+ 3000
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3001
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3002
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3003
+(1 row)
+
+ALTER SEQUENCE test_sequence RESTART WITH 6000;
+INSERT INTO test_table VALUES( (SELECT nextval('test_sequence')) );
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 6001
+(1 row)
+
+ROLLBACK;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+ data
+------
+(0 rows)
+
+-- rollback on table creation with serial column
+BEGIN;
+CREATE TABLE test_table (a SERIAL, b INT);
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+INSERT INTO test_table (b) VALUES (300);
+SELECT setval('test_table_a_seq', 3000);
+ setval
+--------
+ 3000
+(1 row)
+
+INSERT INTO test_table (b) VALUES (400);
+INSERT INTO test_table (b) VALUES (500);
+INSERT INTO test_table (b) VALUES (600);
+ALTER SEQUENCE test_table_a_seq RESTART WITH 6000;
+INSERT INTO test_table (b) VALUES (700);
+INSERT INTO test_table (b) VALUES (800);
+INSERT INTO test_table (b) VALUES (900);
+ROLLBACK;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+ data
+------
+(0 rows)
+
+-- rollback on table with serial column
+CREATE TABLE test_table (a SERIAL, b INT);
+BEGIN;
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+INSERT INTO test_table (b) VALUES (300);
+SELECT setval('test_table_a_seq', 3000);
+ setval
+--------
+ 3000
+(1 row)
+
+INSERT INTO test_table (b) VALUES (400);
+INSERT INTO test_table (b) VALUES (500);
+INSERT INTO test_table (b) VALUES (600);
+ALTER SEQUENCE test_table_a_seq RESTART WITH 6000;
+INSERT INTO test_table (b) VALUES (700);
+INSERT INTO test_table (b) VALUES (800);
+INSERT INTO test_table (b) VALUES (900);
+ROLLBACK;
+-- check table and sequence values after rollback
+SELECT * from test_table_a_seq;
+ last_value | log_cnt | is_called
+------------+---------+-----------
+ 3003 | 30 | t
+(1 row)
+
+SELECT nextval('test_table_a_seq');
+ nextval
+---------
+ 3004
+(1 row)
+
+DROP TABLE test_table;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+ data
+-------------------------------------------------------------------------------------------
+ BEGIN
+ sequence public.test_table_a_seq: transactional:1 last_value: 1 log_cnt: 0 is_called:0
+ COMMIT
+ sequence public.test_table_a_seq: transactional:0 last_value: 33 log_cnt: 0 is_called:1
+ sequence public.test_table_a_seq: transactional:0 last_value: 3000 log_cnt: 0 is_called:1
+ sequence public.test_table_a_seq: transactional:0 last_value: 3033 log_cnt: 0 is_called:1
+(6 rows)
+
+-- savepoint test on table with serial column
+BEGIN;
+CREATE TABLE test_table (a SERIAL, b INT);
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+SAVEPOINT a;
+INSERT INTO test_table (b) VALUES (300);
+ROLLBACK TO SAVEPOINT a;
+DROP TABLE test_table;
+COMMIT;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+ data
+-----------------------------------------------------------------------------------------
+ BEGIN
+ sequence public.test_table_a_seq: transactional:1 last_value: 1 log_cnt: 0 is_called:0
+ sequence public.test_table_a_seq: transactional:1 last_value: 33 log_cnt: 0 is_called:1
+ table public.test_table: INSERT: a[integer]:1 b[integer]:100
+ table public.test_table: INSERT: a[integer]:2 b[integer]:200
+ COMMIT
+(6 rows)
+
+-- savepoint test on table with serial column
+BEGIN;
+CREATE SEQUENCE test_sequence;
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 1
+(1 row)
+
+SELECT setval('test_sequence', 3000);
+ setval
+--------
+ 3000
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3001
+(1 row)
+
+SAVEPOINT a;
+ALTER SEQUENCE test_sequence START WITH 7000;
+SELECT setval('test_sequence', 5000);
+ setval
+--------
+ 5000
+(1 row)
+
+ROLLBACK TO SAVEPOINT a;
+SELECT * FROM test_sequence;
+ last_value | log_cnt | is_called
+------------+---------+-----------
+ 3001 | 32 | t
+(1 row)
+
+DROP SEQUENCE test_sequence;
+COMMIT;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+ data
+----------------------------------------------------------------------------------------
+ BEGIN
+ sequence public.test_sequence: transactional:1 last_value: 1 log_cnt: 0 is_called:0
+ sequence public.test_sequence: transactional:1 last_value: 33 log_cnt: 0 is_called:1
+ sequence public.test_sequence: transactional:1 last_value: 3000 log_cnt: 0 is_called:1
+ sequence public.test_sequence: transactional:1 last_value: 3033 log_cnt: 0 is_called:1
+ COMMIT
+(6 rows)
+
+SELECT pg_drop_replication_slot('regression_slot');
+ pg_drop_replication_slot
+--------------------------
+
+(1 row)
+
diff --git a/contrib/test_decoding/expected/slot.out b/contrib/test_decoding/expected/slot.out
index 63a9940f73..de59d544ae 100644
--- a/contrib/test_decoding/expected/slot.out
+++ b/contrib/test_decoding/expected/slot.out
@@ -95,7 +95,7 @@ SELECT 'init' FROM pg_create_logical_replication_slot('regression_slot2', 'test_
(1 row)
INSERT INTO replication_example(somedata, text) VALUES (1, 3);
-SELECT data FROM pg_logical_slot_get_changes('regression_slot1', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot1', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
---------------------------------------------------------------------------------------------------------
BEGIN
@@ -107,7 +107,7 @@ SELECT data FROM pg_logical_slot_get_changes('regression_slot1', NULL, NULL, 'in
COMMIT
(7 rows)
-SELECT data FROM pg_logical_slot_get_changes('regression_slot2', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot2', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
---------------------------------------------------------------------------------------------------------
BEGIN
@@ -132,7 +132,7 @@ SELECT :'wal_lsn' = :'end_lsn';
t
(1 row)
-SELECT data FROM pg_logical_slot_get_changes('regression_slot1', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot1', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
---------------------------------------------------------------------------------------------------------
BEGIN
@@ -140,7 +140,7 @@ SELECT data FROM pg_logical_slot_get_changes('regression_slot1', NULL, NULL, 'in
COMMIT
(3 rows)
-SELECT data FROM pg_logical_slot_get_changes('regression_slot2', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot2', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
------
(0 rows)
diff --git a/contrib/test_decoding/expected/toast.out b/contrib/test_decoding/expected/toast.out
index a757e7dc8d..d0ab0e0639 100644
--- a/contrib/test_decoding/expected/toast.out
+++ b/contrib/test_decoding/expected/toast.out
@@ -52,7 +52,7 @@ CREATE TABLE toasted_copy (
);
ALTER TABLE toasted_copy ALTER COLUMN data SET STORAGE EXTERNAL;
\copy toasted_copy FROM STDIN
-SELECT substr(data, 1, 200) FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT substr(data, 1, 200) FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
substr
----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
BEGIN
@@ -316,7 +316,7 @@ SELECT pg_column_size(toasted_key) > 2^16 FROM toasted_several;
t
(1 row)
-SELECT regexp_replace(data, '^(.{100}).*(.{100})$', '\1..\2') FROM pg_logical_slot_peek_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT regexp_replace(data, '^(.{100}).*(.{100})$', '\1..\2') FROM pg_logical_slot_peek_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
regexp_replace
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
BEGIN
@@ -327,7 +327,7 @@ SELECT regexp_replace(data, '^(.{100}).*(.{100})$', '\1..\2') FROM pg_logical_sl
-- test update of a toasted key without changing it
UPDATE toasted_several SET toasted_col1 = toasted_key;
UPDATE toasted_several SET toasted_col2 = toasted_col1;
-SELECT regexp_replace(data, '^(.{100}).*(.{100})$', '\1..\2') FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT regexp_replace(data, '^(.{100}).*(.{100})$', '\1..\2') FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
regexp_replace
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
BEGIN
@@ -350,7 +350,7 @@ UPDATE toasted_several SET toasted_col1 = toasted_col2 WHERE id = 1;
DELETE FROM toasted_several WHERE id = 1;
COMMIT;
DROP TABLE toasted_several;
-SELECT regexp_replace(data, '^(.{100}).*(.{100})$', '\1..\2') FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1')
+SELECT regexp_replace(data, '^(.{100}).*(.{100})$', '\1..\2') FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0')
WHERE data NOT LIKE '%INSERT: %';
regexp_replace
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
@@ -373,7 +373,7 @@ INSERT INTO tbl1 VALUES(1, repeat('a', 4000)) ;
ALTER TABLE tbl1 ADD COLUMN id serial primary key;
INSERT INTO tbl2 VALUES(1);
commit;
-SELECT substr(data, 1, 200) FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT substr(data, 1, 200) FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
substr
----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
BEGIN
diff --git a/contrib/test_decoding/expected/truncate.out b/contrib/test_decoding/expected/truncate.out
index e64d377214..961b0ebdc8 100644
--- a/contrib/test_decoding/expected/truncate.out
+++ b/contrib/test_decoding/expected/truncate.out
@@ -11,7 +11,7 @@ CREATE TABLE tab2 (a int primary key, b int);
TRUNCATE tab1;
TRUNCATE tab1, tab1 RESTART IDENTITY CASCADE;
TRUNCATE tab1, tab2;
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
------------------------------------------------------
BEGIN
diff --git a/contrib/test_decoding/specs/mxact.spec b/contrib/test_decoding/specs/mxact.spec
index ea5b1aa2d6..19f3af14b3 100644
--- a/contrib/test_decoding/specs/mxact.spec
+++ b/contrib/test_decoding/specs/mxact.spec
@@ -13,7 +13,7 @@ teardown
session "s0"
setup { SET synchronous_commit=on; }
step "s0init" {SELECT 'init' FROM pg_create_logical_replication_slot('isolation_slot', 'test_decoding');}
-step "s0start" {SELECT data FROM pg_logical_slot_get_changes('isolation_slot', NULL, NULL, 'include-xids', 'false');}
+step "s0start" {SELECT data FROM pg_logical_slot_get_changes('isolation_slot', NULL, NULL, 'include-xids', 'false', 'include-sequences', 'false');}
step "s0alter" {ALTER TABLE do_write ADD column ts timestamptz; }
step "s0w" { INSERT INTO do_write DEFAULT VALUES; }
diff --git a/contrib/test_decoding/specs/ondisk_startup.spec b/contrib/test_decoding/specs/ondisk_startup.spec
index 96ce87f1a4..fb43cbc5e1 100644
--- a/contrib/test_decoding/specs/ondisk_startup.spec
+++ b/contrib/test_decoding/specs/ondisk_startup.spec
@@ -16,7 +16,7 @@ session "s1"
setup { SET synchronous_commit=on; }
step "s1init" {SELECT 'init' FROM pg_create_logical_replication_slot('isolation_slot', 'test_decoding');}
-step "s1start" {SELECT data FROM pg_logical_slot_get_changes('isolation_slot', NULL, NULL, 'include-xids', 'false');}
+step "s1start" {SELECT data FROM pg_logical_slot_get_changes('isolation_slot', NULL, NULL, 'include-xids', 'false', 'include-sequences', 'false');}
step "s1insert" { INSERT INTO do_write DEFAULT VALUES; }
step "s1checkpoint" { CHECKPOINT; }
step "s1alter" { ALTER TABLE do_write ADD COLUMN addedbys1 int; }
diff --git a/contrib/test_decoding/sql/ddl.sql b/contrib/test_decoding/sql/ddl.sql
index 4f76bed72c..807bc56a45 100644
--- a/contrib/test_decoding/sql/ddl.sql
+++ b/contrib/test_decoding/sql/ddl.sql
@@ -19,7 +19,7 @@ SELECT pg_drop_replication_slot('regression_slot');
-- check that we're detecting a streaming rep slot used for logical decoding
SELECT 'init' FROM pg_create_physical_replication_slot('repl');
-SELECT data FROM pg_logical_slot_get_changes('repl', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('repl', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
SELECT pg_drop_replication_slot('repl');
@@ -64,11 +64,11 @@ ALTER TABLE replication_example RENAME COLUMN text TO somenum;
INSERT INTO replication_example(somedata, somenum) VALUES (4, 1);
-- collect all changes
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
ALTER TABLE replication_example ALTER COLUMN somenum TYPE int4 USING (somenum::int4);
-- check that this doesn't produce any changes from the heap rewrite
-SELECT count(data) FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT count(data) FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
INSERT INTO replication_example(somedata, somenum) VALUES (5, 1);
@@ -82,7 +82,7 @@ INSERT INTO replication_example(somedata, somenum, zaphod1) VALUES (6, 4, 2);
COMMIT;
-- show changes
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
-- ON CONFLICT DO UPDATE support
BEGIN;
@@ -91,7 +91,7 @@ INSERT INTO replication_example(id, somedata, somenum) SELECT i, i, i FROM gener
COMMIT;
/* display results */
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
-- MERGE support
BEGIN;
@@ -113,14 +113,14 @@ CREATE TABLE tr_unique(id2 serial unique NOT NULL, data int);
INSERT INTO tr_unique(data) VALUES(10);
ALTER TABLE tr_unique RENAME TO tr_pkey;
ALTER TABLE tr_pkey ADD COLUMN id serial primary key;
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-rewrites', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-rewrites', '1', 'include-sequences', '0');
INSERT INTO tr_pkey(data) VALUES(1);
--show deletion with primary key
DELETE FROM tr_pkey;
/* display results */
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
/*
* check that disk spooling works (also for logical messages)
@@ -137,7 +137,7 @@ COMMIT;
/* display results, but hide most of the output */
SELECT count(*), min(data), max(data)
-FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1')
+FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0')
GROUP BY substring(data, 1, 24)
ORDER BY 1,2;
@@ -152,7 +152,7 @@ DROP TABLE spoolme;
COMMIT;
SELECT data
-FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1')
+FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0')
WHERE data ~ 'UPDATE';
-- check that a large, spooled, upsert works
@@ -161,7 +161,7 @@ SELECT g.i, -g.i FROM generate_series(8000, 12000) g(i)
ON CONFLICT(id) DO UPDATE SET data = EXCLUDED.data;
SELECT substring(data, 1, 29), count(*)
-FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1') WITH ORDINALITY
+FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0') WITH ORDINALITY
GROUP BY 1
ORDER BY min(ordinality);
@@ -189,7 +189,7 @@ INSERT INTO tr_sub(path) VALUES ('1-top-2-#1');
RELEASE SAVEPOINT b;
COMMIT;
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
-- check that we handle xlog assignments correctly
BEGIN;
@@ -218,7 +218,7 @@ RELEASE SAVEPOINT subtop;
INSERT INTO tr_sub(path) VALUES ('2-top-#1');
COMMIT;
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
-- make sure rollbacked subtransactions aren't decoded
BEGIN;
@@ -231,7 +231,7 @@ ROLLBACK TO SAVEPOINT b;
INSERT INTO tr_sub(path) VALUES ('3-top-2-#2');
COMMIT;
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
-- test whether a known, but not yet logged toplevel xact, followed by a
-- subxact commit is handled correctly
@@ -250,7 +250,7 @@ INSERT INTO tr_sub(path) VALUES ('5-top-1-#1');
COMMIT;
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
-- check that DDL in aborted subtransactions handled correctly
CREATE TABLE tr_sub_ddl(data int);
@@ -263,7 +263,7 @@ ALTER TABLE tr_sub_ddl ALTER COLUMN data TYPE bigint;
INSERT INTO tr_sub_ddl VALUES(43);
COMMIT;
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
/*
@@ -303,7 +303,7 @@ ALTER TABLE replication_metadata SET (user_catalog_table = false);
INSERT INTO replication_metadata(relation, options)
VALUES ('zaphod', NULL);
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
/*
* check whether we handle updates/deletes correct with & without a pkey
@@ -414,7 +414,7 @@ UPDATE toasttable
SET toasted_col1 = (SELECT string_agg(g.i::text, '') FROM generate_series(1, 2000) g(i))
WHERE id = 1;
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
INSERT INTO toasttable(toasted_col1) SELECT string_agg(g.i::text, '') FROM generate_series(1, 2000) g(i);
@@ -426,10 +426,10 @@ WHERE id = 1;
-- make sure we decode correctly even if the toast table is gone
DROP TABLE toasttable;
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
-- done, free logical replication slot
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
SELECT pg_drop_replication_slot('regression_slot');
diff --git a/contrib/test_decoding/sql/decoding_in_xact.sql b/contrib/test_decoding/sql/decoding_in_xact.sql
index 108782dc2e..b343b74566 100644
--- a/contrib/test_decoding/sql/decoding_in_xact.sql
+++ b/contrib/test_decoding/sql/decoding_in_xact.sql
@@ -32,10 +32,10 @@ BEGIN;
SELECT pg_current_xact_id() = '0';
-- don't show yet, haven't committed
INSERT INTO nobarf(data) VALUES('2');
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
COMMIT;
INSERT INTO nobarf(data) VALUES('3');
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
SELECT 'stop' FROM pg_drop_replication_slot('regression_slot');
diff --git a/contrib/test_decoding/sql/decoding_into_rel.sql b/contrib/test_decoding/sql/decoding_into_rel.sql
index 1068cec588..27d5d08879 100644
--- a/contrib/test_decoding/sql/decoding_into_rel.sql
+++ b/contrib/test_decoding/sql/decoding_into_rel.sql
@@ -15,14 +15,14 @@ CREATE TABLE somechange(id serial primary key);
INSERT INTO somechange DEFAULT VALUES;
CREATE TABLE changeresult AS
- SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+ SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
SELECT * FROM changeresult;
INSERT INTO changeresult
- SELECT data FROM pg_logical_slot_peek_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+ SELECT data FROM pg_logical_slot_peek_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
INSERT INTO changeresult
- SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+ SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
SELECT * FROM changeresult;
DROP TABLE changeresult;
@@ -32,11 +32,11 @@ DROP TABLE somechange;
CREATE FUNCTION slot_changes_wrapper(slot_name name) RETURNS SETOF TEXT AS $$
BEGIN
RETURN QUERY
- SELECT data FROM pg_logical_slot_peek_changes(slot_name, NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+ SELECT data FROM pg_logical_slot_peek_changes(slot_name, NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
END$$ LANGUAGE plpgsql;
SELECT * FROM slot_changes_wrapper('regression_slot');
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
SELECT 'stop' FROM pg_drop_replication_slot('regression_slot');
diff --git a/contrib/test_decoding/sql/replorigin.sql b/contrib/test_decoding/sql/replorigin.sql
index db06541f56..8f8db49963 100644
--- a/contrib/test_decoding/sql/replorigin.sql
+++ b/contrib/test_decoding/sql/replorigin.sql
@@ -46,10 +46,10 @@ SELECT 'init' FROM pg_create_logical_replication_slot('regression_slot', 'test_d
-- origin tx
INSERT INTO origin_tbl(data) VALUES ('will be replicated and decoded and decoded again');
INSERT INTO target_tbl(data)
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
-- as is normal, the insert into target_tbl shows up
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
INSERT INTO origin_tbl(data) VALUES ('will be replicated, but not decoded again');
@@ -65,7 +65,7 @@ BEGIN;
-- setup transaction origin
SELECT pg_replication_origin_xact_setup('0/aabbccdd', '2013-01-01 00:00');
INSERT INTO target_tbl(data)
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'only-local', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'only-local', '1', 'include-sequences', '0');
COMMIT;
-- check replication progress for the session is correct
@@ -84,11 +84,11 @@ SELECT pg_replication_origin_progress('regress_test_decoding: regression_slot',
SELECT pg_replication_origin_session_reset();
-- and magically the replayed xact will be filtered!
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'only-local', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'only-local', '1', 'include-sequences', '0');
--but new original changes still show up
INSERT INTO origin_tbl(data) VALUES ('will be replicated');
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'only-local', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'only-local', '1', 'include-sequences', '0');
SELECT pg_drop_replication_slot('regression_slot');
SELECT pg_replication_origin_drop('regress_test_decoding: regression_slot');
@@ -119,7 +119,7 @@ SELECT local_id, external_id,
remote_lsn <> '0/0' AS valid_remote_lsn,
local_lsn <> '0/0' AS valid_local_lsn
FROM pg_replication_origin_status;
-SELECT data FROM pg_logical_slot_get_changes('regression_slot_no_lsn', NULL, NULL, 'skip-empty-xacts', '1', 'include-xids', '0');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot_no_lsn', NULL, NULL, 'skip-empty-xacts', '1', 'include-xids', '0', 'include-sequences', '0');
-- Clean up
SELECT pg_replication_origin_session_reset();
SELECT pg_drop_replication_slot('regression_slot_no_lsn');
diff --git a/contrib/test_decoding/sql/rewrite.sql b/contrib/test_decoding/sql/rewrite.sql
index 62dead3a9b..1715bd289d 100644
--- a/contrib/test_decoding/sql/rewrite.sql
+++ b/contrib/test_decoding/sql/rewrite.sql
@@ -35,7 +35,7 @@ SELECT pg_relation_size((SELECT reltoastrelid FROM pg_class WHERE oid = 'pg_shde
SELECT 'init' FROM pg_create_logical_replication_slot('regression_slot', 'test_decoding');
CREATE TABLE replication_example(id SERIAL PRIMARY KEY, somedata int, text varchar(120));
INSERT INTO replication_example(somedata) VALUES (1);
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
BEGIN;
INSERT INTO replication_example(somedata) VALUES (2);
@@ -90,7 +90,7 @@ COMMIT;
-- make old files go away
CHECKPOINT;
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
-- trigger repeated rewrites of a system catalog with a toast table,
-- that previously was buggy: [email protected]
@@ -98,7 +98,7 @@ VACUUM FULL pg_proc; VACUUM FULL pg_description; VACUUM FULL pg_shdescription; V
INSERT INTO replication_example(somedata, testcolumn1, testcolumn3) VALUES (8, 6, 1);
VACUUM FULL pg_proc; VACUUM FULL pg_description; VACUUM FULL pg_shdescription; VACUUM FULL iamalargetable;
INSERT INTO replication_example(somedata, testcolumn1, testcolumn3) VALUES (9, 7, 1);
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
SELECT pg_drop_replication_slot('regression_slot');
DROP TABLE IF EXISTS replication_example;
diff --git a/contrib/test_decoding/sql/sequence.sql b/contrib/test_decoding/sql/sequence.sql
new file mode 100644
index 0000000000..978ee8088e
--- /dev/null
+++ b/contrib/test_decoding/sql/sequence.sql
@@ -0,0 +1,119 @@
+-- predictability
+SET synchronous_commit = on;
+SELECT 'init' FROM pg_create_logical_replication_slot('regression_slot', 'test_decoding');
+
+CREATE SEQUENCE test_sequence;
+
+-- test the sequence changes by several nextval() calls
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+
+-- test the sequence changes by several ALTER commands
+ALTER SEQUENCE test_sequence INCREMENT BY 10;
+SELECT nextval('test_sequence');
+
+ALTER SEQUENCE test_sequence START WITH 3000;
+ALTER SEQUENCE test_sequence MAXVALUE 10000;
+ALTER SEQUENCE test_sequence RESTART WITH 4000;
+SELECT nextval('test_sequence');
+
+-- test the sequence changes by several setval() calls
+SELECT setval('test_sequence', 3500);
+SELECT nextval('test_sequence');
+SELECT setval('test_sequence', 3500, true);
+SELECT nextval('test_sequence');
+SELECT setval('test_sequence', 3500, false);
+SELECT nextval('test_sequence');
+
+-- show results and drop sequence
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+DROP SEQUENCE test_sequence;
+
+-- rollback on sequence creation and update
+BEGIN;
+CREATE SEQUENCE test_sequence;
+CREATE TABLE test_table (a INT);
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT setval('test_sequence', 3000);
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+ALTER SEQUENCE test_sequence RESTART WITH 6000;
+INSERT INTO test_table VALUES( (SELECT nextval('test_sequence')) );
+SELECT nextval('test_sequence');
+ROLLBACK;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+
+-- rollback on table creation with serial column
+BEGIN;
+CREATE TABLE test_table (a SERIAL, b INT);
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+INSERT INTO test_table (b) VALUES (300);
+SELECT setval('test_table_a_seq', 3000);
+INSERT INTO test_table (b) VALUES (400);
+INSERT INTO test_table (b) VALUES (500);
+INSERT INTO test_table (b) VALUES (600);
+ALTER SEQUENCE test_table_a_seq RESTART WITH 6000;
+INSERT INTO test_table (b) VALUES (700);
+INSERT INTO test_table (b) VALUES (800);
+INSERT INTO test_table (b) VALUES (900);
+ROLLBACK;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+
+-- rollback on table with serial column
+CREATE TABLE test_table (a SERIAL, b INT);
+
+BEGIN;
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+INSERT INTO test_table (b) VALUES (300);
+SELECT setval('test_table_a_seq', 3000);
+INSERT INTO test_table (b) VALUES (400);
+INSERT INTO test_table (b) VALUES (500);
+INSERT INTO test_table (b) VALUES (600);
+ALTER SEQUENCE test_table_a_seq RESTART WITH 6000;
+INSERT INTO test_table (b) VALUES (700);
+INSERT INTO test_table (b) VALUES (800);
+INSERT INTO test_table (b) VALUES (900);
+ROLLBACK;
+
+-- check table and sequence values after rollback
+SELECT * from test_table_a_seq;
+SELECT nextval('test_table_a_seq');
+
+DROP TABLE test_table;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+
+-- savepoint test on table with serial column
+BEGIN;
+CREATE TABLE test_table (a SERIAL, b INT);
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+SAVEPOINT a;
+INSERT INTO test_table (b) VALUES (300);
+ROLLBACK TO SAVEPOINT a;
+DROP TABLE test_table;
+COMMIT;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+
+-- savepoint test on table with serial column
+BEGIN;
+CREATE SEQUENCE test_sequence;
+SELECT nextval('test_sequence');
+SELECT setval('test_sequence', 3000);
+SELECT nextval('test_sequence');
+SAVEPOINT a;
+ALTER SEQUENCE test_sequence START WITH 7000;
+SELECT setval('test_sequence', 5000);
+ROLLBACK TO SAVEPOINT a;
+SELECT * FROM test_sequence;
+DROP SEQUENCE test_sequence;
+COMMIT;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+
+SELECT pg_drop_replication_slot('regression_slot');
diff --git a/contrib/test_decoding/sql/slot.sql b/contrib/test_decoding/sql/slot.sql
index 1aa27c5667..52a740c43d 100644
--- a/contrib/test_decoding/sql/slot.sql
+++ b/contrib/test_decoding/sql/slot.sql
@@ -50,8 +50,8 @@ SELECT 'init' FROM pg_create_logical_replication_slot('regression_slot2', 'test_
INSERT INTO replication_example(somedata, text) VALUES (1, 3);
-SELECT data FROM pg_logical_slot_get_changes('regression_slot1', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
-SELECT data FROM pg_logical_slot_get_changes('regression_slot2', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot1', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot2', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
INSERT INTO replication_example(somedata, text) VALUES (1, 4);
INSERT INTO replication_example(somedata, text) VALUES (1, 5);
@@ -65,8 +65,8 @@ SELECT slot_name FROM pg_replication_slot_advance('regression_slot2', pg_current
SELECT :'wal_lsn' = :'end_lsn';
-SELECT data FROM pg_logical_slot_get_changes('regression_slot1', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
-SELECT data FROM pg_logical_slot_get_changes('regression_slot2', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot1', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot2', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
DROP TABLE replication_example;
diff --git a/contrib/test_decoding/sql/toast.sql b/contrib/test_decoding/sql/toast.sql
index d1c560a174..f5d4fc082e 100644
--- a/contrib/test_decoding/sql/toast.sql
+++ b/contrib/test_decoding/sql/toast.sql
@@ -264,7 +264,7 @@ ALTER TABLE toasted_copy ALTER COLUMN data SET STORAGE EXTERNAL;
202 untoasted199
203 untoasted200
\.
-SELECT substr(data, 1, 200) FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT substr(data, 1, 200) FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
-- test we can decode "old" tuples bigger than the max heap tuple size correctly
DROP TABLE IF EXISTS toasted_several;
@@ -287,13 +287,13 @@ UPDATE pg_attribute SET attstorage = 'x' WHERE attrelid = 'toasted_several_pkey'
INSERT INTO toasted_several(toasted_key) VALUES(repeat('9876543210', 10000));
SELECT pg_column_size(toasted_key) > 2^16 FROM toasted_several;
-SELECT regexp_replace(data, '^(.{100}).*(.{100})$', '\1..\2') FROM pg_logical_slot_peek_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT regexp_replace(data, '^(.{100}).*(.{100})$', '\1..\2') FROM pg_logical_slot_peek_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
-- test update of a toasted key without changing it
UPDATE toasted_several SET toasted_col1 = toasted_key;
UPDATE toasted_several SET toasted_col2 = toasted_col1;
-SELECT regexp_replace(data, '^(.{100}).*(.{100})$', '\1..\2') FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT regexp_replace(data, '^(.{100}).*(.{100})$', '\1..\2') FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
/*
* update with large tuplebuf, in a transaction large enough to force to spool to disk
@@ -306,7 +306,7 @@ COMMIT;
DROP TABLE toasted_several;
-SELECT regexp_replace(data, '^(.{100}).*(.{100})$', '\1..\2') FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1')
+SELECT regexp_replace(data, '^(.{100}).*(.{100})$', '\1..\2') FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0')
WHERE data NOT LIKE '%INSERT: %';
/*
@@ -322,6 +322,6 @@ INSERT INTO tbl1 VALUES(1, repeat('a', 4000)) ;
ALTER TABLE tbl1 ADD COLUMN id serial primary key;
INSERT INTO tbl2 VALUES(1);
commit;
-SELECT substr(data, 1, 200) FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT substr(data, 1, 200) FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
SELECT pg_drop_replication_slot('regression_slot');
diff --git a/contrib/test_decoding/sql/truncate.sql b/contrib/test_decoding/sql/truncate.sql
index 5633854e0d..b0f36fd681 100644
--- a/contrib/test_decoding/sql/truncate.sql
+++ b/contrib/test_decoding/sql/truncate.sql
@@ -10,5 +10,5 @@ TRUNCATE tab1;
TRUNCATE tab1, tab1 RESTART IDENTITY CASCADE;
TRUNCATE tab1, tab2;
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
SELECT pg_drop_replication_slot('regression_slot');
diff --git a/contrib/test_decoding/test_decoding.c b/contrib/test_decoding/test_decoding.c
index 3f90ffa32d..e3b131eb29 100644
--- a/contrib/test_decoding/test_decoding.c
+++ b/contrib/test_decoding/test_decoding.c
@@ -31,6 +31,7 @@ typedef struct
bool include_timestamp;
bool skip_empty_xacts;
bool only_local;
+ bool include_sequences;
} TestDecodingData;
/*
@@ -72,6 +73,10 @@ static void pg_decode_message(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn, XLogRecPtr message_lsn,
bool transactional, const char *prefix,
Size sz, const char *message);
+static void pg_decode_sequence(LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn, XLogRecPtr sequence_lsn,
+ Relation rel,
+ int64 last_value, int64 log_cnt, bool is_called);
static bool pg_decode_filter_prepare(LogicalDecodingContext *ctx,
TransactionId xid,
const char *gid);
@@ -112,6 +117,10 @@ static void pg_decode_stream_message(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn, XLogRecPtr message_lsn,
bool transactional, const char *prefix,
Size sz, const char *message);
+static void pg_decode_stream_sequence(LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn, XLogRecPtr sequence_lsn,
+ Relation rel,
+ int64 last_value, int64 log_cnt, bool is_called);
static void pg_decode_stream_truncate(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn,
int nrelations, Relation relations[],
@@ -137,6 +146,7 @@ _PG_output_plugin_init(OutputPluginCallbacks *cb)
cb->filter_by_origin_cb = pg_decode_filter;
cb->shutdown_cb = pg_decode_shutdown;
cb->message_cb = pg_decode_message;
+ cb->sequence_cb = pg_decode_sequence;
cb->filter_prepare_cb = pg_decode_filter_prepare;
cb->begin_prepare_cb = pg_decode_begin_prepare_txn;
cb->prepare_cb = pg_decode_prepare_txn;
@@ -149,6 +159,7 @@ _PG_output_plugin_init(OutputPluginCallbacks *cb)
cb->stream_commit_cb = pg_decode_stream_commit;
cb->stream_change_cb = pg_decode_stream_change;
cb->stream_message_cb = pg_decode_stream_message;
+ cb->stream_sequence_cb = pg_decode_stream_sequence;
cb->stream_truncate_cb = pg_decode_stream_truncate;
}
@@ -169,6 +180,7 @@ pg_decode_startup(LogicalDecodingContext *ctx, OutputPluginOptions *opt,
data->include_xids = true;
data->include_timestamp = false;
data->skip_empty_xacts = false;
+ data->include_sequences = true;
data->only_local = false;
ctx->output_plugin_private = data;
@@ -261,6 +273,17 @@ pg_decode_startup(LogicalDecodingContext *ctx, OutputPluginOptions *opt,
errmsg("could not parse value \"%s\" for parameter \"%s\"",
strVal(elem->arg), elem->defname)));
}
+ else if (strcmp(elem->defname, "include-sequences") == 0)
+ {
+
+ if (elem->arg == NULL)
+ data->include_sequences = false;
+ else if (!parse_bool(strVal(elem->arg), &data->include_sequences))
+ ereport(ERROR,
+ (errcode(ERRCODE_INVALID_PARAMETER_VALUE),
+ errmsg("could not parse value \"%s\" for parameter \"%s\"",
+ strVal(elem->arg), elem->defname)));
+ }
else
{
ereport(ERROR,
@@ -752,6 +775,34 @@ pg_decode_message(LogicalDecodingContext *ctx,
OutputPluginWrite(ctx, true);
}
+static void
+pg_decode_sequence(LogicalDecodingContext *ctx, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ TestDecodingData *data = ctx->output_plugin_private;
+ TestDecodingTxnData *txndata = txn->output_plugin_private;
+
+ if (!data->include_sequences)
+ return;
+
+ /* output BEGIN if we haven't yet, but only for the transactional case */
+ if (data->skip_empty_xacts && !txndata->xact_wrote_changes)
+ {
+ pg_output_begin(ctx, data, txn, false);
+ }
+ txndata->xact_wrote_changes = true;
+
+ OutputPluginPrepareWrite(ctx, true);
+ appendStringInfoString(ctx->out, "sequence ");
+ appendStringInfoString(ctx->out,
+ quote_qualified_identifier(get_namespace_name(get_rel_namespace(RelationGetRelid(rel))),
+ RelationGetRelationName(rel)));
+ appendStringInfo(ctx->out, ": last_value: %zu log_cnt: %zu is_called:%d",
+ last_value, log_cnt, is_called);
+ OutputPluginWrite(ctx, true);
+}
+
static void
pg_decode_stream_start(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn)
@@ -951,6 +1002,36 @@ pg_decode_stream_message(LogicalDecodingContext *ctx,
OutputPluginWrite(ctx, true);
}
+static void
+pg_decode_stream_sequence(LogicalDecodingContext *ctx, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ TestDecodingData *data = ctx->output_plugin_private;
+ TestDecodingTxnData *txndata = txn->output_plugin_private;
+
+ if (!data->include_sequences)
+ return;
+
+ /* output BEGIN if we haven't yet, but only for the transactional case */
+ {
+ if (data->skip_empty_xacts && !txndata->xact_wrote_changes)
+ {
+ pg_output_begin(ctx, data, txn, false);
+ }
+ txndata->xact_wrote_changes = true;
+ }
+
+ OutputPluginPrepareWrite(ctx, true);
+ appendStringInfoString(ctx->out, "streaming sequence ");
+ appendStringInfoString(ctx->out,
+ quote_qualified_identifier(get_namespace_name(get_rel_namespace(RelationGetRelid(rel))),
+ RelationGetRelationName(rel)));
+ appendStringInfo(ctx->out, ": last_value: " INT64_FORMAT " log_cnt: " INT64_FORMAT " is_called:%d",
+ last_value, log_cnt, is_called);
+ OutputPluginWrite(ctx, true);
+}
+
/*
* In streaming mode, we don't display the detailed information of Truncate.
* See pg_decode_stream_change.
diff --git a/doc/src/sgml/catalogs.sgml b/doc/src/sgml/catalogs.sgml
index 2ce539aaf0..786f248817 100644
--- a/doc/src/sgml/catalogs.sgml
+++ b/doc/src/sgml/catalogs.sgml
@@ -6373,6 +6373,16 @@ SCRAM-SHA-256$<replaceable><iteration count></replaceable>:<replaceable>&l
Reference to schema
</para></entry>
</row>
+
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ <structfield>pntype</structfield> <type>char</type>
+ Determines which object type is included from this schema.
+ </para>
+ <para>
+ Reference to schema
+ </para></entry>
+ </row>
</tbody>
</tgroup>
</table>
@@ -9590,6 +9600,11 @@ SCRAM-SHA-256$<replaceable><iteration count></replaceable>:<replaceable>&l
</para></entry>
</row>
+ <row>
+ <entry><link linkend="view-pg-publication-sequences"><structname>pg_publication_sequences</structname></link></entry>
+ <entry>publications and their associated sequences</entry>
+ </row>
+
<row>
<entry role="catalog_table_entry"><para role="column_definition">
<structfield>umoptions</structfield> <type>text[]</type>
diff --git a/doc/src/sgml/logicaldecoding.sgml b/doc/src/sgml/logicaldecoding.sgml
index 38ee69dccc..58d4067cce 100644
--- a/doc/src/sgml/logicaldecoding.sgml
+++ b/doc/src/sgml/logicaldecoding.sgml
@@ -458,6 +458,7 @@ typedef struct OutputPluginCallbacks
LogicalDecodeTruncateCB truncate_cb;
LogicalDecodeCommitCB commit_cb;
LogicalDecodeMessageCB message_cb;
+ LogicalDecodeSequenceCB sequence_cb;
LogicalDecodeFilterByOriginCB filter_by_origin_cb;
LogicalDecodeShutdownCB shutdown_cb;
LogicalDecodeFilterPrepareCB filter_prepare_cb;
@@ -472,6 +473,7 @@ typedef struct OutputPluginCallbacks
LogicalDecodeStreamCommitCB stream_commit_cb;
LogicalDecodeStreamChangeCB stream_change_cb;
LogicalDecodeStreamMessageCB stream_message_cb;
+ LogicalDecodeStreamSequenceCB stream_sequence_cb;
LogicalDecodeStreamTruncateCB stream_truncate_cb;
} OutputPluginCallbacks;
@@ -481,9 +483,11 @@ typedef void (*LogicalOutputPluginInit) (struct OutputPluginCallbacks *cb);
and <function>commit_cb</function> callbacks are required,
while <function>startup_cb</function>,
<function>filter_by_origin_cb</function>, <function>truncate_cb</function>,
- and <function>shutdown_cb</function> are optional.
- If <function>truncate_cb</function> is not set but a
+ <function>sequence_cb</function>, and <function>shutdown_cb</function> are
+ optional. If <function>truncate_cb</function> is not set but a
<command>TRUNCATE</command> is to be decoded, the action will be ignored.
+ Similarly, if <function>sequence_cb</function> is not set and a sequence
+ change is to be decoded, the action will be ignored.
</para>
<para>
@@ -492,7 +496,8 @@ typedef void (*LogicalOutputPluginInit) (struct OutputPluginCallbacks *cb);
<function>stream_stop_cb</function>, <function>stream_abort_cb</function>,
<function>stream_commit_cb</function>, <function>stream_change_cb</function>,
and <function>stream_prepare_cb</function>
- are required, while <function>stream_message_cb</function> and
+ are required, while <function>stream_message_cb</function>,
+ <function>stream_sequence_cb</function>, and
<function>stream_truncate_cb</function> are optional.
</para>
@@ -808,6 +813,35 @@ typedef void (*LogicalDecodeMessageCB) (struct LogicalDecodingContext *ctx,
</para>
</sect3>
+ <sect3 id="logicaldecoding-output-plugin-sequence">
+ <title>Sequence Callback</title>
+
+ <para>
+ The optional <function>sequence_cb</function> callback is called for
+ actions that update a sequence value.
+<programlisting>
+typedef void (*LogicalDecodeSequenceCB) (struct LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional,
+ int64 last_value,
+ int64 log_cnt,
+ bool is_called);
+</programlisting>
+ The <parameter>txn</parameter> parameter contains meta information about
+ the transaction the sequence change is part of. Note however that for
+ non-transactional increments, the transaction may be either NULL or not
+ NULL, depending on if the transaction already has an XID assigned.
+ The <parameter>sequence_lsn</parameter> has the WAL location of the
+ sequence update. <parameter>transactional</parameter> says if the
+ sequence has to be replayed as part of the transaction or directly.
+
+ The <parameter>last_value</parameter>, <parameter>log_cnt</parameter> and
+ <parameter>is_called</parameter> parameters describe the sequence change.
+ </para>
+ </sect3>
+
<sect3 id="logicaldecoding-output-plugin-filter-prepare">
<title>Prepare Filter Callback</title>
@@ -1017,6 +1051,26 @@ typedef void (*LogicalDecodeStreamMessageCB) (struct LogicalDecodingContext *ctx
</para>
</sect3>
+ <sect3 id="logicaldecoding-output-plugin-stream-sequence">
+ <title>Stream Sequence Callback</title>
+ <para>
+ The optional <function>stream_sequence_cb</function> callback is called
+ for actions that change a sequence in a block of streamed changes
+ (demarcated by <function>stream_start_cb</function> and
+ <function>stream_stop_cb</function> calls).
+<programlisting>
+typedef void (*LogicalDecodeStreamSequenceCB) (struct LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional,
+ int64 last_value,
+ int64 log_cnt,
+ bool is_called);
+</programlisting>
+ </para>
+ </sect3>
+
<sect3 id="logicaldecoding-output-plugin-stream-truncate">
<title>Stream Truncate Callback</title>
<para>
@@ -1197,8 +1251,9 @@ OutputPluginWrite(ctx, true);
in-progress transactions. There are multiple required streaming callbacks
(<function>stream_start_cb</function>, <function>stream_stop_cb</function>,
<function>stream_abort_cb</function>, <function>stream_commit_cb</function>
- and <function>stream_change_cb</function>) and two optional callbacks
- (<function>stream_message_cb</function> and <function>stream_truncate_cb</function>).
+ and <function>stream_change_cb</function>) and multiple optional callbacks
+ (<function>stream_message_cb</function>, <function>stream_sequence_cb</function>,
+ and <function>stream_truncate_cb</function>).
Also, if streaming of two-phase commands is to be supported, then additional
callbacks must be provided. (See <xref linkend="logicaldecoding-two-phase-commits"/>
for details).
diff --git a/doc/src/sgml/protocol.sgml b/doc/src/sgml/protocol.sgml
index 87870c5b10..0841ba30bb 100644
--- a/doc/src/sgml/protocol.sgml
+++ b/doc/src/sgml/protocol.sgml
@@ -6397,6 +6397,105 @@ psql "dbname=postgres replication=database" -c "IDENTIFY_SYSTEM;"
</listitem>
</varlistentry>
+ <varlistentry id="protocol-logicalrep-message-formats-Sequence">
+ <term>Sequence</term>
+ <listitem>
+ <variablelist>
+ <varlistentry>
+ <term>Byte1('X')</term>
+ <listitem>
+ <para>
+ Identifies the message as a sequence message.
+ </para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term>Int32 (TransactionId)</term>
+ <listitem>
+ <para>
+ Xid of the transaction (only present for streamed transactions).
+ This field is available since protocol version 2.
+ </para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term>Int8(0)</term>
+ <listitem>
+ <para>
+ Flags; currently unused.
+ </para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term>Int64 (XLogRecPtr)</term>
+ <listitem>
+ <para>
+ The LSN of the sequence increment.
+ </para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term>String</term>
+ <listitem>
+ <para>
+ Namespace (empty string for <literal>pg_catalog</literal>).
+ </para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term>String</term>
+ <listitem>
+ <para>
+ Relation name.
+ </para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term>Int8</term>
+ <listitem>
+ <para>
+ 1 if the sequence update is transactional, 0 otherwise.
+ </para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term>Int64</term>
+ <listitem>
+ <para>
+ <structfield>last_value</structfield> value of the sequence.
+ </para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term>Int64</term>
+ <listitem>
+ <para>
+ <structfield>log_cnt</structfield> value of the sequence.
+ </para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term>Int8</term>
+ <listitem>
+ <para>
+ <structfield>is_called</structfield> value of the sequence.
+ </para>
+ </listitem>
+ </varlistentry>
+
+ </variablelist>
+ </listitem>
+ </varlistentry>
+
<varlistentry id="protocol-logicalrep-message-formats-Type">
<term>Type</term>
<listitem>
diff --git a/doc/src/sgml/ref/alter_publication.sgml b/doc/src/sgml/ref/alter_publication.sgml
index 3e338f4cc5..3f00e9fd4f 100644
--- a/doc/src/sgml/ref/alter_publication.sgml
+++ b/doc/src/sgml/ref/alter_publication.sgml
@@ -31,7 +31,9 @@ ALTER PUBLICATION <replaceable class="parameter">name</replaceable> RENAME TO <r
<phrase>where <replaceable class="parameter">publication_object</replaceable> is one of:</phrase>
TABLE [ ONLY ] <replaceable class="parameter">table_name</replaceable> [ * ] [ ( <replaceable class="parameter">column_name</replaceable> [, ... ] ) ] [ WHERE ( <replaceable class="parameter">expression</replaceable> ) ] [, ... ]
+ SEQUENCE <replaceable class="parameter">sequence_name</replaceable> [, ... ]
ALL TABLES IN SCHEMA { <replaceable class="parameter">schema_name</replaceable> | CURRENT_SCHEMA } [, ... ]
+ ALL SEQUENCES IN SCHEMA { <replaceable class="parameter">schema_name</replaceable> | CURRENT_SCHEMA } [, ... ]
</synopsis>
</refsynopsisdiv>
@@ -44,13 +46,13 @@ ALTER PUBLICATION <replaceable class="parameter">name</replaceable> RENAME TO <r
</para>
<para>
- The first three variants change which tables/schemas are part of the
- publication. The <literal>SET</literal> clause will replace the list of
- tables/schemas in the publication with the specified list; the existing
- tables/schemas that were present in the publication will be removed. The
- <literal>ADD</literal> and <literal>DROP</literal> clauses will add and
- remove one or more tables/schemas from the publication. Note that adding
- tables/schemas to a publication that is already subscribed to will require an
+ The first three variants change which objects (tables, sequences or schemas)
+ are part of the publication. The <literal>SET</literal> clause will replace
+ the list of objects in the publication with the specified list; the existing
+ objects that were present in the publication will be removed.
+ The <literal>ADD</literal> and <literal>DROP</literal> clauses will add and
+ remove one or more objects from the publication. Note that adding objects
+ to a publication that is already subscribed to will require an
<literal>ALTER SUBSCRIPTION ... REFRESH PUBLICATION</literal> action on the
subscribing side in order to become effective. Note also that the combination
of <literal>DROP</literal> with a <literal>WHERE</literal> clause is not
@@ -140,6 +142,15 @@ ALTER PUBLICATION <replaceable class="parameter">name</replaceable> RENAME TO <r
</listitem>
</varlistentry>
+ <varlistentry>
+ <term><replaceable class="parameter">sequence_name</replaceable></term>
+ <listitem>
+ <para>
+ Name of an existing sequence.
+ </para>
+ </listitem>
+ </varlistentry>
+
<varlistentry>
<term><replaceable class="parameter">schema_name</replaceable></term>
<listitem>
diff --git a/doc/src/sgml/ref/alter_subscription.sgml b/doc/src/sgml/ref/alter_subscription.sgml
index 64efc21f53..e34b96ba01 100644
--- a/doc/src/sgml/ref/alter_subscription.sgml
+++ b/doc/src/sgml/ref/alter_subscription.sgml
@@ -150,8 +150,8 @@ ALTER SUBSCRIPTION <replaceable class="parameter">name</replaceable> RENAME TO <
<listitem>
<para>
Fetch missing table information from publisher. This will start
- replication of tables that were added to the subscribed-to publications
- since <command>CREATE SUBSCRIPTION</command> or
+ replication of tables and sequences that were added to the subscribed-to
+ publications since <command>CREATE SUBSCRIPTION</command> or
the last invocation of <command>REFRESH PUBLICATION</command>.
</para>
@@ -169,8 +169,8 @@ ALTER SUBSCRIPTION <replaceable class="parameter">name</replaceable> RENAME TO <
The default is <literal>true</literal>.
</para>
<para>
- Previously subscribed tables are not copied, even if a table's row
- filter <literal>WHERE</literal> clause has since been modified.
+ Previously subscribed tables and sequences are not copied, even if a
+ table's row filter <literal>WHERE</literal> clause has since been modified.
</para>
</listitem>
</varlistentry>
diff --git a/doc/src/sgml/ref/create_publication.sgml b/doc/src/sgml/ref/create_publication.sgml
index 5790d76270..999482c9dd 100644
--- a/doc/src/sgml/ref/create_publication.sgml
+++ b/doc/src/sgml/ref/create_publication.sgml
@@ -22,14 +22,21 @@ PostgreSQL documentation
<refsynopsisdiv>
<synopsis>
CREATE PUBLICATION <replaceable class="parameter">name</replaceable>
- [ FOR ALL TABLES
+ [ FOR ALL <replaceable class="parameter">object_type</replaceable> [, ...]
| FOR <replaceable class="parameter">publication_object</replaceable> [, ... ] ]
[ WITH ( <replaceable class="parameter">publication_parameter</replaceable> [= <replaceable class="parameter">value</replaceable>] [, ... ] ) ]
+<phrase>where <replaceable class="parameter">object type</replaceable> is one of:</phrase>
+
+ TABLES
+ SEQUENCES
+
<phrase>where <replaceable class="parameter">publication_object</replaceable> is one of:</phrase>
TABLE [ ONLY ] <replaceable class="parameter">table_name</replaceable> [ * ] [ ( <replaceable class="parameter">column_name</replaceable> [, ... ] ) ] [ WHERE ( <replaceable class="parameter">expression</replaceable> ) ] [, ... ]
+ SEQUENCE <replaceable class="parameter">sequence_name</replaceable> [ * ] [, ... ]
ALL TABLES IN SCHEMA { <replaceable class="parameter">schema_name</replaceable> | CURRENT_SCHEMA } [, ... ]
+ ALL SEQUENCES IN SCHEMA { <replaceable class="parameter">schema_name</replaceable> | CURRENT_SCHEMA } [, ... ]
</synopsis>
</refsynopsisdiv>
@@ -115,27 +122,43 @@ CREATE PUBLICATION <replaceable class="parameter">name</replaceable>
</listitem>
</varlistentry>
+ <varlistentry>
+ <term><literal>FOR SEQUENCE</literal></term>
+ <listitem>
+ <para>
+ Specifies a list of sequences to add to the publication.
+ </para>
+
+ <para>
+ Specifying a sequence that is part of a schema specified by <literal>FOR
+ ALL SEQUENCES IN SCHEMA</literal> is not supported.
+ </para>
+ </listitem>
+ </varlistentry>
+
<varlistentry>
<term><literal>FOR ALL TABLES</literal></term>
+ <term><literal>FOR ALL SEQUENCES</literal></term>
<listitem>
<para>
- Marks the publication as one that replicates changes for all tables in
- the database, including tables created in the future.
+ Marks the publication as one that replicates changes for all tables/sequences in
+ the database, including tables/sequences created in the future.
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><literal>FOR ALL TABLES IN SCHEMA</literal></term>
+ <term><literal>FOR ALL SEQUENCES IN SCHEMA</literal></term>
<listitem>
<para>
- Marks the publication as one that replicates changes for all tables in
- the specified list of schemas, including tables created in the future.
+ Marks the publication as one that replicates changes for all sequences/tables in
+ the specified list of schemas, including sequences/tables created in the future.
</para>
<para>
- Specifying a schema along with a table which belongs to the specified
- schema using <literal>FOR TABLE</literal> is not supported.
+ Specifying a schema along with a sequence/table which belongs to the specified
+ schema using <literal>FOR SEQUENCE</literal>/<literal>FOR TABLE</literal> is not supported.
</para>
<para>
@@ -221,10 +244,9 @@ CREATE PUBLICATION <replaceable class="parameter">name</replaceable>
<title>Notes</title>
<para>
- If <literal>FOR TABLE</literal>, <literal>FOR ALL TABLES</literal> or
- <literal>FOR ALL TABLES IN SCHEMA</literal> are not specified, then the
- publication starts out with an empty set of tables. That is useful if
- tables or schemas are to be added later.
+ If <literal>FOR TABLE</literal>, <literal>FOR SEQUENCE</literal>, etc. is
+ not specified, then the publication starts out with an empty set of tables
+ and sequences. That is useful if objects are to be added later.
</para>
<para>
@@ -239,10 +261,9 @@ CREATE PUBLICATION <replaceable class="parameter">name</replaceable>
</para>
<para>
- To add a table to a publication, the invoking user must have ownership
- rights on the table. The <command>FOR ALL TABLES</command> and
- <command>FOR ALL TABLES IN SCHEMA</command> clauses require the invoking
- user to be a superuser.
+ To add a table or a sequence to a publication, the invoking user must
+ have ownership rights on the object. The <command>FOR ALL ...</command>
+ clauses require the invoking user to be a superuser.
</para>
<para>
diff --git a/doc/src/sgml/system-views.sgml b/doc/src/sgml/system-views.sgml
index 44aa70a031..344fe5b589 100644
--- a/doc/src/sgml/system-views.sgml
+++ b/doc/src/sgml/system-views.sgml
@@ -126,6 +126,11 @@
<entry>prepared transactions</entry>
</row>
+ <row>
+ <entry><link linkend="view-pg-publication-sequences"><structname>pg_publication_sequences</structname></link></entry>
+ <entry>publications and information of their associated sequences</entry>
+ </row>
+
<row>
<entry><link linkend="view-pg-publication-tables"><structname>pg_publication_tables</structname></link></entry>
<entry>publications and information of their associated tables</entry>
@@ -2077,6 +2082,72 @@ SELECT * FROM pg_locks pl LEFT JOIN pg_prepared_xacts ppx
</sect1>
+ <sect1 id="view-pg-publication-sequences">
+ <title><structname>pg_publication_sequences</structname></title>
+
+ <indexterm zone="view-pg-publication-sequences">
+ <primary>pg_publication_sequences</primary>
+ </indexterm>
+
+ <para>
+ The view <structname>pg_publication_sequences</structname> provides
+ information about the mapping between publications and the sequences they
+ contain. Unlike the underlying catalog
+ <link linkend="catalog-pg-publication-rel"><structname>pg_publication_rel</structname></link>,
+ this view expands
+ publications defined as <literal>FOR ALL SEQUENCES</literal>, so for such
+ publications there will be a row for each eligible sequence.
+ </para>
+
+ <table>
+ <title><structname>pg_publication_sequences</structname> Columns</title>
+ <tgroup cols="1">
+ <thead>
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ Column Type
+ </para>
+ <para>
+ Description
+ </para></entry>
+ </row>
+ </thead>
+
+ <tbody>
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ <structfield>pubname</structfield> <type>name</type>
+ (references <link linkend="catalog-pg-publication"><structname>pg_publication</structname></link>.<structfield>pubname</structfield>)
+ </para>
+ <para>
+ Name of publication
+ </para></entry>
+ </row>
+
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ <structfield>schemaname</structfield> <type>name</type>
+ (references <link linkend="catalog-pg-namespace"><structname>pg_namespace</structname></link>.<structfield>nspname</structfield>)
+ </para>
+ <para>
+ Name of schema containing sequence
+ </para></entry>
+ </row>
+
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ <structfield>sequencename</structfield> <type>name</type>
+ (references <link linkend="catalog-pg-class"><structname>pg_class</structname></link>.<structfield>relname</structfield>)
+ </para>
+ <para>
+ Name of sequence
+ </para></entry>
+ </row>
+ </tbody>
+ </tgroup>
+ </table>
+ </sect1>
+
<sect1 id="view-pg-publication-tables">
<title><structname>pg_publication_tables</structname></title>
diff --git a/src/backend/access/rmgrdesc/logicalmsgdesc.c b/src/backend/access/rmgrdesc/logicalmsgdesc.c
index 08e03aa30d..deb89916c3 100644
--- a/src/backend/access/rmgrdesc/logicalmsgdesc.c
+++ b/src/backend/access/rmgrdesc/logicalmsgdesc.c
@@ -40,6 +40,16 @@ logicalmsg_desc(StringInfo buf, XLogReaderState *record)
sep = " ";
}
}
+ else if (info == XLOG_LOGICAL_SEQUENCE)
+ {
+ xl_logical_sequence *xlrec = (xl_logical_sequence *) rec;
+
+ appendStringInfo(buf, "rel %u/%u/%u last: %lu log_cnt: %lu is_called: %d",
+ xlrec->locator.spcOid,
+ xlrec->locator.dbOid,
+ xlrec->locator.relNumber,
+ xlrec->last, xlrec->log_cnt, xlrec->is_called);
+ }
}
const char *
@@ -48,5 +58,8 @@ logicalmsg_identify(uint8 info)
if ((info & ~XLR_INFO_MASK) == XLOG_LOGICAL_MESSAGE)
return "MESSAGE";
+ if ((info & ~XLR_INFO_MASK) == XLOG_LOGICAL_SEQUENCE)
+ return "SEQUENCE";
+
return NULL;
}
diff --git a/src/backend/access/transam/xact.c b/src/backend/access/transam/xact.c
index 50f092d7eb..376e0ca342 100644
--- a/src/backend/access/transam/xact.c
+++ b/src/backend/access/transam/xact.c
@@ -37,6 +37,7 @@
#include "catalog/storage.h"
#include "commands/async.h"
#include "commands/tablecmds.h"
+#include "commands/sequence.h"
#include "commands/trigger.h"
#include "common/pg_prng.h"
#include "executor/spi.h"
@@ -2237,6 +2238,15 @@ CommitTransaction(void)
/* Prevent cancel/die interrupt while cleaning up */
HOLD_INTERRUPTS();
+ /*
+ * Write state of sequences to WAL. We need to do this early enough so
+ * that we can still write stuff to WAL, but probably before changing
+ * the transaction state.
+ *
+ * XXX Should this happen before/after holding the interrupts?
+ */
+ AtEOXact_Sequences(true);
+
/* Commit updates to the relation map --- do this as late as possible */
AtEOXact_RelationMap(true, is_parallel_worker);
@@ -2513,6 +2523,15 @@ PrepareTransaction(void)
/* Prevent cancel/die interrupt while cleaning up */
HOLD_INTERRUPTS();
+ /*
+ * Write state of sequences to WAL. We need to do this early enough so
+ * that we can still write stuff to WAL, but probably before changing
+ * the transaction state.
+ *
+ * XXX Should this happen before/after holding the interrupts?
+ */
+ AtEOXact_Sequences(true);
+
/*
* set the current transaction state information appropriately during
* prepare processing
@@ -2748,6 +2767,12 @@ AbortTransaction(void)
TransStateAsString(s->state));
Assert(s->parent == NULL);
+ /*
+ * For transaction abort we don't need to write anything to WAL, we just
+ * do cleanup of the hash table etc.
+ */
+ AtEOXact_Sequences(false);
+
/*
* set the current transaction state information appropriately during the
* abort processing
diff --git a/src/backend/catalog/objectaddress.c b/src/backend/catalog/objectaddress.c
index 798c1a2d1e..8a9cad0993 100644
--- a/src/backend/catalog/objectaddress.c
+++ b/src/backend/catalog/objectaddress.c
@@ -1959,12 +1959,14 @@ get_object_address_publication_schema(List *object, bool missing_ok)
char *pubname;
char *schemaname;
Oid schemaid;
+ char *objtype;
ObjectAddressSet(address, PublicationNamespaceRelationId, InvalidOid);
/* Fetch schema name and publication name from input list */
schemaname = strVal(linitial(object));
pubname = strVal(lsecond(object));
+ objtype = strVal(lthird(object));
schemaid = get_namespace_oid(schemaname, missing_ok);
if (!OidIsValid(schemaid))
@@ -1977,10 +1979,12 @@ get_object_address_publication_schema(List *object, bool missing_ok)
/* Find the publication schema mapping in syscache */
address.objectId =
- GetSysCacheOid2(PUBLICATIONNAMESPACEMAP,
+ GetSysCacheOid3(PUBLICATIONNAMESPACEMAP,
Anum_pg_publication_namespace_oid,
ObjectIdGetDatum(schemaid),
- ObjectIdGetDatum(pub->oid));
+ ObjectIdGetDatum(pub->oid),
+ CharGetDatum(objtype[0]));
+
if (!OidIsValid(address.objectId) && !missing_ok)
ereport(ERROR,
(errcode(ERRCODE_UNDEFINED_OBJECT),
@@ -2257,7 +2261,6 @@ pg_get_object_address(PG_FUNCTION_ARGS)
case OBJECT_DOMCONSTRAINT:
case OBJECT_CAST:
case OBJECT_USER_MAPPING:
- case OBJECT_PUBLICATION_NAMESPACE:
case OBJECT_PUBLICATION_REL:
case OBJECT_DEFACL:
case OBJECT_TRANSFORM:
@@ -2282,6 +2285,7 @@ pg_get_object_address(PG_FUNCTION_ARGS)
/* fall through to check args length */
/* FALLTHROUGH */
case OBJECT_OPERATOR:
+ case OBJECT_PUBLICATION_NAMESPACE:
if (list_length(args) != 2)
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
@@ -2353,6 +2357,8 @@ pg_get_object_address(PG_FUNCTION_ARGS)
objnode = (Node *) list_make2(name, linitial(args));
break;
case OBJECT_PUBLICATION_NAMESPACE:
+ objnode = (Node *) list_make3(linitial(name), linitial(args), lsecond(args));
+ break;
case OBJECT_USER_MAPPING:
objnode = (Node *) list_make2(linitial(name), linitial(args));
break;
@@ -2908,11 +2914,12 @@ get_catalog_object_by_oid(Relation catalog, AttrNumber oidcol, Oid objectId)
*
* Get publication name and schema name from the object address into pubname and
* nspname. Both pubname and nspname are palloc'd strings which will be freed by
- * the caller.
+ * the caller. The last parameter specifies which object type is included from
+ * the schema.
*/
static bool
getPublicationSchemaInfo(const ObjectAddress *object, bool missing_ok,
- char **pubname, char **nspname)
+ char **pubname, char **nspname, char **objtype)
{
HeapTuple tup;
Form_pg_publication_namespace pnform;
@@ -2948,6 +2955,13 @@ getPublicationSchemaInfo(const ObjectAddress *object, bool missing_ok,
return false;
}
+ /*
+ * The type is always a single character, but we need to pass it as a string,
+ * so allocate two charaters and set the first one. The second one is \0.
+ */
+ *objtype = palloc0(2);
+ *objtype[0] = pnform->pntype;
+
ReleaseSysCache(tup);
return true;
}
@@ -4021,15 +4035,17 @@ getObjectDescription(const ObjectAddress *object, bool missing_ok)
{
char *pubname;
char *nspname;
+ char *objtype;
if (!getPublicationSchemaInfo(object, missing_ok,
- &pubname, &nspname))
+ &pubname, &nspname, &objtype))
break;
- appendStringInfo(&buffer, _("publication of schema %s in publication %s"),
- nspname, pubname);
+ appendStringInfo(&buffer, _("publication of schema %s in publication %s type %s"),
+ nspname, pubname, objtype);
pfree(pubname);
pfree(nspname);
+ pfree(objtype);
break;
}
@@ -5902,18 +5918,24 @@ getObjectIdentityParts(const ObjectAddress *object,
{
char *pubname;
char *nspname;
+ char *objtype;
if (!getPublicationSchemaInfo(object, missing_ok, &pubname,
- &nspname))
+ &nspname, &objtype))
break;
- appendStringInfo(&buffer, "%s in publication %s",
- nspname, pubname);
+ appendStringInfo(&buffer, "%s in publication %s type %s",
+ nspname, pubname, objtype);
if (objargs)
*objargs = list_make1(pubname);
else
pfree(pubname);
+ if (objargs)
+ *objargs = lappend(*objargs, objtype);
+ else
+ pfree(objtype);
+
if (objname)
*objname = list_make1(nspname);
else
diff --git a/src/backend/catalog/pg_publication.c b/src/backend/catalog/pg_publication.c
index 6af3570005..36e59ecb0b 100644
--- a/src/backend/catalog/pg_publication.c
+++ b/src/backend/catalog/pg_publication.c
@@ -55,9 +55,10 @@ static void publication_translate_columns(Relation targetrel, List *columns,
static void
check_publication_add_relation(Relation targetrel)
{
- /* Must be a regular or partitioned table */
+ /* Must be a regular or partitioned table, or a sequence */
if (RelationGetForm(targetrel)->relkind != RELKIND_RELATION &&
- RelationGetForm(targetrel)->relkind != RELKIND_PARTITIONED_TABLE)
+ RelationGetForm(targetrel)->relkind != RELKIND_PARTITIONED_TABLE &&
+ RelationGetForm(targetrel)->relkind != RELKIND_SEQUENCE)
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("cannot add relation \"%s\" to publication",
@@ -134,7 +135,8 @@ static bool
is_publishable_class(Oid relid, Form_pg_class reltuple)
{
return (reltuple->relkind == RELKIND_RELATION ||
- reltuple->relkind == RELKIND_PARTITIONED_TABLE) &&
+ reltuple->relkind == RELKIND_PARTITIONED_TABLE ||
+ reltuple->relkind == RELKIND_SEQUENCE) &&
!IsCatalogRelationOid(relid) &&
reltuple->relpersistence == RELPERSISTENCE_PERMANENT &&
relid >= FirstNormalObjectId;
@@ -171,6 +173,52 @@ pg_relation_is_publishable(PG_FUNCTION_ARGS)
PG_RETURN_BOOL(result);
}
+/*
+ * Check the character is a valid object type for schema publication.
+ *
+ * This recognizes either 't' for tables or 's' for sequences. Places that
+ * need to handle 'u' for unsupported relkinds need to do that explicitlyl
+ */
+static void
+AssertObjectTypeValid(char objectType)
+{
+#ifdef USE_ASSERT_CHECKING
+ Assert(objectType == PUB_OBJTYPE_SEQUENCE || objectType == PUB_OBJTYPE_TABLE);
+#endif
+}
+
+/*
+ * Determine object type matching a given a relkind value.
+ */
+char
+pub_get_object_type_for_relkind(char relkind)
+{
+ /* sequence maps directly to sequence relkind */
+ if (relkind == RELKIND_SEQUENCE)
+ return PUB_OBJTYPE_SEQUENCE;
+
+ /* for table, we match either regular or partitioned table */
+ if (relkind == RELKIND_RELATION ||
+ relkind == RELKIND_PARTITIONED_TABLE)
+ return PUB_OBJTYPE_TABLE;
+
+ return PUB_OBJTYPE_UNSUPPORTED;
+}
+
+/*
+ * Determine if publication object type matches the relkind.
+ *
+ * Returns true if the relation matches object type replicated by this schema,
+ * false otherwise.
+ */
+static bool
+pub_object_type_matches_relkind(char objectType, char relkind)
+{
+ AssertObjectTypeValid(objectType);
+
+ return (pub_get_object_type_for_relkind(relkind) == objectType);
+}
+
/*
* Filter out the partitions whose parent tables were also specified in
* the publication.
@@ -230,7 +278,7 @@ is_schema_publication(Oid pubid)
ObjectIdGetDatum(pubid));
scan = systable_beginscan(pubschsrel,
- PublicationNamespacePnnspidPnpubidIndexId,
+ PublicationNamespacePnnspidPnpubidPntypeIndexId,
true, NULL, 1, &scankey);
tup = systable_getnext(scan);
result = HeapTupleIsValid(tup);
@@ -316,7 +364,9 @@ GetTopMostAncestorInPublication(Oid puboid, List *ancestors, int *ancestor_level
}
else
{
- aschemaPubids = GetSchemaPublications(get_rel_namespace(ancestor));
+ /* we only search for ancestors of tables, so PUB_OBJTYPE_TABLE */
+ aschemaPubids = GetSchemaPublications(get_rel_namespace(ancestor),
+ PUB_OBJTYPE_TABLE);
if (list_member_oid(aschemaPubids, puboid))
{
topmost_relid = ancestor;
@@ -585,7 +635,7 @@ pub_collist_to_bitmapset(Bitmapset *columns, Datum pubcols, MemoryContext mcxt)
* Insert new publication / schema mapping.
*/
ObjectAddress
-publication_add_schema(Oid pubid, Oid schemaid, bool if_not_exists)
+publication_add_schema(Oid pubid, Oid schemaid, char objectType, bool if_not_exists)
{
Relation rel;
HeapTuple tup;
@@ -597,6 +647,8 @@ publication_add_schema(Oid pubid, Oid schemaid, bool if_not_exists)
ObjectAddress myself,
referenced;
+ AssertObjectTypeValid(objectType);
+
rel = table_open(PublicationNamespaceRelationId, RowExclusiveLock);
/*
@@ -604,9 +656,10 @@ publication_add_schema(Oid pubid, Oid schemaid, bool if_not_exists)
* duplicates, it's here just to provide nicer error message in common
* case. The real protection is the unique key on the catalog.
*/
- if (SearchSysCacheExists2(PUBLICATIONNAMESPACEMAP,
+ if (SearchSysCacheExists3(PUBLICATIONNAMESPACEMAP,
ObjectIdGetDatum(schemaid),
- ObjectIdGetDatum(pubid)))
+ ObjectIdGetDatum(pubid),
+ CharGetDatum(objectType)))
{
table_close(rel, RowExclusiveLock);
@@ -632,6 +685,8 @@ publication_add_schema(Oid pubid, Oid schemaid, bool if_not_exists)
ObjectIdGetDatum(pubid);
values[Anum_pg_publication_namespace_pnnspid - 1] =
ObjectIdGetDatum(schemaid);
+ values[Anum_pg_publication_namespace_pntype - 1] =
+ CharGetDatum(objectType);
tup = heap_form_tuple(RelationGetDescr(rel), values, nulls);
@@ -657,7 +712,7 @@ publication_add_schema(Oid pubid, Oid schemaid, bool if_not_exists)
* publication_add_relation for why we need to consider all the
* partitions.
*/
- schemaRels = GetSchemaPublicationRelations(schemaid,
+ schemaRels = GetSchemaPublicationRelations(schemaid, objectType,
PUBLICATION_PART_ALL);
InvalidatePublicationRels(schemaRels);
@@ -691,11 +746,14 @@ GetRelationPublications(Oid relid)
/*
* Gets list of relation oids for a publication.
*
- * This should only be used FOR TABLE publications, the FOR ALL TABLES
- * should use GetAllTablesPublicationRelations().
+ * This should only be used FOR TABLE / FOR SEQUENCE publications, the FOR
+ * ALL TABLES / SEQUENCES should use GetAllTablesPublicationRelations()
+ * and GetAllSequencesPublicationRelations().
+ *
+ * XXX pub_partopt only matters for tables, not sequences.
*/
List *
-GetPublicationRelations(Oid pubid, PublicationPartOpt pub_partopt)
+GetPublicationRelations(Oid pubid, char objectType, PublicationPartOpt pub_partopt)
{
List *result;
Relation pubrelsrel;
@@ -703,6 +761,8 @@ GetPublicationRelations(Oid pubid, PublicationPartOpt pub_partopt)
SysScanDesc scan;
HeapTuple tup;
+ AssertObjectTypeValid(objectType);
+
/* Find all publications associated with the relation. */
pubrelsrel = table_open(PublicationRelRelationId, AccessShareLock);
@@ -717,11 +777,29 @@ GetPublicationRelations(Oid pubid, PublicationPartOpt pub_partopt)
result = NIL;
while (HeapTupleIsValid(tup = systable_getnext(scan)))
{
+ char relkind;
Form_pg_publication_rel pubrel;
pubrel = (Form_pg_publication_rel) GETSTRUCT(tup);
- result = GetPubPartitionOptionRelations(result, pub_partopt,
- pubrel->prrelid);
+ relkind = get_rel_relkind(pubrel->prrelid);
+
+ /*
+ * If the relkind does not match the requested object type, ignore the
+ * relation. For example we might be interested only in sequences, so
+ * we ignore tables.
+ */
+ if (!pub_object_type_matches_relkind(objectType, relkind))
+ continue;
+
+ /*
+ * We don't have partitioned sequences, so just add them to the list.
+ * Otherwise consider adding all child relations, if requested.
+ */
+ if (relkind == RELKIND_SEQUENCE)
+ result = lappend_oid(result, pubrel->prrelid);
+ else
+ result = GetPubPartitionOptionRelations(result, pub_partopt,
+ pubrel->prrelid);
}
systable_endscan(scan);
@@ -771,6 +849,43 @@ GetAllTablesPublications(void)
return result;
}
+/*
+ * Gets list of publication oids for publications marked as FOR ALL SEQUENCES.
+ */
+List *
+GetAllSequencesPublications(void)
+{
+ List *result;
+ Relation rel;
+ ScanKeyData scankey;
+ SysScanDesc scan;
+ HeapTuple tup;
+
+ /* Find all publications that are marked as for all sequences. */
+ rel = table_open(PublicationRelationId, AccessShareLock);
+
+ ScanKeyInit(&scankey,
+ Anum_pg_publication_puballsequences,
+ BTEqualStrategyNumber, F_BOOLEQ,
+ BoolGetDatum(true));
+
+ scan = systable_beginscan(rel, InvalidOid, false,
+ NULL, 1, &scankey);
+
+ result = NIL;
+ while (HeapTupleIsValid(tup = systable_getnext(scan)))
+ {
+ Oid oid = ((Form_pg_publication) GETSTRUCT(tup))->oid;
+
+ result = lappend_oid(result, oid);
+ }
+
+ systable_endscan(scan);
+ table_close(rel, AccessShareLock);
+
+ return result;
+}
+
/*
* Gets list of all relation published by FOR ALL TABLES publication(s).
*
@@ -837,28 +952,38 @@ GetAllTablesPublicationRelations(bool pubviaroot)
/*
* Gets the list of schema oids for a publication.
*
- * This should only be used FOR ALL TABLES IN SCHEMA publications.
+ * This should only be used FOR ALL TABLES IN SCHEMA and FOR ALL SEQUENCES
+ * publications.
+ *
+ * 'objectType' determines whether to get FOR TABLE or FOR SEQUENCES schemas
*/
List *
-GetPublicationSchemas(Oid pubid)
+GetPublicationSchemas(Oid pubid, char objectType)
{
List *result = NIL;
Relation pubschsrel;
- ScanKeyData scankey;
+ ScanKeyData scankey[2];
SysScanDesc scan;
HeapTuple tup;
+ AssertObjectTypeValid(objectType);
+
/* Find all schemas associated with the publication */
pubschsrel = table_open(PublicationNamespaceRelationId, AccessShareLock);
- ScanKeyInit(&scankey,
+ ScanKeyInit(&scankey[0],
Anum_pg_publication_namespace_pnpubid,
BTEqualStrategyNumber, F_OIDEQ,
ObjectIdGetDatum(pubid));
+ ScanKeyInit(&scankey[1],
+ Anum_pg_publication_namespace_pntype,
+ BTEqualStrategyNumber, F_CHAREQ,
+ CharGetDatum(objectType));
+
scan = systable_beginscan(pubschsrel,
- PublicationNamespacePnnspidPnpubidIndexId,
- true, NULL, 1, &scankey);
+ PublicationNamespacePnnspidPnpubidPntypeIndexId,
+ true, NULL, 2, scankey);
while (HeapTupleIsValid(tup = systable_getnext(scan)))
{
Form_pg_publication_namespace pubsch;
@@ -876,14 +1001,26 @@ GetPublicationSchemas(Oid pubid)
/*
* Gets the list of publication oids associated with a specified schema.
+ *
+ * objectType specifies whether we're looking for schemas including tables or
+ * sequences.
+ *
+ * Note: relcache calls this for all object types, not just tables and sequences.
+ * Which is why we handle the PUB_OBJTYPE_UNSUPPORTED object type too.
*/
List *
-GetSchemaPublications(Oid schemaid)
+GetSchemaPublications(Oid schemaid, char objectType)
{
List *result = NIL;
CatCList *pubschlist;
int i;
+ /* unsupported object type */
+ if (objectType == PUB_OBJTYPE_UNSUPPORTED)
+ return result;
+
+ AssertObjectTypeValid(objectType);
+
/* Find all publications associated with the schema */
pubschlist = SearchSysCacheList1(PUBLICATIONNAMESPACEMAP,
ObjectIdGetDatum(schemaid));
@@ -891,6 +1028,11 @@ GetSchemaPublications(Oid schemaid)
{
HeapTuple tup = &pubschlist->members[i]->tuple;
Oid pubid = ((Form_pg_publication_namespace) GETSTRUCT(tup))->pnpubid;
+ char pntype = ((Form_pg_publication_namespace) GETSTRUCT(tup))->pntype;
+
+ /* Skip schemas publishing a different object type. */
+ if (pntype != objectType)
+ continue;
result = lappend_oid(result, pubid);
}
@@ -902,9 +1044,13 @@ GetSchemaPublications(Oid schemaid)
/*
* Get the list of publishable relation oids for a specified schema.
+ *
+ * objectType specifies whether this is FOR ALL TABLES IN SCHEMA or FOR ALL
+ * SEQUENCES IN SCHEMA
*/
List *
-GetSchemaPublicationRelations(Oid schemaid, PublicationPartOpt pub_partopt)
+GetSchemaPublicationRelations(Oid schemaid, char objectType,
+ PublicationPartOpt pub_partopt)
{
Relation classRel;
ScanKeyData key[1];
@@ -913,6 +1059,7 @@ GetSchemaPublicationRelations(Oid schemaid, PublicationPartOpt pub_partopt)
List *result = NIL;
Assert(OidIsValid(schemaid));
+ AssertObjectTypeValid(objectType);
classRel = table_open(RelationRelationId, AccessShareLock);
@@ -933,9 +1080,16 @@ GetSchemaPublicationRelations(Oid schemaid, PublicationPartOpt pub_partopt)
continue;
relkind = get_rel_relkind(relid);
- if (relkind == RELKIND_RELATION)
- result = lappend_oid(result, relid);
- else if (relkind == RELKIND_PARTITIONED_TABLE)
+
+ /* Skip if the relkind does not match FOR ALL TABLES / SEQUENCES. */
+ if (!pub_object_type_matches_relkind(objectType, relkind))
+ continue;
+
+ /*
+ * If the object is a partitioned table, lookup all the child relations
+ * (if requested). Otherwise just add the object to the list.
+ */
+ if (relkind == RELKIND_PARTITIONED_TABLE)
{
List *partitionrels = NIL;
@@ -948,7 +1102,11 @@ GetSchemaPublicationRelations(Oid schemaid, PublicationPartOpt pub_partopt)
pub_partopt,
relForm->oid);
result = list_concat_unique_oid(result, partitionrels);
+ continue;
}
+
+ /* non-partitioned tables and sequences */
+ result = lappend_oid(result, relid);
}
table_endscan(scan);
@@ -958,27 +1116,67 @@ GetSchemaPublicationRelations(Oid schemaid, PublicationPartOpt pub_partopt)
/*
* Gets the list of all relations published by FOR ALL TABLES IN SCHEMA
- * publication.
+ * or FOR ALL SEQUENCES IN SCHEMA publication.
*/
List *
-GetAllSchemaPublicationRelations(Oid pubid, PublicationPartOpt pub_partopt)
+GetAllSchemaPublicationRelations(Oid pubid, char objectType,
+ PublicationPartOpt pub_partopt)
{
List *result = NIL;
- List *pubschemalist = GetPublicationSchemas(pubid);
+ List *pubschemalist = GetPublicationSchemas(pubid, objectType);
ListCell *cell;
+ AssertObjectTypeValid(objectType);
+
foreach(cell, pubschemalist)
{
Oid schemaid = lfirst_oid(cell);
List *schemaRels = NIL;
- schemaRels = GetSchemaPublicationRelations(schemaid, pub_partopt);
+ schemaRels = GetSchemaPublicationRelations(schemaid, objectType,
+ pub_partopt);
result = list_concat(result, schemaRels);
}
return result;
}
+/*
+ * Gets list of all relation published by FOR ALL SEQUENCES publication(s).
+ */
+List *
+GetAllSequencesPublicationRelations(void)
+{
+ Relation classRel;
+ ScanKeyData key[1];
+ TableScanDesc scan;
+ HeapTuple tuple;
+ List *result = NIL;
+
+ classRel = table_open(RelationRelationId, AccessShareLock);
+
+ ScanKeyInit(&key[0],
+ Anum_pg_class_relkind,
+ BTEqualStrategyNumber, F_CHAREQ,
+ CharGetDatum(RELKIND_SEQUENCE));
+
+ scan = table_beginscan_catalog(classRel, 1, key);
+
+ while ((tuple = heap_getnext(scan, ForwardScanDirection)) != NULL)
+ {
+ Form_pg_class relForm = (Form_pg_class) GETSTRUCT(tuple);
+ Oid relid = relForm->oid;
+
+ if (is_publishable_class(relid, relForm))
+ result = lappend_oid(result, relid);
+ }
+
+ table_endscan(scan);
+
+ table_close(classRel, AccessShareLock);
+ return result;
+}
+
/*
* Get publication using oid
*
@@ -1001,10 +1199,12 @@ GetPublication(Oid pubid)
pub->oid = pubid;
pub->name = pstrdup(NameStr(pubform->pubname));
pub->alltables = pubform->puballtables;
+ pub->allsequences = pubform->puballsequences;
pub->pubactions.pubinsert = pubform->pubinsert;
pub->pubactions.pubupdate = pubform->pubupdate;
pub->pubactions.pubdelete = pubform->pubdelete;
pub->pubactions.pubtruncate = pubform->pubtruncate;
+ pub->pubactions.pubsequence = pubform->pubsequence;
pub->pubviaroot = pubform->pubviaroot;
ReleaseSysCache(tup);
@@ -1066,10 +1266,12 @@ pg_get_publication_tables(PG_FUNCTION_ARGS)
*schemarelids;
relids = GetPublicationRelations(publication->oid,
+ PUB_OBJTYPE_TABLE,
publication->pubviaroot ?
PUBLICATION_PART_ROOT :
PUBLICATION_PART_LEAF);
schemarelids = GetAllSchemaPublicationRelations(publication->oid,
+ PUB_OBJTYPE_TABLE,
publication->pubviaroot ?
PUBLICATION_PART_ROOT :
PUBLICATION_PART_LEAF);
@@ -1151,3 +1353,71 @@ pg_get_publication_tables(PG_FUNCTION_ARGS)
SRF_RETURN_DONE(funcctx);
}
+
+/*
+ * Returns Oids of sequences in a publication.
+ */
+Datum
+pg_get_publication_sequences(PG_FUNCTION_ARGS)
+{
+ FuncCallContext *funcctx;
+ char *pubname = text_to_cstring(PG_GETARG_TEXT_PP(0));
+ Publication *publication;
+ List *sequences;
+
+ /* stuff done only on the first call of the function */
+ if (SRF_IS_FIRSTCALL())
+ {
+ MemoryContext oldcontext;
+
+ /* create a function context for cross-call persistence */
+ funcctx = SRF_FIRSTCALL_INIT();
+
+ /* switch to memory context appropriate for multiple function calls */
+ oldcontext = MemoryContextSwitchTo(funcctx->multi_call_memory_ctx);
+
+ publication = GetPublicationByName(pubname, false);
+
+ /*
+ * Publications support partitioned tables, although all changes are
+ * replicated using leaf partition identity and schema, so we only
+ * need those.
+ */
+ if (publication->allsequences)
+ sequences = GetAllSequencesPublicationRelations();
+ else
+ {
+ List *relids,
+ *schemarelids;
+
+ relids = GetPublicationRelations(publication->oid,
+ PUB_OBJTYPE_SEQUENCE,
+ publication->pubviaroot ?
+ PUBLICATION_PART_ROOT :
+ PUBLICATION_PART_LEAF);
+ schemarelids = GetAllSchemaPublicationRelations(publication->oid,
+ PUB_OBJTYPE_SEQUENCE,
+ publication->pubviaroot ?
+ PUBLICATION_PART_ROOT :
+ PUBLICATION_PART_LEAF);
+ sequences = list_concat_unique_oid(relids, schemarelids);
+ }
+
+ funcctx->user_fctx = (void *) sequences;
+
+ MemoryContextSwitchTo(oldcontext);
+ }
+
+ /* stuff done on every call of the function */
+ funcctx = SRF_PERCALL_SETUP();
+ sequences = (List *) funcctx->user_fctx;
+
+ if (funcctx->call_cntr < list_length(sequences))
+ {
+ Oid relid = list_nth_oid(sequences, funcctx->call_cntr);
+
+ SRF_RETURN_NEXT(funcctx, ObjectIdGetDatum(relid));
+ }
+
+ SRF_RETURN_DONE(funcctx);
+}
diff --git a/src/backend/catalog/system_views.sql b/src/backend/catalog/system_views.sql
index 5a844b63a1..4aaf31680c 100644
--- a/src/backend/catalog/system_views.sql
+++ b/src/backend/catalog/system_views.sql
@@ -382,6 +382,16 @@ CREATE VIEW pg_publication_tables AS
pg_class C JOIN pg_namespace N ON (N.oid = C.relnamespace)
WHERE C.oid = GPT.relid;
+CREATE VIEW pg_publication_sequences AS
+ SELECT
+ P.pubname AS pubname,
+ N.nspname AS schemaname,
+ C.relname AS sequencename
+ FROM pg_publication P,
+ LATERAL pg_get_publication_sequences(P.pubname) GPS,
+ pg_class C JOIN pg_namespace N ON (N.oid = C.relnamespace)
+ WHERE C.oid = GPS.relid;
+
CREATE VIEW pg_locks AS
SELECT * FROM pg_lock_status() AS L;
diff --git a/src/backend/commands/publicationcmds.c b/src/backend/commands/publicationcmds.c
index 8b574b86c4..6c20ab82ca 100644
--- a/src/backend/commands/publicationcmds.c
+++ b/src/backend/commands/publicationcmds.c
@@ -16,6 +16,7 @@
#include "access/genam.h"
#include "access/htup_details.h"
+#include "access/relation.h"
#include "access/table.h"
#include "access/xact.h"
#include "catalog/catalog.h"
@@ -67,15 +68,17 @@ typedef struct rf_context
} rf_context;
static List *OpenRelIdList(List *relids);
-static List *OpenTableList(List *tables);
-static void CloseTableList(List *rels);
+static List *OpenRelationList(List *rels, char objectType);
+static void CloseRelationList(List *rels);
static void LockSchemaList(List *schemalist);
-static void PublicationAddTables(Oid pubid, List *rels, bool if_not_exists,
+static void PublicationAddRelations(Oid pubid, List *rels, bool if_not_exists,
AlterPublicationStmt *stmt);
-static void PublicationDropTables(Oid pubid, List *rels, bool missing_ok);
-static void PublicationAddSchemas(Oid pubid, List *schemas, bool if_not_exists,
- AlterPublicationStmt *stmt);
-static void PublicationDropSchemas(Oid pubid, List *schemas, bool missing_ok);
+static void PublicationDropRelations(Oid pubid, List *rels, bool missing_ok);
+static void PublicationAddSchemas(Oid pubid, List *schemas, char objectType,
+ bool if_not_exists, AlterPublicationStmt *stmt);
+static void PublicationDropSchemas(Oid pubid, List *schemas, char objectType,
+ bool missing_ok);
+
static void
parse_publication_options(ParseState *pstate,
@@ -95,6 +98,7 @@ parse_publication_options(ParseState *pstate,
pubactions->pubupdate = true;
pubactions->pubdelete = true;
pubactions->pubtruncate = true;
+ pubactions->pubsequence = true;
*publish_via_partition_root = false;
/* Parse options */
@@ -119,6 +123,7 @@ parse_publication_options(ParseState *pstate,
pubactions->pubupdate = false;
pubactions->pubdelete = false;
pubactions->pubtruncate = false;
+ pubactions->pubsequence = false;
*publish_given = true;
publish = defGetString(defel);
@@ -141,6 +146,8 @@ parse_publication_options(ParseState *pstate,
pubactions->pubdelete = true;
else if (strcmp(publish_opt, "truncate") == 0)
pubactions->pubtruncate = true;
+ else if (strcmp(publish_opt, "sequence") == 0)
+ pubactions->pubsequence = true;
else
ereport(ERROR,
(errcode(ERRCODE_SYNTAX_ERROR),
@@ -167,7 +174,8 @@ parse_publication_options(ParseState *pstate,
*/
static void
ObjectsInPublicationToOids(List *pubobjspec_list, ParseState *pstate,
- List **rels, List **schemas)
+ List **tables, List **sequences,
+ List **tables_schemas, List **sequences_schemas)
{
ListCell *cell;
PublicationObjSpec *pubobj;
@@ -185,13 +193,22 @@ ObjectsInPublicationToOids(List *pubobjspec_list, ParseState *pstate,
switch (pubobj->pubobjtype)
{
case PUBLICATIONOBJ_TABLE:
- *rels = lappend(*rels, pubobj->pubtable);
+ *tables = lappend(*tables, pubobj->pubtable);
+ break;
+ case PUBLICATIONOBJ_SEQUENCE:
+ *sequences = lappend(*sequences, pubobj->pubtable);
break;
case PUBLICATIONOBJ_TABLES_IN_SCHEMA:
schemaid = get_namespace_oid(pubobj->name, false);
/* Filter out duplicates if user specifies "sch1, sch1" */
- *schemas = list_append_unique_oid(*schemas, schemaid);
+ *tables_schemas = list_append_unique_oid(*tables_schemas, schemaid);
+ break;
+ case PUBLICATIONOBJ_SEQUENCES_IN_SCHEMA:
+ schemaid = get_namespace_oid(pubobj->name, false);
+
+ /* Filter out duplicates if user specifies "sch1, sch1" */
+ *sequences_schemas = list_append_unique_oid(*sequences_schemas, schemaid);
break;
case PUBLICATIONOBJ_TABLES_IN_CUR_SCHEMA:
search_path = fetch_search_path(false);
@@ -204,7 +221,20 @@ ObjectsInPublicationToOids(List *pubobjspec_list, ParseState *pstate,
list_free(search_path);
/* Filter out duplicates if user specifies "sch1, sch1" */
- *schemas = list_append_unique_oid(*schemas, schemaid);
+ *tables_schemas = list_append_unique_oid(*tables_schemas, schemaid);
+ break;
+ case PUBLICATIONOBJ_SEQUENCES_IN_CUR_SCHEMA:
+ search_path = fetch_search_path(false);
+ if (search_path == NIL) /* nothing valid in search_path? */
+ ereport(ERROR,
+ errcode(ERRCODE_UNDEFINED_SCHEMA),
+ errmsg("no schema has been selected for CURRENT_SCHEMA"));
+
+ schemaid = linitial_oid(search_path);
+ list_free(search_path);
+
+ /* Filter out duplicates if user specifies "sch1, sch1" */
+ *sequences_schemas = list_append_unique_oid(*sequences_schemas, schemaid);
break;
default:
/* shouldn't happen */
@@ -240,6 +270,14 @@ CheckObjSchemaNotAlreadyInPublication(List *rels, List *schemaidlist,
errdetail("Table \"%s\" in schema \"%s\" is already part of the publication, adding the same schema is not supported.",
RelationGetRelationName(rel),
get_namespace_name(relSchemaId)));
+ else if (checkobjtype == PUBLICATIONOBJ_SEQUENCES_IN_SCHEMA)
+ ereport(ERROR,
+ errcode(ERRCODE_INVALID_PARAMETER_VALUE),
+ errmsg("cannot add schema \"%s\" to publication",
+ get_namespace_name(relSchemaId)),
+ errdetail("Sequence \"%s\" in schema \"%s\" is already part of the publication, adding the same schema is not supported.",
+ RelationGetRelationName(rel),
+ get_namespace_name(relSchemaId)));
else if (checkobjtype == PUBLICATIONOBJ_TABLE)
ereport(ERROR,
errcode(ERRCODE_INVALID_PARAMETER_VALUE),
@@ -248,6 +286,14 @@ CheckObjSchemaNotAlreadyInPublication(List *rels, List *schemaidlist,
RelationGetRelationName(rel)),
errdetail("Table's schema \"%s\" is already part of the publication or part of the specified schema list.",
get_namespace_name(relSchemaId)));
+ else if (checkobjtype == PUBLICATIONOBJ_SEQUENCE)
+ ereport(ERROR,
+ errcode(ERRCODE_INVALID_PARAMETER_VALUE),
+ errmsg("cannot add relation \"%s.%s\" to publication",
+ get_namespace_name(relSchemaId),
+ RelationGetRelationName(rel)),
+ errdetail("Sequence's schema \"%s\" is already part of the publication or part of the specified schema list.",
+ get_namespace_name(relSchemaId)));
}
}
}
@@ -754,6 +800,7 @@ CheckPubRelationColumnList(List *tables, const char *queryString,
ObjectAddress
CreatePublication(ParseState *pstate, CreatePublicationStmt *stmt)
{
+ ListCell *lc;
Relation rel;
ObjectAddress myself;
Oid puboid;
@@ -765,8 +812,23 @@ CreatePublication(ParseState *pstate, CreatePublicationStmt *stmt)
bool publish_via_partition_root_given;
bool publish_via_partition_root;
AclResult aclresult;
- List *relations = NIL;
- List *schemaidlist = NIL;
+ List *tables = NIL;
+ List *sequences = NIL;
+ List *tables_schemaidlist = NIL;
+ List *sequences_schemaidlist = NIL;
+
+ bool for_all_tables = false;
+ bool for_all_sequences = false;
+
+ /* Translate the list of object types (represented by strings) to bool flags. */
+ foreach (lc, stmt->for_all_objects)
+ {
+ char *val = strVal(lfirst(lc));
+ if (strcmp(val, "tables") == 0)
+ for_all_tables = true;
+ else if (strcmp(val, "sequences") == 0)
+ for_all_sequences = true;
+ }
/* must have CREATE privilege on database */
aclresult = pg_database_aclcheck(MyDatabaseId, GetUserId(), ACL_CREATE);
@@ -775,11 +837,17 @@ CreatePublication(ParseState *pstate, CreatePublicationStmt *stmt)
get_database_name(MyDatabaseId));
/* FOR ALL TABLES requires superuser */
- if (stmt->for_all_tables && !superuser())
+ if (for_all_tables && !superuser())
ereport(ERROR,
(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
errmsg("must be superuser to create FOR ALL TABLES publication")));
+ /* FOR ALL SEQUENCES requires superuser */
+ if (for_all_sequences && !superuser())
+ ereport(ERROR,
+ (errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
+ errmsg("must be superuser to create FOR ALL SEQUENCES publication")));
+
rel = table_open(PublicationRelationId, RowExclusiveLock);
/* Check if name is used */
@@ -811,7 +879,9 @@ CreatePublication(ParseState *pstate, CreatePublicationStmt *stmt)
Anum_pg_publication_oid);
values[Anum_pg_publication_oid - 1] = ObjectIdGetDatum(puboid);
values[Anum_pg_publication_puballtables - 1] =
- BoolGetDatum(stmt->for_all_tables);
+ BoolGetDatum(for_all_tables);
+ values[Anum_pg_publication_puballsequences - 1] =
+ BoolGetDatum(for_all_sequences);
values[Anum_pg_publication_pubinsert - 1] =
BoolGetDatum(pubactions.pubinsert);
values[Anum_pg_publication_pubupdate - 1] =
@@ -820,6 +890,8 @@ CreatePublication(ParseState *pstate, CreatePublicationStmt *stmt)
BoolGetDatum(pubactions.pubdelete);
values[Anum_pg_publication_pubtruncate - 1] =
BoolGetDatum(pubactions.pubtruncate);
+ values[Anum_pg_publication_pubsequence - 1] =
+ BoolGetDatum(pubactions.pubsequence);
values[Anum_pg_publication_pubviaroot - 1] =
BoolGetDatum(publish_via_partition_root);
@@ -837,28 +909,42 @@ CreatePublication(ParseState *pstate, CreatePublicationStmt *stmt)
CommandCounterIncrement();
/* Associate objects with the publication. */
- if (stmt->for_all_tables)
+ if (for_all_tables || for_all_sequences)
{
/* Invalidate relcache so that publication info is rebuilt. */
CacheInvalidateRelcacheAll();
}
- else
+
+ /*
+ * If the publication might have either tables or sequences (directly or
+ * through a schema), process that.
+ */
+ if (!for_all_tables || !for_all_sequences)
{
- ObjectsInPublicationToOids(stmt->pubobjects, pstate, &relations,
- &schemaidlist);
+ ObjectsInPublicationToOids(stmt->pubobjects, pstate,
+ &tables, &sequences,
+ &tables_schemaidlist,
+ &sequences_schemaidlist);
/* FOR ALL TABLES IN SCHEMA requires superuser */
- if (schemaidlist != NIL && !superuser())
+ if (list_length(tables_schemaidlist) > 0 && !superuser())
ereport(ERROR,
errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
errmsg("must be superuser to create FOR ALL TABLES IN SCHEMA publication"));
- if (relations != NIL)
+ /* FOR ALL SEQUENCES IN SCHEMA requires superuser */
+ if (list_length(sequences_schemaidlist) > 0 && !superuser())
+ ereport(ERROR,
+ errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
+ errmsg("must be superuser to create FOR ALL SEQUENCES IN SCHEMA publication"));
+
+ /* tables added directly */
+ if (tables != NIL)
{
List *rels;
- rels = OpenTableList(relations);
- CheckObjSchemaNotAlreadyInPublication(rels, schemaidlist,
+ rels = OpenRelationList(tables, PUB_OBJTYPE_TABLE);
+ CheckObjSchemaNotAlreadyInPublication(rels, tables_schemaidlist,
PUBLICATIONOBJ_TABLE);
TransformPubWhereClauses(rels, pstate->p_sourcetext,
@@ -867,18 +953,51 @@ CreatePublication(ParseState *pstate, CreatePublicationStmt *stmt)
CheckPubRelationColumnList(rels, pstate->p_sourcetext,
publish_via_partition_root);
- PublicationAddTables(puboid, rels, true, NULL);
- CloseTableList(rels);
+ PublicationAddRelations(puboid, rels, true, NULL);
+ CloseRelationList(rels);
}
- if (schemaidlist != NIL)
+ /* sequences added directly */
+ if (sequences != NIL)
+ {
+ List *rels;
+
+ rels = OpenRelationList(sequences, PUB_OBJTYPE_SEQUENCE);
+ CheckObjSchemaNotAlreadyInPublication(rels, sequences_schemaidlist,
+ PUBLICATIONOBJ_SEQUENCE);
+
+ TransformPubWhereClauses(rels, pstate->p_sourcetext,
+ publish_via_partition_root);
+
+ CheckPubRelationColumnList(rels, pstate->p_sourcetext,
+ publish_via_partition_root);
+
+ PublicationAddRelations(puboid, rels, true, NULL);
+ CloseRelationList(rels);
+ }
+
+ /* tables added through a schema */
+ if (list_length(tables_schemaidlist) > 0)
+ {
+ /*
+ * Schema lock is held until the publication is created to prevent
+ * concurrent schema deletion.
+ */
+ LockSchemaList(tables_schemaidlist);
+ PublicationAddSchemas(puboid, tables_schemaidlist,
+ PUB_OBJTYPE_TABLE, true, NULL);
+ }
+
+ /* sequences added through a schema */
+ if (list_length(sequences_schemaidlist) > 0)
{
/*
* Schema lock is held until the publication is created to prevent
* concurrent schema deletion.
*/
- LockSchemaList(schemaidlist);
- PublicationAddSchemas(puboid, schemaidlist, true, NULL);
+ LockSchemaList(sequences_schemaidlist);
+ PublicationAddSchemas(puboid, sequences_schemaidlist,
+ PUB_OBJTYPE_SEQUENCE, true, NULL);
}
}
@@ -943,6 +1062,7 @@ AlterPublicationOptions(ParseState *pstate, AlterPublicationStmt *stmt,
AccessShareLock);
root_relids = GetPublicationRelations(pubform->oid,
+ PUB_OBJTYPE_TABLE,
PUBLICATION_PART_ROOT);
foreach(lc, root_relids)
@@ -1022,6 +1142,9 @@ AlterPublicationOptions(ParseState *pstate, AlterPublicationStmt *stmt,
values[Anum_pg_publication_pubtruncate - 1] = BoolGetDatum(pubactions.pubtruncate);
replaces[Anum_pg_publication_pubtruncate - 1] = true;
+
+ values[Anum_pg_publication_pubsequence - 1] = BoolGetDatum(pubactions.pubsequence);
+ replaces[Anum_pg_publication_pubsequence - 1] = true;
}
if (publish_via_partition_root_given)
@@ -1041,7 +1164,7 @@ AlterPublicationOptions(ParseState *pstate, AlterPublicationStmt *stmt,
pubform = (Form_pg_publication) GETSTRUCT(tup);
/* Invalidate the relcache. */
- if (pubform->puballtables)
+ if (pubform->puballtables || pubform->puballsequences)
{
CacheInvalidateRelcacheAll();
}
@@ -1057,6 +1180,7 @@ AlterPublicationOptions(ParseState *pstate, AlterPublicationStmt *stmt,
*/
if (root_relids == NIL)
relids = GetPublicationRelations(pubform->oid,
+ PUB_OBJTYPE_TABLE,
PUBLICATION_PART_ALL);
else
{
@@ -1070,7 +1194,20 @@ AlterPublicationOptions(ParseState *pstate, AlterPublicationStmt *stmt,
lfirst_oid(lc));
}
+ /* tables */
schemarelids = GetAllSchemaPublicationRelations(pubform->oid,
+ PUB_OBJTYPE_TABLE,
+ PUBLICATION_PART_ALL);
+ relids = list_concat_unique_oid(relids, schemarelids);
+
+ /* sequences */
+ relids = list_concat_unique_oid(relids,
+ GetPublicationRelations(pubform->oid,
+ PUB_OBJTYPE_SEQUENCE,
+ PUBLICATION_PART_ALL));
+
+ schemarelids = GetAllSchemaPublicationRelations(pubform->oid,
+ PUB_OBJTYPE_SEQUENCE,
PUBLICATION_PART_ALL);
relids = list_concat_unique_oid(relids, schemarelids);
@@ -1125,7 +1262,7 @@ AlterPublicationTables(AlterPublicationStmt *stmt, HeapTuple tup,
if (!tables && stmt->action != AP_SetObjects)
return;
- rels = OpenTableList(tables);
+ rels = OpenRelationList(tables, PUB_OBJTYPE_TABLE);
if (stmt->action == AP_AddObjects)
{
@@ -1135,7 +1272,9 @@ AlterPublicationTables(AlterPublicationStmt *stmt, HeapTuple tup,
* Check if the relation is member of the existing schema in the
* publication or member of the schema list specified.
*/
- schemas = list_concat_copy(schemaidlist, GetPublicationSchemas(pubid));
+ schemas = list_concat_copy(schemaidlist,
+ GetPublicationSchemas(pubid,
+ PUB_OBJTYPE_TABLE));
CheckObjSchemaNotAlreadyInPublication(rels, schemas,
PUBLICATIONOBJ_TABLE);
@@ -1143,13 +1282,14 @@ AlterPublicationTables(AlterPublicationStmt *stmt, HeapTuple tup,
CheckPubRelationColumnList(rels, queryString, pubform->pubviaroot);
- PublicationAddTables(pubid, rels, false, stmt);
+ PublicationAddRelations(pubid, rels, false, stmt);
}
else if (stmt->action == AP_DropObjects)
- PublicationDropTables(pubid, rels, false);
+ PublicationDropRelations(pubid, rels, false);
else /* AP_SetObjects */
{
List *oldrelids = GetPublicationRelations(pubid,
+ PUB_OBJTYPE_TABLE,
PUBLICATION_PART_ROOT);
List *delrels = NIL;
ListCell *oldlc;
@@ -1268,18 +1408,18 @@ AlterPublicationTables(AlterPublicationStmt *stmt, HeapTuple tup,
}
/* And drop them. */
- PublicationDropTables(pubid, delrels, true);
+ PublicationDropRelations(pubid, delrels, true);
/*
* Don't bother calculating the difference for adding, we'll catch and
* skip existing ones when doing catalog update.
*/
- PublicationAddTables(pubid, rels, true, stmt);
+ PublicationAddRelations(pubid, rels, true, stmt);
- CloseTableList(delrels);
+ CloseRelationList(delrels);
}
- CloseTableList(rels);
+ CloseRelationList(rels);
}
/*
@@ -1289,7 +1429,8 @@ AlterPublicationTables(AlterPublicationStmt *stmt, HeapTuple tup,
*/
static void
AlterPublicationSchemas(AlterPublicationStmt *stmt,
- HeapTuple tup, List *schemaidlist)
+ HeapTuple tup, List *schemaidlist,
+ char objectType)
{
Form_pg_publication pubform = (Form_pg_publication) GETSTRUCT(tup);
@@ -1311,20 +1452,20 @@ AlterPublicationSchemas(AlterPublicationStmt *stmt,
List *rels;
List *reloids;
- reloids = GetPublicationRelations(pubform->oid, PUBLICATION_PART_ROOT);
+ reloids = GetPublicationRelations(pubform->oid, objectType, PUBLICATION_PART_ROOT);
rels = OpenRelIdList(reloids);
CheckObjSchemaNotAlreadyInPublication(rels, schemaidlist,
PUBLICATIONOBJ_TABLES_IN_SCHEMA);
- CloseTableList(rels);
- PublicationAddSchemas(pubform->oid, schemaidlist, false, stmt);
+ CloseRelationList(rels);
+ PublicationAddSchemas(pubform->oid, schemaidlist, objectType, false, stmt);
}
else if (stmt->action == AP_DropObjects)
- PublicationDropSchemas(pubform->oid, schemaidlist, false);
+ PublicationDropSchemas(pubform->oid, schemaidlist, objectType, false);
else /* AP_SetObjects */
{
- List *oldschemaids = GetPublicationSchemas(pubform->oid);
+ List *oldschemaids = GetPublicationSchemas(pubform->oid, objectType);
List *delschemas = NIL;
/* Identify which schemas should be dropped */
@@ -1337,13 +1478,13 @@ AlterPublicationSchemas(AlterPublicationStmt *stmt,
LockSchemaList(delschemas);
/* And drop them */
- PublicationDropSchemas(pubform->oid, delschemas, true);
+ PublicationDropSchemas(pubform->oid, delschemas, objectType, true);
/*
* Don't bother calculating the difference for adding, we'll catch and
* skip existing ones when doing catalog update.
*/
- PublicationAddSchemas(pubform->oid, schemaidlist, true, stmt);
+ PublicationAddSchemas(pubform->oid, schemaidlist, objectType, true, stmt);
}
}
@@ -1353,12 +1494,13 @@ AlterPublicationSchemas(AlterPublicationStmt *stmt,
*/
static void
CheckAlterPublication(AlterPublicationStmt *stmt, HeapTuple tup,
- List *tables, List *schemaidlist)
+ List *tables, List *tables_schemaidlist,
+ List *sequences, List *sequences_schemaidlist)
{
Form_pg_publication pubform = (Form_pg_publication) GETSTRUCT(tup);
if ((stmt->action == AP_AddObjects || stmt->action == AP_SetObjects) &&
- schemaidlist && !superuser())
+ (tables_schemaidlist || sequences_schemaidlist) && !superuser())
ereport(ERROR,
(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
errmsg("must be superuser to add or set schemas")));
@@ -1367,13 +1509,24 @@ CheckAlterPublication(AlterPublicationStmt *stmt, HeapTuple tup,
* Check that user is allowed to manipulate the publication tables in
* schema
*/
- if (schemaidlist && pubform->puballtables)
+ if (tables_schemaidlist && pubform->puballtables)
ereport(ERROR,
(errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
errmsg("publication \"%s\" is defined as FOR ALL TABLES",
NameStr(pubform->pubname)),
errdetail("Tables from schema cannot be added to, dropped from, or set on FOR ALL TABLES publications.")));
+ /*
+ * Check that user is allowed to manipulate the publication sequences in
+ * schema
+ */
+ if (sequences_schemaidlist && pubform->puballsequences)
+ ereport(ERROR,
+ (errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
+ errmsg("publication \"%s\" is defined as FOR ALL SEQUENCES",
+ NameStr(pubform->pubname)),
+ errdetail("Sequences from schema cannot be added to, dropped from, or set on FOR ALL SEQUENCES publications.")));
+
/* Check that user is allowed to manipulate the publication tables. */
if (tables && pubform->puballtables)
ereport(ERROR,
@@ -1381,6 +1534,108 @@ CheckAlterPublication(AlterPublicationStmt *stmt, HeapTuple tup,
errmsg("publication \"%s\" is defined as FOR ALL TABLES",
NameStr(pubform->pubname)),
errdetail("Tables cannot be added to or dropped from FOR ALL TABLES publications.")));
+
+ /* Check that user is allowed to manipulate the publication sequences. */
+ if (sequences && pubform->puballsequences)
+ ereport(ERROR,
+ (errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
+ errmsg("publication \"%s\" is defined as FOR ALL SEQUENCES",
+ NameStr(pubform->pubname)),
+ errdetail("Sequences cannot be added to or dropped from FOR ALL SEQUENCES publications.")));
+}
+
+/*
+ * Add or remove sequence to/from publication.
+ */
+static void
+AlterPublicationSequences(AlterPublicationStmt *stmt, HeapTuple tup,
+ List *sequences, List *schemaidlist)
+{
+ List *rels = NIL;
+ Form_pg_publication pubform = (Form_pg_publication) GETSTRUCT(tup);
+ Oid pubid = pubform->oid;
+
+ /*
+ * It is quite possible that for the SET case user has not specified any
+ * tables in which case we need to remove all the existing tables.
+ */
+ if (!sequences && stmt->action != AP_SetObjects)
+ return;
+
+ rels = OpenRelationList(sequences, PUB_OBJTYPE_SEQUENCE);
+
+ if (stmt->action == AP_AddObjects)
+ {
+ List *schemas = NIL;
+
+ /*
+ * Check if the relation is member of the existing schema in the
+ * publication or member of the schema list specified.
+ */
+ schemas = list_concat_copy(schemaidlist,
+ GetPublicationSchemas(pubid,
+ PUB_OBJTYPE_SEQUENCE));
+ CheckObjSchemaNotAlreadyInPublication(rels, schemas,
+ PUBLICATIONOBJ_SEQUENCE);
+ PublicationAddRelations(pubid, rels, false, stmt);
+ }
+ else if (stmt->action == AP_DropObjects)
+ PublicationDropRelations(pubid, rels, false);
+ else /* DEFELEM_SET */
+ {
+ List *oldrelids = GetPublicationRelations(pubid,
+ PUB_OBJTYPE_SEQUENCE,
+ PUBLICATION_PART_ROOT);
+ List *delrels = NIL;
+ ListCell *oldlc;
+
+ CheckObjSchemaNotAlreadyInPublication(rels, schemaidlist,
+ PUBLICATIONOBJ_SEQUENCE);
+
+ /* Calculate which relations to drop. */
+ foreach(oldlc, oldrelids)
+ {
+ Oid oldrelid = lfirst_oid(oldlc);
+ ListCell *newlc;
+ PublicationRelInfo *oldrel;
+ bool found = false;
+
+ foreach(newlc, rels)
+ {
+ PublicationRelInfo *newpubrel;
+
+ newpubrel = (PublicationRelInfo *) lfirst(newlc);
+ if (RelationGetRelid(newpubrel->relation) == oldrelid)
+ {
+ found = true;
+ break;
+ }
+ }
+ /* Not yet in the list, open it and add to the list */
+ if (!found)
+ {
+ oldrel = palloc(sizeof(PublicationRelInfo));
+ oldrel->whereClause = NULL;
+ oldrel->columns = NULL;
+ oldrel->relation = table_open(oldrelid,
+ ShareUpdateExclusiveLock);
+ delrels = lappend(delrels, oldrel);
+ }
+ }
+
+ /* And drop them. */
+ PublicationDropRelations(pubid, delrels, true);
+
+ /*
+ * Don't bother calculating the difference for adding, we'll catch and
+ * skip existing ones when doing catalog update.
+ */
+ PublicationAddRelations(pubid, rels, true, stmt);
+
+ CloseRelationList(delrels);
+ }
+
+ CloseRelationList(rels);
}
/*
@@ -1418,14 +1673,20 @@ AlterPublication(ParseState *pstate, AlterPublicationStmt *stmt)
AlterPublicationOptions(pstate, stmt, rel, tup);
else
{
- List *relations = NIL;
- List *schemaidlist = NIL;
+ List *tables = NIL;
+ List *sequences = NIL;
+ List *tables_schemaidlist = NIL;
+ List *sequences_schemaidlist = NIL;
Oid pubid = pubform->oid;
- ObjectsInPublicationToOids(stmt->pubobjects, pstate, &relations,
- &schemaidlist);
+ ObjectsInPublicationToOids(stmt->pubobjects, pstate,
+ &tables, &sequences,
+ &tables_schemaidlist,
+ &sequences_schemaidlist);
- CheckAlterPublication(stmt, tup, relations, schemaidlist);
+ CheckAlterPublication(stmt, tup,
+ tables, tables_schemaidlist,
+ sequences, sequences_schemaidlist);
heap_freetuple(tup);
@@ -1453,9 +1714,16 @@ AlterPublication(ParseState *pstate, AlterPublicationStmt *stmt)
errmsg("publication \"%s\" does not exist",
stmt->pubname));
- AlterPublicationTables(stmt, tup, relations, schemaidlist,
+ AlterPublicationTables(stmt, tup, tables, tables_schemaidlist,
pstate->p_sourcetext);
- AlterPublicationSchemas(stmt, tup, schemaidlist);
+
+ AlterPublicationSequences(stmt, tup, sequences, sequences_schemaidlist);
+
+ AlterPublicationSchemas(stmt, tup, tables_schemaidlist,
+ PUB_OBJTYPE_TABLE);
+
+ AlterPublicationSchemas(stmt, tup, sequences_schemaidlist,
+ PUB_OBJTYPE_SEQUENCE);
}
/* Cleanup. */
@@ -1523,7 +1791,7 @@ RemovePublicationById(Oid pubid)
pubform = (Form_pg_publication) GETSTRUCT(tup);
/* Invalidate relcache so that publication info is rebuilt. */
- if (pubform->puballtables)
+ if (pubform->puballtables || pubform->puballsequences)
CacheInvalidateRelcacheAll();
CatalogTupleDelete(rel, &tup->t_self);
@@ -1559,6 +1827,7 @@ RemovePublicationSchemaById(Oid psoid)
* partitions.
*/
schemaRels = GetSchemaPublicationRelations(pubsch->pnnspid,
+ pubsch->pntype,
PUBLICATION_PART_ALL);
InvalidatePublicationRels(schemaRels);
@@ -1601,10 +1870,10 @@ OpenRelIdList(List *relids)
* add them to a publication.
*/
static List *
-OpenTableList(List *tables)
+OpenRelationList(List *rels, char objectType)
{
List *relids = NIL;
- List *rels = NIL;
+ List *result = NIL;
ListCell *lc;
List *relids_with_rf = NIL;
List *relids_with_collist = NIL;
@@ -1612,19 +1881,35 @@ OpenTableList(List *tables)
/*
* Open, share-lock, and check all the explicitly-specified relations
*/
- foreach(lc, tables)
+ foreach(lc, rels)
{
PublicationTable *t = lfirst_node(PublicationTable, lc);
bool recurse = t->relation->inh;
Relation rel;
Oid myrelid;
PublicationRelInfo *pub_rel;
+ char myrelkind;
/* Allow query cancel in case this takes a long time */
CHECK_FOR_INTERRUPTS();
rel = table_openrv(t->relation, ShareUpdateExclusiveLock);
myrelid = RelationGetRelid(rel);
+ myrelkind = get_rel_relkind(myrelid);
+
+ /*
+ * Make sure the relkind matches the expected object type. This may
+ * happen e.g. when adding a sequence using ADD TABLE or a table
+ * using ADD SEQUENCE).
+ *
+ * XXX We let through unsupported object types (views etc.). Those
+ * will be caught later in check_publication_add_relation.
+ */
+ if (pub_get_object_type_for_relkind(myrelkind) != PUB_OBJTYPE_UNSUPPORTED &&
+ pub_get_object_type_for_relkind(myrelkind) != objectType)
+ ereport(ERROR,
+ errcode(ERRCODE_INVALID_PARAMETER_VALUE),
+ errmsg("object type does not match type expected by command"));
/*
* Filter out duplicates if user specifies "foo, foo".
@@ -1657,7 +1942,7 @@ OpenTableList(List *tables)
pub_rel->relation = rel;
pub_rel->whereClause = t->whereClause;
pub_rel->columns = t->columns;
- rels = lappend(rels, pub_rel);
+ result = lappend(result, pub_rel);
relids = lappend_oid(relids, myrelid);
if (t->whereClause)
@@ -1726,10 +2011,9 @@ OpenTableList(List *tables)
pub_rel->relation = rel;
/* child inherits WHERE clause from parent */
pub_rel->whereClause = t->whereClause;
-
/* child inherits column list from parent */
pub_rel->columns = t->columns;
- rels = lappend(rels, pub_rel);
+ result = lappend(result, pub_rel);
relids = lappend_oid(relids, childrelid);
if (t->whereClause)
@@ -1744,14 +2028,14 @@ OpenTableList(List *tables)
list_free(relids);
list_free(relids_with_rf);
- return rels;
+ return result;
}
/*
* Close all relations in the list.
*/
static void
-CloseTableList(List *rels)
+CloseRelationList(List *rels)
{
ListCell *lc;
@@ -1799,12 +2083,12 @@ LockSchemaList(List *schemalist)
* Add listed tables to the publication.
*/
static void
-PublicationAddTables(Oid pubid, List *rels, bool if_not_exists,
+PublicationAddRelations(Oid pubid, List *rels, bool if_not_exists,
AlterPublicationStmt *stmt)
{
ListCell *lc;
- Assert(!stmt || !stmt->for_all_tables);
+ Assert(!stmt || !stmt->for_all_objects);
foreach(lc, rels)
{
@@ -1833,7 +2117,7 @@ PublicationAddTables(Oid pubid, List *rels, bool if_not_exists,
* Remove listed tables from the publication.
*/
static void
-PublicationDropTables(Oid pubid, List *rels, bool missing_ok)
+PublicationDropRelations(Oid pubid, List *rels, bool missing_ok)
{
ObjectAddress obj;
ListCell *lc;
@@ -1878,19 +2162,19 @@ PublicationDropTables(Oid pubid, List *rels, bool missing_ok)
* Add listed schemas to the publication.
*/
static void
-PublicationAddSchemas(Oid pubid, List *schemas, bool if_not_exists,
- AlterPublicationStmt *stmt)
+PublicationAddSchemas(Oid pubid, List *schemas, char objectType,
+ bool if_not_exists, AlterPublicationStmt *stmt)
{
ListCell *lc;
- Assert(!stmt || !stmt->for_all_tables);
+ Assert(!stmt || !stmt->for_all_objects);
foreach(lc, schemas)
{
Oid schemaid = lfirst_oid(lc);
ObjectAddress obj;
- obj = publication_add_schema(pubid, schemaid, if_not_exists);
+ obj = publication_add_schema(pubid, schemaid, objectType, if_not_exists);
if (stmt)
{
EventTriggerCollectSimpleCommand(obj, InvalidObjectAddress,
@@ -1906,7 +2190,7 @@ PublicationAddSchemas(Oid pubid, List *schemas, bool if_not_exists,
* Remove listed schemas from the publication.
*/
static void
-PublicationDropSchemas(Oid pubid, List *schemas, bool missing_ok)
+PublicationDropSchemas(Oid pubid, List *schemas, char objectType, bool missing_ok)
{
ObjectAddress obj;
ListCell *lc;
@@ -1916,10 +2200,11 @@ PublicationDropSchemas(Oid pubid, List *schemas, bool missing_ok)
{
Oid schemaid = lfirst_oid(lc);
- psid = GetSysCacheOid2(PUBLICATIONNAMESPACEMAP,
+ psid = GetSysCacheOid3(PUBLICATIONNAMESPACEMAP,
Anum_pg_publication_namespace_oid,
ObjectIdGetDatum(schemaid),
- ObjectIdGetDatum(pubid));
+ ObjectIdGetDatum(pubid),
+ CharGetDatum(objectType));
if (!OidIsValid(psid))
{
if (missing_ok)
@@ -1974,6 +2259,13 @@ AlterPublicationOwner_internal(Relation rel, HeapTuple tup, Oid newOwnerId)
NameStr(form->pubname)),
errhint("The owner of a FOR ALL TABLES publication must be a superuser.")));
+ if (form->puballsequences && !superuser_arg(newOwnerId))
+ ereport(ERROR,
+ (errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
+ errmsg("permission denied to change owner of publication \"%s\"",
+ NameStr(form->pubname)),
+ errhint("The owner of a FOR ALL SEQUENCES publication must be a superuser.")));
+
if (!superuser_arg(newOwnerId) && is_schema_publication(form->oid))
ereport(ERROR,
(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
diff --git a/src/backend/commands/sequence.c b/src/backend/commands/sequence.c
index 99c9f91cba..93876b1fd8 100644
--- a/src/backend/commands/sequence.c
+++ b/src/backend/commands/sequence.c
@@ -38,6 +38,7 @@
#include "miscadmin.h"
#include "nodes/makefuncs.h"
#include "parser/parse_type.h"
+#include "replication/message.h"
#include "storage/lmgr.h"
#include "storage/proc.h"
#include "storage/smgr.h"
@@ -76,6 +77,7 @@ typedef struct SeqTableData
{
Oid relid; /* pg_class OID of this sequence (hash key) */
RelFileNumber filenumber; /* last seen relfilenumber of this sequence */
+ Oid tablespace; /* last seen tablespace of this sequence */
LocalTransactionId lxid; /* xact in which we last did a seq op */
bool last_valid; /* do we have a valid "last" value? */
int64 last; /* value last returned by nextval */
@@ -83,6 +85,7 @@ typedef struct SeqTableData
/* if last != cached, we have not used up all the cached values */
int64 increment; /* copy of sequence's increment field */
/* note that increment is zero until we first do nextval_internal() */
+ bool need_log; /* should be written to WAL at commit? */
} SeqTableData;
typedef SeqTableData *SeqTable;
@@ -340,6 +343,69 @@ ResetSequence(Oid seq_relid)
relation_close(seq_rel, NoLock);
}
+/*
+ * Update the sequence state by creating a new relfilenode.
+ *
+ * This creates a new relfilenode, to allow transactional behavior.
+ */
+void
+SetSequence(Oid seq_relid, int64 last_value, int64 log_cnt, bool is_called)
+{
+ SeqTable elm;
+ Relation seqrel;
+ Buffer buf;
+ HeapTupleData seqdatatuple;
+ Form_pg_sequence_data seq;
+ HeapTuple tuple;
+
+ /* open and lock sequence */
+ init_sequence(seq_relid, &elm, &seqrel);
+
+ /* lock page' buffer and read tuple */
+ seq = read_seq_tuple(seqrel, &buf, &seqdatatuple);
+
+ /* Copy the existing sequence tuple. */
+ tuple = heap_copytuple(&seqdatatuple);
+
+ /* Now we're done with the old page */
+ UnlockReleaseBuffer(buf);
+
+ /*
+ * Modify the copied tuple to update the sequence state (similar to what
+ * ResetSequence does).
+ */
+ seq = (Form_pg_sequence_data) GETSTRUCT(tuple);
+ seq->last_value = last_value;
+ seq->is_called = is_called;
+ seq->log_cnt = log_cnt;
+
+ /*
+ * Create a new storage file for the sequence - this is needed for the
+ * transactional behavior.
+ */
+ RelationSetNewRelfilenumber(seqrel, seqrel->rd_rel->relpersistence);
+
+ /*
+ * Ensure sequence's relfrozenxid is at 0, since it won't contain any
+ * unfrozen XIDs. Same with relminmxid, since a sequence will never
+ * contain multixacts.
+ */
+ Assert(seqrel->rd_rel->relfrozenxid == InvalidTransactionId);
+ Assert(seqrel->rd_rel->relminmxid == InvalidMultiXactId);
+
+ /*
+ * Insert the modified tuple into the new storage file. This does all the
+ * necessary WAL-logging etc.
+ */
+ fill_seq_with_data(seqrel, tuple);
+
+ /* Clear local cache so that we don't think we have cached numbers */
+ /* Note that we do not change the currval() state */
+ elm->cached = elm->last;
+
+ relation_close(seqrel, NoLock);
+}
+
/*
* Initialize a sequence's relation with the specified tuple as content
*
@@ -404,10 +470,21 @@ fill_seq_fork_with_data(Relation rel, HeapTuple tuple, ForkNumber forkNum)
tuple->t_data->t_infomask |= HEAP_XMAX_INVALID;
ItemPointerSet(&tuple->t_data->t_ctid, 0, FirstOffsetNumber);
- /* check the comment above nextval_internal()'s equivalent call. */
+ /*
+ * Remember we need to write this sequence to WAL at commit, and make sure
+ * we have XID if we may need to decode this sequence.
+ *
+ * XXX Maybe we should not be doing this except when actually reading a
+ * value from the sequence?
+ */
if (RelationNeedsWAL(rel))
+ {
GetTopTransactionId();
+ if (XLogLogicalInfoActive())
+ GetCurrentTransactionId();
+ }
+
START_CRIT_SECTION();
MarkBufferDirty(buf);
@@ -652,6 +729,23 @@ nextval_internal(Oid relid, bool check_permissions)
/* open and lock sequence */
init_sequence(relid, &elm, &seqrel);
+ /*
+ * Remember we need to write this sequence to WAL at commit, and make sure
+ * we have XID if we may need to decode this sequence.
+ *
+ * XXX Maybe we should not be doing this except when actually reading a
+ * value from the sequence?
+ */
+ if (RelationNeedsWAL(seqrel))
+ {
+ elm->need_log = true;
+
+ GetTopTransactionId();
+
+ if (XLogLogicalInfoActive())
+ GetCurrentTransactionId();
+ }
+
if (check_permissions &&
pg_class_aclcheck(elm->relid, GetUserId(),
ACL_USAGE | ACL_UPDATE) != ACLCHECK_OK)
@@ -806,10 +900,28 @@ nextval_internal(Oid relid, bool check_permissions)
* It's sufficient to ensure the toplevel transaction has an xid, no need
* to assign xids subxacts, that'll already trigger an appropriate wait.
* (Have to do that here, so we're outside the critical section)
+ *
+ * We have to ensure we have a proper XID, which will be included in
+ * the XLOG record by XLogRecordAssemble. Otherwise the first nextval()
+ * in a subxact (without any preceding changes) would get XID 0, and it
+ * would then be impossible to decide which top xact it belongs to.
+ * It'd also trigger assert in DecodeSequence. We only do that with
+ * wal_level=logical, though.
+ *
+ * XXX This might seem unnecessary, because if there's no XID the xact
+ * couldn't have done anything important yet, e.g. it could not have
+ * created a sequence. But that's incorrect, because of subxacts. The
+ * current subtransaction might not have done anything yet (thus no XID),
+ * but an earlier one might have created the sequence.
*/
if (logit && RelationNeedsWAL(seqrel))
+ {
GetTopTransactionId();
+ if (XLogLogicalInfoActive())
+ GetCurrentTransactionId();
+ }
+
/* ready to change the on-disk (or really, in-buffer) tuple */
START_CRIT_SECTION();
@@ -963,6 +1075,23 @@ do_setval(Oid relid, int64 next, bool iscalled)
/* open and lock sequence */
init_sequence(relid, &elm, &seqrel);
+ /*
+ * Remember we need to write this sequence to WAL at commit, and make sure
+ * we have XID if we may need to decode this sequence.
+ *
+ * XXX Maybe we should not be doing this except when actually reading a
+ * value from the sequence?
+ */
+ if (RelationNeedsWAL(seqrel))
+ {
+ elm->need_log = true;
+
+ GetTopTransactionId();
+
+ if (XLogLogicalInfoActive())
+ GetCurrentTransactionId();
+ }
+
if (pg_class_aclcheck(elm->relid, GetUserId(), ACL_UPDATE) != ACLCHECK_OK)
ereport(ERROR,
(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
@@ -1010,8 +1139,13 @@ do_setval(Oid relid, int64 next, bool iscalled)
/* check the comment above nextval_internal()'s equivalent call. */
if (RelationNeedsWAL(seqrel))
+ {
GetTopTransactionId();
+ if (XLogLogicalInfoActive())
+ GetCurrentTransactionId();
+ }
+
/* ready to change the on-disk (or really, in-buffer) tuple */
START_CRIT_SECTION();
@@ -1180,6 +1314,7 @@ init_sequence(Oid relid, SeqTable *p_elm, Relation *p_rel)
if (seqrel->rd_rel->relfilenode != elm->filenumber)
{
elm->filenumber = seqrel->rd_rel->relfilenode;
+ elm->tablespace = seqrel->rd_rel->reltablespace;
elm->cached = elm->last;
}
@@ -1916,3 +2051,121 @@ seq_mask(char *page, BlockNumber blkno)
mask_unused_space(page);
}
+
+
+static void
+read_sequence_info(Relation seqrel, int64 *last_value, int64 *log_cnt, bool *is_called)
+{
+ Buffer buf;
+ Form_pg_sequence_data seq;
+ HeapTupleData seqdatatuple;
+
+ seq = read_seq_tuple(seqrel, &buf, &seqdatatuple);
+
+ *last_value = seq->last_value;
+ *is_called = seq->is_called;
+ *log_cnt = seq->log_cnt;
+
+ UnlockReleaseBuffer(buf);
+}
+
+/* XXX Do this only for wal_level = logical, probably? */
+void
+AtEOXact_Sequences(bool isCommit)
+{
+ SeqTable entry;
+ HASH_SEQ_STATUS scan;
+
+ if (!seqhashtab)
+ return;
+
+ /* only do this with wal_level=logical */
+ if (!XLogLogicalInfoActive())
+ return;
+
+ /*
+ * XXX Maybe we could enforce having XID here? Or is it too late?
+ */
+ // Assert(GetTopTransactionId() != InvalidTransactionId);
+ // Assert(GetCurrentTransactionId() != InvalidTransactionId);
+
+ hash_seq_init(&scan, seqhashtab);
+
+ while ((entry = (SeqTable) hash_seq_search(&scan)))
+ {
+ Relation rel;
+ RelFileLocator rlocator;
+ xl_logical_sequence xlrec;
+
+ /* not commit, we don't guarantee any data to be durable */
+ if (!isCommit)
+ entry->need_log = false;
+
+ /*
+ * If not touched in the current transaction, don't log anything.
+ * We leave needs_log set, so that if future transactions touch
+ * the sequence we'll log it properly.
+ */
+ if (!entry->need_log)
+ continue;
+
+ /* if this is commit, we'll log the */
+ entry->need_log = false;
+
+ /*
+ * does the relation still exist?
+ *
+ * XXX We need to make sure another transaction can't jump ahead of
+ * this one, otherwise the ordering of sequence changes could change.
+ * Imagine T1 and T2, where T1 writes sequence state first, but then
+ * T2 does it too and commits first:
+ *
+ * T1: log sequence state
+ * T2: increment sequence
+ * T2: log sequence state
+ * T2: commit
+ * T1: commit
+ *
+ * If we apply the sequences in this order, it'd be broken as the
+ * value might go backwars. ShareUpdateExclusive protects against
+ * that, but it's also restricting commit throughtput, probably.
+ *
+ * XXX This might be an issue for deadlocks, too, if two xacts try
+ * to write sequences in different ordering. We may need to sort
+ * the OIDs first, to enforce the same lock ordering.
+ */
+ rel = try_relation_open(entry->relid, ShareUpdateExclusiveLock);
+
+ /* XXX relation might have been dropped, maybe we should have logged
+ * the last change? */
+ if (!rel)
+ continue;
+
+ /* tablespace */
+ if (OidIsValid(entry->tablespace))
+ rlocator.spcOid = entry->tablespace;
+ else
+ rlocator.spcOid = MyDatabaseTableSpace;
+
+ rlocator.dbOid = MyDatabaseId; /* database */
+ rlocator.relNumber = entry->filenumber; /* relation */
+
+ xlrec.locator = rlocator;
+ xlrec.reloid = entry->relid;
+
+ /* XXX is it good enough to log values we have in cache? seems
+ * wrong and we may need to re-read that. */
+ read_sequence_info(rel, &xlrec.last, &xlrec.log_cnt, &xlrec.is_called);
+
+ XLogBeginInsert();
+ XLogRegisterData((char *) &xlrec, SizeOfLogicalSequence);
+
+ /* allow origin filtering */
+ XLogSetRecordFlags(XLOG_INCLUDE_ORIGIN);
+
+ (void) XLogInsert(RM_LOGICALMSG_ID, XLOG_LOGICAL_SEQUENCE);
+
+ /* hold the sequence lock until the end of the transaction */
+ relation_close(rel, NoLock);
+ }
+}
diff --git a/src/backend/commands/subscriptioncmds.c b/src/backend/commands/subscriptioncmds.c
index 670b219c8d..53810d8cf9 100644
--- a/src/backend/commands/subscriptioncmds.c
+++ b/src/backend/commands/subscriptioncmds.c
@@ -92,6 +92,7 @@ typedef struct SubOpts
} SubOpts;
static List *fetch_table_list(WalReceiverConn *wrconn, List *publications);
+static List *fetch_sequence_list(WalReceiverConn *wrconn, List *publications);
static void check_duplicates_in_publist(List *publist, Datum *datums);
static List *merge_publications(List *oldpublist, List *newpublist, bool addpub, const char *subname);
static void ReportSlotConnectionError(List *rstates, Oid subid, char *slotname, char *err);
@@ -666,9 +667,9 @@ CreateSubscription(ParseState *pstate, CreateSubscriptionStmt *stmt,
{
char *err;
WalReceiverConn *wrconn;
- List *tables;
+ List *relations;
ListCell *lc;
- char table_state;
+ char sync_state;
/* Try to connect to the publisher. */
wrconn = walrcv_connect(conninfo, true, stmt->subname, &err);
@@ -685,14 +686,17 @@ CreateSubscription(ParseState *pstate, CreateSubscriptionStmt *stmt,
* Set sync state based on if we were asked to do data copy or
* not.
*/
- table_state = opts.copy_data ? SUBREL_STATE_INIT : SUBREL_STATE_READY;
+ sync_state = opts.copy_data ? SUBREL_STATE_INIT : SUBREL_STATE_READY;
/*
- * Get the table list from publisher and build local table status
- * info.
+ * Get the table and sequence list from publisher and build
+ * local relation sync status info.
*/
- tables = fetch_table_list(wrconn, publications);
- foreach(lc, tables)
+ relations = fetch_table_list(wrconn, publications);
+ relations = list_concat(relations,
+ fetch_sequence_list(wrconn, publications));
+
+ foreach(lc, relations)
{
RangeVar *rv = (RangeVar *) lfirst(lc);
Oid relid;
@@ -703,7 +707,7 @@ CreateSubscription(ParseState *pstate, CreateSubscriptionStmt *stmt,
CheckSubscriptionRelkind(get_rel_relkind(relid),
rv->schemaname, rv->relname);
- AddSubscriptionRelState(subid, relid, table_state,
+ AddSubscriptionRelState(subid, relid, sync_state,
InvalidXLogRecPtr);
}
@@ -729,12 +733,12 @@ CreateSubscription(ParseState *pstate, CreateSubscriptionStmt *stmt,
*
* Note that if tables were specified but copy_data is false
* then it is safe to enable two_phase up-front because those
- * tables are already initially in READY state. When the
- * subscription has no tables, we leave the twophase state as
- * PENDING, to allow ALTER SUBSCRIPTION ... REFRESH
+ * relations are already initially in READY state. When the
+ * subscription has no relations, we leave the twophase state
+ * as PENDING, to allow ALTER SUBSCRIPTION ... REFRESH
* PUBLICATION to work.
*/
- if (opts.twophase && !opts.copy_data && tables != NIL)
+ if (opts.twophase && !opts.copy_data && relations != NIL)
twophase_enabled = true;
walrcv_create_slot(wrconn, opts.slot_name, false, twophase_enabled,
@@ -810,8 +814,10 @@ AlterSubscription_refresh(Subscription *sub, bool copy_data,
if (validate_publications)
check_publications(wrconn, validate_publications);
- /* Get the table list from publisher. */
+ /* Get the list of relations from publisher. */
pubrel_names = fetch_table_list(wrconn, sub->publications);
+ pubrel_names = list_concat(pubrel_names,
+ fetch_sequence_list(wrconn, sub->publications));
/* Get local table list. */
subrel_states = GetSubscriptionRelations(sub->oid, false);
@@ -1854,6 +1860,75 @@ fetch_table_list(WalReceiverConn *wrconn, List *publications)
return tablelist;
}
+/*
+ * Get the list of sequences which belong to specified publications on the
+ * publisher connection.
+ */
+static List *
+fetch_sequence_list(WalReceiverConn *wrconn, List *publications)
+{
+ WalRcvExecResult *res;
+ StringInfoData cmd;
+ TupleTableSlot *slot;
+ Oid tableRow[2] = {TEXTOID, TEXTOID};
+ ListCell *lc;
+ bool first;
+ List *tablelist = NIL;
+
+ Assert(list_length(publications) > 0);
+
+ initStringInfo(&cmd);
+ appendStringInfoString(&cmd, "SELECT DISTINCT s.schemaname, s.sequencename\n"
+ " FROM pg_catalog.pg_publication_sequences s\n"
+ " WHERE s.pubname IN (");
+ first = true;
+ foreach(lc, publications)
+ {
+ char *pubname = strVal(lfirst(lc));
+
+ if (first)
+ first = false;
+ else
+ appendStringInfoString(&cmd, ", ");
+
+ appendStringInfoString(&cmd, quote_literal_cstr(pubname));
+ }
+ appendStringInfoChar(&cmd, ')');
+
+ res = walrcv_exec(wrconn, cmd.data, 2, tableRow);
+ pfree(cmd.data);
+
+ if (res->status != WALRCV_OK_TUPLES)
+ ereport(ERROR,
+ (errmsg("could not receive list of replicated sequences from the publisher: %s",
+ res->err)));
+
+ /* Process sequences. */
+ slot = MakeSingleTupleTableSlot(res->tupledesc, &TTSOpsMinimalTuple);
+ while (tuplestore_gettupleslot(res->tuplestore, true, false, slot))
+ {
+ char *nspname;
+ char *relname;
+ bool isnull;
+ RangeVar *rv;
+
+ nspname = TextDatumGetCString(slot_getattr(slot, 1, &isnull));
+ Assert(!isnull);
+ relname = TextDatumGetCString(slot_getattr(slot, 2, &isnull));
+ Assert(!isnull);
+
+ rv = makeRangeVar(nspname, relname, -1);
+ tablelist = lappend(tablelist, rv);
+
+ ExecClearTuple(slot);
+ }
+ ExecDropSingleTupleTableSlot(slot);
+
+ walrcv_clear_result(res);
+
+ return tablelist;
+}
+
/*
* This is to report the connection failure while dropping replication slots.
* Here, we report the WARNING for all tablesync slots so that user can drop
diff --git a/src/backend/commands/tablecmds.c b/src/backend/commands/tablecmds.c
index 9be04c8a1e..4454b21e89 100644
--- a/src/backend/commands/tablecmds.c
+++ b/src/backend/commands/tablecmds.c
@@ -43,6 +43,7 @@
#include "catalog/pg_largeobject.h"
#include "catalog/pg_namespace.h"
#include "catalog/pg_opclass.h"
+#include "catalog/pg_publication_namespace.h"
#include "catalog/pg_statistic_ext.h"
#include "catalog/pg_tablespace.h"
#include "catalog/pg_trigger.h"
@@ -16401,11 +16402,14 @@ AlterTableNamespace(AlterObjectSchemaStmt *stmt, Oid *oldschema)
* Check that setting the relation to a different schema won't result in a
* publication having both a schema and the same schema's table, as this
* is not supported.
+ *
+ * XXX We do this for tables and sequences, but it's better to keep the two
+ * blocks separate, to make the strings easier to translate.
*/
if (stmt->objectType == OBJECT_TABLE)
{
ListCell *lc;
- List *schemaPubids = GetSchemaPublications(nspOid);
+ List *schemaPubids = GetSchemaPublications(nspOid, PUB_OBJTYPE_TABLE);
List *relPubids = GetRelationPublications(RelationGetRelid(rel));
foreach(lc, relPubids)
@@ -16423,6 +16427,27 @@ AlterTableNamespace(AlterObjectSchemaStmt *stmt, Oid *oldschema)
get_publication_name(pubid, false)));
}
}
+ else if (stmt->objectType == OBJECT_SEQUENCE)
+ {
+ ListCell *lc;
+ List *schemaPubids = GetSchemaPublications(nspOid, PUB_OBJTYPE_SEQUENCE);
+ List *relPubids = GetRelationPublications(RelationGetRelid(rel));
+
+ foreach(lc, relPubids)
+ {
+ Oid pubid = lfirst_oid(lc);
+
+ if (list_member_oid(schemaPubids, pubid))
+ ereport(ERROR,
+ errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
+ errmsg("cannot move sequence \"%s\" to schema \"%s\"",
+ RelationGetRelationName(rel), stmt->newschema),
+ errdetail("The schema \"%s\" and same schema's sequence \"%s\" cannot be part of the same publication \"%s\".",
+ stmt->newschema,
+ RelationGetRelationName(rel),
+ get_publication_name(pubid, false)));
+ }
+ }
/* common checks on switching namespaces */
CheckSetNamespace(oldNspOid, nspOid);
diff --git a/src/backend/executor/execReplication.c b/src/backend/executor/execReplication.c
index 6014f2e248..3ae893f221 100644
--- a/src/backend/executor/execReplication.c
+++ b/src/backend/executor/execReplication.c
@@ -655,7 +655,9 @@ void
CheckSubscriptionRelkind(char relkind, const char *nspname,
const char *relname)
{
- if (relkind != RELKIND_RELATION && relkind != RELKIND_PARTITIONED_TABLE)
+ if (relkind != RELKIND_RELATION &&
+ relkind != RELKIND_PARTITIONED_TABLE &&
+ relkind != RELKIND_SEQUENCE)
ereport(ERROR,
(errcode(ERRCODE_WRONG_OBJECT_TYPE),
errmsg("cannot use relation \"%s.%s\" as logical replication target",
diff --git a/src/backend/parser/gram.y b/src/backend/parser/gram.y
index f9037761f9..e2bbbe9bf2 100644
--- a/src/backend/parser/gram.y
+++ b/src/backend/parser/gram.y
@@ -459,7 +459,7 @@ static Node *makeRecursiveViewSelect(char *relname, List *aliases, Node *query);
transform_element_list transform_type_list
TriggerTransitions TriggerReferencing
vacuum_relation_list opt_vacuum_relation_list
- drop_option_list pub_obj_list
+ drop_option_list pub_obj_list pub_obj_type_list
%type <node> opt_routine_body
%type <groupclause> group_clause
@@ -593,6 +593,7 @@ static Node *makeRecursiveViewSelect(char *relname, List *aliases, Node *query);
%type <node> var_value zone_value
%type <rolespec> auth_ident RoleSpec opt_granted_by
%type <publicationobjectspec> PublicationObjSpec
+%type <node> pub_obj_type
%type <keyword> unreserved_keyword type_func_name_keyword
%type <keyword> col_name_keyword reserved_keyword
@@ -10420,12 +10421,9 @@ AlterOwnerStmt: ALTER AGGREGATE aggregate_with_argtypes OWNER TO RoleSpec
*
* CREATE PUBLICATION FOR ALL TABLES [WITH options]
*
- * CREATE PUBLICATION FOR pub_obj [, ...] [WITH options]
- *
- * pub_obj is one of:
+ * CREATE PUBLICATION FOR ALL SEQUENCES [WITH options]
*
- * TABLE table [, ...]
- * ALL TABLES IN SCHEMA schema [, ...]
+ * CREATE PUBLICATION FOR pub_obj [, ...] [WITH options]
*
*****************************************************************************/
@@ -10438,13 +10436,13 @@ CreatePublicationStmt:
n->options = $4;
$$ = (Node *) n;
}
- | CREATE PUBLICATION name FOR ALL TABLES opt_definition
+ | CREATE PUBLICATION name FOR ALL pub_obj_type_list opt_definition
{
CreatePublicationStmt *n = makeNode(CreatePublicationStmt);
n->pubname = $3;
n->options = $7;
- n->for_all_tables = true;
+ n->for_all_objects = $6;
$$ = (Node *) n;
}
| CREATE PUBLICATION name FOR pub_obj_list opt_definition
@@ -10495,6 +10493,26 @@ PublicationObjSpec:
$$->pubobjtype = PUBLICATIONOBJ_TABLES_IN_CUR_SCHEMA;
$$->location = @5;
}
+ | SEQUENCE relation_expr
+ {
+ $$ = makeNode(PublicationObjSpec);
+ $$->pubobjtype = PUBLICATIONOBJ_SEQUENCE;
+ $$->pubtable = makeNode(PublicationTable);
+ $$->pubtable->relation = $2;
+ }
+ | ALL SEQUENCES IN_P SCHEMA ColId
+ {
+ $$ = makeNode(PublicationObjSpec);
+ $$->pubobjtype = PUBLICATIONOBJ_SEQUENCES_IN_SCHEMA;
+ $$->name = $5;
+ $$->location = @5;
+ }
+ | ALL SEQUENCES IN_P SCHEMA CURRENT_SCHEMA
+ {
+ $$ = makeNode(PublicationObjSpec);
+ $$->pubobjtype = PUBLICATIONOBJ_SEQUENCES_IN_CUR_SCHEMA;
+ $$->location = @5;
+ }
| ColId opt_column_list OptWhereClause
{
$$ = makeNode(PublicationObjSpec);
@@ -10556,6 +10574,19 @@ pub_obj_list: PublicationObjSpec
{ $$ = lappend($1, $3); }
;
+pub_obj_type: TABLES
+ { $$ = (Node *) makeString("tables"); }
+ | SEQUENCES
+ { $$ = (Node *) makeString("sequences"); }
+ ;
+
+pub_obj_type_list: pub_obj_type
+ { $$ = list_make1($1); }
+ | pub_obj_type_list ',' pub_obj_type
+ { $$ = lappend($1, $3); }
+ ;
+
+
/*****************************************************************************
*
* ALTER PUBLICATION name SET ( options )
@@ -10566,11 +10597,6 @@ pub_obj_list: PublicationObjSpec
*
* ALTER PUBLICATION name SET pub_obj [, ...]
*
- * pub_obj is one of:
- *
- * TABLE table_name [, ...]
- * ALL TABLES IN SCHEMA schema_name [, ...]
- *
*****************************************************************************/
AlterPublicationStmt:
@@ -19477,7 +19503,8 @@ preprocess_pubobj_list(List *pubobjspec_list, core_yyscan_t yyscanner)
if (pubobj->pubobjtype == PUBLICATIONOBJ_CONTINUATION)
pubobj->pubobjtype = prevobjtype;
- if (pubobj->pubobjtype == PUBLICATIONOBJ_TABLE)
+ if (pubobj->pubobjtype == PUBLICATIONOBJ_TABLE ||
+ pubobj->pubobjtype == PUBLICATIONOBJ_SEQUENCE)
{
/* relation name or pubtable must be set for this type of object */
if (!pubobj->name && !pubobj->pubtable)
@@ -19528,6 +19555,30 @@ preprocess_pubobj_list(List *pubobjspec_list, core_yyscan_t yyscanner)
errmsg("invalid schema name at or near"),
parser_errposition(pubobj->location));
}
+ else if (pubobj->pubobjtype == PUBLICATIONOBJ_SEQUENCES_IN_SCHEMA ||
+ pubobj->pubobjtype == PUBLICATIONOBJ_SEQUENCES_IN_CUR_SCHEMA)
+ {
+ /* WHERE clause is not allowed on a schema object */
+ if (pubobj->pubtable && pubobj->pubtable->whereClause)
+ ereport(ERROR,
+ errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("WHERE clause not allowed for schema"),
+ parser_errposition(pubobj->location));
+
+ /*
+ * We can distinguish between the different type of schema
+ * objects based on whether name and pubtable is set.
+ */
+ if (pubobj->name)
+ pubobj->pubobjtype = PUBLICATIONOBJ_SEQUENCES_IN_SCHEMA;
+ else if (!pubobj->name && !pubobj->pubtable)
+ pubobj->pubobjtype = PUBLICATIONOBJ_SEQUENCES_IN_CUR_SCHEMA;
+ else
+ ereport(ERROR,
+ errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("invalid schema name at or near"),
+ parser_errposition(pubobj->location));
+ }
prevobjtype = pubobj->pubobjtype;
}
diff --git a/src/backend/replication/logical/decode.c b/src/backend/replication/logical/decode.c
index 1667d720b1..ef5f06a040 100644
--- a/src/backend/replication/logical/decode.c
+++ b/src/backend/replication/logical/decode.c
@@ -42,6 +42,10 @@
#include "replication/reorderbuffer.h"
#include "replication/snapbuild.h"
#include "storage/standby.h"
+#include "commands/sequence.h"
+
+static void DecodeLogicalMessage(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
+static void DecodeLogicalSequence(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
/* individual record(group)'s handlers */
static void DecodeInsert(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
@@ -558,6 +562,27 @@ FilterByOrigin(LogicalDecodingContext *ctx, RepOriginId origin_id)
*/
void
logicalmsg_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
+{
+ XLogReaderState *r = buf->record;
+ uint8 info = XLogRecGetInfo(r) & ~XLR_INFO_MASK;
+
+ switch (info)
+ {
+ case XLOG_LOGICAL_MESSAGE:
+ DecodeLogicalMessage(ctx, buf);
+ break;
+
+ case XLOG_LOGICAL_SEQUENCE:
+ DecodeLogicalSequence(ctx, buf);
+ break;
+
+ default:
+ elog(ERROR, "unexpected RM_LOGICALMSG_ID record type: %u", info);
+ }
+}
+
+static void
+DecodeLogicalMessage(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
{
SnapBuild *builder = ctx->snapshot_builder;
XLogReaderState *r = buf->record;
@@ -1251,3 +1276,67 @@ DecodeTXNNeedSkip(LogicalDecodingContext *ctx, XLogRecordBuffer *buf,
(txn_dbid != InvalidOid && txn_dbid != ctx->slot->data.database) ||
ctx->fast_forward || FilterByOrigin(ctx, origin_id));
}
+
+/*
+ * Handle sequence decode
+ *
+ * Decoding sequences is a bit tricky, because while most sequence actions
+ * are non-transactional (not subject to rollback), some need to be handled
+ * as transactional.
+ *
+ * By default, a sequence increment is non-transactional - we must not queue
+ * it in a transaction as other changes, because the transaction might get
+ * rolled back and we'd discard the increment. The downstream would not be
+ * notified about the increment, which is wrong.
+ *
+ * On the other hand, the sequence may be created in a transaction. In this
+ * case we *should* queue the change as other changes in the transaction,
+ * because we don't want to send the increments for unknown sequence to the
+ * plugin - it might get confused about which sequence it's related to etc.
+ */
+void
+DecodeLogicalSequence(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
+{
+ SnapBuild *builder = ctx->snapshot_builder;
+ RelFileLocator target_locator;
+ XLogReaderState *r = buf->record;
+ TransactionId xid = XLogRecGetXid(r);
+ uint8 info = XLogRecGetInfo(buf->record) & ~XLR_INFO_MASK;
+ xl_logical_sequence *xlrec;
+ RepOriginId origin_id = XLogRecGetOrigin(r);
+
+ /* only decode changes flagged with XLOG_SEQ_LOG */
+ if (info != XLOG_LOGICAL_SEQUENCE)
+ elog(ERROR, "unexpected RM_SEQ_ID record type: %u", info);
+
+ ReorderBufferProcessXid(ctx->reorder, XLogRecGetXid(r), buf->origptr);
+
+ /*
+ * If we don't have snapshot or we are just fast-forwarding, there is no
+ * point in decoding messages.
+ */
+ if (SnapBuildCurrentState(builder) < SNAPBUILD_FULL_SNAPSHOT ||
+ ctx->fast_forward)
+ return;
+
+ /* extract the WAL record, with "created" flag */
+ xlrec = (xl_logical_sequence *) XLogRecGetData(r);
+
+ /* only interested in our database */
+ target_locator = xlrec->locator;
+ if (target_locator.dbOid != ctx->slot->data.database)
+ return;
+
+ /* output plugin doesn't look for this origin, no need to queue */
+ if (FilterByOrigin(ctx, XLogRecGetOrigin(r)))
+ return;
+
+ /* Skip the change if already processed (per the snapshot). */
+ if (!SnapBuildProcessChange(builder, xid, buf->origptr))
+ return;
+
+ /* Queue the increment (or send immediately if not transactional). */
+ ReorderBufferQueueSequence(ctx->reorder, xid, buf->endptr,
+ origin_id, xlrec->reloid, target_locator,
+ xlrec->last, xlrec->log_cnt, xlrec->is_called);
+}
diff --git a/src/backend/replication/logical/logical.c b/src/backend/replication/logical/logical.c
index 625a7f4273..b1b4eaa8d8 100644
--- a/src/backend/replication/logical/logical.c
+++ b/src/backend/replication/logical/logical.c
@@ -73,6 +73,9 @@ static void truncate_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
static void message_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
XLogRecPtr message_lsn, bool transactional,
const char *prefix, Size message_size, const char *message);
+static void sequence_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ int64 last_value, int64 log_cnt, bool is_called);
/* streaming callbacks */
static void stream_start_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
@@ -90,6 +93,9 @@ static void stream_change_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn
static void stream_message_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
XLogRecPtr message_lsn, bool transactional,
const char *prefix, Size message_size, const char *message);
+static void stream_sequence_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ int64 last_value, int64 log_cnt, bool is_called);
static void stream_truncate_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
int nrelations, Relation relations[], ReorderBufferChange *change);
@@ -218,6 +224,7 @@ StartupDecodingContext(List *output_plugin_options,
ctx->reorder->apply_truncate = truncate_cb_wrapper;
ctx->reorder->commit = commit_cb_wrapper;
ctx->reorder->message = message_cb_wrapper;
+ ctx->reorder->sequence = sequence_cb_wrapper;
/*
* To support streaming, we require start/stop/abort/commit/change
@@ -234,6 +241,7 @@ StartupDecodingContext(List *output_plugin_options,
(ctx->callbacks.stream_commit_cb != NULL) ||
(ctx->callbacks.stream_change_cb != NULL) ||
(ctx->callbacks.stream_message_cb != NULL) ||
+ (ctx->callbacks.stream_sequence_cb != NULL) ||
(ctx->callbacks.stream_truncate_cb != NULL);
/*
@@ -251,6 +259,7 @@ StartupDecodingContext(List *output_plugin_options,
ctx->reorder->stream_commit = stream_commit_cb_wrapper;
ctx->reorder->stream_change = stream_change_cb_wrapper;
ctx->reorder->stream_message = stream_message_cb_wrapper;
+ ctx->reorder->stream_sequence = stream_sequence_cb_wrapper;
ctx->reorder->stream_truncate = stream_truncate_cb_wrapper;
@@ -1220,6 +1229,42 @@ message_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
error_context_stack = errcallback.previous;
}
+static void
+sequence_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ LogicalDecodingContext *ctx = cache->private_data;
+ LogicalErrorCallbackState state;
+ ErrorContextCallback errcallback;
+
+ Assert(!ctx->fast_forward);
+
+ if (ctx->callbacks.sequence_cb == NULL)
+ return;
+
+ /* Push callback + info on the error context stack */
+ state.ctx = ctx;
+ state.callback_name = "sequence";
+ state.report_location = sequence_lsn;
+ errcallback.callback = output_plugin_error_callback;
+ errcallback.arg = (void *) &state;
+ errcallback.previous = error_context_stack;
+ error_context_stack = &errcallback;
+
+ /* set output state */
+ ctx->accept_writes = true;
+ ctx->write_xid = txn != NULL ? txn->xid : InvalidTransactionId;
+ ctx->write_location = sequence_lsn;
+
+ /* do the actual work: call callback */
+ ctx->callbacks.sequence_cb(ctx, txn, sequence_lsn, rel,
+ last_value, log_cnt, is_called);
+
+ /* Pop the error context stack */
+ error_context_stack = errcallback.previous;
+}
+
static void
stream_start_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
XLogRecPtr first_lsn)
@@ -1535,6 +1580,46 @@ stream_message_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
error_context_stack = errcallback.previous;
}
+static void
+stream_sequence_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ LogicalDecodingContext *ctx = cache->private_data;
+ LogicalErrorCallbackState state;
+ ErrorContextCallback errcallback;
+
+ Assert(!ctx->fast_forward);
+
+ /* We're only supposed to call this when streaming is supported. */
+ Assert(ctx->streaming);
+
+ /* this callback is optional */
+ if (ctx->callbacks.stream_sequence_cb == NULL)
+ return;
+
+ /* Push callback + info on the error context stack */
+ state.ctx = ctx;
+ state.callback_name = "stream_sequence";
+ state.report_location = sequence_lsn;
+ errcallback.callback = output_plugin_error_callback;
+ errcallback.arg = (void *) &state;
+ errcallback.previous = error_context_stack;
+ error_context_stack = &errcallback;
+
+ /* set output state */
+ ctx->accept_writes = true;
+ ctx->write_xid = txn != NULL ? txn->xid : InvalidTransactionId;
+ ctx->write_location = sequence_lsn;
+
+ /* do the actual work: call callback */
+ ctx->callbacks.sequence_cb(ctx, txn, sequence_lsn, rel,
+ last_value, log_cnt, is_called);
+
+ /* Pop the error context stack */
+ error_context_stack = errcallback.previous;
+}
+
static void
stream_truncate_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
int nrelations, Relation relations[],
diff --git a/src/backend/replication/logical/message.c b/src/backend/replication/logical/message.c
index 1c34912610..69178dc1cf 100644
--- a/src/backend/replication/logical/message.c
+++ b/src/backend/replication/logical/message.c
@@ -82,7 +82,8 @@ logicalmsg_redo(XLogReaderState *record)
{
uint8 info = XLogRecGetInfo(record) & ~XLR_INFO_MASK;
- if (info != XLOG_LOGICAL_MESSAGE)
+ if (info != XLOG_LOGICAL_MESSAGE &&
+ info != XLOG_LOGICAL_SEQUENCE)
elog(PANIC, "logicalmsg_redo: unknown op code %u", info);
/* This is only interesting for logical decoding, see decode.c. */
diff --git a/src/backend/replication/logical/proto.c b/src/backend/replication/logical/proto.c
index ff8513e2d2..ae078eed04 100644
--- a/src/backend/replication/logical/proto.c
+++ b/src/backend/replication/logical/proto.c
@@ -662,6 +662,54 @@ logicalrep_write_message(StringInfo out, TransactionId xid, XLogRecPtr lsn,
pq_sendbytes(out, message, sz);
}
+/*
+ * Write SEQUENCE to stream
+ */
+void
+logicalrep_write_sequence(StringInfo out, Relation rel, TransactionId xid,
+ XLogRecPtr lsn,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ uint8 flags = 0;
+ char *relname;
+
+ pq_sendbyte(out, LOGICAL_REP_MSG_SEQUENCE);
+
+ /* transaction ID (if not valid, we're not streaming) */
+ if (TransactionIdIsValid(xid))
+ pq_sendint32(out, xid);
+
+ pq_sendint8(out, flags);
+ pq_sendint64(out, lsn);
+
+ logicalrep_write_namespace(out, RelationGetNamespace(rel));
+ relname = RelationGetRelationName(rel);
+ pq_sendstring(out, relname);
+
+ pq_sendint64(out, last_value);
+ pq_sendint64(out, log_cnt);
+ pq_sendint8(out, is_called);
+}
+
+/*
+ * Read SEQUENCE from the stream.
+ */
+void
+logicalrep_read_sequence(StringInfo in, LogicalRepSequence *seqdata)
+{
+ /* XXX skipping flags and lsn */
+ pq_getmsgint(in, 1);
+ pq_getmsgint64(in);
+
+ /* Read relation name from stream */
+ seqdata->nspname = pstrdup(logicalrep_read_namespace(in));
+ seqdata->seqname = pstrdup(pq_getmsgstring(in));
+
+ seqdata->last_value = pq_getmsgint64(in);
+ seqdata->log_cnt = pq_getmsgint64(in);
+ seqdata->is_called = pq_getmsgint(in, 1);
+}
+
/*
* Write relation description to the output stream.
*/
@@ -1236,6 +1284,8 @@ logicalrep_message_type(LogicalRepMsgType action)
return "STREAM ABORT";
case LOGICAL_REP_MSG_STREAM_PREPARE:
return "STREAM PREPARE";
+ case LOGICAL_REP_MSG_SEQUENCE:
+ return "SEQUENCE";
}
elog(ERROR, "invalid logical replication message type \"%c\"", action);
diff --git a/src/backend/replication/logical/reorderbuffer.c b/src/backend/replication/logical/reorderbuffer.c
index 1c21a1d14b..ca34a5ad99 100644
--- a/src/backend/replication/logical/reorderbuffer.c
+++ b/src/backend/replication/logical/reorderbuffer.c
@@ -77,6 +77,8 @@
* a bit more memory to the oldest subtransactions, because it's likely
* they are the source for the next sequence of changes.
*
+ * FIXME
+ *
* -------------------------------------------------------------------------
*/
#include "postgres.h"
@@ -91,6 +93,7 @@
#include "access/xact.h"
#include "access/xlog_internal.h"
#include "catalog/catalog.h"
+#include "commands/sequence.h"
#include "lib/binaryheap.h"
#include "miscadmin.h"
#include "pgstat.h"
@@ -539,6 +542,7 @@ ReorderBufferReturnChange(ReorderBuffer *rb, ReorderBufferChange *change,
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_ABORT:
case REORDER_BUFFER_CHANGE_INTERNAL_COMMAND_ID:
case REORDER_BUFFER_CHANGE_INTERNAL_TUPLECID:
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
break;
}
@@ -869,6 +873,45 @@ ReorderBufferQueueMessage(ReorderBuffer *rb, TransactionId xid,
}
}
+/*
+ * A transactional sequence increment is queued to be processed upon commit
+ * and a non-transactional increment gets processed immediately.
+ *
+ * A sequence update may be both transactional and non-transactional. When
+ * created in a running transaction, treat it as transactional and queue
+ * the change in it. Otherwise treat it as non-transactional, so that we
+ * don't forget the increment in case of a rollback.
+ */
+void
+ReorderBufferQueueSequence(ReorderBuffer *rb, TransactionId xid,
+ XLogRecPtr lsn, RepOriginId origin_id,
+ Oid reloid, RelFileLocator locator,
+ int64 last, int64 log_cnt, bool is_called)
+{
+ MemoryContext oldcontext;
+ ReorderBufferChange *change;
+
+ /* OK, allocate and queue the change */
+ oldcontext = MemoryContextSwitchTo(rb->context);
+
+ change = ReorderBufferGetChange(rb);
+
+ change->action = REORDER_BUFFER_CHANGE_SEQUENCE;
+ change->origin_id = origin_id;
+
+ memcpy(&change->data.sequence.locator, &locator, sizeof(RelFileLocator));
+
+ change->data.sequence.reloid = reloid;
+ change->data.sequence.last = last;
+ change->data.sequence.log_cnt = log_cnt;
+ change->data.sequence.is_called = is_called;
+
+ /* add it to the same subxact that created the sequence */
+ ReorderBufferQueueChange(rb, xid, lsn, change, false);
+
+ MemoryContextSwitchTo(oldcontext);
+}
+
/*
* AssertTXNLsnOrder
* Verify LSN ordering of transaction lists in the reorderbuffer
@@ -1968,6 +2011,33 @@ ReorderBufferApplyMessage(ReorderBuffer *rb, ReorderBufferTXN *txn,
change->data.msg.message);
}
+/*
+ * Helper function for ReorderBufferProcessTXN for applying sequences.
+ */
+static inline void
+ReorderBufferApplySequence(ReorderBuffer *rb, ReorderBufferTXN *txn,
+ Relation relation, ReorderBufferChange *change,
+ bool streaming)
+{
+ int64 last_value, log_cnt;
+ bool is_called;
+/*
+ * FIXME is it possible that we write sequence state for T1 before T2, but
+ * then end up committing T2 first? If the sequence could be incremented
+ * in between, that might cause data corruption result.
+ */
+ last_value = change->data.sequence.last;
+ log_cnt = change->data.sequence.log_cnt;
+ is_called = change->data.sequence.is_called;
+
+ if (streaming)
+ rb->stream_sequence(rb, txn, change->lsn, relation,
+ last_value, log_cnt, is_called);
+ else
+ rb->sequence(rb, txn, change->lsn, relation,
+ last_value, log_cnt, is_called);
+}
+
/*
* Function to store the command id and snapshot at the end of the current
* stream so that we can reuse the same while sending the next stream.
@@ -2409,6 +2479,33 @@ ReorderBufferProcessTXN(ReorderBuffer *rb, ReorderBufferTXN *txn,
case REORDER_BUFFER_CHANGE_INTERNAL_TUPLECID:
elog(ERROR, "tuplecid value in changequeue");
break;
+
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
+ Assert(snapshot_now);
+
+ /*
+ reloid = RelidByRelfilenode(change->data.sequence.relnode.spcNode,
+ change->data.sequence.relnode.relNode);
+
+ if (reloid == InvalidOid)
+ elog(ERROR, "zz could not map filenode \"%s\" to relation OID",
+ relpathperm(change->data.sequence.relnode,
+ MAIN_FORKNUM));
+
+ */
+ relation = RelationIdGetRelation(change->data.sequence.reloid);
+
+ if (!RelationIsValid(relation))
+ elog(ERROR, "could not open relation with OID %u (for filenode \"%s\")",
+ reloid,
+ relpathperm(change->data.sequence.locator,
+ MAIN_FORKNUM));
+
+ if (RelationIsLogicallyLogged(relation))
+ ReorderBufferApplySequence(rb, txn, relation, change, streaming);
+
+ RelationClose(relation);
+ break;
}
}
@@ -3849,6 +3946,7 @@ ReorderBufferSerializeChange(ReorderBuffer *rb, ReorderBufferTXN *txn,
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_ABORT:
case REORDER_BUFFER_CHANGE_INTERNAL_COMMAND_ID:
case REORDER_BUFFER_CHANGE_INTERNAL_TUPLECID:
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
/* ReorderBufferChange contains everything important */
break;
}
@@ -4113,6 +4211,7 @@ ReorderBufferChangeSize(ReorderBufferChange *change)
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_ABORT:
case REORDER_BUFFER_CHANGE_INTERNAL_COMMAND_ID:
case REORDER_BUFFER_CHANGE_INTERNAL_TUPLECID:
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
/* ReorderBufferChange contains everything important */
break;
}
@@ -4412,6 +4511,7 @@ ReorderBufferRestoreChange(ReorderBuffer *rb, ReorderBufferTXN *txn,
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_ABORT:
case REORDER_BUFFER_CHANGE_INTERNAL_COMMAND_ID:
case REORDER_BUFFER_CHANGE_INTERNAL_TUPLECID:
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
break;
}
diff --git a/src/backend/replication/logical/tablesync.c b/src/backend/replication/logical/tablesync.c
index bfcb80b495..d63e3a0ad2 100644
--- a/src/backend/replication/logical/tablesync.c
+++ b/src/backend/replication/logical/tablesync.c
@@ -100,6 +100,7 @@
#include "catalog/pg_subscription_rel.h"
#include "catalog/pg_type.h"
#include "commands/copy.h"
+#include "commands/sequence.h"
#include "miscadmin.h"
#include "parser/parse_relation.h"
#include "pgstat.h"
@@ -1134,6 +1135,94 @@ copy_table(Relation rel)
logicalrep_rel_close(relmapentry, NoLock);
}
+/*
+ * Fetch sequence data (current state) from the remote node.
+ */
+static void
+fetch_sequence_data(char *nspname, char *relname,
+ int64 *last_value, int64 *log_cnt, bool *is_called)
+{
+ WalRcvExecResult *res;
+ StringInfoData cmd;
+ TupleTableSlot *slot;
+ Oid tableRow[3] = {INT8OID, INT8OID, BOOLOID};
+
+ initStringInfo(&cmd);
+ appendStringInfo(&cmd, "SELECT last_value, log_cnt, is_called\n"
+ " FROM %s", quote_qualified_identifier(nspname, relname));
+
+ res = walrcv_exec(LogRepWorkerWalRcvConn, cmd.data, 3, tableRow);
+ pfree(cmd.data);
+
+ if (res->status != WALRCV_OK_TUPLES)
+ ereport(ERROR,
+ (errmsg("could not receive list of replicated tables from the publisher: %s",
+ res->err)));
+
+ /* Process the sequence. */
+ slot = MakeSingleTupleTableSlot(res->tupledesc, &TTSOpsMinimalTuple);
+ while (tuplestore_gettupleslot(res->tuplestore, true, false, slot))
+ {
+ bool isnull;
+
+ *last_value = DatumGetInt64(slot_getattr(slot, 1, &isnull));
+ Assert(!isnull);
+
+ *log_cnt = DatumGetInt64(slot_getattr(slot, 2, &isnull));
+ Assert(!isnull);
+
+ *is_called = DatumGetBool(slot_getattr(slot, 3, &isnull));
+ Assert(!isnull);
+
+ ExecClearTuple(slot);
+ }
+ ExecDropSingleTupleTableSlot(slot);
+
+ walrcv_clear_result(res);
+}
+
+/*
+ * Copy existing data of a sequence from publisher.
+ *
+ * Caller is responsible for locking the local relation.
+ */
+static void
+copy_sequence(Relation rel)
+{
+ LogicalRepRelMapEntry *relmapentry;
+ LogicalRepRelation lrel;
+ List *qual = NIL;
+ StringInfoData cmd;
+ int64 last_value = 0,
+ log_cnt = 0;
+ bool is_called = 0;
+
+ /* Get the publisher relation info. */
+ fetch_remote_table_info(get_namespace_name(RelationGetNamespace(rel)),
+ RelationGetRelationName(rel), &lrel, &qual);
+
+ /* sequences don't have row filters */
+ Assert(!qual);
+
+ /* Put the relation into relmap. */
+ logicalrep_relmap_update(&lrel);
+
+ /* Map the publisher relation to local one. */
+ relmapentry = logicalrep_rel_open(lrel.remoteid, NoLock);
+ Assert(rel == relmapentry->localrel);
+
+ /* Start copy on the publisher. */
+ initStringInfo(&cmd);
+
+ Assert(lrel.relkind == RELKIND_SEQUENCE);
+
+ fetch_sequence_data(lrel.nspname, lrel.relname, &last_value, &log_cnt, &is_called);
+
+ SetSequence(RelationGetRelid(rel), last_value, log_cnt, is_called);
+
+ logicalrep_rel_close(relmapentry, NoLock);
+}
+
/*
* Determine the tablesync slot name.
*
@@ -1395,10 +1484,21 @@ LogicalRepSyncTableStart(XLogRecPtr *origin_startpos)
originname)));
}
- /* Now do the initial data copy */
- PushActiveSnapshot(GetTransactionSnapshot());
- copy_table(rel);
- PopActiveSnapshot();
+ /* Do the right action depending on the relation kind. */
+ if (get_rel_relkind(RelationGetRelid(rel)) == RELKIND_SEQUENCE)
+ {
+ /* Now do the initial sequence copy */
+ PushActiveSnapshot(GetTransactionSnapshot());
+ copy_sequence(rel);
+ PopActiveSnapshot();
+ }
+ else
+ {
+ /* Now do the initial data copy */
+ PushActiveSnapshot(GetTransactionSnapshot());
+ copy_table(rel);
+ PopActiveSnapshot();
+ }
res = walrcv_exec(LogRepWorkerWalRcvConn, "COMMIT", 0, NULL);
if (res->status != WALRCV_OK_COMMAND)
diff --git a/src/backend/replication/logical/worker.c b/src/backend/replication/logical/worker.c
index 5f8c541763..b1de217626 100644
--- a/src/backend/replication/logical/worker.c
+++ b/src/backend/replication/logical/worker.c
@@ -143,6 +143,7 @@
#include "catalog/pg_subscription.h"
#include "catalog/pg_subscription_rel.h"
#include "catalog/pg_tablespace.h"
+#include "commands/sequence.h"
#include "commands/tablecmds.h"
#include "commands/tablespace.h"
#include "commands/trigger.h"
@@ -1142,6 +1143,47 @@ apply_handle_origin(StringInfo s)
errmsg_internal("ORIGIN message sent out of order")));
}
+/*
+ * Handle SEQUENCE message.
+ */
+static void
+apply_handle_sequence(StringInfo s)
+{
+ LogicalRepSequence seq;
+ Oid relid;
+
+ if (handle_streamed_transaction(LOGICAL_REP_MSG_SEQUENCE, s))
+ return;
+
+ logicalrep_read_sequence(s, &seq);
+
+ /*
+ * We should be in remote transaction.
+ */
+ Assert(in_remote_transaction);
+
+ /*
+ * Make sure we're in a transaction (needed by SetSequence). For
+ * non-transactional updates we're guaranteed to start a new one,
+ * and we'll commit it at the end.
+ */
+ if (!IsTransactionState())
+ {
+ StartTransactionCommand();
+ maybe_reread_subscription();
+ }
+
+ relid = RangeVarGetRelid(makeRangeVar(seq.nspname,
+ seq.seqname, -1),
+ RowExclusiveLock, false);
+
+ /* lock the sequence in AccessExclusiveLock, as expected by SetSequence */
+ LockRelationOid(relid, AccessExclusiveLock);
+
+ /* apply the sequence change */
+ SetSequence(relid, seq.last_value, seq.log_cnt, seq.is_called);
+}
+
/*
* Handle STREAM START message.
*/
@@ -2528,6 +2570,10 @@ apply_dispatch(StringInfo s)
*/
break;
+ case LOGICAL_REP_MSG_SEQUENCE:
+ apply_handle_sequence(s);
+ return;
+
case LOGICAL_REP_MSG_STREAM_START:
apply_handle_stream_start(s);
break;
diff --git a/src/backend/replication/pgoutput/pgoutput.c b/src/backend/replication/pgoutput/pgoutput.c
index 62e0ffecd8..642c93a45f 100644
--- a/src/backend/replication/pgoutput/pgoutput.c
+++ b/src/backend/replication/pgoutput/pgoutput.c
@@ -15,6 +15,7 @@
#include "access/tupconvert.h"
#include "catalog/partition.h"
#include "catalog/pg_publication.h"
+#include "catalog/pg_publication_namespace.h"
#include "catalog/pg_publication_rel.h"
#include "catalog/pg_subscription.h"
#include "commands/defrem.h"
@@ -53,6 +54,10 @@ static void pgoutput_message(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn, XLogRecPtr message_lsn,
bool transactional, const char *prefix,
Size sz, const char *message);
+static void pgoutput_sequence(LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn, XLogRecPtr sequence_lsn,
+ Relation relation,
+ int64 last_value, int64 log_cnt, bool is_called);
static bool pgoutput_origin_filter(LogicalDecodingContext *ctx,
RepOriginId origin_id);
static void pgoutput_begin_prepare_txn(LogicalDecodingContext *ctx,
@@ -256,6 +261,7 @@ _PG_output_plugin_init(OutputPluginCallbacks *cb)
cb->change_cb = pgoutput_change;
cb->truncate_cb = pgoutput_truncate;
cb->message_cb = pgoutput_message;
+ cb->sequence_cb = pgoutput_sequence;
cb->commit_cb = pgoutput_commit_txn;
cb->begin_prepare_cb = pgoutput_begin_prepare_txn;
@@ -272,6 +278,7 @@ _PG_output_plugin_init(OutputPluginCallbacks *cb)
cb->stream_commit_cb = pgoutput_stream_commit;
cb->stream_change_cb = pgoutput_change;
cb->stream_message_cb = pgoutput_message;
+ cb->stream_sequence_cb = pgoutput_sequence;
cb->stream_truncate_cb = pgoutput_truncate;
/* transaction streaming - two-phase commit */
cb->stream_prepare_cb = pgoutput_stream_prepare_txn;
@@ -285,6 +292,7 @@ parse_output_parameters(List *options, PGOutputData *data)
bool publication_names_given = false;
bool binary_option_given = false;
bool messages_option_given = false;
+ bool sequences_option_given = false;
bool streaming_given = false;
bool two_phase_option_given = false;
bool origin_option_given = false;
@@ -293,6 +301,7 @@ parse_output_parameters(List *options, PGOutputData *data)
data->streaming = false;
data->messages = false;
data->two_phase = false;
+ data->sequences = true;
foreach(lc, options)
{
@@ -361,6 +370,16 @@ parse_output_parameters(List *options, PGOutputData *data)
data->messages = defGetBoolean(defel);
}
+ else if (strcmp(defel->defname, "sequences") == 0)
+ {
+ if (sequences_option_given)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("conflicting or redundant options")));
+ sequences_option_given = true;
+
+ data->sequences = defGetBoolean(defel);
+ }
else if (strcmp(defel->defname, "streaming") == 0)
{
if (streaming_given)
@@ -1716,6 +1735,55 @@ pgoutput_message(LogicalDecodingContext *ctx, ReorderBufferTXN *txn,
OutputPluginWrite(ctx, true);
}
+static void
+pgoutput_sequence(LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn, XLogRecPtr sequence_lsn,
+ Relation relation,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ PGOutputData *data = (PGOutputData *) ctx->output_plugin_private;
+ TransactionId xid = InvalidTransactionId;
+ RelationSyncEntry *relentry;
+ PGOutputTxnData *txndata = (PGOutputTxnData *) txn->output_plugin_private;
+
+ if (!data->sequences)
+ return;
+
+ if (!is_publishable_relation(relation))
+ return;
+
+ /*
+ * Remember the xid for the message in streaming mode. See
+ * pgoutput_change.
+ */
+ if (in_streaming)
+ xid = txn->xid;
+
+ relentry = get_rel_sync_entry(data, relation);
+
+ /*
+ * First check the sequence filter.
+ *
+ * We handle just REORDER_BUFFER_CHANGE_SEQUENCE here.
+ */
+ if (!relentry->pubactions.pubsequence)
+ return;
+
+ /* Send BEGIN if we haven't yet */
+ if (txndata && !txndata->sent_begin_txn)
+ pgoutput_send_begin(ctx, txn);
+
+ OutputPluginPrepareWrite(ctx, true);
+ logicalrep_write_sequence(ctx->out,
+ relation,
+ xid,
+ sequence_lsn,
+ last_value,
+ log_cnt,
+ is_called);
+ OutputPluginWrite(ctx, true);
+}
+
/*
* Return true if the data is associated with an origin and the user has
* requested the changes that don't have an origin, false otherwise.
@@ -1997,7 +2065,8 @@ get_rel_sync_entry(PGOutputData *data, Relation relation)
entry->schema_sent = false;
entry->streamed_txns = NIL;
entry->pubactions.pubinsert = entry->pubactions.pubupdate =
- entry->pubactions.pubdelete = entry->pubactions.pubtruncate = false;
+ entry->pubactions.pubdelete = entry->pubactions.pubtruncate =
+ entry->pubactions.pubsequence = false;
entry->new_slot = NULL;
entry->old_slot = NULL;
memset(entry->exprstate, 0, sizeof(entry->exprstate));
@@ -2012,18 +2081,18 @@ get_rel_sync_entry(PGOutputData *data, Relation relation)
{
Oid schemaId = get_rel_namespace(relid);
List *pubids = GetRelationPublications(relid);
-
+ char relkind = get_rel_relkind(relid);
+ char objectType = pub_get_object_type_for_relkind(relkind);
/*
* We don't acquire a lock on the namespace system table as we build
* the cache entry using a historic snapshot and all the later changes
* are absorbed while decoding WAL.
*/
- List *schemaPubids = GetSchemaPublications(schemaId);
+ List *schemaPubids = GetSchemaPublications(schemaId, objectType);
ListCell *lc;
Oid publish_as_relid = relid;
int publish_ancestor_level = 0;
bool am_partition = get_rel_relispartition(relid);
- char relkind = get_rel_relkind(relid);
List *rel_publications = NIL;
/* Reload publications if needed before use. */
@@ -2055,6 +2124,7 @@ get_rel_sync_entry(PGOutputData *data, Relation relation)
entry->pubactions.pubupdate = false;
entry->pubactions.pubdelete = false;
entry->pubactions.pubtruncate = false;
+ entry->pubactions.pubsequence = false;
/*
* Tuple slots cleanups. (Will be rebuilt later if needed).
@@ -2102,9 +2172,11 @@ get_rel_sync_entry(PGOutputData *data, Relation relation)
/*
* If this is a FOR ALL TABLES publication, pick the partition
- * root and set the ancestor level accordingly.
+ * root and set the ancestor level accordingly. If this is a FOR ALL
+ * SEQUENCES publication, we publish it too but we don't need to
+ * pick the partition root etc.
*/
- if (pub->alltables)
+ if (pub->alltables || pub->allsequences)
{
publish = true;
if (pub->pubviaroot && am_partition)
@@ -2168,6 +2240,7 @@ get_rel_sync_entry(PGOutputData *data, Relation relation)
entry->pubactions.pubupdate |= pub->pubactions.pubupdate;
entry->pubactions.pubdelete |= pub->pubactions.pubdelete;
entry->pubactions.pubtruncate |= pub->pubactions.pubtruncate;
+ entry->pubactions.pubsequence |= pub->pubactions.pubsequence;
/*
* We want to publish the changes as the top-most ancestor
diff --git a/src/backend/utils/cache/relcache.c b/src/backend/utils/cache/relcache.c
index 00dc0f2403..e427f05ec2 100644
--- a/src/backend/utils/cache/relcache.c
+++ b/src/backend/utils/cache/relcache.c
@@ -57,6 +57,7 @@
#include "catalog/pg_opclass.h"
#include "catalog/pg_proc.h"
#include "catalog/pg_publication.h"
+#include "catalog/pg_publication_namespace.h"
#include "catalog/pg_rewrite.h"
#include "catalog/pg_shseclabel.h"
#include "catalog/pg_statistic_ext.h"
@@ -5616,6 +5617,8 @@ RelationBuildPublicationDesc(Relation relation, PublicationDesc *pubdesc)
Oid schemaid;
List *ancestors = NIL;
Oid relid = RelationGetRelid(relation);
+ char relkind = relation->rd_rel->relkind;
+ char objType;
/*
* If not publishable, it publishes no actions. (pgoutput_change() will
@@ -5646,8 +5649,15 @@ RelationBuildPublicationDesc(Relation relation, PublicationDesc *pubdesc)
/* Fetch the publication membership info. */
puboids = GetRelationPublications(relid);
schemaid = RelationGetNamespace(relation);
- puboids = list_concat_unique_oid(puboids, GetSchemaPublications(schemaid));
+ objType = pub_get_object_type_for_relkind(relkind);
+ puboids = list_concat_unique_oid(puboids,
+ GetSchemaPublications(schemaid, objType));
+
+ /*
+ * If this is a partion (and thus a table), lookup all ancestors and track
+ * all publications them too.
+ */
if (relation->rd_rel->relispartition)
{
/* Add publications that the ancestors are in too. */
@@ -5659,12 +5669,23 @@ RelationBuildPublicationDesc(Relation relation, PublicationDesc *pubdesc)
puboids = list_concat_unique_oid(puboids,
GetRelationPublications(ancestor));
+
+ /* include all publications publishing schema of all ancestors */
schemaid = get_rel_namespace(ancestor);
puboids = list_concat_unique_oid(puboids,
- GetSchemaPublications(schemaid));
+ GetSchemaPublications(schemaid,
+ PUB_OBJTYPE_TABLE));
}
}
- puboids = list_concat_unique_oid(puboids, GetAllTablesPublications());
+
+ /*
+ * Consider also FOR ALL TABLES and FOR ALL SEQUENCES publications,
+ * depending on the relkind of the relation.
+ */
+ if (relation->rd_rel->relkind == RELKIND_SEQUENCE)
+ puboids = list_concat_unique_oid(puboids, GetAllSequencesPublications());
+ else
+ puboids = list_concat_unique_oid(puboids, GetAllTablesPublications());
foreach(lc, puboids)
{
@@ -5683,6 +5704,7 @@ RelationBuildPublicationDesc(Relation relation, PublicationDesc *pubdesc)
pubdesc->pubactions.pubupdate |= pubform->pubupdate;
pubdesc->pubactions.pubdelete |= pubform->pubdelete;
pubdesc->pubactions.pubtruncate |= pubform->pubtruncate;
+ pubdesc->pubactions.pubsequence |= pubform->pubsequence;
/*
* Check if all columns referenced in the filter expression are part
diff --git a/src/backend/utils/cache/syscache.c b/src/backend/utils/cache/syscache.c
index 1912b12146..8d265f2d23 100644
--- a/src/backend/utils/cache/syscache.c
+++ b/src/backend/utils/cache/syscache.c
@@ -653,12 +653,12 @@ static const struct cachedesc cacheinfo[] = {
64
},
{PublicationNamespaceRelationId, /* PUBLICATIONNAMESPACEMAP */
- PublicationNamespacePnnspidPnpubidIndexId,
- 2,
+ PublicationNamespacePnnspidPnpubidPntypeIndexId,
+ 3,
{
Anum_pg_publication_namespace_pnnspid,
Anum_pg_publication_namespace_pnpubid,
- 0,
+ Anum_pg_publication_namespace_pntype,
0
},
64
diff --git a/src/bin/pg_dump/pg_dump.c b/src/bin/pg_dump/pg_dump.c
index da6605175a..fbd3c22c39 100644
--- a/src/bin/pg_dump/pg_dump.c
+++ b/src/bin/pg_dump/pg_dump.c
@@ -3892,10 +3892,12 @@ getPublications(Archive *fout, int *numPublications)
int i_pubname;
int i_pubowner;
int i_puballtables;
+ int i_puballsequences;
int i_pubinsert;
int i_pubupdate;
int i_pubdelete;
int i_pubtruncate;
+ int i_pubsequence;
int i_pubviaroot;
int i,
ntups;
@@ -3911,23 +3913,29 @@ getPublications(Archive *fout, int *numPublications)
resetPQExpBuffer(query);
/* Get the publications. */
- if (fout->remoteVersion >= 130000)
+ if (fout->remoteVersion >= 160000)
+ appendPQExpBuffer(query,
+ "SELECT p.tableoid, p.oid, p.pubname, "
+ "p.pubowner, "
+ "p.puballtables, p.puballsequences, p.pubinsert, p.pubupdate, p.pubdelete, p.pubtruncate, p.pubsequence, p.pubviaroot "
+ "FROM pg_publication p");
+ else if (fout->remoteVersion >= 130000)
appendPQExpBuffer(query,
"SELECT p.tableoid, p.oid, p.pubname, "
"p.pubowner, "
- "p.puballtables, p.pubinsert, p.pubupdate, p.pubdelete, p.pubtruncate, p.pubviaroot "
+ "p.puballtables, false AS puballsequences, p.pubinsert, p.pubupdate, p.pubdelete, p.pubtruncate, false AS pubsequence, p.pubviaroot "
"FROM pg_publication p");
else if (fout->remoteVersion >= 110000)
appendPQExpBuffer(query,
"SELECT p.tableoid, p.oid, p.pubname, "
"p.pubowner, "
- "p.puballtables, p.pubinsert, p.pubupdate, p.pubdelete, p.pubtruncate, false AS pubviaroot "
+ "p.puballtables, false AS puballsequences, p.pubinsert, p.pubupdate, p.pubdelete, p.pubtruncate, false AS pubsequence, false AS pubviaroot "
"FROM pg_publication p");
else
appendPQExpBuffer(query,
"SELECT p.tableoid, p.oid, p.pubname, "
"p.pubowner, "
- "p.puballtables, p.pubinsert, p.pubupdate, p.pubdelete, false AS pubtruncate, false AS pubviaroot "
+ "p.puballtables, false AS puballsequences, p.pubinsert, p.pubupdate, p.pubdelete, false AS pubtruncate, false AS pubsequence, false AS pubviaroot "
"FROM pg_publication p");
res = ExecuteSqlQuery(fout, query->data, PGRES_TUPLES_OK);
@@ -3939,10 +3947,12 @@ getPublications(Archive *fout, int *numPublications)
i_pubname = PQfnumber(res, "pubname");
i_pubowner = PQfnumber(res, "pubowner");
i_puballtables = PQfnumber(res, "puballtables");
+ i_puballsequences = PQfnumber(res, "puballsequences");
i_pubinsert = PQfnumber(res, "pubinsert");
i_pubupdate = PQfnumber(res, "pubupdate");
i_pubdelete = PQfnumber(res, "pubdelete");
i_pubtruncate = PQfnumber(res, "pubtruncate");
+ i_pubsequence = PQfnumber(res, "pubsequence");
i_pubviaroot = PQfnumber(res, "pubviaroot");
pubinfo = pg_malloc(ntups * sizeof(PublicationInfo));
@@ -3958,6 +3968,8 @@ getPublications(Archive *fout, int *numPublications)
pubinfo[i].rolname = getRoleName(PQgetvalue(res, i, i_pubowner));
pubinfo[i].puballtables =
(strcmp(PQgetvalue(res, i, i_puballtables), "t") == 0);
+ pubinfo[i].puballsequences =
+ (strcmp(PQgetvalue(res, i, i_puballsequences), "t") == 0);
pubinfo[i].pubinsert =
(strcmp(PQgetvalue(res, i, i_pubinsert), "t") == 0);
pubinfo[i].pubupdate =
@@ -3966,6 +3978,8 @@ getPublications(Archive *fout, int *numPublications)
(strcmp(PQgetvalue(res, i, i_pubdelete), "t") == 0);
pubinfo[i].pubtruncate =
(strcmp(PQgetvalue(res, i, i_pubtruncate), "t") == 0);
+ pubinfo[i].pubsequence =
+ (strcmp(PQgetvalue(res, i, i_pubsequence), "t") == 0);
pubinfo[i].pubviaroot =
(strcmp(PQgetvalue(res, i, i_pubviaroot), "t") == 0);
@@ -4011,6 +4025,9 @@ dumpPublication(Archive *fout, const PublicationInfo *pubinfo)
if (pubinfo->puballtables)
appendPQExpBufferStr(query, " FOR ALL TABLES");
+ if (pubinfo->puballsequences)
+ appendPQExpBufferStr(query, " FOR ALL SEQUENCES");
+
appendPQExpBufferStr(query, " WITH (publish = '");
if (pubinfo->pubinsert)
{
@@ -4045,6 +4062,15 @@ dumpPublication(Archive *fout, const PublicationInfo *pubinfo)
first = false;
}
+ if (pubinfo->pubsequence)
+ {
+ if (!first)
+ appendPQExpBufferStr(query, ", ");
+
+ appendPQExpBufferStr(query, "sequence");
+ first = false;
+ }
+
appendPQExpBufferStr(query, "'");
if (pubinfo->pubviaroot)
@@ -4091,6 +4117,7 @@ getPublicationNamespaces(Archive *fout)
int i_oid;
int i_pnpubid;
int i_pnnspid;
+ int i_pntype;
int i,
j,
ntups;
@@ -4102,7 +4129,7 @@ getPublicationNamespaces(Archive *fout)
/* Collect all publication membership info. */
appendPQExpBufferStr(query,
- "SELECT tableoid, oid, pnpubid, pnnspid "
+ "SELECT tableoid, oid, pnpubid, pnnspid, pntype "
"FROM pg_catalog.pg_publication_namespace");
res = ExecuteSqlQuery(fout, query->data, PGRES_TUPLES_OK);
@@ -4112,6 +4139,7 @@ getPublicationNamespaces(Archive *fout)
i_oid = PQfnumber(res, "oid");
i_pnpubid = PQfnumber(res, "pnpubid");
i_pnnspid = PQfnumber(res, "pnnspid");
+ i_pntype = PQfnumber(res, "pntype");
/* this allocation may be more than we need */
pubsinfo = pg_malloc(ntups * sizeof(PublicationSchemaInfo));
@@ -4121,6 +4149,7 @@ getPublicationNamespaces(Archive *fout)
{
Oid pnpubid = atooid(PQgetvalue(res, i, i_pnpubid));
Oid pnnspid = atooid(PQgetvalue(res, i, i_pnnspid));
+ char pntype = PQgetvalue(res, i, i_pntype)[0];
PublicationInfo *pubinfo;
NamespaceInfo *nspinfo;
@@ -4152,6 +4181,7 @@ getPublicationNamespaces(Archive *fout)
pubsinfo[j].dobj.name = nspinfo->dobj.name;
pubsinfo[j].publication = pubinfo;
pubsinfo[j].pubschema = nspinfo;
+ pubsinfo[j].pubtype = pntype;
/* Decide whether we want to dump it */
selectDumpablePublicationObject(&(pubsinfo[j].dobj), fout);
@@ -4317,7 +4347,11 @@ dumpPublicationNamespace(Archive *fout, const PublicationSchemaInfo *pubsinfo)
query = createPQExpBuffer();
appendPQExpBuffer(query, "ALTER PUBLICATION %s ", fmtId(pubinfo->dobj.name));
- appendPQExpBuffer(query, "ADD ALL TABLES IN SCHEMA %s;\n", fmtId(schemainfo->dobj.name));
+
+ if (pubsinfo->pubtype == 't')
+ appendPQExpBuffer(query, "ADD ALL TABLES IN SCHEMA %s;\n", fmtId(schemainfo->dobj.name));
+ else
+ appendPQExpBuffer(query, "ADD ALL SEQUENCES IN SCHEMA %s;\n", fmtId(schemainfo->dobj.name));
/*
* There is no point in creating drop query as the drop is done by schema
@@ -4350,6 +4384,7 @@ dumpPublicationTable(Archive *fout, const PublicationRelInfo *pubrinfo)
TableInfo *tbinfo = pubrinfo->pubtable;
PQExpBuffer query;
char *tag;
+ char *description;
/* Do nothing in data-only dump */
if (dopt->dataOnly)
@@ -4359,8 +4394,19 @@ dumpPublicationTable(Archive *fout, const PublicationRelInfo *pubrinfo)
query = createPQExpBuffer();
- appendPQExpBuffer(query, "ALTER PUBLICATION %s ADD TABLE ONLY",
- fmtId(pubinfo->dobj.name));
+ if (tbinfo->relkind == RELKIND_SEQUENCE)
+ {
+ appendPQExpBuffer(query, "ALTER PUBLICATION %s ADD SEQUENCE",
+ fmtId(pubinfo->dobj.name));
+ description = "PUBLICATION SEQUENCE";
+ }
+ else
+ {
+ appendPQExpBuffer(query, "ALTER PUBLICATION %s ADD TABLE ONLY",
+ fmtId(pubinfo->dobj.name));
+ description = "PUBLICATION TABLE";
+ }
+
appendPQExpBuffer(query, " %s",
fmtQualifiedDumpable(tbinfo));
@@ -4390,7 +4436,7 @@ dumpPublicationTable(Archive *fout, const PublicationRelInfo *pubrinfo)
ARCHIVE_OPTS(.tag = tag,
.namespace = tbinfo->dobj.namespace->dobj.name,
.owner = pubinfo->rolname,
- .description = "PUBLICATION TABLE",
+ .description = description,
.section = SECTION_POST_DATA,
.createStmt = query->data));
diff --git a/src/bin/pg_dump/pg_dump.h b/src/bin/pg_dump/pg_dump.h
index 69ee939d44..1ea4ca7e25 100644
--- a/src/bin/pg_dump/pg_dump.h
+++ b/src/bin/pg_dump/pg_dump.h
@@ -615,10 +615,12 @@ typedef struct _PublicationInfo
DumpableObject dobj;
const char *rolname;
bool puballtables;
+ bool puballsequences;
bool pubinsert;
bool pubupdate;
bool pubdelete;
bool pubtruncate;
+ bool pubsequence;
bool pubviaroot;
} PublicationInfo;
@@ -644,6 +646,7 @@ typedef struct _PublicationSchemaInfo
DumpableObject dobj;
PublicationInfo *publication;
NamespaceInfo *pubschema;
+ char pubtype;
} PublicationSchemaInfo;
/*
diff --git a/src/bin/pg_dump/t/002_pg_dump.pl b/src/bin/pg_dump/t/002_pg_dump.pl
index b10e1c4c0d..c0eda30c55 100644
--- a/src/bin/pg_dump/t/002_pg_dump.pl
+++ b/src/bin/pg_dump/t/002_pg_dump.pl
@@ -2420,7 +2420,7 @@ my %tests = (
create_order => 50,
create_sql => 'CREATE PUBLICATION pub1;',
regexp => qr/^
- \QCREATE PUBLICATION pub1 WITH (publish = 'insert, update, delete, truncate');\E
+ \QCREATE PUBLICATION pub1 WITH (publish = 'insert, update, delete, truncate, sequence');\E
/xm,
like => { %full_runs, section_post_data => 1, },
},
@@ -2440,16 +2440,27 @@ my %tests = (
create_order => 50,
create_sql => 'CREATE PUBLICATION pub3;',
regexp => qr/^
- \QCREATE PUBLICATION pub3 WITH (publish = 'insert, update, delete, truncate');\E
+ \QCREATE PUBLICATION pub3 WITH (publish = 'insert, update, delete, truncate, sequence');\E
/xm,
like => { %full_runs, section_post_data => 1, },
},
'CREATE PUBLICATION pub4' => {
create_order => 50,
- create_sql => 'CREATE PUBLICATION pub4;',
+ create_sql => 'CREATE PUBLICATION pub4
+ FOR ALL SEQUENCES
+ WITH (publish = \'\');',
regexp => qr/^
- \QCREATE PUBLICATION pub4 WITH (publish = 'insert, update, delete, truncate');\E
+ \QCREATE PUBLICATION pub4 FOR ALL SEQUENCES WITH (publish = '');\E
+ /xm,
+ like => { %full_runs, section_post_data => 1, },
+ },
+
+ 'CREATE PUBLICATION pub5' => {
+ create_order => 50,
+ create_sql => 'CREATE PUBLICATION pub5;',
+ regexp => qr/^
+ \QCREATE PUBLICATION pub5 WITH (publish = 'insert, update, delete, truncate, sequence');\E
/xm,
like => { %full_runs, section_post_data => 1, },
},
@@ -2582,6 +2593,27 @@ my %tests = (
unlike => { exclude_dump_test_schema => 1, },
},
+ 'ALTER PUBLICATION pub3 ADD ALL SEQUENCES IN SCHEMA dump_test' => {
+ create_order => 51,
+ create_sql =>
+ 'ALTER PUBLICATION pub3 ADD ALL SEQUENCES IN SCHEMA dump_test;',
+ regexp => qr/^
+ \QALTER PUBLICATION pub3 ADD ALL SEQUENCES IN SCHEMA dump_test;\E
+ /xm,
+ like => { %full_runs, section_post_data => 1, },
+ unlike => { exclude_dump_test_schema => 1, },
+ },
+
+ 'ALTER PUBLICATION pub3 ADD ALL SEQUENCES IN SCHEMA public' => {
+ create_order => 52,
+ create_sql =>
+ 'ALTER PUBLICATION pub3 ADD ALL SEQUENCES IN SCHEMA public;',
+ regexp => qr/^
+ \QALTER PUBLICATION pub3 ADD ALL SEQUENCES IN SCHEMA public;\E
+ /xm,
+ like => { %full_runs, section_post_data => 1, },
+ },
+
'CREATE SCHEMA public' => {
regexp => qr/^CREATE SCHEMA public;/m,
diff --git a/src/bin/psql/describe.c b/src/bin/psql/describe.c
index 327a69487b..ac1c940204 100644
--- a/src/bin/psql/describe.c
+++ b/src/bin/psql/describe.c
@@ -1709,28 +1709,19 @@ describeOneTableDetails(const char *schemaname,
if (tableinfo.relkind == RELKIND_SEQUENCE)
{
PGresult *result = NULL;
- printQueryOpt myopt = pset.popt;
- char *footers[2] = {NULL, NULL};
if (pset.sversion >= 100000)
{
printfPQExpBuffer(&buf,
- "SELECT pg_catalog.format_type(seqtypid, NULL) AS \"%s\",\n"
- " seqstart AS \"%s\",\n"
- " seqmin AS \"%s\",\n"
- " seqmax AS \"%s\",\n"
- " seqincrement AS \"%s\",\n"
- " CASE WHEN seqcycle THEN '%s' ELSE '%s' END AS \"%s\",\n"
- " seqcache AS \"%s\"\n",
- gettext_noop("Type"),
- gettext_noop("Start"),
- gettext_noop("Minimum"),
- gettext_noop("Maximum"),
- gettext_noop("Increment"),
+ "SELECT pg_catalog.format_type(seqtypid, NULL),\n"
+ " seqstart,\n"
+ " seqmin,\n"
+ " seqmax,\n"
+ " seqincrement,\n"
+ " CASE WHEN seqcycle THEN '%s' ELSE '%s' END,\n"
+ " seqcache\n",
gettext_noop("yes"),
- gettext_noop("no"),
- gettext_noop("Cycles?"),
- gettext_noop("Cache"));
+ gettext_noop("no"));
appendPQExpBuffer(&buf,
"FROM pg_catalog.pg_sequence\n"
"WHERE seqrelid = '%s';",
@@ -1739,22 +1730,15 @@ describeOneTableDetails(const char *schemaname,
else
{
printfPQExpBuffer(&buf,
- "SELECT 'bigint' AS \"%s\",\n"
- " start_value AS \"%s\",\n"
- " min_value AS \"%s\",\n"
- " max_value AS \"%s\",\n"
- " increment_by AS \"%s\",\n"
- " CASE WHEN is_cycled THEN '%s' ELSE '%s' END AS \"%s\",\n"
- " cache_value AS \"%s\"\n",
- gettext_noop("Type"),
- gettext_noop("Start"),
- gettext_noop("Minimum"),
- gettext_noop("Maximum"),
- gettext_noop("Increment"),
+ "SELECT 'bigint',\n"
+ " start_value,\n"
+ " min_value,\n"
+ " max_value,\n"
+ " increment_by,\n"
+ " CASE WHEN is_cycled THEN '%s' ELSE '%s' END,\n"
+ " cache_value\n",
gettext_noop("yes"),
- gettext_noop("no"),
- gettext_noop("Cycles?"),
- gettext_noop("Cache"));
+ gettext_noop("no"));
appendPQExpBuffer(&buf, "FROM %s", fmtId(schemaname));
/* must be separate because fmtId isn't reentrant */
appendPQExpBuffer(&buf, ".%s;", fmtId(relationname));
@@ -1764,6 +1748,57 @@ describeOneTableDetails(const char *schemaname,
if (!res)
goto error_return;
+ numrows = PQntuples(res);
+
+ /* XXX reset to use expanded output for sequences (maybe we should
+ * keep this disabled, just like for tables?) */
+ myopt.expanded = pset.popt.topt.expanded;
+
+ printTableInit(&cont, &myopt, title.data, 7, numrows);
+ printTableInitialized = true;
+
+ if (tableinfo.relpersistence == 'u')
+ printfPQExpBuffer(&title, _("Unlogged sequence \"%s.%s\""),
+ schemaname, relationname);
+ else
+ printfPQExpBuffer(&title, _("Sequence \"%s.%s\""),
+ schemaname, relationname);
+
+ printTableAddHeader(&cont, gettext_noop("Type"), true, 'l');
+ printTableAddHeader(&cont, gettext_noop("Start"), true, 'r');
+ printTableAddHeader(&cont, gettext_noop("Minimum"), true, 'r');
+ printTableAddHeader(&cont, gettext_noop("Maximum"), true, 'r');
+ printTableAddHeader(&cont, gettext_noop("Increment"), true, 'r');
+ printTableAddHeader(&cont, gettext_noop("Cycles?"), true, 'l');
+ printTableAddHeader(&cont, gettext_noop("Cache"), true, 'r');
+
+ /* Generate table cells to be printed */
+ for (i = 0; i < numrows; i++)
+ {
+ /* Type */
+ printTableAddCell(&cont, PQgetvalue(res, i, 0), false, false);
+
+ /* Start */
+ printTableAddCell(&cont, PQgetvalue(res, i, 1), false, false);
+
+ /* Minimum */
+ printTableAddCell(&cont, PQgetvalue(res, i, 2), false, false);
+
+ /* Maximum */
+ printTableAddCell(&cont, PQgetvalue(res, i, 3), false, false);
+
+ /* Increment */
+ printTableAddCell(&cont, PQgetvalue(res, i, 4), false, false);
+
+ /* Cycles? */
+ printTableAddCell(&cont, PQgetvalue(res, i, 5), false, false);
+
+ /* Cache */
+ printTableAddCell(&cont, PQgetvalue(res, i, 6), false, false);
+ }
+
+ /* Footer information about a sequence */
+
/* Get the column that owns this sequence */
printfPQExpBuffer(&buf, "SELECT pg_catalog.quote_ident(nspname) || '.' ||"
"\n pg_catalog.quote_ident(relname) || '.' ||"
@@ -1795,32 +1830,63 @@ describeOneTableDetails(const char *schemaname,
switch (PQgetvalue(result, 0, 1)[0])
{
case 'a':
- footers[0] = psprintf(_("Owned by: %s"),
- PQgetvalue(result, 0, 0));
+ printTableAddFooter(&cont,
+ psprintf(_("Owned by: %s"),
+ PQgetvalue(result, 0, 0)));
break;
case 'i':
- footers[0] = psprintf(_("Sequence for identity column: %s"),
- PQgetvalue(result, 0, 0));
+ printTableAddFooter(&cont,
+ psprintf(_("Sequence for identity column: %s"),
+ PQgetvalue(result, 0, 0)));
break;
}
}
PQclear(result);
- if (tableinfo.relpersistence == 'u')
- printfPQExpBuffer(&title, _("Unlogged sequence \"%s.%s\""),
- schemaname, relationname);
- else
- printfPQExpBuffer(&title, _("Sequence \"%s.%s\""),
- schemaname, relationname);
+ /* print any publications */
+ if (pset.sversion >= 160000)
+ {
+ int tuples = 0;
- myopt.footers = footers;
- myopt.topt.default_footer = false;
- myopt.title = title.data;
- myopt.translate_header = true;
+ printfPQExpBuffer(&buf,
+ "SELECT pubname\n"
+ "FROM pg_catalog.pg_publication p\n"
+ " JOIN pg_catalog.pg_publication_namespace pn ON p.oid = pn.pnpubid\n"
+ " JOIN pg_catalog.pg_class pc ON pc.relnamespace = pn.pnnspid\n"
+ "WHERE pc.oid ='%s' and pn.pntype = 's' and pg_catalog.pg_relation_is_publishable('%s')\n"
+ "UNION\n"
+ "SELECT pubname\n"
+ "FROM pg_catalog.pg_publication p\n"
+ " JOIN pg_catalog.pg_publication_rel pr ON p.oid = pr.prpubid\n"
+ "WHERE pr.prrelid = '%s'\n"
+ "UNION\n"
+ "SELECT pubname\n"
+ "FROM pg_catalog.pg_publication p\n"
+ "WHERE p.puballsequences AND pg_catalog.pg_relation_is_publishable('%s')\n"
+ "ORDER BY 1;",
+ oid, oid, oid, oid);
- printQuery(res, &myopt, pset.queryFout, false, pset.logfile);
+ result = PSQLexec(buf.data);
+ if (!result)
+ goto error_return;
+ else
+ tuples = PQntuples(result);
- free(footers[0]);
+ if (tuples > 0)
+ printTableAddFooter(&cont, _("Publications:"));
+
+ /* Might be an empty set - that's ok */
+ for (i = 0; i < tuples; i++)
+ {
+ printfPQExpBuffer(&buf, " \"%s\"",
+ PQgetvalue(result, i, 0));
+
+ printTableAddFooter(&cont, buf.data);
+ }
+ PQclear(result);
+ }
+
+ printTable(&cont, pset.queryFout, false, pset.logfile);
retval = true;
goto error_return; /* not an error, just return early */
@@ -2047,6 +2113,11 @@ describeOneTableDetails(const char *schemaname,
for (i = 0; i < cols; i++)
printTableAddHeader(&cont, headers[i], true, 'l');
+ res = PSQLexec(buf.data);
+ if (!res)
+ goto error_return;
+ numrows = PQntuples(res);
+
/* Generate table cells to be printed */
for (i = 0; i < numrows; i++)
{
@@ -2972,7 +3043,7 @@ describeOneTableDetails(const char *schemaname,
"FROM pg_catalog.pg_publication p\n"
" JOIN pg_catalog.pg_publication_namespace pn ON p.oid = pn.pnpubid\n"
" JOIN pg_catalog.pg_class pc ON pc.relnamespace = pn.pnnspid\n"
- "WHERE pc.oid ='%s' and pg_catalog.pg_relation_is_publishable('%s')\n"
+ "WHERE pc.oid ='%s' and pn.pntype = 't' and pg_catalog.pg_relation_is_publishable('%s')\n"
"UNION\n"
"SELECT pubname\n"
" , pg_get_expr(pr.prqual, c.oid)\n"
@@ -5003,7 +5074,7 @@ listSchemas(const char *pattern, bool verbose, bool showSystem)
int i;
printfPQExpBuffer(&buf,
- "SELECT pubname \n"
+ "SELECT pubname, (CASE WHEN pntype = 't' THEN 'tables' ELSE 'sequences' END) AS pubtype\n"
"FROM pg_catalog.pg_publication p\n"
" JOIN pg_catalog.pg_publication_namespace pn ON p.oid = pn.pnpubid\n"
" JOIN pg_catalog.pg_namespace n ON n.oid = pn.pnnspid \n"
@@ -5029,8 +5100,9 @@ listSchemas(const char *pattern, bool verbose, bool showSystem)
/* Might be an empty set - that's ok */
for (i = 0; i < pub_schema_tuples; i++)
{
- printfPQExpBuffer(&buf, " \"%s\"",
- PQgetvalue(result, i, 0));
+ printfPQExpBuffer(&buf, " \"%s\" (%s)",
+ PQgetvalue(result, i, 0),
+ PQgetvalue(result, i, 1));
footers[i + 1] = pg_strdup(buf.data);
}
@@ -6159,7 +6231,7 @@ listPublications(const char *pattern)
PQExpBufferData buf;
PGresult *res;
printQueryOpt myopt = pset.popt;
- static const bool translate_columns[] = {false, false, false, false, false, false, false, false};
+ static const bool translate_columns[] = {false, false, false, false, false, false, false, false, false, false};
if (pset.sversion < 100000)
{
@@ -6173,23 +6245,45 @@ listPublications(const char *pattern)
initPQExpBuffer(&buf);
- printfPQExpBuffer(&buf,
- "SELECT pubname AS \"%s\",\n"
- " pg_catalog.pg_get_userbyid(pubowner) AS \"%s\",\n"
- " puballtables AS \"%s\",\n"
- " pubinsert AS \"%s\",\n"
- " pubupdate AS \"%s\",\n"
- " pubdelete AS \"%s\"",
- gettext_noop("Name"),
- gettext_noop("Owner"),
- gettext_noop("All tables"),
- gettext_noop("Inserts"),
- gettext_noop("Updates"),
- gettext_noop("Deletes"));
+ if (pset.sversion >= 160000)
+ printfPQExpBuffer(&buf,
+ "SELECT pubname AS \"%s\",\n"
+ " pg_catalog.pg_get_userbyid(pubowner) AS \"%s\",\n"
+ " puballtables AS \"%s\",\n"
+ " puballsequences AS \"%s\",\n"
+ " pubinsert AS \"%s\",\n"
+ " pubupdate AS \"%s\",\n"
+ " pubdelete AS \"%s\"",
+ gettext_noop("Name"),
+ gettext_noop("Owner"),
+ gettext_noop("All tables"),
+ gettext_noop("All sequences"),
+ gettext_noop("Inserts"),
+ gettext_noop("Updates"),
+ gettext_noop("Deletes"));
+ else
+ printfPQExpBuffer(&buf,
+ "SELECT pubname AS \"%s\",\n"
+ " pg_catalog.pg_get_userbyid(pubowner) AS \"%s\",\n"
+ " puballtables AS \"%s\",\n"
+ " pubinsert AS \"%s\",\n"
+ " pubupdate AS \"%s\",\n"
+ " pubdelete AS \"%s\"",
+ gettext_noop("Name"),
+ gettext_noop("Owner"),
+ gettext_noop("All tables"),
+ gettext_noop("Inserts"),
+ gettext_noop("Updates"),
+ gettext_noop("Deletes"));
+
if (pset.sversion >= 110000)
appendPQExpBuffer(&buf,
",\n pubtruncate AS \"%s\"",
gettext_noop("Truncates"));
+ if (pset.sversion >= 160000)
+ appendPQExpBuffer(&buf,
+ ",\n pubsequence AS \"%s\"",
+ gettext_noop("Sequences"));
if (pset.sversion >= 130000)
appendPQExpBuffer(&buf,
",\n pubviaroot AS \"%s\"",
@@ -6284,6 +6378,7 @@ describePublications(const char *pattern)
PGresult *res;
bool has_pubtruncate;
bool has_pubviaroot;
+ bool has_pubsequence;
PQExpBufferData title;
printTableContent cont;
@@ -6300,6 +6395,7 @@ describePublications(const char *pattern)
has_pubtruncate = (pset.sversion >= 110000);
has_pubviaroot = (pset.sversion >= 130000);
+ has_pubsequence = (pset.sversion >= 160000);
initPQExpBuffer(&buf);
@@ -6307,12 +6403,17 @@ describePublications(const char *pattern)
"SELECT oid, pubname,\n"
" pg_catalog.pg_get_userbyid(pubowner) AS owner,\n"
" puballtables, pubinsert, pubupdate, pubdelete");
+
if (has_pubtruncate)
appendPQExpBufferStr(&buf,
", pubtruncate");
if (has_pubviaroot)
appendPQExpBufferStr(&buf,
", pubviaroot");
+ if (has_pubsequence)
+ appendPQExpBufferStr(&buf,
+ ", puballsequences, pubsequence");
+
appendPQExpBufferStr(&buf,
"\nFROM pg_catalog.pg_publication\n");
@@ -6358,6 +6459,7 @@ describePublications(const char *pattern)
char *pubid = PQgetvalue(res, i, 0);
char *pubname = PQgetvalue(res, i, 1);
bool puballtables = strcmp(PQgetvalue(res, i, 3), "t") == 0;
+ bool puballsequences = strcmp(PQgetvalue(res, i, 9), "t") == 0;
printTableOpt myopt = pset.popt.topt;
if (has_pubtruncate)
@@ -6365,29 +6467,43 @@ describePublications(const char *pattern)
if (has_pubviaroot)
ncols++;
+ /* sequences have two extra columns (puballsequences, pubsequences) */
+ if (has_pubsequence)
+ ncols += 2;
+
initPQExpBuffer(&title);
printfPQExpBuffer(&title, _("Publication %s"), pubname);
printTableInit(&cont, &myopt, title.data, ncols, nrows);
printTableAddHeader(&cont, gettext_noop("Owner"), true, align);
printTableAddHeader(&cont, gettext_noop("All tables"), true, align);
+ if (has_pubsequence)
+ printTableAddHeader(&cont, gettext_noop("All sequences"), true, align);
printTableAddHeader(&cont, gettext_noop("Inserts"), true, align);
printTableAddHeader(&cont, gettext_noop("Updates"), true, align);
printTableAddHeader(&cont, gettext_noop("Deletes"), true, align);
if (has_pubtruncate)
printTableAddHeader(&cont, gettext_noop("Truncates"), true, align);
+ if (has_pubsequence)
+ printTableAddHeader(&cont, gettext_noop("Sequences"), true, align);
if (has_pubviaroot)
printTableAddHeader(&cont, gettext_noop("Via root"), true, align);
- printTableAddCell(&cont, PQgetvalue(res, i, 2), false, false);
- printTableAddCell(&cont, PQgetvalue(res, i, 3), false, false);
- printTableAddCell(&cont, PQgetvalue(res, i, 4), false, false);
- printTableAddCell(&cont, PQgetvalue(res, i, 5), false, false);
- printTableAddCell(&cont, PQgetvalue(res, i, 6), false, false);
+ printTableAddCell(&cont, PQgetvalue(res, i, 2), false, false); /* owner */
+ printTableAddCell(&cont, PQgetvalue(res, i, 3), false, false); /* all tables */
+
+ if (has_pubsequence)
+ printTableAddCell(&cont, PQgetvalue(res, i, 9), false, false); /* all sequences */
+
+ printTableAddCell(&cont, PQgetvalue(res, i, 4), false, false); /* insert */
+ printTableAddCell(&cont, PQgetvalue(res, i, 5), false, false); /* update */
+ printTableAddCell(&cont, PQgetvalue(res, i, 6), false, false); /* delete */
if (has_pubtruncate)
- printTableAddCell(&cont, PQgetvalue(res, i, 7), false, false);
+ printTableAddCell(&cont, PQgetvalue(res, i, 7), false, false); /* truncate */
+ if (has_pubsequence)
+ printTableAddCell(&cont, PQgetvalue(res, i, 10), false, false); /* sequence */
if (has_pubviaroot)
- printTableAddCell(&cont, PQgetvalue(res, i, 8), false, false);
+ printTableAddCell(&cont, PQgetvalue(res, i, 8), false, false); /* via root */
if (!puballtables)
{
@@ -6418,6 +6534,7 @@ describePublications(const char *pattern)
"WHERE c.relnamespace = n.oid\n"
" AND c.oid = pr.prrelid\n"
" AND pr.prpubid = '%s'\n"
+ " AND c.relkind != 'S'\n" /* exclude sequences */
"ORDER BY 1,2", pubid);
if (!addFooterToPublicationDesc(&buf, _("Tables:"), false, &cont))
goto error_return;
@@ -6429,7 +6546,7 @@ describePublications(const char *pattern)
"SELECT n.nspname\n"
"FROM pg_catalog.pg_namespace n\n"
" JOIN pg_catalog.pg_publication_namespace pn ON n.oid = pn.pnnspid\n"
- "WHERE pn.pnpubid = '%s'\n"
+ "WHERE pn.pnpubid = '%s' AND pn.pntype = 't'\n"
"ORDER BY 1", pubid);
if (!addFooterToPublicationDesc(&buf, _("Tables from schemas:"),
true, &cont))
@@ -6437,6 +6554,37 @@ describePublications(const char *pattern)
}
}
+ if (!puballsequences)
+ {
+ /* Get the sequences for the specified publication */
+ printfPQExpBuffer(&buf,
+ "SELECT n.nspname, c.relname, NULL, NULL\n"
+ "FROM pg_catalog.pg_class c,\n"
+ " pg_catalog.pg_namespace n,\n"
+ " pg_catalog.pg_publication_rel pr\n"
+ "WHERE c.relnamespace = n.oid\n"
+ " AND c.oid = pr.prrelid\n"
+ " AND pr.prpubid = '%s'\n"
+ " AND c.relkind = 'S'\n" /* only sequences */
+ "ORDER BY 1,2", pubid);
+ if (!addFooterToPublicationDesc(&buf, "Sequences:", false, &cont))
+ goto error_return;
+
+ if (pset.sversion >= 160000)
+ {
+ /* Get the schemas for the specified publication */
+ printfPQExpBuffer(&buf,
+ "SELECT n.nspname\n"
+ "FROM pg_catalog.pg_namespace n\n"
+ " JOIN pg_catalog.pg_publication_namespace pn ON n.oid = pn.pnnspid\n"
+ "WHERE pn.pnpubid = '%s' AND pn.pntype = 's'\n"
+ "ORDER BY 1", pubid);
+ if (!addFooterToPublicationDesc(&buf, "Sequences from schemas:",
+ true, &cont))
+ goto error_return;
+ }
+ }
+
printTable(&cont, pset.queryFout, false, pset.logfile);
printTableCleanup(&cont);
diff --git a/src/bin/psql/tab-complete.c b/src/bin/psql/tab-complete.c
index 62a39779b9..5ad1e517fa 100644
--- a/src/bin/psql/tab-complete.c
+++ b/src/bin/psql/tab-complete.c
@@ -1820,11 +1820,15 @@ psql_completion(const char *text, int start, int end)
COMPLETE_WITH("ADD", "DROP", "OWNER TO", "RENAME TO", "SET");
/* ALTER PUBLICATION <name> ADD */
else if (Matches("ALTER", "PUBLICATION", MatchAny, "ADD"))
- COMPLETE_WITH("ALL TABLES IN SCHEMA", "TABLE");
+ COMPLETE_WITH("ALL TABLES IN SCHEMA", "ALL SEQUENCES IN SCHEMA", "TABLE", "SEQUENCE");
else if (Matches("ALTER", "PUBLICATION", MatchAny, "ADD|SET", "TABLE") ||
(HeadMatches("ALTER", "PUBLICATION", MatchAny, "ADD|SET", "TABLE") &&
ends_with(prev_wd, ',')))
COMPLETE_WITH_SCHEMA_QUERY(Query_for_list_of_tables);
+ else if (Matches("ALTER", "PUBLICATION", MatchAny, "ADD|SET", "SEQUENCE") ||
+ (HeadMatches("ALTER", "PUBLICATION", MatchAny, "ADD|SET", "SEQUENCE") &&
+ ends_with(prev_wd, ',')))
+ COMPLETE_WITH_SCHEMA_QUERY(Query_for_list_of_sequences);
/*
* "ALTER PUBLICATION <name> SET TABLE <name> WHERE (" - complete with
@@ -1844,11 +1848,11 @@ psql_completion(const char *text, int start, int end)
COMPLETE_WITH(",");
/* ALTER PUBLICATION <name> DROP */
else if (Matches("ALTER", "PUBLICATION", MatchAny, "DROP"))
- COMPLETE_WITH("ALL TABLES IN SCHEMA", "TABLE");
+ COMPLETE_WITH("ALL TABLES IN SCHEMA", "ALL SEQUENCES IN SCHEMA", "TABLE", "SEQUENCE");
/* ALTER PUBLICATION <name> SET */
else if (Matches("ALTER", "PUBLICATION", MatchAny, "SET"))
- COMPLETE_WITH("(", "ALL TABLES IN SCHEMA", "TABLE");
- else if (Matches("ALTER", "PUBLICATION", MatchAny, "ADD|DROP|SET", "ALL", "TABLES", "IN", "SCHEMA"))
+ COMPLETE_WITH("(", "ALL TABLES IN SCHEMA", "ALL SEQUENCES IN SCHEMA", "TABLE", "SEQUENCE");
+ else if (Matches("ALTER", "PUBLICATION", MatchAny, "ADD|DROP|SET", "ALL", "TABLES|SEQUENCES", "IN", "SCHEMA"))
COMPLETE_WITH_QUERY_PLUS(Query_for_list_of_schemas
" AND nspname NOT LIKE E'pg\\\\_%%'",
"CURRENT_SCHEMA");
@@ -2981,21 +2985,27 @@ psql_completion(const char *text, int start, int end)
/* CREATE PUBLICATION */
else if (Matches("CREATE", "PUBLICATION", MatchAny))
- COMPLETE_WITH("FOR TABLE", "FOR ALL TABLES", "FOR ALL TABLES IN SCHEMA", "WITH (");
+ COMPLETE_WITH("FOR TABLE", "FOR ALL TABLES", "FOR ALL TABLES IN SCHEMA",
+ "FOR SEQUENCE", "FOR ALL SEQUENCES", "FOR ALL SEQUENCES IN SCHEMA",
+ "WITH (");
else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR"))
- COMPLETE_WITH("TABLE", "ALL TABLES", "ALL TABLES IN SCHEMA");
+ COMPLETE_WITH("TABLE", "ALL TABLES", "ALL TABLES IN SCHEMA",
+ "SEQUENCE", "ALL SEQUENCES", "ALL SEQUENCES IN SCHEMA");
else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "ALL"))
- COMPLETE_WITH("TABLES", "TABLES IN SCHEMA");
- else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "ALL", "TABLES"))
+ COMPLETE_WITH("TABLES", "TABLES IN SCHEMA", "SEQUENCES", "SEQUENCES IN SCHEMA");
+ else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "ALL", "TABLES|SEQUENCES"))
COMPLETE_WITH("IN SCHEMA", "WITH (");
- else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "TABLE", MatchAny) && !ends_with(prev_wd, ','))
+ else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "TABLE|SEQUENCE", MatchAny) && !ends_with(prev_wd, ','))
COMPLETE_WITH("WHERE (", "WITH (");
/* Complete "CREATE PUBLICATION <name> FOR TABLE" with "<table>, ..." */
else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "TABLE"))
COMPLETE_WITH_SCHEMA_QUERY(Query_for_list_of_tables);
+ /* Complete "CREATE PUBLICATION <name> FOR SEQUENCE" with "<sequence>, ..." */
+ else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "SEQUENCE"))
+ COMPLETE_WITH_SCHEMA_QUERY(Query_for_list_of_sequences);
/*
- * "CREATE PUBLICATION <name> FOR TABLE <name> WHERE (" - complete with
+ * "CREATE PUBLICATION <name> FOR TABLE|SEQUENCE <name> WHERE (" - complete with
* table attributes
*/
else if (HeadMatches("CREATE", "PUBLICATION", MatchAny) && TailMatches("WHERE"))
@@ -3006,14 +3016,14 @@ psql_completion(const char *text, int start, int end)
COMPLETE_WITH(" WITH (");
/*
- * Complete "CREATE PUBLICATION <name> FOR ALL TABLES IN SCHEMA <schema>,
+ * Complete "CREATE PUBLICATION <name> FOR ALL TABLES|SEQUENCES IN SCHEMA <schema>,
* ..."
*/
- else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "ALL", "TABLES", "IN", "SCHEMA"))
+ else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "ALL", "TABLES|SEQUENCES", "IN", "SCHEMA"))
COMPLETE_WITH_QUERY_PLUS(Query_for_list_of_schemas
" AND nspname NOT LIKE E'pg\\\\_%%'",
"CURRENT_SCHEMA");
- else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "ALL", "TABLES", "IN", "SCHEMA", MatchAny) && (!ends_with(prev_wd, ',')))
+ else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "ALL", "TABLES|SEQUENCES", "IN", "SCHEMA", MatchAny) && (!ends_with(prev_wd, ',')))
COMPLETE_WITH("WITH (");
/* Complete "CREATE PUBLICATION <name> [...] WITH" */
else if (HeadMatches("CREATE", "PUBLICATION") && TailMatches("WITH", "("))
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index be47583122..9f6e726a41 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -11687,6 +11687,11 @@
proallargtypes => '{text,oid,int2vector,pg_node_tree}',
proargmodes => '{i,o,o,o}', proargnames => '{pubname,relid,attrs,qual}',
prosrc => 'pg_get_publication_tables' },
+{ oid => '8000', descr => 'get OIDs of sequences in a publication',
+ proname => 'pg_get_publication_sequences', prorows => '1000', proretset => 't',
+ provolatile => 's', prorettype => 'oid', proargtypes => 'text',
+ proallargtypes => '{text,oid}', proargmodes => '{i,o}',
+ proargnames => '{pubname,relid}', prosrc => 'pg_get_publication_sequences' },
{ oid => '6121',
descr => 'returns whether a relation can be part of a publication',
proname => 'pg_relation_is_publishable', provolatile => 's',
diff --git a/src/include/catalog/pg_publication.h b/src/include/catalog/pg_publication.h
index c298327f5e..e3526367ed 100644
--- a/src/include/catalog/pg_publication.h
+++ b/src/include/catalog/pg_publication.h
@@ -40,6 +40,12 @@ CATALOG(pg_publication,6104,PublicationRelationId)
*/
bool puballtables;
+ /*
+ * indicates that this is special publication which should encompass all
+ * sequences in the database (except for the unlogged and temp ones)
+ */
+ bool puballsequences;
+
/* true if inserts are published */
bool pubinsert;
@@ -52,6 +58,9 @@ CATALOG(pg_publication,6104,PublicationRelationId)
/* true if truncates are published */
bool pubtruncate;
+ /* true if sequences are published */
+ bool pubsequence;
+
/* true if partition changes are published using root schema */
bool pubviaroot;
} FormData_pg_publication;
@@ -72,6 +81,7 @@ typedef struct PublicationActions
bool pubupdate;
bool pubdelete;
bool pubtruncate;
+ bool pubsequence;
} PublicationActions;
typedef struct PublicationDesc
@@ -99,6 +109,7 @@ typedef struct Publication
Oid oid;
char *name;
bool alltables;
+ bool allsequences;
bool pubviaroot;
PublicationActions pubactions;
} Publication;
@@ -130,14 +141,16 @@ typedef enum PublicationPartOpt
PUBLICATION_PART_ALL,
} PublicationPartOpt;
-extern List *GetPublicationRelations(Oid pubid, PublicationPartOpt pub_partopt);
+extern List *GetPublicationRelations(Oid pubid, char objectType,
+ PublicationPartOpt pub_partopt);
extern List *GetAllTablesPublications(void);
extern List *GetAllTablesPublicationRelations(bool pubviaroot);
-extern List *GetPublicationSchemas(Oid pubid);
-extern List *GetSchemaPublications(Oid schemaid);
-extern List *GetSchemaPublicationRelations(Oid schemaid,
+extern void GetActionsInPublication(Oid pubid, PublicationActions *actions);
+extern List *GetPublicationSchemas(Oid pubid, char objectType);
+extern List *GetSchemaPublications(Oid schemaid, char objectType);
+extern List *GetSchemaPublicationRelations(Oid schemaid, char objectType,
PublicationPartOpt pub_partopt);
-extern List *GetAllSchemaPublicationRelations(Oid puboid,
+extern List *GetAllSchemaPublicationRelations(Oid puboid, char objectType,
PublicationPartOpt pub_partopt);
extern List *GetPubPartitionOptionRelations(List *result,
PublicationPartOpt pub_partopt,
@@ -145,11 +158,15 @@ extern List *GetPubPartitionOptionRelations(List *result,
extern Oid GetTopMostAncestorInPublication(Oid puboid, List *ancestors,
int *ancestor_level);
+extern List *GetAllSequencesPublications(void);
+extern List *GetAllSequencesPublicationRelations(void);
+
extern bool is_publishable_relation(Relation rel);
extern bool is_schema_publication(Oid pubid);
extern ObjectAddress publication_add_relation(Oid pubid, PublicationRelInfo *pri,
bool if_not_exists);
extern ObjectAddress publication_add_schema(Oid pubid, Oid schemaid,
+ char objectType,
bool if_not_exists);
extern Bitmapset *pub_collist_to_bitmapset(Bitmapset *columns, Datum pubcols,
diff --git a/src/include/catalog/pg_publication_namespace.h b/src/include/catalog/pg_publication_namespace.h
index d166600b26..914dde8b61 100644
--- a/src/include/catalog/pg_publication_namespace.h
+++ b/src/include/catalog/pg_publication_namespace.h
@@ -32,6 +32,7 @@ CATALOG(pg_publication_namespace,6237,PublicationNamespaceRelationId)
Oid oid; /* oid */
Oid pnpubid BKI_LOOKUP(pg_publication); /* Oid of the publication */
Oid pnnspid BKI_LOOKUP(pg_namespace); /* Oid of the schema */
+ char pntype; /* object type to include */
} FormData_pg_publication_namespace;
/* ----------------
@@ -42,6 +43,13 @@ CATALOG(pg_publication_namespace,6237,PublicationNamespaceRelationId)
typedef FormData_pg_publication_namespace *Form_pg_publication_namespace;
DECLARE_UNIQUE_INDEX_PKEY(pg_publication_namespace_oid_index, 6238, PublicationNamespaceObjectIndexId, on pg_publication_namespace using btree(oid oid_ops));
-DECLARE_UNIQUE_INDEX(pg_publication_namespace_pnnspid_pnpubid_index, 6239, PublicationNamespacePnnspidPnpubidIndexId, on pg_publication_namespace using btree(pnnspid oid_ops, pnpubid oid_ops));
+DECLARE_UNIQUE_INDEX(pg_publication_namespace_pnnspid_pnpubid_pntype_index, 8903, PublicationNamespacePnnspidPnpubidPntypeIndexId, on pg_publication_namespace using btree(pnnspid oid_ops, pnpubid oid_ops, pntype char_ops));
+
+/* object type to include from a schema, maps to relkind */
+#define PUB_OBJTYPE_TABLE 't' /* table (regular or partitioned) */
+#define PUB_OBJTYPE_SEQUENCE 's' /* sequence object */
+#define PUB_OBJTYPE_UNSUPPORTED 'u' /* used for non-replicated types */
+
+extern char pub_get_object_type_for_relkind(char relkind);
#endif /* PG_PUBLICATION_NAMESPACE_H */
diff --git a/src/include/commands/sequence.h b/src/include/commands/sequence.h
index d38c0e2388..d245bfa38d 100644
--- a/src/include/commands/sequence.h
+++ b/src/include/commands/sequence.h
@@ -60,8 +60,11 @@ extern ObjectAddress AlterSequence(ParseState *pstate, AlterSeqStmt *stmt);
extern void SequenceChangePersistence(Oid relid, char newrelpersistence);
extern void DeleteSequenceTuple(Oid relid);
extern void ResetSequence(Oid seq_relid);
+extern void SetSequence(Oid seq_relid, int64 last_value, int64 log_cnt, bool is_called);
extern void ResetSequenceCaches(void);
+extern void AtEOXact_Sequences(bool isCommit);
+
extern void seq_redo(XLogReaderState *rptr);
extern void seq_desc(StringInfo buf, XLogReaderState *rptr);
extern const char *seq_identify(uint8 info);
diff --git a/src/include/nodes/parsenodes.h b/src/include/nodes/parsenodes.h
index b376031856..cd5db82f19 100644
--- a/src/include/nodes/parsenodes.h
+++ b/src/include/nodes/parsenodes.h
@@ -4028,6 +4028,10 @@ typedef enum PublicationObjSpecType
PUBLICATIONOBJ_TABLES_IN_SCHEMA, /* All tables in schema */
PUBLICATIONOBJ_TABLES_IN_CUR_SCHEMA, /* All tables in first element of
* search_path */
+ PUBLICATIONOBJ_SEQUENCE, /* Sequence type */
+ PUBLICATIONOBJ_SEQUENCES_IN_SCHEMA, /* Sequences in schema type */
+ PUBLICATIONOBJ_SEQUENCES_IN_CUR_SCHEMA, /* Get the first element of
+ * search_path */
PUBLICATIONOBJ_CONTINUATION /* Continuation of previous type */
} PublicationObjSpecType;
@@ -4046,7 +4050,7 @@ typedef struct CreatePublicationStmt
char *pubname; /* Name of the publication */
List *options; /* List of DefElem nodes */
List *pubobjects; /* Optional list of publication objects */
- bool for_all_tables; /* Special publication for all tables in db */
+ List *for_all_objects; /* Special publication for all objects in db */
} CreatePublicationStmt;
typedef enum AlterPublicationAction
@@ -4069,7 +4073,7 @@ typedef struct AlterPublicationStmt
* objects.
*/
List *pubobjects; /* Optional list of publication objects */
- bool for_all_tables; /* Special publication for all tables in db */
+ List *for_all_objects; /* Special publication for all objects in db */
AlterPublicationAction action; /* What action to perform with the given
* objects */
} AlterPublicationStmt;
diff --git a/src/include/replication/decode.h b/src/include/replication/decode.h
index 741bf65cf7..81df40d497 100644
--- a/src/include/replication/decode.h
+++ b/src/include/replication/decode.h
@@ -27,6 +27,7 @@ extern void heap2_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
extern void xact_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
extern void standby_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
extern void logicalmsg_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
+extern void sequence_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
extern void LogicalDecodingProcessRecord(LogicalDecodingContext *ctx,
XLogReaderState *record);
diff --git a/src/include/replication/logicalproto.h b/src/include/replication/logicalproto.h
index a771ab8ff3..a466db394b 100644
--- a/src/include/replication/logicalproto.h
+++ b/src/include/replication/logicalproto.h
@@ -61,6 +61,7 @@ typedef enum LogicalRepMsgType
LOGICAL_REP_MSG_RELATION = 'R',
LOGICAL_REP_MSG_TYPE = 'Y',
LOGICAL_REP_MSG_MESSAGE = 'M',
+ LOGICAL_REP_MSG_SEQUENCE = 'Q',
LOGICAL_REP_MSG_BEGIN_PREPARE = 'b',
LOGICAL_REP_MSG_PREPARE = 'P',
LOGICAL_REP_MSG_COMMIT_PREPARED = 'K',
@@ -118,6 +119,17 @@ typedef struct LogicalRepTyp
char *typname; /* name of the remote type */
} LogicalRepTyp;
+/* Sequence info */
+typedef struct LogicalRepSequence
+{
+ Oid remoteid; /* unique id of the remote sequence */
+ char *nspname; /* schema name of remote sequence */
+ char *seqname; /* name of the remote sequence */
+ int64 last_value;
+ int64 log_cnt;
+ bool is_called;
+} LogicalRepSequence;
+
/* Transaction info */
typedef struct LogicalRepBeginData
{
@@ -230,6 +242,11 @@ extern List *logicalrep_read_truncate(StringInfo in,
bool *cascade, bool *restart_seqs);
extern void logicalrep_write_message(StringInfo out, TransactionId xid, XLogRecPtr lsn,
bool transactional, const char *prefix, Size sz, const char *message);
+extern void logicalrep_write_sequence(StringInfo out, Relation rel,
+ TransactionId xid, XLogRecPtr lsn,
+ int64 last_value, int64 log_cnt,
+ bool is_called);
+extern void logicalrep_read_sequence(StringInfo in, LogicalRepSequence *seqdata);
extern void logicalrep_write_rel(StringInfo out, TransactionId xid,
Relation rel, Bitmapset *columns);
extern LogicalRepRelation *logicalrep_read_rel(StringInfo in);
diff --git a/src/include/replication/message.h b/src/include/replication/message.h
index 0b396c5669..0bfaeb9336 100644
--- a/src/include/replication/message.h
+++ b/src/include/replication/message.h
@@ -27,13 +27,27 @@ typedef struct xl_logical_message
char message[FLEXIBLE_ARRAY_MEMBER];
} xl_logical_message;
+/*
+ * Generic logical decoding sequence wal record.
+ */
+typedef struct xl_logical_sequence
+{
+ RelFileLocator locator;
+ Oid reloid;
+ int64 last; /* last value emitted for sequence */
+ int64 log_cnt; /* last value cached for sequence */
+ bool is_called;
+} xl_logical_sequence;
+
#define SizeOfLogicalMessage (offsetof(xl_logical_message, message))
+#define SizeOfLogicalSequence (sizeof(xl_logical_sequence))
extern XLogRecPtr LogLogicalMessage(const char *prefix, const char *message,
size_t size, bool transactional);
/* RMGR API */
#define XLOG_LOGICAL_MESSAGE 0x00
+#define XLOG_LOGICAL_SEQUENCE 0x10
extern void logicalmsg_redo(XLogReaderState *record);
extern void logicalmsg_desc(StringInfo buf, XLogReaderState *record);
extern const char *logicalmsg_identify(uint8 info);
diff --git a/src/include/replication/output_plugin.h b/src/include/replication/output_plugin.h
index b7d28d7045..959afe3be9 100644
--- a/src/include/replication/output_plugin.h
+++ b/src/include/replication/output_plugin.h
@@ -90,6 +90,17 @@ typedef void (*LogicalDecodeMessageCB) (struct LogicalDecodingContext *ctx,
Size message_size,
const char *message);
+/*
+ * Called for the generic logical decoding sequences.
+ */
+typedef void (*LogicalDecodeSequenceCB) (struct LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ int64 last_value,
+ int64 log_cnt,
+ bool is_called);
+
/*
* Filter changes by origin.
*/
@@ -201,6 +212,18 @@ typedef void (*LogicalDecodeStreamMessageCB) (struct LogicalDecodingContext *ctx
Size message_size,
const char *message);
+/*
+ * Called for the streaming generic logical decoding sequences from in-progress
+ * transactions.
+ */
+typedef void (*LogicalDecodeStreamSequenceCB) (struct LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ int64 last_value,
+ int64 log_cnt,
+ bool is_called);
+
/*
* Callback for streaming truncates from in-progress transactions.
*/
@@ -221,6 +244,7 @@ typedef struct OutputPluginCallbacks
LogicalDecodeTruncateCB truncate_cb;
LogicalDecodeCommitCB commit_cb;
LogicalDecodeMessageCB message_cb;
+ LogicalDecodeSequenceCB sequence_cb;
LogicalDecodeFilterByOriginCB filter_by_origin_cb;
LogicalDecodeShutdownCB shutdown_cb;
@@ -239,6 +263,7 @@ typedef struct OutputPluginCallbacks
LogicalDecodeStreamCommitCB stream_commit_cb;
LogicalDecodeStreamChangeCB stream_change_cb;
LogicalDecodeStreamMessageCB stream_message_cb;
+ LogicalDecodeStreamSequenceCB stream_sequence_cb;
LogicalDecodeStreamTruncateCB stream_truncate_cb;
} OutputPluginCallbacks;
diff --git a/src/include/replication/pgoutput.h b/src/include/replication/pgoutput.h
index 02027550e2..45df2930b7 100644
--- a/src/include/replication/pgoutput.h
+++ b/src/include/replication/pgoutput.h
@@ -30,6 +30,7 @@ typedef struct PGOutputData
bool messages;
bool two_phase;
char *origin;
+ bool sequences;
} PGOutputData;
#endif /* PGOUTPUT_H */
diff --git a/src/include/replication/reorderbuffer.h b/src/include/replication/reorderbuffer.h
index 8695901ba7..f8418d1a7e 100644
--- a/src/include/replication/reorderbuffer.h
+++ b/src/include/replication/reorderbuffer.h
@@ -64,7 +64,8 @@ typedef enum ReorderBufferChangeType
REORDER_BUFFER_CHANGE_INTERNAL_SPEC_INSERT,
REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM,
REORDER_BUFFER_CHANGE_INTERNAL_SPEC_ABORT,
- REORDER_BUFFER_CHANGE_TRUNCATE
+ REORDER_BUFFER_CHANGE_TRUNCATE,
+ REORDER_BUFFER_CHANGE_SEQUENCE
} ReorderBufferChangeType;
/* forward declaration */
@@ -158,6 +159,16 @@ typedef struct ReorderBufferChange
uint32 ninvalidations; /* Number of messages */
SharedInvalidationMessage *invalidations; /* invalidation message */
} inval;
+
+ /* Context data for Sequence changes */
+ struct
+ {
+ RelFileLocator locator;
+ Oid reloid;
+ int64 last;
+ int64 log_cnt;
+ bool is_called;
+ } sequence;
} data;
/*
@@ -435,6 +446,14 @@ typedef void (*ReorderBufferMessageCB) (ReorderBuffer *rb,
const char *prefix, Size sz,
const char *message);
+/* sequence callback signature */
+typedef void (*ReorderBufferSequenceCB) (ReorderBuffer *rb,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ int64 last_value, int64 log_cnt,
+ bool is_called);
+
/* begin prepare callback signature */
typedef void (*ReorderBufferBeginPrepareCB) (ReorderBuffer *rb,
ReorderBufferTXN *txn);
@@ -501,6 +520,14 @@ typedef void (*ReorderBufferStreamMessageCB) (
const char *prefix, Size sz,
const char *message);
+/* stream sequence callback signature */
+typedef void (*ReorderBufferStreamSequenceCB) (ReorderBuffer *rb,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ int64 last_value, int64 log_cnt,
+ bool is_called);
+
/* stream truncate callback signature */
typedef void (*ReorderBufferStreamTruncateCB) (
ReorderBuffer *rb,
@@ -516,6 +543,12 @@ struct ReorderBuffer
*/
HTAB *by_txn;
+ /*
+ * relfilenumber => XID lookup table for sequences created in a transaction
+ * (also includes altered sequences, which assigns new relfilenumber)
+ */
+ HTAB *sequences;
+
/*
* Transactions that could be a toplevel xact, ordered by LSN of the first
* record bearing that xid.
@@ -552,6 +585,7 @@ struct ReorderBuffer
ReorderBufferApplyTruncateCB apply_truncate;
ReorderBufferCommitCB commit;
ReorderBufferMessageCB message;
+ ReorderBufferSequenceCB sequence;
/*
* Callbacks to be called when streaming a transaction at prepare time.
@@ -571,6 +605,7 @@ struct ReorderBuffer
ReorderBufferStreamCommitCB stream_commit;
ReorderBufferStreamChangeCB stream_change;
ReorderBufferStreamMessageCB stream_message;
+ ReorderBufferStreamSequenceCB stream_sequence;
ReorderBufferStreamTruncateCB stream_truncate;
/*
@@ -646,6 +681,10 @@ extern void ReorderBufferQueueChange(ReorderBuffer *, TransactionId,
extern void ReorderBufferQueueMessage(ReorderBuffer *, TransactionId, Snapshot snapshot, XLogRecPtr lsn,
bool transactional, const char *prefix,
Size message_size, const char *message);
+extern void ReorderBufferQueueSequence(ReorderBuffer *rb, TransactionId xid,
+ XLogRecPtr lsn, RepOriginId origin_id,
+ Oid reloid, RelFileLocator locator,
+ int64 last, int64 log_cnt, bool is_called);
extern void ReorderBufferCommit(ReorderBuffer *, TransactionId,
XLogRecPtr commit_lsn, XLogRecPtr end_lsn,
TimestampTz commit_time, RepOriginId origin_id, XLogRecPtr origin_lsn);
diff --git a/src/test/regress/expected/object_address.out b/src/test/regress/expected/object_address.out
index 4117fc27c9..c95d44b3db 100644
--- a/src/test/regress/expected/object_address.out
+++ b/src/test/regress/expected/object_address.out
@@ -46,6 +46,7 @@ CREATE TRANSFORM FOR int LANGUAGE SQL (
SET client_min_messages = 'ERROR';
CREATE PUBLICATION addr_pub FOR TABLE addr_nsp.gentable;
CREATE PUBLICATION addr_pub_schema FOR ALL TABLES IN SCHEMA addr_nsp;
+CREATE PUBLICATION addr_pub_schema2 FOR ALL SEQUENCES IN SCHEMA addr_nsp;
RESET client_min_messages;
CREATE SUBSCRIPTION regress_addr_sub CONNECTION '' PUBLICATION bar WITH (connect = false, slot_name = NONE);
WARNING: tables were not subscribed, you will have to run ALTER SUBSCRIPTION ... REFRESH PUBLICATION to subscribe the tables
@@ -428,7 +429,8 @@ WITH objects (type, name, args) AS (VALUES
('transform', '{int}', '{sql}'),
('access method', '{btree}', '{}'),
('publication', '{addr_pub}', '{}'),
- ('publication namespace', '{addr_nsp}', '{addr_pub_schema}'),
+ ('publication namespace', '{addr_nsp}', '{addr_pub_schema,t}'),
+ ('publication namespace', '{addr_nsp}', '{addr_pub_schema2,s}'),
('publication relation', '{addr_nsp, gentable}', '{addr_pub}'),
('subscription', '{regress_addr_sub}', '{}'),
('statistics object', '{addr_nsp, gentable_stat}', '{}')
@@ -492,8 +494,9 @@ SELECT (pg_identify_object(addr1.classid, addr1.objid, addr1.objsubid)).*,
subscription | | regress_addr_sub | regress_addr_sub | t
publication | | addr_pub | addr_pub | t
publication relation | | | addr_nsp.gentable in publication addr_pub | t
- publication namespace | | | addr_nsp in publication addr_pub_schema | t
-(50 rows)
+ publication namespace | | | addr_nsp in publication addr_pub_schema type t | t
+ publication namespace | | | addr_nsp in publication addr_pub_schema2 type s | t
+(51 rows)
---
--- Cleanup resources
@@ -506,6 +509,7 @@ drop cascades to server integer
drop cascades to user mapping for regress_addr_user on server integer
DROP PUBLICATION addr_pub;
DROP PUBLICATION addr_pub_schema;
+DROP PUBLICATION addr_pub_schema2;
DROP SUBSCRIPTION regress_addr_sub;
DROP SCHEMA addr_nsp CASCADE;
NOTICE: drop cascades to 14 other objects
diff --git a/src/test/regress/expected/psql.out b/src/test/regress/expected/psql.out
index a7f5700edc..ad891f90e3 100644
--- a/src/test/regress/expected/psql.out
+++ b/src/test/regress/expected/psql.out
@@ -5970,9 +5970,9 @@ List of schemas
(0 rows)
\dRp "no.such.publication"
- List of publications
- Name | Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
-------+-------+------------+---------+---------+---------+-----------+----------
+ List of publications
+ Name | Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+------+-------+------------+---------------+---------+---------+---------+-----------+-----------+----------
(0 rows)
\dRs "no.such.subscription"
diff --git a/src/test/regress/expected/publication.out b/src/test/regress/expected/publication.out
index e6e082de2f..34a053ad1f 100644
--- a/src/test/regress/expected/publication.out
+++ b/src/test/regress/expected/publication.out
@@ -30,20 +30,20 @@ ERROR: conflicting or redundant options
LINE 1: ...ub_xxx WITH (publish_via_partition_root = 'true', publish_vi...
^
\dRp
- List of publications
- Name | Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------+--------------------------+------------+---------+---------+---------+-----------+----------
- testpib_ins_trunct | regress_publication_user | f | t | f | f | f | f
- testpub_default | regress_publication_user | f | f | t | f | f | f
+ List of publications
+ Name | Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ testpib_ins_trunct | regress_publication_user | f | f | t | f | f | f | f | f
+ testpub_default | regress_publication_user | f | f | f | t | f | f | f | f
(2 rows)
-ALTER PUBLICATION testpub_default SET (publish = 'insert, update, delete');
+ALTER PUBLICATION testpub_default SET (publish = 'insert, update, delete, sequence');
\dRp
- List of publications
- Name | Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------+--------------------------+------------+---------+---------+---------+-----------+----------
- testpib_ins_trunct | regress_publication_user | f | t | f | f | f | f
- testpub_default | regress_publication_user | f | t | t | t | f | f
+ List of publications
+ Name | Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ testpib_ins_trunct | regress_publication_user | f | f | t | f | f | f | f | f
+ testpub_default | regress_publication_user | f | f | t | t | t | f | t | f
(2 rows)
--- adding tables
@@ -61,6 +61,9 @@ CREATE TABLE testpub_tbl2 (id serial primary key, data text);
ALTER PUBLICATION testpub_foralltables ADD TABLE testpub_tbl2;
ERROR: publication "testpub_foralltables" is defined as FOR ALL TABLES
DETAIL: Tables cannot be added to or dropped from FOR ALL TABLES publications.
+-- fail - can't add a table using ADD SEQUENCE command
+ALTER PUBLICATION testpub_foralltables ADD SEQUENCE testpub_tbl2;
+ERROR: object type does not match type expected by command
-- fail - can't drop from all tables publication
ALTER PUBLICATION testpub_foralltables DROP TABLE testpub_tbl2;
ERROR: publication "testpub_foralltables" is defined as FOR ALL TABLES
@@ -87,10 +90,10 @@ RESET client_min_messages;
-- should be able to add schema to 'FOR TABLE' publication
ALTER PUBLICATION testpub_fortable ADD ALL TABLES IN SCHEMA pub_test;
\dRp+ testpub_fortable
- Publication testpub_fortable
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_fortable
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"public.testpub_tbl1"
Tables from schemas:
@@ -99,20 +102,20 @@ Tables from schemas:
-- should be able to drop schema from 'FOR TABLE' publication
ALTER PUBLICATION testpub_fortable DROP ALL TABLES IN SCHEMA pub_test;
\dRp+ testpub_fortable
- Publication testpub_fortable
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_fortable
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"public.testpub_tbl1"
-- should be able to set schema to 'FOR TABLE' publication
ALTER PUBLICATION testpub_fortable SET ALL TABLES IN SCHEMA pub_test;
\dRp+ testpub_fortable
- Publication testpub_fortable
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_fortable
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test"
@@ -134,10 +137,10 @@ ERROR: relation "testpub_nopk" is not part of the publication
-- should be able to set table to schema publication
ALTER PUBLICATION testpub_forschema SET TABLE pub_test.testpub_nopk;
\dRp+ testpub_forschema
- Publication testpub_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"pub_test.testpub_nopk"
@@ -159,10 +162,10 @@ Publications:
"testpub_foralltables"
\dRp+ testpub_foralltables
- Publication testpub_foralltables
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | t | t | t | f | f | f
+ Publication testpub_foralltables
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | t | f | t | t | f | f | f | f
(1 row)
DROP TABLE testpub_tbl2;
@@ -174,24 +177,545 @@ CREATE PUBLICATION testpub3 FOR TABLE testpub_tbl3;
CREATE PUBLICATION testpub4 FOR TABLE ONLY testpub_tbl3;
RESET client_min_messages;
\dRp+ testpub3
- Publication testpub3
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub3
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"public.testpub_tbl3"
"public.testpub_tbl3a"
\dRp+ testpub4
- Publication testpub4
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub4
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"public.testpub_tbl3"
DROP TABLE testpub_tbl3, testpub_tbl3a;
DROP PUBLICATION testpub3, testpub4;
+--- adding sequences
+CREATE SEQUENCE testpub_seq0;
+CREATE SEQUENCE pub_test.testpub_seq1;
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_forallsequences FOR ALL SEQUENCES WITH (publish = 'sequence');
+RESET client_min_messages;
+ALTER PUBLICATION testpub_forallsequences SET (publish = 'insert, sequence');
+CREATE SEQUENCE testpub_seq2;
+-- fail - can't add to for all sequences publication
+ALTER PUBLICATION testpub_forallsequences ADD SEQUENCE testpub_seq2;
+ERROR: publication "testpub_forallsequences" is defined as FOR ALL SEQUENCES
+DETAIL: Sequences cannot be added to or dropped from FOR ALL SEQUENCES publications.
+-- fail - can't drop from all sequences publication
+ALTER PUBLICATION testpub_forallsequences DROP SEQUENCE testpub_seq2;
+ERROR: publication "testpub_forallsequences" is defined as FOR ALL SEQUENCES
+DETAIL: Sequences cannot be added to or dropped from FOR ALL SEQUENCES publications.
+-- fail - can't add to for all sequences publication
+ALTER PUBLICATION testpub_forallsequences SET SEQUENCE pub_test.testpub_seq1;
+ERROR: publication "testpub_forallsequences" is defined as FOR ALL SEQUENCES
+DETAIL: Sequences cannot be added to or dropped from FOR ALL SEQUENCES publications.
+-- fail - can't add schema to 'FOR ALL SEQUENCES' publication
+ALTER PUBLICATION testpub_forallsequences ADD ALL SEQUENCES IN SCHEMA pub_test;
+ERROR: publication "testpub_forallsequences" is defined as FOR ALL SEQUENCES
+DETAIL: Sequences from schema cannot be added to, dropped from, or set on FOR ALL SEQUENCES publications.
+-- fail - can't drop schema from 'FOR ALL SEQUENCES' publication
+ALTER PUBLICATION testpub_forallsequences DROP ALL SEQUENCES IN SCHEMA pub_test;
+ERROR: publication "testpub_forallsequences" is defined as FOR ALL SEQUENCES
+DETAIL: Sequences from schema cannot be added to, dropped from, or set on FOR ALL SEQUENCES publications.
+-- fail - can't set schema to 'FOR ALL SEQUENCES' publication
+ALTER PUBLICATION testpub_forallsequences SET ALL SEQUENCES IN SCHEMA pub_test;
+ERROR: publication "testpub_forallsequences" is defined as FOR ALL SEQUENCES
+DETAIL: Sequences from schema cannot be added to, dropped from, or set on FOR ALL SEQUENCES publications.
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_forsequence FOR SEQUENCE testpub_seq0;
+RESET client_min_messages;
+-- should be able to add schema to 'FOR SEQUENCE' publication
+ALTER PUBLICATION testpub_forsequence ADD ALL SEQUENCES IN SCHEMA pub_test;
+\dRp+ testpub_forsequence
+ Publication testpub_forsequence
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Sequences:
+ "public.testpub_seq0"
+Sequences from schemas:
+ "pub_test"
+
+-- fail - can't add sequence from the schema we already added
+ALTER PUBLICATION testpub_forsequence ADD SEQUENCE pub_test.testpub_seq1;
+ERROR: cannot add relation "pub_test.testpub_seq1" to publication
+DETAIL: Sequence's schema "pub_test" is already part of the publication or part of the specified schema list.
+-- fail - can't add sequence using ADD TABLE command
+ALTER PUBLICATION testpub_forsequence ADD TABLE pub_test.testpub_seq1;
+ERROR: object type does not match type expected by command
+-- should be able to drop schema from 'FOR SEQUENCE' publication
+ALTER PUBLICATION testpub_forsequence DROP ALL SEQUENCES IN SCHEMA pub_test;
+\dRp+ testpub_forsequence
+ Publication testpub_forsequence
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Sequences:
+ "public.testpub_seq0"
+
+-- should be able to set schema to 'FOR SEQUENCE' publication
+ALTER PUBLICATION testpub_forsequence SET ALL SEQUENCES IN SCHEMA pub_test;
+\dRp+ testpub_forsequence
+ Publication testpub_forsequence
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Sequences from schemas:
+ "pub_test"
+
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_forschema FOR ALL SEQUENCES IN SCHEMA pub_test;
+RESET client_min_messages;
+-- fail - can't create publication with schema and sequence of the same schema
+CREATE PUBLICATION testpub_for_seq_schema FOR ALL SEQUENCES IN SCHEMA pub_test, SEQUENCE pub_test.testpub_seq1;
+ERROR: cannot add relation "pub_test.testpub_seq1" to publication
+DETAIL: Sequence's schema "pub_test" is already part of the publication or part of the specified schema list.
+-- fail - can't add a sequence of the same schema to the schema publication
+ALTER PUBLICATION testpub_forschema ADD SEQUENCE pub_test.testpub_seq1;
+ERROR: cannot add relation "pub_test.testpub_seq1" to publication
+DETAIL: Sequence's schema "pub_test" is already part of the publication or part of the specified schema list.
+-- fail - can't drop a sequence from the schema publication which isn't in the
+-- publication
+ALTER PUBLICATION testpub_forschema DROP SEQUENCE pub_test.testpub_seq1;
+ERROR: relation "testpub_seq1" is not part of the publication
+-- should be able to set sequence to schema publication
+ALTER PUBLICATION testpub_forschema SET SEQUENCE pub_test.testpub_seq1;
+\dRp+ testpub_forschema
+ Publication testpub_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Sequences:
+ "pub_test.testpub_seq1"
+
+SELECT pubname, puballtables, puballsequences FROM pg_publication WHERE pubname = 'testpub_forallsequences';
+ pubname | puballtables | puballsequences
+-------------------------+--------------+-----------------
+ testpub_forallsequences | f | t
+(1 row)
+
+\d+ pub_test.testpub_seq1
+ Sequence "pub_test.testpub_seq1"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+Publications:
+ "testpub_forallsequences"
+ "testpub_forschema"
+ "testpub_forsequence"
+
+\dRp+ testpub_forallsequences
+ Publication testpub_forallsequences
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | t | t | f | f | f | t | f
+(1 row)
+
+DROP SEQUENCE testpub_seq0, pub_test.testpub_seq1, testpub_seq2;
+DROP PUBLICATION testpub_forallsequences, testpub_forsequence, testpub_forschema;
+-- publication testing multiple sequences at the same time
+CREATE SEQUENCE testpub_seq1;
+CREATE SEQUENCE testpub_seq2;
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_multi FOR SEQUENCE testpub_seq1, testpub_seq2;
+RESET client_min_messages;
+\dRp+ testpub_multi
+ Publication testpub_multi
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Sequences:
+ "public.testpub_seq1"
+ "public.testpub_seq2"
+
+DROP PUBLICATION testpub_multi;
+DROP SEQUENCE testpub_seq1;
+DROP SEQUENCE testpub_seq2;
+-- Publication mixing tables and sequences
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_mix;
+RESET client_min_messages;
+CREATE SEQUENCE testpub_seq1;
+CREATE SEQUENCE pub_test.testpub_seq2;
+ALTER PUBLICATION testpub_mix ADD SEQUENCE testpub_seq1, TABLE testpub_tbl1;
+\dRp+ testpub_mix
+ Publication testpub_mix
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Tables:
+ "public.testpub_tbl1"
+Sequences:
+ "public.testpub_seq1"
+
+ALTER PUBLICATION testpub_mix ADD ALL SEQUENCES IN SCHEMA pub_test, ALL TABLES IN SCHEMA pub_test;
+\dRp+ testpub_mix
+ Publication testpub_mix
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Tables:
+ "public.testpub_tbl1"
+Tables from schemas:
+ "pub_test"
+Sequences:
+ "public.testpub_seq1"
+Sequences from schemas:
+ "pub_test"
+
+ALTER PUBLICATION testpub_mix DROP ALL SEQUENCES IN SCHEMA pub_test;
+\dRp+ testpub_mix
+ Publication testpub_mix
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Tables:
+ "public.testpub_tbl1"
+Tables from schemas:
+ "pub_test"
+Sequences:
+ "public.testpub_seq1"
+
+ALTER PUBLICATION testpub_mix DROP ALL TABLES IN SCHEMA pub_test;
+\dRp+ testpub_mix
+ Publication testpub_mix
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Tables:
+ "public.testpub_tbl1"
+Sequences:
+ "public.testpub_seq1"
+
+DROP PUBLICATION testpub_mix;
+DROP SEQUENCE testpub_seq1;
+DROP SEQUENCE pub_test.testpub_seq2;
+-- make sure we replicate only the correct relation type
+CREATE SCHEMA pub_test1;
+CREATE SEQUENCE pub_test1.test_seq1;
+CREATE TABLE pub_test1.test_tbl1 (a int primary key, b int);
+CREATE SCHEMA pub_test2;
+CREATE SEQUENCE pub_test2.test_seq2;
+CREATE TABLE pub_test2.test_tbl2 (a int primary key, b int);
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_schemas;
+RESET client_min_messages;
+-- add tables from one schema, sequences from the other
+ALTER PUBLICATION testpub_schemas ADD ALL TABLES IN SCHEMA pub_test2;
+ALTER PUBLICATION testpub_schemas ADD ALL SEQUENCES IN SCHEMA pub_test1;
+\dRp+ testpub_schemas
+ Publication testpub_schemas
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Tables from schemas:
+ "pub_test2"
+Sequences from schemas:
+ "pub_test1"
+
+\dn+ pub_test1
+ List of schemas
+ Name | Owner | Access privileges | Description
+-----------+--------------------------+-------------------+-------------
+ pub_test1 | regress_publication_user | |
+Publications:
+ "testpub_schemas" (sequences)
+
+\dn+ pub_test2
+ List of schemas
+ Name | Owner | Access privileges | Description
+-----------+--------------------------+-------------------+-------------
+ pub_test2 | regress_publication_user | |
+Publications:
+ "testpub_schemas" (tables)
+
+\d+ pub_test1.test_seq1;
+ Sequence "pub_test1.test_seq1"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test1.test_tbl1;
+ Table "pub_test1.test_tbl1"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl1_pkey" PRIMARY KEY, btree (a)
+
+\d+ pub_test2.test_seq2;
+ Sequence "pub_test2.test_seq2"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+
+\d+ pub_test2.test_tbl2;
+ Table "pub_test2.test_tbl2"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl2_pkey" PRIMARY KEY, btree (a)
+Publications:
+ "testpub_schemas"
+
+-- add the other object type from each schema
+ALTER PUBLICATION testpub_schemas ADD ALL TABLES IN SCHEMA pub_test1;
+ALTER PUBLICATION testpub_schemas ADD ALL SEQUENCES IN SCHEMA pub_test2;
+\dRp+ testpub_schemas
+ Publication testpub_schemas
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Tables from schemas:
+ "pub_test1"
+ "pub_test2"
+Sequences from schemas:
+ "pub_test1"
+ "pub_test2"
+
+\dn+ pub_test1
+ List of schemas
+ Name | Owner | Access privileges | Description
+-----------+--------------------------+-------------------+-------------
+ pub_test1 | regress_publication_user | |
+Publications:
+ "testpub_schemas" (sequences)
+ "testpub_schemas" (tables)
+
+\dn+ pub_test2
+ List of schemas
+ Name | Owner | Access privileges | Description
+-----------+--------------------------+-------------------+-------------
+ pub_test2 | regress_publication_user | |
+Publications:
+ "testpub_schemas" (tables)
+ "testpub_schemas" (sequences)
+
+\d+ pub_test1.test_seq1;
+ Sequence "pub_test1.test_seq1"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test1.test_tbl1;
+ Table "pub_test1.test_tbl1"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl1_pkey" PRIMARY KEY, btree (a)
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test2.test_seq2;
+ Sequence "pub_test2.test_seq2"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test2.test_tbl2;
+ Table "pub_test2.test_tbl2"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl2_pkey" PRIMARY KEY, btree (a)
+Publications:
+ "testpub_schemas"
+
+-- now drop the object type added first
+ALTER PUBLICATION testpub_schemas DROP ALL TABLES IN SCHEMA pub_test2;
+ALTER PUBLICATION testpub_schemas DROP ALL SEQUENCES IN SCHEMA pub_test1;
+\dRp+ testpub_schemas
+ Publication testpub_schemas
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Tables from schemas:
+ "pub_test1"
+Sequences from schemas:
+ "pub_test2"
+
+\dn+ pub_test1
+ List of schemas
+ Name | Owner | Access privileges | Description
+-----------+--------------------------+-------------------+-------------
+ pub_test1 | regress_publication_user | |
+Publications:
+ "testpub_schemas" (tables)
+
+\dn+ pub_test2
+ List of schemas
+ Name | Owner | Access privileges | Description
+-----------+--------------------------+-------------------+-------------
+ pub_test2 | regress_publication_user | |
+Publications:
+ "testpub_schemas" (sequences)
+
+\d+ pub_test1.test_seq1;
+ Sequence "pub_test1.test_seq1"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+
+\d+ pub_test1.test_tbl1;
+ Table "pub_test1.test_tbl1"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl1_pkey" PRIMARY KEY, btree (a)
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test2.test_seq2;
+ Sequence "pub_test2.test_seq2"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test2.test_tbl2;
+ Table "pub_test2.test_tbl2"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl2_pkey" PRIMARY KEY, btree (a)
+
+-- should fail (publication contains the whole schema)
+ALTER PUBLICATION testpub_schemas ADD TABLE pub_test1.test_tbl1;
+ERROR: cannot add relation "pub_test1.test_tbl1" to publication
+DETAIL: Table's schema "pub_test1" is already part of the publication or part of the specified schema list.
+ALTER PUBLICATION testpub_schemas ADD SEQUENCE pub_test2.test_seq2;
+ERROR: cannot add relation "pub_test2.test_seq2" to publication
+DETAIL: Sequence's schema "pub_test2" is already part of the publication or part of the specified schema list.
+-- should work (different schema)
+ALTER PUBLICATION testpub_schemas ADD TABLE pub_test2.test_tbl2;
+ALTER PUBLICATION testpub_schemas ADD SEQUENCE pub_test1.test_seq1;
+\dRp+ testpub_schemas
+ Publication testpub_schemas
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Tables:
+ "pub_test2.test_tbl2"
+Tables from schemas:
+ "pub_test1"
+Sequences:
+ "pub_test1.test_seq1"
+Sequences from schemas:
+ "pub_test2"
+
+\d+ pub_test1.test_seq1;
+ Sequence "pub_test1.test_seq1"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test1.test_tbl1;
+ Table "pub_test1.test_tbl1"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl1_pkey" PRIMARY KEY, btree (a)
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test2.test_seq2;
+ Sequence "pub_test2.test_seq2"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test2.test_tbl2;
+ Table "pub_test2.test_tbl2"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl2_pkey" PRIMARY KEY, btree (a)
+Publications:
+ "testpub_schemas"
+
+-- now drop the explicitly added objects again
+ALTER PUBLICATION testpub_schemas DROP TABLE pub_test2.test_tbl2;
+ALTER PUBLICATION testpub_schemas DROP SEQUENCE pub_test1.test_seq1;
+\dRp+ testpub_schemas
+ Publication testpub_schemas
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Tables from schemas:
+ "pub_test1"
+Sequences from schemas:
+ "pub_test2"
+
+\d+ pub_test1.test_seq1;
+ Sequence "pub_test1.test_seq1"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+
+\d+ pub_test1.test_tbl1;
+ Table "pub_test1.test_tbl1"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl1_pkey" PRIMARY KEY, btree (a)
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test2.test_seq2;
+ Sequence "pub_test2.test_seq2"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test2.test_tbl2;
+ Table "pub_test2.test_tbl2"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl2_pkey" PRIMARY KEY, btree (a)
+
+DROP PUBLICATION testpub_schemas;
+DROP TABLE pub_test1.test_tbl1, pub_test2.test_tbl2;
+DROP SEQUENCE pub_test1.test_seq1, pub_test2.test_seq2;
+DROP SCHEMA pub_test1, pub_test2;
-- Tests for partitioned tables
SET client_min_messages = 'ERROR';
CREATE PUBLICATION testpub_forparted;
@@ -207,10 +731,10 @@ UPDATE testpub_parted1 SET a = 1;
-- only parent is listed as being in publication, not the partition
ALTER PUBLICATION testpub_forparted ADD TABLE testpub_parted;
\dRp+ testpub_forparted
- Publication testpub_forparted
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_forparted
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"public.testpub_parted"
@@ -225,10 +749,10 @@ ALTER TABLE testpub_parted DETACH PARTITION testpub_parted1;
UPDATE testpub_parted1 SET a = 1;
ALTER PUBLICATION testpub_forparted SET (publish_via_partition_root = true);
\dRp+ testpub_forparted
- Publication testpub_forparted
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | t
+ Publication testpub_forparted
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | t
Tables:
"public.testpub_parted"
@@ -257,10 +781,10 @@ SET client_min_messages = 'ERROR';
CREATE PUBLICATION testpub5 FOR TABLE testpub_rf_tbl1, testpub_rf_tbl2 WHERE (c <> 'test' AND d < 5) WITH (publish = 'insert');
RESET client_min_messages;
\dRp+ testpub5
- Publication testpub5
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | f | f | f | f
+ Publication testpub5
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | f | f | f | f | f
Tables:
"public.testpub_rf_tbl1"
"public.testpub_rf_tbl2" WHERE ((c <> 'test'::text) AND (d < 5))
@@ -273,10 +797,10 @@ Tables:
ALTER PUBLICATION testpub5 ADD TABLE testpub_rf_tbl3 WHERE (e > 1000 AND e < 2000);
\dRp+ testpub5
- Publication testpub5
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | f | f | f | f
+ Publication testpub5
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | f | f | f | f | f
Tables:
"public.testpub_rf_tbl1"
"public.testpub_rf_tbl2" WHERE ((c <> 'test'::text) AND (d < 5))
@@ -292,10 +816,10 @@ Publications:
ALTER PUBLICATION testpub5 DROP TABLE testpub_rf_tbl2;
\dRp+ testpub5
- Publication testpub5
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | f | f | f | f
+ Publication testpub5
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | f | f | f | f | f
Tables:
"public.testpub_rf_tbl1"
"public.testpub_rf_tbl3" WHERE ((e > 1000) AND (e < 2000))
@@ -303,10 +827,10 @@ Tables:
-- remove testpub_rf_tbl1 and add testpub_rf_tbl3 again (another WHERE expression)
ALTER PUBLICATION testpub5 SET TABLE testpub_rf_tbl3 WHERE (e > 300 AND e < 500);
\dRp+ testpub5
- Publication testpub5
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | f | f | f | f
+ Publication testpub5
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | f | f | f | f | f
Tables:
"public.testpub_rf_tbl3" WHERE ((e > 300) AND (e < 500))
@@ -339,10 +863,10 @@ SET client_min_messages = 'ERROR';
CREATE PUBLICATION testpub_syntax1 FOR TABLE testpub_rf_tbl1, ONLY testpub_rf_tbl3 WHERE (e < 999) WITH (publish = 'insert');
RESET client_min_messages;
\dRp+ testpub_syntax1
- Publication testpub_syntax1
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | f | f | f | f
+ Publication testpub_syntax1
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | f | f | f | f | f
Tables:
"public.testpub_rf_tbl1"
"public.testpub_rf_tbl3" WHERE (e < 999)
@@ -352,10 +876,10 @@ SET client_min_messages = 'ERROR';
CREATE PUBLICATION testpub_syntax2 FOR TABLE testpub_rf_tbl1, testpub_rf_schema1.testpub_rf_tbl5 WHERE (h < 999) WITH (publish = 'insert');
RESET client_min_messages;
\dRp+ testpub_syntax2
- Publication testpub_syntax2
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | f | f | f | f
+ Publication testpub_syntax2
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | f | f | f | f | f
Tables:
"public.testpub_rf_tbl1"
"testpub_rf_schema1.testpub_rf_tbl5" WHERE (h < 999)
@@ -678,10 +1202,10 @@ CREATE PUBLICATION testpub_table_ins WITH (publish = 'insert, truncate');
RESET client_min_messages;
ALTER PUBLICATION testpub_table_ins ADD TABLE testpub_tbl5 (a); -- ok
\dRp+ testpub_table_ins
- Publication testpub_table_ins
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | f | f | t | f
+ Publication testpub_table_ins
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | f | f | t | f | f
Tables:
"public.testpub_tbl5" (a)
@@ -823,10 +1347,10 @@ CREATE TABLE testpub_tbl_both_filters (a int, b int, c int, PRIMARY KEY (a,c));
ALTER TABLE testpub_tbl_both_filters REPLICA IDENTITY USING INDEX testpub_tbl_both_filters_pkey;
ALTER PUBLICATION testpub_both_filters ADD TABLE testpub_tbl_both_filters (a,c) WHERE (c != 1);
\dRp+ testpub_both_filters
- Publication testpub_both_filters
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_both_filters
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"public.testpub_tbl_both_filters" (a, c) WHERE (c <> 1)
@@ -1031,10 +1555,10 @@ ERROR: relation "testpub_tbl1" is already member of publication "testpub_fortbl
CREATE PUBLICATION testpub_fortbl FOR TABLE testpub_tbl1;
ERROR: publication "testpub_fortbl" already exists
\dRp+ testpub_fortbl
- Publication testpub_fortbl
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_fortbl
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"pub_test.testpub_nopk"
"public.testpub_tbl1"
@@ -1072,10 +1596,10 @@ Publications:
"testpub_fortbl"
\dRp+ testpub_default
- Publication testpub_default
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | f | f
+ Publication testpub_default
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | f | t | f
Tables:
"pub_test.testpub_nopk"
"public.testpub_tbl1"
@@ -1153,10 +1677,10 @@ REVOKE CREATE ON DATABASE regression FROM regress_publication_user2;
DROP TABLE testpub_parted;
DROP TABLE testpub_tbl1;
\dRp+ testpub_default
- Publication testpub_default
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | f | f
+ Publication testpub_default
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | f | t | f
(1 row)
-- fail - must be owner of publication
@@ -1166,20 +1690,20 @@ ERROR: must be owner of publication testpub_default
RESET ROLE;
ALTER PUBLICATION testpub_default RENAME TO testpub_foo;
\dRp testpub_foo
- List of publications
- Name | Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
--------------+--------------------------+------------+---------+---------+---------+-----------+----------
- testpub_foo | regress_publication_user | f | t | t | t | f | f
+ List of publications
+ Name | Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+-------------+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ testpub_foo | regress_publication_user | f | f | t | t | t | f | t | f
(1 row)
-- rename back to keep the rest simple
ALTER PUBLICATION testpub_foo RENAME TO testpub_default;
ALTER PUBLICATION testpub_default OWNER TO regress_publication_user2;
\dRp testpub_default
- List of publications
- Name | Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
------------------+---------------------------+------------+---------+---------+---------+-----------+----------
- testpub_default | regress_publication_user2 | f | t | t | t | f | f
+ List of publications
+ Name | Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+-----------------+---------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ testpub_default | regress_publication_user2 | f | f | t | t | t | f | t | f
(1 row)
-- adding schemas and tables
@@ -1195,19 +1719,19 @@ CREATE TABLE "CURRENT_SCHEMA"."CURRENT_SCHEMA"(id int);
SET client_min_messages = 'ERROR';
CREATE PUBLICATION testpub1_forschema FOR ALL TABLES IN SCHEMA pub_test1;
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
CREATE PUBLICATION testpub2_forschema FOR ALL TABLES IN SCHEMA pub_test1, pub_test2, pub_test3;
\dRp+ testpub2_forschema
- Publication testpub2_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub2_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
"pub_test2"
@@ -1221,44 +1745,44 @@ CREATE PUBLICATION testpub6_forschema FOR ALL TABLES IN SCHEMA "CURRENT_SCHEMA",
CREATE PUBLICATION testpub_fortable FOR TABLE "CURRENT_SCHEMA"."CURRENT_SCHEMA";
RESET client_min_messages;
\dRp+ testpub3_forschema
- Publication testpub3_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub3_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"public"
\dRp+ testpub4_forschema
- Publication testpub4_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub4_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"CURRENT_SCHEMA"
\dRp+ testpub5_forschema
- Publication testpub5_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub5_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"CURRENT_SCHEMA"
"public"
\dRp+ testpub6_forschema
- Publication testpub6_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub6_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"CURRENT_SCHEMA"
"public"
\dRp+ testpub_fortable
- Publication testpub_fortable
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_fortable
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"CURRENT_SCHEMA.CURRENT_SCHEMA"
@@ -1292,10 +1816,10 @@ ERROR: schema "testpub_view" does not exist
-- dropping the schema should reflect the change in publication
DROP SCHEMA pub_test3;
\dRp+ testpub2_forschema
- Publication testpub2_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub2_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
"pub_test2"
@@ -1303,20 +1827,20 @@ Tables from schemas:
-- renaming the schema should reflect the change in publication
ALTER SCHEMA pub_test1 RENAME to pub_test1_renamed;
\dRp+ testpub2_forschema
- Publication testpub2_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub2_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1_renamed"
"pub_test2"
ALTER SCHEMA pub_test1_renamed RENAME to pub_test1;
\dRp+ testpub2_forschema
- Publication testpub2_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub2_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
"pub_test2"
@@ -1324,10 +1848,10 @@ Tables from schemas:
-- alter publication add schema
ALTER PUBLICATION testpub1_forschema ADD ALL TABLES IN SCHEMA pub_test2;
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
"pub_test2"
@@ -1336,10 +1860,10 @@ Tables from schemas:
ALTER PUBLICATION testpub1_forschema ADD ALL TABLES IN SCHEMA non_existent_schema;
ERROR: schema "non_existent_schema" does not exist
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
"pub_test2"
@@ -1348,10 +1872,10 @@ Tables from schemas:
ALTER PUBLICATION testpub1_forschema ADD ALL TABLES IN SCHEMA pub_test1;
ERROR: schema "pub_test1" is already member of publication "testpub1_forschema"
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
"pub_test2"
@@ -1359,10 +1883,10 @@ Tables from schemas:
-- alter publication drop schema
ALTER PUBLICATION testpub1_forschema DROP ALL TABLES IN SCHEMA pub_test2;
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
@@ -1370,10 +1894,10 @@ Tables from schemas:
ALTER PUBLICATION testpub1_forschema DROP ALL TABLES IN SCHEMA pub_test2;
ERROR: tables from schema "pub_test2" are not part of the publication
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
@@ -1381,29 +1905,29 @@ Tables from schemas:
ALTER PUBLICATION testpub1_forschema DROP ALL TABLES IN SCHEMA non_existent_schema;
ERROR: schema "non_existent_schema" does not exist
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
-- drop all schemas
ALTER PUBLICATION testpub1_forschema DROP ALL TABLES IN SCHEMA pub_test1;
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
(1 row)
-- alter publication set multiple schema
ALTER PUBLICATION testpub1_forschema SET ALL TABLES IN SCHEMA pub_test1, pub_test2;
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
"pub_test2"
@@ -1412,10 +1936,10 @@ Tables from schemas:
ALTER PUBLICATION testpub1_forschema SET ALL TABLES IN SCHEMA non_existent_schema;
ERROR: schema "non_existent_schema" does not exist
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
"pub_test2"
@@ -1424,10 +1948,10 @@ Tables from schemas:
-- removing the duplicate schemas
ALTER PUBLICATION testpub1_forschema SET ALL TABLES IN SCHEMA pub_test1, pub_test1;
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
@@ -1506,18 +2030,18 @@ SET client_min_messages = 'ERROR';
CREATE PUBLICATION testpub3_forschema;
RESET client_min_messages;
\dRp+ testpub3_forschema
- Publication testpub3_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub3_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
(1 row)
ALTER PUBLICATION testpub3_forschema SET ALL TABLES IN SCHEMA pub_test1;
\dRp+ testpub3_forschema
- Publication testpub3_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub3_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
@@ -1527,20 +2051,20 @@ CREATE PUBLICATION testpub_forschema_fortable FOR ALL TABLES IN SCHEMA pub_test1
CREATE PUBLICATION testpub_fortable_forschema FOR TABLE pub_test2.tbl1, ALL TABLES IN SCHEMA pub_test1;
RESET client_min_messages;
\dRp+ testpub_forschema_fortable
- Publication testpub_forschema_fortable
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_forschema_fortable
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"pub_test2.tbl1"
Tables from schemas:
"pub_test1"
\dRp+ testpub_fortable_forschema
- Publication testpub_fortable_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_fortable_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"pub_test2.tbl1"
Tables from schemas:
@@ -1584,40 +2108,85 @@ CREATE SCHEMA sch1;
CREATE SCHEMA sch2;
CREATE TABLE sch1.tbl1 (a int) PARTITION BY RANGE(a);
CREATE TABLE sch2.tbl1_part1 PARTITION OF sch1.tbl1 FOR VALUES FROM (1) to (10);
+CREATE SEQUENCE sch1.seq1;
+CREATE SEQUENCE sch2.seq2;
-- Schema publication that does not include the schema that has the parent table
CREATE PUBLICATION pub FOR ALL TABLES IN SCHEMA sch2 WITH (PUBLISH_VIA_PARTITION_ROOT=1);
+ALTER PUBLICATION pub ADD ALL SEQUENCES IN SCHEMA sch2;
SELECT * FROM pg_publication_tables;
pubname | schemaname | tablename | attnames | rowfilter
---------+------------+------------+----------+-----------
pub | sch2 | tbl1_part1 | {a} |
(1 row)
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch2 | seq2
+(1 row)
+
DROP PUBLICATION pub;
-- Table publication that does not include the parent table
CREATE PUBLICATION pub FOR TABLE sch2.tbl1_part1 WITH (PUBLISH_VIA_PARTITION_ROOT=1);
+ALTER PUBLICATION pub ADD SEQUENCE sch2.seq2;
SELECT * FROM pg_publication_tables;
pubname | schemaname | tablename | attnames | rowfilter
---------+------------+------------+----------+-----------
pub | sch2 | tbl1_part1 | {a} |
(1 row)
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch2 | seq2
+(1 row)
+
-- Table publication that includes both the parent table and the child table
ALTER PUBLICATION pub ADD TABLE sch1.tbl1;
+ALTER PUBLICATION pub ADD SEQUENCE sch1.seq1;
SELECT * FROM pg_publication_tables;
pubname | schemaname | tablename | attnames | rowfilter
---------+------------+-----------+----------+-----------
pub | sch1 | tbl1 | {a} |
(1 row)
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch1 | seq1
+ pub | sch2 | seq2
+(2 rows)
+
DROP PUBLICATION pub;
-- Schema publication that does not include the schema that has the parent table
CREATE PUBLICATION pub FOR ALL TABLES IN SCHEMA sch2 WITH (PUBLISH_VIA_PARTITION_ROOT=0);
+ALTER PUBLICATION pub ADD SEQUENCE sch1.seq1;
SELECT * FROM pg_publication_tables;
pubname | schemaname | tablename | attnames | rowfilter
---------+------------+------------+----------+-----------
pub | sch2 | tbl1_part1 | {a} |
(1 row)
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch1 | seq1
+(1 row)
+
+DROP PUBLICATION pub;
+-- Sequence publication
+CREATE PUBLICATION pub FOR SEQUENCE sch2.seq2;
+SELECT * FROM pg_publication_tables;
+ pubname | schemaname | tablename | attnames | rowfilter
+---------+------------+-----------+----------+-----------
+(0 rows)
+
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch2 | seq2
+(1 row)
+
DROP PUBLICATION pub;
-- Table publication that does not include the parent table
CREATE PUBLICATION pub FOR TABLE sch2.tbl1_part1 WITH (PUBLISH_VIA_PARTITION_ROOT=0);
@@ -1627,14 +2196,26 @@ SELECT * FROM pg_publication_tables;
pub | sch2 | tbl1_part1 | {a} |
(1 row)
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+(0 rows)
+
-- Table publication that includes both the parent table and the child table
ALTER PUBLICATION pub ADD TABLE sch1.tbl1;
+ALTER PUBLICATION pub ADD ALL SEQUENCES IN SCHEMA sch2;
SELECT * FROM pg_publication_tables;
pubname | schemaname | tablename | attnames | rowfilter
---------+------------+------------+----------+-----------
pub | sch2 | tbl1_part1 | {a} |
(1 row)
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch2 | seq2
+(1 row)
+
DROP PUBLICATION pub;
DROP TABLE sch2.tbl1_part1;
DROP TABLE sch1.tbl1;
@@ -1650,9 +2231,81 @@ SELECT * FROM pg_publication_tables;
pub | sch1 | tbl1 | {a} |
(1 row)
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+(0 rows)
+
+DROP PUBLICATION pub;
+-- Schema publication
+CREATE PUBLICATION pub FOR SEQUENCE sch2.seq2;
+SELECT * FROM pg_publication_tables;
+ pubname | schemaname | tablename | attnames | rowfilter
+---------+------------+-----------+----------+-----------
+(0 rows)
+
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch2 | seq2
+(1 row)
+
+DROP PUBLICATION pub;
+-- Sequence publication
+CREATE PUBLICATION pub FOR ALL SEQUENCES IN SCHEMA sch2;
+SELECT * FROM pg_publication_tables;
+ pubname | schemaname | tablename | attnames | rowfilter
+---------+------------+-----------+----------+-----------
+(0 rows)
+
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch2 | seq2
+(1 row)
+
+ALTER PUBLICATION pub ADD SEQUENCE sch1.seq1;
+SELECT * FROM pg_publication_tables;
+ pubname | schemaname | tablename | attnames | rowfilter
+---------+------------+-----------+----------+-----------
+(0 rows)
+
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch1 | seq1
+ pub | sch2 | seq2
+(2 rows)
+
+ALTER PUBLICATION pub DROP SEQUENCE sch1.seq1;
+SELECT * FROM pg_publication_tables;
+ pubname | schemaname | tablename | attnames | rowfilter
+---------+------------+-----------+----------+-----------
+(0 rows)
+
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch2 | seq2
+(1 row)
+
+ALTER PUBLICATION pub ADD ALL SEQUENCES IN SCHEMA sch1;
+SELECT * FROM pg_publication_tables;
+ pubname | schemaname | tablename | attnames | rowfilter
+---------+------------+-----------+----------+-----------
+(0 rows)
+
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch1 | seq1
+ pub | sch2 | seq2
+(2 rows)
+
RESET client_min_messages;
DROP PUBLICATION pub;
DROP TABLE sch1.tbl1;
+DROP SEQUENCE sch1.seq1, sch2.seq2;
DROP SCHEMA sch1 cascade;
DROP SCHEMA sch2 cascade;
RESET SESSION AUTHORIZATION;
diff --git a/src/test/regress/expected/rules.out b/src/test/regress/expected/rules.out
index 7ec3d2688f..0b4b388321 100644
--- a/src/test/regress/expected/rules.out
+++ b/src/test/regress/expected/rules.out
@@ -1436,6 +1436,14 @@ pg_prepared_xacts| SELECT p.transaction,
FROM ((pg_prepared_xact() p(transaction, gid, prepared, ownerid, dbid)
LEFT JOIN pg_authid u ON ((p.ownerid = u.oid)))
LEFT JOIN pg_database d ON ((p.dbid = d.oid)));
+pg_publication_sequences| SELECT p.pubname,
+ n.nspname AS schemaname,
+ c.relname AS sequencename
+ FROM pg_publication p,
+ LATERAL pg_get_publication_sequences((p.pubname)::text) gps(relid),
+ (pg_class c
+ JOIN pg_namespace n ON ((n.oid = c.relnamespace)))
+ WHERE (c.oid = gps.relid);
pg_publication_tables| SELECT p.pubname,
n.nspname AS schemaname,
c.relname AS tablename,
diff --git a/src/test/regress/sql/object_address.sql b/src/test/regress/sql/object_address.sql
index acd0468a9d..9d8323468d 100644
--- a/src/test/regress/sql/object_address.sql
+++ b/src/test/regress/sql/object_address.sql
@@ -49,6 +49,7 @@ CREATE TRANSFORM FOR int LANGUAGE SQL (
SET client_min_messages = 'ERROR';
CREATE PUBLICATION addr_pub FOR TABLE addr_nsp.gentable;
CREATE PUBLICATION addr_pub_schema FOR ALL TABLES IN SCHEMA addr_nsp;
+CREATE PUBLICATION addr_pub_schema2 FOR ALL SEQUENCES IN SCHEMA addr_nsp;
RESET client_min_messages;
CREATE SUBSCRIPTION regress_addr_sub CONNECTION '' PUBLICATION bar WITH (connect = false, slot_name = NONE);
CREATE STATISTICS addr_nsp.gentable_stat ON a, b FROM addr_nsp.gentable;
@@ -198,7 +199,8 @@ WITH objects (type, name, args) AS (VALUES
('transform', '{int}', '{sql}'),
('access method', '{btree}', '{}'),
('publication', '{addr_pub}', '{}'),
- ('publication namespace', '{addr_nsp}', '{addr_pub_schema}'),
+ ('publication namespace', '{addr_nsp}', '{addr_pub_schema,t}'),
+ ('publication namespace', '{addr_nsp}', '{addr_pub_schema2,s}'),
('publication relation', '{addr_nsp, gentable}', '{addr_pub}'),
('subscription', '{regress_addr_sub}', '{}'),
('statistics object', '{addr_nsp, gentable_stat}', '{}')
@@ -218,6 +220,7 @@ SELECT (pg_identify_object(addr1.classid, addr1.objid, addr1.objsubid)).*,
DROP FOREIGN DATA WRAPPER addr_fdw CASCADE;
DROP PUBLICATION addr_pub;
DROP PUBLICATION addr_pub_schema;
+DROP PUBLICATION addr_pub_schema2;
DROP SUBSCRIPTION regress_addr_sub;
DROP SCHEMA addr_nsp CASCADE;
diff --git a/src/test/regress/sql/publication.sql b/src/test/regress/sql/publication.sql
index a56387edee..442ac45b94 100644
--- a/src/test/regress/sql/publication.sql
+++ b/src/test/regress/sql/publication.sql
@@ -27,7 +27,7 @@ CREATE PUBLICATION testpub_xxx WITH (publish_via_partition_root = 'true', publis
\dRp
-ALTER PUBLICATION testpub_default SET (publish = 'insert, update, delete');
+ALTER PUBLICATION testpub_default SET (publish = 'insert, update, delete, sequence');
\dRp
@@ -46,6 +46,8 @@ ALTER PUBLICATION testpub_foralltables SET (publish = 'insert, update');
CREATE TABLE testpub_tbl2 (id serial primary key, data text);
-- fail - can't add to for all tables publication
ALTER PUBLICATION testpub_foralltables ADD TABLE testpub_tbl2;
+-- fail - can't add a table using ADD SEQUENCE command
+ALTER PUBLICATION testpub_foralltables ADD SEQUENCE testpub_tbl2;
-- fail - can't drop from all tables publication
ALTER PUBLICATION testpub_foralltables DROP TABLE testpub_tbl2;
-- fail - can't add to for all tables publication
@@ -104,6 +106,199 @@ RESET client_min_messages;
DROP TABLE testpub_tbl3, testpub_tbl3a;
DROP PUBLICATION testpub3, testpub4;
+--- adding sequences
+CREATE SEQUENCE testpub_seq0;
+CREATE SEQUENCE pub_test.testpub_seq1;
+
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_forallsequences FOR ALL SEQUENCES WITH (publish = 'sequence');
+RESET client_min_messages;
+ALTER PUBLICATION testpub_forallsequences SET (publish = 'insert, sequence');
+
+CREATE SEQUENCE testpub_seq2;
+-- fail - can't add to for all sequences publication
+ALTER PUBLICATION testpub_forallsequences ADD SEQUENCE testpub_seq2;
+-- fail - can't drop from all sequences publication
+ALTER PUBLICATION testpub_forallsequences DROP SEQUENCE testpub_seq2;
+-- fail - can't add to for all sequences publication
+ALTER PUBLICATION testpub_forallsequences SET SEQUENCE pub_test.testpub_seq1;
+
+-- fail - can't add schema to 'FOR ALL SEQUENCES' publication
+ALTER PUBLICATION testpub_forallsequences ADD ALL SEQUENCES IN SCHEMA pub_test;
+-- fail - can't drop schema from 'FOR ALL SEQUENCES' publication
+ALTER PUBLICATION testpub_forallsequences DROP ALL SEQUENCES IN SCHEMA pub_test;
+-- fail - can't set schema to 'FOR ALL SEQUENCES' publication
+ALTER PUBLICATION testpub_forallsequences SET ALL SEQUENCES IN SCHEMA pub_test;
+
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_forsequence FOR SEQUENCE testpub_seq0;
+RESET client_min_messages;
+-- should be able to add schema to 'FOR SEQUENCE' publication
+ALTER PUBLICATION testpub_forsequence ADD ALL SEQUENCES IN SCHEMA pub_test;
+\dRp+ testpub_forsequence
+-- fail - can't add sequence from the schema we already added
+ALTER PUBLICATION testpub_forsequence ADD SEQUENCE pub_test.testpub_seq1;
+-- fail - can't add sequence using ADD TABLE command
+ALTER PUBLICATION testpub_forsequence ADD TABLE pub_test.testpub_seq1;
+-- should be able to drop schema from 'FOR SEQUENCE' publication
+ALTER PUBLICATION testpub_forsequence DROP ALL SEQUENCES IN SCHEMA pub_test;
+\dRp+ testpub_forsequence
+-- should be able to set schema to 'FOR SEQUENCE' publication
+ALTER PUBLICATION testpub_forsequence SET ALL SEQUENCES IN SCHEMA pub_test;
+\dRp+ testpub_forsequence
+
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_forschema FOR ALL SEQUENCES IN SCHEMA pub_test;
+RESET client_min_messages;
+-- fail - can't create publication with schema and sequence of the same schema
+CREATE PUBLICATION testpub_for_seq_schema FOR ALL SEQUENCES IN SCHEMA pub_test, SEQUENCE pub_test.testpub_seq1;
+-- fail - can't add a sequence of the same schema to the schema publication
+ALTER PUBLICATION testpub_forschema ADD SEQUENCE pub_test.testpub_seq1;
+-- fail - can't drop a sequence from the schema publication which isn't in the
+-- publication
+ALTER PUBLICATION testpub_forschema DROP SEQUENCE pub_test.testpub_seq1;
+-- should be able to set sequence to schema publication
+ALTER PUBLICATION testpub_forschema SET SEQUENCE pub_test.testpub_seq1;
+\dRp+ testpub_forschema
+
+SELECT pubname, puballtables, puballsequences FROM pg_publication WHERE pubname = 'testpub_forallsequences';
+\d+ pub_test.testpub_seq1
+\dRp+ testpub_forallsequences
+DROP SEQUENCE testpub_seq0, pub_test.testpub_seq1, testpub_seq2;
+DROP PUBLICATION testpub_forallsequences, testpub_forsequence, testpub_forschema;
+
+
+-- publication testing multiple sequences at the same time
+CREATE SEQUENCE testpub_seq1;
+CREATE SEQUENCE testpub_seq2;
+
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_multi FOR SEQUENCE testpub_seq1, testpub_seq2;
+RESET client_min_messages;
+
+\dRp+ testpub_multi
+
+DROP PUBLICATION testpub_multi;
+DROP SEQUENCE testpub_seq1;
+DROP SEQUENCE testpub_seq2;
+
+
+-- Publication mixing tables and sequences
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_mix;
+RESET client_min_messages;
+
+CREATE SEQUENCE testpub_seq1;
+CREATE SEQUENCE pub_test.testpub_seq2;
+
+ALTER PUBLICATION testpub_mix ADD SEQUENCE testpub_seq1, TABLE testpub_tbl1;
+\dRp+ testpub_mix
+
+ALTER PUBLICATION testpub_mix ADD ALL SEQUENCES IN SCHEMA pub_test, ALL TABLES IN SCHEMA pub_test;
+\dRp+ testpub_mix
+
+ALTER PUBLICATION testpub_mix DROP ALL SEQUENCES IN SCHEMA pub_test;
+\dRp+ testpub_mix
+
+ALTER PUBLICATION testpub_mix DROP ALL TABLES IN SCHEMA pub_test;
+\dRp+ testpub_mix
+
+DROP PUBLICATION testpub_mix;
+DROP SEQUENCE testpub_seq1;
+DROP SEQUENCE pub_test.testpub_seq2;
+
+
+-- make sure we replicate only the correct relation type
+CREATE SCHEMA pub_test1;
+CREATE SEQUENCE pub_test1.test_seq1;
+CREATE TABLE pub_test1.test_tbl1 (a int primary key, b int);
+
+CREATE SCHEMA pub_test2;
+CREATE SEQUENCE pub_test2.test_seq2;
+CREATE TABLE pub_test2.test_tbl2 (a int primary key, b int);
+
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_schemas;
+RESET client_min_messages;
+
+-- add tables from one schema, sequences from the other
+ALTER PUBLICATION testpub_schemas ADD ALL TABLES IN SCHEMA pub_test2;
+ALTER PUBLICATION testpub_schemas ADD ALL SEQUENCES IN SCHEMA pub_test1;
+
+\dRp+ testpub_schemas
+
+\dn+ pub_test1
+\dn+ pub_test2
+
+\d+ pub_test1.test_seq1;
+\d+ pub_test1.test_tbl1;
+
+\d+ pub_test2.test_seq2;
+\d+ pub_test2.test_tbl2;
+
+-- add the other object type from each schema
+ALTER PUBLICATION testpub_schemas ADD ALL TABLES IN SCHEMA pub_test1;
+ALTER PUBLICATION testpub_schemas ADD ALL SEQUENCES IN SCHEMA pub_test2;
+
+\dRp+ testpub_schemas
+
+\dn+ pub_test1
+\dn+ pub_test2
+
+\d+ pub_test1.test_seq1;
+\d+ pub_test1.test_tbl1;
+
+\d+ pub_test2.test_seq2;
+\d+ pub_test2.test_tbl2;
+
+-- now drop the object type added first
+ALTER PUBLICATION testpub_schemas DROP ALL TABLES IN SCHEMA pub_test2;
+ALTER PUBLICATION testpub_schemas DROP ALL SEQUENCES IN SCHEMA pub_test1;
+
+\dRp+ testpub_schemas
+
+\dn+ pub_test1
+\dn+ pub_test2
+
+\d+ pub_test1.test_seq1;
+\d+ pub_test1.test_tbl1;
+
+\d+ pub_test2.test_seq2;
+\d+ pub_test2.test_tbl2;
+
+-- should fail (publication contains the whole schema)
+ALTER PUBLICATION testpub_schemas ADD TABLE pub_test1.test_tbl1;
+ALTER PUBLICATION testpub_schemas ADD SEQUENCE pub_test2.test_seq2;
+
+-- should work (different schema)
+ALTER PUBLICATION testpub_schemas ADD TABLE pub_test2.test_tbl2;
+ALTER PUBLICATION testpub_schemas ADD SEQUENCE pub_test1.test_seq1;
+
+\dRp+ testpub_schemas
+
+\d+ pub_test1.test_seq1;
+\d+ pub_test1.test_tbl1;
+
+\d+ pub_test2.test_seq2;
+\d+ pub_test2.test_tbl2;
+
+-- now drop the explicitly added objects again
+ALTER PUBLICATION testpub_schemas DROP TABLE pub_test2.test_tbl2;
+ALTER PUBLICATION testpub_schemas DROP SEQUENCE pub_test1.test_seq1;
+
+\dRp+ testpub_schemas
+
+\d+ pub_test1.test_seq1;
+\d+ pub_test1.test_tbl1;
+
+\d+ pub_test2.test_seq2;
+\d+ pub_test2.test_tbl2;
+
+DROP PUBLICATION testpub_schemas;
+DROP TABLE pub_test1.test_tbl1, pub_test2.test_tbl2;
+DROP SEQUENCE pub_test1.test_seq1, pub_test2.test_seq2;
+DROP SCHEMA pub_test1, pub_test2;
+
-- Tests for partitioned tables
SET client_min_messages = 'ERROR';
CREATE PUBLICATION testpub_forparted;
@@ -1014,32 +1209,51 @@ CREATE SCHEMA sch1;
CREATE SCHEMA sch2;
CREATE TABLE sch1.tbl1 (a int) PARTITION BY RANGE(a);
CREATE TABLE sch2.tbl1_part1 PARTITION OF sch1.tbl1 FOR VALUES FROM (1) to (10);
+CREATE SEQUENCE sch1.seq1;
+CREATE SEQUENCE sch2.seq2;
-- Schema publication that does not include the schema that has the parent table
CREATE PUBLICATION pub FOR ALL TABLES IN SCHEMA sch2 WITH (PUBLISH_VIA_PARTITION_ROOT=1);
+ALTER PUBLICATION pub ADD ALL SEQUENCES IN SCHEMA sch2;
SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
DROP PUBLICATION pub;
-- Table publication that does not include the parent table
CREATE PUBLICATION pub FOR TABLE sch2.tbl1_part1 WITH (PUBLISH_VIA_PARTITION_ROOT=1);
+ALTER PUBLICATION pub ADD SEQUENCE sch2.seq2;
SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
-- Table publication that includes both the parent table and the child table
ALTER PUBLICATION pub ADD TABLE sch1.tbl1;
+ALTER PUBLICATION pub ADD SEQUENCE sch1.seq1;
SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
DROP PUBLICATION pub;
-- Schema publication that does not include the schema that has the parent table
CREATE PUBLICATION pub FOR ALL TABLES IN SCHEMA sch2 WITH (PUBLISH_VIA_PARTITION_ROOT=0);
+ALTER PUBLICATION pub ADD SEQUENCE sch1.seq1;
+SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
+
+DROP PUBLICATION pub;
+-- Sequence publication
+CREATE PUBLICATION pub FOR SEQUENCE sch2.seq2;
SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
DROP PUBLICATION pub;
-- Table publication that does not include the parent table
CREATE PUBLICATION pub FOR TABLE sch2.tbl1_part1 WITH (PUBLISH_VIA_PARTITION_ROOT=0);
SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
-- Table publication that includes both the parent table and the child table
ALTER PUBLICATION pub ADD TABLE sch1.tbl1;
+ALTER PUBLICATION pub ADD ALL SEQUENCES IN SCHEMA sch2;
SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
DROP PUBLICATION pub;
DROP TABLE sch2.tbl1_part1;
@@ -1052,10 +1266,36 @@ CREATE TABLE sch1.tbl1_part3 (a int) PARTITION BY RANGE(a);
ALTER TABLE sch1.tbl1 ATTACH PARTITION sch1.tbl1_part3 FOR VALUES FROM (20) to (30);
CREATE PUBLICATION pub FOR ALL TABLES IN SCHEMA sch1 WITH (PUBLISH_VIA_PARTITION_ROOT=1);
SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
+
+DROP PUBLICATION pub;
+-- Schema publication
+CREATE PUBLICATION pub FOR SEQUENCE sch2.seq2;
+SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
+
+DROP PUBLICATION pub;
+-- Sequence publication
+CREATE PUBLICATION pub FOR ALL SEQUENCES IN SCHEMA sch2;
+SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
+
+ALTER PUBLICATION pub ADD SEQUENCE sch1.seq1;
+SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
+
+ALTER PUBLICATION pub DROP SEQUENCE sch1.seq1;
+SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
+
+ALTER PUBLICATION pub ADD ALL SEQUENCES IN SCHEMA sch1;
+SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
RESET client_min_messages;
DROP PUBLICATION pub;
DROP TABLE sch1.tbl1;
+DROP SEQUENCE sch1.seq1, sch2.seq2;
DROP SCHEMA sch1 cascade;
DROP SCHEMA sch2 cascade;
diff --git a/src/test/subscription/t/030_sequences.pl b/src/test/subscription/t/030_sequences.pl
new file mode 100644
index 0000000000..9ae3c03d7d
--- /dev/null
+++ b/src/test/subscription/t/030_sequences.pl
@@ -0,0 +1,202 @@
+
+# Copyright (c) 2021, PostgreSQL Global Development Group
+
+# This tests that sequences are replicated correctly by logical replication
+use strict;
+use warnings;
+use PostgreSQL::Test::Cluster;
+use PostgreSQL::Test::Utils;
+use Test::More;
+
+# Initialize publisher node
+my $node_publisher = PostgreSQL::Test::Cluster->new('publisher');
+$node_publisher->init(allows_streaming => 'logical');
+$node_publisher->start;
+
+# Create subscriber node
+my $node_subscriber = PostgreSQL::Test::Cluster->new('subscriber');
+$node_subscriber->init(allows_streaming => 'logical');
+$node_subscriber->start;
+
+# Create some preexisting content on publisher
+my $ddl = qq(
+ CREATE TABLE seq_test (v BIGINT);
+ CREATE SEQUENCE s;
+);
+
+# Setup structure on the publisher
+$node_publisher->safe_psql('postgres', $ddl);
+
+# Create some the same structure on subscriber, and an extra sequence that
+# we'll create on the publisher later
+$ddl = qq(
+ CREATE TABLE seq_test (v BIGINT);
+ CREATE SEQUENCE s;
+ CREATE SEQUENCE s2;
+);
+
+$node_subscriber->safe_psql('postgres', $ddl);
+
+# Setup logical replication
+my $publisher_connstr = $node_publisher->connstr . ' dbname=postgres';
+$node_publisher->safe_psql('postgres',
+ "CREATE PUBLICATION seq_pub");
+
+$node_publisher->safe_psql('postgres',
+ "ALTER PUBLICATION seq_pub ADD SEQUENCE s");
+
+$node_subscriber->safe_psql('postgres',
+ "CREATE SUBSCRIPTION seq_sub CONNECTION '$publisher_connstr' PUBLICATION seq_pub"
+);
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Wait for initial sync to finish as well
+my $synced_query =
+ "SELECT count(1) = 0 FROM pg_subscription_rel WHERE srsubstate NOT IN ('s', 'r');";
+$node_subscriber->poll_query_until('postgres', $synced_query)
+ or die "Timed out while waiting for subscriber to synchronize data";
+
+# Insert initial test data
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ -- generate a number of values using the sequence
+ INSERT INTO seq_test SELECT nextval('s') FROM generate_series(1,100);
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+my $result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s;
+));
+
+is( $result, '132|0|t',
+ 'initial test data replicated');
+
+
+# advance the sequence in a rolled-back transaction - the rollback
+# does not wait for the replication, so we could see any intermediate state
+# so do something else after the test, to ensure we wait for everything
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ INSERT INTO seq_test SELECT nextval('s') FROM generate_series(1,100);
+ ROLLBACK;
+ INSERT INTO seq_test VALUES (-1);
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s;
+));
+
+is( $result, '231|0|t',
+ 'advance sequence in rolled-back transaction');
+
+
+# create a new sequence and roll it back - should not be replicated, due to
+# the transactional behavior
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ CREATE SEQUENCE s2;
+ ALTER PUBLICATION seq_pub ADD SEQUENCE s2;
+ INSERT INTO seq_test SELECT nextval('s2') FROM generate_series(1,100);
+ ROLLBACK;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s2;
+));
+
+is( $result, '1|0|f',
+ 'create new sequence and roll it back');
+
+
+# create a new sequence, advance it in a rolled-back transaction, but commit
+# the create - the advance should be replicated nevertheless
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ CREATE SEQUENCE s2;
+ ALTER PUBLICATION seq_pub ADD SEQUENCE s2;
+ SAVEPOINT sp1;
+ INSERT INTO seq_test SELECT nextval('s2') FROM generate_series(1,100);
+ ROLLBACK TO sp1;
+ COMMIT;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Wait for sync of the second sequence we just added to finish
+$synced_query =
+ "SELECT count(1) = 0 FROM pg_subscription_rel WHERE srsubstate NOT IN ('s', 'r');";
+$node_subscriber->poll_query_until('postgres', $synced_query)
+ or die "Timed out while waiting for subscriber to synchronize data";
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s2;
+));
+
+is( $result, '132|0|t',
+ 'create sequence, advance it in rolled-back transaction, but commit the create');
+
+
+# advance the new sequence in a transaction, and roll it back - the rollback
+# does not wait for the replication, so we could see any intermediate state
+# so do something else after the test, to ensure we wait for everything
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ INSERT INTO seq_test SELECT nextval('s2') FROM generate_series(1,100);
+ ROLLBACK;
+ INSERT INTO seq_test VALUES (-1);
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s2;
+));
+
+is( $result, '231|0|t',
+ 'advance the new sequence in a transaction and roll it back');
+
+
+# advance the sequence in a subtransaction - the subtransaction gets rolled
+# back, but commit the main one - the changes should still be replicated
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ SAVEPOINT s1;
+ INSERT INTO seq_test SELECT nextval('s2') FROM generate_series(1,100);
+ ROLLBACK TO s1;
+ COMMIT;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s2;
+));
+
+is( $result, '330|0|t',
+ 'advance sequence in a subtransaction');
+
+
+done_testing();
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2022-08-19 11:11 Tomas Vondra <[email protected]>
parent: Tomas Vondra <[email protected]>
0 siblings, 2 replies; 127+ messages in thread
From: Tomas Vondra @ 2022-08-19 11:11 UTC (permalink / raw)
To: PostgreSQL Hackers <[email protected]>
I've been thinking about the two optimizations mentioned at the end a
bit more, so let me share my thoughts before I forget that:
On 8/18/22 23:10, Tomas Vondra wrote:
>
> ...
>
> And maybe we could then use the LSN to read the increment from the WAL
> during decoding, instead of having to read it and WAL-log it during
> commit. Essentially, we'd run a local XLogReader. Of course, we'd have
> to be careful about checkpoints, not sure what to do about that.
>
I think logging just the LSN is workable.
I was worried about dealing with checkpoints, because imagine you do
nextval() on sequence that was last WAL-logged a couple checkpoints
back. Then you wouldn't be able to read the LSN (when decoding), because
the WAL might have been recycled. But that can't happen, because we
always force WAL-logging the first time nextval() is called after a
checkpoint. So we know the LSN is guaranteed to be available.
Of course, this would not reduce the amount of WAL messages, because
we'd still log all sequences touched by the transaction. We wouldn't
need to read the state from disk, though, and we could ignore "old"
stuff in decoding (with LSN lower than the last LSN we decoded).
For frequently used sequences that seems like a win.
> Another idea that just occurred to me is that if we end up having to
> read the sequence state during commit, maybe we could at least optimize
> it somehow. For example we might track LSN of the last logged state for
> each sequence (in shared memory or something), and the other sessions
> could just skip the WAL-log if their "local" LSN is <= than this LSN.
>
Tracking the last LSN for each sequence (in a SLRU or something) should
work too, I guess. In principle this just moves the skipping of "old"
increments from decoding to writing, so that we don't even have to write
those into WAL.
We don't even need persistence, nor to keep all the records, I think. If
you don't find a record for a given sequence, assume it wasn't logged
yet and just log it. Of course, it requires a bit of shared memory for
each sequence, say ~32B. Not sure about the overhead, but I'd bet if you
have many (~thousands) frequently used sequences, there'll be a lot of
other overhead making this irrelevant.
Of course, if we're doing the skipping when writing the WAL, maybe we
should just read the sequence state - we'd do the I/O, but only in
fraction of the transactions, and we wouldn't need to read old WAL in
logical decoding.
regards
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2022-08-23 04:21 Michael Paquier <[email protected]>
parent: Thomas Munro <[email protected]>
1 sibling, 1 reply; 127+ messages in thread
From: Michael Paquier @ 2022-08-23 04:21 UTC (permalink / raw)
To: Thomas Munro <[email protected]>; +Cc: Noah Misch <[email protected]>; Tomas Vondra <[email protected]>; Amit Kapila <[email protected]>; Petr Jelinek <[email protected]>; Peter Eisentraut <[email protected]>; PostgreSQL Hackers <[email protected]>
On Mon, Aug 08, 2022 at 06:15:46PM +1200, Thomas Munro wrote:
> The attached patch that simply moves the cache invalidation into
> report_invalid_record(), so that it's reached by the above code and
> everything else that reports an error, seems to fix the problem in
> src/bin/pg_ctl/t/003_promote.pl with Noah's spanner-in-the-works patch
> applied, and passes check-world without it. I need to look at this
> some more, though, and figure out if it's the right fix.
Thomas, where are you on this open item? A potential PANIC at
promotion is bad. One possible exit path would be to switch the
default of recovery_prefetch, though that's a kind of last-resort
option seen from here.
--
Michael
Attachments:
[application/pgp-signature] signature.asc (833B, ../../[email protected]/2-signature.asc)
download
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2022-08-23 04:32 Thomas Munro <[email protected]>
parent: Michael Paquier <[email protected]>
0 siblings, 1 reply; 127+ messages in thread
From: Thomas Munro @ 2022-08-23 04:32 UTC (permalink / raw)
To: Michael Paquier <[email protected]>; +Cc: Noah Misch <[email protected]>; Tomas Vondra <[email protected]>; Amit Kapila <[email protected]>; Petr Jelinek <[email protected]>; Peter Eisentraut <[email protected]>; PostgreSQL Hackers <[email protected]>
On Tue, Aug 23, 2022 at 4:21 PM Michael Paquier <[email protected]> wrote:
> On Mon, Aug 08, 2022 at 06:15:46PM +1200, Thomas Munro wrote:
> > The attached patch that simply moves the cache invalidation into
> > report_invalid_record(), so that it's reached by the above code and
> > everything else that reports an error, seems to fix the problem in
> > src/bin/pg_ctl/t/003_promote.pl with Noah's spanner-in-the-works patch
> > applied, and passes check-world without it. I need to look at this
> > some more, though, and figure out if it's the right fix.
>
> Thomas, where are you on this open item? A potential PANIC at
> promotion is bad. One possible exit path would be to switch the
> default of recovery_prefetch, though that's a kind of last-resort
> option seen from here.
I will get a fix committed this week -- I need to study
Horiguchi-san's analysis...
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2022-08-23 15:04 Robert Haas <[email protected]>
parent: Thomas Munro <[email protected]>
0 siblings, 1 reply; 127+ messages in thread
From: Robert Haas @ 2022-08-23 15:04 UTC (permalink / raw)
To: Thomas Munro <[email protected]>; +Cc: Michael Paquier <[email protected]>; Noah Misch <[email protected]>; Tomas Vondra <[email protected]>; Amit Kapila <[email protected]>; Petr Jelinek <[email protected]>; Peter Eisentraut <[email protected]>; PostgreSQL Hackers <[email protected]>
On Tue, Aug 23, 2022 at 12:33 AM Thomas Munro <[email protected]> wrote:
> On Tue, Aug 23, 2022 at 4:21 PM Michael Paquier <[email protected]> wrote:
> > On Mon, Aug 08, 2022 at 06:15:46PM +1200, Thomas Munro wrote:
> > > The attached patch that simply moves the cache invalidation into
> > > report_invalid_record(), so that it's reached by the above code and
> > > everything else that reports an error, seems to fix the problem in
> > > src/bin/pg_ctl/t/003_promote.pl with Noah's spanner-in-the-works patch
> > > applied, and passes check-world without it. I need to look at this
> > > some more, though, and figure out if it's the right fix.
> >
> > Thomas, where are you on this open item? A potential PANIC at
> > promotion is bad. One possible exit path would be to switch the
> > default of recovery_prefetch, though that's a kind of last-resort
> > option seen from here.
>
> I will get a fix committed this week -- I need to study
> Horiguchi-san's analysis...
Hi!
I haven't been paying attention to this thread, but my attention was
just drawn to it, and I'm wondering if the issue you're trying to
track down here is actually the same as what I reported yesterday
here:
https://www.postgresql.org/message-id/[email protected]...
--
Robert Haas
EDB: http://www.enterprisedb.com
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2022-08-24 00:51 Thomas Munro <[email protected]>
parent: Robert Haas <[email protected]>
0 siblings, 0 replies; 127+ messages in thread
From: Thomas Munro @ 2022-08-24 00:51 UTC (permalink / raw)
To: Robert Haas <[email protected]>; +Cc: Michael Paquier <[email protected]>; Noah Misch <[email protected]>; Tomas Vondra <[email protected]>; Amit Kapila <[email protected]>; Petr Jelinek <[email protected]>; Peter Eisentraut <[email protected]>; PostgreSQL Hackers <[email protected]>
On Wed, Aug 24, 2022 at 3:04 AM Robert Haas <[email protected]> wrote:
> I haven't been paying attention to this thread, but my attention was
> just drawn to it, and I'm wondering if the issue you're trying to
> track down here is actually the same as what I reported yesterday
> here:
>
> https://www.postgresql.org/message-id/[email protected]...
Summarising a chat we had about this: Different bug, similar
ingredients. Robert describes a screw-up in what is written, but here
we're talking about a cache invalidation bug while reading.
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2022-08-29 10:21 Thomas Munro <[email protected]>
parent: Kyotaro Horiguchi <[email protected]>
0 siblings, 1 reply; 127+ messages in thread
From: Thomas Munro @ 2022-08-29 10:21 UTC (permalink / raw)
To: Kyotaro Horiguchi <[email protected]>; +Cc: Noah Misch <[email protected]>; Tomas Vondra <[email protected]>; Amit Kapila <[email protected]>; Petr Jelinek <[email protected]>; Peter Eisentraut <[email protected]>; PostgreSQL Hackers <[email protected]>
On Mon, Aug 8, 2022 at 8:56 PM Kyotaro Horiguchi
<[email protected]> wrote:
> At Mon, 08 Aug 2022 17:33:22 +0900 (JST), Kyotaro Horiguchi <[email protected]> wrote in
> > If WaitForWALToBecomeAvailable returned by promotion, ReadPageInteral
> > misses the chance to inavlidate reader-state. That state is not an
> > error while in StandbyMode.
>
> Mmm... Maybe I wanted to say: (Still I'm not sure the rewrite works..)
>
> If WaitForWALToBecomeAvailable returned by promotion, ReadPageInteral
> would miss the chance to invalidate reader-state. When XLogPageRead
> is called in blocking mode while in StandbyMode (that is, the
> traditional condition) , the function continues retrying until it
> succeeds, or returns XLRAD_FAIL if promote is triggered. In other
> words, it was not supposed to return non-failure while the header
> validation is failing while in standby mode. But while in nonblocking
> mode, the function can return non-failure with lastSourceFailed =
> true, which seems wrong.
New ideas:
0001: Instead of figuring out when to invalidate the cache, let's
just invalidate it before every read attempt. It is only marked valid
after success (ie state->readLen > 0). No need to worry about error
cases.
0002: While here, I don't like xlogrecovery.c clobbering
xlogreader.c's internal error state, so I think we should have a
function for that with a documented purpose. It was also a little
inconsistent that it didn't clear a flag (but not buggy AFAICS; kinda
wondering if I should just get rid of that flag, but that's for
another day).
0003: Thinking about your comments above made me realise that I don't
really want XLogReadPage() to be internally retrying for obscure
failures while reading ahead. I think I prefer to give up on
prefetching as soon as anything tricky happens, and deal with
complexities once recovery catches up to that point. I am still
thinking about this point.
Here's the patch set I'm testing.
Attachments:
[text/x-patch] 0001-Fix-cache-invalidation-in-rare-recovery_prefetch-cas.patch (2.1K, ../../CA+hUKGJOnWuQqj42Q8xKfWN9NxGYUvEDEvEd_a+9Y9dF=78Oyw@mail.gmail.com/2-0001-Fix-cache-invalidation-in-rare-recovery_prefetch-cas.patch)
download | inline diff:
From 9f547088e54fcf279c74fbd03e4afaeffec69ea3 Mon Sep 17 00:00:00 2001
From: Thomas Munro <[email protected]>
Date: Mon, 29 Aug 2022 19:56:27 +1200
Subject: [PATCH 1/3] Fix cache invalidation in rare recovery_prefetch case.
Since commit 0668719801 we've had an extra page validation and retry
loop inside XLogPageRead() to handle a rare special case. The new
recovery prefetch code from commit 5dc0418f could lead to a PANIC in
this case, because the WAL page cached internally by xlogreader.c was
not invalidated when it should have been.
Make tracking of the cached page's status more robust by invalidating it
just before we attempt any read. Because it is only marked valid when a
read succeeds, now we don't have to worry about invalidation in error
code paths.
The removed line that was setting errormsg_deferred was redundant
because it's already set by report_invalid_record().
Back-patch to 15.
Reported-by: Noah Misch <[email protected]>
Reviewed-by:
Discussion: https://postgr.es/m/20220807003627.GA4168930%40rfd.leadboat.com
---
src/backend/access/transam/xlogreader.c | 12 +++++++-----
1 file changed, 7 insertions(+), 5 deletions(-)
diff --git a/src/backend/access/transam/xlogreader.c b/src/backend/access/transam/xlogreader.c
index f17e80948d..e883ade607 100644
--- a/src/backend/access/transam/xlogreader.c
+++ b/src/backend/access/transam/xlogreader.c
@@ -987,6 +987,13 @@ ReadPageInternal(XLogReaderState *state, XLogRecPtr pageptr, int reqLen)
targetPageOff == state->segoff && reqLen <= state->readLen)
return state->readLen;
+ /*
+ * Invalidate contents of internal buffer before read attempt. Just set
+ * the length to 0, rather than a full XLogReaderInvalReadState(), so we
+ * don't forget the segment we last read.
+ */
+ state->readLen = 0;
+
/*
* Data is not in our buffer.
*
@@ -1067,11 +1074,6 @@ ReadPageInternal(XLogReaderState *state, XLogRecPtr pageptr, int reqLen)
return readLen;
err:
- if (state->errormsg_buf[0] != '\0')
- {
- state->errormsg_deferred = true;
- XLogReaderInvalReadState(state);
- }
return XLREAD_FAIL;
}
--
2.30.2
[text/x-patch] 0002-Improve-xlogrecovery.c-xlogreader.c-boundary.patch (2.5K, ../../CA+hUKGJOnWuQqj42Q8xKfWN9NxGYUvEDEvEd_a+9Y9dF=78Oyw@mail.gmail.com/3-0002-Improve-xlogrecovery.c-xlogreader.c-boundary.patch)
download | inline diff:
From 72c55dfec6729e0f26a9fbf3ff184d3b4308a025 Mon Sep 17 00:00:00 2001
From: Thomas Munro <[email protected]>
Date: Mon, 29 Aug 2022 21:08:25 +1200
Subject: [PATCH 2/3] Improve xlogrecovery.c/xlogreader.c boundary.
Create a function XLogReaderResetError() to clobber the error message
inside an XLogReaderState, instead of doing it directly from
xlogrecovery.c. This is older code from commit 0668719801, but commit
5dc0418f probably should have tidied this up because it leaves the
internal state a bit out of sync (not a live bug, but a little
confusing).
---
src/backend/access/transam/xlogreader.c | 10 ++++++++++
src/backend/access/transam/xlogrecovery.c | 2 +-
src/include/access/xlogreader.h | 3 +++
3 files changed, 14 insertions(+), 1 deletion(-)
diff --git a/src/backend/access/transam/xlogreader.c b/src/backend/access/transam/xlogreader.c
index e883ade607..c100e18333 100644
--- a/src/backend/access/transam/xlogreader.c
+++ b/src/backend/access/transam/xlogreader.c
@@ -1325,6 +1325,16 @@ XLogReaderValidatePageHeader(XLogReaderState *state, XLogRecPtr recptr,
return true;
}
+/*
+ * Forget about an error produced by XLogReaderValidatePageHeader().
+ */
+void
+XLogReaderResetError(XLogReaderState *state)
+{
+ state->errormsg_buf[0] = '\0';
+ state->errormsg_deferred = false;
+}
+
/*
* Find the first record with an lsn >= RecPtr.
*
diff --git a/src/backend/access/transam/xlogrecovery.c b/src/backend/access/transam/xlogrecovery.c
index a59a0e826b..422322cf5a 100644
--- a/src/backend/access/transam/xlogrecovery.c
+++ b/src/backend/access/transam/xlogrecovery.c
@@ -3335,7 +3335,7 @@ retry:
(errmsg_internal("%s", xlogreader->errormsg_buf)));
/* reset any error XLogReaderValidatePageHeader() might have set */
- xlogreader->errormsg_buf[0] = '\0';
+ XLogReaderResetError(xlogreader);
goto next_record_is_invalid;
}
diff --git a/src/include/access/xlogreader.h b/src/include/access/xlogreader.h
index 87ff00feb7..97b6f00114 100644
--- a/src/include/access/xlogreader.h
+++ b/src/include/access/xlogreader.h
@@ -373,6 +373,9 @@ extern DecodedXLogRecord *XLogReadAhead(XLogReaderState *state,
extern bool XLogReaderValidatePageHeader(XLogReaderState *state,
XLogRecPtr recptr, char *phdr);
+/* Forget error produced by XLogReaderValidatePageHeader(). */
+extern void XLogReaderResetError(XLogReaderState *state);
+
/*
* Error information from WALRead that both backend and frontend caller can
* process. Currently only errors from pread can be reported.
--
2.30.2
[text/x-patch] 0003-Don-t-retry-inside-XLogPageRead-if-reading-ahead.patch (1.2K, ../../CA+hUKGJOnWuQqj42Q8xKfWN9NxGYUvEDEvEd_a+9Y9dF=78Oyw@mail.gmail.com/4-0003-Don-t-retry-inside-XLogPageRead-if-reading-ahead.patch)
download | inline diff:
From bb8a6d52ca91d463993c3e01a48fc776297b0aeb Mon Sep 17 00:00:00 2001
From: Thomas Munro <[email protected]>
Date: Mon, 29 Aug 2022 21:14:25 +1200
Subject: [PATCH 3/3] Don't retry inside XLogPageRead() if reading ahead.
Give up immediately if we have a short read or a page header
invalidation failure inside XLogPageRead(). Once recovery catches up to
this point, it can deal with it. This doesn't fix any known live bug,
but it fits with the general philosophy of giving up fast when
prefetching runs into trouble.
---
src/backend/access/transam/xlogrecovery.c | 8 ++++++++
1 file changed, 8 insertions(+)
diff --git a/src/backend/access/transam/xlogrecovery.c b/src/backend/access/transam/xlogrecovery.c
index 422322cf5a..76ce9704ca 100644
--- a/src/backend/access/transam/xlogrecovery.c
+++ b/src/backend/access/transam/xlogrecovery.c
@@ -3342,6 +3342,14 @@ retry:
return readLen;
next_record_is_invalid:
+ /*
+ * If we're reading ahead, give up fast. Standby retries and error
+ * reporting will be handled by a later read when recovery catches up to
+ * this point.
+ */
+ if (xlogreader->nonblocking)
+ return XLREAD_WOULDBLOCK;
+
lastSourceFailed = true;
if (readFile >= 0)
--
2.30.2
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2022-08-29 18:04 Tomas Vondra <[email protected]>
parent: Thomas Munro <[email protected]>
0 siblings, 1 reply; 127+ messages in thread
From: Tomas Vondra @ 2022-08-29 18:04 UTC (permalink / raw)
To: Thomas Munro <[email protected]>; Kyotaro Horiguchi <[email protected]>; +Cc: Noah Misch <[email protected]>; Amit Kapila <[email protected]>; Petr Jelinek <[email protected]>; Peter Eisentraut <[email protected]>; PostgreSQL Hackers <[email protected]>
On 8/29/22 12:21, Thomas Munro wrote:
> On Mon, Aug 8, 2022 at 8:56 PM Kyotaro Horiguchi
> <[email protected]> wrote:
>> At Mon, 08 Aug 2022 17:33:22 +0900 (JST), Kyotaro Horiguchi <[email protected]> wrote in
>>> If WaitForWALToBecomeAvailable returned by promotion, ReadPageInteral
>>> misses the chance to inavlidate reader-state. That state is not an
>>> error while in StandbyMode.
>>
>> Mmm... Maybe I wanted to say: (Still I'm not sure the rewrite works..)
>>
>> If WaitForWALToBecomeAvailable returned by promotion, ReadPageInteral
>> would miss the chance to invalidate reader-state. When XLogPageRead
>> is called in blocking mode while in StandbyMode (that is, the
>> traditional condition) , the function continues retrying until it
>> succeeds, or returns XLRAD_FAIL if promote is triggered. In other
>> words, it was not supposed to return non-failure while the header
>> validation is failing while in standby mode. But while in nonblocking
>> mode, the function can return non-failure with lastSourceFailed =
>> true, which seems wrong.
>
> New ideas:
>
> 0001: Instead of figuring out when to invalidate the cache, let's
> just invalidate it before every read attempt. It is only marked valid
> after success (ie state->readLen > 0). No need to worry about error
> cases.
>
Maybe I misunderstand how all this works, but won't this have a really
bad performance impact. If not, why do we need the cache at all?
regards
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences
@ 2022-08-29 21:42 Thomas Munro <[email protected]>
parent: Tomas Vondra <[email protected]>
0 siblings, 0 replies; 127+ messages in thread
From: Thomas Munro @ 2022-08-29 21:42 UTC (permalink / raw)
To: Tomas Vondra <[email protected]>; +Cc: Kyotaro Horiguchi <[email protected]>; Noah Misch <[email protected]>; Amit Kapila <[email protected]>; Petr Jelinek <[email protected]>; Peter Eisentraut <[email protected]>; PostgreSQL Hackers <[email protected]>
On Tue, Aug 30, 2022 at 6:04 AM Tomas Vondra
<[email protected]> wrote:
> On 8/29/22 12:21, Thomas Munro wrote:
> > 0001: Instead of figuring out when to invalidate the cache, let's
> > just invalidate it before every read attempt. It is only marked valid
> > after success (ie state->readLen > 0). No need to worry about error
> > cases.
>
> Maybe I misunderstand how all this works, but won't this have a really
> bad performance impact. If not, why do we need the cache at all?
It's a bit confusing because there are several levels of "read". The
cache remains valid as long as the caller of ReadPageInternal() keeps
asking for data that is in range (see early return after comment "/*
check whether we have all the requested data already */"). As soon as
the caller asks for something not in range, this patch marks the cache
invalid before calling the page_read() callback (= XLogPageRead()).
It is only marked valid again after that succeeds. Here's a new
version with no code change, just a better commit message to try to
explain that more clearly.
Attachments:
[text/x-patch] v2-0001-Fix-cache-invalidation-in-rare-recovery_prefetch-.patch (2.5K, ../../CA+hUKG+7z9L-xHwbZWvWkZcrn0KFdtfV5w3LsPdSJN1TrV9N6Q@mail.gmail.com/2-v2-0001-Fix-cache-invalidation-in-rare-recovery_prefetch-.patch)
download | inline diff:
From d02d851ab515d4a1f883f657bc43a9cae600a5d8 Mon Sep 17 00:00:00 2001
From: Thomas Munro <[email protected]>
Date: Mon, 29 Aug 2022 19:56:27 +1200
Subject: [PATCH v2 1/3] Fix cache invalidation in rare recovery_prefetch
cases.
XLogPageRead() can retry internally in rare cases after a read has
succeeded, overwriting the page buffer. Namely, short reads, and page
validation failures while in standby mode (see commit 0668719801). Due
to an oversight in commit 3f1ce973, these cases could leave stale data
in the internal cache of xlogreader.c without marking it invalid. The
main defense against stale cached data was in the error handling path of
ReadPageInternal(), but that wasn't quite enough for errors handled
internally by XLogPageRead()'s retry loop.
Instead of trying to invalidate the cache in the relevant failure cases,
ReadPageInternal() now invalidates it before even calling the page_read
callback (ie XLogPageRead()). It is only marked valid by setting
state->readLen etc after page_read() succeeds. It remains valid until
the caller of ReadPageInternal() eventually asks for something out of
range and we have to go back to page_read() for more.
The removed line that was setting errormsg_deferred was redundant
because it's already set by report_invalid_record().
Back-patch to 15.
Reported-by: Noah Misch <[email protected]>
Reviewed-by:
Discussion: https://postgr.es/m/20220807003627.GA4168930%40rfd.leadboat.com
---
src/backend/access/transam/xlogreader.c | 12 +++++++-----
1 file changed, 7 insertions(+), 5 deletions(-)
diff --git a/src/backend/access/transam/xlogreader.c b/src/backend/access/transam/xlogreader.c
index f17e80948d..e883ade607 100644
--- a/src/backend/access/transam/xlogreader.c
+++ b/src/backend/access/transam/xlogreader.c
@@ -987,6 +987,13 @@ ReadPageInternal(XLogReaderState *state, XLogRecPtr pageptr, int reqLen)
targetPageOff == state->segoff && reqLen <= state->readLen)
return state->readLen;
+ /*
+ * Invalidate contents of internal buffer before read attempt. Just set
+ * the length to 0, rather than a full XLogReaderInvalReadState(), so we
+ * don't forget the segment we last read.
+ */
+ state->readLen = 0;
+
/*
* Data is not in our buffer.
*
@@ -1067,11 +1074,6 @@ ReadPageInternal(XLogReaderState *state, XLogRecPtr pageptr, int reqLen)
return readLen;
err:
- if (state->errormsg_buf[0] != '\0')
- {
- state->errormsg_deferred = true;
- XLogReaderInvalReadState(state);
- }
return XLREAD_FAIL;
}
--
2.30.2
[text/x-patch] v2-0002-Improve-xlogrecovery.c-xlogreader.c-boundary.patch (2.5K, ../../CA+hUKG+7z9L-xHwbZWvWkZcrn0KFdtfV5w3LsPdSJN1TrV9N6Q@mail.gmail.com/3-v2-0002-Improve-xlogrecovery.c-xlogreader.c-boundary.patch)
download | inline diff:
From 6439f5ed5b9cb6524c307f06a43c9c54f51d3b86 Mon Sep 17 00:00:00 2001
From: Thomas Munro <[email protected]>
Date: Mon, 29 Aug 2022 21:08:25 +1200
Subject: [PATCH v2 2/3] Improve xlogrecovery.c/xlogreader.c boundary.
Create a function XLogReaderResetError() to clobber the error message
inside an XLogReaderState, instead of doing it directly from
xlogrecovery.c. This is older code from commit 0668719801, but commit
5dc0418f probably should have tidied this up because it leaves the
internal state a bit out of sync (not a live bug, but a little
confusing).
---
src/backend/access/transam/xlogreader.c | 10 ++++++++++
src/backend/access/transam/xlogrecovery.c | 2 +-
src/include/access/xlogreader.h | 3 +++
3 files changed, 14 insertions(+), 1 deletion(-)
diff --git a/src/backend/access/transam/xlogreader.c b/src/backend/access/transam/xlogreader.c
index e883ade607..c100e18333 100644
--- a/src/backend/access/transam/xlogreader.c
+++ b/src/backend/access/transam/xlogreader.c
@@ -1325,6 +1325,16 @@ XLogReaderValidatePageHeader(XLogReaderState *state, XLogRecPtr recptr,
return true;
}
+/*
+ * Forget about an error produced by XLogReaderValidatePageHeader().
+ */
+void
+XLogReaderResetError(XLogReaderState *state)
+{
+ state->errormsg_buf[0] = '\0';
+ state->errormsg_deferred = false;
+}
+
/*
* Find the first record with an lsn >= RecPtr.
*
diff --git a/src/backend/access/transam/xlogrecovery.c b/src/backend/access/transam/xlogrecovery.c
index a59a0e826b..422322cf5a 100644
--- a/src/backend/access/transam/xlogrecovery.c
+++ b/src/backend/access/transam/xlogrecovery.c
@@ -3335,7 +3335,7 @@ retry:
(errmsg_internal("%s", xlogreader->errormsg_buf)));
/* reset any error XLogReaderValidatePageHeader() might have set */
- xlogreader->errormsg_buf[0] = '\0';
+ XLogReaderResetError(xlogreader);
goto next_record_is_invalid;
}
diff --git a/src/include/access/xlogreader.h b/src/include/access/xlogreader.h
index 87ff00feb7..97b6f00114 100644
--- a/src/include/access/xlogreader.h
+++ b/src/include/access/xlogreader.h
@@ -373,6 +373,9 @@ extern DecodedXLogRecord *XLogReadAhead(XLogReaderState *state,
extern bool XLogReaderValidatePageHeader(XLogReaderState *state,
XLogRecPtr recptr, char *phdr);
+/* Forget error produced by XLogReaderValidatePageHeader(). */
+extern void XLogReaderResetError(XLogReaderState *state);
+
/*
* Error information from WALRead that both backend and frontend caller can
* process. Currently only errors from pread can be reported.
--
2.30.2
[text/x-patch] v2-0003-Don-t-retry-inside-XLogPageRead-if-reading-ahead.patch (1.2K, ../../CA+hUKG+7z9L-xHwbZWvWkZcrn0KFdtfV5w3LsPdSJN1TrV9N6Q@mail.gmail.com/4-v2-0003-Don-t-retry-inside-XLogPageRead-if-reading-ahead.patch)
download | inline diff:
From 394d2b08a896a7e3094ffb052adaa11486b7b14f Mon Sep 17 00:00:00 2001
From: Thomas Munro <[email protected]>
Date: Mon, 29 Aug 2022 21:14:25 +1200
Subject: [PATCH v2 3/3] Don't retry inside XLogPageRead() if reading ahead.
Give up immediately if we have a short read or a page header
invalidation failure inside XLogPageRead(). Once recovery catches up to
this point, it can deal with it. This doesn't fix any known live bug,
but it fits with the general philosophy of giving up fast when
prefetching runs into trouble.
---
src/backend/access/transam/xlogrecovery.c | 8 ++++++++
1 file changed, 8 insertions(+)
diff --git a/src/backend/access/transam/xlogrecovery.c b/src/backend/access/transam/xlogrecovery.c
index 422322cf5a..76ce9704ca 100644
--- a/src/backend/access/transam/xlogrecovery.c
+++ b/src/backend/access/transam/xlogrecovery.c
@@ -3342,6 +3342,14 @@ retry:
return readLen;
next_record_is_invalid:
+ /*
+ * If we're reading ahead, give up fast. Standby retries and error
+ * reporting will be handled by a later read when recovery catches up to
+ * this point.
+ */
+ if (xlogreader->nonblocking)
+ return XLREAD_WOULDBLOCK;
+
lastSourceFailed = true;
if (readFile >= 0)
--
2.30.2
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2022-11-11 22:49 Tomas Vondra <[email protected]>
parent: Tomas Vondra <[email protected]>
1 sibling, 2 replies; 127+ messages in thread
From: Tomas Vondra @ 2022-11-11 22:49 UTC (permalink / raw)
To: PostgreSQL Hackers <[email protected]>
Hi,
I noticed on cfbot the patch no longer applies, so here's a rebased
version. Most of the breakage was due to the column filtering reworks,
grammar changes etc. A lot of bitrot, but mostly mechanical stuff.
I haven't looked into the optimizations / improvements I discussed in my
previous post (logging only LSN of the last WAL-logged increment),
because while fixing "make check-world" I ran into a more serious issue
that I think needs to be discussed first. And I suspect it might also
affect the feasibility of the LSN optimization.
So, what's the issue - the current solution is based on WAL-logging
state of all sequences incremented by the transaction at COMMIT. To do
that, we read the state from disk, and write that into WAL. However,
these WAL messages are not necessarily correlated to COMMIT records, so
stuff like this might happen:
1. transaction T1 increments sequence S
2. transaction T2 increments sequence S
3. both T1 and T2 start to COMMIT
4. T1 reads state of S from disk, writes it into WAL
5. transaction T3 increments sequence S
6. T2 reads state of S from disk, writes it into WAL
7. T2 write COMMIT into WAL
8. T1 write COMMIT into WAL
Because the apply order is determined by ordering of COMMIT records,
this means we'd apply the increments logged by T2, and then by T1. But
that undoes the increment by T3, and the sequence would go backwards.
The previous patch version addressed that by acquiring lock on the
sequence, holding it until transaction end. This effectively ensures the
order of sequence messages and COMMIT matches. But that's problematic
for a number of reasons:
1) throughput reduction, because the COMMIT records need to serialize
2) deadlock risk, if we happen to lock sequences in different order
(in different transactions)
3) problem for prepared transactions - the sequences are locked and
logged in PrepareTransaction, because we may not have seqhashtab
beyond that point. This is a much worse variant of (1).
Note: I also wonder what happens if someone does DISCARD SEQUENCES. I
guess we'll forget the sequences, which is bad - so we'd have to invent
a separate cache that does not have this issue.
I realized (3) because one of the test_decoding TAP tests got stuck
exactly because of a sequence locked by a prepared transaction.
This patch simply releases the lock after writing the WAL message, but
that just makes it vulnerable to the reordering. And this would have
been true even with the LSN optimization.
However, I was thinking that maybe we could use the LSN of the WAL
message (XLOG_LOGICAL_SEQUENCE) to deal with the ordering issue, because
*this* is the sensible sequence increment ordering.
In the example above, we'd first apply the WAL message from T2 (because
that commits first). And then we'd get to apply T1, but the WAL message
has an older LSN, so we'd skip it.
But this requires us remembering LSN of the already applied WAL sequence
messages, which could be tricky - we'd need to persist it in some way
because of restarts, etc. We can't do this while decoding but on the
apply side, I think, because of streaming, aborts.
The other option might be to make these messages non-transactional, in
which case we'd separate the ordering from COMMIT ordering, evading the
reordering problem.
That'd mean we'd ignore rollbacks (which seems fine), we could probably
optimize this by checking if the state actually changed, etc. But we'd
also need to deal with transactions created in the (still uncommitted)
transaction. But I'm also worried it might lead to the same issue with
non-transactional behaviors that forced revert in v15.
regards
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
Attachments:
[text/x-patch] decoding-sequences-tracking-20221112.patch (585.1K, ../../[email protected]/2-decoding-sequences-tracking-20221112.patch)
download | inline diff:
diff --git a/contrib/test_decoding/Makefile b/contrib/test_decoding/Makefile
index c7ce6037064..05c0c5a2f86 100644
--- a/contrib/test_decoding/Makefile
+++ b/contrib/test_decoding/Makefile
@@ -5,7 +5,8 @@ PGFILEDESC = "test_decoding - example of a logical decoding output plugin"
REGRESS = ddl xact rewrite toast permissions decoding_in_xact \
decoding_into_rel binary prepared replorigin time messages \
- spill slot truncate stream stats twophase twophase_stream
+ spill slot truncate stream stats twophase twophase_stream \
+ sequence
ISOLATION = mxact delayed_startup ondisk_startup concurrent_ddl_dml \
oldest_xmin snapshot_transfer subxact_without_top concurrent_stream \
twophase_snapshot slot_creation_error catalog_change_snapshot
diff --git a/contrib/test_decoding/expected/ddl.out b/contrib/test_decoding/expected/ddl.out
index 9a28b5ddc5a..1e37c8c8979 100644
--- a/contrib/test_decoding/expected/ddl.out
+++ b/contrib/test_decoding/expected/ddl.out
@@ -40,7 +40,7 @@ SELECT 'init' FROM pg_create_physical_replication_slot('repl');
init
(1 row)
-SELECT data FROM pg_logical_slot_get_changes('repl', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('repl', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
ERROR: cannot use physical replication slot for logical decoding
SELECT pg_drop_replication_slot('repl');
pg_drop_replication_slot
@@ -89,7 +89,7 @@ COMMIT;
ALTER TABLE replication_example RENAME COLUMN text TO somenum;
INSERT INTO replication_example(somedata, somenum) VALUES (4, 1);
-- collect all changes
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
---------------------------------------------------------------------------------------------------------------------------
BEGIN
@@ -118,7 +118,7 @@ SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'inc
ALTER TABLE replication_example ALTER COLUMN somenum TYPE int4 USING (somenum::int4);
-- check that this doesn't produce any changes from the heap rewrite
-SELECT count(data) FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT count(data) FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
count
-------
0
@@ -134,7 +134,7 @@ INSERT INTO replication_example(somedata, somenum, zaphod2) VALUES (6, 3, 1);
INSERT INTO replication_example(somedata, somenum, zaphod1) VALUES (6, 4, 2);
COMMIT;
-- show changes
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
------------------------------------------------------------------------------------------------------------------------------------------
BEGIN
@@ -154,7 +154,7 @@ INSERT INTO replication_example(id, somedata, somenum) SELECT i, i, i FROM gener
ON CONFLICT (id) DO UPDATE SET somenum = excluded.somenum + 1;
COMMIT;
/* display results */
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
--------------------------------------------------------------------------------------------------------------------------------------------------
BEGIN
@@ -242,7 +242,7 @@ CREATE TABLE tr_unique(id2 serial unique NOT NULL, data int);
INSERT INTO tr_unique(data) VALUES(10);
ALTER TABLE tr_unique RENAME TO tr_pkey;
ALTER TABLE tr_pkey ADD COLUMN id serial primary key;
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-rewrites', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-rewrites', '1', 'include-sequences', '0');
data
-----------------------------------------------------------------------------
BEGIN
@@ -257,7 +257,7 @@ INSERT INTO tr_pkey(data) VALUES(1);
--show deletion with primary key
DELETE FROM tr_pkey;
/* display results */
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
----------------------------------------------------------------------------
BEGIN
@@ -288,7 +288,7 @@ INSERT INTO tr_oddlength VALUES('ab', 'foo');
COMMIT;
/* display results, but hide most of the output */
SELECT count(*), min(data), max(data)
-FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1')
+FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0')
GROUP BY substring(data, 1, 24)
ORDER BY 1,2;
count | min | max
@@ -310,7 +310,7 @@ DELETE FROM spoolme;
DROP TABLE spoolme;
COMMIT;
SELECT data
-FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1')
+FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0')
WHERE data ~ 'UPDATE';
data
-------------------------------------------------------------------------------------------------------------
@@ -324,7 +324,7 @@ INSERT INTO tr_etoomuch (id, data)
SELECT g.i, -g.i FROM generate_series(8000, 12000) g(i)
ON CONFLICT(id) DO UPDATE SET data = EXCLUDED.data;
SELECT substring(data, 1, 29), count(*)
-FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1') WITH ORDINALITY
+FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0') WITH ORDINALITY
GROUP BY 1
ORDER BY min(ordinality);
substring | count
@@ -355,7 +355,7 @@ RELEASE SAVEPOINT c;
INSERT INTO tr_sub(path) VALUES ('1-top-2-#1');
RELEASE SAVEPOINT b;
COMMIT;
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
----------------------------------------------------------------------
BEGIN
@@ -394,7 +394,7 @@ INSERT INTO tr_sub(path) VALUES ('2-top-1...--#3');
RELEASE SAVEPOINT subtop;
INSERT INTO tr_sub(path) VALUES ('2-top-#1');
COMMIT;
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
------------------------------------------------------------------------
BEGIN
@@ -415,7 +415,7 @@ INSERT INTO tr_sub(path) VALUES ('3-top-2-2-#1');
ROLLBACK TO SAVEPOINT b;
INSERT INTO tr_sub(path) VALUES ('3-top-2-#2');
COMMIT;
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
-----------------------------------------------------------------------
BEGIN
@@ -444,7 +444,7 @@ BEGIN;
SAVEPOINT a;
INSERT INTO tr_sub(path) VALUES ('5-top-1-#1');
COMMIT;
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
---------------------------------------------------------------------
BEGIN
@@ -465,7 +465,7 @@ ROLLBACK TO SAVEPOINT a;
ALTER TABLE tr_sub_ddl ALTER COLUMN data TYPE bigint;
INSERT INTO tr_sub_ddl VALUES(43);
COMMIT;
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
--------------------------------------------------
BEGIN
@@ -542,7 +542,7 @@ Options: user_catalog_table=false
INSERT INTO replication_metadata(relation, options)
VALUES ('zaphod', NULL);
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
------------------------------------------------------------------------------------------------------------------------------------
BEGIN
@@ -659,7 +659,7 @@ INSERT INTO toasttable(toasted_col2) SELECT repeat(string_agg(to_char(g.i, 'FM00
UPDATE toasttable
SET toasted_col1 = (SELECT string_agg(g.i::text, '') FROM generate_series(1, 2000) g(i))
WHERE id = 1;
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
BEGIN
@@ -814,7 +814,7 @@ UPDATE toasttable
WHERE id = 1;
-- make sure we decode correctly even if the toast table is gone
DROP TABLE toasttable;
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
BEGIN
@@ -826,7 +826,7 @@ SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'inc
(6 rows)
-- done, free logical replication slot
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
------
(0 rows)
diff --git a/contrib/test_decoding/expected/decoding_in_xact.out b/contrib/test_decoding/expected/decoding_in_xact.out
index b65253f4630..0816c780fe1 100644
--- a/contrib/test_decoding/expected/decoding_in_xact.out
+++ b/contrib/test_decoding/expected/decoding_in_xact.out
@@ -58,7 +58,7 @@ SELECT pg_current_xact_id() = '0';
-- don't show yet, haven't committed
INSERT INTO nobarf(data) VALUES('2');
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
-----------------------------------------------------------
BEGIN
@@ -68,7 +68,7 @@ SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'inc
COMMIT;
INSERT INTO nobarf(data) VALUES('3');
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
-----------------------------------------------------------
BEGIN
diff --git a/contrib/test_decoding/expected/decoding_into_rel.out b/contrib/test_decoding/expected/decoding_into_rel.out
index 8fd3390066d..03966b8b1ca 100644
--- a/contrib/test_decoding/expected/decoding_into_rel.out
+++ b/contrib/test_decoding/expected/decoding_into_rel.out
@@ -19,7 +19,7 @@ SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'inc
CREATE TABLE somechange(id serial primary key);
INSERT INTO somechange DEFAULT VALUES;
CREATE TABLE changeresult AS
- SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+ SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
SELECT * FROM changeresult;
data
------------------------------------------------
@@ -29,9 +29,9 @@ SELECT * FROM changeresult;
(3 rows)
INSERT INTO changeresult
- SELECT data FROM pg_logical_slot_peek_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+ SELECT data FROM pg_logical_slot_peek_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
INSERT INTO changeresult
- SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+ SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
SELECT * FROM changeresult;
data
--------------------------------------------------------------------------------------------------------------------------------------------------
@@ -63,7 +63,7 @@ DROP TABLE somechange;
CREATE FUNCTION slot_changes_wrapper(slot_name name) RETURNS SETOF TEXT AS $$
BEGIN
RETURN QUERY
- SELECT data FROM pg_logical_slot_peek_changes(slot_name, NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+ SELECT data FROM pg_logical_slot_peek_changes(slot_name, NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
END$$ LANGUAGE plpgsql;
SELECT * FROM slot_changes_wrapper('regression_slot');
slot_changes_wrapper
@@ -84,7 +84,7 @@ SELECT * FROM slot_changes_wrapper('regression_slot');
COMMIT
(14 rows)
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
BEGIN
diff --git a/contrib/test_decoding/expected/delayed_startup.out b/contrib/test_decoding/expected/delayed_startup.out
index d10de3658ac..95f3d2ce4c3 100644
--- a/contrib/test_decoding/expected/delayed_startup.out
+++ b/contrib/test_decoding/expected/delayed_startup.out
@@ -20,12 +20,13 @@ step s1b: BEGIN ISOLATION LEVEL SERIALIZABLE;
step s1w: INSERT INTO do_write DEFAULT VALUES;
step s1c: COMMIT;
step s2start: SELECT data FROM pg_logical_slot_get_changes('isolation_slot', NULL, NULL, 'include-xids', 'false');
-data
---------------------------------------------
-BEGIN
-table public.do_write: INSERT: id[integer]:2
-COMMIT
-(3 rows)
+data
+----------------------------------------------------------------------
+BEGIN
+table public.do_write: INSERT: id[integer]:2
+sequence public.do_write_id_seq: last_value: 2 log_cnt: 31 is_called:1
+COMMIT
+(4 rows)
step s1b: BEGIN ISOLATION LEVEL SERIALIZABLE;
step s1w: INSERT INTO do_write DEFAULT VALUES;
@@ -36,12 +37,13 @@ data
step s1c: COMMIT;
step s2start: SELECT data FROM pg_logical_slot_get_changes('isolation_slot', NULL, NULL, 'include-xids', 'false');
-data
---------------------------------------------
-BEGIN
-table public.do_write: INSERT: id[integer]:3
-COMMIT
-(3 rows)
+data
+----------------------------------------------------------------------
+BEGIN
+table public.do_write: INSERT: id[integer]:3
+sequence public.do_write_id_seq: last_value: 3 log_cnt: 30 is_called:1
+COMMIT
+(4 rows)
?column?
--------
diff --git a/contrib/test_decoding/expected/mxact.out b/contrib/test_decoding/expected/mxact.out
index 03ad3df0999..c5bc26c8044 100644
--- a/contrib/test_decoding/expected/mxact.out
+++ b/contrib/test_decoding/expected/mxact.out
@@ -7,7 +7,7 @@ step s0init: SELECT 'init' FROM pg_create_logical_replication_slot('isolation_sl
init
(1 row)
-step s0start: SELECT data FROM pg_logical_slot_get_changes('isolation_slot', NULL, NULL, 'include-xids', 'false');
+step s0start: SELECT data FROM pg_logical_slot_get_changes('isolation_slot', NULL, NULL, 'include-xids', 'false', 'include-sequences', 'false');
data
----
(0 rows)
@@ -27,7 +27,7 @@ t
(1 row)
step s0w: INSERT INTO do_write DEFAULT VALUES;
-step s0start: SELECT data FROM pg_logical_slot_get_changes('isolation_slot', NULL, NULL, 'include-xids', 'false');
+step s0start: SELECT data FROM pg_logical_slot_get_changes('isolation_slot', NULL, NULL, 'include-xids', 'false', 'include-sequences', 'false');
data
--------------------------------------------
BEGIN
@@ -50,7 +50,7 @@ step s0init: SELECT 'init' FROM pg_create_logical_replication_slot('isolation_sl
init
(1 row)
-step s0start: SELECT data FROM pg_logical_slot_get_changes('isolation_slot', NULL, NULL, 'include-xids', 'false');
+step s0start: SELECT data FROM pg_logical_slot_get_changes('isolation_slot', NULL, NULL, 'include-xids', 'false', 'include-sequences', 'false');
data
----
(0 rows)
@@ -71,7 +71,7 @@ t
step s0alter: ALTER TABLE do_write ADD column ts timestamptz;
step s0w: INSERT INTO do_write DEFAULT VALUES;
-step s0start: SELECT data FROM pg_logical_slot_get_changes('isolation_slot', NULL, NULL, 'include-xids', 'false');
+step s0start: SELECT data FROM pg_logical_slot_get_changes('isolation_slot', NULL, NULL, 'include-xids', 'false', 'include-sequences', 'false');
data
------------------------------------------------------------------------------
BEGIN
diff --git a/contrib/test_decoding/expected/ondisk_startup.out b/contrib/test_decoding/expected/ondisk_startup.out
index bc7ff071648..3d2fa4a92e0 100644
--- a/contrib/test_decoding/expected/ondisk_startup.out
+++ b/contrib/test_decoding/expected/ondisk_startup.out
@@ -35,7 +35,7 @@ init
step s2c: COMMIT;
step s1insert: INSERT INTO do_write DEFAULT VALUES;
step s1checkpoint: CHECKPOINT;
-step s1start: SELECT data FROM pg_logical_slot_get_changes('isolation_slot', NULL, NULL, 'include-xids', 'false');
+step s1start: SELECT data FROM pg_logical_slot_get_changes('isolation_slot', NULL, NULL, 'include-xids', 'false', 'include-sequences', 'false');
data
--------------------------------------------------------------------
BEGIN
@@ -46,7 +46,7 @@ COMMIT
step s1insert: INSERT INTO do_write DEFAULT VALUES;
step s1alter: ALTER TABLE do_write ADD COLUMN addedbys1 int;
step s1insert: INSERT INTO do_write DEFAULT VALUES;
-step s1start: SELECT data FROM pg_logical_slot_get_changes('isolation_slot', NULL, NULL, 'include-xids', 'false');
+step s1start: SELECT data FROM pg_logical_slot_get_changes('isolation_slot', NULL, NULL, 'include-xids', 'false', 'include-sequences', 'false');
data
--------------------------------------------------------------------------------------------
BEGIN
diff --git a/contrib/test_decoding/expected/replorigin.out b/contrib/test_decoding/expected/replorigin.out
index 49ffaeea2da..a828a453b1b 100644
--- a/contrib/test_decoding/expected/replorigin.out
+++ b/contrib/test_decoding/expected/replorigin.out
@@ -82,9 +82,9 @@ SELECT 'init' FROM pg_create_logical_replication_slot('regression_slot', 'test_d
-- origin tx
INSERT INTO origin_tbl(data) VALUES ('will be replicated and decoded and decoded again');
INSERT INTO target_tbl(data)
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
-- as is normal, the insert into target_tbl shows up
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
----------------------------------------------------------------------------------------------------------------------------------------------------------------------------
BEGIN
@@ -120,7 +120,7 @@ SELECT pg_replication_origin_xact_setup('0/aabbccdd', '2013-01-01 00:00');
(1 row)
INSERT INTO target_tbl(data)
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'only-local', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'only-local', '1', 'include-sequences', '0');
COMMIT;
-- check replication progress for the session is correct
SELECT pg_replication_origin_session_progress(false);
@@ -164,14 +164,14 @@ SELECT pg_replication_origin_progress('regress_test_decoding: regression_slot',
SELECT pg_replication_origin_session_reset();
ERROR: no replication origin is configured
-- and magically the replayed xact will be filtered!
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'only-local', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'only-local', '1', 'include-sequences', '0');
data
------
(0 rows)
--but new original changes still show up
INSERT INTO origin_tbl(data) VALUES ('will be replicated');
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'only-local', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'only-local', '1', 'include-sequences', '0');
data
--------------------------------------------------------------------------------
BEGIN
@@ -237,7 +237,7 @@ SELECT local_id, external_id,
1 | regress_test_decoding: regression_slot_no_lsn | f | t
(1 row)
-SELECT data FROM pg_logical_slot_get_changes('regression_slot_no_lsn', NULL, NULL, 'skip-empty-xacts', '1', 'include-xids', '0');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot_no_lsn', NULL, NULL, 'skip-empty-xacts', '1', 'include-xids', '0', 'include-sequences', '0');
data
-------------------------------------------------------------------------------------
BEGIN
diff --git a/contrib/test_decoding/expected/rewrite.out b/contrib/test_decoding/expected/rewrite.out
index b30999c436b..5d15b192edc 100644
--- a/contrib/test_decoding/expected/rewrite.out
+++ b/contrib/test_decoding/expected/rewrite.out
@@ -64,7 +64,7 @@ SELECT 'init' FROM pg_create_logical_replication_slot('regression_slot', 'test_d
CREATE TABLE replication_example(id SERIAL PRIMARY KEY, somedata int, text varchar(120));
INSERT INTO replication_example(somedata) VALUES (1);
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
----------------------------------------------------------------------------------------------------------
BEGIN
@@ -115,7 +115,7 @@ INSERT INTO replication_example(somedata, testcolumn1, testcolumn3) VALUES (7, 5
COMMIT;
-- make old files go away
CHECKPOINT;
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
BEGIN
@@ -141,7 +141,7 @@ VACUUM FULL pg_proc; VACUUM FULL pg_description; VACUUM FULL pg_shdescription; V
INSERT INTO replication_example(somedata, testcolumn1, testcolumn3) VALUES (8, 6, 1);
VACUUM FULL pg_proc; VACUUM FULL pg_description; VACUUM FULL pg_shdescription; VACUUM FULL iamalargetable;
INSERT INTO replication_example(somedata, testcolumn1, testcolumn3) VALUES (9, 7, 1);
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
BEGIN
diff --git a/contrib/test_decoding/expected/sequence.out b/contrib/test_decoding/expected/sequence.out
new file mode 100644
index 00000000000..b35f1929b0f
--- /dev/null
+++ b/contrib/test_decoding/expected/sequence.out
@@ -0,0 +1,326 @@
+-- predictability
+SET synchronous_commit = on;
+SELECT 'init' FROM pg_create_logical_replication_slot('regression_slot', 'test_decoding');
+ ?column?
+----------
+ init
+(1 row)
+
+CREATE SEQUENCE test_sequence;
+-- test the sequence changes by several nextval() calls
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 1
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 2
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 4
+(1 row)
+
+-- test the sequence changes by several ALTER commands
+ALTER SEQUENCE test_sequence INCREMENT BY 10;
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 14
+(1 row)
+
+ALTER SEQUENCE test_sequence START WITH 3000;
+ALTER SEQUENCE test_sequence MAXVALUE 10000;
+ALTER SEQUENCE test_sequence RESTART WITH 4000;
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 4000
+(1 row)
+
+-- test the sequence changes by several setval() calls
+SELECT setval('test_sequence', 3500);
+ setval
+--------
+ 3500
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3510
+(1 row)
+
+SELECT setval('test_sequence', 3500, true);
+ setval
+--------
+ 3500
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3510
+(1 row)
+
+SELECT setval('test_sequence', 3500, false);
+ setval
+--------
+ 3500
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3500
+(1 row)
+
+-- show results and drop sequence
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+ data
+-------------------------------------------------------------------------
+ BEGIN
+ sequence public.test_sequence: last_value: 1 log_cnt: 32 is_called:1
+ COMMIT
+ BEGIN
+ sequence public.test_sequence: last_value: 2 log_cnt: 31 is_called:1
+ COMMIT
+ BEGIN
+ sequence public.test_sequence: last_value: 3 log_cnt: 30 is_called:1
+ COMMIT
+ BEGIN
+ sequence public.test_sequence: last_value: 4 log_cnt: 29 is_called:1
+ COMMIT
+ BEGIN
+ sequence public.test_sequence: last_value: 14 log_cnt: 32 is_called:1
+ COMMIT
+ BEGIN
+ sequence public.test_sequence: last_value: 4000 log_cnt: 32 is_called:1
+ COMMIT
+ BEGIN
+ sequence public.test_sequence: last_value: 3500 log_cnt: 0 is_called:1
+ COMMIT
+ BEGIN
+ sequence public.test_sequence: last_value: 3510 log_cnt: 32 is_called:1
+ COMMIT
+ BEGIN
+ sequence public.test_sequence: last_value: 3500 log_cnt: 0 is_called:1
+ COMMIT
+ BEGIN
+ sequence public.test_sequence: last_value: 3510 log_cnt: 32 is_called:1
+ COMMIT
+ BEGIN
+ sequence public.test_sequence: last_value: 3500 log_cnt: 0 is_called:0
+ COMMIT
+ BEGIN
+ sequence public.test_sequence: last_value: 3500 log_cnt: 32 is_called:1
+ COMMIT
+(36 rows)
+
+DROP SEQUENCE test_sequence;
+-- rollback on sequence creation and update
+BEGIN;
+CREATE SEQUENCE test_sequence;
+CREATE TABLE test_table (a INT);
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 1
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 2
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3
+(1 row)
+
+SELECT setval('test_sequence', 3000);
+ setval
+--------
+ 3000
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3001
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3002
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3003
+(1 row)
+
+ALTER SEQUENCE test_sequence RESTART WITH 6000;
+INSERT INTO test_table VALUES( (SELECT nextval('test_sequence')) );
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 6001
+(1 row)
+
+ROLLBACK;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+ data
+------
+(0 rows)
+
+-- rollback on table creation with serial column
+BEGIN;
+CREATE TABLE test_table (a SERIAL, b INT);
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+INSERT INTO test_table (b) VALUES (300);
+SELECT setval('test_table_a_seq', 3000);
+ setval
+--------
+ 3000
+(1 row)
+
+INSERT INTO test_table (b) VALUES (400);
+INSERT INTO test_table (b) VALUES (500);
+INSERT INTO test_table (b) VALUES (600);
+ALTER SEQUENCE test_table_a_seq RESTART WITH 6000;
+INSERT INTO test_table (b) VALUES (700);
+INSERT INTO test_table (b) VALUES (800);
+INSERT INTO test_table (b) VALUES (900);
+ROLLBACK;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+ data
+------
+(0 rows)
+
+-- rollback on table with serial column
+CREATE TABLE test_table (a SERIAL, b INT);
+BEGIN;
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+INSERT INTO test_table (b) VALUES (300);
+SELECT setval('test_table_a_seq', 3000);
+ setval
+--------
+ 3000
+(1 row)
+
+INSERT INTO test_table (b) VALUES (400);
+INSERT INTO test_table (b) VALUES (500);
+INSERT INTO test_table (b) VALUES (600);
+ALTER SEQUENCE test_table_a_seq RESTART WITH 6000;
+INSERT INTO test_table (b) VALUES (700);
+INSERT INTO test_table (b) VALUES (800);
+INSERT INTO test_table (b) VALUES (900);
+ROLLBACK;
+-- check table and sequence values after rollback
+SELECT * from test_table_a_seq;
+ last_value | log_cnt | is_called
+------------+---------+-----------
+ 3003 | 30 | t
+(1 row)
+
+SELECT nextval('test_table_a_seq');
+ nextval
+---------
+ 3004
+(1 row)
+
+DROP TABLE test_table;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+ data
+----------------------------------------------------------------------------
+ BEGIN
+ sequence public.test_table_a_seq: last_value: 3004 log_cnt: 29 is_called:1
+ COMMIT
+(3 rows)
+
+-- savepoint test on table with serial column
+BEGIN;
+CREATE TABLE test_table (a SERIAL, b INT);
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+SAVEPOINT a;
+INSERT INTO test_table (b) VALUES (300);
+ROLLBACK TO SAVEPOINT a;
+DROP TABLE test_table;
+COMMIT;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+ data
+--------------------------------------------------------------
+ BEGIN
+ table public.test_table: INSERT: a[integer]:1 b[integer]:100
+ table public.test_table: INSERT: a[integer]:2 b[integer]:200
+ COMMIT
+(4 rows)
+
+-- savepoint test on table with serial column
+BEGIN;
+CREATE SEQUENCE test_sequence;
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 1
+(1 row)
+
+SELECT setval('test_sequence', 3000);
+ setval
+--------
+ 3000
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3001
+(1 row)
+
+SAVEPOINT a;
+ALTER SEQUENCE test_sequence START WITH 7000;
+SELECT setval('test_sequence', 5000);
+ setval
+--------
+ 5000
+(1 row)
+
+ROLLBACK TO SAVEPOINT a;
+SELECT * FROM test_sequence;
+ last_value | log_cnt | is_called
+------------+---------+-----------
+ 3001 | 32 | t
+(1 row)
+
+DROP SEQUENCE test_sequence;
+COMMIT;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+ data
+------
+(0 rows)
+
+SELECT pg_drop_replication_slot('regression_slot');
+ pg_drop_replication_slot
+--------------------------
+
+(1 row)
+
diff --git a/contrib/test_decoding/expected/slot.out b/contrib/test_decoding/expected/slot.out
index 63a9940f73a..de59d544ae8 100644
--- a/contrib/test_decoding/expected/slot.out
+++ b/contrib/test_decoding/expected/slot.out
@@ -95,7 +95,7 @@ SELECT 'init' FROM pg_create_logical_replication_slot('regression_slot2', 'test_
(1 row)
INSERT INTO replication_example(somedata, text) VALUES (1, 3);
-SELECT data FROM pg_logical_slot_get_changes('regression_slot1', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot1', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
---------------------------------------------------------------------------------------------------------
BEGIN
@@ -107,7 +107,7 @@ SELECT data FROM pg_logical_slot_get_changes('regression_slot1', NULL, NULL, 'in
COMMIT
(7 rows)
-SELECT data FROM pg_logical_slot_get_changes('regression_slot2', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot2', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
---------------------------------------------------------------------------------------------------------
BEGIN
@@ -132,7 +132,7 @@ SELECT :'wal_lsn' = :'end_lsn';
t
(1 row)
-SELECT data FROM pg_logical_slot_get_changes('regression_slot1', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot1', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
---------------------------------------------------------------------------------------------------------
BEGIN
@@ -140,7 +140,7 @@ SELECT data FROM pg_logical_slot_get_changes('regression_slot1', NULL, NULL, 'in
COMMIT
(3 rows)
-SELECT data FROM pg_logical_slot_get_changes('regression_slot2', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot2', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
------
(0 rows)
diff --git a/contrib/test_decoding/expected/toast.out b/contrib/test_decoding/expected/toast.out
index a757e7dc8d5..d0ab0e06394 100644
--- a/contrib/test_decoding/expected/toast.out
+++ b/contrib/test_decoding/expected/toast.out
@@ -52,7 +52,7 @@ CREATE TABLE toasted_copy (
);
ALTER TABLE toasted_copy ALTER COLUMN data SET STORAGE EXTERNAL;
\copy toasted_copy FROM STDIN
-SELECT substr(data, 1, 200) FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT substr(data, 1, 200) FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
substr
----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
BEGIN
@@ -316,7 +316,7 @@ SELECT pg_column_size(toasted_key) > 2^16 FROM toasted_several;
t
(1 row)
-SELECT regexp_replace(data, '^(.{100}).*(.{100})$', '\1..\2') FROM pg_logical_slot_peek_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT regexp_replace(data, '^(.{100}).*(.{100})$', '\1..\2') FROM pg_logical_slot_peek_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
regexp_replace
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
BEGIN
@@ -327,7 +327,7 @@ SELECT regexp_replace(data, '^(.{100}).*(.{100})$', '\1..\2') FROM pg_logical_sl
-- test update of a toasted key without changing it
UPDATE toasted_several SET toasted_col1 = toasted_key;
UPDATE toasted_several SET toasted_col2 = toasted_col1;
-SELECT regexp_replace(data, '^(.{100}).*(.{100})$', '\1..\2') FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT regexp_replace(data, '^(.{100}).*(.{100})$', '\1..\2') FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
regexp_replace
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
BEGIN
@@ -350,7 +350,7 @@ UPDATE toasted_several SET toasted_col1 = toasted_col2 WHERE id = 1;
DELETE FROM toasted_several WHERE id = 1;
COMMIT;
DROP TABLE toasted_several;
-SELECT regexp_replace(data, '^(.{100}).*(.{100})$', '\1..\2') FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1')
+SELECT regexp_replace(data, '^(.{100}).*(.{100})$', '\1..\2') FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0')
WHERE data NOT LIKE '%INSERT: %';
regexp_replace
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
@@ -373,7 +373,7 @@ INSERT INTO tbl1 VALUES(1, repeat('a', 4000)) ;
ALTER TABLE tbl1 ADD COLUMN id serial primary key;
INSERT INTO tbl2 VALUES(1);
commit;
-SELECT substr(data, 1, 200) FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT substr(data, 1, 200) FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
substr
----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
BEGIN
diff --git a/contrib/test_decoding/expected/truncate.out b/contrib/test_decoding/expected/truncate.out
index e64d377214a..961b0ebdc89 100644
--- a/contrib/test_decoding/expected/truncate.out
+++ b/contrib/test_decoding/expected/truncate.out
@@ -11,7 +11,7 @@ CREATE TABLE tab2 (a int primary key, b int);
TRUNCATE tab1;
TRUNCATE tab1, tab1 RESTART IDENTITY CASCADE;
TRUNCATE tab1, tab2;
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
data
------------------------------------------------------
BEGIN
diff --git a/contrib/test_decoding/expected/twophase_snapshot.out b/contrib/test_decoding/expected/twophase_snapshot.out
index f555ffddf74..a9d20ded585 100644
--- a/contrib/test_decoding/expected/twophase_snapshot.out
+++ b/contrib/test_decoding/expected/twophase_snapshot.out
@@ -29,22 +29,24 @@ init
step s1insert: INSERT INTO do_write DEFAULT VALUES;
step s1start: SELECT data FROM pg_logical_slot_get_changes('isolation_slot', NULL, NULL, 'include-xids', 'false', 'skip-empty-xacts', '1');
-data
---------------------------------------------
-BEGIN
-table public.do_write: INSERT: id[integer]:2
-COMMIT
-(3 rows)
+data
+----------------------------------------------------------------------
+BEGIN
+table public.do_write: INSERT: id[integer]:2
+sequence public.do_write_id_seq: last_value: 2 log_cnt: 31 is_called:1
+COMMIT
+(4 rows)
step s2cp: COMMIT PREPARED 'test1';
step s1start: SELECT data FROM pg_logical_slot_get_changes('isolation_slot', NULL, NULL, 'include-xids', 'false', 'skip-empty-xacts', '1');
-data
---------------------------------------------
-BEGIN
-table public.do_write: INSERT: id[integer]:1
-PREPARE TRANSACTION 'test1'
-COMMIT PREPARED 'test1'
-(4 rows)
+data
+----------------------------------------------------------------------
+BEGIN
+table public.do_write: INSERT: id[integer]:1
+sequence public.do_write_id_seq: last_value: 1 log_cnt: 32 is_called:1
+PREPARE TRANSACTION 'test1'
+COMMIT PREPARED 'test1'
+(5 rows)
?column?
--------
diff --git a/contrib/test_decoding/specs/mxact.spec b/contrib/test_decoding/specs/mxact.spec
index ea5b1aa2d67..19f3af14b30 100644
--- a/contrib/test_decoding/specs/mxact.spec
+++ b/contrib/test_decoding/specs/mxact.spec
@@ -13,7 +13,7 @@ teardown
session "s0"
setup { SET synchronous_commit=on; }
step "s0init" {SELECT 'init' FROM pg_create_logical_replication_slot('isolation_slot', 'test_decoding');}
-step "s0start" {SELECT data FROM pg_logical_slot_get_changes('isolation_slot', NULL, NULL, 'include-xids', 'false');}
+step "s0start" {SELECT data FROM pg_logical_slot_get_changes('isolation_slot', NULL, NULL, 'include-xids', 'false', 'include-sequences', 'false');}
step "s0alter" {ALTER TABLE do_write ADD column ts timestamptz; }
step "s0w" { INSERT INTO do_write DEFAULT VALUES; }
diff --git a/contrib/test_decoding/specs/ondisk_startup.spec b/contrib/test_decoding/specs/ondisk_startup.spec
index 96ce87f1a45..fb43cbc5e15 100644
--- a/contrib/test_decoding/specs/ondisk_startup.spec
+++ b/contrib/test_decoding/specs/ondisk_startup.spec
@@ -16,7 +16,7 @@ session "s1"
setup { SET synchronous_commit=on; }
step "s1init" {SELECT 'init' FROM pg_create_logical_replication_slot('isolation_slot', 'test_decoding');}
-step "s1start" {SELECT data FROM pg_logical_slot_get_changes('isolation_slot', NULL, NULL, 'include-xids', 'false');}
+step "s1start" {SELECT data FROM pg_logical_slot_get_changes('isolation_slot', NULL, NULL, 'include-xids', 'false', 'include-sequences', 'false');}
step "s1insert" { INSERT INTO do_write DEFAULT VALUES; }
step "s1checkpoint" { CHECKPOINT; }
step "s1alter" { ALTER TABLE do_write ADD COLUMN addedbys1 int; }
diff --git a/contrib/test_decoding/sql/ddl.sql b/contrib/test_decoding/sql/ddl.sql
index 4f76bed72c1..807bc56a451 100644
--- a/contrib/test_decoding/sql/ddl.sql
+++ b/contrib/test_decoding/sql/ddl.sql
@@ -19,7 +19,7 @@ SELECT pg_drop_replication_slot('regression_slot');
-- check that we're detecting a streaming rep slot used for logical decoding
SELECT 'init' FROM pg_create_physical_replication_slot('repl');
-SELECT data FROM pg_logical_slot_get_changes('repl', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('repl', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
SELECT pg_drop_replication_slot('repl');
@@ -64,11 +64,11 @@ ALTER TABLE replication_example RENAME COLUMN text TO somenum;
INSERT INTO replication_example(somedata, somenum) VALUES (4, 1);
-- collect all changes
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
ALTER TABLE replication_example ALTER COLUMN somenum TYPE int4 USING (somenum::int4);
-- check that this doesn't produce any changes from the heap rewrite
-SELECT count(data) FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT count(data) FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
INSERT INTO replication_example(somedata, somenum) VALUES (5, 1);
@@ -82,7 +82,7 @@ INSERT INTO replication_example(somedata, somenum, zaphod1) VALUES (6, 4, 2);
COMMIT;
-- show changes
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
-- ON CONFLICT DO UPDATE support
BEGIN;
@@ -91,7 +91,7 @@ INSERT INTO replication_example(id, somedata, somenum) SELECT i, i, i FROM gener
COMMIT;
/* display results */
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
-- MERGE support
BEGIN;
@@ -113,14 +113,14 @@ CREATE TABLE tr_unique(id2 serial unique NOT NULL, data int);
INSERT INTO tr_unique(data) VALUES(10);
ALTER TABLE tr_unique RENAME TO tr_pkey;
ALTER TABLE tr_pkey ADD COLUMN id serial primary key;
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-rewrites', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-rewrites', '1', 'include-sequences', '0');
INSERT INTO tr_pkey(data) VALUES(1);
--show deletion with primary key
DELETE FROM tr_pkey;
/* display results */
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
/*
* check that disk spooling works (also for logical messages)
@@ -137,7 +137,7 @@ COMMIT;
/* display results, but hide most of the output */
SELECT count(*), min(data), max(data)
-FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1')
+FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0')
GROUP BY substring(data, 1, 24)
ORDER BY 1,2;
@@ -152,7 +152,7 @@ DROP TABLE spoolme;
COMMIT;
SELECT data
-FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1')
+FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0')
WHERE data ~ 'UPDATE';
-- check that a large, spooled, upsert works
@@ -161,7 +161,7 @@ SELECT g.i, -g.i FROM generate_series(8000, 12000) g(i)
ON CONFLICT(id) DO UPDATE SET data = EXCLUDED.data;
SELECT substring(data, 1, 29), count(*)
-FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1') WITH ORDINALITY
+FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0') WITH ORDINALITY
GROUP BY 1
ORDER BY min(ordinality);
@@ -189,7 +189,7 @@ INSERT INTO tr_sub(path) VALUES ('1-top-2-#1');
RELEASE SAVEPOINT b;
COMMIT;
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
-- check that we handle xlog assignments correctly
BEGIN;
@@ -218,7 +218,7 @@ RELEASE SAVEPOINT subtop;
INSERT INTO tr_sub(path) VALUES ('2-top-#1');
COMMIT;
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
-- make sure rollbacked subtransactions aren't decoded
BEGIN;
@@ -231,7 +231,7 @@ ROLLBACK TO SAVEPOINT b;
INSERT INTO tr_sub(path) VALUES ('3-top-2-#2');
COMMIT;
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
-- test whether a known, but not yet logged toplevel xact, followed by a
-- subxact commit is handled correctly
@@ -250,7 +250,7 @@ INSERT INTO tr_sub(path) VALUES ('5-top-1-#1');
COMMIT;
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
-- check that DDL in aborted subtransactions handled correctly
CREATE TABLE tr_sub_ddl(data int);
@@ -263,7 +263,7 @@ ALTER TABLE tr_sub_ddl ALTER COLUMN data TYPE bigint;
INSERT INTO tr_sub_ddl VALUES(43);
COMMIT;
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
/*
@@ -303,7 +303,7 @@ ALTER TABLE replication_metadata SET (user_catalog_table = false);
INSERT INTO replication_metadata(relation, options)
VALUES ('zaphod', NULL);
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
/*
* check whether we handle updates/deletes correct with & without a pkey
@@ -414,7 +414,7 @@ UPDATE toasttable
SET toasted_col1 = (SELECT string_agg(g.i::text, '') FROM generate_series(1, 2000) g(i))
WHERE id = 1;
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
INSERT INTO toasttable(toasted_col1) SELECT string_agg(g.i::text, '') FROM generate_series(1, 2000) g(i);
@@ -426,10 +426,10 @@ WHERE id = 1;
-- make sure we decode correctly even if the toast table is gone
DROP TABLE toasttable;
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
-- done, free logical replication slot
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
SELECT pg_drop_replication_slot('regression_slot');
diff --git a/contrib/test_decoding/sql/decoding_in_xact.sql b/contrib/test_decoding/sql/decoding_in_xact.sql
index 108782dc2e9..b343b745661 100644
--- a/contrib/test_decoding/sql/decoding_in_xact.sql
+++ b/contrib/test_decoding/sql/decoding_in_xact.sql
@@ -32,10 +32,10 @@ BEGIN;
SELECT pg_current_xact_id() = '0';
-- don't show yet, haven't committed
INSERT INTO nobarf(data) VALUES('2');
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
COMMIT;
INSERT INTO nobarf(data) VALUES('3');
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
SELECT 'stop' FROM pg_drop_replication_slot('regression_slot');
diff --git a/contrib/test_decoding/sql/decoding_into_rel.sql b/contrib/test_decoding/sql/decoding_into_rel.sql
index 1068cec5888..27d5d08879e 100644
--- a/contrib/test_decoding/sql/decoding_into_rel.sql
+++ b/contrib/test_decoding/sql/decoding_into_rel.sql
@@ -15,14 +15,14 @@ CREATE TABLE somechange(id serial primary key);
INSERT INTO somechange DEFAULT VALUES;
CREATE TABLE changeresult AS
- SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+ SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
SELECT * FROM changeresult;
INSERT INTO changeresult
- SELECT data FROM pg_logical_slot_peek_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+ SELECT data FROM pg_logical_slot_peek_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
INSERT INTO changeresult
- SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+ SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
SELECT * FROM changeresult;
DROP TABLE changeresult;
@@ -32,11 +32,11 @@ DROP TABLE somechange;
CREATE FUNCTION slot_changes_wrapper(slot_name name) RETURNS SETOF TEXT AS $$
BEGIN
RETURN QUERY
- SELECT data FROM pg_logical_slot_peek_changes(slot_name, NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+ SELECT data FROM pg_logical_slot_peek_changes(slot_name, NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
END$$ LANGUAGE plpgsql;
SELECT * FROM slot_changes_wrapper('regression_slot');
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
SELECT 'stop' FROM pg_drop_replication_slot('regression_slot');
diff --git a/contrib/test_decoding/sql/replorigin.sql b/contrib/test_decoding/sql/replorigin.sql
index db06541f565..8f8db499639 100644
--- a/contrib/test_decoding/sql/replorigin.sql
+++ b/contrib/test_decoding/sql/replorigin.sql
@@ -46,10 +46,10 @@ SELECT 'init' FROM pg_create_logical_replication_slot('regression_slot', 'test_d
-- origin tx
INSERT INTO origin_tbl(data) VALUES ('will be replicated and decoded and decoded again');
INSERT INTO target_tbl(data)
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
-- as is normal, the insert into target_tbl shows up
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
INSERT INTO origin_tbl(data) VALUES ('will be replicated, but not decoded again');
@@ -65,7 +65,7 @@ BEGIN;
-- setup transaction origin
SELECT pg_replication_origin_xact_setup('0/aabbccdd', '2013-01-01 00:00');
INSERT INTO target_tbl(data)
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'only-local', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'only-local', '1', 'include-sequences', '0');
COMMIT;
-- check replication progress for the session is correct
@@ -84,11 +84,11 @@ SELECT pg_replication_origin_progress('regress_test_decoding: regression_slot',
SELECT pg_replication_origin_session_reset();
-- and magically the replayed xact will be filtered!
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'only-local', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'only-local', '1', 'include-sequences', '0');
--but new original changes still show up
INSERT INTO origin_tbl(data) VALUES ('will be replicated');
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'only-local', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'only-local', '1', 'include-sequences', '0');
SELECT pg_drop_replication_slot('regression_slot');
SELECT pg_replication_origin_drop('regress_test_decoding: regression_slot');
@@ -119,7 +119,7 @@ SELECT local_id, external_id,
remote_lsn <> '0/0' AS valid_remote_lsn,
local_lsn <> '0/0' AS valid_local_lsn
FROM pg_replication_origin_status;
-SELECT data FROM pg_logical_slot_get_changes('regression_slot_no_lsn', NULL, NULL, 'skip-empty-xacts', '1', 'include-xids', '0');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot_no_lsn', NULL, NULL, 'skip-empty-xacts', '1', 'include-xids', '0', 'include-sequences', '0');
-- Clean up
SELECT pg_replication_origin_session_reset();
SELECT pg_drop_replication_slot('regression_slot_no_lsn');
diff --git a/contrib/test_decoding/sql/rewrite.sql b/contrib/test_decoding/sql/rewrite.sql
index 62dead3a9b1..1715bd289dd 100644
--- a/contrib/test_decoding/sql/rewrite.sql
+++ b/contrib/test_decoding/sql/rewrite.sql
@@ -35,7 +35,7 @@ SELECT pg_relation_size((SELECT reltoastrelid FROM pg_class WHERE oid = 'pg_shde
SELECT 'init' FROM pg_create_logical_replication_slot('regression_slot', 'test_decoding');
CREATE TABLE replication_example(id SERIAL PRIMARY KEY, somedata int, text varchar(120));
INSERT INTO replication_example(somedata) VALUES (1);
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
BEGIN;
INSERT INTO replication_example(somedata) VALUES (2);
@@ -90,7 +90,7 @@ COMMIT;
-- make old files go away
CHECKPOINT;
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
-- trigger repeated rewrites of a system catalog with a toast table,
-- that previously was buggy: [email protected]
@@ -98,7 +98,7 @@ VACUUM FULL pg_proc; VACUUM FULL pg_description; VACUUM FULL pg_shdescription; V
INSERT INTO replication_example(somedata, testcolumn1, testcolumn3) VALUES (8, 6, 1);
VACUUM FULL pg_proc; VACUUM FULL pg_description; VACUUM FULL pg_shdescription; VACUUM FULL iamalargetable;
INSERT INTO replication_example(somedata, testcolumn1, testcolumn3) VALUES (9, 7, 1);
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
SELECT pg_drop_replication_slot('regression_slot');
DROP TABLE IF EXISTS replication_example;
diff --git a/contrib/test_decoding/sql/sequence.sql b/contrib/test_decoding/sql/sequence.sql
new file mode 100644
index 00000000000..978ee8088e1
--- /dev/null
+++ b/contrib/test_decoding/sql/sequence.sql
@@ -0,0 +1,119 @@
+-- predictability
+SET synchronous_commit = on;
+SELECT 'init' FROM pg_create_logical_replication_slot('regression_slot', 'test_decoding');
+
+CREATE SEQUENCE test_sequence;
+
+-- test the sequence changes by several nextval() calls
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+
+-- test the sequence changes by several ALTER commands
+ALTER SEQUENCE test_sequence INCREMENT BY 10;
+SELECT nextval('test_sequence');
+
+ALTER SEQUENCE test_sequence START WITH 3000;
+ALTER SEQUENCE test_sequence MAXVALUE 10000;
+ALTER SEQUENCE test_sequence RESTART WITH 4000;
+SELECT nextval('test_sequence');
+
+-- test the sequence changes by several setval() calls
+SELECT setval('test_sequence', 3500);
+SELECT nextval('test_sequence');
+SELECT setval('test_sequence', 3500, true);
+SELECT nextval('test_sequence');
+SELECT setval('test_sequence', 3500, false);
+SELECT nextval('test_sequence');
+
+-- show results and drop sequence
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+DROP SEQUENCE test_sequence;
+
+-- rollback on sequence creation and update
+BEGIN;
+CREATE SEQUENCE test_sequence;
+CREATE TABLE test_table (a INT);
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT setval('test_sequence', 3000);
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+ALTER SEQUENCE test_sequence RESTART WITH 6000;
+INSERT INTO test_table VALUES( (SELECT nextval('test_sequence')) );
+SELECT nextval('test_sequence');
+ROLLBACK;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+
+-- rollback on table creation with serial column
+BEGIN;
+CREATE TABLE test_table (a SERIAL, b INT);
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+INSERT INTO test_table (b) VALUES (300);
+SELECT setval('test_table_a_seq', 3000);
+INSERT INTO test_table (b) VALUES (400);
+INSERT INTO test_table (b) VALUES (500);
+INSERT INTO test_table (b) VALUES (600);
+ALTER SEQUENCE test_table_a_seq RESTART WITH 6000;
+INSERT INTO test_table (b) VALUES (700);
+INSERT INTO test_table (b) VALUES (800);
+INSERT INTO test_table (b) VALUES (900);
+ROLLBACK;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+
+-- rollback on table with serial column
+CREATE TABLE test_table (a SERIAL, b INT);
+
+BEGIN;
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+INSERT INTO test_table (b) VALUES (300);
+SELECT setval('test_table_a_seq', 3000);
+INSERT INTO test_table (b) VALUES (400);
+INSERT INTO test_table (b) VALUES (500);
+INSERT INTO test_table (b) VALUES (600);
+ALTER SEQUENCE test_table_a_seq RESTART WITH 6000;
+INSERT INTO test_table (b) VALUES (700);
+INSERT INTO test_table (b) VALUES (800);
+INSERT INTO test_table (b) VALUES (900);
+ROLLBACK;
+
+-- check table and sequence values after rollback
+SELECT * from test_table_a_seq;
+SELECT nextval('test_table_a_seq');
+
+DROP TABLE test_table;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+
+-- savepoint test on table with serial column
+BEGIN;
+CREATE TABLE test_table (a SERIAL, b INT);
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+SAVEPOINT a;
+INSERT INTO test_table (b) VALUES (300);
+ROLLBACK TO SAVEPOINT a;
+DROP TABLE test_table;
+COMMIT;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+
+-- savepoint test on table with serial column
+BEGIN;
+CREATE SEQUENCE test_sequence;
+SELECT nextval('test_sequence');
+SELECT setval('test_sequence', 3000);
+SELECT nextval('test_sequence');
+SAVEPOINT a;
+ALTER SEQUENCE test_sequence START WITH 7000;
+SELECT setval('test_sequence', 5000);
+ROLLBACK TO SAVEPOINT a;
+SELECT * FROM test_sequence;
+DROP SEQUENCE test_sequence;
+COMMIT;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+
+SELECT pg_drop_replication_slot('regression_slot');
diff --git a/contrib/test_decoding/sql/slot.sql b/contrib/test_decoding/sql/slot.sql
index 1aa27c56674..52a740c43de 100644
--- a/contrib/test_decoding/sql/slot.sql
+++ b/contrib/test_decoding/sql/slot.sql
@@ -50,8 +50,8 @@ SELECT 'init' FROM pg_create_logical_replication_slot('regression_slot2', 'test_
INSERT INTO replication_example(somedata, text) VALUES (1, 3);
-SELECT data FROM pg_logical_slot_get_changes('regression_slot1', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
-SELECT data FROM pg_logical_slot_get_changes('regression_slot2', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot1', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot2', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
INSERT INTO replication_example(somedata, text) VALUES (1, 4);
INSERT INTO replication_example(somedata, text) VALUES (1, 5);
@@ -65,8 +65,8 @@ SELECT slot_name FROM pg_replication_slot_advance('regression_slot2', pg_current
SELECT :'wal_lsn' = :'end_lsn';
-SELECT data FROM pg_logical_slot_get_changes('regression_slot1', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
-SELECT data FROM pg_logical_slot_get_changes('regression_slot2', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot1', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot2', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
DROP TABLE replication_example;
diff --git a/contrib/test_decoding/sql/toast.sql b/contrib/test_decoding/sql/toast.sql
index d1c560a174d..f5d4fc082ed 100644
--- a/contrib/test_decoding/sql/toast.sql
+++ b/contrib/test_decoding/sql/toast.sql
@@ -264,7 +264,7 @@ ALTER TABLE toasted_copy ALTER COLUMN data SET STORAGE EXTERNAL;
202 untoasted199
203 untoasted200
\.
-SELECT substr(data, 1, 200) FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT substr(data, 1, 200) FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
-- test we can decode "old" tuples bigger than the max heap tuple size correctly
DROP TABLE IF EXISTS toasted_several;
@@ -287,13 +287,13 @@ UPDATE pg_attribute SET attstorage = 'x' WHERE attrelid = 'toasted_several_pkey'
INSERT INTO toasted_several(toasted_key) VALUES(repeat('9876543210', 10000));
SELECT pg_column_size(toasted_key) > 2^16 FROM toasted_several;
-SELECT regexp_replace(data, '^(.{100}).*(.{100})$', '\1..\2') FROM pg_logical_slot_peek_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT regexp_replace(data, '^(.{100}).*(.{100})$', '\1..\2') FROM pg_logical_slot_peek_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
-- test update of a toasted key without changing it
UPDATE toasted_several SET toasted_col1 = toasted_key;
UPDATE toasted_several SET toasted_col2 = toasted_col1;
-SELECT regexp_replace(data, '^(.{100}).*(.{100})$', '\1..\2') FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT regexp_replace(data, '^(.{100}).*(.{100})$', '\1..\2') FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
/*
* update with large tuplebuf, in a transaction large enough to force to spool to disk
@@ -306,7 +306,7 @@ COMMIT;
DROP TABLE toasted_several;
-SELECT regexp_replace(data, '^(.{100}).*(.{100})$', '\1..\2') FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1')
+SELECT regexp_replace(data, '^(.{100}).*(.{100})$', '\1..\2') FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0')
WHERE data NOT LIKE '%INSERT: %';
/*
@@ -322,6 +322,6 @@ INSERT INTO tbl1 VALUES(1, repeat('a', 4000)) ;
ALTER TABLE tbl1 ADD COLUMN id serial primary key;
INSERT INTO tbl2 VALUES(1);
commit;
-SELECT substr(data, 1, 200) FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT substr(data, 1, 200) FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
SELECT pg_drop_replication_slot('regression_slot');
diff --git a/contrib/test_decoding/sql/truncate.sql b/contrib/test_decoding/sql/truncate.sql
index 5633854e0df..b0f36fd681b 100644
--- a/contrib/test_decoding/sql/truncate.sql
+++ b/contrib/test_decoding/sql/truncate.sql
@@ -10,5 +10,5 @@ TRUNCATE tab1;
TRUNCATE tab1, tab1 RESTART IDENTITY CASCADE;
TRUNCATE tab1, tab2;
-SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '0');
SELECT pg_drop_replication_slot('regression_slot');
diff --git a/contrib/test_decoding/test_decoding.c b/contrib/test_decoding/test_decoding.c
index e0fd6f17659..87178b05e7a 100644
--- a/contrib/test_decoding/test_decoding.c
+++ b/contrib/test_decoding/test_decoding.c
@@ -31,6 +31,7 @@ typedef struct
bool include_timestamp;
bool skip_empty_xacts;
bool only_local;
+ bool include_sequences;
} TestDecodingData;
/*
@@ -72,6 +73,10 @@ static void pg_decode_message(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn, XLogRecPtr lsn,
bool transactional, const char *prefix,
Size sz, const char *message);
+static void pg_decode_sequence(LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn, XLogRecPtr sequence_lsn,
+ Relation rel,
+ int64 last_value, int64 log_cnt, bool is_called);
static bool pg_decode_filter_prepare(LogicalDecodingContext *ctx,
TransactionId xid,
const char *gid);
@@ -112,6 +117,10 @@ static void pg_decode_stream_message(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn, XLogRecPtr lsn,
bool transactional, const char *prefix,
Size sz, const char *message);
+static void pg_decode_stream_sequence(LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn, XLogRecPtr sequence_lsn,
+ Relation rel,
+ int64 last_value, int64 log_cnt, bool is_called);
static void pg_decode_stream_truncate(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn,
int nrelations, Relation relations[],
@@ -137,6 +146,7 @@ _PG_output_plugin_init(OutputPluginCallbacks *cb)
cb->filter_by_origin_cb = pg_decode_filter;
cb->shutdown_cb = pg_decode_shutdown;
cb->message_cb = pg_decode_message;
+ cb->sequence_cb = pg_decode_sequence;
cb->filter_prepare_cb = pg_decode_filter_prepare;
cb->begin_prepare_cb = pg_decode_begin_prepare_txn;
cb->prepare_cb = pg_decode_prepare_txn;
@@ -149,6 +159,7 @@ _PG_output_plugin_init(OutputPluginCallbacks *cb)
cb->stream_commit_cb = pg_decode_stream_commit;
cb->stream_change_cb = pg_decode_stream_change;
cb->stream_message_cb = pg_decode_stream_message;
+ cb->stream_sequence_cb = pg_decode_stream_sequence;
cb->stream_truncate_cb = pg_decode_stream_truncate;
}
@@ -169,6 +180,7 @@ pg_decode_startup(LogicalDecodingContext *ctx, OutputPluginOptions *opt,
data->include_xids = true;
data->include_timestamp = false;
data->skip_empty_xacts = false;
+ data->include_sequences = true;
data->only_local = false;
ctx->output_plugin_private = data;
@@ -261,6 +273,17 @@ pg_decode_startup(LogicalDecodingContext *ctx, OutputPluginOptions *opt,
errmsg("could not parse value \"%s\" for parameter \"%s\"",
strVal(elem->arg), elem->defname)));
}
+ else if (strcmp(elem->defname, "include-sequences") == 0)
+ {
+
+ if (elem->arg == NULL)
+ data->include_sequences = false;
+ else if (!parse_bool(strVal(elem->arg), &data->include_sequences))
+ ereport(ERROR,
+ (errcode(ERRCODE_INVALID_PARAMETER_VALUE),
+ errmsg("could not parse value \"%s\" for parameter \"%s\"",
+ strVal(elem->arg), elem->defname)));
+ }
else
{
ereport(ERROR,
@@ -752,6 +775,34 @@ pg_decode_message(LogicalDecodingContext *ctx,
OutputPluginWrite(ctx, true);
}
+static void
+pg_decode_sequence(LogicalDecodingContext *ctx, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ TestDecodingData *data = ctx->output_plugin_private;
+ TestDecodingTxnData *txndata = txn->output_plugin_private;
+
+ if (!data->include_sequences)
+ return;
+
+ /* output BEGIN if we haven't yet, but only for the transactional case */
+ if (data->skip_empty_xacts && !txndata->xact_wrote_changes)
+ {
+ pg_output_begin(ctx, data, txn, false);
+ }
+ txndata->xact_wrote_changes = true;
+
+ OutputPluginPrepareWrite(ctx, true);
+ appendStringInfoString(ctx->out, "sequence ");
+ appendStringInfoString(ctx->out,
+ quote_qualified_identifier(get_namespace_name(get_rel_namespace(RelationGetRelid(rel))),
+ RelationGetRelationName(rel)));
+ appendStringInfo(ctx->out, ": last_value: %zu log_cnt: %zu is_called:%d",
+ last_value, log_cnt, is_called);
+ OutputPluginWrite(ctx, true);
+}
+
static void
pg_decode_stream_start(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn)
@@ -951,6 +1002,36 @@ pg_decode_stream_message(LogicalDecodingContext *ctx,
OutputPluginWrite(ctx, true);
}
+static void
+pg_decode_stream_sequence(LogicalDecodingContext *ctx, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ TestDecodingData *data = ctx->output_plugin_private;
+ TestDecodingTxnData *txndata = txn->output_plugin_private;
+
+ if (!data->include_sequences)
+ return;
+
+ /* output BEGIN if we haven't yet, but only for the transactional case */
+ {
+ if (data->skip_empty_xacts && !txndata->xact_wrote_changes)
+ {
+ pg_output_begin(ctx, data, txn, false);
+ }
+ txndata->xact_wrote_changes = true;
+ }
+
+ OutputPluginPrepareWrite(ctx, true);
+ appendStringInfoString(ctx->out, "streaming sequence ");
+ appendStringInfoString(ctx->out,
+ quote_qualified_identifier(get_namespace_name(get_rel_namespace(RelationGetRelid(rel))),
+ RelationGetRelationName(rel)));
+ appendStringInfo(ctx->out, ": last_value: " INT64_FORMAT " log_cnt: " INT64_FORMAT " is_called:%d",
+ last_value, log_cnt, is_called);
+ OutputPluginWrite(ctx, true);
+}
+
/*
* In streaming mode, we don't display the detailed information of Truncate.
* See pg_decode_stream_change.
diff --git a/doc/src/sgml/catalogs.sgml b/doc/src/sgml/catalogs.sgml
index 00f833d210e..c4cb8da8585 100644
--- a/doc/src/sgml/catalogs.sgml
+++ b/doc/src/sgml/catalogs.sgml
@@ -6383,6 +6383,16 @@ SCRAM-SHA-256$<replaceable><iteration count></replaceable>:<replaceable>&l
Reference to schema
</para></entry>
</row>
+
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ <structfield>pntype</structfield> <type>char</type>
+ Determines which object type is included from this schema.
+ </para>
+ <para>
+ Reference to schema
+ </para></entry>
+ </row>
</tbody>
</tgroup>
</table>
@@ -9600,6 +9610,11 @@ SCRAM-SHA-256$<replaceable><iteration count></replaceable>:<replaceable>&l
</para></entry>
</row>
+ <row>
+ <entry><link linkend="view-pg-publication-sequences"><structname>pg_publication_sequences</structname></link></entry>
+ <entry>publications and their associated sequences</entry>
+ </row>
+
<row>
<entry role="catalog_table_entry"><para role="column_definition">
<structfield>umoptions</structfield> <type>text[]</type>
diff --git a/doc/src/sgml/logicaldecoding.sgml b/doc/src/sgml/logicaldecoding.sgml
index 38ee69dcccc..58d4067cced 100644
--- a/doc/src/sgml/logicaldecoding.sgml
+++ b/doc/src/sgml/logicaldecoding.sgml
@@ -458,6 +458,7 @@ typedef struct OutputPluginCallbacks
LogicalDecodeTruncateCB truncate_cb;
LogicalDecodeCommitCB commit_cb;
LogicalDecodeMessageCB message_cb;
+ LogicalDecodeSequenceCB sequence_cb;
LogicalDecodeFilterByOriginCB filter_by_origin_cb;
LogicalDecodeShutdownCB shutdown_cb;
LogicalDecodeFilterPrepareCB filter_prepare_cb;
@@ -472,6 +473,7 @@ typedef struct OutputPluginCallbacks
LogicalDecodeStreamCommitCB stream_commit_cb;
LogicalDecodeStreamChangeCB stream_change_cb;
LogicalDecodeStreamMessageCB stream_message_cb;
+ LogicalDecodeStreamSequenceCB stream_sequence_cb;
LogicalDecodeStreamTruncateCB stream_truncate_cb;
} OutputPluginCallbacks;
@@ -481,9 +483,11 @@ typedef void (*LogicalOutputPluginInit) (struct OutputPluginCallbacks *cb);
and <function>commit_cb</function> callbacks are required,
while <function>startup_cb</function>,
<function>filter_by_origin_cb</function>, <function>truncate_cb</function>,
- and <function>shutdown_cb</function> are optional.
- If <function>truncate_cb</function> is not set but a
+ <function>sequence_cb</function>, and <function>shutdown_cb</function> are
+ optional. If <function>truncate_cb</function> is not set but a
<command>TRUNCATE</command> is to be decoded, the action will be ignored.
+ Similarly, if <function>sequence_cb</function> is not set and a sequence
+ change is to be decoded, the action will be ignored.
</para>
<para>
@@ -492,7 +496,8 @@ typedef void (*LogicalOutputPluginInit) (struct OutputPluginCallbacks *cb);
<function>stream_stop_cb</function>, <function>stream_abort_cb</function>,
<function>stream_commit_cb</function>, <function>stream_change_cb</function>,
and <function>stream_prepare_cb</function>
- are required, while <function>stream_message_cb</function> and
+ are required, while <function>stream_message_cb</function>,
+ <function>stream_sequence_cb</function>, and
<function>stream_truncate_cb</function> are optional.
</para>
@@ -808,6 +813,35 @@ typedef void (*LogicalDecodeMessageCB) (struct LogicalDecodingContext *ctx,
</para>
</sect3>
+ <sect3 id="logicaldecoding-output-plugin-sequence">
+ <title>Sequence Callback</title>
+
+ <para>
+ The optional <function>sequence_cb</function> callback is called for
+ actions that update a sequence value.
+<programlisting>
+typedef void (*LogicalDecodeSequenceCB) (struct LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional,
+ int64 last_value,
+ int64 log_cnt,
+ bool is_called);
+</programlisting>
+ The <parameter>txn</parameter> parameter contains meta information about
+ the transaction the sequence change is part of. Note however that for
+ non-transactional increments, the transaction may be either NULL or not
+ NULL, depending on if the transaction already has an XID assigned.
+ The <parameter>sequence_lsn</parameter> has the WAL location of the
+ sequence update. <parameter>transactional</parameter> says if the
+ sequence has to be replayed as part of the transaction or directly.
+
+ The <parameter>last_value</parameter>, <parameter>log_cnt</parameter> and
+ <parameter>is_called</parameter> parameters describe the sequence change.
+ </para>
+ </sect3>
+
<sect3 id="logicaldecoding-output-plugin-filter-prepare">
<title>Prepare Filter Callback</title>
@@ -1017,6 +1051,26 @@ typedef void (*LogicalDecodeStreamMessageCB) (struct LogicalDecodingContext *ctx
</para>
</sect3>
+ <sect3 id="logicaldecoding-output-plugin-stream-sequence">
+ <title>Stream Sequence Callback</title>
+ <para>
+ The optional <function>stream_sequence_cb</function> callback is called
+ for actions that change a sequence in a block of streamed changes
+ (demarcated by <function>stream_start_cb</function> and
+ <function>stream_stop_cb</function> calls).
+<programlisting>
+typedef void (*LogicalDecodeStreamSequenceCB) (struct LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional,
+ int64 last_value,
+ int64 log_cnt,
+ bool is_called);
+</programlisting>
+ </para>
+ </sect3>
+
<sect3 id="logicaldecoding-output-plugin-stream-truncate">
<title>Stream Truncate Callback</title>
<para>
@@ -1197,8 +1251,9 @@ OutputPluginWrite(ctx, true);
in-progress transactions. There are multiple required streaming callbacks
(<function>stream_start_cb</function>, <function>stream_stop_cb</function>,
<function>stream_abort_cb</function>, <function>stream_commit_cb</function>
- and <function>stream_change_cb</function>) and two optional callbacks
- (<function>stream_message_cb</function> and <function>stream_truncate_cb</function>).
+ and <function>stream_change_cb</function>) and multiple optional callbacks
+ (<function>stream_message_cb</function>, <function>stream_sequence_cb</function>,
+ and <function>stream_truncate_cb</function>).
Also, if streaming of two-phase commands is to be supported, then additional
callbacks must be provided. (See <xref linkend="logicaldecoding-two-phase-commits"/>
for details).
diff --git a/doc/src/sgml/protocol.sgml b/doc/src/sgml/protocol.sgml
index 5fdd429e05d..993433863d4 100644
--- a/doc/src/sgml/protocol.sgml
+++ b/doc/src/sgml/protocol.sgml
@@ -6403,6 +6403,105 @@ psql "dbname=postgres replication=database" -c "IDENTIFY_SYSTEM;"
</listitem>
</varlistentry>
+ <varlistentry id="protocol-logicalrep-message-formats-Sequence">
+ <term>Sequence</term>
+ <listitem>
+ <variablelist>
+ <varlistentry>
+ <term>Byte1('X')</term>
+ <listitem>
+ <para>
+ Identifies the message as a sequence message.
+ </para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term>Int32 (TransactionId)</term>
+ <listitem>
+ <para>
+ Xid of the transaction (only present for streamed transactions).
+ This field is available since protocol version 2.
+ </para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term>Int8(0)</term>
+ <listitem>
+ <para>
+ Flags; currently unused.
+ </para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term>Int64 (XLogRecPtr)</term>
+ <listitem>
+ <para>
+ The LSN of the sequence increment.
+ </para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term>String</term>
+ <listitem>
+ <para>
+ Namespace (empty string for <literal>pg_catalog</literal>).
+ </para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term>String</term>
+ <listitem>
+ <para>
+ Relation name.
+ </para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term>Int8</term>
+ <listitem>
+ <para>
+ 1 if the sequence update is transactional, 0 otherwise.
+ </para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term>Int64</term>
+ <listitem>
+ <para>
+ <structfield>last_value</structfield> value of the sequence.
+ </para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term>Int64</term>
+ <listitem>
+ <para>
+ <structfield>log_cnt</structfield> value of the sequence.
+ </para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term>Int8</term>
+ <listitem>
+ <para>
+ <structfield>is_called</structfield> value of the sequence.
+ </para>
+ </listitem>
+ </varlistentry>
+
+ </variablelist>
+ </listitem>
+ </varlistentry>
+
<varlistentry id="protocol-logicalrep-message-formats-Type">
<term>Type</term>
<listitem>
diff --git a/doc/src/sgml/ref/alter_publication.sgml b/doc/src/sgml/ref/alter_publication.sgml
index c84b11f47a7..b7dee52a0a7 100644
--- a/doc/src/sgml/ref/alter_publication.sgml
+++ b/doc/src/sgml/ref/alter_publication.sgml
@@ -31,7 +31,9 @@ ALTER PUBLICATION <replaceable class="parameter">name</replaceable> RENAME TO <r
<phrase>where <replaceable class="parameter">publication_object</replaceable> is one of:</phrase>
TABLE [ ONLY ] <replaceable class="parameter">table_name</replaceable> [ * ] [ ( <replaceable class="parameter">column_name</replaceable> [, ... ] ) ] [ WHERE ( <replaceable class="parameter">expression</replaceable> ) ] [, ... ]
+ SEQUENCE <replaceable class="parameter">sequence_name</replaceable> [, ... ]
TABLES IN SCHEMA { <replaceable class="parameter">schema_name</replaceable> | CURRENT_SCHEMA } [, ... ]
+ SEQUENCES IN SCHEMA { <replaceable class="parameter">schema_name</replaceable> | CURRENT_SCHEMA } [, ... ]
</synopsis>
</refsynopsisdiv>
@@ -44,13 +46,13 @@ ALTER PUBLICATION <replaceable class="parameter">name</replaceable> RENAME TO <r
</para>
<para>
- The first three variants change which tables/schemas are part of the
- publication. The <literal>SET</literal> clause will replace the list of
- tables/schemas in the publication with the specified list; the existing
- tables/schemas that were present in the publication will be removed. The
- <literal>ADD</literal> and <literal>DROP</literal> clauses will add and
- remove one or more tables/schemas from the publication. Note that adding
- tables/schemas to a publication that is already subscribed to will require an
+ The first three variants change which objects (tables, sequences or schemas)
+ are part of the publication. The <literal>SET</literal> clause will replace
+ the list of objects in the publication with the specified list; the existing
+ objects that were present in the publication will be removed.
+ The <literal>ADD</literal> and <literal>DROP</literal> clauses will add and
+ remove one or more objects from the publication. Note that adding objects
+ to a publication that is already subscribed to will require an
<literal>ALTER SUBSCRIPTION ... REFRESH PUBLICATION</literal> action on the
subscribing side in order to become effective. Note also that
<literal>DROP TABLES IN SCHEMA</literal> will not drop any schema tables
@@ -133,6 +135,15 @@ ALTER PUBLICATION <replaceable class="parameter">name</replaceable> RENAME TO <r
</listitem>
</varlistentry>
+ <varlistentry>
+ <term><replaceable class="parameter">sequence_name</replaceable></term>
+ <listitem>
+ <para>
+ Name of an existing sequence.
+ </para>
+ </listitem>
+ </varlistentry>
+
<varlistentry>
<term><replaceable class="parameter">schema_name</replaceable></term>
<listitem>
diff --git a/doc/src/sgml/ref/alter_subscription.sgml b/doc/src/sgml/ref/alter_subscription.sgml
index 1e8d72062b3..c9f2db4a5ca 100644
--- a/doc/src/sgml/ref/alter_subscription.sgml
+++ b/doc/src/sgml/ref/alter_subscription.sgml
@@ -150,8 +150,8 @@ ALTER SUBSCRIPTION <replaceable class="parameter">name</replaceable> RENAME TO <
<listitem>
<para>
Fetch missing table information from publisher. This will start
- replication of tables that were added to the subscribed-to publications
- since <command>CREATE SUBSCRIPTION</command> or
+ replication of tables and sequences that were added to the subscribed-to
+ publications since <command>CREATE SUBSCRIPTION</command> or
the last invocation of <command>REFRESH PUBLICATION</command>.
</para>
@@ -169,8 +169,8 @@ ALTER SUBSCRIPTION <replaceable class="parameter">name</replaceable> RENAME TO <
The default is <literal>true</literal>.
</para>
<para>
- Previously subscribed tables are not copied, even if a table's row
- filter <literal>WHERE</literal> clause has since been modified.
+ Previously subscribed tables and sequences are not copied, even if a
+ table's row filter <literal>WHERE</literal> clause has since been modified.
</para>
<para>
See <xref linkend="sql-createsubscription-notes"/> for details of
diff --git a/doc/src/sgml/ref/create_publication.sgml b/doc/src/sgml/ref/create_publication.sgml
index e229384e6ff..93c2ffbcf5d 100644
--- a/doc/src/sgml/ref/create_publication.sgml
+++ b/doc/src/sgml/ref/create_publication.sgml
@@ -23,6 +23,7 @@ PostgreSQL documentation
<synopsis>
CREATE PUBLICATION <replaceable class="parameter">name</replaceable>
[ FOR ALL TABLES
+ | FOR ALL SEQUENCES
| FOR <replaceable class="parameter">publication_object</replaceable> [, ... ] ]
[ WITH ( <replaceable class="parameter">publication_parameter</replaceable> [= <replaceable class="parameter">value</replaceable>] [, ... ] ) ]
@@ -30,6 +31,8 @@ CREATE PUBLICATION <replaceable class="parameter">name</replaceable>
TABLE [ ONLY ] <replaceable class="parameter">table_name</replaceable> [ * ] [ ( <replaceable class="parameter">column_name</replaceable> [, ... ] ) ] [ WHERE ( <replaceable class="parameter">expression</replaceable> ) ] [, ... ]
TABLES IN SCHEMA { <replaceable class="parameter">schema_name</replaceable> | CURRENT_SCHEMA } [, ... ]
+ SEQUENCE <replaceable class="parameter">sequence_name</replaceable> [ * ] [, ... ]
+ SEQUENCES IN SCHEMA { <replaceable class="parameter">schema_name</replaceable> | CURRENT_SCHEMA } [, ... ]
</synopsis>
</refsynopsisdiv>
@@ -117,31 +120,47 @@ CREATE PUBLICATION <replaceable class="parameter">name</replaceable>
</listitem>
</varlistentry>
+ <varlistentry>
+ <term><literal>FOR SEQUENCE</literal></term>
+ <listitem>
+ <para>
+ Specifies a list of sequences to add to the publication.
+ </para>
+
+ <para>
+ Specifying a sequence that is part of a schema specified by <literal>FOR
+ ALL SEQUENCES IN SCHEMA</literal> is not supported.
+ </para>
+ </listitem>
+ </varlistentry>
+
<varlistentry>
<term><literal>FOR ALL TABLES</literal></term>
+ <term><literal>FOR ALL SEQUENCES</literal></term>
<listitem>
<para>
- Marks the publication as one that replicates changes for all tables in
- the database, including tables created in the future.
+ Marks the publication as one that replicates changes for all tables/sequences in
+ the database, including tables/sequences created in the future.
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><literal>FOR TABLES IN SCHEMA</literal></term>
+ <term><literal>FOR SEQUENCES IN SCHEMA</literal></term>
<listitem>
<para>
- Marks the publication as one that replicates changes for all tables in
+ Marks the publication as one that replicates changes for all tables/sequences in
the specified list of schemas, including tables created in the future.
</para>
<para>
- Specifying a schema when the publication also publishes a table with a
+ Specifying a schema when the publication also publishes a table/sequence with a
column list is not supported.
</para>
<para>
- Only persistent base tables and partitioned tables present in the schema
+ Only persistent base tables and partitioned tables/sequences present in the schema
will be included as part of the publication. Temporary tables, unlogged
tables, foreign tables, materialized views, and regular views from the
schema will not be part of the publication.
@@ -223,10 +242,12 @@ CREATE PUBLICATION <replaceable class="parameter">name</replaceable>
<title>Notes</title>
<para>
- If <literal>FOR TABLE</literal>, <literal>FOR ALL TABLES</literal> or
- <literal>FOR TABLES IN SCHEMA</literal> are not specified, then the
- publication starts out with an empty set of tables. That is useful if
- tables or schemas are to be added later.
+ If <literal>FOR TABLE</literal>, <literal>FOR ALL TABLES</literal>,
+ <literal>FOR TABLES IN SCHEMA</literal>, <literal>FOR SEQUENCE</literal>,
+ <literal>FOR ALL SEQUENCES </literal> or <literal>FOR SEQUENCES IN SCHEMA</literal>
+ are not specified, then the publication starts out with an empty set
+ of tables. That is useful if tables, sequences or schemas are to be
+ added later.
</para>
<para>
@@ -241,9 +262,10 @@ CREATE PUBLICATION <replaceable class="parameter">name</replaceable>
</para>
<para>
- To add a table to a publication, the invoking user must have ownership
- rights on the table. The <command>FOR ALL TABLES</command> and
- <command>FOR TABLES IN SCHEMA</command> clauses require the invoking
+ To add a table or a sequence to a publication, the invoking user must
+ have ownership rights on the object. The <command>FOR ALL TABLES</command>,
+ <command>FOR TABLES IN SCHEMA</command>, <command>FOR ALL SEQUENCES</command>
+ and <command>FOR SEQUENCES IN SCHEMA</command> clauses require the invoking
user to be a superuser.
</para>
diff --git a/doc/src/sgml/system-views.sgml b/doc/src/sgml/system-views.sgml
index 7c716fe3276..cf858908d25 100644
--- a/doc/src/sgml/system-views.sgml
+++ b/doc/src/sgml/system-views.sgml
@@ -126,6 +126,11 @@
<entry>prepared transactions</entry>
</row>
+ <row>
+ <entry><link linkend="view-pg-publication-sequences"><structname>pg_publication_sequences</structname></link></entry>
+ <entry>publications and information of their associated sequences</entry>
+ </row>
+
<row>
<entry><link linkend="view-pg-publication-tables"><structname>pg_publication_tables</structname></link></entry>
<entry>publications and information of their associated tables</entry>
@@ -2098,6 +2103,72 @@ SELECT * FROM pg_locks pl LEFT JOIN pg_prepared_xacts ppx
</sect1>
+ <sect1 id="view-pg-publication-sequences">
+ <title><structname>pg_publication_sequences</structname></title>
+
+ <indexterm zone="view-pg-publication-sequences">
+ <primary>pg_publication_sequences</primary>
+ </indexterm>
+
+ <para>
+ The view <structname>pg_publication_sequences</structname> provides
+ information about the mapping between publications and the sequences they
+ contain. Unlike the underlying catalog
+ <link linkend="catalog-pg-publication-rel"><structname>pg_publication_rel</structname></link>,
+ this view expands
+ publications defined as <literal>FOR ALL SEQUENCES</literal>, so for such
+ publications there will be a row for each eligible sequence.
+ </para>
+
+ <table>
+ <title><structname>pg_publication_sequences</structname> Columns</title>
+ <tgroup cols="1">
+ <thead>
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ Column Type
+ </para>
+ <para>
+ Description
+ </para></entry>
+ </row>
+ </thead>
+
+ <tbody>
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ <structfield>pubname</structfield> <type>name</type>
+ (references <link linkend="catalog-pg-publication"><structname>pg_publication</structname></link>.<structfield>pubname</structfield>)
+ </para>
+ <para>
+ Name of publication
+ </para></entry>
+ </row>
+
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ <structfield>schemaname</structfield> <type>name</type>
+ (references <link linkend="catalog-pg-namespace"><structname>pg_namespace</structname></link>.<structfield>nspname</structfield>)
+ </para>
+ <para>
+ Name of schema containing sequence
+ </para></entry>
+ </row>
+
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ <structfield>sequencename</structfield> <type>name</type>
+ (references <link linkend="catalog-pg-class"><structname>pg_class</structname></link>.<structfield>relname</structfield>)
+ </para>
+ <para>
+ Name of sequence
+ </para></entry>
+ </row>
+ </tbody>
+ </tgroup>
+ </table>
+ </sect1>
+
<sect1 id="view-pg-publication-tables">
<title><structname>pg_publication_tables</structname></title>
diff --git a/src/backend/access/rmgrdesc/logicalmsgdesc.c b/src/backend/access/rmgrdesc/logicalmsgdesc.c
index 08e03aa30d1..deb89916c3f 100644
--- a/src/backend/access/rmgrdesc/logicalmsgdesc.c
+++ b/src/backend/access/rmgrdesc/logicalmsgdesc.c
@@ -40,6 +40,16 @@ logicalmsg_desc(StringInfo buf, XLogReaderState *record)
sep = " ";
}
}
+ else if (info == XLOG_LOGICAL_SEQUENCE)
+ {
+ xl_logical_sequence *xlrec = (xl_logical_sequence *) rec;
+
+ appendStringInfo(buf, "rel %u/%u/%u last: %lu log_cnt: %lu is_called: %d",
+ xlrec->locator.spcOid,
+ xlrec->locator.dbOid,
+ xlrec->locator.relNumber,
+ xlrec->last, xlrec->log_cnt, xlrec->is_called);
+ }
}
const char *
@@ -48,5 +58,8 @@ logicalmsg_identify(uint8 info)
if ((info & ~XLR_INFO_MASK) == XLOG_LOGICAL_MESSAGE)
return "MESSAGE";
+ if ((info & ~XLR_INFO_MASK) == XLOG_LOGICAL_SEQUENCE)
+ return "SEQUENCE";
+
return NULL;
}
diff --git a/src/backend/access/transam/xact.c b/src/backend/access/transam/xact.c
index 8086b857b96..df16e279b74 100644
--- a/src/backend/access/transam/xact.c
+++ b/src/backend/access/transam/xact.c
@@ -37,6 +37,7 @@
#include "catalog/storage.h"
#include "commands/async.h"
#include "commands/tablecmds.h"
+#include "commands/sequence.h"
#include "commands/trigger.h"
#include "common/pg_prng.h"
#include "executor/spi.h"
@@ -2237,6 +2238,15 @@ CommitTransaction(void)
/* Prevent cancel/die interrupt while cleaning up */
HOLD_INTERRUPTS();
+ /*
+ * Write state of sequences to WAL. We need to do this early enough so
+ * that we can still write stuff to WAL, but probably before changing
+ * the transaction state.
+ *
+ * XXX Should this happen before/after holding the interrupts?
+ */
+ AtEOXact_Sequences(true);
+
/* Commit updates to the relation map --- do this as late as possible */
AtEOXact_RelationMap(true, is_parallel_worker);
@@ -2513,6 +2523,15 @@ PrepareTransaction(void)
/* Prevent cancel/die interrupt while cleaning up */
HOLD_INTERRUPTS();
+ /*
+ * Write state of sequences to WAL. We need to do this early enough so
+ * that we can still write stuff to WAL, but probably before changing
+ * the transaction state.
+ *
+ * XXX Should this happen before/after holding the interrupts?
+ */
+ AtEOXact_Sequences(true);
+
/*
* set the current transaction state information appropriately during
* prepare processing
@@ -2748,6 +2767,12 @@ AbortTransaction(void)
TransStateAsString(s->state));
Assert(s->parent == NULL);
+ /*
+ * For transaction abort we don't need to write anything to WAL, we just
+ * do cleanup of the hash table etc.
+ */
+ AtEOXact_Sequences(false);
+
/*
* set the current transaction state information appropriately during the
* abort processing
diff --git a/src/backend/catalog/objectaddress.c b/src/backend/catalog/objectaddress.c
index c7de7232b89..706cb6ddd73 100644
--- a/src/backend/catalog/objectaddress.c
+++ b/src/backend/catalog/objectaddress.c
@@ -1959,12 +1959,14 @@ get_object_address_publication_schema(List *object, bool missing_ok)
char *pubname;
char *schemaname;
Oid schemaid;
+ char *objtype;
ObjectAddressSet(address, PublicationNamespaceRelationId, InvalidOid);
/* Fetch schema name and publication name from input list */
schemaname = strVal(linitial(object));
pubname = strVal(lsecond(object));
+ objtype = strVal(lthird(object));
schemaid = get_namespace_oid(schemaname, missing_ok);
if (!OidIsValid(schemaid))
@@ -1977,10 +1979,12 @@ get_object_address_publication_schema(List *object, bool missing_ok)
/* Find the publication schema mapping in syscache */
address.objectId =
- GetSysCacheOid2(PUBLICATIONNAMESPACEMAP,
+ GetSysCacheOid3(PUBLICATIONNAMESPACEMAP,
Anum_pg_publication_namespace_oid,
ObjectIdGetDatum(schemaid),
- ObjectIdGetDatum(pub->oid));
+ ObjectIdGetDatum(pub->oid),
+ CharGetDatum(objtype[0]));
+
if (!OidIsValid(address.objectId) && !missing_ok)
ereport(ERROR,
(errcode(ERRCODE_UNDEFINED_OBJECT),
@@ -2254,7 +2258,6 @@ pg_get_object_address(PG_FUNCTION_ARGS)
*/
switch (type)
{
- case OBJECT_PUBLICATION_NAMESPACE:
case OBJECT_USER_MAPPING:
if (list_length(name) != 1)
ereport(ERROR,
@@ -2293,6 +2296,16 @@ pg_get_object_address(PG_FUNCTION_ARGS)
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("argument list length must be exactly %d", 2)));
break;
+ case OBJECT_PUBLICATION_NAMESPACE:
+ if (list_length(name) != 1)
+ ereport(ERROR,
+ (errcode(ERRCODE_INVALID_PARAMETER_VALUE),
+ errmsg("name list length must be exactly %d", 1)));
+ if (list_length(args) != 2)
+ ereport(ERROR,
+ (errcode(ERRCODE_INVALID_PARAMETER_VALUE),
+ errmsg("argument list length must be exactly %d", 2)));
+ break;
default:
break;
}
@@ -2359,6 +2372,8 @@ pg_get_object_address(PG_FUNCTION_ARGS)
objnode = (Node *) list_make2(name, linitial(args));
break;
case OBJECT_PUBLICATION_NAMESPACE:
+ objnode = (Node *) list_make3(linitial(name), linitial(args), lsecond(args));
+ break;
case OBJECT_USER_MAPPING:
objnode = (Node *) list_make2(linitial(name), linitial(args));
break;
@@ -2914,11 +2929,12 @@ get_catalog_object_by_oid(Relation catalog, AttrNumber oidcol, Oid objectId)
*
* Get publication name and schema name from the object address into pubname and
* nspname. Both pubname and nspname are palloc'd strings which will be freed by
- * the caller.
+ * the caller. The last parameter specifies which object type is included from
+ * the schema.
*/
static bool
getPublicationSchemaInfo(const ObjectAddress *object, bool missing_ok,
- char **pubname, char **nspname)
+ char **pubname, char **nspname, char **objtype)
{
HeapTuple tup;
Form_pg_publication_namespace pnform;
@@ -2954,6 +2970,13 @@ getPublicationSchemaInfo(const ObjectAddress *object, bool missing_ok,
return false;
}
+ /*
+ * The type is always a single character, but we need to pass it as a string,
+ * so allocate two charaters and set the first one. The second one is \0.
+ */
+ *objtype = palloc0(2);
+ *objtype[0] = pnform->pntype;
+
ReleaseSysCache(tup);
return true;
}
@@ -4027,15 +4050,17 @@ getObjectDescription(const ObjectAddress *object, bool missing_ok)
{
char *pubname;
char *nspname;
+ char *objtype;
if (!getPublicationSchemaInfo(object, missing_ok,
- &pubname, &nspname))
+ &pubname, &nspname, &objtype))
break;
- appendStringInfo(&buffer, _("publication of schema %s in publication %s"),
- nspname, pubname);
+ appendStringInfo(&buffer, _("publication of schema %s in publication %s type %s"),
+ nspname, pubname, objtype);
pfree(pubname);
pfree(nspname);
+ pfree(objtype);
break;
}
@@ -5908,18 +5933,24 @@ getObjectIdentityParts(const ObjectAddress *object,
{
char *pubname;
char *nspname;
+ char *objtype;
if (!getPublicationSchemaInfo(object, missing_ok, &pubname,
- &nspname))
+ &nspname, &objtype))
break;
- appendStringInfo(&buffer, "%s in publication %s",
- nspname, pubname);
+ appendStringInfo(&buffer, "%s in publication %s type %s",
+ nspname, pubname, objtype);
if (objargs)
*objargs = list_make1(pubname);
else
pfree(pubname);
+ if (objargs)
+ *objargs = lappend(*objargs, objtype);
+ else
+ pfree(objtype);
+
if (objname)
*objname = list_make1(nspname);
else
diff --git a/src/backend/catalog/pg_publication.c b/src/backend/catalog/pg_publication.c
index 59967098b3f..249a998f1ea 100644
--- a/src/backend/catalog/pg_publication.c
+++ b/src/backend/catalog/pg_publication.c
@@ -55,9 +55,10 @@ static void publication_translate_columns(Relation targetrel, List *columns,
static void
check_publication_add_relation(Relation targetrel)
{
- /* Must be a regular or partitioned table */
+ /* Must be a regular or partitioned table, or a sequence */
if (RelationGetForm(targetrel)->relkind != RELKIND_RELATION &&
- RelationGetForm(targetrel)->relkind != RELKIND_PARTITIONED_TABLE)
+ RelationGetForm(targetrel)->relkind != RELKIND_PARTITIONED_TABLE &&
+ RelationGetForm(targetrel)->relkind != RELKIND_SEQUENCE)
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("cannot add relation \"%s\" to publication",
@@ -134,7 +135,8 @@ static bool
is_publishable_class(Oid relid, Form_pg_class reltuple)
{
return (reltuple->relkind == RELKIND_RELATION ||
- reltuple->relkind == RELKIND_PARTITIONED_TABLE) &&
+ reltuple->relkind == RELKIND_PARTITIONED_TABLE ||
+ reltuple->relkind == RELKIND_SEQUENCE) &&
!IsCatalogRelationOid(relid) &&
reltuple->relpersistence == RELPERSISTENCE_PERMANENT &&
relid >= FirstNormalObjectId;
@@ -171,6 +173,52 @@ pg_relation_is_publishable(PG_FUNCTION_ARGS)
PG_RETURN_BOOL(result);
}
+/*
+ * Check the character is a valid object type for schema publication.
+ *
+ * This recognizes either 't' for tables or 's' for sequences. Places that
+ * need to handle 'u' for unsupported relkinds need to do that explicitlyl
+ */
+static void
+AssertObjectTypeValid(char objectType)
+{
+#ifdef USE_ASSERT_CHECKING
+ Assert(objectType == PUB_OBJTYPE_SEQUENCE || objectType == PUB_OBJTYPE_TABLE);
+#endif
+}
+
+/*
+ * Determine object type matching a given a relkind value.
+ */
+char
+pub_get_object_type_for_relkind(char relkind)
+{
+ /* sequence maps directly to sequence relkind */
+ if (relkind == RELKIND_SEQUENCE)
+ return PUB_OBJTYPE_SEQUENCE;
+
+ /* for table, we match either regular or partitioned table */
+ if (relkind == RELKIND_RELATION ||
+ relkind == RELKIND_PARTITIONED_TABLE)
+ return PUB_OBJTYPE_TABLE;
+
+ return PUB_OBJTYPE_UNSUPPORTED;
+}
+
+/*
+ * Determine if publication object type matches the relkind.
+ *
+ * Returns true if the relation matches object type replicated by this schema,
+ * false otherwise.
+ */
+static bool
+pub_object_type_matches_relkind(char objectType, char relkind)
+{
+ AssertObjectTypeValid(objectType);
+
+ return (pub_get_object_type_for_relkind(relkind) == objectType);
+}
+
/*
* Filter out the partitions whose parent tables were also specified in
* the publication.
@@ -213,6 +261,8 @@ filter_partitions(List *relids)
/*
* Returns true if any schema is associated with the publication, false if no
* schema is associated with the publication.
+ *
+ * FIXME probably needs to check the object type added from the schema
*/
bool
is_schema_publication(Oid pubid)
@@ -230,7 +280,7 @@ is_schema_publication(Oid pubid)
ObjectIdGetDatum(pubid));
scan = systable_beginscan(pubschsrel,
- PublicationNamespacePnnspidPnpubidIndexId,
+ PublicationNamespacePnnspidPnpubidPntypeIndexId,
true, NULL, 1, &scankey);
tup = systable_getnext(scan);
result = HeapTupleIsValid(tup);
@@ -316,7 +366,9 @@ GetTopMostAncestorInPublication(Oid puboid, List *ancestors, int *ancestor_level
}
else
{
- aschemaPubids = GetSchemaPublications(get_rel_namespace(ancestor));
+ /* we only search for ancestors of tables, so PUB_OBJTYPE_TABLE */
+ aschemaPubids = GetSchemaPublications(get_rel_namespace(ancestor),
+ PUB_OBJTYPE_TABLE);
if (list_member_oid(aschemaPubids, puboid))
{
topmost_relid = ancestor;
@@ -585,7 +637,7 @@ pub_collist_to_bitmapset(Bitmapset *columns, Datum pubcols, MemoryContext mcxt)
* Insert new publication / schema mapping.
*/
ObjectAddress
-publication_add_schema(Oid pubid, Oid schemaid, bool if_not_exists)
+publication_add_schema(Oid pubid, Oid schemaid, char objectType, bool if_not_exists)
{
Relation rel;
HeapTuple tup;
@@ -597,6 +649,8 @@ publication_add_schema(Oid pubid, Oid schemaid, bool if_not_exists)
ObjectAddress myself,
referenced;
+ AssertObjectTypeValid(objectType);
+
rel = table_open(PublicationNamespaceRelationId, RowExclusiveLock);
/*
@@ -604,9 +658,10 @@ publication_add_schema(Oid pubid, Oid schemaid, bool if_not_exists)
* duplicates, it's here just to provide nicer error message in common
* case. The real protection is the unique key on the catalog.
*/
- if (SearchSysCacheExists2(PUBLICATIONNAMESPACEMAP,
+ if (SearchSysCacheExists3(PUBLICATIONNAMESPACEMAP,
ObjectIdGetDatum(schemaid),
- ObjectIdGetDatum(pubid)))
+ ObjectIdGetDatum(pubid),
+ CharGetDatum(objectType)))
{
table_close(rel, RowExclusiveLock);
@@ -632,6 +687,8 @@ publication_add_schema(Oid pubid, Oid schemaid, bool if_not_exists)
ObjectIdGetDatum(pubid);
values[Anum_pg_publication_namespace_pnnspid - 1] =
ObjectIdGetDatum(schemaid);
+ values[Anum_pg_publication_namespace_pntype - 1] =
+ CharGetDatum(objectType);
tup = heap_form_tuple(RelationGetDescr(rel), values, nulls);
@@ -657,7 +714,7 @@ publication_add_schema(Oid pubid, Oid schemaid, bool if_not_exists)
* publication_add_relation for why we need to consider all the
* partitions.
*/
- schemaRels = GetSchemaPublicationRelations(schemaid,
+ schemaRels = GetSchemaPublicationRelations(schemaid, objectType,
PUBLICATION_PART_ALL);
InvalidatePublicationRels(schemaRels);
@@ -691,11 +748,14 @@ GetRelationPublications(Oid relid)
/*
* Gets list of relation oids for a publication.
*
- * This should only be used FOR TABLE publications, the FOR ALL TABLES
- * should use GetAllTablesPublicationRelations().
+ * This should only be used FOR TABLE / FOR SEQUENCE publications, the FOR
+ * ALL TABLES / SEQUENCES should use GetAllTablesPublicationRelations()
+ * and GetAllSequencesPublicationRelations().
+ *
+ * XXX pub_partopt only matters for tables, not sequences.
*/
List *
-GetPublicationRelations(Oid pubid, PublicationPartOpt pub_partopt)
+GetPublicationRelations(Oid pubid, char objectType, PublicationPartOpt pub_partopt)
{
List *result;
Relation pubrelsrel;
@@ -703,6 +763,8 @@ GetPublicationRelations(Oid pubid, PublicationPartOpt pub_partopt)
SysScanDesc scan;
HeapTuple tup;
+ AssertObjectTypeValid(objectType);
+
/* Find all publications associated with the relation. */
pubrelsrel = table_open(PublicationRelRelationId, AccessShareLock);
@@ -717,11 +779,29 @@ GetPublicationRelations(Oid pubid, PublicationPartOpt pub_partopt)
result = NIL;
while (HeapTupleIsValid(tup = systable_getnext(scan)))
{
+ char relkind;
Form_pg_publication_rel pubrel;
pubrel = (Form_pg_publication_rel) GETSTRUCT(tup);
- result = GetPubPartitionOptionRelations(result, pub_partopt,
- pubrel->prrelid);
+ relkind = get_rel_relkind(pubrel->prrelid);
+
+ /*
+ * If the relkind does not match the requested object type, ignore the
+ * relation. For example we might be interested only in sequences, so
+ * we ignore tables.
+ */
+ if (!pub_object_type_matches_relkind(objectType, relkind))
+ continue;
+
+ /*
+ * We don't have partitioned sequences, so just add them to the list.
+ * Otherwise consider adding all child relations, if requested.
+ */
+ if (relkind == RELKIND_SEQUENCE)
+ result = lappend_oid(result, pubrel->prrelid);
+ else
+ result = GetPubPartitionOptionRelations(result, pub_partopt,
+ pubrel->prrelid);
}
systable_endscan(scan);
@@ -771,6 +851,43 @@ GetAllTablesPublications(void)
return result;
}
+/*
+ * Gets list of publication oids for publications marked as FOR ALL SEQUENCES.
+ */
+List *
+GetAllSequencesPublications(void)
+{
+ List *result;
+ Relation rel;
+ ScanKeyData scankey;
+ SysScanDesc scan;
+ HeapTuple tup;
+
+ /* Find all publications that are marked as for all sequences. */
+ rel = table_open(PublicationRelationId, AccessShareLock);
+
+ ScanKeyInit(&scankey,
+ Anum_pg_publication_puballsequences,
+ BTEqualStrategyNumber, F_BOOLEQ,
+ BoolGetDatum(true));
+
+ scan = systable_beginscan(rel, InvalidOid, false,
+ NULL, 1, &scankey);
+
+ result = NIL;
+ while (HeapTupleIsValid(tup = systable_getnext(scan)))
+ {
+ Oid oid = ((Form_pg_publication) GETSTRUCT(tup))->oid;
+
+ result = lappend_oid(result, oid);
+ }
+
+ systable_endscan(scan);
+ table_close(rel, AccessShareLock);
+
+ return result;
+}
+
/*
* Gets list of all relation published by FOR ALL TABLES publication(s).
*
@@ -837,28 +954,38 @@ GetAllTablesPublicationRelations(bool pubviaroot)
/*
* Gets the list of schema oids for a publication.
*
- * This should only be used FOR TABLES IN SCHEMA publications.
+ * This should only be used FOR TABLES IN SCHEMA and FOR SEQUENCES IN SCHEMA publications.
+ *
+ * 'objectType' determines whether to get FOR TABLE or FOR SEQUENCES schemas
*/
List *
-GetPublicationSchemas(Oid pubid)
+GetPublicationSchemas(Oid pubid, char objectType)
{
List *result = NIL;
Relation pubschsrel;
- ScanKeyData scankey;
+ ScanKeyData scankey[2];
SysScanDesc scan;
HeapTuple tup;
+ AssertObjectTypeValid(objectType);
+
/* Find all schemas associated with the publication */
pubschsrel = table_open(PublicationNamespaceRelationId, AccessShareLock);
- ScanKeyInit(&scankey,
+ ScanKeyInit(&scankey[0],
Anum_pg_publication_namespace_pnpubid,
BTEqualStrategyNumber, F_OIDEQ,
ObjectIdGetDatum(pubid));
+ ScanKeyInit(&scankey[1],
+ Anum_pg_publication_namespace_pntype,
+ BTEqualStrategyNumber, F_CHAREQ,
+ CharGetDatum(objectType));
+
+
scan = systable_beginscan(pubschsrel,
- PublicationNamespacePnnspidPnpubidIndexId,
- true, NULL, 1, &scankey);
+ PublicationNamespacePnnspidPnpubidPntypeIndexId,
+ true, NULL, 2, scankey);
while (HeapTupleIsValid(tup = systable_getnext(scan)))
{
Form_pg_publication_namespace pubsch;
@@ -876,14 +1003,26 @@ GetPublicationSchemas(Oid pubid)
/*
* Gets the list of publication oids associated with a specified schema.
+ *
+ * objectType specifies whether we're looking for schemas including tables or
+ * sequences.
+ *
+ * Note: relcache calls this for all object types, not just tables and sequences.
+ * Which is why we handle the PUB_OBJTYPE_UNSUPPORTED object type too.
*/
List *
-GetSchemaPublications(Oid schemaid)
+GetSchemaPublications(Oid schemaid, char objectType)
{
List *result = NIL;
CatCList *pubschlist;
int i;
+ /* unsupported object type */
+ if (objectType == PUB_OBJTYPE_UNSUPPORTED)
+ return result;
+
+ AssertObjectTypeValid(objectType);
+
/* Find all publications associated with the schema */
pubschlist = SearchSysCacheList1(PUBLICATIONNAMESPACEMAP,
ObjectIdGetDatum(schemaid));
@@ -891,6 +1030,11 @@ GetSchemaPublications(Oid schemaid)
{
HeapTuple tup = &pubschlist->members[i]->tuple;
Oid pubid = ((Form_pg_publication_namespace) GETSTRUCT(tup))->pnpubid;
+ char pntype = ((Form_pg_publication_namespace) GETSTRUCT(tup))->pntype;
+
+ /* Skip schemas publishing a different object type. */
+ if (pntype != objectType)
+ continue;
result = lappend_oid(result, pubid);
}
@@ -902,9 +1046,12 @@ GetSchemaPublications(Oid schemaid)
/*
* Get the list of publishable relation oids for a specified schema.
+ *
+ * objectType specifies whether this is FOR TABLES / SEQUENCES IN SCHEMA
*/
List *
-GetSchemaPublicationRelations(Oid schemaid, PublicationPartOpt pub_partopt)
+GetSchemaPublicationRelations(Oid schemaid, char objectType,
+ PublicationPartOpt pub_partopt)
{
Relation classRel;
ScanKeyData key[1];
@@ -913,6 +1060,7 @@ GetSchemaPublicationRelations(Oid schemaid, PublicationPartOpt pub_partopt)
List *result = NIL;
Assert(OidIsValid(schemaid));
+ AssertObjectTypeValid(objectType);
classRel = table_open(RelationRelationId, AccessShareLock);
@@ -933,9 +1081,16 @@ GetSchemaPublicationRelations(Oid schemaid, PublicationPartOpt pub_partopt)
continue;
relkind = get_rel_relkind(relid);
- if (relkind == RELKIND_RELATION)
- result = lappend_oid(result, relid);
- else if (relkind == RELKIND_PARTITIONED_TABLE)
+
+ /* Skip if the relkind does not match FOR ALL TABLES / SEQUENCES. */
+ if (!pub_object_type_matches_relkind(objectType, relkind))
+ continue;
+
+ /*
+ * If the object is a partitioned table, lookup all the child relations
+ * (if requested). Otherwise just add the object to the list.
+ */
+ if (relkind == RELKIND_PARTITIONED_TABLE)
{
List *partitionrels = NIL;
@@ -949,6 +1104,8 @@ GetSchemaPublicationRelations(Oid schemaid, PublicationPartOpt pub_partopt)
relForm->oid);
result = list_concat_unique_oid(result, partitionrels);
}
+ else /* non-partitioned tables and sequences */
+ result = lappend_oid(result, relid);
}
table_endscan(scan);
@@ -957,28 +1114,68 @@ GetSchemaPublicationRelations(Oid schemaid, PublicationPartOpt pub_partopt)
}
/*
- * Gets the list of all relations published by FOR TABLES IN SCHEMA
- * publication.
+ * Gets the list of all relations published by FOR TABLES IN SCHEMA of
+ * FOR SEQUENCES IN SCHEMA publication.
*/
List *
-GetAllSchemaPublicationRelations(Oid pubid, PublicationPartOpt pub_partopt)
+GetAllSchemaPublicationRelations(Oid pubid, char objectType,
+ PublicationPartOpt pub_partopt)
{
List *result = NIL;
- List *pubschemalist = GetPublicationSchemas(pubid);
+ List *pubschemalist = GetPublicationSchemas(pubid, objectType);
ListCell *cell;
+ AssertObjectTypeValid(objectType);
+
foreach(cell, pubschemalist)
{
Oid schemaid = lfirst_oid(cell);
List *schemaRels = NIL;
- schemaRels = GetSchemaPublicationRelations(schemaid, pub_partopt);
+ schemaRels = GetSchemaPublicationRelations(schemaid, objectType,
+ pub_partopt);
result = list_concat(result, schemaRels);
}
return result;
}
+/*
+ * Gets list of all relation published by FOR ALL SEQUENCES publication(s).
+ */
+List *
+GetAllSequencesPublicationRelations(void)
+{
+ Relation classRel;
+ ScanKeyData key[1];
+ TableScanDesc scan;
+ HeapTuple tuple;
+ List *result = NIL;
+
+ classRel = table_open(RelationRelationId, AccessShareLock);
+
+ ScanKeyInit(&key[0],
+ Anum_pg_class_relkind,
+ BTEqualStrategyNumber, F_CHAREQ,
+ CharGetDatum(RELKIND_SEQUENCE));
+
+ scan = table_beginscan_catalog(classRel, 1, key);
+
+ while ((tuple = heap_getnext(scan, ForwardScanDirection)) != NULL)
+ {
+ Form_pg_class relForm = (Form_pg_class) GETSTRUCT(tuple);
+ Oid relid = relForm->oid;
+
+ if (is_publishable_class(relid, relForm))
+ result = lappend_oid(result, relid);
+ }
+
+ table_endscan(scan);
+
+ table_close(classRel, AccessShareLock);
+ return result;
+}
+
/*
* Get publication using oid
*
@@ -1001,10 +1198,12 @@ GetPublication(Oid pubid)
pub->oid = pubid;
pub->name = pstrdup(NameStr(pubform->pubname));
pub->alltables = pubform->puballtables;
+ pub->allsequences = pubform->puballsequences;
pub->pubactions.pubinsert = pubform->pubinsert;
pub->pubactions.pubupdate = pubform->pubupdate;
pub->pubactions.pubdelete = pubform->pubdelete;
pub->pubactions.pubtruncate = pubform->pubtruncate;
+ pub->pubactions.pubsequence = pubform->pubsequence;
pub->pubviaroot = pubform->pubviaroot;
ReleaseSysCache(tup);
@@ -1066,10 +1265,12 @@ pg_get_publication_tables(PG_FUNCTION_ARGS)
*schemarelids;
relids = GetPublicationRelations(publication->oid,
+ PUB_OBJTYPE_TABLE,
publication->pubviaroot ?
PUBLICATION_PART_ROOT :
PUBLICATION_PART_LEAF);
schemarelids = GetAllSchemaPublicationRelations(publication->oid,
+ PUB_OBJTYPE_TABLE,
publication->pubviaroot ?
PUBLICATION_PART_ROOT :
PUBLICATION_PART_LEAF);
@@ -1128,9 +1329,10 @@ pg_get_publication_tables(PG_FUNCTION_ARGS)
* FOR TABLES IN SCHEMA publications.
*/
if (!publication->alltables &&
- !SearchSysCacheExists2(PUBLICATIONNAMESPACEMAP,
+ !SearchSysCacheExists3(PUBLICATIONNAMESPACEMAP,
ObjectIdGetDatum(schemaid),
- ObjectIdGetDatum(publication->oid)))
+ ObjectIdGetDatum(publication->oid),
+ CharGetDatum(PUB_OBJTYPE_TABLE)))
pubtuple = SearchSysCacheCopy2(PUBLICATIONRELMAP,
ObjectIdGetDatum(relid),
ObjectIdGetDatum(publication->oid));
@@ -1160,3 +1362,71 @@ pg_get_publication_tables(PG_FUNCTION_ARGS)
SRF_RETURN_DONE(funcctx);
}
+
+/*
+ * Returns Oids of sequences in a publication.
+ */
+Datum
+pg_get_publication_sequences(PG_FUNCTION_ARGS)
+{
+ FuncCallContext *funcctx;
+ char *pubname = text_to_cstring(PG_GETARG_TEXT_PP(0));
+ Publication *publication;
+ List *sequences;
+
+ /* stuff done only on the first call of the function */
+ if (SRF_IS_FIRSTCALL())
+ {
+ MemoryContext oldcontext;
+
+ /* create a function context for cross-call persistence */
+ funcctx = SRF_FIRSTCALL_INIT();
+
+ /* switch to memory context appropriate for multiple function calls */
+ oldcontext = MemoryContextSwitchTo(funcctx->multi_call_memory_ctx);
+
+ publication = GetPublicationByName(pubname, false);
+
+ /*
+ * Publications support partitioned tables, although all changes are
+ * replicated using leaf partition identity and schema, so we only
+ * need those.
+ */
+ if (publication->allsequences)
+ sequences = GetAllSequencesPublicationRelations();
+ else
+ {
+ List *relids,
+ *schemarelids;
+
+ relids = GetPublicationRelations(publication->oid,
+ PUB_OBJTYPE_SEQUENCE,
+ publication->pubviaroot ?
+ PUBLICATION_PART_ROOT :
+ PUBLICATION_PART_LEAF);
+ schemarelids = GetAllSchemaPublicationRelations(publication->oid,
+ PUB_OBJTYPE_SEQUENCE,
+ publication->pubviaroot ?
+ PUBLICATION_PART_ROOT :
+ PUBLICATION_PART_LEAF);
+ sequences = list_concat_unique_oid(relids, schemarelids);
+ }
+
+ funcctx->user_fctx = (void *) sequences;
+
+ MemoryContextSwitchTo(oldcontext);
+ }
+
+ /* stuff done on every call of the function */
+ funcctx = SRF_PERCALL_SETUP();
+ sequences = (List *) funcctx->user_fctx;
+
+ if (funcctx->call_cntr < list_length(sequences))
+ {
+ Oid relid = list_nth_oid(sequences, funcctx->call_cntr);
+
+ SRF_RETURN_NEXT(funcctx, ObjectIdGetDatum(relid));
+ }
+
+ SRF_RETURN_DONE(funcctx);
+}
diff --git a/src/backend/catalog/system_views.sql b/src/backend/catalog/system_views.sql
index 2d8104b0907..a6ef8ceaf0f 100644
--- a/src/backend/catalog/system_views.sql
+++ b/src/backend/catalog/system_views.sql
@@ -381,6 +381,16 @@ CREATE VIEW pg_publication_tables AS
pg_class C JOIN pg_namespace N ON (N.oid = C.relnamespace)
WHERE C.oid = GPT.relid;
+CREATE VIEW pg_publication_sequences AS
+ SELECT
+ P.pubname AS pubname,
+ N.nspname AS schemaname,
+ C.relname AS sequencename
+ FROM pg_publication P,
+ LATERAL pg_get_publication_sequences(P.pubname) GPS,
+ pg_class C JOIN pg_namespace N ON (N.oid = C.relnamespace)
+ WHERE C.oid = GPS.relid;
+
CREATE VIEW pg_locks AS
SELECT * FROM pg_lock_status() AS L;
diff --git a/src/backend/commands/publicationcmds.c b/src/backend/commands/publicationcmds.c
index a8b75eb1be8..9ef6c425dcb 100644
--- a/src/backend/commands/publicationcmds.c
+++ b/src/backend/commands/publicationcmds.c
@@ -16,6 +16,7 @@
#include "access/genam.h"
#include "access/htup_details.h"
+#include "access/relation.h"
#include "access/table.h"
#include "access/xact.h"
#include "catalog/catalog.h"
@@ -67,16 +68,16 @@ typedef struct rf_context
Oid parentid; /* relid of the parent relation */
} rf_context;
-static List *OpenTableList(List *tables);
-static void CloseTableList(List *rels);
+static List *OpenRelationList(List *rels, char objectType);
+static void CloseRelationList(List *rels);
static void LockSchemaList(List *schemalist);
-static void PublicationAddTables(Oid pubid, List *rels, bool if_not_exists,
- AlterPublicationStmt *stmt);
-static void PublicationDropTables(Oid pubid, List *rels, bool missing_ok);
-static void PublicationAddSchemas(Oid pubid, List *schemas, bool if_not_exists,
- AlterPublicationStmt *stmt);
-static void PublicationDropSchemas(Oid pubid, List *schemas, bool missing_ok);
-
+static void PublicationAddRelations(Oid pubid, List *rels, bool if_not_exists,
+ AlterPublicationStmt *stmt);
+static void PublicationDropRelations(Oid pubid, List *rels, bool missing_ok);
+static void PublicationAddSchemas(Oid pubid, List *schemas, char objectType,
+ bool if_not_exists, AlterPublicationStmt *stmt);
+static void PublicationDropSchemas(Oid pubid, List *schemas, char objectType,
+ bool missing_ok);
static void
parse_publication_options(ParseState *pstate,
@@ -96,6 +97,7 @@ parse_publication_options(ParseState *pstate,
pubactions->pubupdate = true;
pubactions->pubdelete = true;
pubactions->pubtruncate = true;
+ pubactions->pubsequence = true;
*publish_via_partition_root = false;
/* Parse options */
@@ -120,6 +122,7 @@ parse_publication_options(ParseState *pstate,
pubactions->pubupdate = false;
pubactions->pubdelete = false;
pubactions->pubtruncate = false;
+ pubactions->pubsequence = false;
*publish_given = true;
publish = defGetString(defel);
@@ -143,6 +146,8 @@ parse_publication_options(ParseState *pstate,
pubactions->pubdelete = true;
else if (strcmp(publish_opt, "truncate") == 0)
pubactions->pubtruncate = true;
+ else if (strcmp(publish_opt, "sequence") == 0)
+ pubactions->pubsequence = true;
else
ereport(ERROR,
(errcode(ERRCODE_SYNTAX_ERROR),
@@ -170,7 +175,8 @@ parse_publication_options(ParseState *pstate,
*/
static void
ObjectsInPublicationToOids(List *pubobjspec_list, ParseState *pstate,
- List **rels, List **schemas)
+ List **tables, List **sequences,
+ List **tables_schemas, List **sequences_schemas)
{
ListCell *cell;
PublicationObjSpec *pubobj;
@@ -188,13 +194,22 @@ ObjectsInPublicationToOids(List *pubobjspec_list, ParseState *pstate,
switch (pubobj->pubobjtype)
{
case PUBLICATIONOBJ_TABLE:
- *rels = lappend(*rels, pubobj->pubtable);
+ *tables = lappend(*tables, pubobj->pubtable);
+ break;
+ case PUBLICATIONOBJ_SEQUENCE:
+ *sequences = lappend(*sequences, pubobj->pubtable);
break;
case PUBLICATIONOBJ_TABLES_IN_SCHEMA:
schemaid = get_namespace_oid(pubobj->name, false);
/* Filter out duplicates if user specifies "sch1, sch1" */
- *schemas = list_append_unique_oid(*schemas, schemaid);
+ *tables_schemas = list_append_unique_oid(*tables_schemas, schemaid);
+ break;
+ case PUBLICATIONOBJ_SEQUENCES_IN_SCHEMA:
+ schemaid = get_namespace_oid(pubobj->name, false);
+
+ /* Filter out duplicates if user specifies "sch1, sch1" */
+ *sequences_schemas = list_append_unique_oid(*sequences_schemas, schemaid);
break;
case PUBLICATIONOBJ_TABLES_IN_CUR_SCHEMA:
search_path = fetch_search_path(false);
@@ -207,7 +222,20 @@ ObjectsInPublicationToOids(List *pubobjspec_list, ParseState *pstate,
list_free(search_path);
/* Filter out duplicates if user specifies "sch1, sch1" */
- *schemas = list_append_unique_oid(*schemas, schemaid);
+ *tables_schemas = list_append_unique_oid(*tables_schemas, schemaid);
+ break;
+ case PUBLICATIONOBJ_SEQUENCES_IN_CUR_SCHEMA:
+ search_path = fetch_search_path(false);
+ if (search_path == NIL) /* nothing valid in search_path? */
+ ereport(ERROR,
+ errcode(ERRCODE_UNDEFINED_SCHEMA),
+ errmsg("no schema has been selected for CURRENT_SCHEMA"));
+
+ schemaid = linitial_oid(search_path);
+ list_free(search_path);
+
+ /* Filter out duplicates if user specifies "sch1, sch1" */
+ *sequences_schemas = list_append_unique_oid(*sequences_schemas, schemaid);
break;
default:
/* shouldn't happen */
@@ -733,6 +761,7 @@ CheckPubRelationColumnList(char *pubname, List *tables,
ObjectAddress
CreatePublication(ParseState *pstate, CreatePublicationStmt *stmt)
{
+ ListCell *lc;
Relation rel;
ObjectAddress myself;
Oid puboid;
@@ -744,8 +773,23 @@ CreatePublication(ParseState *pstate, CreatePublicationStmt *stmt)
bool publish_via_partition_root_given;
bool publish_via_partition_root;
AclResult aclresult;
- List *relations = NIL;
- List *schemaidlist = NIL;
+ List *tables = NIL;
+ List *sequences = NIL;
+ List *tables_schemaidlist = NIL;
+ List *sequences_schemaidlist = NIL;
+
+ bool for_all_tables = false;
+ bool for_all_sequences = false;
+
+ /* Translate the list of object types (represented by strings) to bool flags. */
+ foreach (lc, stmt->for_all_objects)
+ {
+ char *val = strVal(lfirst(lc));
+ if (strcmp(val, "tables") == 0)
+ for_all_tables = true;
+ else if (strcmp(val, "sequences") == 0)
+ for_all_sequences = true;
+ }
/* must have CREATE privilege on database */
aclresult = pg_database_aclcheck(MyDatabaseId, GetUserId(), ACL_CREATE);
@@ -754,11 +798,17 @@ CreatePublication(ParseState *pstate, CreatePublicationStmt *stmt)
get_database_name(MyDatabaseId));
/* FOR ALL TABLES requires superuser */
- if (stmt->for_all_tables && !superuser())
+ if (for_all_tables && !superuser())
ereport(ERROR,
(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
errmsg("must be superuser to create FOR ALL TABLES publication")));
+ /* FOR ALL SEQUENCES requires superuser */
+ if (for_all_sequences && !superuser())
+ ereport(ERROR,
+ (errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
+ errmsg("must be superuser to create FOR ALL SEQUENCES publication")));
+
rel = table_open(PublicationRelationId, RowExclusiveLock);
/* Check if name is used */
@@ -788,7 +838,9 @@ CreatePublication(ParseState *pstate, CreatePublicationStmt *stmt)
Anum_pg_publication_oid);
values[Anum_pg_publication_oid - 1] = ObjectIdGetDatum(puboid);
values[Anum_pg_publication_puballtables - 1] =
- BoolGetDatum(stmt->for_all_tables);
+ BoolGetDatum(for_all_tables);
+ values[Anum_pg_publication_puballsequences - 1] =
+ BoolGetDatum(for_all_sequences);
values[Anum_pg_publication_pubinsert - 1] =
BoolGetDatum(pubactions.pubinsert);
values[Anum_pg_publication_pubupdate - 1] =
@@ -797,6 +849,8 @@ CreatePublication(ParseState *pstate, CreatePublicationStmt *stmt)
BoolGetDatum(pubactions.pubdelete);
values[Anum_pg_publication_pubtruncate - 1] =
BoolGetDatum(pubactions.pubtruncate);
+ values[Anum_pg_publication_pubsequence - 1] =
+ BoolGetDatum(pubactions.pubsequence);
values[Anum_pg_publication_pubviaroot - 1] =
BoolGetDatum(publish_via_partition_root);
@@ -814,46 +868,90 @@ CreatePublication(ParseState *pstate, CreatePublicationStmt *stmt)
CommandCounterIncrement();
/* Associate objects with the publication. */
- if (stmt->for_all_tables)
+ if (for_all_tables || for_all_sequences)
{
/* Invalidate relcache so that publication info is rebuilt. */
CacheInvalidateRelcacheAll();
}
- else
+
+ /*
+ * If the publication might have either tables or sequences (directly or
+ * through a schema), process that.
+ */
+ if (!for_all_tables || !for_all_sequences)
{
- ObjectsInPublicationToOids(stmt->pubobjects, pstate, &relations,
- &schemaidlist);
+ ObjectsInPublicationToOids(stmt->pubobjects, pstate,
+ &tables, &sequences,
+ &tables_schemaidlist,
+ &sequences_schemaidlist);
/* FOR TABLES IN SCHEMA requires superuser */
- if (schemaidlist != NIL && !superuser())
+ if (list_length(tables_schemaidlist) > 0 && !superuser())
ereport(ERROR,
errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
errmsg("must be superuser to create FOR TABLES IN SCHEMA publication"));
- if (relations != NIL)
+ /* FOR SEQUENCES IN SCHEMA requires superuser */
+ if (list_length(sequences_schemaidlist) > 0 && !superuser())
+ ereport(ERROR,
+ errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
+ errmsg("must be superuser to create FOR SEQUENCES IN SCHEMA publication"));
+
+ if (tables != NIL)
+ {
+ List *rels;
+
+ rels = OpenRelationList(tables, PUB_OBJTYPE_TABLE);
+ TransformPubWhereClauses(rels, pstate->p_sourcetext,
+ publish_via_partition_root);
+
+ CheckPubRelationColumnList(stmt->pubname, rels,
+ tables_schemaidlist != NIL,
+ publish_via_partition_root);
+
+ PublicationAddRelations(puboid, rels, true, NULL);
+ CloseRelationList(rels);
+ }
+
+ /* sequences added directly */
+ if (sequences != NIL)
{
List *rels;
- rels = OpenTableList(relations);
+ rels = OpenRelationList(sequences, PUB_OBJTYPE_SEQUENCE);
TransformPubWhereClauses(rels, pstate->p_sourcetext,
publish_via_partition_root);
CheckPubRelationColumnList(stmt->pubname, rels,
- schemaidlist != NIL,
+ sequences_schemaidlist != NIL,
publish_via_partition_root);
- PublicationAddTables(puboid, rels, true, NULL);
- CloseTableList(rels);
+ PublicationAddRelations(puboid, rels, true, NULL);
+ CloseRelationList(rels);
+ }
+
+ /* tables added through a schema */
+ if (list_length(tables_schemaidlist) > 0)
+ {
+ /*
+ * Schema lock is held until the publication is created to prevent
+ * concurrent schema deletion.
+ */
+ LockSchemaList(tables_schemaidlist);
+ PublicationAddSchemas(puboid, tables_schemaidlist,
+ PUB_OBJTYPE_TABLE, true, NULL);
}
- if (schemaidlist != NIL)
+ /* sequences added through a schema */
+ if (list_length(sequences_schemaidlist) > 0)
{
/*
* Schema lock is held until the publication is created to prevent
* concurrent schema deletion.
*/
- LockSchemaList(schemaidlist);
- PublicationAddSchemas(puboid, schemaidlist, true, NULL);
+ LockSchemaList(sequences_schemaidlist);
+ PublicationAddSchemas(puboid, sequences_schemaidlist,
+ PUB_OBJTYPE_SEQUENCE, true, NULL);
}
}
@@ -916,6 +1014,7 @@ AlterPublicationOptions(ParseState *pstate, AlterPublicationStmt *stmt,
AccessShareLock);
root_relids = GetPublicationRelations(pubform->oid,
+ PUB_OBJTYPE_TABLE,
PUBLICATION_PART_ROOT);
foreach(lc, root_relids)
@@ -995,6 +1094,9 @@ AlterPublicationOptions(ParseState *pstate, AlterPublicationStmt *stmt,
values[Anum_pg_publication_pubtruncate - 1] = BoolGetDatum(pubactions.pubtruncate);
replaces[Anum_pg_publication_pubtruncate - 1] = true;
+
+ values[Anum_pg_publication_pubsequence - 1] = BoolGetDatum(pubactions.pubsequence);
+ replaces[Anum_pg_publication_pubsequence - 1] = true;
}
if (publish_via_partition_root_given)
@@ -1014,7 +1116,7 @@ AlterPublicationOptions(ParseState *pstate, AlterPublicationStmt *stmt,
pubform = (Form_pg_publication) GETSTRUCT(tup);
/* Invalidate the relcache. */
- if (pubform->puballtables)
+ if (pubform->puballtables || pubform->puballsequences)
{
CacheInvalidateRelcacheAll();
}
@@ -1030,6 +1132,7 @@ AlterPublicationOptions(ParseState *pstate, AlterPublicationStmt *stmt,
*/
if (root_relids == NIL)
relids = GetPublicationRelations(pubform->oid,
+ PUB_OBJTYPE_TABLE,
PUBLICATION_PART_ALL);
else
{
@@ -1043,7 +1146,20 @@ AlterPublicationOptions(ParseState *pstate, AlterPublicationStmt *stmt,
lfirst_oid(lc));
}
+ /* tables */
+ schemarelids = GetAllSchemaPublicationRelations(pubform->oid,
+ PUB_OBJTYPE_TABLE,
+ PUBLICATION_PART_ALL);
+ relids = list_concat_unique_oid(relids, schemarelids);
+
+ /* sequences */
+ relids = list_concat_unique_oid(relids,
+ GetPublicationRelations(pubform->oid,
+ PUB_OBJTYPE_SEQUENCE,
+ PUBLICATION_PART_ALL));
+
schemarelids = GetAllSchemaPublicationRelations(pubform->oid,
+ PUB_OBJTYPE_SEQUENCE,
PUBLICATION_PART_ALL);
relids = list_concat_unique_oid(relids, schemarelids);
@@ -1098,7 +1214,7 @@ AlterPublicationTables(AlterPublicationStmt *stmt, HeapTuple tup,
if (!tables && stmt->action != AP_SetObjects)
return;
- rels = OpenTableList(tables);
+ rels = OpenRelationList(tables, PUB_OBJTYPE_TABLE);
if (stmt->action == AP_AddObjects)
{
@@ -1109,13 +1225,14 @@ AlterPublicationTables(AlterPublicationStmt *stmt, HeapTuple tup,
CheckPubRelationColumnList(stmt->pubname, rels, publish_schema,
pubform->pubviaroot);
- PublicationAddTables(pubid, rels, false, stmt);
+ PublicationAddRelations(pubid, rels, false, stmt);
}
else if (stmt->action == AP_DropObjects)
- PublicationDropTables(pubid, rels, false);
+ PublicationDropRelations(pubid, rels, false);
else /* AP_SetObjects */
{
List *oldrelids = GetPublicationRelations(pubid,
+ PUB_OBJTYPE_TABLE,
PUBLICATION_PART_ROOT);
List *delrels = NIL;
ListCell *oldlc;
@@ -1232,18 +1349,18 @@ AlterPublicationTables(AlterPublicationStmt *stmt, HeapTuple tup,
}
/* And drop them. */
- PublicationDropTables(pubid, delrels, true);
+ PublicationDropRelations(pubid, delrels, true);
/*
* Don't bother calculating the difference for adding, we'll catch and
* skip existing ones when doing catalog update.
*/
- PublicationAddTables(pubid, rels, true, stmt);
+ PublicationAddRelations(pubid, rels, true, stmt);
- CloseTableList(delrels);
+ CloseRelationList(delrels);
}
- CloseTableList(rels);
+ CloseRelationList(rels);
}
/*
@@ -1253,7 +1370,8 @@ AlterPublicationTables(AlterPublicationStmt *stmt, HeapTuple tup,
*/
static void
AlterPublicationSchemas(AlterPublicationStmt *stmt,
- HeapTuple tup, List *schemaidlist)
+ HeapTuple tup, List *schemaidlist,
+ char objectType)
{
Form_pg_publication pubform = (Form_pg_publication) GETSTRUCT(tup);
@@ -1275,7 +1393,7 @@ AlterPublicationSchemas(AlterPublicationStmt *stmt,
ListCell *lc;
List *reloids;
- reloids = GetPublicationRelations(pubform->oid, PUBLICATION_PART_ROOT);
+ reloids = GetPublicationRelations(pubform->oid, objectType, PUBLICATION_PART_ROOT);
foreach(lc, reloids)
{
@@ -1302,13 +1420,13 @@ AlterPublicationSchemas(AlterPublicationStmt *stmt,
ReleaseSysCache(coltuple);
}
- PublicationAddSchemas(pubform->oid, schemaidlist, false, stmt);
+ PublicationAddSchemas(pubform->oid, schemaidlist, objectType, false, stmt);
}
else if (stmt->action == AP_DropObjects)
- PublicationDropSchemas(pubform->oid, schemaidlist, false);
+ PublicationDropSchemas(pubform->oid, schemaidlist, objectType, false);
else /* AP_SetObjects */
{
- List *oldschemaids = GetPublicationSchemas(pubform->oid);
+ List *oldschemaids = GetPublicationSchemas(pubform->oid, objectType);
List *delschemas = NIL;
/* Identify which schemas should be dropped */
@@ -1321,13 +1439,13 @@ AlterPublicationSchemas(AlterPublicationStmt *stmt,
LockSchemaList(delschemas);
/* And drop them */
- PublicationDropSchemas(pubform->oid, delschemas, true);
+ PublicationDropSchemas(pubform->oid, delschemas, objectType, true);
/*
* Don't bother calculating the difference for adding, we'll catch and
* skip existing ones when doing catalog update.
*/
- PublicationAddSchemas(pubform->oid, schemaidlist, true, stmt);
+ PublicationAddSchemas(pubform->oid, schemaidlist, objectType, true, stmt);
}
}
@@ -1337,12 +1455,13 @@ AlterPublicationSchemas(AlterPublicationStmt *stmt,
*/
static void
CheckAlterPublication(AlterPublicationStmt *stmt, HeapTuple tup,
- List *tables, List *schemaidlist)
+ List *tables, List *tables_schemaidlist,
+ List *sequences, List *sequences_schemaidlist)
{
Form_pg_publication pubform = (Form_pg_publication) GETSTRUCT(tup);
if ((stmt->action == AP_AddObjects || stmt->action == AP_SetObjects) &&
- schemaidlist && !superuser())
+ (tables_schemaidlist || sequences_schemaidlist) && !superuser())
ereport(ERROR,
(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
errmsg("must be superuser to add or set schemas")));
@@ -1351,13 +1470,24 @@ CheckAlterPublication(AlterPublicationStmt *stmt, HeapTuple tup,
* Check that user is allowed to manipulate the publication tables in
* schema
*/
- if (schemaidlist && pubform->puballtables)
+ if (tables_schemaidlist && pubform->puballtables)
ereport(ERROR,
(errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
errmsg("publication \"%s\" is defined as FOR ALL TABLES",
NameStr(pubform->pubname)),
errdetail("Schemas cannot be added to or dropped from FOR ALL TABLES publications.")));
+ /*
+ * Check that user is allowed to manipulate the publication sequences in
+ * schema
+ */
+ if (sequences_schemaidlist && pubform->puballsequences)
+ ereport(ERROR,
+ (errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
+ errmsg("publication \"%s\" is defined as FOR ALL SEQUENCES",
+ NameStr(pubform->pubname)),
+ errdetail("Schemas cannot be added to or dropped from FOR ALL SEQUENCES publications.")));
+
/* Check that user is allowed to manipulate the publication tables. */
if (tables && pubform->puballtables)
ereport(ERROR,
@@ -1365,6 +1495,95 @@ CheckAlterPublication(AlterPublicationStmt *stmt, HeapTuple tup,
errmsg("publication \"%s\" is defined as FOR ALL TABLES",
NameStr(pubform->pubname)),
errdetail("Tables cannot be added to or dropped from FOR ALL TABLES publications.")));
+
+ /* Check that user is allowed to manipulate the publication sequences. */
+ if (sequences && pubform->puballsequences)
+ ereport(ERROR,
+ (errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
+ errmsg("publication \"%s\" is defined as FOR ALL SEQUENCES",
+ NameStr(pubform->pubname)),
+ errdetail("Sequences cannot be added to or dropped from FOR ALL SEQUENCES publications.")));
+}
+
+/*
+ * Add or remove sequence to/from publication.
+ */
+static void
+AlterPublicationSequences(AlterPublicationStmt *stmt, HeapTuple tup,
+ List *sequences, const char *queryString,
+ bool publish_schema)
+{
+ List *rels = NIL;
+ Form_pg_publication pubform = (Form_pg_publication) GETSTRUCT(tup);
+ Oid pubid = pubform->oid;
+
+ /*
+ * It is quite possible that for the SET case user has not specified any
+ * tables in which case we need to remove all the existing tables.
+ */
+ if (!sequences && stmt->action != AP_SetObjects)
+ return;
+
+ rels = OpenRelationList(sequences, PUB_OBJTYPE_SEQUENCE);
+
+ if (stmt->action == AP_AddObjects)
+ {
+ PublicationAddRelations(pubid, rels, false, stmt);
+ }
+ else if (stmt->action == AP_DropObjects)
+ PublicationDropRelations(pubid, rels, false);
+ else /* DEFELEM_SET */
+ {
+ List *oldrelids = GetPublicationRelations(pubid,
+ PUB_OBJTYPE_SEQUENCE,
+ PUBLICATION_PART_ROOT);
+ List *delrels = NIL;
+ ListCell *oldlc;
+
+ /* Calculate which relations to drop. */
+ foreach(oldlc, oldrelids)
+ {
+ Oid oldrelid = lfirst_oid(oldlc);
+ ListCell *newlc;
+ PublicationRelInfo *oldrel;
+ bool found = false;
+
+ foreach(newlc, rels)
+ {
+ PublicationRelInfo *newpubrel;
+
+ newpubrel = (PublicationRelInfo *) lfirst(newlc);
+ if (RelationGetRelid(newpubrel->relation) == oldrelid)
+ {
+ found = true;
+ break;
+ }
+ }
+ /* Not yet in the list, open it and add to the list */
+ if (!found)
+ {
+ oldrel = palloc(sizeof(PublicationRelInfo));
+ oldrel->whereClause = NULL;
+ oldrel->columns = NULL;
+ oldrel->relation = table_open(oldrelid,
+ ShareUpdateExclusiveLock);
+ delrels = lappend(delrels, oldrel);
+ }
+ }
+
+ /* And drop them. */
+ PublicationDropRelations(pubid, delrels, true);
+
+ /*
+ * Don't bother calculating the difference for adding, we'll catch and
+ * skip existing ones when doing catalog update.
+ */
+ PublicationAddRelations(pubid, rels, true, stmt);
+
+ CloseRelationList(delrels);
+ }
+
+ CloseRelationList(rels);
}
/*
@@ -1402,14 +1621,20 @@ AlterPublication(ParseState *pstate, AlterPublicationStmt *stmt)
AlterPublicationOptions(pstate, stmt, rel, tup);
else
{
- List *relations = NIL;
- List *schemaidlist = NIL;
+ List *tables = NIL;
+ List *sequences = NIL;
+ List *tables_schemaidlist = NIL;
+ List *sequences_schemaidlist = NIL;
Oid pubid = pubform->oid;
- ObjectsInPublicationToOids(stmt->pubobjects, pstate, &relations,
- &schemaidlist);
+ ObjectsInPublicationToOids(stmt->pubobjects, pstate,
+ &tables, &sequences,
+ &tables_schemaidlist,
+ &sequences_schemaidlist);
- CheckAlterPublication(stmt, tup, relations, schemaidlist);
+ CheckAlterPublication(stmt, tup,
+ tables, tables_schemaidlist,
+ sequences, sequences_schemaidlist);
heap_freetuple(tup);
@@ -1430,9 +1655,15 @@ AlterPublication(ParseState *pstate, AlterPublicationStmt *stmt)
errmsg("publication \"%s\" does not exist",
stmt->pubname));
- AlterPublicationTables(stmt, tup, relations, pstate->p_sourcetext,
- schemaidlist != NIL);
- AlterPublicationSchemas(stmt, tup, schemaidlist);
+ AlterPublicationTables(stmt, tup, tables, pstate->p_sourcetext,
+ tables_schemaidlist != NIL);
+ AlterPublicationSequences(stmt, tup, sequences, pstate->p_sourcetext,
+ sequences_schemaidlist != NIL);
+
+ AlterPublicationSchemas(stmt, tup, tables_schemaidlist,
+ PUB_OBJTYPE_TABLE);
+ AlterPublicationSchemas(stmt, tup, sequences_schemaidlist,
+ PUB_OBJTYPE_SEQUENCE);
}
/* Cleanup. */
@@ -1500,7 +1731,7 @@ RemovePublicationById(Oid pubid)
pubform = (Form_pg_publication) GETSTRUCT(tup);
/* Invalidate relcache so that publication info is rebuilt. */
- if (pubform->puballtables)
+ if (pubform->puballtables || pubform->puballsequences)
CacheInvalidateRelcacheAll();
CatalogTupleDelete(rel, &tup->t_self);
@@ -1536,6 +1767,7 @@ RemovePublicationSchemaById(Oid psoid)
* partitions.
*/
schemaRels = GetSchemaPublicationRelations(pubsch->pnnspid,
+ pubsch->pntype,
PUBLICATION_PART_ALL);
InvalidatePublicationRels(schemaRels);
@@ -1552,10 +1784,10 @@ RemovePublicationSchemaById(Oid psoid)
* add them to a publication.
*/
static List *
-OpenTableList(List *tables)
+OpenRelationList(List *rels, char objectType)
{
List *relids = NIL;
- List *rels = NIL;
+ List *result = NIL;
ListCell *lc;
List *relids_with_rf = NIL;
List *relids_with_collist = NIL;
@@ -1563,19 +1795,35 @@ OpenTableList(List *tables)
/*
* Open, share-lock, and check all the explicitly-specified relations
*/
- foreach(lc, tables)
+ foreach(lc, rels)
{
PublicationTable *t = lfirst_node(PublicationTable, lc);
bool recurse = t->relation->inh;
Relation rel;
Oid myrelid;
PublicationRelInfo *pub_rel;
+ char myrelkind;
/* Allow query cancel in case this takes a long time */
CHECK_FOR_INTERRUPTS();
rel = table_openrv(t->relation, ShareUpdateExclusiveLock);
myrelid = RelationGetRelid(rel);
+ myrelkind = get_rel_relkind(myrelid);
+
+ /*
+ * Make sure the relkind matches the expected object type. This may
+ * happen e.g. when adding a sequence using ADD TABLE or a table
+ * using ADD SEQUENCE).
+ *
+ * XXX We let through unsupported object types (views etc.). Those
+ * will be caught later in check_publication_add_relation.
+ */
+ if (pub_get_object_type_for_relkind(myrelkind) != PUB_OBJTYPE_UNSUPPORTED &&
+ pub_get_object_type_for_relkind(myrelkind) != objectType)
+ ereport(ERROR,
+ errcode(ERRCODE_INVALID_PARAMETER_VALUE),
+ errmsg("object type does not match type expected by command"));
/*
* Filter out duplicates if user specifies "foo, foo".
@@ -1608,7 +1856,7 @@ OpenTableList(List *tables)
pub_rel->relation = rel;
pub_rel->whereClause = t->whereClause;
pub_rel->columns = t->columns;
- rels = lappend(rels, pub_rel);
+ result = lappend(result, pub_rel);
relids = lappend_oid(relids, myrelid);
if (t->whereClause)
@@ -1680,7 +1928,7 @@ OpenTableList(List *tables)
/* child inherits column list from parent */
pub_rel->columns = t->columns;
- rels = lappend(rels, pub_rel);
+ result = lappend(result, pub_rel);
relids = lappend_oid(relids, childrelid);
if (t->whereClause)
@@ -1695,14 +1943,14 @@ OpenTableList(List *tables)
list_free(relids);
list_free(relids_with_rf);
- return rels;
+ return result;
}
/*
* Close all relations in the list.
*/
static void
-CloseTableList(List *rels)
+CloseRelationList(List *rels)
{
ListCell *lc;
@@ -1750,12 +1998,12 @@ LockSchemaList(List *schemalist)
* Add listed tables to the publication.
*/
static void
-PublicationAddTables(Oid pubid, List *rels, bool if_not_exists,
+PublicationAddRelations(Oid pubid, List *rels, bool if_not_exists,
AlterPublicationStmt *stmt)
{
ListCell *lc;
- Assert(!stmt || !stmt->for_all_tables);
+ Assert(!stmt || !stmt->for_all_objects);
foreach(lc, rels)
{
@@ -1784,7 +2032,7 @@ PublicationAddTables(Oid pubid, List *rels, bool if_not_exists,
* Remove listed tables from the publication.
*/
static void
-PublicationDropTables(Oid pubid, List *rels, bool missing_ok)
+PublicationDropRelations(Oid pubid, List *rels, bool missing_ok)
{
ObjectAddress obj;
ListCell *lc;
@@ -1829,19 +2077,19 @@ PublicationDropTables(Oid pubid, List *rels, bool missing_ok)
* Add listed schemas to the publication.
*/
static void
-PublicationAddSchemas(Oid pubid, List *schemas, bool if_not_exists,
- AlterPublicationStmt *stmt)
+PublicationAddSchemas(Oid pubid, List *schemas, char objectType,
+ bool if_not_exists, AlterPublicationStmt *stmt)
{
ListCell *lc;
- Assert(!stmt || !stmt->for_all_tables);
+ Assert(!stmt || !stmt->for_all_objects);
foreach(lc, schemas)
{
Oid schemaid = lfirst_oid(lc);
ObjectAddress obj;
- obj = publication_add_schema(pubid, schemaid, if_not_exists);
+ obj = publication_add_schema(pubid, schemaid, objectType, if_not_exists);
if (stmt)
{
EventTriggerCollectSimpleCommand(obj, InvalidObjectAddress,
@@ -1857,7 +2105,7 @@ PublicationAddSchemas(Oid pubid, List *schemas, bool if_not_exists,
* Remove listed schemas from the publication.
*/
static void
-PublicationDropSchemas(Oid pubid, List *schemas, bool missing_ok)
+PublicationDropSchemas(Oid pubid, List *schemas, char objectType, bool missing_ok)
{
ObjectAddress obj;
ListCell *lc;
@@ -1867,10 +2115,11 @@ PublicationDropSchemas(Oid pubid, List *schemas, bool missing_ok)
{
Oid schemaid = lfirst_oid(lc);
- psid = GetSysCacheOid2(PUBLICATIONNAMESPACEMAP,
+ psid = GetSysCacheOid3(PUBLICATIONNAMESPACEMAP,
Anum_pg_publication_namespace_oid,
ObjectIdGetDatum(schemaid),
- ObjectIdGetDatum(pubid));
+ ObjectIdGetDatum(pubid),
+ CharGetDatum(objectType));
if (!OidIsValid(psid))
{
if (missing_ok)
@@ -1925,6 +2174,13 @@ AlterPublicationOwner_internal(Relation rel, HeapTuple tup, Oid newOwnerId)
NameStr(form->pubname)),
errhint("The owner of a FOR ALL TABLES publication must be a superuser.")));
+ if (form->puballsequences && !superuser_arg(newOwnerId))
+ ereport(ERROR,
+ (errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
+ errmsg("permission denied to change owner of publication \"%s\"",
+ NameStr(form->pubname)),
+ errhint("The owner of a FOR ALL SEQUENCES publication must be a superuser.")));
+
if (!superuser_arg(newOwnerId) && is_schema_publication(form->oid))
ereport(ERROR,
(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
diff --git a/src/backend/commands/sequence.c b/src/backend/commands/sequence.c
index 99c9f91cba5..b0cb0d2bec0 100644
--- a/src/backend/commands/sequence.c
+++ b/src/backend/commands/sequence.c
@@ -38,6 +38,7 @@
#include "miscadmin.h"
#include "nodes/makefuncs.h"
#include "parser/parse_type.h"
+#include "replication/message.h"
#include "storage/lmgr.h"
#include "storage/proc.h"
#include "storage/smgr.h"
@@ -76,6 +77,7 @@ typedef struct SeqTableData
{
Oid relid; /* pg_class OID of this sequence (hash key) */
RelFileNumber filenumber; /* last seen relfilenumber of this sequence */
+ Oid tablespace; /* last seen tablespace of this sequence */
LocalTransactionId lxid; /* xact in which we last did a seq op */
bool last_valid; /* do we have a valid "last" value? */
int64 last; /* value last returned by nextval */
@@ -83,6 +85,7 @@ typedef struct SeqTableData
/* if last != cached, we have not used up all the cached values */
int64 increment; /* copy of sequence's increment field */
/* note that increment is zero until we first do nextval_internal() */
+ bool need_log; /* should be written to WAL at commit? */
} SeqTableData;
typedef SeqTableData *SeqTable;
@@ -340,6 +343,69 @@ ResetSequence(Oid seq_relid)
relation_close(seq_rel, NoLock);
}
+/*
+ * Update the sequence state by creating a new relfilenode.
+ *
+ * This creates a new relfilenode, to allow transactional behavior.
+ */
+void
+SetSequence(Oid seq_relid, int64 last_value, int64 log_cnt, bool is_called)
+{
+ SeqTable elm;
+ Relation seqrel;
+ Buffer buf;
+ HeapTupleData seqdatatuple;
+ Form_pg_sequence_data seq;
+ HeapTuple tuple;
+
+ /* open and lock sequence */
+ init_sequence(seq_relid, &elm, &seqrel);
+
+ /* lock page' buffer and read tuple */
+ seq = read_seq_tuple(seqrel, &buf, &seqdatatuple);
+
+ /* Copy the existing sequence tuple. */
+ tuple = heap_copytuple(&seqdatatuple);
+
+ /* Now we're done with the old page */
+ UnlockReleaseBuffer(buf);
+
+ /*
+ * Modify the copied tuple to update the sequence state (similar to what
+ * ResetSequence does).
+ */
+ seq = (Form_pg_sequence_data) GETSTRUCT(tuple);
+ seq->last_value = last_value;
+ seq->is_called = is_called;
+ seq->log_cnt = log_cnt;
+
+ /*
+ * Create a new storage file for the sequence - this is needed for the
+ * transactional behavior.
+ */
+ RelationSetNewRelfilenumber(seqrel, seqrel->rd_rel->relpersistence);
+
+ /*
+ * Ensure sequence's relfrozenxid is at 0, since it won't contain any
+ * unfrozen XIDs. Same with relminmxid, since a sequence will never
+ * contain multixacts.
+ */
+ Assert(seqrel->rd_rel->relfrozenxid == InvalidTransactionId);
+ Assert(seqrel->rd_rel->relminmxid == InvalidMultiXactId);
+
+ /*
+ * Insert the modified tuple into the new storage file. This does all the
+ * necessary WAL-logging etc.
+ */
+ fill_seq_with_data(seqrel, tuple);
+
+ /* Clear local cache so that we don't think we have cached numbers */
+ /* Note that we do not change the currval() state */
+ elm->cached = elm->last;
+
+ relation_close(seqrel, NoLock);
+}
+
/*
* Initialize a sequence's relation with the specified tuple as content
*
@@ -404,10 +470,21 @@ fill_seq_fork_with_data(Relation rel, HeapTuple tuple, ForkNumber forkNum)
tuple->t_data->t_infomask |= HEAP_XMAX_INVALID;
ItemPointerSet(&tuple->t_data->t_ctid, 0, FirstOffsetNumber);
- /* check the comment above nextval_internal()'s equivalent call. */
+ /*
+ * Remember we need to write this sequence to WAL at commit, and make sure
+ * we have XID if we may need to decode this sequence.
+ *
+ * XXX Maybe we should not be doing this except when actually reading a
+ * value from the sequence?
+ */
if (RelationNeedsWAL(rel))
+ {
GetTopTransactionId();
+ if (XLogLogicalInfoActive())
+ GetCurrentTransactionId();
+ }
+
START_CRIT_SECTION();
MarkBufferDirty(buf);
@@ -652,6 +729,23 @@ nextval_internal(Oid relid, bool check_permissions)
/* open and lock sequence */
init_sequence(relid, &elm, &seqrel);
+ /*
+ * Remember we need to write this sequence to WAL at commit, and make sure
+ * we have XID if we may need to decode this sequence.
+ *
+ * XXX Maybe we should not be doing this except when actually reading a
+ * value from the sequence?
+ */
+ if (RelationNeedsWAL(seqrel))
+ {
+ elm->need_log = true;
+
+ GetTopTransactionId();
+
+ if (XLogLogicalInfoActive())
+ GetCurrentTransactionId();
+ }
+
if (check_permissions &&
pg_class_aclcheck(elm->relid, GetUserId(),
ACL_USAGE | ACL_UPDATE) != ACLCHECK_OK)
@@ -806,10 +900,28 @@ nextval_internal(Oid relid, bool check_permissions)
* It's sufficient to ensure the toplevel transaction has an xid, no need
* to assign xids subxacts, that'll already trigger an appropriate wait.
* (Have to do that here, so we're outside the critical section)
+ *
+ * We have to ensure we have a proper XID, which will be included in
+ * the XLOG record by XLogRecordAssemble. Otherwise the first nextval()
+ * in a subxact (without any preceding changes) would get XID 0, and it
+ * would then be impossible to decide which top xact it belongs to.
+ * It'd also trigger assert in DecodeSequence. We only do that with
+ * wal_level=logical, though.
+ *
+ * XXX This might seem unnecessary, because if there's no XID the xact
+ * couldn't have done anything important yet, e.g. it could not have
+ * created a sequence. But that's incorrect, because of subxacts. The
+ * current subtransaction might not have done anything yet (thus no XID),
+ * but an earlier one might have created the sequence.
*/
if (logit && RelationNeedsWAL(seqrel))
+ {
GetTopTransactionId();
+ if (XLogLogicalInfoActive())
+ GetCurrentTransactionId();
+ }
+
/* ready to change the on-disk (or really, in-buffer) tuple */
START_CRIT_SECTION();
@@ -963,6 +1075,23 @@ do_setval(Oid relid, int64 next, bool iscalled)
/* open and lock sequence */
init_sequence(relid, &elm, &seqrel);
+ /*
+ * Remember we need to write this sequence to WAL at commit, and make sure
+ * we have XID if we may need to decode this sequence.
+ *
+ * XXX Maybe we should not be doing this except when actually reading a
+ * value from the sequence?
+ */
+ if (RelationNeedsWAL(seqrel))
+ {
+ elm->need_log = true;
+
+ GetTopTransactionId();
+
+ if (XLogLogicalInfoActive())
+ GetCurrentTransactionId();
+ }
+
if (pg_class_aclcheck(elm->relid, GetUserId(), ACL_UPDATE) != ACLCHECK_OK)
ereport(ERROR,
(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
@@ -1010,8 +1139,13 @@ do_setval(Oid relid, int64 next, bool iscalled)
/* check the comment above nextval_internal()'s equivalent call. */
if (RelationNeedsWAL(seqrel))
+ {
GetTopTransactionId();
+ if (XLogLogicalInfoActive())
+ GetCurrentTransactionId();
+ }
+
/* ready to change the on-disk (or really, in-buffer) tuple */
START_CRIT_SECTION();
@@ -1180,6 +1314,7 @@ init_sequence(Oid relid, SeqTable *p_elm, Relation *p_rel)
if (seqrel->rd_rel->relfilenode != elm->filenumber)
{
elm->filenumber = seqrel->rd_rel->relfilenode;
+ elm->tablespace = seqrel->rd_rel->reltablespace;
elm->cached = elm->last;
}
@@ -1916,3 +2051,138 @@ seq_mask(char *page, BlockNumber blkno)
mask_unused_space(page);
}
+
+
+static void
+read_sequence_info(Relation seqrel, int64 *last_value, int64 *log_cnt, bool *is_called)
+{
+ Buffer buf;
+ Form_pg_sequence_data seq;
+ HeapTupleData seqdatatuple;
+
+ seq = read_seq_tuple(seqrel, &buf, &seqdatatuple);
+
+ *last_value = seq->last_value;
+ *is_called = seq->is_called;
+ *log_cnt = seq->log_cnt;
+
+ UnlockReleaseBuffer(buf);
+}
+
+/* XXX Do this only for wal_level = logical, probably? */
+void
+AtEOXact_Sequences(bool isCommit)
+{
+ SeqTable entry;
+ HASH_SEQ_STATUS scan;
+
+ if (!seqhashtab)
+ return;
+
+ /* only do this with wal_level=logical */
+ if (!XLogLogicalInfoActive())
+ return;
+
+ /*
+ * XXX Maybe we could enforce having XID here? Or is it too late?
+ */
+ // Assert(GetTopTransactionId() != InvalidTransactionId);
+ // Assert(GetCurrentTransactionId() != InvalidTransactionId);
+
+ hash_seq_init(&scan, seqhashtab);
+
+ while ((entry = (SeqTable) hash_seq_search(&scan)))
+ {
+ Relation rel;
+ RelFileLocator rlocator;
+ xl_logical_sequence xlrec;
+
+ /* not commit, we don't guarantee any data to be durable */
+ if (!isCommit)
+ entry->need_log = false;
+
+ /*
+ * If not touched in the current transaction, don't log anything.
+ * We leave needs_log set, so that if future transactions touch
+ * the sequence we'll log it properly.
+ */
+ if (!entry->need_log)
+ continue;
+
+ /* if this is commit, we'll log the */
+ entry->need_log = false;
+
+ /*
+ * does the relation still exist?
+ *
+ * XXX We need to make sure another transaction can't jump ahead of
+ * this one, otherwise the ordering of sequence changes could change.
+ * Imagine T1 and T2, where T1 writes sequence state first, but then
+ * T2 does it too and commits first:
+ *
+ * T1: log sequence state
+ * T2: increment sequence
+ * T2: log sequence state
+ * T2: commit
+ * T1: commit
+ *
+ * If we apply the sequences in this order, it'd be broken as the
+ * value might go backwars. ShareUpdateExclusive protects against
+ * that, but it's also restricting commit throughtput, probably.
+ *
+ * XXX This might be an issue for deadlocks, too, if two xacts try
+ * to write sequences in different ordering. We may need to sort
+ * the OIDs first, to enforce the same lock ordering.
+ *
+ * XXX Keeping the SUE lock also poses issue with prepared xacts,
+ * because no other transaction touching the sequence will be able
+ * to commit until the prepared xact does (due to waiting on the
+ * lock). That's a significant escalation compared to now, so we
+ * now release the lock, but that makes this vulnerable to the
+ * reordering issue.
+ *
+ * XXX But I think we could use LSN of the XLOG_LOGICAL_SEQUENCE
+ * because that correlates with the sequence state changes - the
+ * commit order may result in applying the changes in a different
+ * order, but we could discard changes where the LSN would go
+ * backwards (if we remember the LSN). But we'd have to persist
+ * that (sequence,LSN) somewhere on the subscriber, because of
+ * restarts, though.
+ */
+ rel = try_relation_open(entry->relid, ShareUpdateExclusiveLock);
+
+ /* XXX relation might have been dropped, maybe we should have logged
+ * the last change? */
+ if (!rel)
+ continue;
+
+ /* tablespace */
+ if (OidIsValid(entry->tablespace))
+ rlocator.spcOid = entry->tablespace;
+ else
+ rlocator.spcOid = MyDatabaseTableSpace;
+
+ rlocator.dbOid = MyDatabaseId; /* database */
+ rlocator.relNumber = entry->filenumber; /* relation */
+
+ xlrec.locator = rlocator;
+ xlrec.reloid = entry->relid;
+
+ /*
+ * Re-read sequence info from disk (so that we always log the state
+ * which might have been modified by some other backend).
+ */
+ read_sequence_info(rel, &xlrec.last, &xlrec.log_cnt, &xlrec.is_called);
+
+ XLogBeginInsert();
+ XLogRegisterData((char *) &xlrec, SizeOfLogicalSequence);
+
+ /* allow origin filtering */
+ XLogSetRecordFlags(XLOG_INCLUDE_ORIGIN);
+
+ (void) XLogInsert(RM_LOGICALMSG_ID, XLOG_LOGICAL_SEQUENCE);
+
+ /* release the sequence lock */
+ relation_close(rel, ShareUpdateExclusiveLock);
+ }
+}
diff --git a/src/backend/commands/subscriptioncmds.c b/src/backend/commands/subscriptioncmds.c
index f0cec2ad5e7..de57f0f9336 100644
--- a/src/backend/commands/subscriptioncmds.c
+++ b/src/backend/commands/subscriptioncmds.c
@@ -92,6 +92,7 @@ typedef struct SubOpts
} SubOpts;
static List *fetch_table_list(WalReceiverConn *wrconn, List *publications);
+static List *fetch_sequence_list(WalReceiverConn *wrconn, List *publications);
static void check_publications_origin(WalReceiverConn *wrconn,
List *publications, bool copydata,
char *origin, Oid *subrel_local_oids,
@@ -668,9 +669,9 @@ CreateSubscription(ParseState *pstate, CreateSubscriptionStmt *stmt,
{
char *err;
WalReceiverConn *wrconn;
- List *tables;
+ List *relations;
ListCell *lc;
- char table_state;
+ char sync_state;
/* Try to connect to the publisher. */
wrconn = walrcv_connect(conninfo, true, stmt->subname, &err);
@@ -689,14 +690,17 @@ CreateSubscription(ParseState *pstate, CreateSubscriptionStmt *stmt,
* Set sync state based on if we were asked to do data copy or
* not.
*/
- table_state = opts.copy_data ? SUBREL_STATE_INIT : SUBREL_STATE_READY;
+ sync_state = opts.copy_data ? SUBREL_STATE_INIT : SUBREL_STATE_READY;
/*
- * Get the table list from publisher and build local table status
- * info.
+ * Get the table and sequence list from publisher and build
+ * local relation sync status info.
*/
- tables = fetch_table_list(wrconn, publications);
- foreach(lc, tables)
+ relations = fetch_table_list(wrconn, publications);
+ relations = list_concat(relations,
+ fetch_sequence_list(wrconn, publications));
+
+ foreach(lc, relations)
{
RangeVar *rv = (RangeVar *) lfirst(lc);
Oid relid;
@@ -707,7 +711,7 @@ CreateSubscription(ParseState *pstate, CreateSubscriptionStmt *stmt,
CheckSubscriptionRelkind(get_rel_relkind(relid),
rv->schemaname, rv->relname);
- AddSubscriptionRelState(subid, relid, table_state,
+ AddSubscriptionRelState(subid, relid, sync_state,
InvalidXLogRecPtr);
}
@@ -733,12 +737,12 @@ CreateSubscription(ParseState *pstate, CreateSubscriptionStmt *stmt,
*
* Note that if tables were specified but copy_data is false
* then it is safe to enable two_phase up-front because those
- * tables are already initially in READY state. When the
- * subscription has no tables, we leave the twophase state as
- * PENDING, to allow ALTER SUBSCRIPTION ... REFRESH
+ * relations are already initially in READY state. When the
+ * subscription has no relations, we leave the twophase state
+ * as PENDING, to allow ALTER SUBSCRIPTION ... REFRESH
* PUBLICATION to work.
*/
- if (opts.twophase && !opts.copy_data && tables != NIL)
+ if (opts.twophase && !opts.copy_data && relations != NIL)
twophase_enabled = true;
walrcv_create_slot(wrconn, opts.slot_name, false, twophase_enabled,
@@ -814,8 +818,10 @@ AlterSubscription_refresh(Subscription *sub, bool copy_data,
if (validate_publications)
check_publications(wrconn, validate_publications);
- /* Get the table list from publisher. */
+ /* Get the list of relations from publisher. */
pubrel_names = fetch_table_list(wrconn, sub->publications);
+ pubrel_names = list_concat(pubrel_names,
+ fetch_sequence_list(wrconn, sub->publications));
/* Get local table list. */
subrel_states = GetSubscriptionRelations(sub->oid, false);
@@ -1980,6 +1986,75 @@ fetch_table_list(WalReceiverConn *wrconn, List *publications)
return tablelist;
}
+/*
+ * Get the list of sequences which belong to specified publications on the
+ * publisher connection.
+ */
+static List *
+fetch_sequence_list(WalReceiverConn *wrconn, List *publications)
+{
+ WalRcvExecResult *res;
+ StringInfoData cmd;
+ TupleTableSlot *slot;
+ Oid tableRow[2] = {TEXTOID, TEXTOID};
+ ListCell *lc;
+ bool first;
+ List *tablelist = NIL;
+
+ Assert(list_length(publications) > 0);
+
+ initStringInfo(&cmd);
+ appendStringInfoString(&cmd, "SELECT DISTINCT s.schemaname, s.sequencename\n"
+ " FROM pg_catalog.pg_publication_sequences s\n"
+ " WHERE s.pubname IN (");
+ first = true;
+ foreach(lc, publications)
+ {
+ char *pubname = strVal(lfirst(lc));
+
+ if (first)
+ first = false;
+ else
+ appendStringInfoString(&cmd, ", ");
+
+ appendStringInfoString(&cmd, quote_literal_cstr(pubname));
+ }
+ appendStringInfoChar(&cmd, ')');
+
+ res = walrcv_exec(wrconn, cmd.data, 2, tableRow);
+ pfree(cmd.data);
+
+ if (res->status != WALRCV_OK_TUPLES)
+ ereport(ERROR,
+ (errmsg("could not receive list of replicated sequences from the publisher: %s",
+ res->err)));
+
+ /* Process sequences. */
+ slot = MakeSingleTupleTableSlot(res->tupledesc, &TTSOpsMinimalTuple);
+ while (tuplestore_gettupleslot(res->tuplestore, true, false, slot))
+ {
+ char *nspname;
+ char *relname;
+ bool isnull;
+ RangeVar *rv;
+
+ nspname = TextDatumGetCString(slot_getattr(slot, 1, &isnull));
+ Assert(!isnull);
+ relname = TextDatumGetCString(slot_getattr(slot, 2, &isnull));
+ Assert(!isnull);
+
+ rv = makeRangeVar(nspname, relname, -1);
+ tablelist = lappend(tablelist, rv);
+
+ ExecClearTuple(slot);
+ }
+ ExecDropSingleTupleTableSlot(slot);
+
+ walrcv_clear_result(res);
+
+ return tablelist;
+}
+
/*
* This is to report the connection failure while dropping replication slots.
* Here, we report the WARNING for all tablesync slots so that user can drop
diff --git a/src/backend/executor/execReplication.c b/src/backend/executor/execReplication.c
index 6014f2e2482..3ae893f2211 100644
--- a/src/backend/executor/execReplication.c
+++ b/src/backend/executor/execReplication.c
@@ -655,7 +655,9 @@ void
CheckSubscriptionRelkind(char relkind, const char *nspname,
const char *relname)
{
- if (relkind != RELKIND_RELATION && relkind != RELKIND_PARTITIONED_TABLE)
+ if (relkind != RELKIND_RELATION &&
+ relkind != RELKIND_PARTITIONED_TABLE &&
+ relkind != RELKIND_SEQUENCE)
ereport(ERROR,
(errcode(ERRCODE_WRONG_OBJECT_TYPE),
errmsg("cannot use relation \"%s.%s\" as logical replication target",
diff --git a/src/backend/parser/gram.y b/src/backend/parser/gram.y
index 2a910ded15e..0f5022fd628 100644
--- a/src/backend/parser/gram.y
+++ b/src/backend/parser/gram.y
@@ -452,7 +452,7 @@ static Node *makeRecursiveViewSelect(char *relname, List *aliases, Node *query);
transform_element_list transform_type_list
TriggerTransitions TriggerReferencing
vacuum_relation_list opt_vacuum_relation_list
- drop_option_list pub_obj_list
+ drop_option_list pub_obj_list pub_obj_type_list
%type <node> opt_routine_body
%type <groupclause> group_clause
@@ -586,6 +586,7 @@ static Node *makeRecursiveViewSelect(char *relname, List *aliases, Node *query);
%type <node> var_value zone_value
%type <rolespec> auth_ident RoleSpec opt_granted_by
%type <publicationobjectspec> PublicationObjSpec
+%type <node> pub_obj_type
%type <keyword> unreserved_keyword type_func_name_keyword
%type <keyword> col_name_keyword reserved_keyword
@@ -10337,12 +10338,16 @@ AlterOwnerStmt: ALTER AGGREGATE aggregate_with_argtypes OWNER TO RoleSpec
*
* CREATE PUBLICATION FOR ALL TABLES [WITH options]
*
+ * CREATE PUBLICATION FOR ALL SEQUENCES [WITH options]
+ *
* CREATE PUBLICATION FOR pub_obj [, ...] [WITH options]
*
* pub_obj is one of:
*
* TABLE table [, ...]
* TABLES IN SCHEMA schema [, ...]
+ * SEQUENCE sequence [, ...]
+ * SEQUENCES IN SCHEMA schema [, ...]
*
*****************************************************************************/
@@ -10355,13 +10360,13 @@ CreatePublicationStmt:
n->options = $4;
$$ = (Node *) n;
}
- | CREATE PUBLICATION name FOR ALL TABLES opt_definition
+ | CREATE PUBLICATION name FOR ALL pub_obj_type_list opt_definition
{
CreatePublicationStmt *n = makeNode(CreatePublicationStmt);
n->pubname = $3;
n->options = $7;
- n->for_all_tables = true;
+ n->for_all_objects = $6;
$$ = (Node *) n;
}
| CREATE PUBLICATION name FOR pub_obj_list opt_definition
@@ -10412,6 +10417,26 @@ PublicationObjSpec:
$$->pubobjtype = PUBLICATIONOBJ_TABLES_IN_CUR_SCHEMA;
$$->location = @4;
}
+ | SEQUENCE relation_expr
+ {
+ $$ = makeNode(PublicationObjSpec);
+ $$->pubobjtype = PUBLICATIONOBJ_SEQUENCE;
+ $$->pubtable = makeNode(PublicationTable);
+ $$->pubtable->relation = $2;
+ }
+ | SEQUENCES IN_P SCHEMA ColId
+ {
+ $$ = makeNode(PublicationObjSpec);
+ $$->pubobjtype = PUBLICATIONOBJ_SEQUENCES_IN_SCHEMA;
+ $$->name = $4;
+ $$->location = @4;
+ }
+ | SEQUENCES IN_P SCHEMA CURRENT_SCHEMA
+ {
+ $$ = makeNode(PublicationObjSpec);
+ $$->pubobjtype = PUBLICATIONOBJ_SEQUENCES_IN_CUR_SCHEMA;
+ $$->location = @4;
+ }
| ColId opt_column_list OptWhereClause
{
$$ = makeNode(PublicationObjSpec);
@@ -10473,6 +10498,19 @@ pub_obj_list: PublicationObjSpec
{ $$ = lappend($1, $3); }
;
+pub_obj_type: TABLES
+ { $$ = (Node *) makeString("tables"); }
+ | SEQUENCES
+ { $$ = (Node *) makeString("sequences"); }
+ ;
+
+pub_obj_type_list: pub_obj_type
+ { $$ = list_make1($1); }
+ | pub_obj_type_list ',' pub_obj_type
+ { $$ = lappend($1, $3); }
+ ;
+
+
/*****************************************************************************
*
* ALTER PUBLICATION name SET ( options )
@@ -18457,7 +18495,7 @@ preprocess_pubobj_list(List *pubobjspec_list, core_yyscan_t yyscanner)
ereport(ERROR,
errcode(ERRCODE_SYNTAX_ERROR),
errmsg("invalid publication object list"),
- errdetail("One of TABLE or TABLES IN SCHEMA must be specified before a standalone table or schema name."),
+ errdetail("One of TABLE, SEQUENCE, TABLES IN SCHEMA or SEQUENCES IN SCHEMA must be specified before a standalone table, sequence or schema name."),
parser_errposition(pubobj->location));
foreach(cell, pubobjspec_list)
@@ -18467,7 +18505,8 @@ preprocess_pubobj_list(List *pubobjspec_list, core_yyscan_t yyscanner)
if (pubobj->pubobjtype == PUBLICATIONOBJ_CONTINUATION)
pubobj->pubobjtype = prevobjtype;
- if (pubobj->pubobjtype == PUBLICATIONOBJ_TABLE)
+ if (pubobj->pubobjtype == PUBLICATIONOBJ_TABLE ||
+ pubobj->pubobjtype == PUBLICATIONOBJ_SEQUENCE)
{
/* relation name or pubtable must be set for this type of object */
if (!pubobj->name && !pubobj->pubtable)
@@ -18518,6 +18557,30 @@ preprocess_pubobj_list(List *pubobjspec_list, core_yyscan_t yyscanner)
errmsg("invalid schema name at or near"),
parser_errposition(pubobj->location));
}
+ else if (pubobj->pubobjtype == PUBLICATIONOBJ_SEQUENCES_IN_SCHEMA ||
+ pubobj->pubobjtype == PUBLICATIONOBJ_SEQUENCES_IN_CUR_SCHEMA)
+ {
+ /* WHERE clause is not allowed on a schema object */
+ if (pubobj->pubtable && pubobj->pubtable->whereClause)
+ ereport(ERROR,
+ errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("WHERE clause not allowed for schema"),
+ parser_errposition(pubobj->location));
+
+ /*
+ * We can distinguish between the different type of schema
+ * objects based on whether name and pubtable is set.
+ */
+ if (pubobj->name)
+ pubobj->pubobjtype = PUBLICATIONOBJ_SEQUENCES_IN_SCHEMA;
+ else if (!pubobj->name && !pubobj->pubtable)
+ pubobj->pubobjtype = PUBLICATIONOBJ_SEQUENCES_IN_CUR_SCHEMA;
+ else
+ ereport(ERROR,
+ errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("invalid schema name at or near"),
+ parser_errposition(pubobj->location));
+ }
prevobjtype = pubobj->pubobjtype;
}
diff --git a/src/backend/replication/logical/decode.c b/src/backend/replication/logical/decode.c
index 2cc0ac9eb09..59536c50b44 100644
--- a/src/backend/replication/logical/decode.c
+++ b/src/backend/replication/logical/decode.c
@@ -42,6 +42,10 @@
#include "replication/reorderbuffer.h"
#include "replication/snapbuild.h"
#include "storage/standby.h"
+#include "commands/sequence.h"
+
+static void DecodeLogicalMessage(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
+static void DecodeLogicalSequence(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
/* individual record(group)'s handlers */
static void DecodeInsert(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
@@ -558,6 +562,27 @@ FilterByOrigin(LogicalDecodingContext *ctx, RepOriginId origin_id)
*/
void
logicalmsg_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
+{
+ XLogReaderState *r = buf->record;
+ uint8 info = XLogRecGetInfo(r) & ~XLR_INFO_MASK;
+
+ switch (info)
+ {
+ case XLOG_LOGICAL_MESSAGE:
+ DecodeLogicalMessage(ctx, buf);
+ break;
+
+ case XLOG_LOGICAL_SEQUENCE:
+ DecodeLogicalSequence(ctx, buf);
+ break;
+
+ default:
+ elog(ERROR, "unexpected RM_LOGICALMSG_ID record type: %u", info);
+ }
+}
+
+static void
+DecodeLogicalMessage(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
{
SnapBuild *builder = ctx->snapshot_builder;
XLogReaderState *r = buf->record;
@@ -1251,3 +1276,67 @@ DecodeTXNNeedSkip(LogicalDecodingContext *ctx, XLogRecordBuffer *buf,
(txn_dbid != InvalidOid && txn_dbid != ctx->slot->data.database) ||
ctx->fast_forward || FilterByOrigin(ctx, origin_id));
}
+
+/*
+ * Handle sequence decode
+ *
+ * Decoding sequences is a bit tricky, because while most sequence actions
+ * are non-transactional (not subject to rollback), some need to be handled
+ * as transactional.
+ *
+ * By default, a sequence increment is non-transactional - we must not queue
+ * it in a transaction as other changes, because the transaction might get
+ * rolled back and we'd discard the increment. The downstream would not be
+ * notified about the increment, which is wrong.
+ *
+ * On the other hand, the sequence may be created in a transaction. In this
+ * case we *should* queue the change as other changes in the transaction,
+ * because we don't want to send the increments for unknown sequence to the
+ * plugin - it might get confused about which sequence it's related to etc.
+ */
+void
+DecodeLogicalSequence(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
+{
+ SnapBuild *builder = ctx->snapshot_builder;
+ RelFileLocator target_locator;
+ XLogReaderState *r = buf->record;
+ TransactionId xid = XLogRecGetXid(r);
+ uint8 info = XLogRecGetInfo(buf->record) & ~XLR_INFO_MASK;
+ xl_logical_sequence *xlrec;
+ RepOriginId origin_id = XLogRecGetOrigin(r);
+
+ /* only decode changes flagged with XLOG_SEQ_LOG */
+ if (info != XLOG_LOGICAL_SEQUENCE)
+ elog(ERROR, "unexpected RM_SEQ_ID record type: %u", info);
+
+ ReorderBufferProcessXid(ctx->reorder, XLogRecGetXid(r), buf->origptr);
+
+ /*
+ * If we don't have snapshot or we are just fast-forwarding, there is no
+ * point in decoding messages.
+ */
+ if (SnapBuildCurrentState(builder) < SNAPBUILD_FULL_SNAPSHOT ||
+ ctx->fast_forward)
+ return;
+
+ /* extract the WAL record, with "created" flag */
+ xlrec = (xl_logical_sequence *) XLogRecGetData(r);
+
+ /* only interested in our database */
+ target_locator = xlrec->locator;
+ if (target_locator.dbOid != ctx->slot->data.database)
+ return;
+
+ /* output plugin doesn't look for this origin, no need to queue */
+ if (FilterByOrigin(ctx, XLogRecGetOrigin(r)))
+ return;
+
+ /* Skip the change if already processed (per the snapshot). */
+ if (!SnapBuildProcessChange(builder, xid, buf->origptr))
+ return;
+
+ /* Queue the increment (or send immediately if not transactional). */
+ ReorderBufferQueueSequence(ctx->reorder, xid, buf->endptr,
+ origin_id, xlrec->reloid, target_locator,
+ xlrec->last, xlrec->log_cnt, xlrec->is_called);
+}
diff --git a/src/backend/replication/logical/logical.c b/src/backend/replication/logical/logical.c
index 625a7f42730..b1b4eaa8d89 100644
--- a/src/backend/replication/logical/logical.c
+++ b/src/backend/replication/logical/logical.c
@@ -73,6 +73,9 @@ static void truncate_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
static void message_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
XLogRecPtr message_lsn, bool transactional,
const char *prefix, Size message_size, const char *message);
+static void sequence_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ int64 last_value, int64 log_cnt, bool is_called);
/* streaming callbacks */
static void stream_start_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
@@ -90,6 +93,9 @@ static void stream_change_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn
static void stream_message_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
XLogRecPtr message_lsn, bool transactional,
const char *prefix, Size message_size, const char *message);
+static void stream_sequence_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ int64 last_value, int64 log_cnt, bool is_called);
static void stream_truncate_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
int nrelations, Relation relations[], ReorderBufferChange *change);
@@ -218,6 +224,7 @@ StartupDecodingContext(List *output_plugin_options,
ctx->reorder->apply_truncate = truncate_cb_wrapper;
ctx->reorder->commit = commit_cb_wrapper;
ctx->reorder->message = message_cb_wrapper;
+ ctx->reorder->sequence = sequence_cb_wrapper;
/*
* To support streaming, we require start/stop/abort/commit/change
@@ -234,6 +241,7 @@ StartupDecodingContext(List *output_plugin_options,
(ctx->callbacks.stream_commit_cb != NULL) ||
(ctx->callbacks.stream_change_cb != NULL) ||
(ctx->callbacks.stream_message_cb != NULL) ||
+ (ctx->callbacks.stream_sequence_cb != NULL) ||
(ctx->callbacks.stream_truncate_cb != NULL);
/*
@@ -251,6 +259,7 @@ StartupDecodingContext(List *output_plugin_options,
ctx->reorder->stream_commit = stream_commit_cb_wrapper;
ctx->reorder->stream_change = stream_change_cb_wrapper;
ctx->reorder->stream_message = stream_message_cb_wrapper;
+ ctx->reorder->stream_sequence = stream_sequence_cb_wrapper;
ctx->reorder->stream_truncate = stream_truncate_cb_wrapper;
@@ -1220,6 +1229,42 @@ message_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
error_context_stack = errcallback.previous;
}
+static void
+sequence_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ LogicalDecodingContext *ctx = cache->private_data;
+ LogicalErrorCallbackState state;
+ ErrorContextCallback errcallback;
+
+ Assert(!ctx->fast_forward);
+
+ if (ctx->callbacks.sequence_cb == NULL)
+ return;
+
+ /* Push callback + info on the error context stack */
+ state.ctx = ctx;
+ state.callback_name = "sequence";
+ state.report_location = sequence_lsn;
+ errcallback.callback = output_plugin_error_callback;
+ errcallback.arg = (void *) &state;
+ errcallback.previous = error_context_stack;
+ error_context_stack = &errcallback;
+
+ /* set output state */
+ ctx->accept_writes = true;
+ ctx->write_xid = txn != NULL ? txn->xid : InvalidTransactionId;
+ ctx->write_location = sequence_lsn;
+
+ /* do the actual work: call callback */
+ ctx->callbacks.sequence_cb(ctx, txn, sequence_lsn, rel,
+ last_value, log_cnt, is_called);
+
+ /* Pop the error context stack */
+ error_context_stack = errcallback.previous;
+}
+
static void
stream_start_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
XLogRecPtr first_lsn)
@@ -1535,6 +1580,46 @@ stream_message_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
error_context_stack = errcallback.previous;
}
+static void
+stream_sequence_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ LogicalDecodingContext *ctx = cache->private_data;
+ LogicalErrorCallbackState state;
+ ErrorContextCallback errcallback;
+
+ Assert(!ctx->fast_forward);
+
+ /* We're only supposed to call this when streaming is supported. */
+ Assert(ctx->streaming);
+
+ /* this callback is optional */
+ if (ctx->callbacks.stream_sequence_cb == NULL)
+ return;
+
+ /* Push callback + info on the error context stack */
+ state.ctx = ctx;
+ state.callback_name = "stream_sequence";
+ state.report_location = sequence_lsn;
+ errcallback.callback = output_plugin_error_callback;
+ errcallback.arg = (void *) &state;
+ errcallback.previous = error_context_stack;
+ error_context_stack = &errcallback;
+
+ /* set output state */
+ ctx->accept_writes = true;
+ ctx->write_xid = txn != NULL ? txn->xid : InvalidTransactionId;
+ ctx->write_location = sequence_lsn;
+
+ /* do the actual work: call callback */
+ ctx->callbacks.sequence_cb(ctx, txn, sequence_lsn, rel,
+ last_value, log_cnt, is_called);
+
+ /* Pop the error context stack */
+ error_context_stack = errcallback.previous;
+}
+
static void
stream_truncate_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
int nrelations, Relation relations[],
diff --git a/src/backend/replication/logical/message.c b/src/backend/replication/logical/message.c
index 1c34912610e..69178dc1cfb 100644
--- a/src/backend/replication/logical/message.c
+++ b/src/backend/replication/logical/message.c
@@ -82,7 +82,8 @@ logicalmsg_redo(XLogReaderState *record)
{
uint8 info = XLogRecGetInfo(record) & ~XLR_INFO_MASK;
- if (info != XLOG_LOGICAL_MESSAGE)
+ if (info != XLOG_LOGICAL_MESSAGE &&
+ info != XLOG_LOGICAL_SEQUENCE)
elog(PANIC, "logicalmsg_redo: unknown op code %u", info);
/* This is only interesting for logical decoding, see decode.c. */
diff --git a/src/backend/replication/logical/proto.c b/src/backend/replication/logical/proto.c
index ff8513e2d29..ae078eed048 100644
--- a/src/backend/replication/logical/proto.c
+++ b/src/backend/replication/logical/proto.c
@@ -662,6 +662,54 @@ logicalrep_write_message(StringInfo out, TransactionId xid, XLogRecPtr lsn,
pq_sendbytes(out, message, sz);
}
+/*
+ * Write SEQUENCE to stream
+ */
+void
+logicalrep_write_sequence(StringInfo out, Relation rel, TransactionId xid,
+ XLogRecPtr lsn,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ uint8 flags = 0;
+ char *relname;
+
+ pq_sendbyte(out, LOGICAL_REP_MSG_SEQUENCE);
+
+ /* transaction ID (if not valid, we're not streaming) */
+ if (TransactionIdIsValid(xid))
+ pq_sendint32(out, xid);
+
+ pq_sendint8(out, flags);
+ pq_sendint64(out, lsn);
+
+ logicalrep_write_namespace(out, RelationGetNamespace(rel));
+ relname = RelationGetRelationName(rel);
+ pq_sendstring(out, relname);
+
+ pq_sendint64(out, last_value);
+ pq_sendint64(out, log_cnt);
+ pq_sendint8(out, is_called);
+}
+
+/*
+ * Read SEQUENCE from the stream.
+ */
+void
+logicalrep_read_sequence(StringInfo in, LogicalRepSequence *seqdata)
+{
+ /* XXX skipping flags and lsn */
+ pq_getmsgint(in, 1);
+ pq_getmsgint64(in);
+
+ /* Read relation name from stream */
+ seqdata->nspname = pstrdup(logicalrep_read_namespace(in));
+ seqdata->seqname = pstrdup(pq_getmsgstring(in));
+
+ seqdata->last_value = pq_getmsgint64(in);
+ seqdata->log_cnt = pq_getmsgint64(in);
+ seqdata->is_called = pq_getmsgint(in, 1);
+}
+
/*
* Write relation description to the output stream.
*/
@@ -1236,6 +1284,8 @@ logicalrep_message_type(LogicalRepMsgType action)
return "STREAM ABORT";
case LOGICAL_REP_MSG_STREAM_PREPARE:
return "STREAM PREPARE";
+ case LOGICAL_REP_MSG_SEQUENCE:
+ return "SEQUENCE";
}
elog(ERROR, "invalid logical replication message type \"%c\"", action);
diff --git a/src/backend/replication/logical/reorderbuffer.c b/src/backend/replication/logical/reorderbuffer.c
index 31f7381f2d6..b5b6b803b7d 100644
--- a/src/backend/replication/logical/reorderbuffer.c
+++ b/src/backend/replication/logical/reorderbuffer.c
@@ -77,6 +77,8 @@
* a bit more memory to the oldest subtransactions, because it's likely
* they are the source for the next sequence of changes.
*
+ * FIXME
+ *
* -------------------------------------------------------------------------
*/
#include "postgres.h"
@@ -91,6 +93,7 @@
#include "access/xact.h"
#include "access/xlog_internal.h"
#include "catalog/catalog.h"
+#include "commands/sequence.h"
#include "lib/binaryheap.h"
#include "miscadmin.h"
#include "pgstat.h"
@@ -537,6 +540,7 @@ ReorderBufferReturnChange(ReorderBuffer *rb, ReorderBufferChange *change,
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_ABORT:
case REORDER_BUFFER_CHANGE_INTERNAL_COMMAND_ID:
case REORDER_BUFFER_CHANGE_INTERNAL_TUPLECID:
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
break;
}
@@ -867,6 +871,45 @@ ReorderBufferQueueMessage(ReorderBuffer *rb, TransactionId xid,
}
}
+/*
+ * A transactional sequence increment is queued to be processed upon commit
+ * and a non-transactional increment gets processed immediately.
+ *
+ * A sequence update may be both transactional and non-transactional. When
+ * created in a running transaction, treat it as transactional and queue
+ * the change in it. Otherwise treat it as non-transactional, so that we
+ * don't forget the increment in case of a rollback.
+ */
+void
+ReorderBufferQueueSequence(ReorderBuffer *rb, TransactionId xid,
+ XLogRecPtr lsn, RepOriginId origin_id,
+ Oid reloid, RelFileLocator locator,
+ int64 last, int64 log_cnt, bool is_called)
+{
+ MemoryContext oldcontext;
+ ReorderBufferChange *change;
+
+ /* OK, allocate and queue the change */
+ oldcontext = MemoryContextSwitchTo(rb->context);
+
+ change = ReorderBufferGetChange(rb);
+
+ change->action = REORDER_BUFFER_CHANGE_SEQUENCE;
+ change->origin_id = origin_id;
+
+ memcpy(&change->data.sequence.locator, &locator, sizeof(RelFileLocator));
+
+ change->data.sequence.reloid = reloid;
+ change->data.sequence.last = last;
+ change->data.sequence.log_cnt = log_cnt;
+ change->data.sequence.is_called = is_called;
+
+ /* add it to the same subxact that created the sequence */
+ ReorderBufferQueueChange(rb, xid, lsn, change, false);
+
+ MemoryContextSwitchTo(oldcontext);
+}
+
/*
* AssertTXNLsnOrder
* Verify LSN ordering of transaction lists in the reorderbuffer
@@ -1975,6 +2018,33 @@ ReorderBufferApplyMessage(ReorderBuffer *rb, ReorderBufferTXN *txn,
change->data.msg.message);
}
+/*
+ * Helper function for ReorderBufferProcessTXN for applying sequences.
+ */
+static inline void
+ReorderBufferApplySequence(ReorderBuffer *rb, ReorderBufferTXN *txn,
+ Relation relation, ReorderBufferChange *change,
+ bool streaming)
+{
+ int64 last_value, log_cnt;
+ bool is_called;
+/*
+ * FIXME is it possible that we write sequence state for T1 before T2, but
+ * then end up committing T2 first? If the sequence could be incremented
+ * in between, that might cause data corruption result.
+ */
+ last_value = change->data.sequence.last;
+ log_cnt = change->data.sequence.log_cnt;
+ is_called = change->data.sequence.is_called;
+
+ if (streaming)
+ rb->stream_sequence(rb, txn, change->lsn, relation,
+ last_value, log_cnt, is_called);
+ else
+ rb->sequence(rb, txn, change->lsn, relation,
+ last_value, log_cnt, is_called);
+}
+
/*
* Function to store the command id and snapshot at the end of the current
* stream so that we can reuse the same while sending the next stream.
@@ -2418,6 +2488,33 @@ ReorderBufferProcessTXN(ReorderBuffer *rb, ReorderBufferTXN *txn,
case REORDER_BUFFER_CHANGE_INTERNAL_TUPLECID:
elog(ERROR, "tuplecid value in changequeue");
break;
+
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
+ Assert(snapshot_now);
+
+ /*
+ reloid = RelidByRelfilenode(change->data.sequence.relnode.spcNode,
+ change->data.sequence.relnode.relNode);
+
+ if (reloid == InvalidOid)
+ elog(ERROR, "zz could not map filenode \"%s\" to relation OID",
+ relpathperm(change->data.sequence.relnode,
+ MAIN_FORKNUM));
+
+ */
+ relation = RelationIdGetRelation(change->data.sequence.reloid);
+
+ if (!RelationIsValid(relation))
+ elog(ERROR, "could not open relation with OID %u (for filenode \"%s\")",
+ RelationGetRelid(relation),
+ relpathperm(change->data.sequence.locator,
+ MAIN_FORKNUM));
+
+ if (RelationIsLogicallyLogged(relation))
+ ReorderBufferApplySequence(rb, txn, relation, change, streaming);
+
+ RelationClose(relation);
+ break;
}
}
@@ -3856,6 +3953,7 @@ ReorderBufferSerializeChange(ReorderBuffer *rb, ReorderBufferTXN *txn,
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_ABORT:
case REORDER_BUFFER_CHANGE_INTERNAL_COMMAND_ID:
case REORDER_BUFFER_CHANGE_INTERNAL_TUPLECID:
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
/* ReorderBufferChange contains everything important */
break;
}
@@ -4120,6 +4218,7 @@ ReorderBufferChangeSize(ReorderBufferChange *change)
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_ABORT:
case REORDER_BUFFER_CHANGE_INTERNAL_COMMAND_ID:
case REORDER_BUFFER_CHANGE_INTERNAL_TUPLECID:
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
/* ReorderBufferChange contains everything important */
break;
}
@@ -4421,6 +4520,7 @@ ReorderBufferRestoreChange(ReorderBuffer *rb, ReorderBufferTXN *txn,
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_ABORT:
case REORDER_BUFFER_CHANGE_INTERNAL_COMMAND_ID:
case REORDER_BUFFER_CHANGE_INTERNAL_TUPLECID:
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
break;
}
diff --git a/src/backend/replication/logical/tablesync.c b/src/backend/replication/logical/tablesync.c
index 94e813ac53c..9bdfcd4fa40 100644
--- a/src/backend/replication/logical/tablesync.c
+++ b/src/backend/replication/logical/tablesync.c
@@ -100,6 +100,7 @@
#include "catalog/pg_subscription_rel.h"
#include "catalog/pg_type.h"
#include "commands/copy.h"
+#include "commands/sequence.h"
#include "miscadmin.h"
#include "parser/parse_relation.h"
#include "pgstat.h"
@@ -1168,6 +1169,94 @@ copy_table(Relation rel)
logicalrep_rel_close(relmapentry, NoLock);
}
+/*
+ * Fetch sequence data (current state) from the remote node.
+ */
+static void
+fetch_sequence_data(char *nspname, char *relname,
+ int64 *last_value, int64 *log_cnt, bool *is_called)
+{
+ WalRcvExecResult *res;
+ StringInfoData cmd;
+ TupleTableSlot *slot;
+ Oid tableRow[3] = {INT8OID, INT8OID, BOOLOID};
+
+ initStringInfo(&cmd);
+ appendStringInfo(&cmd, "SELECT last_value, log_cnt, is_called\n"
+ " FROM %s", quote_qualified_identifier(nspname, relname));
+
+ res = walrcv_exec(LogRepWorkerWalRcvConn, cmd.data, 3, tableRow);
+ pfree(cmd.data);
+
+ if (res->status != WALRCV_OK_TUPLES)
+ ereport(ERROR,
+ (errmsg("could not receive list of replicated tables from the publisher: %s",
+ res->err)));
+
+ /* Process the sequence. */
+ slot = MakeSingleTupleTableSlot(res->tupledesc, &TTSOpsMinimalTuple);
+ while (tuplestore_gettupleslot(res->tuplestore, true, false, slot))
+ {
+ bool isnull;
+
+ *last_value = DatumGetInt64(slot_getattr(slot, 1, &isnull));
+ Assert(!isnull);
+
+ *log_cnt = DatumGetInt64(slot_getattr(slot, 2, &isnull));
+ Assert(!isnull);
+
+ *is_called = DatumGetBool(slot_getattr(slot, 3, &isnull));
+ Assert(!isnull);
+
+ ExecClearTuple(slot);
+ }
+ ExecDropSingleTupleTableSlot(slot);
+
+ walrcv_clear_result(res);
+}
+
+/*
+ * Copy existing data of a sequence from publisher.
+ *
+ * Caller is responsible for locking the local relation.
+ */
+static void
+copy_sequence(Relation rel)
+{
+ LogicalRepRelMapEntry *relmapentry;
+ LogicalRepRelation lrel;
+ List *qual = NIL;
+ StringInfoData cmd;
+ int64 last_value = 0,
+ log_cnt = 0;
+ bool is_called = 0;
+
+ /* Get the publisher relation info. */
+ fetch_remote_table_info(get_namespace_name(RelationGetNamespace(rel)),
+ RelationGetRelationName(rel), &lrel, &qual);
+
+ /* sequences don't have row filters */
+ Assert(!qual);
+
+ /* Put the relation into relmap. */
+ logicalrep_relmap_update(&lrel);
+
+ /* Map the publisher relation to local one. */
+ relmapentry = logicalrep_rel_open(lrel.remoteid, NoLock);
+ Assert(rel == relmapentry->localrel);
+
+ /* Start copy on the publisher. */
+ initStringInfo(&cmd);
+
+ Assert(lrel.relkind == RELKIND_SEQUENCE);
+
+ fetch_sequence_data(lrel.nspname, lrel.relname, &last_value, &log_cnt, &is_called);
+
+ SetSequence(RelationGetRelid(rel), last_value, log_cnt, is_called);
+
+ logicalrep_rel_close(relmapentry, NoLock);
+}
+
/*
* Determine the tablesync slot name.
*
@@ -1417,10 +1506,21 @@ LogicalRepSyncTableStart(XLogRecPtr *origin_startpos)
originname)));
}
- /* Now do the initial data copy */
- PushActiveSnapshot(GetTransactionSnapshot());
- copy_table(rel);
- PopActiveSnapshot();
+ /* Do the right action depending on the relation kind. */
+ if (get_rel_relkind(RelationGetRelid(rel)) == RELKIND_SEQUENCE)
+ {
+ /* Now do the initial sequence copy */
+ PushActiveSnapshot(GetTransactionSnapshot());
+ copy_sequence(rel);
+ PopActiveSnapshot();
+ }
+ else
+ {
+ /* Now do the initial data copy */
+ PushActiveSnapshot(GetTransactionSnapshot());
+ copy_table(rel);
+ PopActiveSnapshot();
+ }
res = walrcv_exec(LogRepWorkerWalRcvConn, "COMMIT", 0, NULL);
if (res->status != WALRCV_OK_COMMAND)
diff --git a/src/backend/replication/logical/worker.c b/src/backend/replication/logical/worker.c
index e48a3f589ae..4c7ab140213 100644
--- a/src/backend/replication/logical/worker.c
+++ b/src/backend/replication/logical/worker.c
@@ -143,6 +143,7 @@
#include "catalog/pg_subscription.h"
#include "catalog/pg_subscription_rel.h"
#include "catalog/pg_tablespace.h"
+#include "commands/sequence.h"
#include "commands/tablecmds.h"
#include "commands/tablespace.h"
#include "commands/trigger.h"
@@ -1167,6 +1168,47 @@ apply_handle_origin(StringInfo s)
errmsg_internal("ORIGIN message sent out of order")));
}
+/*
+ * Handle SEQUENCE message.
+ */
+static void
+apply_handle_sequence(StringInfo s)
+{
+ LogicalRepSequence seq;
+ Oid relid;
+
+ if (handle_streamed_transaction(LOGICAL_REP_MSG_SEQUENCE, s))
+ return;
+
+ logicalrep_read_sequence(s, &seq);
+
+ /*
+ * We should be in remote transaction.
+ */
+ Assert(in_remote_transaction);
+
+ /*
+ * Make sure we're in a transaction (needed by SetSequence). For
+ * non-transactional updates we're guaranteed to start a new one,
+ * and we'll commit it at the end.
+ */
+ if (!IsTransactionState())
+ {
+ StartTransactionCommand();
+ maybe_reread_subscription();
+ }
+
+ relid = RangeVarGetRelid(makeRangeVar(seq.nspname,
+ seq.seqname, -1),
+ RowExclusiveLock, false);
+
+ /* lock the sequence in AccessExclusiveLock, as expected by SetSequence */
+ LockRelationOid(relid, AccessExclusiveLock);
+
+ /* apply the sequence change */
+ SetSequence(relid, seq.last_value, seq.log_cnt, seq.is_called);
+}
+
/*
* Handle STREAM START message.
*/
@@ -2567,6 +2609,10 @@ apply_dispatch(StringInfo s)
*/
break;
+ case LOGICAL_REP_MSG_SEQUENCE:
+ apply_handle_sequence(s);
+ return;
+
case LOGICAL_REP_MSG_STREAM_START:
apply_handle_stream_start(s);
break;
diff --git a/src/backend/replication/pgoutput/pgoutput.c b/src/backend/replication/pgoutput/pgoutput.c
index 2ecaa5b9074..648ac178d49 100644
--- a/src/backend/replication/pgoutput/pgoutput.c
+++ b/src/backend/replication/pgoutput/pgoutput.c
@@ -15,6 +15,7 @@
#include "access/tupconvert.h"
#include "catalog/partition.h"
#include "catalog/pg_publication.h"
+#include "catalog/pg_publication_namespace.h"
#include "catalog/pg_publication_rel.h"
#include "catalog/pg_subscription.h"
#include "commands/defrem.h"
@@ -53,6 +54,10 @@ static void pgoutput_message(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn, XLogRecPtr message_lsn,
bool transactional, const char *prefix,
Size sz, const char *message);
+static void pgoutput_sequence(LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn, XLogRecPtr sequence_lsn,
+ Relation relation,
+ int64 last_value, int64 log_cnt, bool is_called);
static bool pgoutput_origin_filter(LogicalDecodingContext *ctx,
RepOriginId origin_id);
static void pgoutput_begin_prepare_txn(LogicalDecodingContext *ctx,
@@ -256,6 +261,7 @@ _PG_output_plugin_init(OutputPluginCallbacks *cb)
cb->change_cb = pgoutput_change;
cb->truncate_cb = pgoutput_truncate;
cb->message_cb = pgoutput_message;
+ cb->sequence_cb = pgoutput_sequence;
cb->commit_cb = pgoutput_commit_txn;
cb->begin_prepare_cb = pgoutput_begin_prepare_txn;
@@ -272,6 +278,7 @@ _PG_output_plugin_init(OutputPluginCallbacks *cb)
cb->stream_commit_cb = pgoutput_stream_commit;
cb->stream_change_cb = pgoutput_change;
cb->stream_message_cb = pgoutput_message;
+ cb->stream_sequence_cb = pgoutput_sequence;
cb->stream_truncate_cb = pgoutput_truncate;
/* transaction streaming - two-phase commit */
cb->stream_prepare_cb = pgoutput_stream_prepare_txn;
@@ -285,6 +292,7 @@ parse_output_parameters(List *options, PGOutputData *data)
bool publication_names_given = false;
bool binary_option_given = false;
bool messages_option_given = false;
+ bool sequences_option_given = false;
bool streaming_given = false;
bool two_phase_option_given = false;
bool origin_option_given = false;
@@ -293,6 +301,7 @@ parse_output_parameters(List *options, PGOutputData *data)
data->streaming = false;
data->messages = false;
data->two_phase = false;
+ data->sequences = true;
foreach(lc, options)
{
@@ -361,6 +370,16 @@ parse_output_parameters(List *options, PGOutputData *data)
data->messages = defGetBoolean(defel);
}
+ else if (strcmp(defel->defname, "sequences") == 0)
+ {
+ if (sequences_option_given)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("conflicting or redundant options")));
+ sequences_option_given = true;
+
+ data->sequences = defGetBoolean(defel);
+ }
else if (strcmp(defel->defname, "streaming") == 0)
{
if (streaming_given)
@@ -885,9 +904,10 @@ pgoutput_row_filter_init(PGOutputData *data, List *publications,
* (even if other publications have a row filter).
*/
if (!pub->alltables &&
- !SearchSysCacheExists2(PUBLICATIONNAMESPACEMAP,
+ !SearchSysCacheExists3(PUBLICATIONNAMESPACEMAP,
ObjectIdGetDatum(schemaid),
- ObjectIdGetDatum(pub->oid)))
+ ObjectIdGetDatum(pub->oid),
+ CharGetDatum(PUB_OBJTYPE_TABLE)))
{
/*
* Check for the presence of a row filter in this publication.
@@ -1712,6 +1732,55 @@ pgoutput_message(LogicalDecodingContext *ctx, ReorderBufferTXN *txn,
OutputPluginWrite(ctx, true);
}
+static void
+pgoutput_sequence(LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn, XLogRecPtr sequence_lsn,
+ Relation relation,
+ int64 last_value, int64 log_cnt, bool is_called)
+{
+ PGOutputData *data = (PGOutputData *) ctx->output_plugin_private;
+ TransactionId xid = InvalidTransactionId;
+ RelationSyncEntry *relentry;
+ PGOutputTxnData *txndata = (PGOutputTxnData *) txn->output_plugin_private;
+
+ if (!data->sequences)
+ return;
+
+ if (!is_publishable_relation(relation))
+ return;
+
+ /*
+ * Remember the xid for the message in streaming mode. See
+ * pgoutput_change.
+ */
+ if (in_streaming)
+ xid = txn->xid;
+
+ relentry = get_rel_sync_entry(data, relation);
+
+ /*
+ * First check the sequence filter.
+ *
+ * We handle just REORDER_BUFFER_CHANGE_SEQUENCE here.
+ */
+ if (!relentry->pubactions.pubsequence)
+ return;
+
+ /* Send BEGIN if we haven't yet */
+ if (txndata && !txndata->sent_begin_txn)
+ pgoutput_send_begin(ctx, txn);
+
+ OutputPluginPrepareWrite(ctx, true);
+ logicalrep_write_sequence(ctx->out,
+ relation,
+ xid,
+ sequence_lsn,
+ last_value,
+ log_cnt,
+ is_called);
+ OutputPluginWrite(ctx, true);
+}
+
/*
* Return true if the data is associated with an origin and the user has
* requested the changes that don't have an origin, false otherwise.
@@ -1993,7 +2062,8 @@ get_rel_sync_entry(PGOutputData *data, Relation relation)
entry->schema_sent = false;
entry->streamed_txns = NIL;
entry->pubactions.pubinsert = entry->pubactions.pubupdate =
- entry->pubactions.pubdelete = entry->pubactions.pubtruncate = false;
+ entry->pubactions.pubdelete = entry->pubactions.pubtruncate =
+ entry->pubactions.pubsequence = false;
entry->new_slot = NULL;
entry->old_slot = NULL;
memset(entry->exprstate, 0, sizeof(entry->exprstate));
@@ -2008,18 +2078,18 @@ get_rel_sync_entry(PGOutputData *data, Relation relation)
{
Oid schemaId = get_rel_namespace(relid);
List *pubids = GetRelationPublications(relid);
-
+ char relkind = get_rel_relkind(relid);
+ char objectType = pub_get_object_type_for_relkind(relkind);
/*
* We don't acquire a lock on the namespace system table as we build
* the cache entry using a historic snapshot and all the later changes
* are absorbed while decoding WAL.
*/
- List *schemaPubids = GetSchemaPublications(schemaId);
+ List *schemaPubids = GetSchemaPublications(schemaId, objectType);
ListCell *lc;
Oid publish_as_relid = relid;
int publish_ancestor_level = 0;
bool am_partition = get_rel_relispartition(relid);
- char relkind = get_rel_relkind(relid);
List *rel_publications = NIL;
/* Reload publications if needed before use. */
@@ -2051,6 +2121,7 @@ get_rel_sync_entry(PGOutputData *data, Relation relation)
entry->pubactions.pubupdate = false;
entry->pubactions.pubdelete = false;
entry->pubactions.pubtruncate = false;
+ entry->pubactions.pubsequence = false;
/*
* Tuple slots cleanups. (Will be rebuilt later if needed).
@@ -2098,9 +2169,11 @@ get_rel_sync_entry(PGOutputData *data, Relation relation)
/*
* If this is a FOR ALL TABLES publication, pick the partition
- * root and set the ancestor level accordingly.
+ * root and set the ancestor level accordingly. If this is a FOR ALL
+ * SEQUENCES publication, we publish it too but we don't need to
+ * pick the partition root etc.
*/
- if (pub->alltables)
+ if (pub->alltables || pub->allsequences)
{
publish = true;
if (pub->pubviaroot && am_partition)
@@ -2164,6 +2237,7 @@ get_rel_sync_entry(PGOutputData *data, Relation relation)
entry->pubactions.pubupdate |= pub->pubactions.pubupdate;
entry->pubactions.pubdelete |= pub->pubactions.pubdelete;
entry->pubactions.pubtruncate |= pub->pubactions.pubtruncate;
+ entry->pubactions.pubsequence |= pub->pubactions.pubsequence;
/*
* We want to publish the changes as the top-most ancestor
diff --git a/src/backend/utils/cache/relcache.c b/src/backend/utils/cache/relcache.c
index bd6cd4e47b5..d295fed27b3 100644
--- a/src/backend/utils/cache/relcache.c
+++ b/src/backend/utils/cache/relcache.c
@@ -57,6 +57,7 @@
#include "catalog/pg_opclass.h"
#include "catalog/pg_proc.h"
#include "catalog/pg_publication.h"
+#include "catalog/pg_publication_namespace.h"
#include "catalog/pg_rewrite.h"
#include "catalog/pg_shseclabel.h"
#include "catalog/pg_statistic_ext.h"
@@ -5616,6 +5617,8 @@ RelationBuildPublicationDesc(Relation relation, PublicationDesc *pubdesc)
Oid schemaid;
List *ancestors = NIL;
Oid relid = RelationGetRelid(relation);
+ char relkind = relation->rd_rel->relkind;
+ char objType;
/*
* If not publishable, it publishes no actions. (pgoutput_change() will
@@ -5646,8 +5649,15 @@ RelationBuildPublicationDesc(Relation relation, PublicationDesc *pubdesc)
/* Fetch the publication membership info. */
puboids = GetRelationPublications(relid);
schemaid = RelationGetNamespace(relation);
- puboids = list_concat_unique_oid(puboids, GetSchemaPublications(schemaid));
+ objType = pub_get_object_type_for_relkind(relkind);
+ puboids = list_concat_unique_oid(puboids,
+ GetSchemaPublications(schemaid, objType));
+
+ /*
+ * If this is a partion (and thus a table), lookup all ancestors and track
+ * all publications them too.
+ */
if (relation->rd_rel->relispartition)
{
/* Add publications that the ancestors are in too. */
@@ -5659,12 +5669,23 @@ RelationBuildPublicationDesc(Relation relation, PublicationDesc *pubdesc)
puboids = list_concat_unique_oid(puboids,
GetRelationPublications(ancestor));
+
+ /* include all publications publishing schema of all ancestors */
schemaid = get_rel_namespace(ancestor);
puboids = list_concat_unique_oid(puboids,
- GetSchemaPublications(schemaid));
+ GetSchemaPublications(schemaid,
+ PUB_OBJTYPE_TABLE));
}
}
- puboids = list_concat_unique_oid(puboids, GetAllTablesPublications());
+
+ /*
+ * Consider also FOR ALL TABLES and FOR ALL SEQUENCES publications,
+ * depending on the relkind of the relation.
+ */
+ if (relation->rd_rel->relkind == RELKIND_SEQUENCE)
+ puboids = list_concat_unique_oid(puboids, GetAllSequencesPublications());
+ else
+ puboids = list_concat_unique_oid(puboids, GetAllTablesPublications());
foreach(lc, puboids)
{
@@ -5683,6 +5704,7 @@ RelationBuildPublicationDesc(Relation relation, PublicationDesc *pubdesc)
pubdesc->pubactions.pubupdate |= pubform->pubupdate;
pubdesc->pubactions.pubdelete |= pubform->pubdelete;
pubdesc->pubactions.pubtruncate |= pubform->pubtruncate;
+ pubdesc->pubactions.pubsequence |= pubform->pubsequence;
/*
* Check if all columns referenced in the filter expression are part
diff --git a/src/backend/utils/cache/syscache.c b/src/backend/utils/cache/syscache.c
index eec644ec848..dbbb31a0900 100644
--- a/src/backend/utils/cache/syscache.c
+++ b/src/backend/utils/cache/syscache.c
@@ -653,12 +653,12 @@ static const struct cachedesc cacheinfo[] = {
64
},
{PublicationNamespaceRelationId, /* PUBLICATIONNAMESPACEMAP */
- PublicationNamespacePnnspidPnpubidIndexId,
- 2,
+ PublicationNamespacePnnspidPnpubidPntypeIndexId,
+ 3,
{
Anum_pg_publication_namespace_pnnspid,
Anum_pg_publication_namespace_pnpubid,
- 0,
+ Anum_pg_publication_namespace_pntype,
0
},
64
diff --git a/src/bin/pg_dump/pg_dump.c b/src/bin/pg_dump/pg_dump.c
index da427f4d4a1..ad8e255ee9d 100644
--- a/src/bin/pg_dump/pg_dump.c
+++ b/src/bin/pg_dump/pg_dump.c
@@ -3893,10 +3893,12 @@ getPublications(Archive *fout, int *numPublications)
int i_pubname;
int i_pubowner;
int i_puballtables;
+ int i_puballsequences;
int i_pubinsert;
int i_pubupdate;
int i_pubdelete;
int i_pubtruncate;
+ int i_pubsequence;
int i_pubviaroot;
int i,
ntups;
@@ -3912,23 +3914,32 @@ getPublications(Archive *fout, int *numPublications)
resetPQExpBuffer(query);
/* Get the publications. */
- if (fout->remoteVersion >= 130000)
+ if (fout->remoteVersion >= 160000)
+ appendPQExpBufferStr(query,
+ "SELECT p.tableoid, p.oid, p.pubname, "
+ "p.pubowner, "
+ "p.puballtables, p.puballsequences, p.pubinsert, p.pubupdate, p.pubdelete, p.pubtruncate, p.pubsequence, p.pubviaroot "
+ "FROM pg_publication p");
+ else if (fout->remoteVersion >= 130000)
appendPQExpBufferStr(query,
"SELECT p.tableoid, p.oid, p.pubname, "
"p.pubowner, "
"p.puballtables, p.pubinsert, p.pubupdate, p.pubdelete, p.pubtruncate, p.pubviaroot "
+ "p.puballtables, false AS puballsequences, p.pubinsert, p.pubupdate, p.pubdelete, p.pubtruncate, false AS pubsequence, p.pubviaroot "
"FROM pg_publication p");
else if (fout->remoteVersion >= 110000)
appendPQExpBufferStr(query,
"SELECT p.tableoid, p.oid, p.pubname, "
"p.pubowner, "
"p.puballtables, p.pubinsert, p.pubupdate, p.pubdelete, p.pubtruncate, false AS pubviaroot "
+ "p.puballtables, false AS puballsequences, p.pubinsert, p.pubupdate, p.pubdelete, p.pubtruncate, false AS pubsequence, false AS pubviaroot "
"FROM pg_publication p");
else
appendPQExpBufferStr(query,
"SELECT p.tableoid, p.oid, p.pubname, "
"p.pubowner, "
"p.puballtables, p.pubinsert, p.pubupdate, p.pubdelete, false AS pubtruncate, false AS pubviaroot "
+ "p.puballtables, false AS puballsequences, p.pubinsert, p.pubupdate, p.pubdelete, false AS pubtruncate, false AS pubsequence, false AS pubviaroot "
"FROM pg_publication p");
res = ExecuteSqlQuery(fout, query->data, PGRES_TUPLES_OK);
@@ -3940,10 +3951,12 @@ getPublications(Archive *fout, int *numPublications)
i_pubname = PQfnumber(res, "pubname");
i_pubowner = PQfnumber(res, "pubowner");
i_puballtables = PQfnumber(res, "puballtables");
+ i_puballsequences = PQfnumber(res, "puballsequences");
i_pubinsert = PQfnumber(res, "pubinsert");
i_pubupdate = PQfnumber(res, "pubupdate");
i_pubdelete = PQfnumber(res, "pubdelete");
i_pubtruncate = PQfnumber(res, "pubtruncate");
+ i_pubsequence = PQfnumber(res, "pubsequence");
i_pubviaroot = PQfnumber(res, "pubviaroot");
pubinfo = pg_malloc(ntups * sizeof(PublicationInfo));
@@ -3959,6 +3972,8 @@ getPublications(Archive *fout, int *numPublications)
pubinfo[i].rolname = getRoleName(PQgetvalue(res, i, i_pubowner));
pubinfo[i].puballtables =
(strcmp(PQgetvalue(res, i, i_puballtables), "t") == 0);
+ pubinfo[i].puballsequences =
+ (strcmp(PQgetvalue(res, i, i_puballsequences), "t") == 0);
pubinfo[i].pubinsert =
(strcmp(PQgetvalue(res, i, i_pubinsert), "t") == 0);
pubinfo[i].pubupdate =
@@ -3967,6 +3982,8 @@ getPublications(Archive *fout, int *numPublications)
(strcmp(PQgetvalue(res, i, i_pubdelete), "t") == 0);
pubinfo[i].pubtruncate =
(strcmp(PQgetvalue(res, i, i_pubtruncate), "t") == 0);
+ pubinfo[i].pubsequence =
+ (strcmp(PQgetvalue(res, i, i_pubsequence), "t") == 0);
pubinfo[i].pubviaroot =
(strcmp(PQgetvalue(res, i, i_pubviaroot), "t") == 0);
@@ -4012,6 +4029,9 @@ dumpPublication(Archive *fout, const PublicationInfo *pubinfo)
if (pubinfo->puballtables)
appendPQExpBufferStr(query, " FOR ALL TABLES");
+ if (pubinfo->puballsequences)
+ appendPQExpBufferStr(query, " FOR ALL SEQUENCES");
+
appendPQExpBufferStr(query, " WITH (publish = '");
if (pubinfo->pubinsert)
{
@@ -4046,6 +4066,15 @@ dumpPublication(Archive *fout, const PublicationInfo *pubinfo)
first = false;
}
+ if (pubinfo->pubsequence)
+ {
+ if (!first)
+ appendPQExpBufferStr(query, ", ");
+
+ appendPQExpBufferStr(query, "sequence");
+ first = false;
+ }
+
appendPQExpBufferChar(query, '\'');
if (pubinfo->pubviaroot)
@@ -4092,6 +4121,7 @@ getPublicationNamespaces(Archive *fout)
int i_oid;
int i_pnpubid;
int i_pnnspid;
+ int i_pntype;
int i,
j,
ntups;
@@ -4103,7 +4133,7 @@ getPublicationNamespaces(Archive *fout)
/* Collect all publication membership info. */
appendPQExpBufferStr(query,
- "SELECT tableoid, oid, pnpubid, pnnspid "
+ "SELECT tableoid, oid, pnpubid, pnnspid, pntype "
"FROM pg_catalog.pg_publication_namespace");
res = ExecuteSqlQuery(fout, query->data, PGRES_TUPLES_OK);
@@ -4113,6 +4143,7 @@ getPublicationNamespaces(Archive *fout)
i_oid = PQfnumber(res, "oid");
i_pnpubid = PQfnumber(res, "pnpubid");
i_pnnspid = PQfnumber(res, "pnnspid");
+ i_pntype = PQfnumber(res, "pntype");
/* this allocation may be more than we need */
pubsinfo = pg_malloc(ntups * sizeof(PublicationSchemaInfo));
@@ -4122,6 +4153,7 @@ getPublicationNamespaces(Archive *fout)
{
Oid pnpubid = atooid(PQgetvalue(res, i, i_pnpubid));
Oid pnnspid = atooid(PQgetvalue(res, i, i_pnnspid));
+ char pntype = PQgetvalue(res, i, i_pntype)[0];
PublicationInfo *pubinfo;
NamespaceInfo *nspinfo;
@@ -4153,6 +4185,7 @@ getPublicationNamespaces(Archive *fout)
pubsinfo[j].dobj.name = nspinfo->dobj.name;
pubsinfo[j].publication = pubinfo;
pubsinfo[j].pubschema = nspinfo;
+ pubsinfo[j].pubtype = pntype;
/* Decide whether we want to dump it */
selectDumpablePublicationObject(&(pubsinfo[j].dobj), fout);
@@ -4318,7 +4351,10 @@ dumpPublicationNamespace(Archive *fout, const PublicationSchemaInfo *pubsinfo)
query = createPQExpBuffer();
appendPQExpBuffer(query, "ALTER PUBLICATION %s ", fmtId(pubinfo->dobj.name));
- appendPQExpBuffer(query, "ADD TABLES IN SCHEMA %s;\n", fmtId(schemainfo->dobj.name));
+ if (pubsinfo->pubtype == 't')
+ appendPQExpBuffer(query, "ADD TABLES IN SCHEMA %s;\n", fmtId(schemainfo->dobj.name));
+ else
+ appendPQExpBuffer(query, "ADD SEQUENCES IN SCHEMA %s;\n", fmtId(schemainfo->dobj.name));
/*
* There is no point in creating drop query as the drop is done by schema
@@ -4351,6 +4387,7 @@ dumpPublicationTable(Archive *fout, const PublicationRelInfo *pubrinfo)
TableInfo *tbinfo = pubrinfo->pubtable;
PQExpBuffer query;
char *tag;
+ char *description;
/* Do nothing in data-only dump */
if (dopt->dataOnly)
@@ -4360,8 +4397,19 @@ dumpPublicationTable(Archive *fout, const PublicationRelInfo *pubrinfo)
query = createPQExpBuffer();
- appendPQExpBuffer(query, "ALTER PUBLICATION %s ADD TABLE ONLY",
- fmtId(pubinfo->dobj.name));
+ if (tbinfo->relkind == RELKIND_SEQUENCE)
+ {
+ appendPQExpBuffer(query, "ALTER PUBLICATION %s ADD SEQUENCE",
+ fmtId(pubinfo->dobj.name));
+ description = "PUBLICATION SEQUENCE";
+ }
+ else
+ {
+ appendPQExpBuffer(query, "ALTER PUBLICATION %s ADD TABLE ONLY",
+ fmtId(pubinfo->dobj.name));
+ description = "PUBLICATION TABLE";
+ }
+
appendPQExpBuffer(query, " %s",
fmtQualifiedDumpable(tbinfo));
@@ -4391,7 +4439,7 @@ dumpPublicationTable(Archive *fout, const PublicationRelInfo *pubrinfo)
ARCHIVE_OPTS(.tag = tag,
.namespace = tbinfo->dobj.namespace->dobj.name,
.owner = pubinfo->rolname,
- .description = "PUBLICATION TABLE",
+ .description = description,
.section = SECTION_POST_DATA,
.createStmt = query->data));
diff --git a/src/bin/pg_dump/pg_dump.h b/src/bin/pg_dump/pg_dump.h
index 427f5d45f65..45ebc39131b 100644
--- a/src/bin/pg_dump/pg_dump.h
+++ b/src/bin/pg_dump/pg_dump.h
@@ -615,10 +615,12 @@ typedef struct _PublicationInfo
DumpableObject dobj;
const char *rolname;
bool puballtables;
+ bool puballsequences;
bool pubinsert;
bool pubupdate;
bool pubdelete;
bool pubtruncate;
+ bool pubsequence;
bool pubviaroot;
} PublicationInfo;
@@ -644,6 +646,7 @@ typedef struct _PublicationSchemaInfo
DumpableObject dobj;
PublicationInfo *publication;
NamespaceInfo *pubschema;
+ char pubtype;
} PublicationSchemaInfo;
/*
diff --git a/src/bin/pg_dump/t/002_pg_dump.pl b/src/bin/pg_dump/t/002_pg_dump.pl
index 8dc1f0eccb5..f34415e48bb 100644
--- a/src/bin/pg_dump/t/002_pg_dump.pl
+++ b/src/bin/pg_dump/t/002_pg_dump.pl
@@ -2432,7 +2432,7 @@ my %tests = (
create_order => 50,
create_sql => 'CREATE PUBLICATION pub1;',
regexp => qr/^
- \QCREATE PUBLICATION pub1 WITH (publish = 'insert, update, delete, truncate');\E
+ \QCREATE PUBLICATION pub1 WITH (publish = 'insert, update, delete, truncate, sequence');\E
/xm,
like => { %full_runs, section_post_data => 1, },
},
@@ -2452,16 +2452,27 @@ my %tests = (
create_order => 50,
create_sql => 'CREATE PUBLICATION pub3;',
regexp => qr/^
- \QCREATE PUBLICATION pub3 WITH (publish = 'insert, update, delete, truncate');\E
+ \QCREATE PUBLICATION pub3 WITH (publish = 'insert, update, delete, truncate, sequence');\E
/xm,
like => { %full_runs, section_post_data => 1, },
},
'CREATE PUBLICATION pub4' => {
create_order => 50,
- create_sql => 'CREATE PUBLICATION pub4;',
+ create_sql => 'CREATE PUBLICATION pub4
+ FOR ALL SEQUENCES
+ WITH (publish = \'\');',
regexp => qr/^
- \QCREATE PUBLICATION pub4 WITH (publish = 'insert, update, delete, truncate');\E
+ \QCREATE PUBLICATION pub4 FOR ALL SEQUENCES WITH (publish = '');\E
+ /xm,
+ like => { %full_runs, section_post_data => 1, },
+ },
+
+ 'CREATE PUBLICATION pub5' => {
+ create_order => 50,
+ create_sql => 'CREATE PUBLICATION pub5;',
+ regexp => qr/^
+ \QCREATE PUBLICATION pub5 WITH (publish = 'insert, update, delete, truncate, sequence');\E
/xm,
like => { %full_runs, section_post_data => 1, },
},
@@ -2607,6 +2618,27 @@ my %tests = (
unlike => { exclude_dump_test_schema => 1, },
},
+ 'ALTER PUBLICATION pub3 ADD SEQUENCES IN SCHEMA dump_test' => {
+ create_order => 51,
+ create_sql =>
+ 'ALTER PUBLICATION pub3 ADD SEQUENCES IN SCHEMA dump_test;',
+ regexp => qr/^
+ \QALTER PUBLICATION pub3 ADD SEQUENCES IN SCHEMA dump_test;\E
+ /xm,
+ like => { %full_runs, section_post_data => 1, },
+ unlike => { exclude_dump_test_schema => 1, },
+ },
+
+ 'ALTER PUBLICATION pub3 ADD SEQUENCES IN SCHEMA public' => {
+ create_order => 52,
+ create_sql =>
+ 'ALTER PUBLICATION pub3 ADD SEQUENCES IN SCHEMA public;',
+ regexp => qr/^
+ \QALTER PUBLICATION pub3 ADD SEQUENCES IN SCHEMA public;\E
+ /xm,
+ like => { %full_runs, section_post_data => 1, },
+ },
+
'CREATE SCHEMA public' => {
regexp => qr/^CREATE SCHEMA public;/m,
diff --git a/src/bin/psql/describe.c b/src/bin/psql/describe.c
index 2eae519b1dd..f8bbd28def5 100644
--- a/src/bin/psql/describe.c
+++ b/src/bin/psql/describe.c
@@ -1709,28 +1709,19 @@ describeOneTableDetails(const char *schemaname,
if (tableinfo.relkind == RELKIND_SEQUENCE)
{
PGresult *result = NULL;
- printQueryOpt myopt = pset.popt;
- char *footers[2] = {NULL, NULL};
if (pset.sversion >= 100000)
{
printfPQExpBuffer(&buf,
- "SELECT pg_catalog.format_type(seqtypid, NULL) AS \"%s\",\n"
- " seqstart AS \"%s\",\n"
- " seqmin AS \"%s\",\n"
- " seqmax AS \"%s\",\n"
- " seqincrement AS \"%s\",\n"
- " CASE WHEN seqcycle THEN '%s' ELSE '%s' END AS \"%s\",\n"
- " seqcache AS \"%s\"\n",
- gettext_noop("Type"),
- gettext_noop("Start"),
- gettext_noop("Minimum"),
- gettext_noop("Maximum"),
- gettext_noop("Increment"),
+ "SELECT pg_catalog.format_type(seqtypid, NULL),\n"
+ " seqstart,\n"
+ " seqmin,\n"
+ " seqmax,\n"
+ " seqincrement,\n"
+ " CASE WHEN seqcycle THEN '%s' ELSE '%s' END,\n"
+ " seqcache\n",
gettext_noop("yes"),
- gettext_noop("no"),
- gettext_noop("Cycles?"),
- gettext_noop("Cache"));
+ gettext_noop("no"));
appendPQExpBuffer(&buf,
"FROM pg_catalog.pg_sequence\n"
"WHERE seqrelid = '%s';",
@@ -1739,22 +1730,15 @@ describeOneTableDetails(const char *schemaname,
else
{
printfPQExpBuffer(&buf,
- "SELECT 'bigint' AS \"%s\",\n"
- " start_value AS \"%s\",\n"
- " min_value AS \"%s\",\n"
- " max_value AS \"%s\",\n"
- " increment_by AS \"%s\",\n"
- " CASE WHEN is_cycled THEN '%s' ELSE '%s' END AS \"%s\",\n"
- " cache_value AS \"%s\"\n",
- gettext_noop("Type"),
- gettext_noop("Start"),
- gettext_noop("Minimum"),
- gettext_noop("Maximum"),
- gettext_noop("Increment"),
+ "SELECT 'bigint',\n"
+ " start_value,\n"
+ " min_value,\n"
+ " max_value,\n"
+ " increment_by,\n"
+ " CASE WHEN is_cycled THEN '%s' ELSE '%s' END,\n"
+ " cache_value\n",
gettext_noop("yes"),
- gettext_noop("no"),
- gettext_noop("Cycles?"),
- gettext_noop("Cache"));
+ gettext_noop("no"));
appendPQExpBuffer(&buf, "FROM %s", fmtId(schemaname));
/* must be separate because fmtId isn't reentrant */
appendPQExpBuffer(&buf, ".%s;", fmtId(relationname));
@@ -1764,6 +1748,57 @@ describeOneTableDetails(const char *schemaname,
if (!res)
goto error_return;
+ numrows = PQntuples(res);
+
+ /* XXX reset to use expanded output for sequences (maybe we should
+ * keep this disabled, just like for tables?) */
+ myopt.expanded = pset.popt.topt.expanded;
+
+ printTableInit(&cont, &myopt, title.data, 7, numrows);
+ printTableInitialized = true;
+
+ if (tableinfo.relpersistence == 'u')
+ printfPQExpBuffer(&title, _("Unlogged sequence \"%s.%s\""),
+ schemaname, relationname);
+ else
+ printfPQExpBuffer(&title, _("Sequence \"%s.%s\""),
+ schemaname, relationname);
+
+ printTableAddHeader(&cont, gettext_noop("Type"), true, 'l');
+ printTableAddHeader(&cont, gettext_noop("Start"), true, 'r');
+ printTableAddHeader(&cont, gettext_noop("Minimum"), true, 'r');
+ printTableAddHeader(&cont, gettext_noop("Maximum"), true, 'r');
+ printTableAddHeader(&cont, gettext_noop("Increment"), true, 'r');
+ printTableAddHeader(&cont, gettext_noop("Cycles?"), true, 'l');
+ printTableAddHeader(&cont, gettext_noop("Cache"), true, 'r');
+
+ /* Generate table cells to be printed */
+ for (i = 0; i < numrows; i++)
+ {
+ /* Type */
+ printTableAddCell(&cont, PQgetvalue(res, i, 0), false, false);
+
+ /* Start */
+ printTableAddCell(&cont, PQgetvalue(res, i, 1), false, false);
+
+ /* Minimum */
+ printTableAddCell(&cont, PQgetvalue(res, i, 2), false, false);
+
+ /* Maximum */
+ printTableAddCell(&cont, PQgetvalue(res, i, 3), false, false);
+
+ /* Increment */
+ printTableAddCell(&cont, PQgetvalue(res, i, 4), false, false);
+
+ /* Cycles? */
+ printTableAddCell(&cont, PQgetvalue(res, i, 5), false, false);
+
+ /* Cache */
+ printTableAddCell(&cont, PQgetvalue(res, i, 6), false, false);
+ }
+
+ /* Footer information about a sequence */
+
/* Get the column that owns this sequence */
printfPQExpBuffer(&buf, "SELECT pg_catalog.quote_ident(nspname) || '.' ||"
"\n pg_catalog.quote_ident(relname) || '.' ||"
@@ -1795,32 +1830,63 @@ describeOneTableDetails(const char *schemaname,
switch (PQgetvalue(result, 0, 1)[0])
{
case 'a':
- footers[0] = psprintf(_("Owned by: %s"),
- PQgetvalue(result, 0, 0));
+ printTableAddFooter(&cont,
+ psprintf(_("Owned by: %s"),
+ PQgetvalue(result, 0, 0)));
break;
case 'i':
- footers[0] = psprintf(_("Sequence for identity column: %s"),
- PQgetvalue(result, 0, 0));
+ printTableAddFooter(&cont,
+ psprintf(_("Sequence for identity column: %s"),
+ PQgetvalue(result, 0, 0)));
break;
}
}
PQclear(result);
- if (tableinfo.relpersistence == 'u')
- printfPQExpBuffer(&title, _("Unlogged sequence \"%s.%s\""),
- schemaname, relationname);
- else
- printfPQExpBuffer(&title, _("Sequence \"%s.%s\""),
- schemaname, relationname);
+ /* print any publications */
+ if (pset.sversion >= 160000)
+ {
+ int tuples = 0;
- myopt.footers = footers;
- myopt.topt.default_footer = false;
- myopt.title = title.data;
- myopt.translate_header = true;
+ printfPQExpBuffer(&buf,
+ "SELECT pubname\n"
+ "FROM pg_catalog.pg_publication p\n"
+ " JOIN pg_catalog.pg_publication_namespace pn ON p.oid = pn.pnpubid\n"
+ " JOIN pg_catalog.pg_class pc ON pc.relnamespace = pn.pnnspid\n"
+ "WHERE pc.oid ='%s' and pn.pntype = 's' and pg_catalog.pg_relation_is_publishable('%s')\n"
+ "UNION\n"
+ "SELECT pubname\n"
+ "FROM pg_catalog.pg_publication p\n"
+ " JOIN pg_catalog.pg_publication_rel pr ON p.oid = pr.prpubid\n"
+ "WHERE pr.prrelid = '%s'\n"
+ "UNION\n"
+ "SELECT pubname\n"
+ "FROM pg_catalog.pg_publication p\n"
+ "WHERE p.puballsequences AND pg_catalog.pg_relation_is_publishable('%s')\n"
+ "ORDER BY 1;",
+ oid, oid, oid, oid);
- printQuery(res, &myopt, pset.queryFout, false, pset.logfile);
+ result = PSQLexec(buf.data);
+ if (!result)
+ goto error_return;
+ else
+ tuples = PQntuples(result);
- free(footers[0]);
+ if (tuples > 0)
+ printTableAddFooter(&cont, _("Publications:"));
+
+ /* Might be an empty set - that's ok */
+ for (i = 0; i < tuples; i++)
+ {
+ printfPQExpBuffer(&buf, " \"%s\"",
+ PQgetvalue(result, i, 0));
+
+ printTableAddFooter(&cont, buf.data);
+ }
+ PQclear(result);
+ }
+
+ printTable(&cont, pset.queryFout, false, pset.logfile);
retval = true;
goto error_return; /* not an error, just return early */
@@ -2047,6 +2113,11 @@ describeOneTableDetails(const char *schemaname,
for (i = 0; i < cols; i++)
printTableAddHeader(&cont, headers[i], true, 'l');
+ res = PSQLexec(buf.data);
+ if (!res)
+ goto error_return;
+ numrows = PQntuples(res);
+
/* Generate table cells to be printed */
for (i = 0; i < numrows; i++)
{
@@ -2972,7 +3043,7 @@ describeOneTableDetails(const char *schemaname,
"FROM pg_catalog.pg_publication p\n"
" JOIN pg_catalog.pg_publication_namespace pn ON p.oid = pn.pnpubid\n"
" JOIN pg_catalog.pg_class pc ON pc.relnamespace = pn.pnnspid\n"
- "WHERE pc.oid ='%s' and pg_catalog.pg_relation_is_publishable('%s')\n"
+ "WHERE pc.oid ='%s' and pn.pntype = 't' and pg_catalog.pg_relation_is_publishable('%s')\n"
"UNION\n"
"SELECT pubname\n"
" , pg_get_expr(pr.prqual, c.oid)\n"
@@ -5005,7 +5076,7 @@ listSchemas(const char *pattern, bool verbose, bool showSystem)
int i;
printfPQExpBuffer(&buf,
- "SELECT pubname \n"
+ "SELECT pubname, (CASE WHEN pntype = 't' THEN 'tables' ELSE 'sequences' END) AS pubtype\n"
"FROM pg_catalog.pg_publication p\n"
" JOIN pg_catalog.pg_publication_namespace pn ON p.oid = pn.pnpubid\n"
" JOIN pg_catalog.pg_namespace n ON n.oid = pn.pnnspid \n"
@@ -5031,8 +5102,9 @@ listSchemas(const char *pattern, bool verbose, bool showSystem)
/* Might be an empty set - that's ok */
for (i = 0; i < pub_schema_tuples; i++)
{
- printfPQExpBuffer(&buf, " \"%s\"",
- PQgetvalue(result, i, 0));
+ printfPQExpBuffer(&buf, " \"%s\" (%s)",
+ PQgetvalue(result, i, 0),
+ PQgetvalue(result, i, 1));
footers[i + 1] = pg_strdup(buf.data);
}
@@ -6161,7 +6233,7 @@ listPublications(const char *pattern)
PQExpBufferData buf;
PGresult *res;
printQueryOpt myopt = pset.popt;
- static const bool translate_columns[] = {false, false, false, false, false, false, false, false};
+ static const bool translate_columns[] = {false, false, false, false, false, false, false, false, false, false};
if (pset.sversion < 100000)
{
@@ -6175,23 +6247,45 @@ listPublications(const char *pattern)
initPQExpBuffer(&buf);
- printfPQExpBuffer(&buf,
- "SELECT pubname AS \"%s\",\n"
- " pg_catalog.pg_get_userbyid(pubowner) AS \"%s\",\n"
- " puballtables AS \"%s\",\n"
- " pubinsert AS \"%s\",\n"
- " pubupdate AS \"%s\",\n"
- " pubdelete AS \"%s\"",
- gettext_noop("Name"),
- gettext_noop("Owner"),
- gettext_noop("All tables"),
- gettext_noop("Inserts"),
- gettext_noop("Updates"),
- gettext_noop("Deletes"));
+ if (pset.sversion >= 160000)
+ printfPQExpBuffer(&buf,
+ "SELECT pubname AS \"%s\",\n"
+ " pg_catalog.pg_get_userbyid(pubowner) AS \"%s\",\n"
+ " puballtables AS \"%s\",\n"
+ " puballsequences AS \"%s\",\n"
+ " pubinsert AS \"%s\",\n"
+ " pubupdate AS \"%s\",\n"
+ " pubdelete AS \"%s\"",
+ gettext_noop("Name"),
+ gettext_noop("Owner"),
+ gettext_noop("All tables"),
+ gettext_noop("All sequences"),
+ gettext_noop("Inserts"),
+ gettext_noop("Updates"),
+ gettext_noop("Deletes"));
+ else
+ printfPQExpBuffer(&buf,
+ "SELECT pubname AS \"%s\",\n"
+ " pg_catalog.pg_get_userbyid(pubowner) AS \"%s\",\n"
+ " puballtables AS \"%s\",\n"
+ " pubinsert AS \"%s\",\n"
+ " pubupdate AS \"%s\",\n"
+ " pubdelete AS \"%s\"",
+ gettext_noop("Name"),
+ gettext_noop("Owner"),
+ gettext_noop("All tables"),
+ gettext_noop("Inserts"),
+ gettext_noop("Updates"),
+ gettext_noop("Deletes"));
+
if (pset.sversion >= 110000)
appendPQExpBuffer(&buf,
",\n pubtruncate AS \"%s\"",
gettext_noop("Truncates"));
+ if (pset.sversion >= 160000)
+ appendPQExpBuffer(&buf,
+ ",\n pubsequence AS \"%s\"",
+ gettext_noop("Sequences"));
if (pset.sversion >= 130000)
appendPQExpBuffer(&buf,
",\n pubviaroot AS \"%s\"",
@@ -6286,6 +6380,7 @@ describePublications(const char *pattern)
PGresult *res;
bool has_pubtruncate;
bool has_pubviaroot;
+ bool has_pubsequence;
PQExpBufferData title;
printTableContent cont;
@@ -6302,6 +6397,7 @@ describePublications(const char *pattern)
has_pubtruncate = (pset.sversion >= 110000);
has_pubviaroot = (pset.sversion >= 130000);
+ has_pubsequence = (pset.sversion >= 160000);
initPQExpBuffer(&buf);
@@ -6309,12 +6405,17 @@ describePublications(const char *pattern)
"SELECT oid, pubname,\n"
" pg_catalog.pg_get_userbyid(pubowner) AS owner,\n"
" puballtables, pubinsert, pubupdate, pubdelete");
+
if (has_pubtruncate)
appendPQExpBufferStr(&buf,
", pubtruncate");
if (has_pubviaroot)
appendPQExpBufferStr(&buf,
", pubviaroot");
+ if (has_pubsequence)
+ appendPQExpBufferStr(&buf,
+ ", puballsequences, pubsequence");
+
appendPQExpBufferStr(&buf,
"\nFROM pg_catalog.pg_publication\n");
@@ -6360,6 +6461,7 @@ describePublications(const char *pattern)
char *pubid = PQgetvalue(res, i, 0);
char *pubname = PQgetvalue(res, i, 1);
bool puballtables = strcmp(PQgetvalue(res, i, 3), "t") == 0;
+ bool puballsequences = strcmp(PQgetvalue(res, i, 9), "t") == 0;
printTableOpt myopt = pset.popt.topt;
if (has_pubtruncate)
@@ -6367,29 +6469,43 @@ describePublications(const char *pattern)
if (has_pubviaroot)
ncols++;
+ /* sequences have two extra columns (puballsequences, pubsequences) */
+ if (has_pubsequence)
+ ncols += 2;
+
initPQExpBuffer(&title);
printfPQExpBuffer(&title, _("Publication %s"), pubname);
printTableInit(&cont, &myopt, title.data, ncols, nrows);
printTableAddHeader(&cont, gettext_noop("Owner"), true, align);
printTableAddHeader(&cont, gettext_noop("All tables"), true, align);
+ if (has_pubsequence)
+ printTableAddHeader(&cont, gettext_noop("All sequences"), true, align);
printTableAddHeader(&cont, gettext_noop("Inserts"), true, align);
printTableAddHeader(&cont, gettext_noop("Updates"), true, align);
printTableAddHeader(&cont, gettext_noop("Deletes"), true, align);
if (has_pubtruncate)
printTableAddHeader(&cont, gettext_noop("Truncates"), true, align);
+ if (has_pubsequence)
+ printTableAddHeader(&cont, gettext_noop("Sequences"), true, align);
if (has_pubviaroot)
printTableAddHeader(&cont, gettext_noop("Via root"), true, align);
- printTableAddCell(&cont, PQgetvalue(res, i, 2), false, false);
- printTableAddCell(&cont, PQgetvalue(res, i, 3), false, false);
- printTableAddCell(&cont, PQgetvalue(res, i, 4), false, false);
- printTableAddCell(&cont, PQgetvalue(res, i, 5), false, false);
- printTableAddCell(&cont, PQgetvalue(res, i, 6), false, false);
+ printTableAddCell(&cont, PQgetvalue(res, i, 2), false, false); /* owner */
+ printTableAddCell(&cont, PQgetvalue(res, i, 3), false, false); /* all tables */
+
+ if (has_pubsequence)
+ printTableAddCell(&cont, PQgetvalue(res, i, 9), false, false); /* all sequences */
+
+ printTableAddCell(&cont, PQgetvalue(res, i, 4), false, false); /* insert */
+ printTableAddCell(&cont, PQgetvalue(res, i, 5), false, false); /* update */
+ printTableAddCell(&cont, PQgetvalue(res, i, 6), false, false); /* delete */
if (has_pubtruncate)
- printTableAddCell(&cont, PQgetvalue(res, i, 7), false, false);
+ printTableAddCell(&cont, PQgetvalue(res, i, 7), false, false); /* truncate */
+ if (has_pubsequence)
+ printTableAddCell(&cont, PQgetvalue(res, i, 10), false, false); /* sequence */
if (has_pubviaroot)
- printTableAddCell(&cont, PQgetvalue(res, i, 8), false, false);
+ printTableAddCell(&cont, PQgetvalue(res, i, 8), false, false); /* via root */
if (!puballtables)
{
@@ -6420,6 +6536,7 @@ describePublications(const char *pattern)
"WHERE c.relnamespace = n.oid\n"
" AND c.oid = pr.prrelid\n"
" AND pr.prpubid = '%s'\n"
+ " AND c.relkind != 'S'\n" /* exclude sequences */
"ORDER BY 1,2", pubid);
if (!addFooterToPublicationDesc(&buf, _("Tables:"), false, &cont))
goto error_return;
@@ -6431,7 +6548,7 @@ describePublications(const char *pattern)
"SELECT n.nspname\n"
"FROM pg_catalog.pg_namespace n\n"
" JOIN pg_catalog.pg_publication_namespace pn ON n.oid = pn.pnnspid\n"
- "WHERE pn.pnpubid = '%s'\n"
+ "WHERE pn.pnpubid = '%s' AND pn.pntype = 't'\n"
"ORDER BY 1", pubid);
if (!addFooterToPublicationDesc(&buf, _("Tables from schemas:"),
true, &cont))
@@ -6439,6 +6556,37 @@ describePublications(const char *pattern)
}
}
+ if (!puballsequences)
+ {
+ /* Get the sequences for the specified publication */
+ printfPQExpBuffer(&buf,
+ "SELECT n.nspname, c.relname, NULL, NULL\n"
+ "FROM pg_catalog.pg_class c,\n"
+ " pg_catalog.pg_namespace n,\n"
+ " pg_catalog.pg_publication_rel pr\n"
+ "WHERE c.relnamespace = n.oid\n"
+ " AND c.oid = pr.prrelid\n"
+ " AND pr.prpubid = '%s'\n"
+ " AND c.relkind = 'S'\n" /* only sequences */
+ "ORDER BY 1,2", pubid);
+ if (!addFooterToPublicationDesc(&buf, "Sequences:", false, &cont))
+ goto error_return;
+
+ if (pset.sversion >= 160000)
+ {
+ /* Get the schemas for the specified publication */
+ printfPQExpBuffer(&buf,
+ "SELECT n.nspname\n"
+ "FROM pg_catalog.pg_namespace n\n"
+ " JOIN pg_catalog.pg_publication_namespace pn ON n.oid = pn.pnnspid\n"
+ "WHERE pn.pnpubid = '%s' AND pn.pntype = 's'\n"
+ "ORDER BY 1", pubid);
+ if (!addFooterToPublicationDesc(&buf, "Sequences from schemas:",
+ true, &cont))
+ goto error_return;
+ }
+ }
+
printTable(&cont, pset.queryFout, false, pset.logfile);
printTableCleanup(&cont);
diff --git a/src/bin/psql/tab-complete.c b/src/bin/psql/tab-complete.c
index 7b73886ce19..9a5f37c4a5d 100644
--- a/src/bin/psql/tab-complete.c
+++ b/src/bin/psql/tab-complete.c
@@ -1820,11 +1820,15 @@ psql_completion(const char *text, int start, int end)
COMPLETE_WITH("ADD", "DROP", "OWNER TO", "RENAME TO", "SET");
/* ALTER PUBLICATION <name> ADD */
else if (Matches("ALTER", "PUBLICATION", MatchAny, "ADD"))
- COMPLETE_WITH("TABLES IN SCHEMA", "TABLE");
+ COMPLETE_WITH("TABLES IN SCHEMA", "TABLE", "SEQUENCES IN SCHEMA", "SEQUENCE");
else if (Matches("ALTER", "PUBLICATION", MatchAny, "ADD|SET", "TABLE") ||
(HeadMatches("ALTER", "PUBLICATION", MatchAny, "ADD|SET", "TABLE") &&
ends_with(prev_wd, ',')))
COMPLETE_WITH_SCHEMA_QUERY(Query_for_list_of_tables);
+ else if (Matches("ALTER", "PUBLICATION", MatchAny, "ADD|SET", "SEQUENCE") ||
+ (HeadMatches("ALTER", "PUBLICATION", MatchAny, "ADD|SET", "SEQUENCE") &&
+ ends_with(prev_wd, ',')))
+ COMPLETE_WITH_SCHEMA_QUERY(Query_for_list_of_sequences);
/*
* "ALTER PUBLICATION <name> SET TABLE <name> WHERE (" - complete with
@@ -1844,10 +1848,10 @@ psql_completion(const char *text, int start, int end)
COMPLETE_WITH(",");
/* ALTER PUBLICATION <name> DROP */
else if (Matches("ALTER", "PUBLICATION", MatchAny, "DROP"))
- COMPLETE_WITH("TABLES IN SCHEMA", "TABLE");
+ COMPLETE_WITH("TABLES IN SCHEMA", "TABLE", "SEQUENCES IN SCHEMA", "SEQUENCE");
/* ALTER PUBLICATION <name> SET */
else if (Matches("ALTER", "PUBLICATION", MatchAny, "SET"))
- COMPLETE_WITH("(", "TABLES IN SCHEMA", "TABLE");
+ COMPLETE_WITH("(", "TABLES IN SCHEMA", "TABLE", "SEQUENCES IN SCHEMA", "SEQUENCE");
else if (Matches("ALTER", "PUBLICATION", MatchAny, "ADD|DROP|SET", "TABLES", "IN", "SCHEMA"))
COMPLETE_WITH_QUERY_PLUS(Query_for_list_of_schemas
" AND nspname NOT LIKE E'pg\\\\_%%'",
@@ -3017,23 +3021,32 @@ psql_completion(const char *text, int start, int end)
/* CREATE PUBLICATION */
else if (Matches("CREATE", "PUBLICATION", MatchAny))
- COMPLETE_WITH("FOR TABLE", "FOR ALL TABLES", "FOR TABLES IN SCHEMA", "WITH (");
+ COMPLETE_WITH("FOR TABLE", "FOR ALL TABLES", "FOR TABLES IN SCHEMA", "FOR SEQUENCE", "FOR ALL SEQUENCES", "FOR SEQUENCES IN SCHEMA", "WITH (");
else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR"))
- COMPLETE_WITH("TABLE", "ALL TABLES", "TABLES IN SCHEMA");
+ COMPLETE_WITH("TABLE", "ALL TABLES", "TABLES IN SCHEMA", "SEQUENCE", "ALL SEQUENCES", "SEQUENCES IN SCHEMA");
else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "ALL"))
- COMPLETE_WITH("TABLES");
+ COMPLETE_WITH("TABLES", "SEQUENCES");
else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "ALL", "TABLES"))
COMPLETE_WITH("WITH (");
+ else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "ALL", "SEQUENCES"))
+ COMPLETE_WITH("WITH (");
else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "TABLES"))
COMPLETE_WITH("IN SCHEMA");
+ else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "SEQUENCES"))
+ COMPLETE_WITH("IN SCHEMA");
else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "TABLE", MatchAny) && !ends_with(prev_wd, ','))
COMPLETE_WITH("WHERE (", "WITH (");
+ else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "SEQUENCE", MatchAny) && !ends_with(prev_wd, ','))
+ COMPLETE_WITH("WHERE (", "WITH (");
/* Complete "CREATE PUBLICATION <name> FOR TABLE" with "<table>, ..." */
else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "TABLE"))
COMPLETE_WITH_SCHEMA_QUERY(Query_for_list_of_tables);
+ /* Complete "CREATE PUBLICATION <name> FOR SEQUENCE" with "<sequence>, ..." */
+ else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "SEQUENCE"))
+ COMPLETE_WITH_SCHEMA_QUERY(Query_for_list_of_sequences);
/*
- * "CREATE PUBLICATION <name> FOR TABLE <name> WHERE (" - complete with
+ * "CREATE PUBLICATION <name> FOR TABLE|SEQUENCE <name> WHERE (" - complete with
* table attributes
*/
else if (HeadMatches("CREATE", "PUBLICATION", MatchAny) && TailMatches("WHERE"))
@@ -3044,13 +3057,13 @@ psql_completion(const char *text, int start, int end)
COMPLETE_WITH(" WITH (");
/*
- * Complete "CREATE PUBLICATION <name> FOR TABLES IN SCHEMA <schema>, ..."
+ * Complete "CREATE PUBLICATION <name> FOR TABLES|SEQUENCES IN SCHEMA <schema>, ..."
*/
- else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "TABLES", "IN", "SCHEMA"))
+ else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "TABLES|SEQUENCES", "IN", "SCHEMA"))
COMPLETE_WITH_QUERY_PLUS(Query_for_list_of_schemas
" AND nspname NOT LIKE E'pg\\\\_%%'",
"CURRENT_SCHEMA");
- else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "TABLES", "IN", "SCHEMA", MatchAny) && (!ends_with(prev_wd, ',')))
+ else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "TABLES|SEQUENCES", "IN", "SCHEMA", MatchAny) && (!ends_with(prev_wd, ',')))
COMPLETE_WITH("WITH (");
/* Complete "CREATE PUBLICATION <name> [...] WITH" */
else if (HeadMatches("CREATE", "PUBLICATION") && TailMatches("WITH", "("))
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 20f5aa56eab..a06e3c5eb2b 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -11621,6 +11621,11 @@
proallargtypes => '{text,oid,int2vector,pg_node_tree}',
proargmodes => '{i,o,o,o}', proargnames => '{pubname,relid,attrs,qual}',
prosrc => 'pg_get_publication_tables' },
+{ oid => '8000', descr => 'get OIDs of sequences in a publication',
+ proname => 'pg_get_publication_sequences', prorows => '1000', proretset => 't',
+ provolatile => 's', prorettype => 'oid', proargtypes => 'text',
+ proallargtypes => '{text,oid}', proargmodes => '{i,o}',
+ proargnames => '{pubname,relid}', prosrc => 'pg_get_publication_sequences' },
{ oid => '6121',
descr => 'returns whether a relation can be part of a publication',
proname => 'pg_relation_is_publishable', provolatile => 's',
diff --git a/src/include/catalog/pg_publication.h b/src/include/catalog/pg_publication.h
index ecf5a28e002..e3526367ed5 100644
--- a/src/include/catalog/pg_publication.h
+++ b/src/include/catalog/pg_publication.h
@@ -40,6 +40,12 @@ CATALOG(pg_publication,6104,PublicationRelationId)
*/
bool puballtables;
+ /*
+ * indicates that this is special publication which should encompass all
+ * sequences in the database (except for the unlogged and temp ones)
+ */
+ bool puballsequences;
+
/* true if inserts are published */
bool pubinsert;
@@ -52,6 +58,9 @@ CATALOG(pg_publication,6104,PublicationRelationId)
/* true if truncates are published */
bool pubtruncate;
+ /* true if sequences are published */
+ bool pubsequence;
+
/* true if partition changes are published using root schema */
bool pubviaroot;
} FormData_pg_publication;
@@ -72,6 +81,7 @@ typedef struct PublicationActions
bool pubupdate;
bool pubdelete;
bool pubtruncate;
+ bool pubsequence;
} PublicationActions;
typedef struct PublicationDesc
@@ -99,6 +109,7 @@ typedef struct Publication
Oid oid;
char *name;
bool alltables;
+ bool allsequences;
bool pubviaroot;
PublicationActions pubactions;
} Publication;
@@ -130,14 +141,16 @@ typedef enum PublicationPartOpt
PUBLICATION_PART_ALL,
} PublicationPartOpt;
-extern List *GetPublicationRelations(Oid pubid, PublicationPartOpt pub_partopt);
+extern List *GetPublicationRelations(Oid pubid, char objectType,
+ PublicationPartOpt pub_partopt);
extern List *GetAllTablesPublications(void);
extern List *GetAllTablesPublicationRelations(bool pubviaroot);
-extern List *GetPublicationSchemas(Oid pubid);
-extern List *GetSchemaPublications(Oid schemaid);
-extern List *GetSchemaPublicationRelations(Oid schemaid,
+extern void GetActionsInPublication(Oid pubid, PublicationActions *actions);
+extern List *GetPublicationSchemas(Oid pubid, char objectType);
+extern List *GetSchemaPublications(Oid schemaid, char objectType);
+extern List *GetSchemaPublicationRelations(Oid schemaid, char objectType,
PublicationPartOpt pub_partopt);
-extern List *GetAllSchemaPublicationRelations(Oid pubid,
+extern List *GetAllSchemaPublicationRelations(Oid puboid, char objectType,
PublicationPartOpt pub_partopt);
extern List *GetPubPartitionOptionRelations(List *result,
PublicationPartOpt pub_partopt,
@@ -145,11 +158,15 @@ extern List *GetPubPartitionOptionRelations(List *result,
extern Oid GetTopMostAncestorInPublication(Oid puboid, List *ancestors,
int *ancestor_level);
+extern List *GetAllSequencesPublications(void);
+extern List *GetAllSequencesPublicationRelations(void);
+
extern bool is_publishable_relation(Relation rel);
extern bool is_schema_publication(Oid pubid);
extern ObjectAddress publication_add_relation(Oid pubid, PublicationRelInfo *pri,
bool if_not_exists);
extern ObjectAddress publication_add_schema(Oid pubid, Oid schemaid,
+ char objectType,
bool if_not_exists);
extern Bitmapset *pub_collist_to_bitmapset(Bitmapset *columns, Datum pubcols,
diff --git a/src/include/catalog/pg_publication_namespace.h b/src/include/catalog/pg_publication_namespace.h
index d166600b260..914dde8b617 100644
--- a/src/include/catalog/pg_publication_namespace.h
+++ b/src/include/catalog/pg_publication_namespace.h
@@ -32,6 +32,7 @@ CATALOG(pg_publication_namespace,6237,PublicationNamespaceRelationId)
Oid oid; /* oid */
Oid pnpubid BKI_LOOKUP(pg_publication); /* Oid of the publication */
Oid pnnspid BKI_LOOKUP(pg_namespace); /* Oid of the schema */
+ char pntype; /* object type to include */
} FormData_pg_publication_namespace;
/* ----------------
@@ -42,6 +43,13 @@ CATALOG(pg_publication_namespace,6237,PublicationNamespaceRelationId)
typedef FormData_pg_publication_namespace *Form_pg_publication_namespace;
DECLARE_UNIQUE_INDEX_PKEY(pg_publication_namespace_oid_index, 6238, PublicationNamespaceObjectIndexId, on pg_publication_namespace using btree(oid oid_ops));
-DECLARE_UNIQUE_INDEX(pg_publication_namespace_pnnspid_pnpubid_index, 6239, PublicationNamespacePnnspidPnpubidIndexId, on pg_publication_namespace using btree(pnnspid oid_ops, pnpubid oid_ops));
+DECLARE_UNIQUE_INDEX(pg_publication_namespace_pnnspid_pnpubid_pntype_index, 8903, PublicationNamespacePnnspidPnpubidPntypeIndexId, on pg_publication_namespace using btree(pnnspid oid_ops, pnpubid oid_ops, pntype char_ops));
+
+/* object type to include from a schema, maps to relkind */
+#define PUB_OBJTYPE_TABLE 't' /* table (regular or partitioned) */
+#define PUB_OBJTYPE_SEQUENCE 's' /* sequence object */
+#define PUB_OBJTYPE_UNSUPPORTED 'u' /* used for non-replicated types */
+
+extern char pub_get_object_type_for_relkind(char relkind);
#endif /* PG_PUBLICATION_NAMESPACE_H */
diff --git a/src/include/commands/sequence.h b/src/include/commands/sequence.h
index b3b04ccfa95..6f5a2ec63d7 100644
--- a/src/include/commands/sequence.h
+++ b/src/include/commands/sequence.h
@@ -60,8 +60,11 @@ extern ObjectAddress AlterSequence(ParseState *pstate, AlterSeqStmt *stmt);
extern void SequenceChangePersistence(Oid relid, char newrelpersistence);
extern void DeleteSequenceTuple(Oid relid);
extern void ResetSequence(Oid seq_relid);
+extern void SetSequence(Oid seq_relid, int64 last_value, int64 log_cnt, bool is_called);
extern void ResetSequenceCaches(void);
+extern void AtEOXact_Sequences(bool isCommit);
+
extern void seq_redo(XLogReaderState *record);
extern void seq_desc(StringInfo buf, XLogReaderState *record);
extern const char *seq_identify(uint8 info);
diff --git a/src/include/nodes/parsenodes.h b/src/include/nodes/parsenodes.h
index 7caff62af7f..207c8ce84e1 100644
--- a/src/include/nodes/parsenodes.h
+++ b/src/include/nodes/parsenodes.h
@@ -3743,6 +3743,10 @@ typedef enum PublicationObjSpecType
PUBLICATIONOBJ_TABLES_IN_SCHEMA, /* All tables in schema */
PUBLICATIONOBJ_TABLES_IN_CUR_SCHEMA, /* All tables in first element of
* search_path */
+ PUBLICATIONOBJ_SEQUENCE, /* Sequence type */
+ PUBLICATIONOBJ_SEQUENCES_IN_SCHEMA, /* Sequences in schema type */
+ PUBLICATIONOBJ_SEQUENCES_IN_CUR_SCHEMA, /* Get the first element of
+ * search_path */
PUBLICATIONOBJ_CONTINUATION /* Continuation of previous type */
} PublicationObjSpecType;
@@ -3761,7 +3765,7 @@ typedef struct CreatePublicationStmt
char *pubname; /* Name of the publication */
List *options; /* List of DefElem nodes */
List *pubobjects; /* Optional list of publication objects */
- bool for_all_tables; /* Special publication for all tables in db */
+ List *for_all_objects; /* Special publication for all objects in db */
} CreatePublicationStmt;
typedef enum AlterPublicationAction
@@ -3784,7 +3788,7 @@ typedef struct AlterPublicationStmt
* objects.
*/
List *pubobjects; /* Optional list of publication objects */
- bool for_all_tables; /* Special publication for all tables in db */
+ List *for_all_objects; /* Special publication for all objects in db */
AlterPublicationAction action; /* What action to perform with the given
* objects */
} AlterPublicationStmt;
diff --git a/src/include/replication/decode.h b/src/include/replication/decode.h
index 741bf65cf7a..81df40d4974 100644
--- a/src/include/replication/decode.h
+++ b/src/include/replication/decode.h
@@ -27,6 +27,7 @@ extern void heap2_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
extern void xact_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
extern void standby_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
extern void logicalmsg_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
+extern void sequence_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
extern void LogicalDecodingProcessRecord(LogicalDecodingContext *ctx,
XLogReaderState *record);
diff --git a/src/include/replication/logicalproto.h b/src/include/replication/logicalproto.h
index 7eaa4c97eda..57c2aaa2efe 100644
--- a/src/include/replication/logicalproto.h
+++ b/src/include/replication/logicalproto.h
@@ -61,6 +61,7 @@ typedef enum LogicalRepMsgType
LOGICAL_REP_MSG_RELATION = 'R',
LOGICAL_REP_MSG_TYPE = 'Y',
LOGICAL_REP_MSG_MESSAGE = 'M',
+ LOGICAL_REP_MSG_SEQUENCE = 'Q',
LOGICAL_REP_MSG_BEGIN_PREPARE = 'b',
LOGICAL_REP_MSG_PREPARE = 'P',
LOGICAL_REP_MSG_COMMIT_PREPARED = 'K',
@@ -118,6 +119,17 @@ typedef struct LogicalRepTyp
char *typname; /* name of the remote type */
} LogicalRepTyp;
+/* Sequence info */
+typedef struct LogicalRepSequence
+{
+ Oid remoteid; /* unique id of the remote sequence */
+ char *nspname; /* schema name of remote sequence */
+ char *seqname; /* name of the remote sequence */
+ int64 last_value;
+ int64 log_cnt;
+ bool is_called;
+} LogicalRepSequence;
+
/* Transaction info */
typedef struct LogicalRepBeginData
{
@@ -230,6 +242,11 @@ extern List *logicalrep_read_truncate(StringInfo in,
bool *cascade, bool *restart_seqs);
extern void logicalrep_write_message(StringInfo out, TransactionId xid, XLogRecPtr lsn,
bool transactional, const char *prefix, Size sz, const char *message);
+extern void logicalrep_write_sequence(StringInfo out, Relation rel,
+ TransactionId xid, XLogRecPtr lsn,
+ int64 last_value, int64 log_cnt,
+ bool is_called);
+extern void logicalrep_read_sequence(StringInfo in, LogicalRepSequence *seqdata);
extern void logicalrep_write_rel(StringInfo out, TransactionId xid,
Relation rel, Bitmapset *columns);
extern LogicalRepRelation *logicalrep_read_rel(StringInfo in);
diff --git a/src/include/replication/message.h b/src/include/replication/message.h
index 0b396c56693..0bfaeb9336c 100644
--- a/src/include/replication/message.h
+++ b/src/include/replication/message.h
@@ -27,13 +27,27 @@ typedef struct xl_logical_message
char message[FLEXIBLE_ARRAY_MEMBER];
} xl_logical_message;
+/*
+ * Generic logical decoding sequence wal record.
+ */
+typedef struct xl_logical_sequence
+{
+ RelFileLocator locator;
+ Oid reloid;
+ int64 last; /* last value emitted for sequence */
+ int64 log_cnt; /* last value cached for sequence */
+ bool is_called;
+} xl_logical_sequence;
+
#define SizeOfLogicalMessage (offsetof(xl_logical_message, message))
+#define SizeOfLogicalSequence (sizeof(xl_logical_sequence))
extern XLogRecPtr LogLogicalMessage(const char *prefix, const char *message,
size_t size, bool transactional);
/* RMGR API */
#define XLOG_LOGICAL_MESSAGE 0x00
+#define XLOG_LOGICAL_SEQUENCE 0x10
extern void logicalmsg_redo(XLogReaderState *record);
extern void logicalmsg_desc(StringInfo buf, XLogReaderState *record);
extern const char *logicalmsg_identify(uint8 info);
diff --git a/src/include/replication/output_plugin.h b/src/include/replication/output_plugin.h
index b7d28d7045c..959afe3be98 100644
--- a/src/include/replication/output_plugin.h
+++ b/src/include/replication/output_plugin.h
@@ -90,6 +90,17 @@ typedef void (*LogicalDecodeMessageCB) (struct LogicalDecodingContext *ctx,
Size message_size,
const char *message);
+/*
+ * Called for the generic logical decoding sequences.
+ */
+typedef void (*LogicalDecodeSequenceCB) (struct LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ int64 last_value,
+ int64 log_cnt,
+ bool is_called);
+
/*
* Filter changes by origin.
*/
@@ -201,6 +212,18 @@ typedef void (*LogicalDecodeStreamMessageCB) (struct LogicalDecodingContext *ctx
Size message_size,
const char *message);
+/*
+ * Called for the streaming generic logical decoding sequences from in-progress
+ * transactions.
+ */
+typedef void (*LogicalDecodeStreamSequenceCB) (struct LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ int64 last_value,
+ int64 log_cnt,
+ bool is_called);
+
/*
* Callback for streaming truncates from in-progress transactions.
*/
@@ -221,6 +244,7 @@ typedef struct OutputPluginCallbacks
LogicalDecodeTruncateCB truncate_cb;
LogicalDecodeCommitCB commit_cb;
LogicalDecodeMessageCB message_cb;
+ LogicalDecodeSequenceCB sequence_cb;
LogicalDecodeFilterByOriginCB filter_by_origin_cb;
LogicalDecodeShutdownCB shutdown_cb;
@@ -239,6 +263,7 @@ typedef struct OutputPluginCallbacks
LogicalDecodeStreamCommitCB stream_commit_cb;
LogicalDecodeStreamChangeCB stream_change_cb;
LogicalDecodeStreamMessageCB stream_message_cb;
+ LogicalDecodeStreamSequenceCB stream_sequence_cb;
LogicalDecodeStreamTruncateCB stream_truncate_cb;
} OutputPluginCallbacks;
diff --git a/src/include/replication/pgoutput.h b/src/include/replication/pgoutput.h
index 02027550e25..45df2930b78 100644
--- a/src/include/replication/pgoutput.h
+++ b/src/include/replication/pgoutput.h
@@ -30,6 +30,7 @@ typedef struct PGOutputData
bool messages;
bool two_phase;
char *origin;
+ bool sequences;
} PGOutputData;
#endif /* PGOUTPUT_H */
diff --git a/src/include/replication/reorderbuffer.h b/src/include/replication/reorderbuffer.h
index b23d8cc4f9f..f1e2a697c34 100644
--- a/src/include/replication/reorderbuffer.h
+++ b/src/include/replication/reorderbuffer.h
@@ -64,7 +64,8 @@ typedef enum ReorderBufferChangeType
REORDER_BUFFER_CHANGE_INTERNAL_SPEC_INSERT,
REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM,
REORDER_BUFFER_CHANGE_INTERNAL_SPEC_ABORT,
- REORDER_BUFFER_CHANGE_TRUNCATE
+ REORDER_BUFFER_CHANGE_TRUNCATE,
+ REORDER_BUFFER_CHANGE_SEQUENCE
} ReorderBufferChangeType;
/* forward declaration */
@@ -158,6 +159,16 @@ typedef struct ReorderBufferChange
uint32 ninvalidations; /* Number of messages */
SharedInvalidationMessage *invalidations; /* invalidation message */
} inval;
+
+ /* Context data for Sequence changes */
+ struct
+ {
+ RelFileLocator locator;
+ Oid reloid;
+ int64 last;
+ int64 log_cnt;
+ bool is_called;
+ } sequence;
} data;
/*
@@ -435,6 +446,14 @@ typedef void (*ReorderBufferMessageCB) (ReorderBuffer *rb,
const char *prefix, Size sz,
const char *message);
+/* sequence callback signature */
+typedef void (*ReorderBufferSequenceCB) (ReorderBuffer *rb,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ int64 last_value, int64 log_cnt,
+ bool is_called);
+
/* begin prepare callback signature */
typedef void (*ReorderBufferBeginPrepareCB) (ReorderBuffer *rb,
ReorderBufferTXN *txn);
@@ -501,6 +520,14 @@ typedef void (*ReorderBufferStreamMessageCB) (
const char *prefix, Size sz,
const char *message);
+/* stream sequence callback signature */
+typedef void (*ReorderBufferStreamSequenceCB) (ReorderBuffer *rb,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ int64 last_value, int64 log_cnt,
+ bool is_called);
+
/* stream truncate callback signature */
typedef void (*ReorderBufferStreamTruncateCB) (
ReorderBuffer *rb,
@@ -516,6 +543,12 @@ struct ReorderBuffer
*/
HTAB *by_txn;
+ /*
+ * relfilenumber => XID lookup table for sequences created in a transaction
+ * (also includes altered sequences, which assigns new relfilenumber)
+ */
+ HTAB *sequences;
+
/*
* Transactions that could be a toplevel xact, ordered by LSN of the first
* record bearing that xid.
@@ -551,6 +584,7 @@ struct ReorderBuffer
ReorderBufferApplyTruncateCB apply_truncate;
ReorderBufferCommitCB commit;
ReorderBufferMessageCB message;
+ ReorderBufferSequenceCB sequence;
/*
* Callbacks to be called when streaming a transaction at prepare time.
@@ -570,6 +604,7 @@ struct ReorderBuffer
ReorderBufferStreamCommitCB stream_commit;
ReorderBufferStreamChangeCB stream_change;
ReorderBufferStreamMessageCB stream_message;
+ ReorderBufferStreamSequenceCB stream_sequence;
ReorderBufferStreamTruncateCB stream_truncate;
/*
@@ -649,6 +684,10 @@ extern void ReorderBufferQueueMessage(ReorderBuffer *rb, TransactionId xid,
Snapshot snap, XLogRecPtr lsn,
bool transactional, const char *prefix,
Size message_size, const char *message);
+extern void ReorderBufferQueueSequence(ReorderBuffer *rb, TransactionId xid,
+ XLogRecPtr lsn, RepOriginId origin_id,
+ Oid reloid, RelFileLocator locator,
+ int64 last, int64 log_cnt, bool is_called);
extern void ReorderBufferCommit(ReorderBuffer *rb, TransactionId xid,
XLogRecPtr commit_lsn, XLogRecPtr end_lsn,
TimestampTz commit_time, RepOriginId origin_id, XLogRecPtr origin_lsn);
diff --git a/src/test/regress/expected/object_address.out b/src/test/regress/expected/object_address.out
index dadd58e8b9c..046107fe137 100644
--- a/src/test/regress/expected/object_address.out
+++ b/src/test/regress/expected/object_address.out
@@ -16,6 +16,7 @@ CREATE TEXT SEARCH CONFIGURATION addr_ts_conf (copy=english);
CREATE TEXT SEARCH TEMPLATE addr_ts_temp (lexize=dsimple_lexize);
CREATE TEXT SEARCH PARSER addr_ts_prs
(start = prsd_start, gettoken = prsd_nexttoken, end = prsd_end, lextypes = prsd_lextype);
+CREATE SEQUENCE addr_nsp.gensequence;
CREATE TABLE addr_nsp.gentable (
a serial primary key CONSTRAINT a_chk CHECK (a > 0),
b text DEFAULT 'hello');
@@ -45,7 +46,9 @@ CREATE TRANSFORM FOR int LANGUAGE SQL (
-- suppress warning that depends on wal_level
SET client_min_messages = 'ERROR';
CREATE PUBLICATION addr_pub FOR TABLE addr_nsp.gentable;
+CREATE PUBLICATION addr_pub2 FOR SEQUENCE addr_nsp.gensequence;
CREATE PUBLICATION addr_pub_schema FOR TABLES IN SCHEMA addr_nsp;
+CREATE PUBLICATION addr_pub_schema2 FOR SEQUENCES IN SCHEMA addr_nsp;
RESET client_min_messages;
CREATE SUBSCRIPTION regress_addr_sub CONNECTION '' PUBLICATION bar WITH (connect = false, slot_name = NONE);
WARNING: subscription was created, but is not connected
@@ -314,8 +317,8 @@ WARNING: error for function of access method,{addr_nsp,zwei},{}: name list leng
WARNING: error for function of access method,{addr_nsp,zwei},{integer}: name list length must be at least 3
WARNING: error for function of access method,{eins,zwei,drei},{}: argument list length must be exactly 2
WARNING: error for function of access method,{eins,zwei,drei},{integer}: argument list length must be exactly 2
-WARNING: error for publication namespace,{eins},{}: argument list length must be exactly 1
-WARNING: error for publication namespace,{eins},{integer}: schema "eins" does not exist
+WARNING: error for publication namespace,{eins},{}: argument list length must be exactly 2
+WARNING: error for publication namespace,{eins},{integer}: argument list length must be exactly 2
WARNING: error for publication namespace,{addr_nsp,zwei},{}: name list length must be exactly 1
WARNING: error for publication namespace,{addr_nsp,zwei},{integer}: name list length must be exactly 1
WARNING: error for publication namespace,{eins,zwei,drei},{}: name list length must be exactly 1
@@ -435,7 +438,8 @@ WITH objects (type, name, args) AS (VALUES
('transform', '{int}', '{sql}'),
('access method', '{btree}', '{}'),
('publication', '{addr_pub}', '{}'),
- ('publication namespace', '{addr_nsp}', '{addr_pub_schema}'),
+ ('publication namespace', '{addr_nsp}', '{addr_pub_schema,t}'),
+ ('publication namespace', '{addr_nsp}', '{addr_pub_schema2,s}'),
('publication relation', '{addr_nsp, gentable}', '{addr_pub}'),
('subscription', '{regress_addr_sub}', '{}'),
('statistics object', '{addr_nsp, gentable_stat}', '{}')
@@ -499,8 +503,9 @@ SELECT (pg_identify_object(addr1.classid, addr1.objid, addr1.objsubid)).*,
subscription | | regress_addr_sub | regress_addr_sub | t
publication | | addr_pub | addr_pub | t
publication relation | | | addr_nsp.gentable in publication addr_pub | t
- publication namespace | | | addr_nsp in publication addr_pub_schema | t
-(50 rows)
+ publication namespace | | | addr_nsp in publication addr_pub_schema type t | t
+ publication namespace | | | addr_nsp in publication addr_pub_schema2 type s | t
+(51 rows)
---
--- Cleanup resources
@@ -512,14 +517,17 @@ drop cascades to foreign table genftable
drop cascades to server integer
drop cascades to user mapping for regress_addr_user on server integer
DROP PUBLICATION addr_pub;
+DROP PUBLICATION addr_pub2;
DROP PUBLICATION addr_pub_schema;
+DROP PUBLICATION addr_pub_schema2;
DROP SUBSCRIPTION regress_addr_sub;
DROP SCHEMA addr_nsp CASCADE;
-NOTICE: drop cascades to 14 other objects
+NOTICE: drop cascades to 15 other objects
DETAIL: drop cascades to text search dictionary addr_ts_dict
drop cascades to text search configuration addr_ts_conf
drop cascades to text search template addr_ts_temp
drop cascades to text search parser addr_ts_prs
+drop cascades to sequence gensequence
drop cascades to table gentable
drop cascades to table parttable
drop cascades to view genview
diff --git a/src/test/regress/expected/psql.out b/src/test/regress/expected/psql.out
index a7f5700edc1..ad891f90e34 100644
--- a/src/test/regress/expected/psql.out
+++ b/src/test/regress/expected/psql.out
@@ -5970,9 +5970,9 @@ List of schemas
(0 rows)
\dRp "no.such.publication"
- List of publications
- Name | Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
-------+-------+------------+---------+---------+---------+-----------+----------
+ List of publications
+ Name | Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+------+-------+------------+---------------+---------+---------+---------+-----------+-----------+----------
(0 rows)
\dRs "no.such.subscription"
diff --git a/src/test/regress/expected/publication.out b/src/test/regress/expected/publication.out
index 427f87ea077..cf70bfcda17 100644
--- a/src/test/regress/expected/publication.out
+++ b/src/test/regress/expected/publication.out
@@ -30,20 +30,20 @@ ERROR: conflicting or redundant options
LINE 1: ...ub_xxx WITH (publish_via_partition_root = 'true', publish_vi...
^
\dRp
- List of publications
- Name | Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------+--------------------------+------------+---------+---------+---------+-----------+----------
- testpib_ins_trunct | regress_publication_user | f | t | f | f | f | f
- testpub_default | regress_publication_user | f | f | t | f | f | f
+ List of publications
+ Name | Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ testpib_ins_trunct | regress_publication_user | f | f | t | f | f | f | f | f
+ testpub_default | regress_publication_user | f | f | f | t | f | f | f | f
(2 rows)
-ALTER PUBLICATION testpub_default SET (publish = 'insert, update, delete');
+ALTER PUBLICATION testpub_default SET (publish = 'insert, update, delete, sequence');
\dRp
- List of publications
- Name | Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------+--------------------------+------------+---------+---------+---------+-----------+----------
- testpib_ins_trunct | regress_publication_user | f | t | f | f | f | f
- testpub_default | regress_publication_user | f | t | t | t | f | f
+ List of publications
+ Name | Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ testpib_ins_trunct | regress_publication_user | f | f | t | f | f | f | f | f
+ testpub_default | regress_publication_user | f | f | t | t | t | f | t | f
(2 rows)
--- adding tables
@@ -61,6 +61,9 @@ CREATE TABLE testpub_tbl2 (id serial primary key, data text);
ALTER PUBLICATION testpub_foralltables ADD TABLE testpub_tbl2;
ERROR: publication "testpub_foralltables" is defined as FOR ALL TABLES
DETAIL: Tables cannot be added to or dropped from FOR ALL TABLES publications.
+-- fail - can't add a table using ADD SEQUENCE command
+ALTER PUBLICATION testpub_foralltables ADD SEQUENCE testpub_tbl2;
+ERROR: object type does not match type expected by command
-- fail - can't drop from all tables publication
ALTER PUBLICATION testpub_foralltables DROP TABLE testpub_tbl2;
ERROR: publication "testpub_foralltables" is defined as FOR ALL TABLES
@@ -87,10 +90,10 @@ RESET client_min_messages;
-- should be able to add schema to 'FOR TABLE' publication
ALTER PUBLICATION testpub_fortable ADD TABLES IN SCHEMA pub_test;
\dRp+ testpub_fortable
- Publication testpub_fortable
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_fortable
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"public.testpub_tbl1"
Tables from schemas:
@@ -99,20 +102,20 @@ Tables from schemas:
-- should be able to drop schema from 'FOR TABLE' publication
ALTER PUBLICATION testpub_fortable DROP TABLES IN SCHEMA pub_test;
\dRp+ testpub_fortable
- Publication testpub_fortable
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_fortable
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"public.testpub_tbl1"
-- should be able to set schema to 'FOR TABLE' publication
ALTER PUBLICATION testpub_fortable SET TABLES IN SCHEMA pub_test;
\dRp+ testpub_fortable
- Publication testpub_fortable
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_fortable
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test"
@@ -123,10 +126,10 @@ CREATE PUBLICATION testpub_forschema FOR TABLES IN SCHEMA pub_test;
CREATE PUBLICATION testpub_for_tbl_schema FOR TABLES IN SCHEMA pub_test, TABLE pub_test.testpub_nopk;
RESET client_min_messages;
\dRp+ testpub_for_tbl_schema
- Publication testpub_for_tbl_schema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_for_tbl_schema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"pub_test.testpub_nopk"
Tables from schemas:
@@ -135,10 +138,10 @@ Tables from schemas:
-- should be able to add a table of the same schema to the schema publication
ALTER PUBLICATION testpub_forschema ADD TABLE pub_test.testpub_nopk;
\dRp+ testpub_forschema
- Publication testpub_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"pub_test.testpub_nopk"
Tables from schemas:
@@ -147,10 +150,10 @@ Tables from schemas:
-- should be able to drop the table
ALTER PUBLICATION testpub_forschema DROP TABLE pub_test.testpub_nopk;
\dRp+ testpub_forschema
- Publication testpub_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test"
@@ -161,10 +164,10 @@ ERROR: relation "testpub_nopk" is not part of the publication
-- should be able to set table to schema publication
ALTER PUBLICATION testpub_forschema SET TABLE pub_test.testpub_nopk;
\dRp+ testpub_forschema
- Publication testpub_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"pub_test.testpub_nopk"
@@ -186,10 +189,10 @@ Publications:
"testpub_foralltables"
\dRp+ testpub_foralltables
- Publication testpub_foralltables
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | t | t | t | f | f | f
+ Publication testpub_foralltables
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | t | f | t | t | f | f | f | f
(1 row)
DROP TABLE testpub_tbl2;
@@ -201,24 +204,547 @@ CREATE PUBLICATION testpub3 FOR TABLE testpub_tbl3;
CREATE PUBLICATION testpub4 FOR TABLE ONLY testpub_tbl3;
RESET client_min_messages;
\dRp+ testpub3
- Publication testpub3
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub3
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"public.testpub_tbl3"
"public.testpub_tbl3a"
\dRp+ testpub4
- Publication testpub4
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub4
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"public.testpub_tbl3"
DROP TABLE testpub_tbl3, testpub_tbl3a;
DROP PUBLICATION testpub3, testpub4;
+--- adding sequences
+CREATE SEQUENCE testpub_seq0;
+CREATE SEQUENCE pub_test.testpub_seq1;
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_forallsequences FOR ALL SEQUENCES WITH (publish = 'sequence');
+RESET client_min_messages;
+ALTER PUBLICATION testpub_forallsequences SET (publish = 'insert, sequence');
+CREATE SEQUENCE testpub_seq2;
+-- fail - can't add to for all sequences publication
+ALTER PUBLICATION testpub_forallsequences ADD SEQUENCE testpub_seq2;
+ERROR: publication "testpub_forallsequences" is defined as FOR ALL SEQUENCES
+DETAIL: Sequences cannot be added to or dropped from FOR ALL SEQUENCES publications.
+-- fail - can't drop from all sequences publication
+ALTER PUBLICATION testpub_forallsequences DROP SEQUENCE testpub_seq2;
+ERROR: publication "testpub_forallsequences" is defined as FOR ALL SEQUENCES
+DETAIL: Sequences cannot be added to or dropped from FOR ALL SEQUENCES publications.
+-- fail - can't add to for all sequences publication
+ALTER PUBLICATION testpub_forallsequences SET SEQUENCE pub_test.testpub_seq1;
+ERROR: publication "testpub_forallsequences" is defined as FOR ALL SEQUENCES
+DETAIL: Sequences cannot be added to or dropped from FOR ALL SEQUENCES publications.
+-- fail - can't add schema to 'FOR ALL SEQUENCES' publication
+ALTER PUBLICATION testpub_forallsequences ADD SEQUENCES IN SCHEMA pub_test;
+ERROR: publication "testpub_forallsequences" is defined as FOR ALL SEQUENCES
+DETAIL: Schemas cannot be added to or dropped from FOR ALL SEQUENCES publications.
+-- fail - can't drop schema from 'FOR ALL SEQUENCES' publication
+ALTER PUBLICATION testpub_forallsequences DROP SEQUENCES IN SCHEMA pub_test;
+ERROR: publication "testpub_forallsequences" is defined as FOR ALL SEQUENCES
+DETAIL: Schemas cannot be added to or dropped from FOR ALL SEQUENCES publications.
+-- fail - can't set schema to 'FOR ALL SEQUENCES' publication
+ALTER PUBLICATION testpub_forallsequences SET SEQUENCES IN SCHEMA pub_test;
+ERROR: publication "testpub_forallsequences" is defined as FOR ALL SEQUENCES
+DETAIL: Schemas cannot be added to or dropped from FOR ALL SEQUENCES publications.
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_forsequence FOR SEQUENCE testpub_seq0;
+RESET client_min_messages;
+-- should be able to add schema to 'FOR SEQUENCE' publication
+ALTER PUBLICATION testpub_forsequence ADD SEQUENCES IN SCHEMA pub_test;
+\dRp+ testpub_forsequence
+ Publication testpub_forsequence
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Sequences:
+ "public.testpub_seq0"
+Sequences from schemas:
+ "pub_test"
+
+-- add sequence from the schema we already added
+ALTER PUBLICATION testpub_forsequence ADD SEQUENCE pub_test.testpub_seq1;
+-- fail - can't add sequence using ADD TABLE command
+ALTER PUBLICATION testpub_forsequence ADD TABLE pub_test.testpub_seq1;
+ERROR: object type does not match type expected by command
+-- should be able to drop schema from 'FOR SEQUENCE' publication
+ALTER PUBLICATION testpub_forsequence DROP SEQUENCES IN SCHEMA pub_test;
+\dRp+ testpub_forsequence
+ Publication testpub_forsequence
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Sequences:
+ "pub_test.testpub_seq1"
+ "public.testpub_seq0"
+
+-- should be able to set schema to 'FOR SEQUENCE' publication
+ALTER PUBLICATION testpub_forsequence SET SEQUENCES IN SCHEMA pub_test;
+\dRp+ testpub_forsequence
+ Publication testpub_forsequence
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Sequences from schemas:
+ "pub_test"
+
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_forschema FOR SEQUENCES IN SCHEMA pub_test;
+RESET client_min_messages;
+-- create publication with schema and sequence of the same schema
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_for_seq_schema FOR SEQUENCES IN SCHEMA pub_test, SEQUENCE pub_test.testpub_seq1;
+RESET client_min_messages;
+-- drop the publication again
+DROP PUBLICATION testpub_for_seq_schema;
+-- add a sequence of the same schema to the schema publication
+ALTER PUBLICATION testpub_forschema ADD SEQUENCE pub_test.testpub_seq1;
+-- drop a sequence from the publication again
+ALTER PUBLICATION testpub_forschema DROP SEQUENCE pub_test.testpub_seq1;
+-- drop a sequence from the schema publication which isn't in the publication
+ALTER PUBLICATION testpub_forschema DROP SEQUENCE pub_test.testpub_seq1;
+ERROR: relation "testpub_seq1" is not part of the publication
+-- should be able to set sequence to schema publication
+ALTER PUBLICATION testpub_forschema SET SEQUENCE pub_test.testpub_seq1;
+\dRp+ testpub_forschema
+ Publication testpub_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Sequences:
+ "pub_test.testpub_seq1"
+
+SELECT pubname, puballtables, puballsequences FROM pg_publication WHERE pubname = 'testpub_forallsequences';
+ pubname | puballtables | puballsequences
+-------------------------+--------------+-----------------
+ testpub_forallsequences | f | t
+(1 row)
+
+\d+ pub_test.testpub_seq1
+ Sequence "pub_test.testpub_seq1"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+Publications:
+ "testpub_forallsequences"
+ "testpub_forschema"
+ "testpub_forsequence"
+
+\dRp+ testpub_forallsequences
+ Publication testpub_forallsequences
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | t | t | f | f | f | t | f
+(1 row)
+
+DROP SEQUENCE testpub_seq0, pub_test.testpub_seq1, testpub_seq2;
+DROP PUBLICATION testpub_forallsequences, testpub_forsequence, testpub_forschema;
+-- publication testing multiple sequences at the same time
+CREATE SEQUENCE testpub_seq1;
+CREATE SEQUENCE testpub_seq2;
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_multi FOR SEQUENCE testpub_seq1, testpub_seq2;
+RESET client_min_messages;
+\dRp+ testpub_multi
+ Publication testpub_multi
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Sequences:
+ "public.testpub_seq1"
+ "public.testpub_seq2"
+
+DROP PUBLICATION testpub_multi;
+DROP SEQUENCE testpub_seq1;
+DROP SEQUENCE testpub_seq2;
+-- Publication mixing tables and sequences
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_mix;
+RESET client_min_messages;
+CREATE SEQUENCE testpub_seq1;
+CREATE SEQUENCE pub_test.testpub_seq2;
+ALTER PUBLICATION testpub_mix ADD SEQUENCE testpub_seq1, TABLE testpub_tbl1;
+\dRp+ testpub_mix
+ Publication testpub_mix
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Tables:
+ "public.testpub_tbl1"
+Sequences:
+ "public.testpub_seq1"
+
+ALTER PUBLICATION testpub_mix ADD SEQUENCES IN SCHEMA pub_test, TABLES IN SCHEMA pub_test;
+\dRp+ testpub_mix
+ Publication testpub_mix
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Tables:
+ "public.testpub_tbl1"
+Tables from schemas:
+ "pub_test"
+Sequences:
+ "public.testpub_seq1"
+Sequences from schemas:
+ "pub_test"
+
+ALTER PUBLICATION testpub_mix DROP SEQUENCES IN SCHEMA pub_test;
+\dRp+ testpub_mix
+ Publication testpub_mix
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Tables:
+ "public.testpub_tbl1"
+Tables from schemas:
+ "pub_test"
+Sequences:
+ "public.testpub_seq1"
+
+ALTER PUBLICATION testpub_mix DROP TABLES IN SCHEMA pub_test;
+\dRp+ testpub_mix
+ Publication testpub_mix
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Tables:
+ "public.testpub_tbl1"
+Sequences:
+ "public.testpub_seq1"
+
+DROP PUBLICATION testpub_mix;
+DROP SEQUENCE testpub_seq1;
+DROP SEQUENCE pub_test.testpub_seq2;
+-- make sure we replicate only the correct relation type
+CREATE SCHEMA pub_test1;
+CREATE SEQUENCE pub_test1.test_seq1;
+CREATE TABLE pub_test1.test_tbl1 (a int primary key, b int);
+CREATE SCHEMA pub_test2;
+CREATE SEQUENCE pub_test2.test_seq2;
+CREATE TABLE pub_test2.test_tbl2 (a int primary key, b int);
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_schemas;
+RESET client_min_messages;
+-- add tables from one schema, sequences from the other
+ALTER PUBLICATION testpub_schemas ADD TABLES IN SCHEMA pub_test2;
+ALTER PUBLICATION testpub_schemas ADD SEQUENCES IN SCHEMA pub_test1;
+\dRp+ testpub_schemas
+ Publication testpub_schemas
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Tables from schemas:
+ "pub_test2"
+Sequences from schemas:
+ "pub_test1"
+
+\dn+ pub_test1
+ List of schemas
+ Name | Owner | Access privileges | Description
+-----------+--------------------------+-------------------+-------------
+ pub_test1 | regress_publication_user | |
+Publications:
+ "testpub_schemas" (sequences)
+
+\dn+ pub_test2
+ List of schemas
+ Name | Owner | Access privileges | Description
+-----------+--------------------------+-------------------+-------------
+ pub_test2 | regress_publication_user | |
+Publications:
+ "testpub_schemas" (tables)
+
+\d+ pub_test1.test_seq1;
+ Sequence "pub_test1.test_seq1"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test1.test_tbl1;
+ Table "pub_test1.test_tbl1"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl1_pkey" PRIMARY KEY, btree (a)
+
+\d+ pub_test2.test_seq2;
+ Sequence "pub_test2.test_seq2"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+
+\d+ pub_test2.test_tbl2;
+ Table "pub_test2.test_tbl2"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl2_pkey" PRIMARY KEY, btree (a)
+Publications:
+ "testpub_schemas"
+
+-- add the other object type from each schema
+ALTER PUBLICATION testpub_schemas ADD TABLES IN SCHEMA pub_test1;
+ALTER PUBLICATION testpub_schemas ADD SEQUENCES IN SCHEMA pub_test2;
+\dRp+ testpub_schemas
+ Publication testpub_schemas
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Tables from schemas:
+ "pub_test1"
+ "pub_test2"
+Sequences from schemas:
+ "pub_test1"
+ "pub_test2"
+
+\dn+ pub_test1
+ List of schemas
+ Name | Owner | Access privileges | Description
+-----------+--------------------------+-------------------+-------------
+ pub_test1 | regress_publication_user | |
+Publications:
+ "testpub_schemas" (sequences)
+ "testpub_schemas" (tables)
+
+\dn+ pub_test2
+ List of schemas
+ Name | Owner | Access privileges | Description
+-----------+--------------------------+-------------------+-------------
+ pub_test2 | regress_publication_user | |
+Publications:
+ "testpub_schemas" (tables)
+ "testpub_schemas" (sequences)
+
+\d+ pub_test1.test_seq1;
+ Sequence "pub_test1.test_seq1"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test1.test_tbl1;
+ Table "pub_test1.test_tbl1"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl1_pkey" PRIMARY KEY, btree (a)
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test2.test_seq2;
+ Sequence "pub_test2.test_seq2"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test2.test_tbl2;
+ Table "pub_test2.test_tbl2"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl2_pkey" PRIMARY KEY, btree (a)
+Publications:
+ "testpub_schemas"
+
+-- now drop the object type added first
+ALTER PUBLICATION testpub_schemas DROP TABLES IN SCHEMA pub_test2;
+ALTER PUBLICATION testpub_schemas DROP SEQUENCES IN SCHEMA pub_test1;
+\dRp+ testpub_schemas
+ Publication testpub_schemas
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Tables from schemas:
+ "pub_test1"
+Sequences from schemas:
+ "pub_test2"
+
+\dn+ pub_test1
+ List of schemas
+ Name | Owner | Access privileges | Description
+-----------+--------------------------+-------------------+-------------
+ pub_test1 | regress_publication_user | |
+Publications:
+ "testpub_schemas" (tables)
+
+\dn+ pub_test2
+ List of schemas
+ Name | Owner | Access privileges | Description
+-----------+--------------------------+-------------------+-------------
+ pub_test2 | regress_publication_user | |
+Publications:
+ "testpub_schemas" (sequences)
+
+\d+ pub_test1.test_seq1;
+ Sequence "pub_test1.test_seq1"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+
+\d+ pub_test1.test_tbl1;
+ Table "pub_test1.test_tbl1"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl1_pkey" PRIMARY KEY, btree (a)
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test2.test_seq2;
+ Sequence "pub_test2.test_seq2"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test2.test_tbl2;
+ Table "pub_test2.test_tbl2"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl2_pkey" PRIMARY KEY, btree (a)
+
+-- should work, we're not enforcing there's no overlap
+ALTER PUBLICATION testpub_schemas ADD TABLE pub_test1.test_tbl1;
+ALTER PUBLICATION testpub_schemas ADD SEQUENCE pub_test2.test_seq2;
+-- should work (different schema)
+ALTER PUBLICATION testpub_schemas ADD TABLE pub_test2.test_tbl2;
+ALTER PUBLICATION testpub_schemas ADD SEQUENCE pub_test1.test_seq1;
+\dRp+ testpub_schemas
+ Publication testpub_schemas
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Tables:
+ "pub_test1.test_tbl1"
+ "pub_test2.test_tbl2"
+Tables from schemas:
+ "pub_test1"
+Sequences:
+ "pub_test1.test_seq1"
+ "pub_test2.test_seq2"
+Sequences from schemas:
+ "pub_test2"
+
+\d+ pub_test1.test_seq1;
+ Sequence "pub_test1.test_seq1"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test1.test_tbl1;
+ Table "pub_test1.test_tbl1"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl1_pkey" PRIMARY KEY, btree (a)
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test2.test_seq2;
+ Sequence "pub_test2.test_seq2"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test2.test_tbl2;
+ Table "pub_test2.test_tbl2"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl2_pkey" PRIMARY KEY, btree (a)
+Publications:
+ "testpub_schemas"
+
+-- now drop the explicitly added objects again
+ALTER PUBLICATION testpub_schemas DROP TABLE pub_test2.test_tbl2;
+ALTER PUBLICATION testpub_schemas DROP SEQUENCE pub_test1.test_seq1;
+\dRp+ testpub_schemas
+ Publication testpub_schemas
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Tables:
+ "pub_test1.test_tbl1"
+Tables from schemas:
+ "pub_test1"
+Sequences:
+ "pub_test2.test_seq2"
+Sequences from schemas:
+ "pub_test2"
+
+\d+ pub_test1.test_seq1;
+ Sequence "pub_test1.test_seq1"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+
+\d+ pub_test1.test_tbl1;
+ Table "pub_test1.test_tbl1"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl1_pkey" PRIMARY KEY, btree (a)
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test2.test_seq2;
+ Sequence "pub_test2.test_seq2"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test2.test_tbl2;
+ Table "pub_test2.test_tbl2"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl2_pkey" PRIMARY KEY, btree (a)
+
+DROP PUBLICATION testpub_schemas;
+DROP TABLE pub_test1.test_tbl1, pub_test2.test_tbl2;
+DROP SEQUENCE pub_test1.test_seq1, pub_test2.test_seq2;
+DROP SCHEMA pub_test1, pub_test2;
-- Tests for partitioned tables
SET client_min_messages = 'ERROR';
CREATE PUBLICATION testpub_forparted;
@@ -234,10 +760,10 @@ UPDATE testpub_parted1 SET a = 1;
-- only parent is listed as being in publication, not the partition
ALTER PUBLICATION testpub_forparted ADD TABLE testpub_parted;
\dRp+ testpub_forparted
- Publication testpub_forparted
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_forparted
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"public.testpub_parted"
@@ -252,10 +778,10 @@ ALTER TABLE testpub_parted DETACH PARTITION testpub_parted1;
UPDATE testpub_parted1 SET a = 1;
ALTER PUBLICATION testpub_forparted SET (publish_via_partition_root = true);
\dRp+ testpub_forparted
- Publication testpub_forparted
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | t
+ Publication testpub_forparted
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | t
Tables:
"public.testpub_parted"
@@ -284,10 +810,10 @@ SET client_min_messages = 'ERROR';
CREATE PUBLICATION testpub5 FOR TABLE testpub_rf_tbl1, testpub_rf_tbl2 WHERE (c <> 'test' AND d < 5) WITH (publish = 'insert');
RESET client_min_messages;
\dRp+ testpub5
- Publication testpub5
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | f | f | f | f
+ Publication testpub5
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | f | f | f | f | f
Tables:
"public.testpub_rf_tbl1"
"public.testpub_rf_tbl2" WHERE ((c <> 'test'::text) AND (d < 5))
@@ -300,10 +826,10 @@ Tables:
ALTER PUBLICATION testpub5 ADD TABLE testpub_rf_tbl3 WHERE (e > 1000 AND e < 2000);
\dRp+ testpub5
- Publication testpub5
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | f | f | f | f
+ Publication testpub5
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | f | f | f | f | f
Tables:
"public.testpub_rf_tbl1"
"public.testpub_rf_tbl2" WHERE ((c <> 'test'::text) AND (d < 5))
@@ -319,10 +845,10 @@ Publications:
ALTER PUBLICATION testpub5 DROP TABLE testpub_rf_tbl2;
\dRp+ testpub5
- Publication testpub5
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | f | f | f | f
+ Publication testpub5
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | f | f | f | f | f
Tables:
"public.testpub_rf_tbl1"
"public.testpub_rf_tbl3" WHERE ((e > 1000) AND (e < 2000))
@@ -330,10 +856,10 @@ Tables:
-- remove testpub_rf_tbl1 and add testpub_rf_tbl3 again (another WHERE expression)
ALTER PUBLICATION testpub5 SET TABLE testpub_rf_tbl3 WHERE (e > 300 AND e < 500);
\dRp+ testpub5
- Publication testpub5
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | f | f | f | f
+ Publication testpub5
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | f | f | f | f | f
Tables:
"public.testpub_rf_tbl3" WHERE ((e > 300) AND (e < 500))
@@ -366,10 +892,10 @@ SET client_min_messages = 'ERROR';
CREATE PUBLICATION testpub_syntax1 FOR TABLE testpub_rf_tbl1, ONLY testpub_rf_tbl3 WHERE (e < 999) WITH (publish = 'insert');
RESET client_min_messages;
\dRp+ testpub_syntax1
- Publication testpub_syntax1
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | f | f | f | f
+ Publication testpub_syntax1
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | f | f | f | f | f
Tables:
"public.testpub_rf_tbl1"
"public.testpub_rf_tbl3" WHERE (e < 999)
@@ -379,10 +905,10 @@ SET client_min_messages = 'ERROR';
CREATE PUBLICATION testpub_syntax2 FOR TABLE testpub_rf_tbl1, testpub_rf_schema1.testpub_rf_tbl5 WHERE (h < 999) WITH (publish = 'insert');
RESET client_min_messages;
\dRp+ testpub_syntax2
- Publication testpub_syntax2
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | f | f | f | f
+ Publication testpub_syntax2
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | f | f | f | f | f
Tables:
"public.testpub_rf_tbl1"
"testpub_rf_schema1.testpub_rf_tbl5" WHERE (h < 999)
@@ -497,10 +1023,10 @@ CREATE PUBLICATION testpub6 FOR TABLES IN SCHEMA testpub_rf_schema2;
ALTER PUBLICATION testpub6 SET TABLES IN SCHEMA testpub_rf_schema2, TABLE testpub_rf_schema2.testpub_rf_tbl6 WHERE (i < 99);
RESET client_min_messages;
\dRp+ testpub6
- Publication testpub6
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub6
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"testpub_rf_schema2.testpub_rf_tbl6" WHERE (i < 99)
Tables from schemas:
@@ -714,10 +1240,10 @@ CREATE PUBLICATION testpub_table_ins WITH (publish = 'insert, truncate');
RESET client_min_messages;
ALTER PUBLICATION testpub_table_ins ADD TABLE testpub_tbl5 (a); -- ok
\dRp+ testpub_table_ins
- Publication testpub_table_ins
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | f | f | t | f
+ Publication testpub_table_ins
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | f | f | t | f | f
Tables:
"public.testpub_tbl5" (a)
@@ -891,10 +1417,10 @@ CREATE TABLE testpub_tbl_both_filters (a int, b int, c int, PRIMARY KEY (a,c));
ALTER TABLE testpub_tbl_both_filters REPLICA IDENTITY USING INDEX testpub_tbl_both_filters_pkey;
ALTER PUBLICATION testpub_both_filters ADD TABLE testpub_tbl_both_filters (a,c) WHERE (c != 1);
\dRp+ testpub_both_filters
- Publication testpub_both_filters
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_both_filters
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"public.testpub_tbl_both_filters" (a, c) WHERE (c <> 1)
@@ -1099,10 +1625,10 @@ ERROR: relation "testpub_tbl1" is already member of publication "testpub_fortbl
CREATE PUBLICATION testpub_fortbl FOR TABLE testpub_tbl1;
ERROR: publication "testpub_fortbl" already exists
\dRp+ testpub_fortbl
- Publication testpub_fortbl
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_fortbl
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"pub_test.testpub_nopk"
"public.testpub_tbl1"
@@ -1140,10 +1666,10 @@ Publications:
"testpub_fortbl"
\dRp+ testpub_default
- Publication testpub_default
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | f | f
+ Publication testpub_default
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | f | t | f
Tables:
"pub_test.testpub_nopk"
"public.testpub_tbl1"
@@ -1221,10 +1747,10 @@ REVOKE CREATE ON DATABASE regression FROM regress_publication_user2;
DROP TABLE testpub_parted;
DROP TABLE testpub_tbl1;
\dRp+ testpub_default
- Publication testpub_default
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | f | f
+ Publication testpub_default
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | f | t | f
(1 row)
-- fail - must be owner of publication
@@ -1234,20 +1760,20 @@ ERROR: must be owner of publication testpub_default
RESET ROLE;
ALTER PUBLICATION testpub_default RENAME TO testpub_foo;
\dRp testpub_foo
- List of publications
- Name | Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
--------------+--------------------------+------------+---------+---------+---------+-----------+----------
- testpub_foo | regress_publication_user | f | t | t | t | f | f
+ List of publications
+ Name | Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+-------------+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ testpub_foo | regress_publication_user | f | f | t | t | t | f | t | f
(1 row)
-- rename back to keep the rest simple
ALTER PUBLICATION testpub_foo RENAME TO testpub_default;
ALTER PUBLICATION testpub_default OWNER TO regress_publication_user2;
\dRp testpub_default
- List of publications
- Name | Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
------------------+---------------------------+------------+---------+---------+---------+-----------+----------
- testpub_default | regress_publication_user2 | f | t | t | t | f | f
+ List of publications
+ Name | Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+-----------------+---------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ testpub_default | regress_publication_user2 | f | f | t | t | t | f | t | f
(1 row)
-- adding schemas and tables
@@ -1263,19 +1789,19 @@ CREATE TABLE "CURRENT_SCHEMA"."CURRENT_SCHEMA"(id int);
SET client_min_messages = 'ERROR';
CREATE PUBLICATION testpub1_forschema FOR TABLES IN SCHEMA pub_test1;
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
CREATE PUBLICATION testpub2_forschema FOR TABLES IN SCHEMA pub_test1, pub_test2, pub_test3;
\dRp+ testpub2_forschema
- Publication testpub2_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub2_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
"pub_test2"
@@ -1289,44 +1815,44 @@ CREATE PUBLICATION testpub6_forschema FOR TABLES IN SCHEMA "CURRENT_SCHEMA", CUR
CREATE PUBLICATION testpub_fortable FOR TABLE "CURRENT_SCHEMA"."CURRENT_SCHEMA";
RESET client_min_messages;
\dRp+ testpub3_forschema
- Publication testpub3_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub3_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"public"
\dRp+ testpub4_forschema
- Publication testpub4_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub4_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"CURRENT_SCHEMA"
\dRp+ testpub5_forschema
- Publication testpub5_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub5_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"CURRENT_SCHEMA"
"public"
\dRp+ testpub6_forschema
- Publication testpub6_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub6_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"CURRENT_SCHEMA"
"public"
\dRp+ testpub_fortable
- Publication testpub_fortable
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_fortable
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"CURRENT_SCHEMA.CURRENT_SCHEMA"
@@ -1341,7 +1867,7 @@ CREATE PUBLICATION testpub_forschema1 FOR CURRENT_SCHEMA;
ERROR: invalid publication object list
LINE 1: CREATE PUBLICATION testpub_forschema1 FOR CURRENT_SCHEMA;
^
-DETAIL: One of TABLE or TABLES IN SCHEMA must be specified before a standalone table or schema name.
+DETAIL: One of TABLE, SEQUENCE, TABLES IN SCHEMA or SEQUENCES IN SCHEMA must be specified before a standalone table, sequence or schema name.
-- check create publication on CURRENT_SCHEMA along with FOR TABLE
CREATE PUBLICATION testpub_forschema1 FOR TABLE CURRENT_SCHEMA;
ERROR: syntax error at or near "CURRENT_SCHEMA"
@@ -1360,10 +1886,10 @@ ERROR: schema "testpub_view" does not exist
-- dropping the schema should reflect the change in publication
DROP SCHEMA pub_test3;
\dRp+ testpub2_forschema
- Publication testpub2_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub2_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
"pub_test2"
@@ -1371,20 +1897,20 @@ Tables from schemas:
-- renaming the schema should reflect the change in publication
ALTER SCHEMA pub_test1 RENAME to pub_test1_renamed;
\dRp+ testpub2_forschema
- Publication testpub2_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub2_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1_renamed"
"pub_test2"
ALTER SCHEMA pub_test1_renamed RENAME to pub_test1;
\dRp+ testpub2_forschema
- Publication testpub2_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub2_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
"pub_test2"
@@ -1392,10 +1918,10 @@ Tables from schemas:
-- alter publication add schema
ALTER PUBLICATION testpub1_forschema ADD TABLES IN SCHEMA pub_test2;
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
"pub_test2"
@@ -1404,10 +1930,10 @@ Tables from schemas:
ALTER PUBLICATION testpub1_forschema ADD TABLES IN SCHEMA non_existent_schema;
ERROR: schema "non_existent_schema" does not exist
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
"pub_test2"
@@ -1416,10 +1942,10 @@ Tables from schemas:
ALTER PUBLICATION testpub1_forschema ADD TABLES IN SCHEMA pub_test1;
ERROR: schema "pub_test1" is already member of publication "testpub1_forschema"
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
"pub_test2"
@@ -1427,10 +1953,10 @@ Tables from schemas:
-- alter publication drop schema
ALTER PUBLICATION testpub1_forschema DROP TABLES IN SCHEMA pub_test2;
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
@@ -1438,10 +1964,10 @@ Tables from schemas:
ALTER PUBLICATION testpub1_forschema DROP TABLES IN SCHEMA pub_test2;
ERROR: tables from schema "pub_test2" are not part of the publication
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
@@ -1449,29 +1975,29 @@ Tables from schemas:
ALTER PUBLICATION testpub1_forschema DROP TABLES IN SCHEMA non_existent_schema;
ERROR: schema "non_existent_schema" does not exist
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
-- drop all schemas
ALTER PUBLICATION testpub1_forschema DROP TABLES IN SCHEMA pub_test1;
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
(1 row)
-- alter publication set multiple schema
ALTER PUBLICATION testpub1_forschema SET TABLES IN SCHEMA pub_test1, pub_test2;
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
"pub_test2"
@@ -1480,10 +2006,10 @@ Tables from schemas:
ALTER PUBLICATION testpub1_forschema SET TABLES IN SCHEMA non_existent_schema;
ERROR: schema "non_existent_schema" does not exist
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
"pub_test2"
@@ -1492,10 +2018,10 @@ Tables from schemas:
-- removing the duplicate schemas
ALTER PUBLICATION testpub1_forschema SET TABLES IN SCHEMA pub_test1, pub_test1;
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
@@ -1574,18 +2100,18 @@ SET client_min_messages = 'ERROR';
CREATE PUBLICATION testpub3_forschema;
RESET client_min_messages;
\dRp+ testpub3_forschema
- Publication testpub3_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub3_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
(1 row)
ALTER PUBLICATION testpub3_forschema SET TABLES IN SCHEMA pub_test1;
\dRp+ testpub3_forschema
- Publication testpub3_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub3_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
@@ -1595,20 +2121,20 @@ CREATE PUBLICATION testpub_forschema_fortable FOR TABLES IN SCHEMA pub_test1, TA
CREATE PUBLICATION testpub_fortable_forschema FOR TABLE pub_test2.tbl1, TABLES IN SCHEMA pub_test1;
RESET client_min_messages;
\dRp+ testpub_forschema_fortable
- Publication testpub_forschema_fortable
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_forschema_fortable
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"pub_test2.tbl1"
Tables from schemas:
"pub_test1"
\dRp+ testpub_fortable_forschema
- Publication testpub_fortable_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_fortable_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"pub_test2.tbl1"
Tables from schemas:
@@ -1620,7 +2146,7 @@ CREATE PUBLICATION testpub_error FOR pub_test2.tbl1;
ERROR: invalid publication object list
LINE 1: CREATE PUBLICATION testpub_error FOR pub_test2.tbl1;
^
-DETAIL: One of TABLE or TABLES IN SCHEMA must be specified before a standalone table or schema name.
+DETAIL: One of TABLE, SEQUENCE, TABLES IN SCHEMA or SEQUENCES IN SCHEMA must be specified before a standalone table, sequence or schema name.
DROP VIEW testpub_view;
DROP PUBLICATION testpub_default;
DROP PUBLICATION testpib_ins_trunct;
@@ -1652,40 +2178,85 @@ CREATE SCHEMA sch1;
CREATE SCHEMA sch2;
CREATE TABLE sch1.tbl1 (a int) PARTITION BY RANGE(a);
CREATE TABLE sch2.tbl1_part1 PARTITION OF sch1.tbl1 FOR VALUES FROM (1) to (10);
+CREATE SEQUENCE sch1.seq1;
+CREATE SEQUENCE sch2.seq2;
-- Schema publication that does not include the schema that has the parent table
CREATE PUBLICATION pub FOR TABLES IN SCHEMA sch2 WITH (PUBLISH_VIA_PARTITION_ROOT=1);
+ALTER PUBLICATION pub ADD SEQUENCES IN SCHEMA sch2;
SELECT * FROM pg_publication_tables;
pubname | schemaname | tablename | attnames | rowfilter
---------+------------+------------+----------+-----------
pub | sch2 | tbl1_part1 | {a} |
(1 row)
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch2 | seq2
+(1 row)
+
DROP PUBLICATION pub;
-- Table publication that does not include the parent table
CREATE PUBLICATION pub FOR TABLE sch2.tbl1_part1 WITH (PUBLISH_VIA_PARTITION_ROOT=1);
+ALTER PUBLICATION pub ADD SEQUENCE sch2.seq2;
SELECT * FROM pg_publication_tables;
pubname | schemaname | tablename | attnames | rowfilter
---------+------------+------------+----------+-----------
pub | sch2 | tbl1_part1 | {a} |
(1 row)
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch2 | seq2
+(1 row)
+
-- Table publication that includes both the parent table and the child table
ALTER PUBLICATION pub ADD TABLE sch1.tbl1;
+ALTER PUBLICATION pub ADD SEQUENCE sch1.seq1;
SELECT * FROM pg_publication_tables;
pubname | schemaname | tablename | attnames | rowfilter
---------+------------+-----------+----------+-----------
pub | sch1 | tbl1 | {a} |
(1 row)
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch1 | seq1
+ pub | sch2 | seq2
+(2 rows)
+
DROP PUBLICATION pub;
-- Schema publication that does not include the schema that has the parent table
CREATE PUBLICATION pub FOR TABLES IN SCHEMA sch2 WITH (PUBLISH_VIA_PARTITION_ROOT=0);
+ALTER PUBLICATION pub ADD SEQUENCE sch1.seq1;
SELECT * FROM pg_publication_tables;
pubname | schemaname | tablename | attnames | rowfilter
---------+------------+------------+----------+-----------
pub | sch2 | tbl1_part1 | {a} |
(1 row)
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch1 | seq1
+(1 row)
+
+DROP PUBLICATION pub;
+-- Sequence publication
+CREATE PUBLICATION pub FOR SEQUENCE sch2.seq2;
+SELECT * FROM pg_publication_tables;
+ pubname | schemaname | tablename | attnames | rowfilter
+---------+------------+-----------+----------+-----------
+(0 rows)
+
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch2 | seq2
+(1 row)
+
DROP PUBLICATION pub;
-- Table publication that does not include the parent table
CREATE PUBLICATION pub FOR TABLE sch2.tbl1_part1 WITH (PUBLISH_VIA_PARTITION_ROOT=0);
@@ -1695,14 +2266,26 @@ SELECT * FROM pg_publication_tables;
pub | sch2 | tbl1_part1 | {a} |
(1 row)
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+(0 rows)
+
-- Table publication that includes both the parent table and the child table
ALTER PUBLICATION pub ADD TABLE sch1.tbl1;
+ALTER PUBLICATION pub ADD SEQUENCES IN SCHEMA sch2;
SELECT * FROM pg_publication_tables;
pubname | schemaname | tablename | attnames | rowfilter
---------+------------+------------+----------+-----------
pub | sch2 | tbl1_part1 | {a} |
(1 row)
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch2 | seq2
+(1 row)
+
DROP PUBLICATION pub;
DROP TABLE sch2.tbl1_part1;
DROP TABLE sch1.tbl1;
@@ -1718,9 +2301,81 @@ SELECT * FROM pg_publication_tables;
pub | sch1 | tbl1 | {a} |
(1 row)
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+(0 rows)
+
+DROP PUBLICATION pub;
+-- Schema publication
+CREATE PUBLICATION pub FOR SEQUENCE sch2.seq2;
+SELECT * FROM pg_publication_tables;
+ pubname | schemaname | tablename | attnames | rowfilter
+---------+------------+-----------+----------+-----------
+(0 rows)
+
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch2 | seq2
+(1 row)
+
+DROP PUBLICATION pub;
+-- Sequence publication
+CREATE PUBLICATION pub FOR SEQUENCES IN SCHEMA sch2;
+SELECT * FROM pg_publication_tables;
+ pubname | schemaname | tablename | attnames | rowfilter
+---------+------------+-----------+----------+-----------
+(0 rows)
+
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch2 | seq2
+(1 row)
+
+ALTER PUBLICATION pub ADD SEQUENCE sch1.seq1;
+SELECT * FROM pg_publication_tables;
+ pubname | schemaname | tablename | attnames | rowfilter
+---------+------------+-----------+----------+-----------
+(0 rows)
+
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch1 | seq1
+ pub | sch2 | seq2
+(2 rows)
+
+ALTER PUBLICATION pub DROP SEQUENCE sch1.seq1;
+SELECT * FROM pg_publication_tables;
+ pubname | schemaname | tablename | attnames | rowfilter
+---------+------------+-----------+----------+-----------
+(0 rows)
+
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch2 | seq2
+(1 row)
+
+ALTER PUBLICATION pub ADD SEQUENCES IN SCHEMA sch1;
+SELECT * FROM pg_publication_tables;
+ pubname | schemaname | tablename | attnames | rowfilter
+---------+------------+-----------+----------+-----------
+(0 rows)
+
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch1 | seq1
+ pub | sch2 | seq2
+(2 rows)
+
RESET client_min_messages;
DROP PUBLICATION pub;
DROP TABLE sch1.tbl1;
+DROP SEQUENCE sch1.seq1, sch2.seq2;
DROP SCHEMA sch1 cascade;
DROP SCHEMA sch2 cascade;
RESET SESSION AUTHORIZATION;
diff --git a/src/test/regress/expected/rules.out b/src/test/regress/expected/rules.out
index 624d0e5aae1..cc8a9cc36e6 100644
--- a/src/test/regress/expected/rules.out
+++ b/src/test/regress/expected/rules.out
@@ -1438,6 +1438,14 @@ pg_prepared_xacts| SELECT p.transaction,
FROM ((pg_prepared_xact() p(transaction, gid, prepared, ownerid, dbid)
LEFT JOIN pg_authid u ON ((p.ownerid = u.oid)))
LEFT JOIN pg_database d ON ((p.dbid = d.oid)));
+pg_publication_sequences| SELECT p.pubname,
+ n.nspname AS schemaname,
+ c.relname AS sequencename
+ FROM pg_publication p,
+ LATERAL pg_get_publication_sequences((p.pubname)::text) gps(relid),
+ (pg_class c
+ JOIN pg_namespace n ON ((n.oid = c.relnamespace)))
+ WHERE (c.oid = gps.relid);
pg_publication_tables| SELECT p.pubname,
n.nspname AS schemaname,
c.relname AS tablename,
diff --git a/src/test/regress/sql/object_address.sql b/src/test/regress/sql/object_address.sql
index e91072a75d1..b33e42c976b 100644
--- a/src/test/regress/sql/object_address.sql
+++ b/src/test/regress/sql/object_address.sql
@@ -19,6 +19,7 @@ CREATE TEXT SEARCH CONFIGURATION addr_ts_conf (copy=english);
CREATE TEXT SEARCH TEMPLATE addr_ts_temp (lexize=dsimple_lexize);
CREATE TEXT SEARCH PARSER addr_ts_prs
(start = prsd_start, gettoken = prsd_nexttoken, end = prsd_end, lextypes = prsd_lextype);
+CREATE SEQUENCE addr_nsp.gensequence;
CREATE TABLE addr_nsp.gentable (
a serial primary key CONSTRAINT a_chk CHECK (a > 0),
b text DEFAULT 'hello');
@@ -48,7 +49,9 @@ CREATE TRANSFORM FOR int LANGUAGE SQL (
-- suppress warning that depends on wal_level
SET client_min_messages = 'ERROR';
CREATE PUBLICATION addr_pub FOR TABLE addr_nsp.gentable;
+CREATE PUBLICATION addr_pub2 FOR SEQUENCE addr_nsp.gensequence;
CREATE PUBLICATION addr_pub_schema FOR TABLES IN SCHEMA addr_nsp;
+CREATE PUBLICATION addr_pub_schema2 FOR SEQUENCES IN SCHEMA addr_nsp;
RESET client_min_messages;
CREATE SUBSCRIPTION regress_addr_sub CONNECTION '' PUBLICATION bar WITH (connect = false, slot_name = NONE);
CREATE STATISTICS addr_nsp.gentable_stat ON a, b FROM addr_nsp.gentable;
@@ -198,7 +201,8 @@ WITH objects (type, name, args) AS (VALUES
('transform', '{int}', '{sql}'),
('access method', '{btree}', '{}'),
('publication', '{addr_pub}', '{}'),
- ('publication namespace', '{addr_nsp}', '{addr_pub_schema}'),
+ ('publication namespace', '{addr_nsp}', '{addr_pub_schema,t}'),
+ ('publication namespace', '{addr_nsp}', '{addr_pub_schema2,s}'),
('publication relation', '{addr_nsp, gentable}', '{addr_pub}'),
('subscription', '{regress_addr_sub}', '{}'),
('statistics object', '{addr_nsp, gentable_stat}', '{}')
@@ -217,7 +221,9 @@ SELECT (pg_identify_object(addr1.classid, addr1.objid, addr1.objsubid)).*,
---
DROP FOREIGN DATA WRAPPER addr_fdw CASCADE;
DROP PUBLICATION addr_pub;
+DROP PUBLICATION addr_pub2;
DROP PUBLICATION addr_pub_schema;
+DROP PUBLICATION addr_pub_schema2;
DROP SUBSCRIPTION regress_addr_sub;
DROP SCHEMA addr_nsp CASCADE;
diff --git a/src/test/regress/sql/publication.sql b/src/test/regress/sql/publication.sql
index a47c5939d58..53883b8ecd3 100644
--- a/src/test/regress/sql/publication.sql
+++ b/src/test/regress/sql/publication.sql
@@ -27,7 +27,7 @@ CREATE PUBLICATION testpub_xxx WITH (publish_via_partition_root = 'true', publis
\dRp
-ALTER PUBLICATION testpub_default SET (publish = 'insert, update, delete');
+ALTER PUBLICATION testpub_default SET (publish = 'insert, update, delete, sequence');
\dRp
@@ -46,6 +46,8 @@ ALTER PUBLICATION testpub_foralltables SET (publish = 'insert, update');
CREATE TABLE testpub_tbl2 (id serial primary key, data text);
-- fail - can't add to for all tables publication
ALTER PUBLICATION testpub_foralltables ADD TABLE testpub_tbl2;
+-- fail - can't add a table using ADD SEQUENCE command
+ALTER PUBLICATION testpub_foralltables ADD SEQUENCE testpub_tbl2;
-- fail - can't drop from all tables publication
ALTER PUBLICATION testpub_foralltables DROP TABLE testpub_tbl2;
-- fail - can't add to for all tables publication
@@ -113,6 +115,204 @@ RESET client_min_messages;
DROP TABLE testpub_tbl3, testpub_tbl3a;
DROP PUBLICATION testpub3, testpub4;
+--- adding sequences
+CREATE SEQUENCE testpub_seq0;
+CREATE SEQUENCE pub_test.testpub_seq1;
+
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_forallsequences FOR ALL SEQUENCES WITH (publish = 'sequence');
+RESET client_min_messages;
+ALTER PUBLICATION testpub_forallsequences SET (publish = 'insert, sequence');
+
+CREATE SEQUENCE testpub_seq2;
+-- fail - can't add to for all sequences publication
+ALTER PUBLICATION testpub_forallsequences ADD SEQUENCE testpub_seq2;
+-- fail - can't drop from all sequences publication
+ALTER PUBLICATION testpub_forallsequences DROP SEQUENCE testpub_seq2;
+-- fail - can't add to for all sequences publication
+ALTER PUBLICATION testpub_forallsequences SET SEQUENCE pub_test.testpub_seq1;
+
+-- fail - can't add schema to 'FOR ALL SEQUENCES' publication
+ALTER PUBLICATION testpub_forallsequences ADD SEQUENCES IN SCHEMA pub_test;
+-- fail - can't drop schema from 'FOR ALL SEQUENCES' publication
+ALTER PUBLICATION testpub_forallsequences DROP SEQUENCES IN SCHEMA pub_test;
+-- fail - can't set schema to 'FOR ALL SEQUENCES' publication
+ALTER PUBLICATION testpub_forallsequences SET SEQUENCES IN SCHEMA pub_test;
+
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_forsequence FOR SEQUENCE testpub_seq0;
+RESET client_min_messages;
+-- should be able to add schema to 'FOR SEQUENCE' publication
+ALTER PUBLICATION testpub_forsequence ADD SEQUENCES IN SCHEMA pub_test;
+\dRp+ testpub_forsequence
+-- add sequence from the schema we already added
+ALTER PUBLICATION testpub_forsequence ADD SEQUENCE pub_test.testpub_seq1;
+-- fail - can't add sequence using ADD TABLE command
+ALTER PUBLICATION testpub_forsequence ADD TABLE pub_test.testpub_seq1;
+-- should be able to drop schema from 'FOR SEQUENCE' publication
+ALTER PUBLICATION testpub_forsequence DROP SEQUENCES IN SCHEMA pub_test;
+\dRp+ testpub_forsequence
+-- should be able to set schema to 'FOR SEQUENCE' publication
+ALTER PUBLICATION testpub_forsequence SET SEQUENCES IN SCHEMA pub_test;
+\dRp+ testpub_forsequence
+
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_forschema FOR SEQUENCES IN SCHEMA pub_test;
+RESET client_min_messages;
+-- create publication with schema and sequence of the same schema
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_for_seq_schema FOR SEQUENCES IN SCHEMA pub_test, SEQUENCE pub_test.testpub_seq1;
+RESET client_min_messages;
+-- drop the publication again
+DROP PUBLICATION testpub_for_seq_schema;
+-- add a sequence of the same schema to the schema publication
+ALTER PUBLICATION testpub_forschema ADD SEQUENCE pub_test.testpub_seq1;
+-- drop a sequence from the publication again
+ALTER PUBLICATION testpub_forschema DROP SEQUENCE pub_test.testpub_seq1;
+-- drop a sequence from the schema publication which isn't in the publication
+ALTER PUBLICATION testpub_forschema DROP SEQUENCE pub_test.testpub_seq1;
+-- should be able to set sequence to schema publication
+ALTER PUBLICATION testpub_forschema SET SEQUENCE pub_test.testpub_seq1;
+\dRp+ testpub_forschema
+
+SELECT pubname, puballtables, puballsequences FROM pg_publication WHERE pubname = 'testpub_forallsequences';
+\d+ pub_test.testpub_seq1
+\dRp+ testpub_forallsequences
+DROP SEQUENCE testpub_seq0, pub_test.testpub_seq1, testpub_seq2;
+DROP PUBLICATION testpub_forallsequences, testpub_forsequence, testpub_forschema;
+
+
+-- publication testing multiple sequences at the same time
+CREATE SEQUENCE testpub_seq1;
+CREATE SEQUENCE testpub_seq2;
+
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_multi FOR SEQUENCE testpub_seq1, testpub_seq2;
+RESET client_min_messages;
+
+\dRp+ testpub_multi
+
+DROP PUBLICATION testpub_multi;
+DROP SEQUENCE testpub_seq1;
+DROP SEQUENCE testpub_seq2;
+
+
+-- Publication mixing tables and sequences
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_mix;
+RESET client_min_messages;
+
+CREATE SEQUENCE testpub_seq1;
+CREATE SEQUENCE pub_test.testpub_seq2;
+
+ALTER PUBLICATION testpub_mix ADD SEQUENCE testpub_seq1, TABLE testpub_tbl1;
+\dRp+ testpub_mix
+
+ALTER PUBLICATION testpub_mix ADD SEQUENCES IN SCHEMA pub_test, TABLES IN SCHEMA pub_test;
+\dRp+ testpub_mix
+
+ALTER PUBLICATION testpub_mix DROP SEQUENCES IN SCHEMA pub_test;
+\dRp+ testpub_mix
+
+ALTER PUBLICATION testpub_mix DROP TABLES IN SCHEMA pub_test;
+\dRp+ testpub_mix
+
+DROP PUBLICATION testpub_mix;
+DROP SEQUENCE testpub_seq1;
+DROP SEQUENCE pub_test.testpub_seq2;
+
+
+-- make sure we replicate only the correct relation type
+CREATE SCHEMA pub_test1;
+CREATE SEQUENCE pub_test1.test_seq1;
+CREATE TABLE pub_test1.test_tbl1 (a int primary key, b int);
+
+CREATE SCHEMA pub_test2;
+CREATE SEQUENCE pub_test2.test_seq2;
+CREATE TABLE pub_test2.test_tbl2 (a int primary key, b int);
+
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_schemas;
+RESET client_min_messages;
+
+-- add tables from one schema, sequences from the other
+ALTER PUBLICATION testpub_schemas ADD TABLES IN SCHEMA pub_test2;
+ALTER PUBLICATION testpub_schemas ADD SEQUENCES IN SCHEMA pub_test1;
+
+\dRp+ testpub_schemas
+
+\dn+ pub_test1
+\dn+ pub_test2
+
+\d+ pub_test1.test_seq1;
+\d+ pub_test1.test_tbl1;
+
+\d+ pub_test2.test_seq2;
+\d+ pub_test2.test_tbl2;
+
+-- add the other object type from each schema
+ALTER PUBLICATION testpub_schemas ADD TABLES IN SCHEMA pub_test1;
+ALTER PUBLICATION testpub_schemas ADD SEQUENCES IN SCHEMA pub_test2;
+
+\dRp+ testpub_schemas
+
+\dn+ pub_test1
+\dn+ pub_test2
+
+\d+ pub_test1.test_seq1;
+\d+ pub_test1.test_tbl1;
+
+\d+ pub_test2.test_seq2;
+\d+ pub_test2.test_tbl2;
+
+-- now drop the object type added first
+ALTER PUBLICATION testpub_schemas DROP TABLES IN SCHEMA pub_test2;
+ALTER PUBLICATION testpub_schemas DROP SEQUENCES IN SCHEMA pub_test1;
+
+\dRp+ testpub_schemas
+
+\dn+ pub_test1
+\dn+ pub_test2
+
+\d+ pub_test1.test_seq1;
+\d+ pub_test1.test_tbl1;
+
+\d+ pub_test2.test_seq2;
+\d+ pub_test2.test_tbl2;
+
+-- should work, we're not enforcing there's no overlap
+ALTER PUBLICATION testpub_schemas ADD TABLE pub_test1.test_tbl1;
+ALTER PUBLICATION testpub_schemas ADD SEQUENCE pub_test2.test_seq2;
+
+-- should work (different schema)
+ALTER PUBLICATION testpub_schemas ADD TABLE pub_test2.test_tbl2;
+ALTER PUBLICATION testpub_schemas ADD SEQUENCE pub_test1.test_seq1;
+
+\dRp+ testpub_schemas
+
+\d+ pub_test1.test_seq1;
+\d+ pub_test1.test_tbl1;
+
+\d+ pub_test2.test_seq2;
+\d+ pub_test2.test_tbl2;
+
+-- now drop the explicitly added objects again
+ALTER PUBLICATION testpub_schemas DROP TABLE pub_test2.test_tbl2;
+ALTER PUBLICATION testpub_schemas DROP SEQUENCE pub_test1.test_seq1;
+
+\dRp+ testpub_schemas
+
+\d+ pub_test1.test_seq1;
+\d+ pub_test1.test_tbl1;
+
+\d+ pub_test2.test_seq2;
+\d+ pub_test2.test_tbl2;
+
+DROP PUBLICATION testpub_schemas;
+DROP TABLE pub_test1.test_tbl1, pub_test2.test_tbl2;
+DROP SEQUENCE pub_test1.test_seq1, pub_test2.test_seq2;
+DROP SCHEMA pub_test1, pub_test2;
+
-- Tests for partitioned tables
SET client_min_messages = 'ERROR';
CREATE PUBLICATION testpub_forparted;
@@ -1048,32 +1248,51 @@ CREATE SCHEMA sch1;
CREATE SCHEMA sch2;
CREATE TABLE sch1.tbl1 (a int) PARTITION BY RANGE(a);
CREATE TABLE sch2.tbl1_part1 PARTITION OF sch1.tbl1 FOR VALUES FROM (1) to (10);
+CREATE SEQUENCE sch1.seq1;
+CREATE SEQUENCE sch2.seq2;
-- Schema publication that does not include the schema that has the parent table
CREATE PUBLICATION pub FOR TABLES IN SCHEMA sch2 WITH (PUBLISH_VIA_PARTITION_ROOT=1);
+ALTER PUBLICATION pub ADD SEQUENCES IN SCHEMA sch2;
SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
DROP PUBLICATION pub;
-- Table publication that does not include the parent table
CREATE PUBLICATION pub FOR TABLE sch2.tbl1_part1 WITH (PUBLISH_VIA_PARTITION_ROOT=1);
+ALTER PUBLICATION pub ADD SEQUENCE sch2.seq2;
SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
-- Table publication that includes both the parent table and the child table
ALTER PUBLICATION pub ADD TABLE sch1.tbl1;
+ALTER PUBLICATION pub ADD SEQUENCE sch1.seq1;
SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
DROP PUBLICATION pub;
-- Schema publication that does not include the schema that has the parent table
CREATE PUBLICATION pub FOR TABLES IN SCHEMA sch2 WITH (PUBLISH_VIA_PARTITION_ROOT=0);
+ALTER PUBLICATION pub ADD SEQUENCE sch1.seq1;
+SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
+
+DROP PUBLICATION pub;
+-- Sequence publication
+CREATE PUBLICATION pub FOR SEQUENCE sch2.seq2;
SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
DROP PUBLICATION pub;
-- Table publication that does not include the parent table
CREATE PUBLICATION pub FOR TABLE sch2.tbl1_part1 WITH (PUBLISH_VIA_PARTITION_ROOT=0);
SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
-- Table publication that includes both the parent table and the child table
ALTER PUBLICATION pub ADD TABLE sch1.tbl1;
+ALTER PUBLICATION pub ADD SEQUENCES IN SCHEMA sch2;
SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
DROP PUBLICATION pub;
DROP TABLE sch2.tbl1_part1;
@@ -1086,10 +1305,37 @@ CREATE TABLE sch1.tbl1_part3 (a int) PARTITION BY RANGE(a);
ALTER TABLE sch1.tbl1 ATTACH PARTITION sch1.tbl1_part3 FOR VALUES FROM (20) to (30);
CREATE PUBLICATION pub FOR TABLES IN SCHEMA sch1 WITH (PUBLISH_VIA_PARTITION_ROOT=1);
SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
+
+DROP PUBLICATION pub;
+-- Schema publication
+CREATE PUBLICATION pub FOR SEQUENCE sch2.seq2;
+SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
+
+DROP PUBLICATION pub;
+-- Sequence publication
+CREATE PUBLICATION pub FOR SEQUENCES IN SCHEMA sch2;
+SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
+
+ALTER PUBLICATION pub ADD SEQUENCE sch1.seq1;
+SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
+
+ALTER PUBLICATION pub DROP SEQUENCE sch1.seq1;
+SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
+
+ALTER PUBLICATION pub ADD SEQUENCES IN SCHEMA sch1;
+SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
RESET client_min_messages;
DROP PUBLICATION pub;
+
DROP TABLE sch1.tbl1;
+DROP SEQUENCE sch1.seq1, sch2.seq2;
DROP SCHEMA sch1 cascade;
DROP SCHEMA sch2 cascade;
diff --git a/src/test/subscription/t/030_sequences.pl b/src/test/subscription/t/030_sequences.pl
new file mode 100644
index 00000000000..1d94e78ae9f
--- /dev/null
+++ b/src/test/subscription/t/030_sequences.pl
@@ -0,0 +1,205 @@
+
+# Copyright (c) 2021, PostgreSQL Global Development Group
+
+# This tests that sequences are replicated correctly by logical replication
+use strict;
+use warnings;
+use PostgreSQL::Test::Cluster;
+use PostgreSQL::Test::Utils;
+use Test::More;
+
+# Initialize publisher node
+my $node_publisher = PostgreSQL::Test::Cluster->new('publisher');
+$node_publisher->init(allows_streaming => 'logical');
+$node_publisher->start;
+
+# Create subscriber node
+my $node_subscriber = PostgreSQL::Test::Cluster->new('subscriber');
+$node_subscriber->init(allows_streaming => 'logical');
+$node_subscriber->start;
+
+# Create some preexisting content on publisher
+my $ddl = qq(
+ CREATE TABLE seq_test (v BIGINT);
+ CREATE SEQUENCE s;
+);
+
+# Setup structure on the publisher
+$node_publisher->safe_psql('postgres', $ddl);
+
+# Create some the same structure on subscriber, and an extra sequence that
+# we'll create on the publisher later
+$ddl = qq(
+ CREATE TABLE seq_test (v BIGINT);
+ CREATE SEQUENCE s;
+ CREATE SEQUENCE s2;
+);
+
+$node_subscriber->safe_psql('postgres', $ddl);
+
+# Setup logical replication
+my $publisher_connstr = $node_publisher->connstr . ' dbname=postgres';
+$node_publisher->safe_psql('postgres',
+ "CREATE PUBLICATION seq_pub");
+
+$node_publisher->safe_psql('postgres',
+ "ALTER PUBLICATION seq_pub ADD SEQUENCE s");
+
+$node_subscriber->safe_psql('postgres',
+ "CREATE SUBSCRIPTION seq_sub CONNECTION '$publisher_connstr' PUBLICATION seq_pub"
+);
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Wait for initial sync to finish as well
+my $synced_query =
+ "SELECT count(1) = 0 FROM pg_subscription_rel WHERE srsubstate NOT IN ('s', 'r');";
+$node_subscriber->poll_query_until('postgres', $synced_query)
+ or die "Timed out while waiting for subscriber to synchronize data";
+
+# Insert initial test data
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ -- generate a number of values using the sequence
+ INSERT INTO seq_test SELECT nextval('s') FROM generate_series(1,100);
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+my $result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s;
+));
+
+is( $result, '100|32|t',
+ 'initial test data replicated');
+
+
+# advance the sequence in a rolled-back transaction - the rollback
+# does not wait for the replication, so we could see any intermediate state
+# so do something else after the test, to ensure we wait for everything
+# (but make sure not to use the sequence)
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ INSERT INTO seq_test SELECT nextval('s') FROM generate_series(1,100);
+ ROLLBACK;
+ INSERT INTO seq_test VALUES (-1);
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s;
+));
+
+is( $result, '100|32|t',
+ 'sequence advanced in rolled-back transaction does not replicate');
+
+
+# create a new sequence and roll it back - should not be replicated, due to
+# the transactional behavior
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ CREATE SEQUENCE s2;
+ ALTER PUBLICATION seq_pub ADD SEQUENCE s2;
+ INSERT INTO seq_test SELECT nextval('s2') FROM generate_series(1,100);
+ ROLLBACK;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s2;
+));
+
+is( $result, '1|0|f',
+ 'create new sequence and roll it back');
+
+
+# create a new sequence, advance it in a rolled-back transaction, but commit
+# the create - the advance should be replicated nevertheless
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ CREATE SEQUENCE s2;
+ ALTER PUBLICATION seq_pub ADD SEQUENCE s2;
+ SAVEPOINT sp1;
+ INSERT INTO seq_test SELECT nextval('s2') FROM generate_series(1,100);
+ ROLLBACK TO sp1;
+ COMMIT;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Wait for sync of the second sequence we just added to finish
+$synced_query =
+ "SELECT count(1) = 0 FROM pg_subscription_rel WHERE srsubstate NOT IN ('s', 'r');";
+$node_subscriber->poll_query_until('postgres', $synced_query)
+ or die "Timed out while waiting for subscriber to synchronize data";
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s2;
+));
+
+is( $result, '100|32|t',
+ 'create sequence, advance it in rolled-back transaction, but commit the create');
+
+
+# advance the new sequence in a transaction, and roll it back - the rollback
+# does not wait for the replication, so we could see any intermediate state
+# so do something else after the test, to ensure we wait for everything
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ INSERT INTO seq_test SELECT nextval('s2') FROM generate_series(1,100);
+ ROLLBACK;
+ INSERT INTO seq_test VALUES (-1);
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s2;
+));
+
+# we did touch the sequence in the transaction, so even tough the advance
+# was rolled back, we should have replicated it at the end
+is( $result, '100|32|t',
+ 'advance the new sequence in a transaction and roll it back');
+
+
+# advance the sequence in a subtransaction - the subtransaction gets rolled
+# back, but commit the main one - the changes should still be replicated
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ SAVEPOINT s1;
+ INSERT INTO seq_test SELECT nextval('s2') FROM generate_series(1,100);
+ ROLLBACK TO s1;
+ COMMIT;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s2;
+));
+
+is( $result, '300|30|t',
+ 'advance sequence in a subtransaction');
+
+
+done_testing();
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2022-11-16 04:43 Ian Lawrence Barwick <[email protected]>
parent: Tomas Vondra <[email protected]>
1 sibling, 0 replies; 127+ messages in thread
From: Ian Lawrence Barwick @ 2022-11-16 04:43 UTC (permalink / raw)
To: Tomas Vondra <[email protected]>; +Cc: PostgreSQL Hackers <[email protected]>
2022年11月12日(土) 7:49 Tomas Vondra <[email protected]>:
>
> Hi,
>
> I noticed on cfbot the patch no longer applies, so here's a rebased
> version. Most of the breakage was due to the column filtering reworks,
> grammar changes etc. A lot of bitrot, but mostly mechanical stuff.
(...)
Hi
Thanks for the update patch.
While reviewing the patch backlog, we have determined that this patch adds
one or more TAP tests but has not added the test to the "meson.build" file.
To do this, locate the relevant "meson.build" file for each test and add it
in the 'tests' dictionary, which will look something like this:
'tap': {
'tests': [
't/001_basic.pl',
],
},
For some additional details please see this Wiki article:
https://wiki.postgresql.org/wiki/Meson_for_patch_authors
For more information on the meson build system for PostgreSQL see:
https://wiki.postgresql.org/wiki/Meson
Regards
Ian Barwick
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2022-11-16 21:05 Robert Haas <[email protected]>
parent: Tomas Vondra <[email protected]>
1 sibling, 1 reply; 127+ messages in thread
From: Robert Haas @ 2022-11-16 21:05 UTC (permalink / raw)
To: Tomas Vondra <[email protected]>; +Cc: PostgreSQL Hackers <[email protected]>
On Fri, Nov 11, 2022 at 5:49 PM Tomas Vondra
<[email protected]> wrote:
> The other option might be to make these messages non-transactional, in
> which case we'd separate the ordering from COMMIT ordering, evading the
> reordering problem.
>
> That'd mean we'd ignore rollbacks (which seems fine), we could probably
> optimize this by checking if the state actually changed, etc. But we'd
> also need to deal with transactions created in the (still uncommitted)
> transaction. But I'm also worried it might lead to the same issue with
> non-transactional behaviors that forced revert in v15.
I think it might be a good idea to step back slightly from
implementation details and try to agree on a theoretical model of
what's happening here. Let's start by banishing the words
transactional and non-transactional from the conversation and talk
about what logical replication is trying to do.
We can imagine that the replicated objects on the primary pass through
a series of states S1, S2, ..., Sn, where n keeps going up as new
state changes occur. The state, for our purposes here, is the contents
of the database as they could be observed by a user running SELECT
queries at some moment in time chosen by the user. For instance, if
the initial state of the database is S1, and then the user executes
BEGIN, 2 single-row INSERT statements, and a COMMIT, then S2 is the
state that differs from S1 in that both of those rows are now part of
the database contents. There is no state where one of those rows is
visible and the other is not. That was never observable by the user,
except from within the transaction as it was executing, which we can
and should discount. I believe that the goal of logical replication is
to bring about a state of affairs where the set of states observable
on the standby is a subset of the states observable on the primary.
That is, if the primary goes from S1 to S2 to S3, the standby can do
the same thing, or it can go straight from S1 to S3 without ever
making it possible for the user to observe S2. Either is correct
behavior. But the standby cannot invent any new states that didn't
occur on the primary. It can't decide to go from S1 to S1.5 to S2.5 to
S3, or something like that. It can only consolidate changes that
occurred separately on the primary, never split them up. Neither can
it reorder them.
Now, if you accept this as a reasonable definition of correctness,
then the next question is what consequences it has for transactional
and non-transactional behavior. If all behavior is transactional, then
we've basically got to replay each primary transaction in a single
standby transaction, and commit those transactions in the same order
that the corresponding primary transactions committed. We could
legally choose to merge a group of transactions that committed one
after the other on the primary into a single transaction on the
standby, and it might even be a good idea if they're all very tiny,
but it's not required. But if there are non-transactional things
happening, then there are changes that become visible at some time
other than at a transaction commit. For example, consider this
sequence of events, in which each "thing" that happens is
transactional except where the contrary is noted:
T1: BEGIN;
T2: BEGIN;
T1: Do thing 1;
T2: Do thing 2;
T1: Do a non-transactional thing;
T1: Do thing 3;
T2: Do thing 4;
T2: COMMIT;
T1: COMMIT;
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2022-11-17 01:41 Tomas Vondra <[email protected]>
parent: Robert Haas <[email protected]>
0 siblings, 2 replies; 127+ messages in thread
From: Tomas Vondra @ 2022-11-17 01:41 UTC (permalink / raw)
To: Robert Haas <[email protected]>; +Cc: PostgreSQL Hackers <[email protected]>
On 11/16/22 22:05, Robert Haas wrote:
> On Fri, Nov 11, 2022 at 5:49 PM Tomas Vondra
> <[email protected]> wrote:
>> The other option might be to make these messages non-transactional, in
>> which case we'd separate the ordering from COMMIT ordering, evading the
>> reordering problem.
>>
>> That'd mean we'd ignore rollbacks (which seems fine), we could probably
>> optimize this by checking if the state actually changed, etc. But we'd
>> also need to deal with transactions created in the (still uncommitted)
>> transaction. But I'm also worried it might lead to the same issue with
>> non-transactional behaviors that forced revert in v15.
>
> I think it might be a good idea to step back slightly from
> implementation details and try to agree on a theoretical model of
> what's happening here. Let's start by banishing the words
> transactional and non-transactional from the conversation and talk
> about what logical replication is trying to do.
>
OK, let's try.
> We can imagine that the replicated objects on the primary pass through
> a series of states S1, S2, ..., Sn, where n keeps going up as new
> state changes occur. The state, for our purposes here, is the contents
> of the database as they could be observed by a user running SELECT
> queries at some moment in time chosen by the user. For instance, if
> the initial state of the database is S1, and then the user executes
> BEGIN, 2 single-row INSERT statements, and a COMMIT, then S2 is the
> state that differs from S1 in that both of those rows are now part of
> the database contents. There is no state where one of those rows is
> visible and the other is not. That was never observable by the user,
> except from within the transaction as it was executing, which we can
> and should discount. I believe that the goal of logical replication is
> to bring about a state of affairs where the set of states observable
> on the standby is a subset of the states observable on the primary.
> That is, if the primary goes from S1 to S2 to S3, the standby can do
> the same thing, or it can go straight from S1 to S3 without ever
> making it possible for the user to observe S2. Either is correct
> behavior. But the standby cannot invent any new states that didn't
> occur on the primary. It can't decide to go from S1 to S1.5 to S2.5 to
> S3, or something like that. It can only consolidate changes that
> occurred separately on the primary, never split them up. Neither can
> it reorder them.
>
I mostly agree, and in a way the last patch aims to do roughly this,
i.e. make sure that the state after each transaction matches the state a
user might observe on the primary (modulo implementation challenges).
There's a couple of caveats, though:
1) Maybe we should focus more on "actually observed" state instead of
"observable". Who cares if the sequence moved forward in a transaction
that was ultimately rolled back? No committed transaction should have
observer those values - in a way, the last "valid" state of the sequence
is the last value generated in a transaction that ultimately committed.
2) I think what matters more is that we never generate duplicate value.
That is, if you generate a value from a sequence, commit a transaction
and replicate it, then the logical standby should not generate the same
value from the sequence. This guarantee seems necessary for "failover"
to logical standby.
> Now, if you accept this as a reasonable definition of correctness,
> then the next question is what consequences it has for transactional
> and non-transactional behavior. If all behavior is transactional, then
> we've basically got to replay each primary transaction in a single
> standby transaction, and commit those transactions in the same order
> that the corresponding primary transactions committed. We could
> legally choose to merge a group of transactions that committed one
> after the other on the primary into a single transaction on the
> standby, and it might even be a good idea if they're all very tiny,
> but it's not required. But if there are non-transactional things
> happening, then there are changes that become visible at some time
> other than at a transaction commit. For example, consider this
> sequence of events, in which each "thing" that happens is
> transactional except where the contrary is noted:
>
> T1: BEGIN;
> T2: BEGIN;
> T1: Do thing 1;
> T2: Do thing 2;
> T1: Do a non-transactional thing;
> T1: Do thing 3;
> T2: Do thing 4;
> T2: COMMIT;
> T1: COMMIT;
>
> From the point of the user here, there are 4 observable states here:
>
> S1: Initiate state.
> S2: State after the non-transactional thing happens.
> S3: State after T2 commits (reflects the non-transactional thing plus
> things 2 and 4).
> S4: State after T1 commits.
>
> Basically, the non-transactional thing behaves a whole lot like a
> separate transaction. That non-transactional operation ought to be
> replicated before T2, which ought to be replicated before T1. Maybe
> logical replication ought to treat it in exactly that way: as a
> separate operation that needs to be replicated after any earlier
> transactions that completed prior to the history shown here, but
> before T2 or T1. Alternatively, you can merge the non-transactional
> change into T2, i.e. the first transaction that committed after it
> happened. But you can't merge it into T1, even though it happened in
> T1. If you do that, then you're creating states on the standby that
> never existed on the primary, which is wrong. You could argue that
> this is just nitpicking: who cares if the change in the sequence value
> doesn't get replicated at exactly the right moment? But I don't think
> it's a technicality at all: I think if we don't make the operation
> appear to happen at the same point in the sequence as it became
> visible on the master, then there will be endless artifacts and corner
> cases to the bottom of which we will never get. Just like if we
> replicated the actual transactions out of order, chaos would ensue,
> because there can be logical dependencies between them, so too can
> there be logical dependencies between non-transactional operations, or
> between a non-transactional operation and a transactional operation.
>
Well, yeah - we can either try to perform the stuff independently of the
transactions that triggered it, or we can try making it part of some of
the transactions. Each of those options has problems, though :-(
The first version of the patch tried the first approach, i.e. decode the
increments and apply that independently. But:
(a) What would you do with increments of sequences created/reset in a
transaction? Can't apply those outside the transaction, because it
might be rolled back (and that state is not visible on primary).
(b) What about increments created before we have a proper snapshot?
There may be transactions dependent on the increment. This is what
ultimately led to revert of the patch.
This version of the patch tries to do the opposite thing - make sure
that the state after each commit matches what the transaction might have
seen (for sequences it accessed). It's imperfect, because it might log a
state generated "after" the sequence got accessed - it focuses on the
guarantee not to generate duplicate values.
> To make it more concrete, consider two sessions concurrently running this SQL:
>
> insert into t1 select nextval('s1') from generate_series(1,1000000) g;
>
> There are, in effect, 2000002 transaction-like things here. The
> sequence gets incremented 2 million times, and then there are 2
> commits that each insert a million rows. Perhaps the actual order of
> events looks something like this:
>
> 1. nextval the sequence N times, where N >= 1 million
> 2. commit the first transaction, adding a million rows to t1
> 3. nextval the sequence 2 million - N times
> 4. commit the second transaction, adding another million rows to t1
>
> Unless we replicate all of the nextval operations that occur in step 1
> at the same time or prior to replicating the first transaction in step
> 2, we might end up making visible a state where the next value of the
> sequence is less than the highest value present in the table, which
> would be bad.
>
Right, that's the "guarantee" I've mentioned above, more or less.
> With that perhaps overly-long set of preliminaries, I'm going to move
> on to talking about the implementation ideas which you mention. You
> write that "the current solution is based on WAL-logging state of all
> sequences incremented by the transaction at COMMIT" and then, it seems
> to me, go on to demonstrate that it's simply incorrect. In my opinion,
> the fundamental problem is that it doesn't look at the order that
> things happened on the primary and do them in the same order on the
> standby. Instead, it accepts that the non-transactional operations are
> going to be replicated at the wrong time, and then tries to patch
> around the issue by attempting to scrounge up the correct values at
> some convenient point and use that data to compensate for our failure
> to do the right thing at an earlier point. That doesn't seem like a
> satisfying solution, and I think it will be hard to make it fully
> correct.
>
I understand what you're saying, but I'm not sure I agree with you.
Yes, this would mean we accept we may end up with something like this:
1: T1 logs sequence state S1
2: someone increments sequence
3: T2 logs sequence stats S2
4: T2 commits
5: T1 commits
which "inverts" the apply order of S1 vs. S2, because we first apply S2
and then the "old" S1. But as long as we're smart enough to "discard"
applying S1, I think that's acceptable - because it guarantees we'll not
generate duplicate values (with values in the committed transaction).
I'd also argue it does not actually generate invalid state, because once
we commit either transaction, S2 is what's visible.
Yes, if you so "SELECT * FROM sequence" you'll see some intermediate
state, but that's not how sequences are accessed. And you can't do
currval('s') from a transaction that never accessed the sequence.
And if it did, we'd write S2 (or whatever it saw) as part of it's commits.
So I think the main issue of this approach is how to decide which
sequence states are obsolete and should be skipped.
> Your alternative proposal says "The other option might be to make
> these messages non-transactional, in which case we'd separate the
> ordering from COMMIT ordering, evading the reordering problem." But I
> don't think that avoids the reordering problem at all.
I don't understand why. Why would it not address the reordering issue?
> Nor do I think it's correct.
Nor do I understand this. I mean, isn't it essentially the option you
mentioned earlier - treating the non-transactional actions as
independent transactions? Yes, we'd be batching them so that we'd not
see "intermediate" states, but those are not observed by abyone.
> I don't think you *can* separate the ordering of these
> operations from the COMMIT ordering. They are, as I argue here,
> essentially mini-commits that only bump the sequence value, and they
> need to be replicated after the transactions that commit prior to the
> sequence value bump and before those that commit afterward. If they
> aren't handled that way, I don't think you're going to get fully
> correct behavior.
I'm confused. Isn't that pretty much exactly what I'm proposing? Imagine
you have something like this:
1: T1 does something and also increments a sequence
2: T1 logs state of the sequence (right before commit)
3: T1 writes COMMIT
Now when we decode/apply this, we end up doing this:
1: decode all T1 changes, stash them
2: decode the sequence state and apply it separately
3: decode COMMIT, apply all T1 changes
There might be other transactions interleaving with this, but I think
it'd behave correctly. What example would not work?
>
> I'm going to confess that I have no really specific idea how to
> implement that. I'm just not sufficiently familiar with this code.
> However, I suspect that the solution lies in changing things on the
> decoding side rather than in the WAL format. I feel like the
> information that we need in order to do the right thing must already
> be present in the WAL. If it weren't, then how could crash recovery
> work correctly, or physical replication? At any given moment, you can
> choose to promote a physical standby, and at that point the state you
> observe on the new primary had better be some state that existed on
> the primary at some point in its history. At any moment, you can
> unplug the primary, restart it, and run crash recovery, and if you do,
> you had better end up with some state that existed on the primary at
> some point shortly before the crash. I think that there are actually a
> few subtle inaccuracies in the last two sentences, because actually
> the order in which transactions become visible on a physical standby
> can differ from the order in which it happens on the primary, but I
> don't think that actually changes the picture much. The point is that
> the WAL is the definitive source of information about what happened
> and in what order it happened, and we use it in that way already in
> the context of physical replication, and of standbys. If logical
> decoding has a problem with some case that those systems handle
> correctly, the problem is with logical decoding, not the WAL format.
>
The problem lies in how we log sequences. If we wrote each individual
increment to WAL, it might work the way you propose (except for cases
with sequences created in a transaction, etc.). But that's not what we
do - we log sequence increments in batches of 32 values, and then only
modify the sequence relfilenode.
This works for physical replication, because the WAL describes the
"next" state of the sequence (so if you do "SELECT * FROM sequence"
you'll not see the same state, and the sequence value may "jump ahead"
after a failover).
But for logical replication this does not work, because the transaction
might depend on a state created (WAL-logged) by some other transaction.
And perhaps that transaction actually happened *before* we even built
the first snapshot for decoding :-/
There's also the issue with what snapshot to use when decoding these
transactional changes in logical decoding (see
> In particular, I think it's likely that the "non-transactional
> messages" that you mention earlier don't get applied at the point in
> the commit sequence where they were found in the WAL. Not sure why
> exactly, but perhaps the point at which we're reading WAL runs ahead
> of the decoding per se, or something like that, and thus those
> non-transactional messages arrive too early relative to the commit
> ordering. Possibly that could be changed, and they could be buffered
I'm not sure which case of "non-transactional messages" this refers to,
so I can't quite respond to these comments. Perhaps you mean the
problems that killed the previous patch [1]?
[1]
https://www.postgresql.org/message-id/00708727-d856-1886-48e3-811296c7ba8c%40enterprisedb.com
> until earlier commits are replicated. Or else, when we see a WAL
> record for a non-transactional sequence operation, we could arrange to
> bundle that operation into an "adjacent" replicated transaction i.e.
IIRC moving stuff between transactions during decoding is problematic,
because of snapshots.
> the transaction whose commit record occurs most nearly prior to, or
> most nearly after, the WAL record for the operation itself. Or else,
> we could create "virtual" transactions for such operations and make
> sure those get replayed at the right point in the commit sequence. Or
> else, I don't know, maybe something else. But I think the overall
> picture is that we need to approach the problem by replicating changes
> in WAL order, as a physical standby would do. Saying that a change is
> "nontransactional" doesn't mean that it's exempt from ordering
> requirements; rather, it means that that change has its own place in
> that ordering, distinct from the transaction in which it occurred.
>
But doesn't the approach with WAL-logging sequence state before COMMIT,
and then applying it independently in WAL-order, do pretty much this?
regards
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2022-11-17 02:43 Andres Freund <[email protected]>
parent: Tomas Vondra <[email protected]>
1 sibling, 1 reply; 127+ messages in thread
From: Andres Freund @ 2022-11-17 02:43 UTC (permalink / raw)
To: Tomas Vondra <[email protected]>; +Cc: Robert Haas <[email protected]>; PostgreSQL Hackers <[email protected]>
Hi,
On 2022-11-17 02:41:14 +0100, Tomas Vondra wrote:
> Well, yeah - we can either try to perform the stuff independently of the
> transactions that triggered it, or we can try making it part of some of
> the transactions. Each of those options has problems, though :-(
>
> The first version of the patch tried the first approach, i.e. decode the
> increments and apply that independently. But:
>
> (a) What would you do with increments of sequences created/reset in a
> transaction? Can't apply those outside the transaction, because it
> might be rolled back (and that state is not visible on primary).
I think a reasonable approach could be to actually perform different WAL
logging for that case. It'll require a bit of machinery, but could actually
result in *less* WAL logging overall, because we don't need to emit a WAL
record for each SEQ_LOG_VALS sequence values.
> (b) What about increments created before we have a proper snapshot?
> There may be transactions dependent on the increment. This is what
> ultimately led to revert of the patch.
I don't understand this - why would we ever need to process those increments
from before we have a snapshot? Wouldn't they, by definition, be before the
slot was active?
To me this is the rough equivalent of logical decoding not giving the initial
state of all tables. You need some process outside of logical decoding to get
that (obviously we have some support for that via the exported data snapshot
during slot creation).
I assume that part of the initial sync would have to be a new sequence
synchronization step that reads all the sequence states on the publisher and
ensures that the subscriber sequences are at the same point. There's a bit of
trickiness there, but it seems entirely doable. The logical replication replay
support for sequences will have to be a bit careful about not decreasing the
subscriber's sequence values - the standby initially will be ahead of the
increments we'll see in the WAL. But that seems inevitable given the
non-transactional nature of sequences.
> This version of the patch tries to do the opposite thing - make sure
> that the state after each commit matches what the transaction might have
> seen (for sequences it accessed). It's imperfect, because it might log a
> state generated "after" the sequence got accessed - it focuses on the
> guarantee not to generate duplicate values.
That approach seems quite wrong to me.
> > I'm going to confess that I have no really specific idea how to
> > implement that. I'm just not sufficiently familiar with this code.
> > However, I suspect that the solution lies in changing things on the
> > decoding side rather than in the WAL format. I feel like the
> > information that we need in order to do the right thing must already
> > be present in the WAL. If it weren't, then how could crash recovery
> > work correctly, or physical replication? At any given moment, you can
> > choose to promote a physical standby, and at that point the state you
> > observe on the new primary had better be some state that existed on
> > the primary at some point in its history. At any moment, you can
> > unplug the primary, restart it, and run crash recovery, and if you do,
> > you had better end up with some state that existed on the primary at
> > some point shortly before the crash.
One minor exception here is that there's no real time bound to see the last
few sequence increments if nothing after the XLOG_SEQ_LOG records forces a WAL
flush.
> > I think that there are actually a
> > few subtle inaccuracies in the last two sentences, because actually
> > the order in which transactions become visible on a physical standby
> > can differ from the order in which it happens on the primary, but I
> > don't think that actually changes the picture much. The point is that
> > the WAL is the definitive source of information about what happened
> > and in what order it happened, and we use it in that way already in
> > the context of physical replication, and of standbys. If logical
> > decoding has a problem with some case that those systems handle
> > correctly, the problem is with logical decoding, not the WAL format.
> >
>
> The problem lies in how we log sequences. If we wrote each individual
> increment to WAL, it might work the way you propose (except for cases
> with sequences created in a transaction, etc.). But that's not what we
> do - we log sequence increments in batches of 32 values, and then only
> modify the sequence relfilenode.
> This works for physical replication, because the WAL describes the
> "next" state of the sequence (so if you do "SELECT * FROM sequence"
> you'll not see the same state, and the sequence value may "jump ahead"
> after a failover).
>
> But for logical replication this does not work, because the transaction
> might depend on a state created (WAL-logged) by some other transaction.
> And perhaps that transaction actually happened *before* we even built
> the first snapshot for decoding :-/
I really can't follow the "depend on state ... by some other transaction"
aspect.
Even the case of a sequence that is renamed inside a transaction that did
*not* create / reset the sequence and then also triggers increment of the
sequence seems to be dealt with reasonably by processing sequence increments
outside a transaction - the old name will be used for the increments, replay
of the renaming transaction would then implement the rename in a hypothetical
DDL-replay future.
> There's also the issue with what snapshot to use when decoding these
> transactional changes in logical decoding (see
Incomplete parenthetical? Or were you referencing the next paragraph?
What are the transactional changes you're referring to here?
I did some skimming of the referenced thread about the reversal of the last
approach, but I couldn't really understand what the fundamental issues were
with the reverted implementation - it's a very long thread and references
other threads.
Greetings,
Andres Freund
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2022-11-17 11:39 Tomas Vondra <[email protected]>
parent: Andres Freund <[email protected]>
0 siblings, 1 reply; 127+ messages in thread
From: Tomas Vondra @ 2022-11-17 11:39 UTC (permalink / raw)
To: Andres Freund <[email protected]>; +Cc: Robert Haas <[email protected]>; PostgreSQL Hackers <[email protected]>
On 11/17/22 03:43, Andres Freund wrote:
> Hi,
>
>
> On 2022-11-17 02:41:14 +0100, Tomas Vondra wrote:
>> Well, yeah - we can either try to perform the stuff independently of the
>> transactions that triggered it, or we can try making it part of some of
>> the transactions. Each of those options has problems, though :-(
>>
>> The first version of the patch tried the first approach, i.e. decode the
>> increments and apply that independently. But:
>>
>> (a) What would you do with increments of sequences created/reset in a
>> transaction? Can't apply those outside the transaction, because it
>> might be rolled back (and that state is not visible on primary).
>
> I think a reasonable approach could be to actually perform different WAL
> logging for that case. It'll require a bit of machinery, but could actually
> result in *less* WAL logging overall, because we don't need to emit a WAL
> record for each SEQ_LOG_VALS sequence values.
>
Could you elaborate? Hard to comment without knowing more ...
My point was that stuff like this (creating a new sequence or at least a
new relfilenode) means we can't apply that independently of the
transaction (unlike regular increments). I'm not sure how a change to
WAL logging would make that go away.
>
>
>> (b) What about increments created before we have a proper snapshot?
>> There may be transactions dependent on the increment. This is what
>> ultimately led to revert of the patch.
>
> I don't understand this - why would we ever need to process those increments
> from before we have a snapshot? Wouldn't they, by definition, be before the
> slot was active?
>
> To me this is the rough equivalent of logical decoding not giving the initial
> state of all tables. You need some process outside of logical decoding to get
> that (obviously we have some support for that via the exported data snapshot
> during slot creation).
>
Which is what already happens during tablesync, no? We more or less copy
sequences as if they were tables.
> I assume that part of the initial sync would have to be a new sequence
> synchronization step that reads all the sequence states on the publisher and
> ensures that the subscriber sequences are at the same point. There's a bit of
> trickiness there, but it seems entirely doable. The logical replication replay
> support for sequences will have to be a bit careful about not decreasing the
> subscriber's sequence values - the standby initially will be ahead of the
> increments we'll see in the WAL. But that seems inevitable given the
> non-transactional nature of sequences.
>
See fetch_sequence_data / copy_sequence in the patch. The bit about
ensuring the sequence does not go away (say, using page LSN and/or LSN
of the increment) is not there, however isn't that pretty much what I
proposed doing for "reconciling" the sequence state logged at COMMIT?
>
>> This version of the patch tries to do the opposite thing - make sure
>> that the state after each commit matches what the transaction might have
>> seen (for sequences it accessed). It's imperfect, because it might log a
>> state generated "after" the sequence got accessed - it focuses on the
>> guarantee not to generate duplicate values.
>
> That approach seems quite wrong to me.
>
Why? Because it might log a state for sequence as of COMMIT, when the
transaction accessed the sequence much earlier? That is, this may happen:
T1: nextval('s') -> 1
T2: call nextval('s') 1000000x
T1: commit
and T1 will log sequence state ~1000001, give or take. I don't think
there's way around that, given the non-transactional nature of
sequences. And I'm not convinced this is an issue, as it ensures
uniqueness of values generated on the subscriber. And I think it's
reasonable to replicate the sequence state as of the commit (because
that's what you'd see on the primary).
>
>>> I'm going to confess that I have no really specific idea how to
>>> implement that. I'm just not sufficiently familiar with this code.
>>> However, I suspect that the solution lies in changing things on the
>>> decoding side rather than in the WAL format. I feel like the
>>> information that we need in order to do the right thing must already
>>> be present in the WAL. If it weren't, then how could crash recovery
>>> work correctly, or physical replication? At any given moment, you can
>>> choose to promote a physical standby, and at that point the state you
>>> observe on the new primary had better be some state that existed on
>>> the primary at some point in its history. At any moment, you can
>>> unplug the primary, restart it, and run crash recovery, and if you do,
>>> you had better end up with some state that existed on the primary at
>>> some point shortly before the crash.
>
> One minor exception here is that there's no real time bound to see the last
> few sequence increments if nothing after the XLOG_SEQ_LOG records forces a WAL
> flush.
>
Right. Another issue is we ignore stuff that happened in aborted
transactions, so then nextval('s') in another transaction may not wait
for syncrep to confirm receiving that WAL. Which is a data loss case,
see [1]:
[1]
https://www.postgresql.org/message-id/712cad46-a9c8-1389-aef8-faf0203c9be9%40enterprisedb.com
>
>>> I think that there are actually a
>>> few subtle inaccuracies in the last two sentences, because actually
>>> the order in which transactions become visible on a physical standby
>>> can differ from the order in which it happens on the primary, but I
>>> don't think that actually changes the picture much. The point is that
>>> the WAL is the definitive source of information about what happened
>>> and in what order it happened, and we use it in that way already in
>>> the context of physical replication, and of standbys. If logical
>>> decoding has a problem with some case that those systems handle
>>> correctly, the problem is with logical decoding, not the WAL format.
>>>
>>
>> The problem lies in how we log sequences. If we wrote each individual
>> increment to WAL, it might work the way you propose (except for cases
>> with sequences created in a transaction, etc.). But that's not what we
>> do - we log sequence increments in batches of 32 values, and then only
>> modify the sequence relfilenode.
>
>> This works for physical replication, because the WAL describes the
>> "next" state of the sequence (so if you do "SELECT * FROM sequence"
>> you'll not see the same state, and the sequence value may "jump ahead"
>> after a failover).
>>
>> But for logical replication this does not work, because the transaction
>> might depend on a state created (WAL-logged) by some other transaction.
>> And perhaps that transaction actually happened *before* we even built
>> the first snapshot for decoding :-/
>
> I really can't follow the "depend on state ... by some other transaction"
> aspect.
>
T1: nextval('s') -> writes WAL, covering by the next 32 increments
T2: nextval('s') -> no WAL generated, covered by T1 WAL
This is what I mean by "dependency" on state logged by another
transaction. It already causes problems with streaming replication (see
the reference to syncrep above), logical replication has the same issue.
>
> Even the case of a sequence that is renamed inside a transaction that did
> *not* create / reset the sequence and then also triggers increment of the
> sequence seems to be dealt with reasonably by processing sequence increments
> outside a transaction - the old name will be used for the increments, replay
> of the renaming transaction would then implement the rename in a hypothetical
> DDL-replay future.
>
>
>> There's also the issue with what snapshot to use when decoding these
>> transactional changes in logical decoding (see
>
> Incomplete parenthetical? Or were you referencing the next paragraph?
>
> What are the transactional changes you're referring to here?
>
Sorry, IIRC I merely wanted to mention/reference the snapshot issue in
the thread [2] that I ended up referencing in the next paragraph.
[2]
https://www.postgresql.org/message-id/00708727-d856-1886-48e3-811296c7ba8c%40enterprisedb.com
>
> I did some skimming of the referenced thread about the reversal of the last
> approach, but I couldn't really understand what the fundamental issues were
> with the reverted implementation - it's a very long thread and references
> other threads.
>
Yes, it's long/complex, but I intentionally linked to a specific message
which describes the issue ...
It's entirely possible there is a simple fix for the issue, and I just
got confused / unable to see the solution. The whole issue was due to
having a mix of transactional and non-transactional cases, similarly to
logical messages - and logicalmsg_decode() has the same issue, so maybe
let's talk about that for a moment.
See [3] and imagine you're dealing with a transactional message, but
you're still building a consistent snapshot. So the first branch applies:
if (transactional &&
!SnapBuildProcessChange(builder, xid, buf->origptr))
return;
but because we don't have a snapshot, SnapBuildProcessChange does this:
if (builder->state < SNAPBUILD_FULL_SNAPSHOT)
return false;
which however means logicalmsg_decode() does
snapshot = SnapBuildGetOrBuildSnapshot(builder);
which crashes, because it hits this assert:
Assert(builder->state == SNAPBUILD_CONSISTENT);
The sequence decoding did almost the same thing, with the same issue.
Maybe the correct thing to do is to just ignore the change in this case?
Presumably it'd be replicated by tablesync. But we've been unable to
convince ourselves that's correct, or what snapshot to pass to
ReorderBufferQueueMessage/ReorderBufferQueueSequence.
[3]
https://github.com/postgres/postgres/blob/master/src/backend/replication/logical/decode.c#L585
regards
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2022-11-17 16:04 Robert Haas <[email protected]>
parent: Tomas Vondra <[email protected]>
1 sibling, 0 replies; 127+ messages in thread
From: Robert Haas @ 2022-11-17 16:04 UTC (permalink / raw)
To: Tomas Vondra <[email protected]>; +Cc: PostgreSQL Hackers <[email protected]>
On Wed, Nov 16, 2022 at 8:41 PM Tomas Vondra
<[email protected]> wrote:
> There's a couple of caveats, though:
>
> 1) Maybe we should focus more on "actually observed" state instead of
> "observable". Who cares if the sequence moved forward in a transaction
> that was ultimately rolled back? No committed transaction should have
> observer those values - in a way, the last "valid" state of the sequence
> is the last value generated in a transaction that ultimately committed.
When I say "observable" I mean from a separate transaction, not one
that is making changes to things.
I said "observable" rather than "actually observed" because we neither
know nor care whether someone actually ran a SELECT statement at any
given moment in time, just what they would have seen if they did.
> 2) I think what matters more is that we never generate duplicate value.
> That is, if you generate a value from a sequence, commit a transaction
> and replicate it, then the logical standby should not generate the same
> value from the sequence. This guarantee seems necessary for "failover"
> to logical standby.
I think that matters, but I don't think it's sufficient. We need to
preserve the order in which things appear to happen, and which changes
are and are not atomic, not just the final result.
> Well, yeah - we can either try to perform the stuff independently of the
> transactions that triggered it, or we can try making it part of some of
> the transactions. Each of those options has problems, though :-(
>
> The first version of the patch tried the first approach, i.e. decode the
> increments and apply that independently. But:
>
> (a) What would you do with increments of sequences created/reset in a
> transaction? Can't apply those outside the transaction, because it
> might be rolled back (and that state is not visible on primary).
If the state isn't going to be visible until the transaction commits,
it has to be replicated as part of the transaction. If I create a
sequence and then nextval it a bunch of times, I can't replicate that
by first creating the sequence, and then later, as a separate
operation, replicating the nextvals. If I do that, then there's an
intermediate state visible on the replica that was never visible on
the origin server. That's broken.
> (b) What about increments created before we have a proper snapshot?
> There may be transactions dependent on the increment. This is what
> ultimately led to revert of the patch.
Whatever problem exists here is with the implementation, not the
concept. If you copy the initial state as it exists at some moment in
time to a replica, and then replicate all the changes that happen
afterward to that replica without messing up the order, the replica
WILL be in sync with the origin server. The things that happen before
you copy the initial state do not and cannot matter.
But what you're describing sounds like the changes aren't really
replicated in visibility order, and then it is easy to see how a
problem like this can happen. Because now, an operation that actually
became visible just before or just after the initial copy was taken
might be thought to belong on the other side of that boundary, and
then everything will break. And it sounds like that is what you are
describing.
> This version of the patch tries to do the opposite thing - make sure
> that the state after each commit matches what the transaction might have
> seen (for sequences it accessed). It's imperfect, because it might log a
> state generated "after" the sequence got accessed - it focuses on the
> guarantee not to generate duplicate values.
Like Andres, I just can't imagine this being correct. It feels like
it's trying to paper over the failure to do the replication properly
during the transaction by overwriting state at the end.
> Yes, this would mean we accept we may end up with something like this:
>
> 1: T1 logs sequence state S1
> 2: someone increments sequence
> 3: T2 logs sequence stats S2
> 4: T2 commits
> 5: T1 commits
>
> which "inverts" the apply order of S1 vs. S2, because we first apply S2
> and then the "old" S1. But as long as we're smart enough to "discard"
> applying S1, I think that's acceptable - because it guarantees we'll not
> generate duplicate values (with values in the committed transaction).
>
> I'd also argue it does not actually generate invalid state, because once
> we commit either transaction, S2 is what's visible.
I agree that it's OK if the sequence increment gets merged into the
commit that immediately follows. However, I disagree with the idea of
discarding the second update on the grounds that it would make the
sequence go backward and we know that can't be right. That algorithm
works in the really specific case where the only operations are
increments. As soon as anyone does anything else to the sequence, such
an algorithm can no longer work. Nor can it work for objects that are
not sequences. The alternative strategy of replicating each change
exactly once and in the correct order works for all current and future
object types in all cases.
> > Your alternative proposal says "The other option might be to make
> > these messages non-transactional, in which case we'd separate the
> > ordering from COMMIT ordering, evading the reordering problem." But I
> > don't think that avoids the reordering problem at all.
>
> I don't understand why. Why would it not address the reordering issue?
>
> > Nor do I think it's correct.
>
> Nor do I understand this. I mean, isn't it essentially the option you
> mentioned earlier - treating the non-transactional actions as
> independent transactions? Yes, we'd be batching them so that we'd not
> see "intermediate" states, but those are not observed by abyone.
I don't think that batching them is a bad idea, in fact I think it's
necessary. But those batches still have to be applied at the right
time relative to the sequence of commits.
> I'm confused. Isn't that pretty much exactly what I'm proposing? Imagine
> you have something like this:
>
> 1: T1 does something and also increments a sequence
> 2: T1 logs state of the sequence (right before commit)
> 3: T1 writes COMMIT
>
> Now when we decode/apply this, we end up doing this:
>
> 1: decode all T1 changes, stash them
> 2: decode the sequence state and apply it separately
> 3: decode COMMIT, apply all T1 changes
>
> There might be other transactions interleaving with this, but I think
> it'd behave correctly. What example would not work?
What if one of the other transactions renames the sequence, or changes
the current value, or does basically anything to it other than
nextval?
> The problem lies in how we log sequences. If we wrote each individual
> increment to WAL, it might work the way you propose (except for cases
> with sequences created in a transaction, etc.). But that's not what we
> do - we log sequence increments in batches of 32 values, and then only
> modify the sequence relfilenode.
>
> This works for physical replication, because the WAL describes the
> "next" state of the sequence (so if you do "SELECT * FROM sequence"
> you'll not see the same state, and the sequence value may "jump ahead"
> after a failover).
>
> But for logical replication this does not work, because the transaction
> might depend on a state created (WAL-logged) by some other transaction.
> And perhaps that transaction actually happened *before* we even built
> the first snapshot for decoding :-/
I agree that there's a problem here but I don't think that it's a huge
problem. I think that it's not QUITE right to think about what state
is visible on the primary. It's better to think about what state would
be visible on the primary if it crashed and restarted after writing
any given amount of WAL, or what would be visible on a physical
standby after replaying any given amount of WAL. If logical
replication mimics that, I think it's as correct as it needs to be. If
not, those other systems are broken, too.
So I think what should happen is that when we write a WAL record
saying that the sequence has been incremented by 32, that should be
logically replicated after all commits whose commit record precedes
that WAL record and before commits whose commit record follows that
WAL record. It is OK to merge the replication of that record into one
of either the immediately preceding or the immediately following
commit, but you can't do it as part of any other commit because then
you're changing the order of operations.
For instance, consider:
T1: BEGIN; INSERT; COMMIT;
T2: BEGIN; nextval('a_seq') causing a logged advancement to the sequence;
T3: BEGIN; nextval('b_seq') causing a logged advancement to the sequence;
T4: BEGIN; INSERT; COMMIT;
T2: COMMIT;
T3: COMMIT;
The sequence increments can be replicated as part of T1 or part of T4
or in between applying T1 and T4. They cannot be applied as part of T2
or T3. Otherwise, suppose T4 read the current value of one of those
sequences and included that value in the inserted row, and the target
table happened to be the sequence_value_at_end_of_period table. Then
imagine that after receiving the data for T4 and replicating it, the
primary server is hit by a meteor and the replica is promoted. Well,
it's now possible for some new transaction to get a value from that
sequence than what has already been written to the
sequence_value_at_end_of_period table, which will presumably break the
application.
> > In particular, I think it's likely that the "non-transactional
> > messages" that you mention earlier don't get applied at the point in
> > the commit sequence where they were found in the WAL. Not sure why
> > exactly, but perhaps the point at which we're reading WAL runs ahead
> > of the decoding per se, or something like that, and thus those
> > non-transactional messages arrive too early relative to the commit
> > ordering. Possibly that could be changed, and they could be buffered
>
> I'm not sure which case of "non-transactional messages" this refers to,
> so I can't quite respond to these comments. Perhaps you mean the
> problems that killed the previous patch [1]?
In http://postgr.es/m/[email protected]
you said "The other option might be to make these messages
non-transactional". I was referring to that.
> > the transaction whose commit record occurs most nearly prior to, or
> > most nearly after, the WAL record for the operation itself. Or else,
> > we could create "virtual" transactions for such operations and make
> > sure those get replayed at the right point in the commit sequence. Or
> > else, I don't know, maybe something else. But I think the overall
> > picture is that we need to approach the problem by replicating changes
> > in WAL order, as a physical standby would do. Saying that a change is
> > "nontransactional" doesn't mean that it's exempt from ordering
> > requirements; rather, it means that that change has its own place in
> > that ordering, distinct from the transaction in which it occurred.
>
> But doesn't the approach with WAL-logging sequence state before COMMIT,
> and then applying it independently in WAL-order, do pretty much this?
I'm sort of repeating myself here, but: only if the only operations
that ever get performed on sequences are increments. Which is just not
true.
--
Robert Haas
EDB: http://www.enterprisedb.com
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2022-11-17 17:07 Andres Freund <[email protected]>
parent: Tomas Vondra <[email protected]>
0 siblings, 1 reply; 127+ messages in thread
From: Andres Freund @ 2022-11-17 17:07 UTC (permalink / raw)
To: Tomas Vondra <[email protected]>; +Cc: Robert Haas <[email protected]>; PostgreSQL Hackers <[email protected]>
Hi,
On 2022-11-17 12:39:49 +0100, Tomas Vondra wrote:
> On 11/17/22 03:43, Andres Freund wrote:
> > On 2022-11-17 02:41:14 +0100, Tomas Vondra wrote:
> >> Well, yeah - we can either try to perform the stuff independently of the
> >> transactions that triggered it, or we can try making it part of some of
> >> the transactions. Each of those options has problems, though :-(
> >>
> >> The first version of the patch tried the first approach, i.e. decode the
> >> increments and apply that independently. But:
> >>
> >> (a) What would you do with increments of sequences created/reset in a
> >> transaction? Can't apply those outside the transaction, because it
> >> might be rolled back (and that state is not visible on primary).
> >
> > I think a reasonable approach could be to actually perform different WAL
> > logging for that case. It'll require a bit of machinery, but could actually
> > result in *less* WAL logging overall, because we don't need to emit a WAL
> > record for each SEQ_LOG_VALS sequence values.
> >
>
> Could you elaborate? Hard to comment without knowing more ...
>
> My point was that stuff like this (creating a new sequence or at least a
> new relfilenode) means we can't apply that independently of the
> transaction (unlike regular increments). I'm not sure how a change to
> WAL logging would make that go away.
Different WAL logging would make it easy to handle that on the logical
decoding level. We don't need to emit WAL records each time a
created-in-this-toplevel-xact sequences gets incremented as they're not
persisting anyway if the surrounding xact aborts. We already need to remember
the filenode so it can be dropped at the end of the transaction, so we could
emit a single record for each sequence at that point.
> >> (b) What about increments created before we have a proper snapshot?
> >> There may be transactions dependent on the increment. This is what
> >> ultimately led to revert of the patch.
> >
> > I don't understand this - why would we ever need to process those increments
> > from before we have a snapshot? Wouldn't they, by definition, be before the
> > slot was active?
> >
> > To me this is the rough equivalent of logical decoding not giving the initial
> > state of all tables. You need some process outside of logical decoding to get
> > that (obviously we have some support for that via the exported data snapshot
> > during slot creation).
> >
>
> Which is what already happens during tablesync, no? We more or less copy
> sequences as if they were tables.
I think you might have to copy sequences after tables, but I'm not sure. But
otherwise, yea.
> > I assume that part of the initial sync would have to be a new sequence
> > synchronization step that reads all the sequence states on the publisher and
> > ensures that the subscriber sequences are at the same point. There's a bit of
> > trickiness there, but it seems entirely doable. The logical replication replay
> > support for sequences will have to be a bit careful about not decreasing the
> > subscriber's sequence values - the standby initially will be ahead of the
> > increments we'll see in the WAL. But that seems inevitable given the
> > non-transactional nature of sequences.
> >
>
> See fetch_sequence_data / copy_sequence in the patch. The bit about
> ensuring the sequence does not go away (say, using page LSN and/or LSN
> of the increment) is not there, however isn't that pretty much what I
> proposed doing for "reconciling" the sequence state logged at COMMIT?
Well, I think the approach of logging all sequence increments at commit is the
wrong idea...
Creating a new relfilenode whenever a sequence is incremented seems like a
complete no-go to me. That increases sequence overhead by several orders of
magnitude and will lead to *awful* catalog bloat on the subscriber.
> >
> >> This version of the patch tries to do the opposite thing - make sure
> >> that the state after each commit matches what the transaction might have
> >> seen (for sequences it accessed). It's imperfect, because it might log a
> >> state generated "after" the sequence got accessed - it focuses on the
> >> guarantee not to generate duplicate values.
> >
> > That approach seems quite wrong to me.
> >
>
> Why? Because it might log a state for sequence as of COMMIT, when the
> transaction accessed the sequence much earlier?
Mainly because sequences aren't transactional and trying to make them will
require awful contortions.
While there are cases where we don't flush the WAL / wait for syncrep for
sequences, we do replicate their state correctly on physical replication. If
an LSN has been acknowledged as having been replicated, we won't just loose a
prior sequence increment after promotion, even if the transaction didn't [yet]
commit.
It's completely valid for an application to call nextval() in one transaction,
potentially even abort it, and then only use that sequence value in another
transaction.
> > I did some skimming of the referenced thread about the reversal of the last
> > approach, but I couldn't really understand what the fundamental issues were
> > with the reverted implementation - it's a very long thread and references
> > other threads.
> >
>
> Yes, it's long/complex, but I intentionally linked to a specific message
> which describes the issue ...
>
> It's entirely possible there is a simple fix for the issue, and I just
> got confused / unable to see the solution. The whole issue was due to
> having a mix of transactional and non-transactional cases, similarly to
> logical messages - and logicalmsg_decode() has the same issue, so maybe
> let's talk about that for a moment.
>
> See [3] and imagine you're dealing with a transactional message, but
> you're still building a consistent snapshot. So the first branch applies:
>
> if (transactional &&
> !SnapBuildProcessChange(builder, xid, buf->origptr))
> return;
>
> but because we don't have a snapshot, SnapBuildProcessChange does this:
>
> if (builder->state < SNAPBUILD_FULL_SNAPSHOT)
> return false;
In this case we'd just return without further work in logicalmsg_decode(). The
problematic case presumably is is when we have a full snapshot but aren't yet
consistent, but xid is >= next_phase_at. Then SnapBuildProcessChange() returns
true. And we reach:
> which however means logicalmsg_decode() does
>
> snapshot = SnapBuildGetOrBuildSnapshot(builder);
>
> which crashes, because it hits this assert:
>
> Assert(builder->state == SNAPBUILD_CONSISTENT);
I think the problem here is just that we shouldn't even try to get a snapshot
in the transactional case - note that it's not even used in
ReorderBufferQueueMessage() for transactional message. The transactional case
needs to behave like a "normal" change - we might never decode the message if
the transaction ends up committing before we've reached a consistent point.
> The sequence decoding did almost the same thing, with the same issue.
> Maybe the correct thing to do is to just ignore the change in this case?
No, I don't think that'd be correct, the message | sequence needs to be queued
for the transaction. If the transaction ends up committing after we've reached
consistency, we'll get the correct snapshot from the base snapshot set in
SnapBuildProcessChange().
Greetings,
Andres Freund
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2022-11-17 21:13 Tomas Vondra <[email protected]>
parent: Andres Freund <[email protected]>
0 siblings, 1 reply; 127+ messages in thread
From: Tomas Vondra @ 2022-11-17 21:13 UTC (permalink / raw)
To: Andres Freund <[email protected]>; +Cc: Robert Haas <[email protected]>; PostgreSQL Hackers <[email protected]>
On 11/17/22 18:07, Andres Freund wrote:
> Hi,
>
> On 2022-11-17 12:39:49 +0100, Tomas Vondra wrote:
>> On 11/17/22 03:43, Andres Freund wrote:
>>> On 2022-11-17 02:41:14 +0100, Tomas Vondra wrote:
>>>> Well, yeah - we can either try to perform the stuff independently of the
>>>> transactions that triggered it, or we can try making it part of some of
>>>> the transactions. Each of those options has problems, though :-(
>>>>
>>>> The first version of the patch tried the first approach, i.e. decode the
>>>> increments and apply that independently. But:
>>>>
>>>> (a) What would you do with increments of sequences created/reset in a
>>>> transaction? Can't apply those outside the transaction, because it
>>>> might be rolled back (and that state is not visible on primary).
>>>
>>> I think a reasonable approach could be to actually perform different WAL
>>> logging for that case. It'll require a bit of machinery, but could actually
>>> result in *less* WAL logging overall, because we don't need to emit a WAL
>>> record for each SEQ_LOG_VALS sequence values.
>>>
>>
>> Could you elaborate? Hard to comment without knowing more ...
>>
>> My point was that stuff like this (creating a new sequence or at least a
>> new relfilenode) means we can't apply that independently of the
>> transaction (unlike regular increments). I'm not sure how a change to
>> WAL logging would make that go away.
>
> Different WAL logging would make it easy to handle that on the logical
> decoding level. We don't need to emit WAL records each time a
> created-in-this-toplevel-xact sequences gets incremented as they're not
> persisting anyway if the surrounding xact aborts. We already need to remember
> the filenode so it can be dropped at the end of the transaction, so we could
> emit a single record for each sequence at that point.
>
>
>>>> (b) What about increments created before we have a proper snapshot?
>>>> There may be transactions dependent on the increment. This is what
>>>> ultimately led to revert of the patch.
>>>
>>> I don't understand this - why would we ever need to process those increments
>>> from before we have a snapshot? Wouldn't they, by definition, be before the
>>> slot was active?
>>>
>>> To me this is the rough equivalent of logical decoding not giving the initial
>>> state of all tables. You need some process outside of logical decoding to get
>>> that (obviously we have some support for that via the exported data snapshot
>>> during slot creation).
>>>
>>
>> Which is what already happens during tablesync, no? We more or less copy
>> sequences as if they were tables.
>
> I think you might have to copy sequences after tables, but I'm not sure. But
> otherwise, yea.
>
>
>>> I assume that part of the initial sync would have to be a new sequence
>>> synchronization step that reads all the sequence states on the publisher and
>>> ensures that the subscriber sequences are at the same point. There's a bit of
>>> trickiness there, but it seems entirely doable. The logical replication replay
>>> support for sequences will have to be a bit careful about not decreasing the
>>> subscriber's sequence values - the standby initially will be ahead of the
>>> increments we'll see in the WAL. But that seems inevitable given the
>>> non-transactional nature of sequences.
>>>
>>
>> See fetch_sequence_data / copy_sequence in the patch. The bit about
>> ensuring the sequence does not go away (say, using page LSN and/or LSN
>> of the increment) is not there, however isn't that pretty much what I
>> proposed doing for "reconciling" the sequence state logged at COMMIT?
>
> Well, I think the approach of logging all sequence increments at commit is the
> wrong idea...
>
But we're not logging all sequence increments, no?
We're logging the state for each sequence touched by the transaction,
but only once - if the transaction incremented the sequence 1000000x
times, we'll still log it just once (at least for this particular purpose).
Yes, if transactions touch each sequence just once, then we're logging
individual increments.
The only more efficient solution would be to decode the existing WAL
(every ~32 increments), and perhaps also tracking which sequences were
accessed by a transaction. And then simply stashing the increments in a
global reorderbuffer hash table, and then applying only the last one at
commit time. This would require the transactional / non-transactional
behavior (I think), but perhaps we can make that work.
Or are you thinking about some other scheme?
> Creating a new relfilenode whenever a sequence is incremented seems like a
> complete no-go to me. That increases sequence overhead by several orders of
> magnitude and will lead to *awful* catalog bloat on the subscriber.
>
You mean on the the apply side? Yes, I agree this needs a better
approach, I've focused on the decoding side so far.
>
>>>
>>>> This version of the patch tries to do the opposite thing - make sure
>>>> that the state after each commit matches what the transaction might have
>>>> seen (for sequences it accessed). It's imperfect, because it might log a
>>>> state generated "after" the sequence got accessed - it focuses on the
>>>> guarantee not to generate duplicate values.
>>>
>>> That approach seems quite wrong to me.
>>>
>>
>> Why? Because it might log a state for sequence as of COMMIT, when the
>> transaction accessed the sequence much earlier?
>
> Mainly because sequences aren't transactional and trying to make them will
> require awful contortions.
>
> While there are cases where we don't flush the WAL / wait for syncrep for
> sequences, we do replicate their state correctly on physical replication. If
> an LSN has been acknowledged as having been replicated, we won't just loose a
> prior sequence increment after promotion, even if the transaction didn't [yet]
> commit.
>
True, I agree we should aim to achieve that.
> It's completely valid for an application to call nextval() in one transaction,
> potentially even abort it, and then only use that sequence value in another
> transaction.
>
I don't quite agree with that - we make no promises about what happens
to sequence changes in aborted transactions. I don't think I've ever
seen an application using such pattern either.
And I'd argue we already fail to uphold such guarantee, because we don't
wait for syncrep if the sequence WAL happened in aborted transaction. So
if you use the value elsewhere (outside PG), you may lose it.
Anyway, I think the scheme I outlined above (with stashing decoded
increments, logged once every 32 values) and applying the latest
increment for all sequences at commit, would work.
>
>
>>> I did some skimming of the referenced thread about the reversal of the last
>>> approach, but I couldn't really understand what the fundamental issues were
>>> with the reverted implementation - it's a very long thread and references
>>> other threads.
>>>
>>
>> Yes, it's long/complex, but I intentionally linked to a specific message
>> which describes the issue ...
>>
>> It's entirely possible there is a simple fix for the issue, and I just
>> got confused / unable to see the solution. The whole issue was due to
>> having a mix of transactional and non-transactional cases, similarly to
>> logical messages - and logicalmsg_decode() has the same issue, so maybe
>> let's talk about that for a moment.
>>
>> See [3] and imagine you're dealing with a transactional message, but
>> you're still building a consistent snapshot. So the first branch applies:
>>
>> if (transactional &&
>> !SnapBuildProcessChange(builder, xid, buf->origptr))
>> return;
>>
>> but because we don't have a snapshot, SnapBuildProcessChange does this:
>>
>> if (builder->state < SNAPBUILD_FULL_SNAPSHOT)
>> return false;
>
> In this case we'd just return without further work in logicalmsg_decode(). The
> problematic case presumably is is when we have a full snapshot but aren't yet
> consistent, but xid is >= next_phase_at. Then SnapBuildProcessChange() returns
> true. And we reach:
>
>> which however means logicalmsg_decode() does
>>
>> snapshot = SnapBuildGetOrBuildSnapshot(builder);
>>
>> which crashes, because it hits this assert:
>>
>> Assert(builder->state == SNAPBUILD_CONSISTENT);
>
> I think the problem here is just that we shouldn't even try to get a snapshot
> in the transactional case - note that it's not even used in
> ReorderBufferQueueMessage() for transactional message. The transactional case
> needs to behave like a "normal" change - we might never decode the message if
> the transaction ends up committing before we've reached a consistent point.
>
>
>> The sequence decoding did almost the same thing, with the same issue.
>> Maybe the correct thing to do is to just ignore the change in this case?
>
> No, I don't think that'd be correct, the message | sequence needs to be queued
> for the transaction. If the transaction ends up committing after we've reached
> consistency, we'll get the correct snapshot from the base snapshot set in
> SnapBuildProcessChange().
>
Yeah, I think you're right. I looked at this again, with fresh mind, and
I came to the same conclusion. Roughly what the attached patch does.
regards
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
Attachments:
[text/x-patch] logical-msg-fix.patch (2.7K, ../../[email protected]/2-logical-msg-fix.patch)
download | inline diff:
diff --git a/src/backend/replication/logical/decode.c b/src/backend/replication/logical/decode.c
index 2cc0ac9eb09..66771ac35a2 100644
--- a/src/backend/replication/logical/decode.c
+++ b/src/backend/replication/logical/decode.c
@@ -564,7 +564,6 @@ logicalmsg_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
TransactionId xid = XLogRecGetXid(r);
uint8 info = XLogRecGetInfo(r) & ~XLR_INFO_MASK;
RepOriginId origin_id = XLogRecGetOrigin(r);
- Snapshot snapshot;
xl_logical_message *message;
if (info != XLOG_LOGICAL_MESSAGE)
@@ -594,8 +593,7 @@ logicalmsg_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
SnapBuildXactNeedsSkip(builder, buf->origptr)))
return;
- snapshot = SnapBuildGetOrBuildSnapshot(builder);
- ReorderBufferQueueMessage(ctx->reorder, xid, snapshot, buf->endptr,
+ ReorderBufferQueueMessage(ctx->reorder, xid, buf->endptr,
message->transactional,
message->message, /* first part of message is
* prefix */
diff --git a/src/backend/replication/logical/reorderbuffer.c b/src/backend/replication/logical/reorderbuffer.c
index 31f7381f2d6..0a34aebdd58 100644
--- a/src/backend/replication/logical/reorderbuffer.c
+++ b/src/backend/replication/logical/reorderbuffer.c
@@ -818,7 +818,7 @@ ReorderBufferQueueChange(ReorderBuffer *rb, TransactionId xid, XLogRecPtr lsn,
*/
void
ReorderBufferQueueMessage(ReorderBuffer *rb, TransactionId xid,
- Snapshot snap, XLogRecPtr lsn,
+ XLogRecPtr lsn,
bool transactional, const char *prefix,
Size message_size, const char *message)
{
@@ -845,7 +845,11 @@ ReorderBufferQueueMessage(ReorderBuffer *rb, TransactionId xid,
else
{
ReorderBufferTXN *txn = NULL;
- volatile Snapshot snapshot_now = snap;
+
+ LogicalDecodingContext *ctx = rb->private_data;
+ SnapBuild *builder = ctx->snapshot_builder;
+
+ volatile Snapshot snapshot_now = SnapBuildGetOrBuildSnapshot(builder);
if (xid != InvalidTransactionId)
txn = ReorderBufferTXNByXid(rb, xid, true, NULL, lsn, true);
diff --git a/src/include/replication/reorderbuffer.h b/src/include/replication/reorderbuffer.h
index b23d8cc4f9f..50b48d52b5a 100644
--- a/src/include/replication/reorderbuffer.h
+++ b/src/include/replication/reorderbuffer.h
@@ -646,7 +646,7 @@ extern void ReorderBufferQueueChange(ReorderBuffer *rb, TransactionId xid,
XLogRecPtr lsn, ReorderBufferChange *change,
bool toast_insert);
extern void ReorderBufferQueueMessage(ReorderBuffer *rb, TransactionId xid,
- Snapshot snap, XLogRecPtr lsn,
+ XLogRecPtr lsn,
bool transactional, const char *prefix,
Size message_size, const char *message);
extern void ReorderBufferCommit(ReorderBuffer *rb, TransactionId xid,
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2022-11-18 03:03 Andres Freund <[email protected]>
parent: Tomas Vondra <[email protected]>
0 siblings, 0 replies; 127+ messages in thread
From: Andres Freund @ 2022-11-18 03:03 UTC (permalink / raw)
To: Tomas Vondra <[email protected]>; +Cc: Robert Haas <[email protected]>; PostgreSQL Hackers <[email protected]>
Hi,
On 2022-11-17 22:13:23 +0100, Tomas Vondra wrote:
> On 11/17/22 18:07, Andres Freund wrote:
> > On 2022-11-17 12:39:49 +0100, Tomas Vondra wrote:
> >> On 11/17/22 03:43, Andres Freund wrote:
> >>> I assume that part of the initial sync would have to be a new sequence
> >>> synchronization step that reads all the sequence states on the publisher and
> >>> ensures that the subscriber sequences are at the same point. There's a bit of
> >>> trickiness there, but it seems entirely doable. The logical replication replay
> >>> support for sequences will have to be a bit careful about not decreasing the
> >>> subscriber's sequence values - the standby initially will be ahead of the
> >>> increments we'll see in the WAL. But that seems inevitable given the
> >>> non-transactional nature of sequences.
> >>>
> >>
> >> See fetch_sequence_data / copy_sequence in the patch. The bit about
> >> ensuring the sequence does not go away (say, using page LSN and/or LSN
> >> of the increment) is not there, however isn't that pretty much what I
> >> proposed doing for "reconciling" the sequence state logged at COMMIT?
> >
> > Well, I think the approach of logging all sequence increments at commit is the
> > wrong idea...
> >
>
> But we're not logging all sequence increments, no?
I was imprecise - I meant streaming them out at commit.
> Yeah, I think you're right. I looked at this again, with fresh mind, and
> I came to the same conclusion. Roughly what the attached patch does.
To me it seems a bit nicer to keep the SnapBuildGetOrBuildSnapshot() call in
decode.c instead of moving it to reorderbuffer.c. Perhaps we should add a
snapbuild.c helper similar to SnapBuildProcessChange() for non-transactional
changes that also gets a snapshot?
Could look something like
Snapshot snapshot = NULL;
if (message->transactional &&
!SnapBuildProcessChange(builder, xid, buf->origptr))
return;
else if (!SnapBuildProcessStateNonTx(builder, &snapshot))
return;
...
Or perhaps we should just bite the bullet and add an argument to
SnapBuildProcessChange to deal with that?
Greetings,
Andres Freund
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2023-11-29 14:41 Tomas Vondra <[email protected]>
parent: Tomas Vondra <[email protected]>
1 sibling, 1 reply; 127+ messages in thread
From: Tomas Vondra @ 2023-11-29 14:41 UTC (permalink / raw)
To: Amit Kapila <[email protected]>; +Cc: Hayato Kuroda (Fujitsu) <[email protected]>; Zhijie Hou (Fujitsu) <[email protected]>; Ashutosh Bapat <[email protected]>; PostgreSQL Hackers <[email protected]>; Masahiko Sawada <[email protected]>; Peter Eisentraut <[email protected]>; Dilip Kumar <[email protected]>
On 11/29/23 14:42, Amit Kapila wrote:
> On Wed, Nov 29, 2023 at 2:59 AM Tomas Vondra
> <[email protected]> wrote:
>>
>> I have been hacking on improving the improvements outlined in my
>> preceding e-mail, but I have some bad news - I ran into an issue that I
>> don't know how to solve :-(
>>
>> Consider this transaction:
>>
>> BEGIN;
>> ALTER SEQUENCE s RESTART 1000;
>>
>> SAVEPOINT s1;
>> ALTER SEQUENCE s RESTART 2000;
>> ROLLBACK TO s1;
>>
>> INSERT INTO seq_test SELECT nextval('s') FROM generate_series(1,40);
>> COMMIT;
>>
>> If you try this with the approach relying on rd_newRelfilelocatorSubid
>> and rd_createSubid, it fails like this on the subscriber:
>>
>> ERROR: could not map filenode "base/5/16394" to relation OID
>>
>> This happens because ReorderBufferQueueSequence tries to do this in the
>> non-transactional branch:
>>
>> reloid = RelidByRelfilenumber(rlocator.spcOid, rlocator.relNumber);
>>
>> and the relfilenode is the one created by the first ALTER. But this is
>> obviously wrong - the changes should have been treated as transactional,
>> because they are tied to the first ALTER. So how did we get there?
>>
>> Well, the whole problem is that in case of abort, AtEOSubXact_cleanup
>> resets the two fields to InvalidSubTransactionId. Which means the
>> rollback in the above transaction also forgets about the first ALTER.
>> Now that I look at the RelationData comments, it actually describes
>> exactly this situation:
>>
>> *
>> * rd_newRelfilelocatorSubid is the ID of the highest subtransaction
>> * the most-recent relfilenumber change has survived into or zero if
>> * not changed in the current transaction (or we have forgotten
>> * changing it). This field is accurate when non-zero, but it can be
>> * zero when a relation has multiple new relfilenumbers within a
>> * single transaction, with one of them occurring in a subsequently
>> * aborted subtransaction, e.g.
>> * BEGIN;
>> * TRUNCATE t;
>> * SAVEPOINT save;
>> * TRUNCATE t;
>> * ROLLBACK TO save;
>> * -- rd_newRelfilelocatorSubid is now forgotten
>> *
>>
>> The root of this problem is that we'd need some sort of "history" for
>> the field, so that when a subxact aborts, we can restore the previous
>> value. But we obviously don't have that, and I doubt we want to add that
>> to relcache - for example, it'd either need to impose some limit on the
>> history (and thus a failure when we reach the limit), or it'd need to
>> handle histories of arbitrary length.
>>
>
> Yeah, I think that would be really tricky and we may not want to go there.
>
>> At this point I don't see a solution for this, which means the best way
>> forward with the sequence decoding patch seems to be the original
>> approach, on the decoding side.
>>
>
> One thing that worries me about that approach is that it can suck with
> the workload that has a lot of DDLs that create XLOG_SMGR_CREATE
> records. We have previously fixed some such workloads in logical
> decoding where decoding a transaction containing truncation of a table
> with a lot of partitions (1000 or more) used to take a very long time.
> Don't we face performance issues in such scenarios?
>
I don't think we do, really. We will have to decode the SMGR records and
add the relfilenodes to the hash table(s), but I think that affects the
lookup performance too much. What I think might be a problem is if we
have many top-level transactions, especially if those transactions do
something that creates a relfilenode. Because then we'll have to do a
hash_search for each of them, and that might be measurable even if each
lookup is O(1). And we do the lookup for every sequence change ...
> How do we see this work w.r.t to some sort of global sequences? There
> is some recent discussion where I have raised a similar point [1].
>
> [1] - https://www.postgresql.org/message-id/CAA4eK1JF%3D4_Eoq7FFjHSe98-_ooJ5QWd0s2_pj8gR%2B_dvwKxvA%40mail...
>
I think those are very different things, even though called "sequences".
AFAIK solutions like snowflakeID or UUIDs don't require replication of
any shared state (that's kinda the whole point), so I don't see why
would it need some special support in logical decoding.
regards
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2023-11-29 23:58 Tomas Vondra <[email protected]>
parent: Tomas Vondra <[email protected]>
0 siblings, 2 replies; 127+ messages in thread
From: Tomas Vondra @ 2023-11-29 23:58 UTC (permalink / raw)
To: Amit Kapila <[email protected]>; +Cc: Hayato Kuroda (Fujitsu) <[email protected]>; Zhijie Hou (Fujitsu) <[email protected]>; Ashutosh Bapat <[email protected]>; PostgreSQL Hackers <[email protected]>; Masahiko Sawada <[email protected]>; Peter Eisentraut <[email protected]>; Dilip Kumar <[email protected]>
On 11/29/23 15:41, Tomas Vondra wrote:
> ...
>>
>> One thing that worries me about that approach is that it can suck with
>> the workload that has a lot of DDLs that create XLOG_SMGR_CREATE
>> records. We have previously fixed some such workloads in logical
>> decoding where decoding a transaction containing truncation of a table
>> with a lot of partitions (1000 or more) used to take a very long time.
>> Don't we face performance issues in such scenarios?
>>
>
> I don't think we do, really. We will have to decode the SMGR records and
> add the relfilenodes to the hash table(s), but I think that affects the
> lookup performance too much. What I think might be a problem is if we
> have many top-level transactions, especially if those transactions do
> something that creates a relfilenode. Because then we'll have to do a
> hash_search for each of them, and that might be measurable even if each
> lookup is O(1). And we do the lookup for every sequence change ...
>
I did some micro-benchmarking today, trying to identify cases where this
would cause unexpected problems, either due to having to maintain all
the relfilenodes, or due to having to do hash lookups for every sequence
change. But I think it's fine, mostly ...
I did all the following tests with 64 clients. I may try more, but even
with this there should be fair number of concurrent transactions, which
determines the number of top-level transactions in reorderbuffer. I'll
try with more clients tomorrow, but I don't think it'll change stuff.
The test is fairly simple - run a particular number of transactions
(might be 1000 * 64, or more). And then measure how long it takes to
decode the changes using test_decoding.
Now, the various workloads I tried:
1) "good case" - small OLTP transactions, a couple nextval('s') calls
begin;
insert into t (1);
select nextval('s');
insert into t (1);
commit;
This is pretty fine, the sequence part of reorderbuffer is really not
measurable, it's like 1% of the total CPU time. Which is expected,
because we only wal-log every 32-nd increment or so.
2) "good case" - same as (1) but more nextval calls to always do wal
begin;
insert into t (1);
select nextval('s') from generate_series(1,40);
insert into t (1);
commit;
Here sequences are more measurable, it's like 15% of CPU time, but most
of that comes to AbortCurrentTransaction() in the non-transactional
branch of ReorderBufferQueueSequence. I don't think there's a way around
that, and it's entirely unrelated to relfilenodes. The function checking
if the change is transactional (ReorderBufferSequenceIsTransactional) is
less than 1% of the profile - and this is the version that always walks
all top-level transactions.
3) "bad case" - small transactions that generate a lot of relfilenodes
select alter_sequence();
where the function is defined like this (I did create 1000 sequences
before the test):
CREATE OR REPLACE FUNCTION alter_sequence() RETURNS void AS $$
DECLARE
v INT;
BEGIN
v := 1 + (random() * 999)::int;
execute format('alter sequence s%s restart with 1000', v);
perform nextval('s');
END;
$$ LANGUAGE plpgsql;
This performs terribly, but it's entirely unrelated to sequences.
Current master has exactly the same problem, if transactions do DDL.
Like this, for example:
CREATE OR REPLACE FUNCTION create_table() RETURNS void AS $$
DECLARE
v INT;
BEGIN
v := 1 + (random() * 999)::int;
execute format('create table t%s (a int)', v);
execute format('drop table t%s', v);
insert into t values (1);
END;
$$ LANGUAGE plpgsql;
This has the same impact on master. The perf report shows this:
--98.06%--pg_logical_slot_get_changes_guts
|
--97.88%--LogicalDecodingProcessRecord
|
--97.56%--xact_decode
|
--97.51%--DecodeCommit
|
|--91.92%--SnapBuildCommitTxn
| |
| --91.65%--SnapBuildBuildSnapshot
| |
| --91.14%--pg_qsort
The sequence decoding is maybe ~1%. The reason why SnapBuildSnapshot
takes so long is because:
-----------------
Breakpoint 1, SnapBuildBuildSnapshot (builder=0x21f60f8)
at snapbuild.c:498
498 + sizeof(TransactionId) * builder->committed.xcnt
(gdb) p builder->committed.xcnt
$4 = 11532
-----------------
And with each iteration it grows by 1. That looks quite weird, possibly
a bug worth fixing, but unrelated to this patch. I can't investigate
this more at the moment, not sure when/if I'll get to that.
regards
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2023-11-30 11:56 Amit Kapila <[email protected]>
parent: Tomas Vondra <[email protected]>
1 sibling, 1 reply; 127+ messages in thread
From: Amit Kapila @ 2023-11-30 11:56 UTC (permalink / raw)
To: Tomas Vondra <[email protected]>; +Cc: Hayato Kuroda (Fujitsu) <[email protected]>; Zhijie Hou (Fujitsu) <[email protected]>; Ashutosh Bapat <[email protected]>; PostgreSQL Hackers <[email protected]>; Masahiko Sawada <[email protected]>; Peter Eisentraut <[email protected]>; Dilip Kumar <[email protected]>
On Thu, Nov 30, 2023 at 5:28 AM Tomas Vondra
<[email protected]> wrote:
>
> 3) "bad case" - small transactions that generate a lot of relfilenodes
>
> select alter_sequence();
>
> where the function is defined like this (I did create 1000 sequences
> before the test):
>
> CREATE OR REPLACE FUNCTION alter_sequence() RETURNS void AS $$
> DECLARE
> v INT;
> BEGIN
> v := 1 + (random() * 999)::int;
> execute format('alter sequence s%s restart with 1000', v);
> perform nextval('s');
> END;
> $$ LANGUAGE plpgsql;
>
> This performs terribly, but it's entirely unrelated to sequences.
> Current master has exactly the same problem, if transactions do DDL.
> Like this, for example:
>
> CREATE OR REPLACE FUNCTION create_table() RETURNS void AS $$
> DECLARE
> v INT;
> BEGIN
> v := 1 + (random() * 999)::int;
> execute format('create table t%s (a int)', v);
> execute format('drop table t%s', v);
> insert into t values (1);
> END;
> $$ LANGUAGE plpgsql;
>
> This has the same impact on master. The perf report shows this:
>
> --98.06%--pg_logical_slot_get_changes_guts
> |
> --97.88%--LogicalDecodingProcessRecord
> |
> --97.56%--xact_decode
> |
> --97.51%--DecodeCommit
> |
> |--91.92%--SnapBuildCommitTxn
> | |
> | --91.65%--SnapBuildBuildSnapshot
> | |
> | --91.14%--pg_qsort
>
> The sequence decoding is maybe ~1%. The reason why SnapBuildSnapshot
> takes so long is because:
>
> -----------------
> Breakpoint 1, SnapBuildBuildSnapshot (builder=0x21f60f8)
> at snapbuild.c:498
> 498 + sizeof(TransactionId) * builder->committed.xcnt
> (gdb) p builder->committed.xcnt
> $4 = 11532
> -----------------
>
> And with each iteration it grows by 1.
>
Can we somehow avoid this either by keeping DDL-related xacts open or
aborting them? Also, will it make any difference to use setval as
do_setval() seems to be logging each time?
If possible, can you share the scripts? Kuroda-San has access to the
performance machine, he may be able to try it as well.
--
With Regards,
Amit Kapila.
^ permalink raw reply [nested|flat] 127+ messages in thread
* RE: logical decoding and replication of sequences, take 2
@ 2023-12-01 11:08 Hayato Kuroda (Fujitsu) <[email protected]>
parent: Tomas Vondra <[email protected]>
1 sibling, 1 reply; 127+ messages in thread
From: Hayato Kuroda (Fujitsu) @ 2023-12-01 11:08 UTC (permalink / raw)
To: 'Tomas Vondra' <[email protected]>; Amit Kapila <[email protected]>; +Cc: Zhijie Hou (Fujitsu) <[email protected]>; Ashutosh Bapat <[email protected]>; PostgreSQL Hackers <[email protected]>; Masahiko Sawada <[email protected]>; Peter Eisentraut <[email protected]>; Dilip Kumar <[email protected]>
Dear Tomas,
> I did some micro-benchmarking today, trying to identify cases where this
> would cause unexpected problems, either due to having to maintain all
> the relfilenodes, or due to having to do hash lookups for every sequence
> change. But I think it's fine, mostly ...
>
I did also performance tests (especially case 3). First of all, there are some
variants from yours.
1. patch 0002 was reverted because it has an issue. So this test checks whether
refactoring around ReorderBufferSequenceIsTransactional seems really needed.
2. per comments from Amit, I also measured the abort case. In this case, the
alter_sequence() is called but the transaction is aborted.
3. I measured with changing number of clients {8, 16, 32, 64, 128}. In any cases,
clients executed 1000 transactions. The performance machine has 128 core so that
result for 128 clients might be saturated.
4. a short sleep (0.1s) was added in alter_sequence(), especially between
"alter sequence" and nextval(). Because while testing, I found that the
transaction is too short to execute in parallel. I think it is reasonable
because ReorderBufferSequenceIsTransactional() might be worse when the parallelism
is increased.
I attached one backend process via perf and executed pg_slot_logical_get_changes().
Attached txt file shows which function occupied CPU time, especially from
pg_logical_slot_get_changes_guts() and ReorderBufferSequenceIsTransactional().
Here are my observations about them.
* In case of commit, as you said, SnapBuildCommitTxn() seems dominant for 8-64
clients case.
* For (commit, 128 clients) case, however, ReorderBufferRestoreChanges() waste
many times. I think this is because changes exceed logical_decoding_work_mem,
so we do not have to analyze anymore.
* In case of abort, CPU time used by ReorderBufferSequenceIsTransactional() is linearly
longer. This means that we need to think some solution to avoid the overhead by
ReorderBufferSequenceIsTransactional().
```
8 clients 3.73% occupied time
16 7.26%
32 15.82%
64 29.14%
128 46.27%
```
* In case of abort, I also checked CPU time used by ReorderBufferAddRelFileLocator(), but
it seems not so depends on the number of clients.
```
8 clients 3.66% occupied time
16 6.94%
32 4.65%
64 5.39%
128 3.06%
```
As next step, I've planned to run the case which uses setval() function, because it
generates more WALs than normal nextval();
How do you think?
Best Regards,
Hayato Kuroda
FUJITSU LIMITED
# commit case
## 128 clients
- 97.77% 0.00% postgres postgres [.] pg_logical_slot_get_changes_guts
- 97.77% pg_logical_slot_get_changes_guts
- 97.77% LogicalDecodingProcessRecord
- 97.71% xact_decode
- 54.95% ReorderBufferProcessTXN
- 32.64% ReorderBufferRestoreChanges
+ 23.37% FileRead
4.58% __memcpy_ssse3_back
+ 4.14% MemoryContextAllocZero
+ 7.73% ReorderBufferCopySnap.isra.21
+ 5.94% ReorderBufferSerializeTXN
+ 5.78% ReorderBufferCleanupTXN
+ 2.04% RelationIdGetRelation
- 42.75% SnapBuildCommitTxn
- 34.56% ReorderBufferQueueChange
- 34.36% ReorderBufferSerializeTXN
27.64% __write_nocancel
4.42% __memcpy_ssse3_back
+ 1.52% OpenTransientFilePerm
+ 7.97% SnapBuildBuildSnapshot
0.04% 0.00% postgres postgres [.] ReorderBufferSequenceIsTransactional
## 64 clients
- 86.49% 0.04% postgres postgres [.] pg_logical_slot_get_changes_guts
- 86.45% pg_logical_slot_get_changes_guts
- 86.37% LogicalDecodingProcessRecord
- 84.79% xact_decode
- 51.05% SnapBuildCommitTxn
+ 49.77% SnapBuildBuildSnapshot
0.53% ReorderBufferXidHasBaseSnapshot
- 33.45% ReorderBufferProcessTXN
+ 21.37% ReorderBufferCopySnap.isra.21
+ 4.86% RelationIdGetRelation
+ 2.31% RelidByRelfilenumber
+ 0.84% AbortCurrentTransaction
0.81% ReorderBufferCleanupTXN
- 1.10% seq_decode
0.65% ReorderBufferSequenceIsTransactional
- 1.04% 0.09% postgres postgres [.] ReorderBufferSequenceIsTransactional
## 32 clients
- 82.20% 0.13% postgres postgres [.] pg_logical_slot_get_changes_guts
- 82.09% pg_logical_slot_get_changes_guts
- 81.86% LogicalDecodingProcessRecord
- 80.01% xact_decode
- 49.38% SnapBuildCommitTxn
+ 48.10% SnapBuildBuildSnapshot
0.50% ReorderBufferXidHasBaseSnapshot
- 29.98% ReorderBufferProcessTXN
+ 10.52% ReorderBufferCopySnap.isra.21
+ 7.64% RelationIdGetRelation
+ 4.11% RelidByRelfilenumber
+ 1.70% AbortCurrentTransaction
1.01% ReorderBufferCleanupTXN
0.92% LocalExecuteInvalidationMessage
0.87% commit_cb_wrapper
- 1.01% seq_decode
0.56% ReorderBufferSequenceIsTransactional
+ 0.85% 0.08% postgres postgres [.] ReorderBufferSequenceIsTransactional
## 16 clients
- 78.30% 0.10% postgres postgres [.] pg_logical_slot_get_changes_guts
- 78.24% pg_logical_slot_get_changes_guts
- 77.89% LogicalDecodingProcessRecord
- 75.11% xact_decode
- 38.84% SnapBuildCommitTxn
- 37.54% SnapBuildBuildSnapshot
+ 37.09% pg_qsort
0.56% ReorderBufferXidHasBaseSnapshot
- 35.34% ReorderBufferProcessTXN
+ 11.17% RelationIdGetRelation
+ 7.05% RelidByRelfilenumber
+ 6.33% ReorderBufferCopySnap.isra.21
+ 2.78% AbortCurrentTransaction
+ 1.27% commit_cb_wrapper
+ 1.22% LocalExecuteInvalidationMessage
0.93% ReorderBufferCleanupTXN
0.79% begin_cb_wrapper
0.74% hash_create
0.57% UpdateDecodingStats
+ 0.99% seq_decode
+ 0.74% ReorderBufferAddRelFileLocator
0.77% 0.08% postgres postgres [.] ReorderBufferSequenceIsTransactional
## 8 clients
- 75.95% 0.16% postgres postgres [.] pg_logical_slot_get_changes_guts
- 75.78% pg_logical_slot_get_changes_guts
- 75.06% LogicalDecodingProcessRecord
- 71.67% xact_decode
- 40.29% ReorderBufferProcessTXN
+ 14.88% RelationIdGetRelation
+ 8.87% RelidByRelfilenumber
+ 3.36% AbortCurrentTransaction
+ 2.05% ReorderBufferCopySnap.isra.21
+ 1.71% LocalExecuteInvalidationMessage
+ 1.70% begin_cb_wrapper
+ 1.52% ReorderBufferCleanupTXN
+ 1.46% commit_cb_wrapper
1.10% hash_create
0.54% BeginInternalSubTransaction
+ 29.97% SnapBuildCommitTxn
1.07% seq_decode
+ 0.82% ReorderBufferAddRelFileLocator
0.71% 0.11% postgres postgres [.] ReorderBufferSequenceIsTransactional
# abort case
## 128 clients
- 64.15% 0.62% postgres postgres [.] pg_logical_slot_get_changes_guts
- 63.71% pg_logical_slot_get_changes_guts
- 62.17% LogicalDecodingProcessRecord
- 46.95% seq_decode
- 23.71% ReorderBufferSequenceIsTransactional
+ 13.63% hash_search
8.64% hash_search_with_hash_value
- 22.34% ReorderBufferQueueSequence
+ 20.95% ReorderBufferSequenceIsTransactional
+ 0.53% AbortCurrentTransaction
+ 6.23% xact_decode
+ 3.06% ReorderBufferAddRelFileLocator
+ 2.25% heap2_decode
+ 1.65% heap_decode
+ 0.76% smgr_decode
1.27% XLogReadRecord
- 46.27% 4.60% postgres postgres [.] ReorderBufferSequenceIsTransactional
- 42.57% ReorderBufferSequenceIsTransactional
+ 25.68% hash_search
15.98% hash_search_with_hash_value
+ 1.86% _start
+ 1.13% ServerLoop
+ 0.71% 0x34
## 64 clients
- 38.08% 1.45% postgres postgres [.] pg_logical_slot_get_changes_guts
- 36.89% pg_logical_slot_get_changes_guts
- 34.41% LogicalDecodingProcessRecord
+ 9.88% xact_decode
- 6.35% seq_decode
+ 2.47% ReorderBufferQueueSequence
- 1.73% ReorderBufferSequenceIsTransactional
+ 0.91% hash_search
0.72% hash_search_with_hash_value
+ 0.85% ReorderBufferGetTupleBuf
0.62% ReorderBufferTXNByXid
- 5.39% ReorderBufferAddRelFileLocator
+ 5.14% hash_create
+ 4.91% heap2_decode
+ 3.26% heap_decode
+ 1.76% smgr_decode
1.09% ReorderBufferTXNByXid
+ 1.97% XLogReadRecord
+ 1.19% _start
- 29.14% 3.89% postgres postgres [.] ReorderBufferSequenceIsTransactional
- 25.60% ReorderBufferSequenceIsTransactional
+ 14.29% hash_search
10.94% hash_search_with_hash_value
+ 2.89% _start
+ 0.60% 0x34
## 32 clients
- 51.65% 1.37% postgres postgres [.] pg_logical_slot_get_changes_guts
- 50.81% pg_logical_slot_get_changes_guts
- 47.23% LogicalDecodingProcessRecord
- 19.72% seq_decode
+ 9.36% ReorderBufferSequenceIsTransactional
- 8.73% ReorderBufferQueueSequence
+ 6.20% ReorderBufferSequenceIsTransactional
+ 1.49% AbortCurrentTransaction
0.50% ReorderBufferGetTupleBuf
+ 10.70% xact_decode
+ 6.26% heap2_decode
+ 4.65% ReorderBufferAddRelFileLocator
2.05% heap_decode
+ 1.19% smgr_decode
+ 0.98% standby_decode
+ 2.54% XLogReadRecord
+ 0.84% _start
- 15.82% 1.76% postgres postgres [.] ReorderBufferSequenceIsTransactional
- 14.05% ReorderBufferSequenceIsTransactional
+ 8.10% hash_search
5.95% hash_search_with_hash_value
+ 1.51% _start
## 16 clients
- 47.89% 0.62% postgres postgres [.] pg_logical_slot_get_changes_guts
- 47.51% pg_logical_slot_get_changes_guts
- 44.85% LogicalDecodingProcessRecord
- 13.51% seq_decode
- 5.84% ReorderBufferQueueSequence
+ 3.84% ReorderBufferSequenceIsTransactional
+ 0.71% AbortCurrentTransaction
+ 2.88% ReorderBufferSequenceIsTransactional
+ 1.83% ReorderBufferGetTupleBuf
1.22% __memcpy_ssse3_back
+ 0.52% ReorderBufferTXNByXid
+ 11.06% xact_decode
+ 6.94% ReorderBufferAddRelFileLocator
+ 5.00% heap2_decode
+ 2.73% heap_decode
+ 2.58% standby_decode
+ 1.82% smgr_decode
+ 1.66% XLogReadRecord
- 7.26% 1.04% postgres postgres [.] ReorderBufferSequenceIsTransactional
- 6.49% ReorderBufferSequenceIsTransactional
+ 3.95% hash_search
2.27% hash_search_with_hash_value
+ 0.50% _start
## 8 clients
- 36.33% 1.49% postgres postgres [.] pg_logical_slot_get_changes_guts
- 35.35% pg_logical_slot_get_changes_guts
- 32.91% LogicalDecodingProcessRecord
- 8.31% seq_decode
- 4.54% ReorderBufferQueueSequence
+ 2.35% ReorderBufferSequenceIsTransactional
+ 0.99% AbortCurrentTransaction
+ 1.38% ReorderBufferSequenceIsTransactional
0.97% __memcpy_ssse3_back
0.95% ReorderBufferGetTupleBuf
+ 8.02% xact_decode
+ 5.63% heap2_decode
+ 3.66% ReorderBufferAddRelFileLocator
+ 3.53% standby_decode
+ 2.25% heap_decode
0.99% ReorderBufferTXNByXid
+ 1.22% CreateDecodingContext
+ 0.51% InvalidateSystemCachesExtended
+ 0.98% _start
- 3.73% 0.00% postgres postgres [.] ReorderBufferSequenceIsTransactional
- ReorderBufferSequenceIsTransactional
2.37% hash_search_with_hash_value
+ 1.36% hash_search
Attachments:
[text/plain] perf_results.txt (9.8K, ../../TY3PR01MB9889D457278B254CA87D1325F581A@TY3PR01MB9889.jpnprd01.prod.outlook.com/2-perf_results.txt)
download | inline:
# commit case
## 128 clients
- 97.77% 0.00% postgres postgres [.] pg_logical_slot_get_changes_guts
- 97.77% pg_logical_slot_get_changes_guts
- 97.77% LogicalDecodingProcessRecord
- 97.71% xact_decode
- 54.95% ReorderBufferProcessTXN
- 32.64% ReorderBufferRestoreChanges
+ 23.37% FileRead
4.58% __memcpy_ssse3_back
+ 4.14% MemoryContextAllocZero
+ 7.73% ReorderBufferCopySnap.isra.21
+ 5.94% ReorderBufferSerializeTXN
+ 5.78% ReorderBufferCleanupTXN
+ 2.04% RelationIdGetRelation
- 42.75% SnapBuildCommitTxn
- 34.56% ReorderBufferQueueChange
- 34.36% ReorderBufferSerializeTXN
27.64% __write_nocancel
4.42% __memcpy_ssse3_back
+ 1.52% OpenTransientFilePerm
+ 7.97% SnapBuildBuildSnapshot
0.04% 0.00% postgres postgres [.] ReorderBufferSequenceIsTransactional
## 64 clients
- 86.49% 0.04% postgres postgres [.] pg_logical_slot_get_changes_guts
- 86.45% pg_logical_slot_get_changes_guts
- 86.37% LogicalDecodingProcessRecord
- 84.79% xact_decode
- 51.05% SnapBuildCommitTxn
+ 49.77% SnapBuildBuildSnapshot
0.53% ReorderBufferXidHasBaseSnapshot
- 33.45% ReorderBufferProcessTXN
+ 21.37% ReorderBufferCopySnap.isra.21
+ 4.86% RelationIdGetRelation
+ 2.31% RelidByRelfilenumber
+ 0.84% AbortCurrentTransaction
0.81% ReorderBufferCleanupTXN
- 1.10% seq_decode
0.65% ReorderBufferSequenceIsTransactional
- 1.04% 0.09% postgres postgres [.] ReorderBufferSequenceIsTransactional
## 32 clients
- 82.20% 0.13% postgres postgres [.] pg_logical_slot_get_changes_guts
- 82.09% pg_logical_slot_get_changes_guts
- 81.86% LogicalDecodingProcessRecord
- 80.01% xact_decode
- 49.38% SnapBuildCommitTxn
+ 48.10% SnapBuildBuildSnapshot
0.50% ReorderBufferXidHasBaseSnapshot
- 29.98% ReorderBufferProcessTXN
+ 10.52% ReorderBufferCopySnap.isra.21
+ 7.64% RelationIdGetRelation
+ 4.11% RelidByRelfilenumber
+ 1.70% AbortCurrentTransaction
1.01% ReorderBufferCleanupTXN
0.92% LocalExecuteInvalidationMessage
0.87% commit_cb_wrapper
- 1.01% seq_decode
0.56% ReorderBufferSequenceIsTransactional
+ 0.85% 0.08% postgres postgres [.] ReorderBufferSequenceIsTransactional
## 16 clients
- 78.30% 0.10% postgres postgres [.] pg_logical_slot_get_changes_guts
- 78.24% pg_logical_slot_get_changes_guts
- 77.89% LogicalDecodingProcessRecord
- 75.11% xact_decode
- 38.84% SnapBuildCommitTxn
- 37.54% SnapBuildBuildSnapshot
+ 37.09% pg_qsort
0.56% ReorderBufferXidHasBaseSnapshot
- 35.34% ReorderBufferProcessTXN
+ 11.17% RelationIdGetRelation
+ 7.05% RelidByRelfilenumber
+ 6.33% ReorderBufferCopySnap.isra.21
+ 2.78% AbortCurrentTransaction
+ 1.27% commit_cb_wrapper
+ 1.22% LocalExecuteInvalidationMessage
0.93% ReorderBufferCleanupTXN
0.79% begin_cb_wrapper
0.74% hash_create
0.57% UpdateDecodingStats
+ 0.99% seq_decode
+ 0.74% ReorderBufferAddRelFileLocator
0.77% 0.08% postgres postgres [.] ReorderBufferSequenceIsTransactional
## 8 clients
- 75.95% 0.16% postgres postgres [.] pg_logical_slot_get_changes_guts
- 75.78% pg_logical_slot_get_changes_guts
- 75.06% LogicalDecodingProcessRecord
- 71.67% xact_decode
- 40.29% ReorderBufferProcessTXN
+ 14.88% RelationIdGetRelation
+ 8.87% RelidByRelfilenumber
+ 3.36% AbortCurrentTransaction
+ 2.05% ReorderBufferCopySnap.isra.21
+ 1.71% LocalExecuteInvalidationMessage
+ 1.70% begin_cb_wrapper
+ 1.52% ReorderBufferCleanupTXN
+ 1.46% commit_cb_wrapper
1.10% hash_create
0.54% BeginInternalSubTransaction
+ 29.97% SnapBuildCommitTxn
1.07% seq_decode
+ 0.82% ReorderBufferAddRelFileLocator
0.71% 0.11% postgres postgres [.] ReorderBufferSequenceIsTransactional
# abort case
## 128 clients
- 64.15% 0.62% postgres postgres [.] pg_logical_slot_get_changes_guts
- 63.71% pg_logical_slot_get_changes_guts
- 62.17% LogicalDecodingProcessRecord
- 46.95% seq_decode
- 23.71% ReorderBufferSequenceIsTransactional
+ 13.63% hash_search
8.64% hash_search_with_hash_value
- 22.34% ReorderBufferQueueSequence
+ 20.95% ReorderBufferSequenceIsTransactional
+ 0.53% AbortCurrentTransaction
+ 6.23% xact_decode
+ 3.06% ReorderBufferAddRelFileLocator
+ 2.25% heap2_decode
+ 1.65% heap_decode
+ 0.76% smgr_decode
1.27% XLogReadRecord
- 46.27% 4.60% postgres postgres [.] ReorderBufferSequenceIsTransactional
- 42.57% ReorderBufferSequenceIsTransactional
+ 25.68% hash_search
15.98% hash_search_with_hash_value
+ 1.86% _start
+ 1.13% ServerLoop
+ 0.71% 0x34
## 64 clients
- 38.08% 1.45% postgres postgres [.] pg_logical_slot_get_changes_guts
- 36.89% pg_logical_slot_get_changes_guts
- 34.41% LogicalDecodingProcessRecord
+ 9.88% xact_decode
- 6.35% seq_decode
+ 2.47% ReorderBufferQueueSequence
- 1.73% ReorderBufferSequenceIsTransactional
+ 0.91% hash_search
0.72% hash_search_with_hash_value
+ 0.85% ReorderBufferGetTupleBuf
0.62% ReorderBufferTXNByXid
- 5.39% ReorderBufferAddRelFileLocator
+ 5.14% hash_create
+ 4.91% heap2_decode
+ 3.26% heap_decode
+ 1.76% smgr_decode
1.09% ReorderBufferTXNByXid
+ 1.97% XLogReadRecord
+ 1.19% _start
- 29.14% 3.89% postgres postgres [.] ReorderBufferSequenceIsTransactional
- 25.60% ReorderBufferSequenceIsTransactional
+ 14.29% hash_search
10.94% hash_search_with_hash_value
+ 2.89% _start
+ 0.60% 0x34
## 32 clients
- 51.65% 1.37% postgres postgres [.] pg_logical_slot_get_changes_guts
- 50.81% pg_logical_slot_get_changes_guts
- 47.23% LogicalDecodingProcessRecord
- 19.72% seq_decode
+ 9.36% ReorderBufferSequenceIsTransactional
- 8.73% ReorderBufferQueueSequence
+ 6.20% ReorderBufferSequenceIsTransactional
+ 1.49% AbortCurrentTransaction
0.50% ReorderBufferGetTupleBuf
+ 10.70% xact_decode
+ 6.26% heap2_decode
+ 4.65% ReorderBufferAddRelFileLocator
2.05% heap_decode
+ 1.19% smgr_decode
+ 0.98% standby_decode
+ 2.54% XLogReadRecord
+ 0.84% _start
- 15.82% 1.76% postgres postgres [.] ReorderBufferSequenceIsTransactional
- 14.05% ReorderBufferSequenceIsTransactional
+ 8.10% hash_search
5.95% hash_search_with_hash_value
+ 1.51% _start
## 16 clients
- 47.89% 0.62% postgres postgres [.] pg_logical_slot_get_changes_guts
- 47.51% pg_logical_slot_get_changes_guts
- 44.85% LogicalDecodingProcessRecord
- 13.51% seq_decode
- 5.84% ReorderBufferQueueSequence
+ 3.84% ReorderBufferSequenceIsTransactional
+ 0.71% AbortCurrentTransaction
+ 2.88% ReorderBufferSequenceIsTransactional
+ 1.83% ReorderBufferGetTupleBuf
1.22% __memcpy_ssse3_back
+ 0.52% ReorderBufferTXNByXid
+ 11.06% xact_decode
+ 6.94% ReorderBufferAddRelFileLocator
+ 5.00% heap2_decode
+ 2.73% heap_decode
+ 2.58% standby_decode
+ 1.82% smgr_decode
+ 1.66% XLogReadRecord
- 7.26% 1.04% postgres postgres [.] ReorderBufferSequenceIsTransactional
- 6.49% ReorderBufferSequenceIsTransactional
+ 3.95% hash_search
2.27% hash_search_with_hash_value
+ 0.50% _start
## 8 clients
- 36.33% 1.49% postgres postgres [.] pg_logical_slot_get_changes_guts
- 35.35% pg_logical_slot_get_changes_guts
- 32.91% LogicalDecodingProcessRecord
- 8.31% seq_decode
- 4.54% ReorderBufferQueueSequence
+ 2.35% ReorderBufferSequenceIsTransactional
+ 0.99% AbortCurrentTransaction
+ 1.38% ReorderBufferSequenceIsTransactional
0.97% __memcpy_ssse3_back
0.95% ReorderBufferGetTupleBuf
+ 8.02% xact_decode
+ 5.63% heap2_decode
+ 3.66% ReorderBufferAddRelFileLocator
+ 3.53% standby_decode
+ 2.25% heap_decode
0.99% ReorderBufferTXNByXid
+ 1.22% CreateDecodingContext
+ 0.51% InvalidateSystemCachesExtended
+ 0.98% _start
- 3.73% 0.00% postgres postgres [.] ReorderBufferSequenceIsTransactional
- ReorderBufferSequenceIsTransactional
2.37% hash_search_with_hash_value
+ 1.36% hash_search
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2023-12-02 00:10 Tomas Vondra <[email protected]>
parent: Amit Kapila <[email protected]>
0 siblings, 0 replies; 127+ messages in thread
From: Tomas Vondra @ 2023-12-02 00:10 UTC (permalink / raw)
To: Amit Kapila <[email protected]>; +Cc: Hayato Kuroda (Fujitsu) <[email protected]>; Zhijie Hou (Fujitsu) <[email protected]>; Ashutosh Bapat <[email protected]>; PostgreSQL Hackers <[email protected]>; Masahiko Sawada <[email protected]>; Peter Eisentraut <[email protected]>; Dilip Kumar <[email protected]>
On 11/30/23 12:56, Amit Kapila wrote:
> On Thu, Nov 30, 2023 at 5:28 AM Tomas Vondra
> <[email protected]> wrote:
>>
>> 3) "bad case" - small transactions that generate a lot of relfilenodes
>>
>> select alter_sequence();
>>
>> where the function is defined like this (I did create 1000 sequences
>> before the test):
>>
>> CREATE OR REPLACE FUNCTION alter_sequence() RETURNS void AS $$
>> DECLARE
>> v INT;
>> BEGIN
>> v := 1 + (random() * 999)::int;
>> execute format('alter sequence s%s restart with 1000', v);
>> perform nextval('s');
>> END;
>> $$ LANGUAGE plpgsql;
>>
>> This performs terribly, but it's entirely unrelated to sequences.
>> Current master has exactly the same problem, if transactions do DDL.
>> Like this, for example:
>>
>> CREATE OR REPLACE FUNCTION create_table() RETURNS void AS $$
>> DECLARE
>> v INT;
>> BEGIN
>> v := 1 + (random() * 999)::int;
>> execute format('create table t%s (a int)', v);
>> execute format('drop table t%s', v);
>> insert into t values (1);
>> END;
>> $$ LANGUAGE plpgsql;
>>
>> This has the same impact on master. The perf report shows this:
>>
>> --98.06%--pg_logical_slot_get_changes_guts
>> |
>> --97.88%--LogicalDecodingProcessRecord
>> |
>> --97.56%--xact_decode
>> |
>> --97.51%--DecodeCommit
>> |
>> |--91.92%--SnapBuildCommitTxn
>> | |
>> | --91.65%--SnapBuildBuildSnapshot
>> | |
>> | --91.14%--pg_qsort
>>
>> The sequence decoding is maybe ~1%. The reason why SnapBuildSnapshot
>> takes so long is because:
>>
>> -----------------
>> Breakpoint 1, SnapBuildBuildSnapshot (builder=0x21f60f8)
>> at snapbuild.c:498
>> 498 + sizeof(TransactionId) * builder->committed.xcnt
>> (gdb) p builder->committed.xcnt
>> $4 = 11532
>> -----------------
>>
>> And with each iteration it grows by 1.
>>
>
> Can we somehow avoid this either by keeping DDL-related xacts open or
> aborting them?
I
I'm not sure why the snapshot builder does this, i.e. why we end up
accumulating that many xids, and I didn't have time to look closer. So I
don't know if this would be a solution or not.
> Also, will it make any difference to use setval as
> do_setval() seems to be logging each time?
>
I think that's pretty much what case (2) does, as it calls nextval()
enough time for each transaction do generate WAL. But I don't think this
is a very sensible benchmark - it's an extreme case, but practical cases
are far closer to case (1) because sequences are intermixed with other
activity. No one really does just nextval() calls.
> If possible, can you share the scripts? Kuroda-San has access to the
> performance machine, he may be able to try it as well.
>
Sure, attached. But it's a very primitive script, nothing fancy.
regards
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
Attachments:
[application/x-compressed-tar] scripts.tgz (1.3K, ../../[email protected]/2-scripts.tgz)
download
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2023-12-02 00:23 Tomas Vondra <[email protected]>
parent: Hayato Kuroda (Fujitsu) <[email protected]>
0 siblings, 1 reply; 127+ messages in thread
From: Tomas Vondra @ 2023-12-02 00:23 UTC (permalink / raw)
To: Hayato Kuroda (Fujitsu) <[email protected]>; Amit Kapila <[email protected]>; +Cc: Zhijie Hou (Fujitsu) <[email protected]>; Ashutosh Bapat <[email protected]>; PostgreSQL Hackers <[email protected]>; Masahiko Sawada <[email protected]>; Peter Eisentraut <[email protected]>; Dilip Kumar <[email protected]>
On 12/1/23 12:08, Hayato Kuroda (Fujitsu) wrote:
> Dear Tomas,
>
>> I did some micro-benchmarking today, trying to identify cases where this
>> would cause unexpected problems, either due to having to maintain all
>> the relfilenodes, or due to having to do hash lookups for every sequence
>> change. But I think it's fine, mostly ...
>>
>
> I did also performance tests (especially case 3). First of all, there are some
> variants from yours.
>
> 1. patch 0002 was reverted because it has an issue. So this test checks whether
> refactoring around ReorderBufferSequenceIsTransactional seems really needed.
FWIW I also did the benchmarks without the 0002 patch, for the same
reason. I forgot to mention that.
> 2. per comments from Amit, I also measured the abort case. In this case, the
> alter_sequence() is called but the transaction is aborted.
> 3. I measured with changing number of clients {8, 16, 32, 64, 128}. In any cases,
> clients executed 1000 transactions. The performance machine has 128 core so that
> result for 128 clients might be saturated.
> 4. a short sleep (0.1s) was added in alter_sequence(), especially between
> "alter sequence" and nextval(). Because while testing, I found that the
> transaction is too short to execute in parallel. I think it is reasonable
> because ReorderBufferSequenceIsTransactional() might be worse when the parallelism
> is increased.
>
> I attached one backend process via perf and executed pg_slot_logical_get_changes().
> Attached txt file shows which function occupied CPU time, especially from
> pg_logical_slot_get_changes_guts() and ReorderBufferSequenceIsTransactional().
> Here are my observations about them.
>
> * In case of commit, as you said, SnapBuildCommitTxn() seems dominant for 8-64
> clients case.
> * For (commit, 128 clients) case, however, ReorderBufferRestoreChanges() waste
> many times. I think this is because changes exceed logical_decoding_work_mem,
> so we do not have to analyze anymore.
> * In case of abort, CPU time used by ReorderBufferSequenceIsTransactional() is linearly
> longer. This means that we need to think some solution to avoid the overhead by
> ReorderBufferSequenceIsTransactional().
>
> ```
> 8 clients 3.73% occupied time
> 16 7.26%
> 32 15.82%
> 64 29.14%
> 128 46.27%
> ```
Interesting, so what exactly does the transaction do? Anyway, I don't
think this is very surprising - I believe it behaves like this because
of having to search in many hash tables (one in each toplevel xact). And
I think the solution I explained before (maintaining a single toplevel
hash, instead of many per-top-level hashes).
FWIW I find this case interesting, but not very practical, because no
practical workload has that many aborts.
>
> * In case of abort, I also checked CPU time used by ReorderBufferAddRelFileLocator(), but
> it seems not so depends on the number of clients.
>
> ```
> 8 clients 3.66% occupied time
> 16 6.94%
> 32 4.65%
> 64 5.39%
> 128 3.06%
> ```
>
> As next step, I've planned to run the case which uses setval() function, because it
> generates more WALs than normal nextval();
> How do you think?
>
Sure, although I don't think it's much different from the test selecting
40 values from the sequence (in each transaction). That generates about
the same amount of WAL.
regards
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
^ permalink raw reply [nested|flat] 127+ messages in thread
* RE: logical decoding and replication of sequences, take 2
@ 2023-12-03 12:55 Hayato Kuroda (Fujitsu) <[email protected]>
parent: Tomas Vondra <[email protected]>
0 siblings, 1 reply; 127+ messages in thread
From: Hayato Kuroda (Fujitsu) @ 2023-12-03 12:55 UTC (permalink / raw)
To: 'Tomas Vondra' <[email protected]>; Amit Kapila <[email protected]>; +Cc: Zhijie Hou (Fujitsu) <[email protected]>; Ashutosh Bapat <[email protected]>; PostgreSQL Hackers <[email protected]>; Masahiko Sawada <[email protected]>; Peter Eisentraut <[email protected]>; Dilip Kumar <[email protected]>
Dear Tomas,
> > I did also performance tests (especially case 3). First of all, there are some
> > variants from yours.
> >
> > 1. patch 0002 was reverted because it has an issue. So this test checks whether
> > refactoring around ReorderBufferSequenceIsTransactional seems really
> needed.
>
> FWIW I also did the benchmarks without the 0002 patch, for the same
> reason. I forgot to mention that.
Oh, good news. So your bench markings are quite meaningful.
>
> Interesting, so what exactly does the transaction do?
It is quite simple - PSA the script file. It was executed with 64 multiplicity.
The definition of alter_sequence() is same as you said.
(I did use normal bash script for running them, but your approach may be smarter)
> Anyway, I don't
> think this is very surprising - I believe it behaves like this because
> of having to search in many hash tables (one in each toplevel xact). And
> I think the solution I explained before (maintaining a single toplevel
> hash, instead of many per-top-level hashes).
Agreed. And I can benchmark again for new ones, maybe when we decide new
approach.
Best Regards,
Hayato Kuroda
FUJITSU LIMITED
Attachments:
[application/octet-stream] one_client.sh (317B, ../../OS3PR01MB9882CCE9A9F037309078DC3DF587A@OS3PR01MB9882.jpnprd01.prod.outlook.com/2-one_client.sh)
download
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2023-12-03 17:52 Tomas Vondra <[email protected]>
parent: Hayato Kuroda (Fujitsu) <[email protected]>
0 siblings, 3 replies; 127+ messages in thread
From: Tomas Vondra @ 2023-12-03 17:52 UTC (permalink / raw)
To: Hayato Kuroda (Fujitsu) <[email protected]>; Amit Kapila <[email protected]>; +Cc: Zhijie Hou (Fujitsu) <[email protected]>; Ashutosh Bapat <[email protected]>; PostgreSQL Hackers <[email protected]>; Masahiko Sawada <[email protected]>; Peter Eisentraut <[email protected]>; Dilip Kumar <[email protected]>
On 12/3/23 13:55, Hayato Kuroda (Fujitsu) wrote:
> Dear Tomas,
>
>>> I did also performance tests (especially case 3). First of all, there are some
>>> variants from yours.
>>>
>>> 1. patch 0002 was reverted because it has an issue. So this test checks whether
>>> refactoring around ReorderBufferSequenceIsTransactional seems really
>> needed.
>>
>> FWIW I also did the benchmarks without the 0002 patch, for the same
>> reason. I forgot to mention that.
>
> Oh, good news. So your bench markings are quite meaningful.
>
>>
>> Interesting, so what exactly does the transaction do?
>
> It is quite simple - PSA the script file. It was executed with 64 multiplicity.
> The definition of alter_sequence() is same as you said.
> (I did use normal bash script for running them, but your approach may be smarter)
>
>> Anyway, I don't
>> think this is very surprising - I believe it behaves like this because
>> of having to search in many hash tables (one in each toplevel xact). And
>> I think the solution I explained before (maintaining a single toplevel
>> hash, instead of many per-top-level hashes).
>
> Agreed. And I can benchmark again for new ones, maybe when we decide new
> approach.
>
Thanks for the script. Are you also measuring the time it takes to
decode this using test_decoding?
FWIW I did more comprehensive suite of tests over the weekend, with a
couple more variations. I'm attaching the updated scripts, running it
should be as simple as
./run.sh BRANCH TRANSACTIONS RUNS
so perhaps
./run.sh master 1000 3
to do 3 runs with 1000 transactions per client. And it'll run a bunch of
combinations hard-coded in the script, and write the timings into a CSV
file (with "master" in each row).
I did this on two machines (i5 with 4 cores, xeon with 16/32 cores). I
did this with current master, the basic patch (without the 0002 part),
and then with the optimized approach (single global hash table, see the
0004 part). That's what master / patched / optimized in the results is.
Interestingly enough, the i5 handled this much faster, it seems to be
better in single-core tasks. The xeon is still running, so the results
for "optimized" only have one run (out of 3), but shouldn't change much.
Attached is also a table summarizing this, and visualizing the timing
change (vs. master) in the last couple columns. Green is "faster" than
master (but we don't really expect that), and "red" means slower than
master (the more red, the slower).
There results are grouped by script (see the attached .tgz), with either
32 or 96 clients (which does affect the timing, but not between master
and patch). Some executions have no pg_sleep() calls, some have 0.001
wait (but that doesn't seem to make much difference).
Overall, I'd group the results into about three groups:
1) good cases [nextval, nextval-40, nextval-abort]
These are cases that slow down a bit, but the slowdown is mostly within
reasonable bounds (we're making the decoding to do more stuff, so it'd
be a bit silly to require that extra work to make no impact). And I do
think this is reasonable, because this is pretty much an extreme / worst
case behavior. People don't really do just nextval() calls, without
doing anything else. Not to mention doing aborts for 100% transactions.
So in practice this is going to be within noise (and in those cases the
results even show speedup, which seems a bit surprising). It's somewhat
dependent on CPU too - on xeon there's hardly any regression.
2) nextval-40-abort
Here the slowdown is clear, but I'd argue it generally falls in the same
group as (1). Yes, I'd be happier if it didn't behave like this, but if
someone can show me a practical workload affected by this ...
3) irrelevant cases [all the alters taking insane amounts of time]
I absolutely refuse to care about these extreme cases where decoding
100k transactions takes 5-10 minutes (on i5), or up to 30 minutes (on
xeon). If this was a problem for some practical workload, we'd have
already heard about it I guess. And even if there was such workload, it
wouldn't be up to this patch to fix that. There's clearly something
misbehaving in the snapshot builder.
I was hopeful the global hash table would be an improvement, but that
doesn't seem to be the case. I haven't done much profiling yet, but I'd
guess most of the overhead is due to ReorderBufferQueueSequence()
starting and aborting a transaction in the non-transactinal case. Which
is unfortunate, but I don't know if there's a way to optimize that.
Some time ago I floated the idea of maybe "queuing" the sequence changes
and only replay them on the next commit, somehow. But we did ran into
problems with which snapshot to use, that I didn't know how to solve.
Maybe we should try again. The idea is we'd queue the non-transactional
changes somewhere (can't be in the transaction, because we must keep
them even if it aborts), and then "inject" them into the next commit.
That'd mean we wouldn't do the separate start/abort for each change.
regards
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
Attachments:
[application/x-compressed-tar] sequence-scripts.tgz (1.8K, ../../[email protected]/2-sequence-scripts.tgz)
download
[application/pdf] seq-decoding-benchmark.pdf (52.3K, ../../[email protected]/3-seq-decoding-benchmark.pdf)
download
[text/csv] xeon.csv (15.2K, ../../[email protected]/4-xeon.csv)
download | inline:
branch run clients transactions sleep script timing
master 1 32 100000 0 nextval.sql 623.028
master 1 32 100000 0 nextval-40.sql 655.193
master 1 32 100000 0 create-table.sql 203132.708
master 1 32 100000 0 alter-sequence.sql 384457.824
master 1 32 100000 0 alter-sequence-2.sql 382544.738
master 1 32 100000 0 nextval-abort.sql 84.931
master 1 32 100000 0 nextval-40-abort.sql 189.953
master 1 32 100000 0 create-table-abort.sql 1662.349
master 1 32 100000 0 alter-sequence-abort.sql 324.404
master 1 32 100000 0 alter-sequence-2-abort.sql 382084.142
master 1 32 100000 0.001 nextval.sql 621.664
master 1 32 100000 0.001 nextval-40.sql 658.105
master 1 32 100000 0.001 create-table.sql 847962.454
master 1 32 100000 0.001 alter-sequence.sql 931680.072
master 1 32 100000 0.001 alter-sequence-2.sql 942723.206
master 1 32 100000 0.001 nextval-abort.sql 118.114
master 1 32 100000 0.001 nextval-40-abort.sql 107.572
master 1 32 100000 0.001 create-table-abort.sql 1659.758
master 1 32 100000 0.001 alter-sequence-abort.sql 317.642
master 1 32 100000 0.001 alter-sequence-2-abort.sql 309907.767
master 1 96 100000 0 nextval.sql 620.488
master 1 96 100000 0 nextval-40.sql 664.853
master 1 96 100000 0 create-table.sql 785955.642
master 1 96 100000 0 alter-sequence.sql 416292.711
master 1 96 100000 0 alter-sequence-2.sql 420279.458
master 1 96 100000 0 nextval-abort.sql 81.838
master 1 96 100000 0 nextval-40-abort.sql 183.521
master 1 96 100000 0 create-table-abort.sql 1700.027
master 1 96 100000 0 alter-sequence-abort.sql 323.321
master 1 96 100000 0 alter-sequence-2-abort.sql 204939.351
master 1 96 100000 0.001 nextval.sql 635.048
master 1 96 100000 0.001 nextval-40.sql 667.729
master 1 96 100000 0.001 create-table.sql 1213062.393
master 1 96 100000 0.001 alter-sequence.sql 1764115.249
master 1 96 100000 0.001 alter-sequence-2.sql 1804703.323
master 1 96 100000 0.001 nextval-abort.sql 142.321
master 1 96 100000 0.001 nextval-40-abort.sql 107.555
master 1 96 100000 0.001 create-table-abort.sql 1785.01
master 1 96 100000 0.001 alter-sequence-abort.sql 335.631
master 1 96 100000 0.001 alter-sequence-2-abort.sql 1067857.044
master 2 32 100000 0 nextval.sql 633.909
master 2 32 100000 0 nextval-40.sql 651.664
master 2 32 100000 0 create-table.sql 223647.957
master 2 32 100000 0 alter-sequence.sql 386095.205
master 2 32 100000 0 alter-sequence-2.sql 382203.441
master 2 32 100000 0 nextval-abort.sql 180.156
master 2 32 100000 0 nextval-40-abort.sql 137.725
master 2 32 100000 0 create-table-abort.sql 1691.456
master 2 32 100000 0 alter-sequence-abort.sql 320.58
master 2 32 100000 0 alter-sequence-2-abort.sql 345024.461
master 2 32 100000 0.001 nextval.sql 629.496
master 2 32 100000 0.001 nextval-40.sql 657.461
master 2 32 100000 0.001 create-table.sql 856956.245
master 2 32 100000 0.001 alter-sequence.sql 802099.219
master 2 32 100000 0.001 alter-sequence-2.sql 934687.475
master 2 32 100000 0.001 nextval-abort.sql 175.991
master 2 32 100000 0.001 nextval-40-abort.sql 107.497
master 2 32 100000 0.001 create-table-abort.sql 1659.163
master 2 32 100000 0.001 alter-sequence-abort.sql 328.077
master 2 32 100000 0.001 alter-sequence-2-abort.sql 274636.942
master 2 96 100000 0 nextval.sql 660.764
master 2 96 100000 0 nextval-40.sql 661.686
master 2 96 100000 0 create-table.sql 782779.064
master 2 96 100000 0 alter-sequence.sql 418178.117
master 2 96 100000 0 alter-sequence-2.sql 418967.685
master 2 96 100000 0 nextval-abort.sql 83.675
master 2 96 100000 0 nextval-40-abort.sql 106.25
master 2 96 100000 0 create-table-abort.sql 1699.298
master 2 96 100000 0 alter-sequence-abort.sql 326.064
master 2 96 100000 0 alter-sequence-2-abort.sql 204349.435
master 2 96 100000 0.001 nextval.sql 629.36
master 2 96 100000 0.001 nextval-40.sql 667.915
master 2 96 100000 0.001 create-table.sql 1248478.739
master 2 96 100000 0.001 alter-sequence.sql 1680748.995
master 2 96 100000 0.001 alter-sequence-2.sql 1821028.125
master 2 96 100000 0.001 nextval-abort.sql 91.008
master 2 96 100000 0.001 nextval-40-abort.sql 110.73
master 2 96 100000 0.001 create-table-abort.sql 1769.19
master 2 96 100000 0.001 alter-sequence-abort.sql 334.953
master 2 96 100000 0.001 alter-sequence-2-abort.sql 1081497.995
master 3 32 100000 0 nextval.sql 634.269
master 3 32 100000 0 nextval-40.sql 652.172
master 3 32 100000 0 create-table.sql 232797.109
master 3 32 100000 0 alter-sequence.sql 384339.053
master 3 32 100000 0 alter-sequence-2.sql 382868.001
master 3 32 100000 0 nextval-abort.sql 90.109
master 3 32 100000 0 nextval-40-abort.sql 104.856
master 3 32 100000 0 create-table-abort.sql 1664.04
master 3 32 100000 0 alter-sequence-abort.sql 321.284
master 3 32 100000 0 alter-sequence-2-abort.sql 382257.942
master 3 32 100000 0.001 nextval.sql 634.248
master 3 32 100000 0.001 nextval-40.sql 653.863
master 3 32 100000 0.001 create-table.sql 799216.081
master 3 32 100000 0.001 alter-sequence.sql 890561.388
master 3 32 100000 0.001 alter-sequence-2.sql 946368.525
master 3 32 100000 0.001 nextval-abort.sql 89.157
master 3 32 100000 0.001 nextval-40-abort.sql 137.648
master 3 32 100000 0.001 create-table-abort.sql 1674.031
master 3 32 100000 0.001 alter-sequence-abort.sql 315.631
master 3 32 100000 0.001 alter-sequence-2-abort.sql 367522.915
master 3 96 100000 0 nextval.sql 627.831
master 3 96 100000 0 nextval-40.sql 670.078
master 3 96 100000 0 create-table.sql 850366.975
master 3 96 100000 0 alter-sequence.sql 416125.588
master 3 96 100000 0 alter-sequence-2.sql 419438.818
master 3 96 100000 0 nextval-abort.sql 80.946
master 3 96 100000 0 nextval-40-abort.sql 111.971
master 3 96 100000 0 create-table-abort.sql 1677.581
master 3 96 100000 0 alter-sequence-abort.sql 326.4
master 3 96 100000 0 alter-sequence-2-abort.sql 206609.934
master 3 96 100000 0.001 nextval.sql 636.162
master 3 96 100000 0.001 nextval-40.sql 666.666
master 3 96 100000 0.001 create-table.sql 1195355.42
master 3 96 100000 0.001 alter-sequence.sql 1762692.981
master 3 96 100000 0.001 alter-sequence-2.sql 1766023.526
master 3 96 100000 0.001 nextval-abort.sql 103.923
master 3 96 100000 0.001 nextval-40-abort.sql 110.86
master 3 96 100000 0.001 create-table-abort.sql 1772.726
master 3 96 100000 0.001 alter-sequence-abort.sql 329.202
master 3 96 100000 0.001 alter-sequence-2-abort.sql 972002.504
master 4 32 100000 0 nextval.sql 637.332
master 4 32 100000 0 nextval-40.sql 697.911
patched 1 32 100000 0 nextval.sql 582.811
patched 1 32 100000 0 nextval-40.sql 768.982
patched 1 32 100000 0 create-table.sql 504418.908
patched 1 32 100000 0 alter-sequence.sql 391041.74
patched 1 32 100000 0 alter-sequence-2.sql 388431.858
patched 1 32 100000 0 nextval-abort.sql 86.357
patched 1 32 100000 0 nextval-40-abort.sql 242.381
patched 1 32 100000 0 create-table-abort.sql 1663.813
patched 1 32 100000 0 alter-sequence-abort.sql 402.007
patched 1 32 100000 0 alter-sequence-2-abort.sql 353347.152
patched 1 32 100000 0.001 nextval.sql 583.371
patched 1 32 100000 0.001 nextval-40.sql 742.705
patched 1 32 100000 0.001 create-table.sql 800608.208
patched 1 32 100000 0.001 alter-sequence.sql 911449.794
patched 1 32 100000 0.001 alter-sequence-2.sql 930452.39
patched 1 32 100000 0.001 nextval-abort.sql 171.333
patched 1 32 100000 0.001 nextval-40-abort.sql 241.795
patched 1 32 100000 0.001 create-table-abort.sql 1674.296
patched 1 32 100000 0.001 alter-sequence-abort.sql 519.697
patched 1 32 100000 0.001 alter-sequence-2-abort.sql 401755.108
patched 1 96 100000 0 nextval.sql 572.773
patched 1 96 100000 0 nextval-40.sql 788.99
patched 1 96 100000 0 create-table.sql 834872.858
patched 1 96 100000 0 alter-sequence.sql 426334.896
patched 1 96 100000 0 alter-sequence-2.sql 425846.849
patched 1 96 100000 0 nextval-abort.sql 89.158
patched 1 96 100000 0 nextval-40-abort.sql 278.338
patched 1 96 100000 0 create-table-abort.sql 1696.007
patched 1 96 100000 0 alter-sequence-abort.sql 414.074
patched 1 96 100000 0 alter-sequence-2-abort.sql 215396.952
patched 1 96 100000 0.001 nextval.sql 580.34
patched 1 96 100000 0.001 nextval-40.sql 775.923
patched 1 96 100000 0.001 create-table.sql 1233740.788
patched 1 96 100000 0.001 alter-sequence.sql 1799104.873
patched 1 96 100000 0.001 alter-sequence-2.sql 1853112.016
patched 1 96 100000 0.001 nextval-abort.sql 116.561
patched 1 96 100000 0.001 nextval-40-abort.sql 277.966
patched 1 96 100000 0.001 create-table-abort.sql 1792.763
patched 1 96 100000 0.001 alter-sequence-abort.sql 699.433
patched 1 96 100000 0.001 alter-sequence-2-abort.sql 1083022.66
patched 2 32 100000 0 nextval.sql 578.803
patched 2 32 100000 0 nextval-40.sql 725.015
patched 2 32 100000 0 create-table.sql 231988.224
patched 2 32 100000 0 alter-sequence.sql 391229.125
patched 2 32 100000 0 alter-sequence-2.sql 388150.028
patched 2 32 100000 0 nextval-abort.sql 91.322
patched 2 32 100000 0 nextval-40-abort.sql 240.441
patched 2 32 100000 0 create-table-abort.sql 1671.314
patched 2 32 100000 0 alter-sequence-abort.sql 404.341
patched 2 32 100000 0 alter-sequence-2-abort.sql 353094.219
patched 2 32 100000 0.001 nextval.sql 594.606
patched 2 32 100000 0.001 nextval-40.sql 745.626
patched 2 32 100000 0.001 create-table.sql 831198.496
patched 2 32 100000 0.001 alter-sequence.sql 937281.704
patched 2 32 100000 0.001 alter-sequence-2.sql 952451.819
patched 2 32 100000 0.001 nextval-abort.sql 201.736
patched 2 32 100000 0.001 nextval-40-abort.sql 239.869
patched 2 32 100000 0.001 create-table-abort.sql 1695.108
patched 2 32 100000 0.001 alter-sequence-abort.sql 516.011
patched 2 32 100000 0.001 alter-sequence-2-abort.sql 406210.715
patched 2 96 100000 0 nextval.sql 582.956
patched 2 96 100000 0 nextval-40.sql 783.476
patched 2 96 100000 0 create-table.sql 836805.677
patched 2 96 100000 0 alter-sequence.sql 425948.742
patched 2 96 100000 0 alter-sequence-2.sql 426474.015
patched 2 96 100000 0 nextval-abort.sql 98.754
patched 2 96 100000 0 nextval-40-abort.sql 272.858
patched 2 96 100000 0 create-table-abort.sql 1693.037
patched 2 96 100000 0 alter-sequence-abort.sql 409.655
patched 2 96 100000 0 alter-sequence-2-abort.sql 210151.276
patched 2 96 100000 0.001 nextval.sql 588.948
patched 2 96 100000 0.001 nextval-40.sql 781.928
patched 2 96 100000 0.001 create-table.sql 1211019.807
patched 2 96 100000 0.001 alter-sequence.sql 1783066.319
patched 2 96 100000 0.001 alter-sequence-2.sql 1789958.586
patched 2 96 100000 0.001 nextval-abort.sql 120.88
patched 2 96 100000 0.001 nextval-40-abort.sql 276.215
patched 2 96 100000 0.001 create-table-abort.sql 1790.2
patched 2 96 100000 0.001 alter-sequence-abort.sql 693.773
patched 2 96 100000 0.001 alter-sequence-2-abort.sql 1181206.568
patched 3 32 100000 0 nextval.sql 584.777
patched 3 32 100000 0 nextval-40.sql 740.645
patched 3 32 100000 0 create-table.sql 223208.024
patched 3 32 100000 0 alter-sequence.sql 389546.695
patched 3 32 100000 0 alter-sequence-2.sql 388280.629
patched 3 32 100000 0 nextval-abort.sql 89.683
patched 3 32 100000 0 nextval-40-abort.sql 237.569
patched 3 32 100000 0 create-table-abort.sql 1690.902
patched 3 32 100000 0 alter-sequence-abort.sql 405.102
patched 3 32 100000 0 alter-sequence-2-abort.sql 346352.469
patched 3 32 100000 0.001 nextval.sql 586.807
patched 3 32 100000 0.001 nextval-40.sql 738.844
patched 3 32 100000 0.001 create-table.sql 814163.33
patched 3 32 100000 0.001 alter-sequence.sql 931786.185
patched 3 32 100000 0.001 alter-sequence-2.sql 891870.938
patched 3 32 100000 0.001 nextval-abort.sql 95.77
patched 3 32 100000 0.001 nextval-40-abort.sql 244.585
patched 3 32 100000 0.001 create-table-abort.sql 1679.509
patched 3 32 100000 0.001 alter-sequence-abort.sql 516.661
patched 3 32 100000 0.001 alter-sequence-2-abort.sql 428329.35
patched 3 96 100000 0 nextval.sql 582.784
patched 3 96 100000 0 nextval-40.sql 789.597
patched 3 96 100000 0 create-table.sql 817410.662
patched 3 96 100000 0 alter-sequence.sql 422896.539
patched 3 96 100000 0 alter-sequence-2.sql 428197.491
patched 3 96 100000 0 nextval-abort.sql 87.072
patched 3 96 100000 0 nextval-40-abort.sql 270.847
patched 3 96 100000 0 create-table-abort.sql 1708.159
patched 3 96 100000 0 alter-sequence-abort.sql 414.376
patched 3 96 100000 0 alter-sequence-2-abort.sql 223596.193
patched 3 96 100000 0.001 nextval.sql 592.26
patched 3 96 100000 0.001 nextval-40.sql 786.391
patched 3 96 100000 0.001 create-table.sql 1229672.022
patched 3 96 100000 0.001 alter-sequence.sql 1788533.373
patched 3 96 100000 0.001 alter-sequence-2.sql 1830162.435
patched 3 96 100000 0.001 nextval-abort.sql 167.015
patched 3 96 100000 0.001 nextval-40-abort.sql 268.889
patched 3 96 100000 0.001 create-table-abort.sql 1788.174
patched 3 96 100000 0.001 alter-sequence-abort.sql 714.544
patched 3 96 100000 0.001 alter-sequence-2-abort.sql 1103484.773
optimized 1 32 100000 0 nextval.sql 619.319
optimized 1 32 100000 0 nextval-40.sql 787.225
optimized 1 32 100000 0 create-table.sql 205991.756
optimized 1 32 100000 0 alter-sequence.sql 206724.823
optimized 1 32 100000 0 alter-sequence-2.sql 388796.523
optimized 1 32 100000 0 nextval-abort.sql 88.662
optimized 1 32 100000 0 nextval-40-abort.sql 257.685
optimized 1 32 100000 0 create-table-abort.sql 1696.909
optimized 1 32 100000 0 alter-sequence-abort.sql 416.988
optimized 1 32 100000 0 alter-sequence-2-abort.sql 383156.9
optimized 1 32 100000 0.001 nextval.sql 624.847
optimized 1 32 100000 0.001 nextval-40.sql 776.512
optimized 1 32 100000 0.001 create-table.sql 841509.604
optimized 1 32 100000 0.001 alter-sequence.sql 940885.532
optimized 1 32 100000 0.001 alter-sequence-2.sql 964950.171
optimized 1 32 100000 0.001 nextval-abort.sql 93.634
optimized 1 32 100000 0.001 nextval-40-abort.sql 252.261
optimized 1 32 100000 0.001 create-table-abort.sql 1693.693
optimized 1 32 100000 0.001 alter-sequence-abort.sql 421.481
optimized 1 32 100000 0.001 alter-sequence-2-abort.sql 394728.448
optimized 1 96 100000 0 nextval.sql 617.308
optimized 1 96 100000 0 nextval-40.sql 793.195
optimized 1 96 100000 0 create-table.sql 819194.515
optimized 1 96 100000 0 alter-sequence.sql 424567.77
optimized 1 96 100000 0 alter-sequence-2.sql 427492.528
optimized 1 96 100000 0 nextval-abort.sql 86.414
optimized 1 96 100000 0 nextval-40-abort.sql 258.088
optimized 1 96 100000 0 create-table-abort.sql 1753.341
optimized 1 96 100000 0 alter-sequence-abort.sql 415.611
optimized 1 96 100000 0 alter-sequence-2-abort.sql 222861.994
optimized 1 96 100000 0.001 nextval.sql 634.778
optimized 1 96 100000 0.001 nextval-40.sql 786.588
optimized 1 96 100000 0.001 create-table.sql 1206783.249
optimized 1 96 100000 0.001 alter-sequence.sql 1735665.882
optimized 1 96 100000 0.001 alter-sequence-2.sql 1792988.121
optimized 1 96 100000 0.001 nextval-abort.sql 98.396
optimized 1 96 100000 0.001 nextval-40-abort.sql 252.04
optimized 1 96 100000 0.001 create-table-abort.sql 1770.059
optimized 1 96 100000 0.001 alter-sequence-abort.sql 452.714
optimized 1 96 100000 0.001 alter-sequence-2-abort.sql 1068880.432
optimized 2 32 100000 0 nextval.sql 615.274
optimized 2 32 100000 0 nextval-40.sql 790.594
optimized 2 32 100000 0 create-table.sql 223994.562
optimized 2 32 100000 0 alter-sequence.sql 390434.754
optimized 2 32 100000 0 alter-sequence-2.sql 389221.619
optimized 2 32 100000 0 nextval-abort.sql 89.345
optimized 2 32 100000 0 nextval-40-abort.sql 254.626
optimized 2 32 100000 0 create-table-abort.sql 1736.071
optimized 2 32 100000 0 alter-sequence-abort.sql 417.376
optimized 2 32 100000 0 alter-sequence-2-abort.sql 380100.252
optimized 2 32 100000 0.001 nextval.sql 645.032
optimized 2 32 100000 0.001 nextval-40.sql 775.543
[text/csv] i5.csv (19.5K, ../../[email protected]/5-i5.csv)
download | inline:
branch run clients transactions sleep script timing
1-master 1 32 100000 0 nextval.sql 728.167
1-master 1 32 100000 0 nextval-40.sql 753.546
1-master 1 32 100000 0 create-table.sql 172928.949
1-master 1 32 100000 0 alter-sequence.sql 181075.668
1-master 1 32 100000 0 alter-sequence-2.sql 269768.004
1-master 1 32 100000 0 nextval-abort.sql 76.587
1-master 1 32 100000 0 nextval-40-abort.sql 91.105
1-master 1 32 100000 0 create-table-abort.sql 1446.957
1-master 1 32 100000 0 alter-sequence-abort.sql 297.452
1-master 1 32 100000 0 alter-sequence-2-abort.sql 181518.673
1-master 1 32 100000 0.001 nextval.sql 735.422
1-master 1 32 100000 0.001 nextval-40.sql 759.748
1-master 1 32 100000 0.001 create-table.sql 138989.97
1-master 1 32 100000 0.001 alter-sequence.sql 258707.666
1-master 1 32 100000 0.001 alter-sequence-2.sql 262026.109
1-master 1 32 100000 0.001 nextval-abort.sql 189.946
1-master 1 32 100000 0.001 nextval-40-abort.sql 95.296
1-master 1 32 100000 0.001 create-table-abort.sql 1448.126
1-master 1 32 100000 0.001 alter-sequence-abort.sql 297.924
1-master 1 32 100000 0.001 alter-sequence-2-abort.sql 285404.813
1-master 1 96 100000 0 nextval.sql 729.897
1-master 1 96 100000 0 nextval-40.sql 755.923
1-master 1 96 100000 0 create-table.sql 123178.649
1-master 1 96 100000 0 alter-sequence.sql 227157.73
1-master 1 96 100000 0 alter-sequence-2.sql 495684.235
1-master 1 96 100000 0 nextval-abort.sql 81.693
1-master 1 96 100000 0 nextval-40-abort.sql 101.058
1-master 1 96 100000 0 create-table-abort.sql 1451.299
1-master 1 96 100000 0 alter-sequence-abort.sql 301.943
1-master 1 96 100000 0 alter-sequence-2-abort.sql 578909.506
1-master 1 96 100000 0.001 nextval.sql 752.748
1-master 1 96 100000 0.001 nextval-40.sql 765.174
1-master 1 96 100000 0.001 create-table.sql 132873.637
1-master 1 96 100000 0.001 alter-sequence.sql 534810.485
1-master 1 96 100000 0.001 alter-sequence-2.sql 606835.115
1-master 1 96 100000 0.001 nextval-abort.sql 80.457
1-master 1 96 100000 0.001 nextval-40-abort.sql 169.033
1-master 1 96 100000 0.001 create-table-abort.sql 1460.329
1-master 1 96 100000 0.001 alter-sequence-abort.sql 297.76
1-master 1 96 100000 0.001 alter-sequence-2-abort.sql 538469.484
1-master 2 32 100000 0 nextval.sql 731.446
1-master 2 32 100000 0 nextval-40.sql 751.254
1-master 2 32 100000 0 create-table.sql 172400.662
1-master 2 32 100000 0 alter-sequence.sql 284383.441
1-master 2 32 100000 0 alter-sequence-2.sql 269557.89
1-master 2 32 100000 0 nextval-abort.sql 77.227
1-master 2 32 100000 0 nextval-40-abort.sql 143.786
1-master 2 32 100000 0 create-table-abort.sql 1486.622
1-master 2 32 100000 0 alter-sequence-abort.sql 298.072
1-master 2 32 100000 0 alter-sequence-2-abort.sql 182096.898
1-master 2 32 100000 0.001 nextval.sql 737.937
1-master 2 32 100000 0.001 nextval-40.sql 762.05
1-master 2 32 100000 0.001 create-table.sql 139173.98
1-master 2 32 100000 0.001 alter-sequence.sql 286577.966
1-master 2 32 100000 0.001 alter-sequence-2.sql 264698.696
1-master 2 32 100000 0.001 nextval-abort.sql 135.294
1-master 2 32 100000 0.001 nextval-40-abort.sql 97.865
1-master 2 32 100000 0.001 create-table-abort.sql 1485.83
1-master 2 32 100000 0.001 alter-sequence-abort.sql 288.976
1-master 2 32 100000 0.001 alter-sequence-2-abort.sql 272963.736
1-master 2 96 100000 0 nextval.sql 732.143
1-master 2 96 100000 0 nextval-40.sql 762.154
1-master 2 96 100000 0 create-table.sql 119050.528
1-master 2 96 100000 0 alter-sequence.sql 310053.732
1-master 2 96 100000 0 alter-sequence-2.sql 460864.75
1-master 2 96 100000 0 nextval-abort.sql 75.672
1-master 2 96 100000 0 nextval-40-abort.sql 92.65
1-master 2 96 100000 0 create-table-abort.sql 1460.708
1-master 2 96 100000 0 alter-sequence-abort.sql 301.75
1-master 2 96 100000 0 alter-sequence-2-abort.sql 515394.956
1-master 2 96 100000 0.001 nextval.sql 738.655
1-master 2 96 100000 0.001 nextval-40.sql 764.778
1-master 2 96 100000 0.001 create-table.sql 114543.592
1-master 2 96 100000 0.001 alter-sequence.sql 364767.771
1-master 2 96 100000 0.001 alter-sequence-2.sql 609877.622
1-master 2 96 100000 0.001 nextval-abort.sql 79.167
1-master 2 96 100000 0.001 nextval-40-abort.sql 135.904
1-master 2 96 100000 0.001 create-table-abort.sql 1476.578
1-master 2 96 100000 0.001 alter-sequence-abort.sql 296.619
1-master 2 96 100000 0.001 alter-sequence-2-abort.sql 556344.68
1-master 3 32 100000 0 nextval.sql 730.639
1-master 3 32 100000 0 nextval-40.sql 750.252
1-master 3 32 100000 0 create-table.sql 172517.696
1-master 3 32 100000 0 alter-sequence.sql 194280.947
1-master 3 32 100000 0 alter-sequence-2.sql 264652.601
1-master 3 32 100000 0 nextval-abort.sql 170.833
1-master 3 32 100000 0 nextval-40-abort.sql 90.1
1-master 3 32 100000 0 create-table-abort.sql 1430.336
1-master 3 32 100000 0 alter-sequence-abort.sql 296.324
1-master 3 32 100000 0 alter-sequence-2-abort.sql 182050.709
1-master 3 32 100000 0.001 nextval.sql 746.739
1-master 3 32 100000 0.001 nextval-40.sql 768.951
1-master 3 32 100000 0.001 create-table.sql 139038.527
1-master 3 32 100000 0.001 alter-sequence.sql 257413.541
1-master 3 32 100000 0.001 alter-sequence-2.sql 259065.183
1-master 3 32 100000 0.001 nextval-abort.sql 76.961
1-master 3 32 100000 0.001 nextval-40-abort.sql 97.925
1-master 3 32 100000 0.001 create-table-abort.sql 1499.66
1-master 3 32 100000 0.001 alter-sequence-abort.sql 288.158
1-master 3 32 100000 0.001 alter-sequence-2-abort.sql 285997.012
1-master 3 96 100000 0 nextval.sql 739.755
1-master 3 96 100000 0 nextval-40.sql 770.252
1-master 3 96 100000 0 create-table.sql 116443.91
1-master 3 96 100000 0 alter-sequence.sql 213266.518
1-master 3 96 100000 0 alter-sequence-2.sql 499323.51
1-master 3 96 100000 0 nextval-abort.sql 73.947
1-master 3 96 100000 0 nextval-40-abort.sql 90.821
1-master 3 96 100000 0 create-table-abort.sql 1434.674
1-master 3 96 100000 0 alter-sequence-abort.sql 302.117
1-master 3 96 100000 0 alter-sequence-2-abort.sql 525310.12
1-master 3 96 100000 0.001 nextval.sql 751.435
1-master 3 96 100000 0.001 nextval-40.sql 776.285
1-master 3 96 100000 0.001 create-table.sql 115242.189
1-master 3 96 100000 0.001 alter-sequence.sql 446144.591
1-master 3 96 100000 0.001 alter-sequence-2.sql 604509.412
1-master 3 96 100000 0.001 nextval-abort.sql 107.621
1-master 3 96 100000 0.001 nextval-40-abort.sql 95.478
1-master 3 96 100000 0.001 create-table-abort.sql 1484.295
1-master 3 96 100000 0.001 alter-sequence-abort.sql 297.634
1-master 3 96 100000 0.001 alter-sequence-2-abort.sql 675396.969
2-patched 1 32 100000 0 nextval.sql 695.735
2-patched 1 32 100000 0 nextval-40.sql 1048.029
2-patched 1 32 100000 0 create-table.sql 162181.947
2-patched 1 32 100000 0 alter-sequence.sql 297332.922
2-patched 1 32 100000 0 alter-sequence-2.sql 277341.427
2-patched 1 32 100000 0 nextval-abort.sql 197.08
2-patched 1 32 100000 0 nextval-40-abort.sql 429.385
2-patched 1 32 100000 0 create-table-abort.sql 1502.787
2-patched 1 32 100000 0 alter-sequence-abort.sql 399.596
2-patched 1 32 100000 0 alter-sequence-2-abort.sql 189904.667
2-patched 1 32 100000 0.001 nextval.sql 703.33
2-patched 1 32 100000 0.001 nextval-40.sql 1050.797
2-patched 1 32 100000 0.001 create-table.sql 143872.668
2-patched 1 32 100000 0.001 alter-sequence.sql 273918.802
2-patched 1 32 100000 0.001 alter-sequence-2.sql 276422.119
2-patched 1 32 100000 0.001 nextval-abort.sql 91.486
2-patched 1 32 100000 0.001 nextval-40-abort.sql 437.644
2-patched 1 32 100000 0.001 create-table-abort.sql 1467.934
2-patched 1 32 100000 0.001 alter-sequence-abort.sql 493.006
2-patched 1 32 100000 0.001 alter-sequence-2-abort.sql 323036.514
2-patched 1 96 100000 0 nextval.sql 698.808
2-patched 1 96 100000 0 nextval-40.sql 1055.27
2-patched 1 96 100000 0 create-table.sql 125236.032
2-patched 1 96 100000 0 alter-sequence.sql 335566.142
2-patched 1 96 100000 0 alter-sequence-2.sql 516360.343
2-patched 1 96 100000 0 nextval-abort.sql 92.701
2-patched 1 96 100000 0 nextval-40-abort.sql 448.173
2-patched 1 96 100000 0 create-table-abort.sql 1466.764
2-patched 1 96 100000 0 alter-sequence-abort.sql 463.433
2-patched 1 96 100000 0 alter-sequence-2-abort.sql 547195.101
2-patched 1 96 100000 0.001 nextval.sql 699.215
2-patched 1 96 100000 0.001 nextval-40.sql 1061.92
2-patched 1 96 100000 0.001 create-table.sql 120813.522
2-patched 1 96 100000 0.001 alter-sequence.sql 401987.151
2-patched 1 96 100000 0.001 alter-sequence-2.sql 657523.689
2-patched 1 96 100000 0.001 nextval-abort.sql 95.266
2-patched 1 96 100000 0.001 nextval-40-abort.sql 451.257
2-patched 1 96 100000 0.001 create-table-abort.sql 1485.894
2-patched 1 96 100000 0.001 alter-sequence-abort.sql 560.839
2-patched 1 96 100000 0.001 alter-sequence-2-abort.sql 715900.871
2-patched 2 32 100000 0 nextval.sql 696.763
2-patched 2 32 100000 0 nextval-40.sql 1036.543
2-patched 2 32 100000 0 create-table.sql 175842.387
2-patched 2 32 100000 0 alter-sequence.sql 292754.941
2-patched 2 32 100000 0 alter-sequence-2.sql 190665.954
2-patched 2 32 100000 0 nextval-abort.sql 88.254
2-patched 2 32 100000 0 nextval-40-abort.sql 426.328
2-patched 2 32 100000 0 create-table-abort.sql 1455.159
2-patched 2 32 100000 0 alter-sequence-abort.sql 400.633
2-patched 2 32 100000 0 alter-sequence-2-abort.sql 190216.152
2-patched 2 32 100000 0.001 nextval.sql 696.664
2-patched 2 32 100000 0.001 nextval-40.sql 1050.48
2-patched 2 32 100000 0.001 create-table.sql 223190.066
2-patched 2 32 100000 0.001 alter-sequence.sql 241653.795
2-patched 2 32 100000 0.001 alter-sequence-2.sql 275550.132
2-patched 2 32 100000 0.001 nextval-abort.sql 199.31
2-patched 2 32 100000 0.001 nextval-40-abort.sql 435.768
2-patched 2 32 100000 0.001 create-table-abort.sql 1483.037
2-patched 2 32 100000 0.001 alter-sequence-abort.sql 495.563
2-patched 2 32 100000 0.001 alter-sequence-2-abort.sql 300394.075
2-patched 2 96 100000 0 nextval.sql 703.99
2-patched 2 96 100000 0 nextval-40.sql 1054.942
2-patched 2 96 100000 0 create-table.sql 127371.575
2-patched 2 96 100000 0 alter-sequence.sql 372535.837
2-patched 2 96 100000 0 alter-sequence-2.sql 485113.441
2-patched 2 96 100000 0 nextval-abort.sql 93.926
2-patched 2 96 100000 0 nextval-40-abort.sql 443.013
2-patched 2 96 100000 0 create-table-abort.sql 1466.567
2-patched 2 96 100000 0 alter-sequence-abort.sql 461.256
2-patched 2 96 100000 0 alter-sequence-2-abort.sql 550689.382
2-patched 2 96 100000 0.001 nextval.sql 700.361
2-patched 2 96 100000 0.001 nextval-40.sql 1053.895
2-patched 2 96 100000 0.001 create-table.sql 119762.077
2-patched 2 96 100000 0.001 alter-sequence.sql 489062.233
2-patched 2 96 100000 0.001 alter-sequence-2.sql 669849.22
2-patched 2 96 100000 0.001 nextval-abort.sql 95.041
2-patched 2 96 100000 0.001 nextval-40-abort.sql 451.233
2-patched 2 96 100000 0.001 create-table-abort.sql 1483.516
2-patched 2 96 100000 0.001 alter-sequence-abort.sql 555.718
2-patched 2 96 100000 0.001 alter-sequence-2-abort.sql 566282.914
2-patched 3 32 100000 0 nextval.sql 696.75
2-patched 3 32 100000 0 nextval-40.sql 1041.715
2-patched 3 32 100000 0 create-table.sql 146404.068
2-patched 3 32 100000 0 alter-sequence.sql 294223.589
2-patched 3 32 100000 0 alter-sequence-2.sql 208693.477
2-patched 3 32 100000 0 nextval-abort.sql 172.037
2-patched 3 32 100000 0 nextval-40-abort.sql 428.415
2-patched 3 32 100000 0 create-table-abort.sql 1460.485
2-patched 3 32 100000 0 alter-sequence-abort.sql 401.919
2-patched 3 32 100000 0 alter-sequence-2-abort.sql 189780.79
2-patched 3 32 100000 0.001 nextval.sql 698.105
2-patched 3 32 100000 0.001 nextval-40.sql 1049.018
2-patched 3 32 100000 0.001 create-table.sql 143399.48
2-patched 3 32 100000 0.001 alter-sequence.sql 272708.948
2-patched 3 32 100000 0.001 alter-sequence-2.sql 277884.588
2-patched 3 32 100000 0.001 nextval-abort.sql 176.967
2-patched 3 32 100000 0.001 nextval-40-abort.sql 434.658
2-patched 3 32 100000 0.001 create-table-abort.sql 1449.505
2-patched 3 32 100000 0.001 alter-sequence-abort.sql 494.287
2-patched 3 32 100000 0.001 alter-sequence-2-abort.sql 303062.426
2-patched 3 96 100000 0 nextval.sql 702.931
2-patched 3 96 100000 0 nextval-40.sql 1073.771
2-patched 3 96 100000 0 create-table.sql 125039.393
2-patched 3 96 100000 0 alter-sequence.sql 226768.632
2-patched 3 96 100000 0 alter-sequence-2.sql 515514.121
2-patched 3 96 100000 0 nextval-abort.sql 134.791
2-patched 3 96 100000 0 nextval-40-abort.sql 450.192
2-patched 3 96 100000 0 create-table-abort.sql 1474.572
2-patched 3 96 100000 0 alter-sequence-abort.sql 463.484
2-patched 3 96 100000 0 alter-sequence-2-abort.sql 578914.392
2-patched 3 96 100000 0.001 nextval.sql 719.772
2-patched 3 96 100000 0.001 nextval-40.sql 1070.341
2-patched 3 96 100000 0.001 create-table.sql 132422.062
2-patched 3 96 100000 0.001 alter-sequence.sql 537647.41
2-patched 3 96 100000 0.001 alter-sequence-2.sql 652513.541
2-patched 3 96 100000 0.001 nextval-abort.sql 93.968
2-patched 3 96 100000 0.001 nextval-40-abort.sql 458.387
2-patched 3 96 100000 0.001 create-table-abort.sql 1530.094
2-patched 3 96 100000 0.001 alter-sequence-abort.sql 558.702
2-patched 3 96 100000 0.001 alter-sequence-2-abort.sql 563909.028
3-optimized 1 32 100000 0 nextval.sql 707.398
3-optimized 1 32 100000 0 nextval-40.sql 1041.247
3-optimized 1 32 100000 0 create-table.sql 177558.436
3-optimized 1 32 100000 0 alter-sequence.sql 298707.15
3-optimized 1 32 100000 0 alter-sequence-2.sql 192773.686
3-optimized 1 32 100000 0 nextval-abort.sql 88.375
3-optimized 1 32 100000 0 nextval-40-abort.sql 416.698
3-optimized 1 32 100000 0 create-table-abort.sql 1468.954
3-optimized 1 32 100000 0 alter-sequence-abort.sql 408.026
3-optimized 1 32 100000 0 alter-sequence-2-abort.sql 189761.999
3-optimized 1 32 100000 0.001 nextval.sql 708.911
3-optimized 1 32 100000 0.001 nextval-40.sql 1054.583
3-optimized 1 32 100000 0.001 create-table.sql 142599.159
3-optimized 1 32 100000 0.001 alter-sequence.sql 271439.001
3-optimized 1 32 100000 0.001 alter-sequence-2.sql 275135.834
3-optimized 1 32 100000 0.001 nextval-abort.sql 90.306
3-optimized 1 32 100000 0.001 nextval-40-abort.sql 423.531
3-optimized 1 32 100000 0.001 create-table-abort.sql 1509.739
3-optimized 1 32 100000 0.001 alter-sequence-abort.sql 408.7
3-optimized 1 32 100000 0.001 alter-sequence-2-abort.sql 300625.411
3-optimized 1 96 100000 0 nextval.sql 706.1
3-optimized 1 96 100000 0 nextval-40.sql 1046.686
3-optimized 1 96 100000 0 create-table.sql 124887.325
3-optimized 1 96 100000 0 alter-sequence.sql 235187.403
3-optimized 1 96 100000 0 alter-sequence-2.sql 509939.154
3-optimized 1 96 100000 0 nextval-abort.sql 126.603
3-optimized 1 96 100000 0 nextval-40-abort.sql 413.421
3-optimized 1 96 100000 0 create-table-abort.sql 1462.081
3-optimized 1 96 100000 0 alter-sequence-abort.sql 417.714
3-optimized 1 96 100000 0 alter-sequence-2-abort.sql 547545.836
3-optimized 1 96 100000 0.001 nextval.sql 712.181
3-optimized 1 96 100000 0.001 nextval-40.sql 1045.539
3-optimized 1 96 100000 0.001 create-table.sql 121172.119
3-optimized 1 96 100000 0.001 alter-sequence.sql 583259.279
3-optimized 1 96 100000 0.001 alter-sequence-2.sql 679324.221
3-optimized 1 96 100000 0.001 nextval-abort.sql 92.946
3-optimized 1 96 100000 0.001 nextval-40-abort.sql 418.094
3-optimized 1 96 100000 0.001 create-table-abort.sql 1493.84
3-optimized 1 96 100000 0.001 alter-sequence-abort.sql 418.023
3-optimized 1 96 100000 0.001 alter-sequence-2-abort.sql 731359.163
3-optimized 2 32 100000 0 nextval.sql 710.917
3-optimized 2 32 100000 0 nextval-40.sql 1041.635
3-optimized 2 32 100000 0 create-table.sql 174984.078
3-optimized 2 32 100000 0 alter-sequence.sql 299229.181
3-optimized 2 32 100000 0 alter-sequence-2.sql 317322.65
3-optimized 2 32 100000 0 nextval-abort.sql 184.164
3-optimized 2 32 100000 0 nextval-40-abort.sql 411.176
3-optimized 2 32 100000 0 create-table-abort.sql 1499.139
3-optimized 2 32 100000 0 alter-sequence-abort.sql 408.866
3-optimized 2 32 100000 0 alter-sequence-2-abort.sql 189292.533
3-optimized 2 32 100000 0.001 nextval.sql 710.211
3-optimized 2 32 100000 0.001 nextval-40.sql 1055.835
3-optimized 2 32 100000 0.001 create-table.sql 142660.077
3-optimized 2 32 100000 0.001 alter-sequence.sql 271600.461
3-optimized 2 32 100000 0.001 alter-sequence-2.sql 272361.617
3-optimized 2 32 100000 0.001 nextval-abort.sql 89.191
3-optimized 2 32 100000 0.001 nextval-40-abort.sql 420.114
3-optimized 2 32 100000 0.001 create-table-abort.sql 1513.827
3-optimized 2 32 100000 0.001 alter-sequence-abort.sql 420.713
3-optimized 2 32 100000 0.001 alter-sequence-2-abort.sql 301259.082
3-optimized 2 96 100000 0 nextval.sql 709.969
3-optimized 2 96 100000 0 nextval-40.sql 1049.466
3-optimized 2 96 100000 0 create-table.sql 125034.778
3-optimized 2 96 100000 0 alter-sequence.sql 367980.442
3-optimized 2 96 100000 0 alter-sequence-2.sql 498984.736
3-optimized 2 96 100000 0 nextval-abort.sql 142.791
3-optimized 2 96 100000 0 nextval-40-abort.sql 418.018
3-optimized 2 96 100000 0 create-table-abort.sql 1476.474
3-optimized 2 96 100000 0 alter-sequence-abort.sql 415.457
3-optimized 2 96 100000 0 alter-sequence-2-abort.sql 596458.974
3-optimized 2 96 100000 0.001 nextval.sql 717.91
3-optimized 2 96 100000 0.001 nextval-40.sql 1039.969
3-optimized 2 96 100000 0.001 create-table.sql 132937.812
3-optimized 2 96 100000 0.001 alter-sequence.sql 533475.204
3-optimized 2 96 100000 0.001 alter-sequence-2.sql 613236.53
3-optimized 2 96 100000 0.001 nextval-abort.sql 91.842
3-optimized 2 96 100000 0.001 nextval-40-abort.sql 423.901
3-optimized 2 96 100000 0.001 create-table-abort.sql 1484.899
3-optimized 2 96 100000 0.001 alter-sequence-abort.sql 417.592
3-optimized 2 96 100000 0.001 alter-sequence-2-abort.sql 582522.834
3-optimized 3 32 100000 0 nextval.sql 714.669
3-optimized 3 32 100000 0 nextval-40.sql 1043.856
3-optimized 3 32 100000 0 create-table.sql 175181.771
3-optimized 3 32 100000 0 alter-sequence.sql 293323.076
3-optimized 3 32 100000 0 alter-sequence-2.sql 192391.494
3-optimized 3 32 100000 0 nextval-abort.sql 191.2
3-optimized 3 32 100000 0 nextval-40-abort.sql 406.298
3-optimized 3 32 100000 0 create-table-abort.sql 1450.291
3-optimized 3 32 100000 0 alter-sequence-abort.sql 406.079
3-optimized 3 32 100000 0 alter-sequence-2-abort.sql 189837.542
3-optimized 3 32 100000 0.001 nextval.sql 715.922
3-optimized 3 32 100000 0.001 nextval-40.sql 1043.398
3-optimized 3 32 100000 0.001 create-table.sql 143181.08
3-optimized 3 32 100000 0.001 alter-sequence.sql 271055.783
3-optimized 3 32 100000 0.001 alter-sequence-2.sql 275353.624
3-optimized 3 32 100000 0.001 nextval-abort.sql 161.438
3-optimized 3 32 100000 0.001 nextval-40-abort.sql 419.931
3-optimized 3 32 100000 0.001 create-table-abort.sql 1462.551
3-optimized 3 32 100000 0.001 alter-sequence-abort.sql 421.507
3-optimized 3 32 100000 0.001 alter-sequence-2-abort.sql 311076.403
3-optimized 3 96 100000 0 nextval.sql 713.33
3-optimized 3 96 100000 0 nextval-40.sql 1063.056
3-optimized 3 96 100000 0 create-table.sql 125511.585
3-optimized 3 96 100000 0 alter-sequence.sql 354536.222
3-optimized 3 96 100000 0 alter-sequence-2.sql 506427.979
3-optimized 3 96 100000 0 nextval-abort.sql 90.122
3-optimized 3 96 100000 0 nextval-40-abort.sql 416.569
3-optimized 3 96 100000 0 create-table-abort.sql 1477.368
3-optimized 3 96 100000 0 alter-sequence-abort.sql 415.234
3-optimized 3 96 100000 0 alter-sequence-2-abort.sql 568071.055
3-optimized 3 96 100000 0.001 nextval.sql 718.803
3-optimized 3 96 100000 0.001 nextval-40.sql 1058.091
3-optimized 3 96 100000 0.001 create-table.sql 123880.338
3-optimized 3 96 100000 0.001 alter-sequence.sql 577954.533
3-optimized 3 96 100000 0.001 alter-sequence-2.sql 550099.583
3-optimized 3 96 100000 0.001 nextval-abort.sql 93.142
3-optimized 3 96 100000 0.001 nextval-40-abort.sql 424.906
3-optimized 3 96 100000 0.001 create-table-abort.sql 1491.27
3-optimized 3 96 100000 0.001 alter-sequence-abort.sql 417.14
3-optimized 3 96 100000 0.001 alter-sequence-2-abort.sql 708543.886
[text/x-patch] v20231203-0001-Logical-decoding-of-sequences.patch (68.9K, ../../[email protected]/6-v20231203-0001-Logical-decoding-of-sequences.patch)
download | inline diff:
From 02e8ff50e064fc1c8d6236acadbccb935028c352 Mon Sep 17 00:00:00 2001
From: Tomas Vondra <[email protected]>
Date: Thu, 2 Nov 2023 15:03:13 +0100
Subject: [PATCH v20231203 1/4] Logical decoding of sequences
Extend the logical decoding to allow decoding sequence changes.
Each sequence change is either transactional or non-transactional,
depending on what whether the sequence relfilenode is visible outside
the transaction.
When a transaction creates a new relfilenode to the sequence (usually
because of CREATE SEQUENCE or ALTER SEQUENCE), all subsequent sequence
changes are treated as transactional - queued just like other changes in
the transaction, and either replayed at commit time or discarded if the
transaction aborts.
All other sequence changes, performed on sequences / relfilenodes
created by earlier transactions, are non-transactional. These changes
are not queued but replayed immediately after decoding. This also means
these changes are not subject to abort.
This mixed behavior is necessary, and is consistent with how sequences
behave in general - calling nextval(s) makes the change visible to
concurrent transactions, and can't be rolled back. But if DDL is
performed on a sequence, changes become transactional - invisible to
concurrent transactions until commit, and subject to abort.
To identify changes that need transactional handling, each transaction
tracks relfilenodes it created in a hash table. Then when a change is
decoded, this hash table is used to check if the relfilenode is new.
For each relfilenode we track the XID of (sub)transaction that created
it, which is needed for cleanup at the transaction end. We don't need to
check the XID to decide if an increment is transactional - if we find a
match in the hash table, it has to be the same transaction.
This requires only minor changes to WAL-logging. We need to ensure the
sequence record is always associated with the subxact XID - until now
the change might have XID 0 if it was the first change in a subxact. But
the sequence might have been created in the same top-level transaction,
in which case it still needs transactional treatment. So we ensure the
XID is assigned when WAL-logging sequence changes, if wal_level=logical.
A patch adding decoding of sequences was originally submitted by Cary
Huang. This commit reworks many important aspects (e.g. the WAL logging
and transactional/non-transactional handling). However, the original
patch and reviews were very useful.
Author: Tomas Vondra, Cary Huang
Reviewed-by: Ashutosh Bapat, Amit Kapila, Peter Eisentraut, Hannu
Krosing, Andres Freund, Zhijie Hou, Peter Smith
Discussion: https://postgr.es/m/[email protected]
Discussion: https://postgr.es/m/[email protected]
Discussion: https://postgr.es/m/[email protected]
---
doc/src/sgml/logicaldecoding.sgml | 62 +-
src/backend/commands/sequence.c | 89 +++
src/backend/replication/logical/decode.c | 228 +++++-
src/backend/replication/logical/logical.c | 100 ++-
.../replication/logical/reorderbuffer.c | 738 +++++++++++++++++-
src/include/access/rmgrlist.h | 4 +-
src/include/replication/decode.h | 2 +
src/include/replication/logical.h | 5 +
src/include/replication/output_plugin.h | 23 +
src/include/replication/reorderbuffer.h | 42 +
src/tools/pgindent/typedefs.list | 1 +
11 files changed, 1282 insertions(+), 12 deletions(-)
diff --git a/doc/src/sgml/logicaldecoding.sgml b/doc/src/sgml/logicaldecoding.sgml
index cd152d4ced9..90e99a95128 100644
--- a/doc/src/sgml/logicaldecoding.sgml
+++ b/doc/src/sgml/logicaldecoding.sgml
@@ -488,6 +488,7 @@ typedef struct OutputPluginCallbacks
LogicalDecodeTruncateCB truncate_cb;
LogicalDecodeCommitCB commit_cb;
LogicalDecodeMessageCB message_cb;
+ LogicalDecodeSequenceCB sequence_cb;
LogicalDecodeFilterByOriginCB filter_by_origin_cb;
LogicalDecodeShutdownCB shutdown_cb;
LogicalDecodeFilterPrepareCB filter_prepare_cb;
@@ -502,6 +503,7 @@ typedef struct OutputPluginCallbacks
LogicalDecodeStreamCommitCB stream_commit_cb;
LogicalDecodeStreamChangeCB stream_change_cb;
LogicalDecodeStreamMessageCB stream_message_cb;
+ LogicalDecodeStreamSequenceCB stream_sequence_cb;
LogicalDecodeStreamTruncateCB stream_truncate_cb;
} OutputPluginCallbacks;
@@ -510,10 +512,13 @@ typedef void (*LogicalOutputPluginInit) (struct OutputPluginCallbacks *cb);
The <function>begin_cb</function>, <function>change_cb</function>
and <function>commit_cb</function> callbacks are required,
while <function>startup_cb</function>, <function>truncate_cb</function>,
- <function>message_cb</function>, <function>filter_by_origin_cb</function>,
+ <function>message_cb</function>, <function>sequence_cb</function>,
+ <function>filter_by_origin_cb</function>,
and <function>shutdown_cb</function> are optional.
If <function>truncate_cb</function> is not set but a
<command>TRUNCATE</command> is to be decoded, the action will be ignored.
+ Similarly, if <function>sequence_cb</function> is not set and a sequence
+ change is to be decoded, the action will be ignored.
</para>
<para>
@@ -521,7 +526,8 @@ typedef void (*LogicalOutputPluginInit) (struct OutputPluginCallbacks *cb);
in-progress transactions. The <function>stream_start_cb</function>,
<function>stream_stop_cb</function>, <function>stream_abort_cb</function>,
<function>stream_commit_cb</function>, and <function>stream_change_cb</function>
- are required, while <function>stream_message_cb</function> and
+ are required, while <function>stream_message_cb</function>,
+ <function>stream_sequence_cb</function> and
<function>stream_truncate_cb</function> are optional. The
<function>stream_prepare_cb</function> is also required if the output
plugin also support two-phase commits.
@@ -840,6 +846,35 @@ typedef void (*LogicalDecodeMessageCB) (struct LogicalDecodingContext *ctx,
</para>
</sect3>
+ <sect3 id="logicaldecoding-output-plugin-sequence">
+ <title>Sequence Callback</title>
+
+ <para>
+ The optional <function>sequence_cb</function> callback is called for
+ actions that update a sequence value.
+<programlisting>
+typedef void (*LogicalDecodeSequenceCB) (struct LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional,
+ int64 value);
+</programlisting>
+ The <parameter>txn</parameter> parameter contains meta information about
+ the transaction the sequence change is part of. Note however that for
+ non-transactional updates, the transaction may be NULL, depending on
+ if the transaction already has an XID assigned.
+ The <parameter>sequence_lsn</parameter> has the WAL location of the
+ sequence update. <parameter>transactional</parameter> determines whether
+ the sequence has to be replayed as part of the transaction or directly.
+
+ The <parameter>value</parameter> parameter describes the sequence change.
+ Note that this may not be the value obtained by the process updating the
+ process, but the future sequence value written to WAL (typically about
+ 32 values ahead).
+ </para>
+ </sect3>
+
<sect3 id="logicaldecoding-output-plugin-filter-prepare">
<title>Prepare Filter Callback</title>
@@ -1051,6 +1086,24 @@ typedef void (*LogicalDecodeStreamMessageCB) (struct LogicalDecodingContext *ctx
</para>
</sect3>
+ <sect3 id="logicaldecoding-output-plugin-stream-sequence">
+ <title>Stream Sequence Callback</title>
+ <para>
+ The optional <function>stream_sequence_cb</function> callback is called
+ for actions that change a sequence in a block of streamed changes
+ (demarcated by <function>stream_start_cb</function> and
+ <function>stream_stop_cb</function> calls).
+<programlisting>
+typedef void (*LogicalDecodeStreamSequenceCB) (struct LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional,
+ int64 value);
+</programlisting>
+ </para>
+ </sect3>
+
<sect3 id="logicaldecoding-output-plugin-stream-truncate">
<title>Stream Truncate Callback</title>
<para>
@@ -1230,8 +1283,9 @@ OutputPluginWrite(ctx, true);
in-progress transactions. There are multiple required streaming callbacks
(<function>stream_start_cb</function>, <function>stream_stop_cb</function>,
<function>stream_abort_cb</function>, <function>stream_commit_cb</function>
- and <function>stream_change_cb</function>) and two optional callbacks
- (<function>stream_message_cb</function> and <function>stream_truncate_cb</function>).
+ and <function>stream_change_cb</function>) and three optional callbacks
+ (<function>stream_message_cb</function>, <function>stream_sequence_cb</function>,
+ and <function>stream_truncate_cb</function>).
Also, if streaming of two-phase commands is to be supported, then additional
callbacks must be provided. (See <xref linkend="logicaldecoding-two-phase-commits"/>
for details).
diff --git a/src/backend/commands/sequence.c b/src/backend/commands/sequence.c
index da2ace79ccc..616b8685e39 100644
--- a/src/backend/commands/sequence.c
+++ b/src/backend/commands/sequence.c
@@ -204,6 +204,14 @@ DefineSequence(ParseState *pstate, CreateSeqStmt *seq)
stmt->tablespacename = NULL;
stmt->if_not_exists = seq->if_not_exists;
+ /*
+ * Make sure the relfilenode creation (in DefineRelation) is associated
+ * with the XID, so that decoding of sequences can handle this as a
+ * transactional change.
+ */
+ if (XLogLogicalInfoActive())
+ GetCurrentTransactionId();
+
address = DefineRelation(stmt, RELKIND_SEQUENCE, seq->ownerId, NULL, NULL);
seqoid = address.objectId;
Assert(seqoid != InvalidOid);
@@ -302,6 +310,13 @@ ResetSequence(Oid seq_relid)
seq->is_called = false;
seq->log_cnt = 0;
+ /*
+ * Make sure the relfilenode creation is associated with the XID, so that
+ * decoding of sequences can handle this as a transactional change.
+ */
+ if (XLogLogicalInfoActive())
+ GetCurrentTransactionId();
+
/*
* Create a new storage file for the sequence.
*/
@@ -392,8 +407,19 @@ fill_seq_fork_with_data(Relation rel, HeapTuple tuple, ForkNumber forkNum)
/* check the comment above nextval_internal()'s equivalent call. */
if (RelationNeedsWAL(rel))
+ {
GetTopTransactionId();
+ /*
+ * Make sure the subtransaction has a XID assigned, so that the
+ * sequence change WAL record is properly associated with it. This
+ * matters for sequences created/altered in a transaction - the
+ * subsequent changes need to be handled as transactional.
+ */
+ if (XLogLogicalInfoActive())
+ GetCurrentTransactionId();
+ }
+
START_CRIT_SECTION();
MarkBufferDirty(buf);
@@ -417,6 +443,9 @@ fill_seq_fork_with_data(Relation rel, HeapTuple tuple, ForkNumber forkNum)
XLogRegisterData((char *) &xlrec, sizeof(xl_seq_rec));
XLogRegisterData((char *) tuple->t_data, tuple->t_len);
+ /* allow filtering by origin on a sequence update */
+ XLogSetRecordFlags(XLOG_INCLUDE_ORIGIN);
+
recptr = XLogInsert(RM_SEQ_ID, XLOG_SEQ_LOG);
PageSetLSN(page, recptr);
@@ -499,6 +528,14 @@ AlterSequence(ParseState *pstate, AlterSeqStmt *stmt)
if (RelationNeedsWAL(seqrel))
GetTopTransactionId();
+ /*
+ * Make sure the relfilenode creation is associated with the XID, so
+ * that decoding of sequences can handle this as a transactional
+ * change.
+ */
+ if (XLogLogicalInfoActive())
+ GetCurrentTransactionId();
+
/*
* Create a new storage file for the sequence, making the state
* changes transactional.
@@ -548,8 +585,19 @@ SequenceChangePersistence(Oid relid, char newrelpersistence)
/* check the comment above nextval_internal()'s equivalent call. */
if (RelationNeedsWAL(seqrel))
+ {
GetTopTransactionId();
+ /*
+ * Make sure the subtransaction has a XID assigned, so that the
+ * sequence change WAL record is properly associated with it. This
+ * matters for sequences created/altered in a transaction - the
+ * subsequent changes need to be handled as transactional.
+ */
+ if (XLogLogicalInfoActive())
+ GetCurrentTransactionId();
+ }
+
(void) read_seq_tuple(seqrel, &buf, &seqdatatuple);
RelationSetNewRelfilenumber(seqrel, newrelpersistence);
fill_seq_with_data(seqrel, &seqdatatuple);
@@ -792,10 +840,34 @@ nextval_internal(Oid relid, bool check_permissions)
* It's sufficient to ensure the toplevel transaction has an xid, no need
* to assign xids subxacts, that'll already trigger an appropriate wait.
* (Have to do that here, so we're outside the critical section)
+ *
+ * We have to ensure we have a proper XID, which will be included in the
+ * XLOG record by XLogRecordAssemble. Otherwise the first nextval() in a
+ * subxact (without any preceding changes) would get XID 0, and it would
+ * then be impossible to decide which top xact it belongs to. It'd also
+ * trigger assert in DecodeSequence. We only do that with
+ * wal_level=logical, though.
+ *
+ * XXX This might seem unnecessary, because if there's no XID the xact
+ * couldn't have done anything important yet, e.g. it could not have
+ * created a sequence. But that's incorrect, because of subxacts. The
+ * current subtransaction might not have done anything yet (thus no XID),
+ * but an earlier one might have created the sequence.
*/
if (logit && RelationNeedsWAL(seqrel))
+ {
GetTopTransactionId();
+ /*
+ * Make sure the subtransaction has a XID assigned, so that the
+ * sequence change WAL record is properly associated with it. This
+ * matters for sequences created/altered in a transaction - the
+ * subsequent changes need to be handled as transactional.
+ */
+ if (XLogLogicalInfoActive())
+ GetCurrentTransactionId();
+ }
+
/* ready to change the on-disk (or really, in-buffer) tuple */
START_CRIT_SECTION();
@@ -835,6 +907,9 @@ nextval_internal(Oid relid, bool check_permissions)
XLogRegisterData((char *) &xlrec, sizeof(xl_seq_rec));
XLogRegisterData((char *) seqdatatuple.t_data, seqdatatuple.t_len);
+ /* allow filtering by origin on a sequence update */
+ XLogSetRecordFlags(XLOG_INCLUDE_ORIGIN);
+
recptr = XLogInsert(RM_SEQ_ID, XLOG_SEQ_LOG);
PageSetLSN(page, recptr);
@@ -996,8 +1071,19 @@ do_setval(Oid relid, int64 next, bool iscalled)
/* check the comment above nextval_internal()'s equivalent call. */
if (RelationNeedsWAL(seqrel))
+ {
GetTopTransactionId();
+ /*
+ * Make sure the subtransaction has a XID assigned, so that the
+ * sequence change WAL record is properly associated with it. This
+ * matters for sequences created/altered in a transaction - the
+ * subsequent changes need to be handled as transactional.
+ */
+ if (XLogLogicalInfoActive())
+ GetCurrentTransactionId();
+ }
+
/* ready to change the on-disk (or really, in-buffer) tuple */
START_CRIT_SECTION();
@@ -1021,6 +1107,9 @@ do_setval(Oid relid, int64 next, bool iscalled)
XLogRegisterData((char *) &xlrec, sizeof(xl_seq_rec));
XLogRegisterData((char *) seqdatatuple.t_data, seqdatatuple.t_len);
+ /* allow filtering by origin on a sequence update */
+ XLogSetRecordFlags(XLOG_INCLUDE_ORIGIN);
+
recptr = XLogInsert(RM_SEQ_ID, XLOG_SEQ_LOG);
PageSetLSN(page, recptr);
diff --git a/src/backend/replication/logical/decode.c b/src/backend/replication/logical/decode.c
index 1237118e84f..d48d88081f4 100644
--- a/src/backend/replication/logical/decode.c
+++ b/src/backend/replication/logical/decode.c
@@ -35,6 +35,7 @@
#include "access/xlogrecord.h"
#include "access/xlogutils.h"
#include "catalog/pg_control.h"
+#include "catalog/storage_xlog.h"
#include "replication/decode.h"
#include "replication/logical.h"
#include "replication/message.h"
@@ -42,6 +43,7 @@
#include "replication/reorderbuffer.h"
#include "replication/snapbuild.h"
#include "storage/standby.h"
+#include "commands/sequence.h"
/* individual record(group)'s handlers */
static void DecodeInsert(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
@@ -63,6 +65,7 @@ static void DecodePrepare(LogicalDecodingContext *ctx, XLogRecordBuffer *buf,
/* common function to decode tuples */
static void DecodeXLogTuple(char *data, Size len, ReorderBufferTupleBuf *tuple);
+static void DecodeSeqTuple(char *data, Size len, ReorderBufferTupleBuf *tuple);
/* helper functions for decoding transactions */
static inline bool FilterPrepare(LogicalDecodingContext *ctx,
@@ -629,7 +632,7 @@ logicalmsg_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
* We need to set processing_required flag to notify the message's
* existence to the caller. Usually, the flag is set when either the
* COMMIT or ABORT records are decoded, but this must be turned on
- * here because the non-transactional logical message is decoded
+ * here because the non-transactional sequence change is decoded
* without waiting for these records.
*/
if (!message->transactional)
@@ -1320,3 +1323,226 @@ DecodeTXNNeedSkip(LogicalDecodingContext *ctx, XLogRecordBuffer *buf,
return false;
}
+
+/*
+ * DecodeSeqTuple
+ * decode tuple describing the sequence change
+ *
+ * Sequences are represented as a table with a single row, which gets updated by
+ * nextval() etc. The tuple is stored in WAL right after the xl_seq_rec, so we
+ * simply copy it into the tuplebuf (similar to seq_redo).
+ */
+static void
+DecodeSeqTuple(char *data, Size len, ReorderBufferTupleBuf *tuple)
+{
+ int datalen = len - sizeof(xl_seq_rec) - SizeofHeapTupleHeader;
+
+ Assert(datalen >= 0);
+
+ tuple->tuple.t_len = datalen + SizeofHeapTupleHeader;
+
+ ItemPointerSetInvalid(&tuple->tuple.t_self);
+
+ tuple->tuple.t_tableOid = InvalidOid;
+
+ memcpy(((char *) tuple->tuple.t_data),
+ data + sizeof(xl_seq_rec),
+ SizeofHeapTupleHeader + datalen);
+}
+
+/*
+ * Handle sequence decode
+ *
+ * Decoding sequences is a bit tricky, because while most sequence actions
+ * are non-transactional (immediately visible and not subject to rollback),
+ * some need to be handled as transactional (e.g. for sequences created in
+ * a transaction are invisible until that transaction commits).
+ *
+ * By default, a sequence change is non-transactional - we must not queue
+ * it in a transaction as other changes, because the transaction might get
+ * rolled back and we'd discard the change. But that'd be incorrect, as the
+ * change is already visible to other processes accessing the sequence. If
+ * we discard the change, the downstream will not be notified about the
+ * change, and may fall behind.
+ *
+ * On the other hand, the sequence may be created in a transaction. In this
+ * case we *have to* queue it in the transaction just like other changes,
+ * because we don't want to process changes for sequences that may be unknown
+ * outside the transaction - the downstream might get confused about which
+ * sequence it's related to etc.
+ */
+void
+seq_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
+{
+ SnapBuild *builder = ctx->snapshot_builder;
+ ReorderBufferTupleBuf *tuplebuf;
+ RelFileLocator target_locator;
+ XLogReaderState *r = buf->record;
+ char *tupledata = NULL;
+ Size tuplelen;
+ Size datalen = 0;
+ TransactionId xid = XLogRecGetXid(r);
+ uint8 info = XLogRecGetInfo(buf->record) & ~XLR_INFO_MASK;
+ Snapshot snapshot = NULL;
+ RepOriginId origin_id = XLogRecGetOrigin(r);
+ bool transactional;
+
+ /* ignore sequences when the plugin does not have the callbacks */
+ if (!ctx->sequences)
+ return;
+
+ /* only decode changes flagged with XLOG_SEQ_LOG */
+ if (info != XLOG_SEQ_LOG)
+ elog(ERROR, "unexpected RM_SEQ_ID record type: %u", info);
+
+ ReorderBufferProcessXid(ctx->reorder, XLogRecGetXid(r), buf->origptr);
+
+ /*
+ * If we don't have snapshot or we are just fast-forwarding, there is no
+ * point in decoding sequences.
+ */
+ if (SnapBuildCurrentState(builder) < SNAPBUILD_FULL_SNAPSHOT)
+ return;
+
+ /* only interested in our database */
+ XLogRecGetBlockTag(r, 0, &target_locator, NULL, NULL);
+ if (target_locator.dbOid != ctx->slot->data.database)
+ return;
+
+ /* output plugin doesn't look for this origin, no need to queue */
+ if (FilterByOrigin(ctx, XLogRecGetOrigin(r)))
+ return;
+
+ /* time to decode the sequence data from the tuple */
+ tupledata = XLogRecGetData(r);
+ datalen = XLogRecGetDataLen(r);
+ tuplelen = datalen - SizeOfHeapHeader - sizeof(xl_seq_rec);
+
+ /* the sequence should not have changed without data */
+ if (!datalen || !tupledata)
+ elog(ERROR, "sequence decode missing tuple data");
+
+ tuplebuf = ReorderBufferGetTupleBuf(ctx->reorder, tuplelen);
+ DecodeSeqTuple(tupledata, datalen, tuplebuf);
+
+ /*
+ * Should we handle the sequence change as transactional or not?
+ *
+ * If the relfilenode was created in the current transaction, treat the
+ * change as transactional and queue it. Otherwise it needs to be treated
+ * as non-transactional, in which case we just send it to the plugin right
+ * away.
+ */
+ transactional = ReorderBufferSequenceIsTransactional(ctx->reorder,
+ target_locator,
+ NULL);
+
+ /* Skip the change if already processed (per the snapshot). */
+ if (transactional &&
+ !SnapBuildProcessChange(builder, xid, buf->origptr))
+ return;
+ else if (!transactional &&
+ (SnapBuildCurrentState(builder) != SNAPBUILD_CONSISTENT ||
+ SnapBuildXactNeedsSkip(builder, buf->origptr)))
+ return;
+
+ /*
+ * We also skip decoding in fast_forward mode. This check must be last
+ * because we don't want to set the processing_required flag unless we
+ * have a decodable sequence change.
+ */
+ if (ctx->fast_forward)
+ {
+ /*
+ * We need to set processing_required flag to notify the sequence
+ * change existence to the caller. Usually, the flag is set when
+ * either the COMMIT or ABORT records are decoded, but this must be
+ * turned on here because the non-transactional logical message is
+ * decoded without waiting for these records.
+ */
+ if (!transactional)
+ ctx->processing_required = true;
+
+ return;
+ }
+
+ /*
+ * If this is a non-transactional change, get the snapshot we're expected
+ * to use. We only get here when the snapshot is consistent, and the
+ * change is not meant to be skipped.
+ *
+ * For transactional changes we don't need a snapshot, we'll use the
+ * regular snapshot maintained by ReorderBuffer. We just leave it NULL.
+ */
+ if (!transactional)
+ snapshot = SnapBuildGetOrBuildSnapshot(builder);
+
+ /* Queue the change (or send immediately if not transactional). */
+ ReorderBufferQueueSequence(ctx->reorder, xid, snapshot, buf->endptr,
+ origin_id, target_locator, transactional,
+ tuplebuf);
+}
+
+/*
+ * Decode relfilenode change
+ *
+ * Decode SMGR records, so that we can later decide which sequence changes
+ * need to be treated as transactional. Most sequence changes are going to
+ * be non-transactional (applied as if outside the decoded transaction, not
+ * subject to rollback, etc.). Changes for sequences created/altered in the
+ * transaction need to be handled as transactional (i.e. applied as part of
+ * the decoded transaction, same as all other changes).
+ *
+ * To decide which of those cases is it we decode XLOG_SMGR_CREATE records
+ * and track relfilenodes created in each (sub)transaction. Changes for
+ * these relfilenodes are then queued and treated as transactional, while
+ * remaining changes are treated as non-transactional.
+ *
+ * We only care about XLOG_SMGR_CREATE records for "our" database (logical
+ * decoding is restricted to a single database), and we do the filtering
+ * and skipping, as appropriate.
+ */
+void
+smgr_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
+{
+ SnapBuild *builder = ctx->snapshot_builder;
+ XLogReaderState *r = buf->record;
+ TransactionId xid = XLogRecGetXid(r);
+ uint8 info = XLogRecGetInfo(buf->record) & ~XLR_INFO_MASK;
+ xl_smgr_create *xlrec;
+
+ /*
+ * Bail out when not decoding sequences, which is currently the only case
+ * when we need to know about relfilenodes created in a transaction.
+ */
+ if (!ctx->sequences)
+ return;
+
+ /*
+ * We only care about XLOG_SMGR_CREATE, because that's what determines if
+ * the following sequence changes are transactional.
+ */
+ if (info != XLOG_SMGR_CREATE)
+ return;
+
+ ReorderBufferProcessXid(ctx->reorder, XLogRecGetXid(r), buf->origptr);
+
+ /*
+ * If we don't have snapshot or we are just fast-forwarding, there is no
+ * point in decoding relfilenode information.
+ */
+ if (SnapBuildCurrentState(builder) < SNAPBUILD_FULL_SNAPSHOT ||
+ ctx->fast_forward)
+ return;
+
+ /* only interested in our database */
+ xlrec = (xl_smgr_create *) XLogRecGetData(r);
+ if (xlrec->rlocator.dbOid != ctx->slot->data.database)
+ return;
+
+ /* output plugin doesn't look for this origin, no need to queue */
+ if (FilterByOrigin(ctx, XLogRecGetOrigin(r)))
+ return;
+
+ ReorderBufferAddRelFileLocator(ctx->reorder, xid, buf->endptr, xlrec->rlocator);
+}
diff --git a/src/backend/replication/logical/logical.c b/src/backend/replication/logical/logical.c
index 8288da5277f..3a7b3b9be0c 100644
--- a/src/backend/replication/logical/logical.c
+++ b/src/backend/replication/logical/logical.c
@@ -75,6 +75,9 @@ static void truncate_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
static void message_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
XLogRecPtr message_lsn, bool transactional,
const char *prefix, Size message_size, const char *message);
+static void sequence_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ bool transactional, int64 value);
/* streaming callbacks */
static void stream_start_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
@@ -92,6 +95,9 @@ static void stream_change_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn
static void stream_message_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
XLogRecPtr message_lsn, bool transactional,
const char *prefix, Size message_size, const char *message);
+static void stream_sequence_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ bool transactional, int64 value);
static void stream_truncate_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
int nrelations, Relation relations[], ReorderBufferChange *change);
@@ -223,13 +229,14 @@ StartupDecodingContext(List *output_plugin_options,
ctx->reorder->apply_truncate = truncate_cb_wrapper;
ctx->reorder->commit = commit_cb_wrapper;
ctx->reorder->message = message_cb_wrapper;
+ ctx->reorder->sequence = sequence_cb_wrapper;
/*
* To support streaming, we require start/stop/abort/commit/change
- * callbacks. The message and truncate callbacks are optional, similar to
- * regular output plugins. We however enable streaming when at least one
- * of the methods is enabled so that we can easily identify missing
- * methods.
+ * callbacks. The message, sequence and truncate callbacks are optional,
+ * similar to regular output plugins. We however enable streaming when at
+ * least one of the methods is enabled so that we can easily identify
+ * missing methods.
*
* We decide it here, but only check it later in the wrappers.
*/
@@ -239,6 +246,7 @@ StartupDecodingContext(List *output_plugin_options,
(ctx->callbacks.stream_commit_cb != NULL) ||
(ctx->callbacks.stream_change_cb != NULL) ||
(ctx->callbacks.stream_message_cb != NULL) ||
+ (ctx->callbacks.stream_sequence_cb != NULL) ||
(ctx->callbacks.stream_truncate_cb != NULL);
/*
@@ -256,6 +264,7 @@ StartupDecodingContext(List *output_plugin_options,
ctx->reorder->stream_commit = stream_commit_cb_wrapper;
ctx->reorder->stream_change = stream_change_cb_wrapper;
ctx->reorder->stream_message = stream_message_cb_wrapper;
+ ctx->reorder->stream_sequence = stream_sequence_cb_wrapper;
ctx->reorder->stream_truncate = stream_truncate_cb_wrapper;
@@ -289,6 +298,13 @@ StartupDecodingContext(List *output_plugin_options,
*/
ctx->reorder->update_progress_txn = update_progress_txn_cb_wrapper;
+ /*
+ * To support logical decoding of sequences, we require the sequence
+ * callback. We decide it here, but only check it later in the wrappers.
+ */
+ ctx->sequences = ((ctx->callbacks.sequence_cb != NULL) ||
+ (ctx->callbacks.stream_sequence_cb != NULL));
+
ctx->out = makeStringInfo();
ctx->prepare_write = prepare_write;
ctx->write = do_write;
@@ -1262,6 +1278,42 @@ message_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
error_context_stack = errcallback.previous;
}
+static void
+sequence_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel, bool transactional,
+ int64 value)
+{
+ LogicalDecodingContext *ctx = cache->private_data;
+ LogicalErrorCallbackState state;
+ ErrorContextCallback errcallback;
+
+ Assert(!ctx->fast_forward);
+
+ if (ctx->callbacks.sequence_cb == NULL)
+ return;
+
+ /* Push callback + info on the error context stack */
+ state.ctx = ctx;
+ state.callback_name = "sequence";
+ state.report_location = sequence_lsn;
+ errcallback.callback = output_plugin_error_callback;
+ errcallback.arg = (void *) &state;
+ errcallback.previous = error_context_stack;
+ error_context_stack = &errcallback;
+
+ /* set output state */
+ ctx->accept_writes = true;
+ ctx->write_xid = txn != NULL ? txn->xid : InvalidTransactionId;
+ ctx->write_location = sequence_lsn;
+
+ /* do the actual work: call callback */
+ ctx->callbacks.sequence_cb(ctx, txn, sequence_lsn, rel,
+ transactional, value);
+
+ /* Pop the error context stack */
+ error_context_stack = errcallback.previous;
+}
+
static void
stream_start_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
XLogRecPtr first_lsn)
@@ -1577,6 +1629,46 @@ stream_message_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
error_context_stack = errcallback.previous;
}
+static void
+stream_sequence_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ bool transactional, int64 value)
+{
+ LogicalDecodingContext *ctx = cache->private_data;
+ LogicalErrorCallbackState state;
+ ErrorContextCallback errcallback;
+
+ Assert(!ctx->fast_forward);
+
+ /* We're only supposed to call this when streaming is supported. */
+ Assert(ctx->streaming);
+
+ /* this callback is optional */
+ if (ctx->callbacks.stream_sequence_cb == NULL)
+ return;
+
+ /* Push callback + info on the error context stack */
+ state.ctx = ctx;
+ state.callback_name = "stream_sequence";
+ state.report_location = sequence_lsn;
+ errcallback.callback = output_plugin_error_callback;
+ errcallback.arg = (void *) &state;
+ errcallback.previous = error_context_stack;
+ error_context_stack = &errcallback;
+
+ /* set output state */
+ ctx->accept_writes = true;
+ ctx->write_xid = txn != NULL ? txn->xid : InvalidTransactionId;
+ ctx->write_location = sequence_lsn;
+
+ /* do the actual work: call callback */
+ ctx->callbacks.sequence_cb(ctx, txn, sequence_lsn, rel,
+ transactional, value);
+
+ /* Pop the error context stack */
+ error_context_stack = errcallback.previous;
+}
+
static void
stream_truncate_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
int nrelations, Relation relations[],
diff --git a/src/backend/replication/logical/reorderbuffer.c b/src/backend/replication/logical/reorderbuffer.c
index 12edc5772a9..bb16ca4f7b9 100644
--- a/src/backend/replication/logical/reorderbuffer.c
+++ b/src/backend/replication/logical/reorderbuffer.c
@@ -77,6 +77,41 @@
* a bit more memory to the oldest subtransactions, because it's likely
* they are the source for the next sequence of changes.
*
+ * When decoding sequences, we differentiate between transactional and
+ * non-transactional changes. Sequence change is "non-transactional" if
+ * it was visible to outside transactions immediately, and would not be
+ * undone by a rollback. Most sequence changes are non-transactional, as
+ * the changes performed by nextval() are visible and can't be undone.
+ * Non-transactional changes are replayed immediately, as if outside any
+ * transaction (and thus also not subject to rollback etc.).
+ *
+ * However, if a transaction modifies the sequence in a way that creates
+ * a new relfilenode (e.g. by running ALTER SEQUENCE), we consider the
+ * subsequent changes "transactional" and queue them in the transaction
+ * that performed them, just like other changes. If that transaction rolls
+ * back, these sequence changes are discarded too (together with the
+ * new relfilenode).
+ *
+ * This mixed behavior is necessary - sequences are non-transactional
+ * (e.g. ROLLBACK does not undo that the sequence moved forward). But
+ * changes on new relfilenodes need to be handled as transactional.
+ *
+ * To decide if a sequence change is transactional, we maintain a hash
+ * table of relfilenodes created in each (sub)transaction, along with
+ * the XID of the (sub)transaction that created the relfilenode. The
+ * entries from subtransactions are copied to the top-level transaction
+ * to make checks cheaper. The hash table gets cleaned up when the
+ * transaction completes (commit/abort).
+ *
+ * The XID of the subxact that created the relfilenode is necessary, as
+ * that's where the sequence changes need to be queued - if this subxact
+ * rolls back, we want to discard the changes tied to this relfilenode
+ * (even if done in some other subtransaction).
+ *
+ * The XID may be valid even for non-transactional sequences - we simply
+ * keep the XID decoded from WAL, it's up to the reorderbuffer to decide
+ * if the change is transactional.
+ *
* -------------------------------------------------------------------------
*/
#include "postgres.h"
@@ -91,6 +126,7 @@
#include "access/xact.h"
#include "access/xlog_internal.h"
#include "catalog/catalog.h"
+#include "commands/sequence.h"
#include "lib/binaryheap.h"
#include "miscadmin.h"
#include "pgstat.h"
@@ -116,6 +152,13 @@ typedef struct ReorderBufferTXNByIdEnt
ReorderBufferTXN *txn;
} ReorderBufferTXNByIdEnt;
+/* entry for hash table we use to track sequences created in running xacts */
+typedef struct ReorderBufferSequenceEnt
+{
+ RelFileLocator rlocator;
+ ReorderBufferTXN *txn;
+} ReorderBufferSequenceEnt;
+
/* data structures for (relfilelocator, ctid) => (cmin, cmax) mapping */
typedef struct ReorderBufferTupleCidKey
{
@@ -225,6 +268,7 @@ static void ReorderBufferTransferSnapToParent(ReorderBufferTXN *txn,
ReorderBufferTXN *subtxn);
static void AssertTXNLsnOrder(ReorderBuffer *rb);
+static void AssertCheckSequences(ReorderBuffer *rb);
/* ---------------------------------------
* support functions for lsn-order iterating over the ->changes of a
@@ -456,12 +500,48 @@ ReorderBufferReturnTXN(ReorderBuffer *rb, ReorderBufferTXN *txn)
txn->invalidations = NULL;
}
+ if (txn->sequences_hash != NULL)
+ {
+ hash_destroy(txn->sequences_hash);
+ txn->sequences_hash = NULL;
+ }
+
/* Reset the toast hash */
ReorderBufferToastReset(rb, txn);
pfree(txn);
}
+/*
+ * Initialize hash table of relfilenodes created by the transaction.
+ *
+ * Each entry maps the relfilenode to the (sub)transaction that created the
+ * relfilenode - which is also the transaction the sequence change needs to
+ * be part of (in transactional case).
+ *
+ * We don't do this in ReorderBufferGetTXN because that'd allocate the hash
+ * for all transactions, and we expect new relfilenodes to be fairly rare.
+ * So only do that when adding the first entry.
+ */
+static void
+ReorderBufferTXNSequencesInit(ReorderBuffer *rb, ReorderBufferTXN *txn)
+{
+ HASHCTL hash_ctl;
+
+ /* bail out if already initialized */
+ if (txn->sequences_hash)
+ return;
+
+ /* hash table of sequences, mapping relfilelocator to transaction */
+ hash_ctl.keysize = sizeof(RelFileLocator);
+ hash_ctl.entrysize = sizeof(ReorderBufferSequenceEnt);
+ hash_ctl.hcxt = rb->context;
+
+ /* we expect relfilenodes to be created only rarely, so 32 seems enough */
+ txn->sequences_hash = hash_create("ReorderBufferTXNSequenceHash", 32, &hash_ctl,
+ HASH_ELEM | HASH_BLOBS | HASH_CONTEXT);
+}
+
/*
* Get a fresh ReorderBufferChange.
*/
@@ -536,6 +616,13 @@ ReorderBufferReturnChange(ReorderBuffer *rb, ReorderBufferChange *change,
change->data.truncate.relids = NULL;
}
break;
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
+ if (change->data.sequence.tuple)
+ {
+ ReorderBufferReturnTupleBuf(rb, change->data.sequence.tuple);
+ change->data.sequence.tuple = NULL;
+ }
+ break;
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM:
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_ABORT:
case REORDER_BUFFER_CHANGE_INTERNAL_COMMAND_ID:
@@ -804,7 +891,8 @@ ReorderBufferQueueChange(ReorderBuffer *rb, TransactionId xid, XLogRecPtr lsn,
change->action == REORDER_BUFFER_CHANGE_DELETE ||
change->action == REORDER_BUFFER_CHANGE_INTERNAL_SPEC_INSERT ||
change->action == REORDER_BUFFER_CHANGE_TRUNCATE ||
- change->action == REORDER_BUFFER_CHANGE_MESSAGE)
+ change->action == REORDER_BUFFER_CHANGE_MESSAGE ||
+ change->action == REORDER_BUFFER_CHANGE_SEQUENCE)
{
ReorderBufferTXN *toptxn = rbtxn_get_toptxn(txn);
@@ -895,6 +983,352 @@ ReorderBufferQueueMessage(ReorderBuffer *rb, TransactionId xid,
}
}
+/*
+ * Treat the sequence change as transactional?
+ *
+ * To decide if a sequence change should be handled as transactional or applied
+ * immediately, we need to decide if the relfilenode was created by a running
+ * transaction. Each transaction has a hash table of such relfilenodes.
+ *
+ * A naive approach would be to just loop through all transactions and check
+ * each of them, but there may be many (easily thousands of) subtransactions,
+ * and the check happens for each sequence change. So this could be very costly.
+ *
+ * To limit the number of transactions we need to check, we transfer the
+ * relfilenodes to the top-level xact, which allows us to check just the
+ * top-level xacts (and there should be a very limited number of those). We
+ * expect relfilenode creation to be a quite rare thing, so this should not
+ * waste too much memory. OTOH sequence changes are very common, so making
+ * the check cheaper seems like a good trade off.
+ *
+ * Returns true for transactional changes, false otherwise.
+ *
+ * Optionaly (when the "xid" parameter is set) returns XID of the (sub)xact
+ * to use for queueing transactional changes.
+ *
+ * XXX As an optimization, maybe we should try searching the current xact (or
+ * it's top-level xact) first. If the change is transactional, where we'd
+ * find the match.
+ */
+bool
+ReorderBufferSequenceIsTransactional(ReorderBuffer *rb,
+ RelFileLocator rlocator,
+ TransactionId *xid)
+{
+ bool found = false;
+ dlist_iter iter;
+
+ AssertCheckSequences(rb);
+
+ /*
+ * Walk all top-level transactions (some of which may be subxacts, except
+ * that we haven't processed the assignments yet), and check if any of
+ * them created the relfilenode.
+ */
+ dlist_foreach(iter, &rb->toplevel_by_lsn)
+ {
+ ReorderBufferSequenceEnt *entry;
+ ReorderBufferTXN *txn = dlist_container(ReorderBufferTXN, node,
+ iter.cur);
+
+ /* transaction has no relfilenodes at all */
+ if (!txn->sequences_hash)
+ continue;
+
+ entry = hash_search(txn->sequences_hash,
+ (void *) &rlocator,
+ HASH_FIND,
+ &found);
+
+ /*
+ * If we found an entry with matching relfilenode, we're done - we
+ * have to treat the sequence change as transactional, and replay it
+ * in the same (sub)transaction just like any other change.
+ *
+ * Optionally set XID of the (sub)xact that created the relfilenode.
+ */
+ if (found)
+ {
+ if (xid)
+ *xid = entry->txn->xid;
+
+ break;
+ }
+ }
+
+ return found;
+}
+
+/*
+ * Cleanup sequences after a subtransaction got aborted.
+ *
+ * The hash table will get destroyed in ReorderBufferReturnTXN, so we don't
+ * need to worry about that. But the entries were copied to the parent xact,
+ * and that's still being decoded - we make sure to remove the entries from
+ * the aborted one.
+ */
+static void
+ReorderBufferSequenceCleanup(ReorderBufferTXN *txn)
+{
+ HASH_SEQ_STATUS scan_status;
+ ReorderBufferSequenceEnt *ent;
+
+ /* Bail out if not a subxact, or if there are no entries. */
+ if (!rbtxn_is_known_subxact(txn))
+ return;
+
+ if (!txn->sequences_hash)
+ return;
+
+ /*
+ * Scan the top-level transaction hash and remove the entries from it. If
+ * we have entries for subxact, the top-level hash must have been
+ * initialized.
+ */
+ hash_seq_init(&scan_status, txn->sequences_hash);
+ while ((ent = (ReorderBufferSequenceEnt *) hash_seq_search(&scan_status)) != NULL)
+ {
+ if (hash_search(txn->toptxn->sequences_hash,
+ (void *) &ent->rlocator,
+ HASH_REMOVE, NULL) == NULL)
+ elog(ERROR, "hash table corrupted");
+ }
+}
+
+/*
+ * A transactional sequence change is queued to be processed upon commit
+ * and a non-transactional change gets processed immediately.
+ *
+ * A sequence update may be both transactional and non-transactional. When
+ * created in a running transaction, treat it as transactional and queue
+ * the change in it. Otherwise treat it as non-transactional, so that we
+ * don't forget the change in case of a rollback.
+ */
+void
+ReorderBufferQueueSequence(ReorderBuffer *rb, TransactionId xid,
+ Snapshot snapshot, XLogRecPtr lsn, RepOriginId origin_id,
+ RelFileLocator rlocator, bool transactional,
+ ReorderBufferTupleBuf *tuplebuf)
+{
+ AssertCheckSequences(rb);
+
+ /*
+ * Change needs to be handled as transactional, because the sequence was
+ * created in a transaction that is still seen as running. In that case
+ * all the changes need to be queued in that transaction, we must not send
+ * the changes to the downstream until the transaction commits.
+ *
+ * There's a bit of a trouble with subtransactions - we can't queue it
+ * into the subxact, because it might be rolled back and we'd lose the
+ * change (and that'd be wrong, as sequences are not transactional). But
+ * we also can't queue it into the top-level transaction, because the
+ * relfilenode creation is transactional - if that subxact gets aborted,
+ * we must throw away the changes too.
+ *
+ * So we need to queue it into the same (sub)xact that created the new
+ * sequence relfilenode, which is why we have the XID in the hash table.
+ */
+ if (transactional)
+ {
+ MemoryContext oldcontext;
+ ReorderBufferChange *change;
+
+ /* allocate and queue the transactional sequence change */
+ oldcontext = MemoryContextSwitchTo(rb->context);
+
+ change = ReorderBufferGetChange(rb);
+
+ change->action = REORDER_BUFFER_CHANGE_SEQUENCE;
+ change->origin_id = origin_id;
+
+ memcpy(&change->data.sequence.locator, &rlocator, sizeof(RelFileLocator));
+
+ change->data.sequence.tuple = tuplebuf;
+
+ /* lookup the XID for transaction that created the relfilenode */
+ ReorderBufferSequenceIsTransactional(rb, rlocator, &xid);
+
+ /* the XID should be valid for a transactional change */
+ Assert(TransactionIdIsValid(xid));
+
+ /* add it to the same (sub)xact that created that relfilenode */
+ ReorderBufferQueueChange(rb, xid, lsn, change, false);
+
+ MemoryContextSwitchTo(oldcontext);
+ }
+ else
+ {
+ /*
+ * This change is for a sequence that was not created in any running
+ * transaction, so we treat it as non-transactional and just send it
+ * to the output plugin directly.
+ */
+ ReorderBufferTXN *txn = NULL;
+ volatile Snapshot snapshot_now = snapshot;
+ bool using_subtxn;
+
+ /* non-transactional changes require a valid snapshot and XID */
+ Assert(snapshot_now);
+ Assert(TransactionIdIsValid(xid));
+
+ /* Make sure the sequence is not in any of the hash tables */
+ Assert(!ReorderBufferSequenceIsTransactional(rb, rlocator, NULL));
+
+ txn = ReorderBufferTXNByXid(rb, xid, true, NULL, lsn, true);
+
+ /* setup snapshot to allow catalog access */
+ SetupHistoricSnapshot(snapshot_now, NULL);
+
+ /*
+ * Decoding needs access to syscaches et al., which in turn use
+ * heavyweight locks and such. Thus we need to have enough state
+ * around to keep track of those. The easiest way is to simply use a
+ * transaction internally. That also allows us to easily enforce that
+ * nothing writes to the database by checking for xid assignments.
+ *
+ * When we're called via the SQL SRF there's already a transaction
+ * started, so start an explicit subtransaction there.
+ */
+ using_subtxn = IsTransactionOrTransactionBlock();
+
+ PG_TRY();
+ {
+ Relation relation;
+ HeapTuple tuple;
+ Form_pg_sequence_data seq;
+ Oid reloid;
+ int64 value;
+
+ if (using_subtxn)
+ BeginInternalSubTransaction("sequence");
+ else
+ StartTransactionCommand();
+
+ reloid = RelidByRelfilenumber(rlocator.spcOid, rlocator.relNumber);
+
+ if (reloid == InvalidOid)
+ elog(ERROR, "could not map filenode \"%s\" to relation OID",
+ relpathperm(rlocator,
+ MAIN_FORKNUM));
+
+ relation = RelationIdGetRelation(reloid);
+
+ if (!RelationIsValid(relation))
+ elog(ERROR, "could not open relation with OID %u (for filenode \"%s\")",
+ reloid,
+ relpathperm(rlocator, MAIN_FORKNUM));
+
+ tuple = &tuplebuf->tuple;
+ seq = (Form_pg_sequence_data) GETSTRUCT(tuple);
+
+ /*
+ * Calculate the first value of the next batch (at which point we
+ * generate and decode another WAL record).
+ */
+ value = seq->last_value;
+ value += (seq->is_called) ? seq->log_cnt : 0;
+
+ rb->sequence(rb, txn, lsn, relation, transactional, value);
+
+ RelationClose(relation);
+
+ TeardownHistoricSnapshot(false);
+
+ AbortCurrentTransaction();
+
+ if (using_subtxn)
+ RollbackAndReleaseCurrentSubTransaction();
+ }
+ PG_CATCH();
+ {
+ TeardownHistoricSnapshot(true);
+
+ AbortCurrentTransaction();
+
+ if (using_subtxn)
+ RollbackAndReleaseCurrentSubTransaction();
+
+ PG_RE_THROW();
+ }
+ PG_END_TRY();
+ }
+}
+
+/*
+ * ReorderBufferAddRelFileLocator
+ * Add newly created relfilenode to the hash table for transaction.
+ *
+ * If the transaction is already known to be a subtransaction, we add the same
+ * entry into the parent transaction (but it still points at the subxact, so
+ * that we know where to queue changes or what to discard in case of an abort).
+ */
+void
+ReorderBufferAddRelFileLocator(ReorderBuffer *rb, TransactionId xid,
+ XLogRecPtr lsn, RelFileLocator rlocator)
+{
+ bool found;
+ ReorderBufferSequenceEnt *entry;
+ ReorderBufferTXN *txn;
+
+ AssertCheckSequences(rb);
+
+ /*
+ * We only care about sequence relfilenodes for now, and those always have
+ * a XID. So if there's no XID, don't bother adding them to the hash.
+ */
+ if (!TransactionIdIsValid(xid))
+ return;
+
+ /*
+ * This might be the first change decoded for this transaction, so make
+ * sure we create it if needed.
+ */
+ txn = ReorderBufferTXNByXid(rb, xid, true, NULL, lsn, true);
+
+ /*
+ * First add it to the top-level transaction, but make sure it links to
+ * the correct subtransaction (so that we add later changes to it).
+ */
+ if (txn->toptxn)
+ {
+ /* make sure the hash table is initialized */
+ ReorderBufferTXNSequencesInit(rb, txn->toptxn);
+
+ /* search the lookup table */
+ entry = hash_search(txn->toptxn->sequences_hash,
+ (void *) &rlocator,
+ HASH_ENTER,
+ &found);
+
+ /*
+ * We've just decoded creation of the relfilenode, so if we found it
+ * in the hash table, something is wrong.
+ */
+ Assert(!found);
+
+ entry->txn = txn;
+ }
+
+ /* make sure the hash table is initialized */
+ ReorderBufferTXNSequencesInit(rb, txn);
+
+ /* search the lookup table */
+ entry = hash_search(txn->sequences_hash,
+ (void *) &rlocator,
+ HASH_ENTER,
+ &found);
+
+ /*
+ * We've just decoded creation of the relfilenode, so if we found it in
+ * the hash table, something is wrong.
+ */
+ Assert(!found);
+
+ entry->txn = txn;
+
+ AssertCheckSequences(rb);
+}
+
/*
* AssertTXNLsnOrder
* Verify LSN ordering of transaction lists in the reorderbuffer
@@ -969,6 +1403,161 @@ AssertTXNLsnOrder(ReorderBuffer *rb)
#endif
}
+/*
+ * AssertCheckSequences
+ * Verify consistency of hash tables tracking new relfilenodes.
+ *
+ * We check that the hash table in the top-level transaction is consistent with
+ * respect to the hash tables in it's subxacts. That is, each entry has to be
+ * either from the top-level xact or one of it's subtransactions. Likewise, each
+ * entry in subxact hash table has to have a match in the top-level hash table.
+ */
+static void
+AssertCheckSequences(ReorderBuffer *rb)
+{
+#ifdef USE_ASSERT_CHECKING
+ LogicalDecodingContext *ctx = rb->private_data;
+ dlist_iter iter;
+
+ /*
+ * Skip the verification if we don't reach the LSN at which we start
+ * decoding the contents of transactions yet because until we reach the
+ * LSN, we could have transactions that don't have the association between
+ * the top-level transaction and subtransaction yet and consequently have
+ * the same LSN. We don't guarantee this association until we try to
+ * decode the actual contents of transaction. The ordering of the records
+ * prior to the start_decoding_at LSN should have been checked before the
+ * restart.
+ */
+ if (SnapBuildXactNeedsSkip(ctx->snapshot_builder, ctx->reader->EndRecPtr))
+ return;
+
+ /*
+ * Make sure the relfilenodes from subxacts are properly recorded in the
+ * top-level transaction hash table.
+ */
+ dlist_foreach(iter, &rb->toplevel_by_lsn)
+ {
+ int nentries = 0,
+ nsubentries = 0;
+ dlist_iter subiter;
+ ReorderBufferTXN *txn = dlist_container(ReorderBufferTXN, node,
+ iter.cur);
+
+ /*
+ * We don't skip top-level transactions without relfilenodes, because
+ * there might be a subtransaction with some, and we want to detect
+ * such cases too.
+ */
+ if (txn->sequences_hash)
+ nentries = hash_get_num_entries(txn->sequences_hash);
+
+ /* walk all subxacts, and check the hash table in each one */
+ dlist_foreach(subiter, &txn->subtxns)
+ {
+ HASH_SEQ_STATUS scan_status;
+ ReorderBufferSequenceEnt *entry;
+
+ ReorderBufferTXN *subtxn = dlist_container(ReorderBufferTXN, node,
+ subiter.cur);
+
+ /*
+ * If this subxact has no relfilenodes, skip it. We'll do the
+ * check in the opposite direction (that all top-level
+ * relfilenodes are in the correct subxact) later.
+ */
+ if (!subtxn->sequences_hash)
+ continue;
+
+ /* add number of relfilenodes in this subxact */
+ nsubentries += hash_get_num_entries(subtxn->sequences_hash);
+
+ /*
+ * Check that all subxact relfilenodes are in the top-level txn
+ * too, and are pointing to this subtransaction.
+ */
+ hash_seq_init(&scan_status, subtxn->sequences_hash);
+ while ((entry = (ReorderBufferSequenceEnt *) hash_seq_search(&scan_status)) != NULL)
+ {
+ bool found = false;
+ ReorderBufferSequenceEnt *subentry;
+
+ /* search for the same relfilenode in the top-level txn */
+ subentry = hash_search(txn->sequences_hash,
+ (void *) &entry->rlocator,
+ HASH_FIND,
+ &found);
+
+ /*
+ * The top-level txn hash should have the relfilenode too, and
+ * it should point to this subxact.
+ */
+ Assert(found);
+
+ /*
+ * The entry has to point to the subxact - there's no subxact
+ * "below" this one to which the relfilenode could belong.
+ */
+ Assert(subentry->txn == subtxn);
+ }
+ }
+
+ /*
+ * We shouldn't have more relfilenodes in subtransactions than in the
+ * top-level one. There might be relfilenodes in the top-level one
+ * directly, so this needs to be inequality.
+ */
+ Assert(nentries >= nsubentries);
+
+ /*
+ * Now do the check in the opposite direction - check that every entry
+ * in the top-level txn (except those pointing to the top-level txn
+ * itself) point to one of the subxacts, and there's an entry in the
+ * subxact hash.
+ */
+ if (txn->sequences_hash)
+ {
+ HASH_SEQ_STATUS scan_status;
+ ReorderBufferSequenceEnt *entry;
+
+ hash_seq_init(&scan_status, txn->sequences_hash);
+ while ((entry = (ReorderBufferSequenceEnt *) hash_seq_search(&scan_status)) != NULL)
+ {
+ bool found = false;
+ ReorderBufferSequenceEnt *subentry;
+
+ /* Skip entries for the top-level txn itself. */
+ if (entry->txn == txn)
+ continue;
+
+ /* Is it a subxact of this txn? */
+ Assert(rbtxn_is_known_subxact(entry->txn));
+ Assert(entry->txn->toptxn == txn);
+
+ /*
+ * Search for the same relfilenode in the subxact (it should
+ * be initialized, as we expect it to contain the
+ * relfilenode).
+ */
+ subentry = hash_search(entry->txn->sequences_hash,
+ (void *) &entry->rlocator,
+ HASH_FIND,
+ &found);
+
+ Assert(found);
+
+ /*
+ * Check the txn pointer in the top-level hash table entry is
+ * consistent with the subxact hash table (we already checked
+ * the subxact entry points to the subxact itself).
+ */
+ Assert(subentry->txn = entry->txn);
+ }
+ }
+ }
+#endif
+}
+
/*
* AssertChangeLsnOrder
*
@@ -1094,6 +1683,9 @@ ReorderBufferAssignChild(ReorderBuffer *rb, TransactionId xid,
subtxn->toplevel_xid = xid;
Assert(subtxn->nsubtxns == 0);
+ /* There should be no sequence relfilenodes in the subxact yet. */
+ Assert(!subtxn->sequences_hash);
+
/* set the reference to top-level transaction */
subtxn->toptxn = txn;
@@ -1997,6 +2589,35 @@ ReorderBufferApplyMessage(ReorderBuffer *rb, ReorderBufferTXN *txn,
change->data.msg.message);
}
+/*
+ * Helper function for ReorderBufferProcessTXN for applying sequences.
+ */
+static inline void
+ReorderBufferApplySequence(ReorderBuffer *rb, ReorderBufferTXN *txn,
+ Relation relation, ReorderBufferChange *change,
+ bool streaming)
+{
+ HeapTuple tuple;
+ Form_pg_sequence_data seq;
+ int64 value;
+
+ tuple = &change->data.sequence.tuple->tuple;
+ seq = (Form_pg_sequence_data) GETSTRUCT(tuple);
+
+ value = seq->last_value;
+ value += (seq->is_called) ? seq->log_cnt : 0;
+
+ /*
+ * When called from ReorderBufferApplySequence, we're applying changes
+ * accumulated in a ReorderBufferTXN, so all those are transactional
+ * changes of sequences.
+ */
+ if (streaming)
+ rb->stream_sequence(rb, txn, change->lsn, relation, true, value);
+ else
+ rb->sequence(rb, txn, change->lsn, relation, true, value);
+}
+
/*
* Function to store the command id and snapshot at the end of the current
* stream so that we can reuse the same while sending the next stream.
@@ -2442,6 +3063,31 @@ ReorderBufferProcessTXN(ReorderBuffer *rb, ReorderBufferTXN *txn,
case REORDER_BUFFER_CHANGE_INTERNAL_TUPLECID:
elog(ERROR, "tuplecid value in changequeue");
break;
+
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
+ Assert(snapshot_now);
+
+ reloid = RelidByRelfilenumber(change->data.sequence.locator.spcOid,
+ change->data.sequence.locator.relNumber);
+
+ if (reloid == InvalidOid)
+ elog(ERROR, "could not map filenode \"%s\" to relation OID",
+ relpathperm(change->data.sequence.locator,
+ MAIN_FORKNUM));
+
+ relation = RelationIdGetRelation(reloid);
+
+ if (!RelationIsValid(relation))
+ elog(ERROR, "could not open relation with OID %u (for filenode \"%s\")",
+ reloid,
+ relpathperm(change->data.sequence.locator,
+ MAIN_FORKNUM));
+
+ if (RelationIsLogicallyLogged(relation))
+ ReorderBufferApplySequence(rb, txn, relation, change, streaming);
+
+ RelationClose(relation);
+ break;
}
/*
@@ -2872,6 +3518,10 @@ ReorderBufferFinishPrepared(ReorderBuffer *rb, TransactionId xid,
/* cleanup: make sure there's no cache pollution */
ReorderBufferExecuteInvalidations(txn->ninvalidations,
txn->invalidations);
+
+ /* cleanup: remove sequence relfilenodes from the top-level txn */
+ ReorderBufferSequenceCleanup(txn);
+
ReorderBufferCleanupTXN(rb, txn);
}
@@ -2921,8 +3571,13 @@ ReorderBufferAbort(ReorderBuffer *rb, TransactionId xid, XLogRecPtr lsn,
/* cosmetic... */
txn->final_lsn = lsn;
+ /* remove sequence relfilenodes from the top-level txn */
+ ReorderBufferSequenceCleanup(txn);
+
/* remove potential on-disk data, and deallocate */
ReorderBufferCleanupTXN(rb, txn);
+
+ AssertCheckSequences(rb);
}
/*
@@ -2958,8 +3613,13 @@ ReorderBufferAbortOld(ReorderBuffer *rb, TransactionId oldestRunningXid)
if (rbtxn_is_streamed(txn))
rb->stream_abort(rb, txn, InvalidXLogRecPtr);
+ /* remove sequence relfilenodes from the top-level txn */
+ ReorderBufferSequenceCleanup(txn);
+
/* remove potential on-disk data, and deallocate this tx */
ReorderBufferCleanupTXN(rb, txn);
+
+ AssertCheckSequences(rb);
}
else
return;
@@ -3008,6 +3668,9 @@ ReorderBufferForget(ReorderBuffer *rb, TransactionId xid, XLogRecPtr lsn)
else
Assert(txn->ninvalidations == 0);
+ /* remove sequence relfilenodes from a top-level txn */
+ ReorderBufferSequenceCleanup(txn);
+
/* remove potential on-disk data, and deallocate */
ReorderBufferCleanupTXN(rb, txn);
}
@@ -3909,6 +4572,39 @@ ReorderBufferSerializeChange(ReorderBuffer *rb, ReorderBufferTXN *txn,
memcpy(data, change->data.truncate.relids, size);
data += size;
+ break;
+ }
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
+ {
+ char *data;
+ ReorderBufferTupleBuf *tup;
+ Size len = 0;
+
+ tup = change->data.sequence.tuple;
+
+ if (tup)
+ {
+ sz += sizeof(HeapTupleData);
+ len = tup->tuple.t_len;
+ sz += len;
+ }
+
+ /* make sure we have enough space */
+ ReorderBufferSerializeReserve(rb, sz);
+
+ data = ((char *) rb->outbuf) + sizeof(ReorderBufferDiskChange);
+ /* might have been reallocated above */
+ ondisk = (ReorderBufferDiskChange *) rb->outbuf;
+
+ if (len)
+ {
+ memcpy(data, &tup->tuple, sizeof(HeapTupleData));
+ data += sizeof(HeapTupleData);
+
+ memcpy(data, tup->tuple.t_data, len);
+ data += len;
+ }
+
break;
}
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM:
@@ -4173,6 +4869,22 @@ ReorderBufferChangeSize(ReorderBufferChange *change)
{
sz += sizeof(Oid) * change->data.truncate.nrelids;
+ break;
+ }
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
+ {
+ ReorderBufferTupleBuf *tup;
+ Size len = 0;
+
+ tup = change->data.sequence.tuple;
+
+ if (tup)
+ {
+ sz += sizeof(HeapTupleData);
+ len = tup->tuple.t_len;
+ sz += len;
+ }
+
break;
}
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM:
@@ -4476,6 +5188,30 @@ ReorderBufferRestoreChange(ReorderBuffer *rb, ReorderBufferTXN *txn,
break;
}
+
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
+ if (change->data.sequence.tuple)
+ {
+ uint32 tuplelen = ((HeapTuple) data)->t_len;
+
+ change->data.sequence.tuple =
+ ReorderBufferGetTupleBuf(rb, tuplelen - SizeofHeapTupleHeader);
+
+ /* restore ->tuple */
+ memcpy(&change->data.sequence.tuple->tuple, data,
+ sizeof(HeapTupleData));
+ data += sizeof(HeapTupleData);
+
+ /* reset t_data pointer into the new tuplebuf */
+ change->data.sequence.tuple->tuple.t_data =
+ ReorderBufferTupleBufData(change->data.sequence.tuple);
+
+ /* restore tuple data itself */
+ memcpy(change->data.sequence.tuple->tuple.t_data, data, tuplelen);
+ data += tuplelen;
+ }
+ break;
+
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM:
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_ABORT:
case REORDER_BUFFER_CHANGE_INTERNAL_COMMAND_ID:
diff --git a/src/include/access/rmgrlist.h b/src/include/access/rmgrlist.h
index 463bcb67c57..305ca49a55f 100644
--- a/src/include/access/rmgrlist.h
+++ b/src/include/access/rmgrlist.h
@@ -27,7 +27,7 @@
/* symbol name, textual name, redo, desc, identify, startup, cleanup, mask, decode */
PG_RMGR(RM_XLOG_ID, "XLOG", xlog_redo, xlog_desc, xlog_identify, NULL, NULL, NULL, xlog_decode)
PG_RMGR(RM_XACT_ID, "Transaction", xact_redo, xact_desc, xact_identify, NULL, NULL, NULL, xact_decode)
-PG_RMGR(RM_SMGR_ID, "Storage", smgr_redo, smgr_desc, smgr_identify, NULL, NULL, NULL, NULL)
+PG_RMGR(RM_SMGR_ID, "Storage", smgr_redo, smgr_desc, smgr_identify, NULL, NULL, NULL, smgr_decode)
PG_RMGR(RM_CLOG_ID, "CLOG", clog_redo, clog_desc, clog_identify, NULL, NULL, NULL, NULL)
PG_RMGR(RM_DBASE_ID, "Database", dbase_redo, dbase_desc, dbase_identify, NULL, NULL, NULL, NULL)
PG_RMGR(RM_TBLSPC_ID, "Tablespace", tblspc_redo, tblspc_desc, tblspc_identify, NULL, NULL, NULL, NULL)
@@ -40,7 +40,7 @@ PG_RMGR(RM_BTREE_ID, "Btree", btree_redo, btree_desc, btree_identify, btree_xlog
PG_RMGR(RM_HASH_ID, "Hash", hash_redo, hash_desc, hash_identify, NULL, NULL, hash_mask, NULL)
PG_RMGR(RM_GIN_ID, "Gin", gin_redo, gin_desc, gin_identify, gin_xlog_startup, gin_xlog_cleanup, gin_mask, NULL)
PG_RMGR(RM_GIST_ID, "Gist", gist_redo, gist_desc, gist_identify, gist_xlog_startup, gist_xlog_cleanup, gist_mask, NULL)
-PG_RMGR(RM_SEQ_ID, "Sequence", seq_redo, seq_desc, seq_identify, NULL, NULL, seq_mask, NULL)
+PG_RMGR(RM_SEQ_ID, "Sequence", seq_redo, seq_desc, seq_identify, NULL, NULL, seq_mask, seq_decode)
PG_RMGR(RM_SPGIST_ID, "SPGist", spg_redo, spg_desc, spg_identify, spg_xlog_startup, spg_xlog_cleanup, spg_mask, NULL)
PG_RMGR(RM_BRIN_ID, "BRIN", brin_redo, brin_desc, brin_identify, NULL, NULL, brin_mask, NULL)
PG_RMGR(RM_COMMIT_TS_ID, "CommitTs", commit_ts_redo, commit_ts_desc, commit_ts_identify, NULL, NULL, NULL, NULL)
diff --git a/src/include/replication/decode.h b/src/include/replication/decode.h
index 14fa921ab47..8d06247165b 100644
--- a/src/include/replication/decode.h
+++ b/src/include/replication/decode.h
@@ -27,6 +27,8 @@ extern void heap2_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
extern void xact_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
extern void standby_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
extern void logicalmsg_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
+extern void seq_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
+extern void smgr_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
extern void LogicalDecodingProcessRecord(LogicalDecodingContext *ctx,
XLogReaderState *record);
diff --git a/src/include/replication/logical.h b/src/include/replication/logical.h
index dffc0d15648..72a17112efe 100644
--- a/src/include/replication/logical.h
+++ b/src/include/replication/logical.h
@@ -90,6 +90,11 @@ typedef struct LogicalDecodingContext
*/
bool twophase;
+ /*
+ * Does the output pluging support decoding of sequence changes?
+ */
+ bool sequences;
+
/*
* Is two-phase option given by output plugin?
*
diff --git a/src/include/replication/output_plugin.h b/src/include/replication/output_plugin.h
index 2ffcf17505f..7b5085f7d00 100644
--- a/src/include/replication/output_plugin.h
+++ b/src/include/replication/output_plugin.h
@@ -90,6 +90,16 @@ typedef void (*LogicalDecodeMessageCB) (struct LogicalDecodingContext *ctx,
Size message_size,
const char *message);
+/*
+ * Called for the generic logical decoding sequences.
+ */
+typedef void (*LogicalDecodeSequenceCB) (struct LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional,
+ int64 value);
+
/*
* Filter changes by origin.
*/
@@ -201,6 +211,17 @@ typedef void (*LogicalDecodeStreamMessageCB) (struct LogicalDecodingContext *ctx
Size message_size,
const char *message);
+/*
+ * Called for the streaming generic logical decoding sequences from in-progress
+ * transactions.
+ */
+typedef void (*LogicalDecodeStreamSequenceCB) (struct LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional,
+ int64 value);
+
/*
* Callback for streaming truncates from in-progress transactions.
*/
@@ -221,6 +242,7 @@ typedef struct OutputPluginCallbacks
LogicalDecodeTruncateCB truncate_cb;
LogicalDecodeCommitCB commit_cb;
LogicalDecodeMessageCB message_cb;
+ LogicalDecodeSequenceCB sequence_cb;
LogicalDecodeFilterByOriginCB filter_by_origin_cb;
LogicalDecodeShutdownCB shutdown_cb;
@@ -239,6 +261,7 @@ typedef struct OutputPluginCallbacks
LogicalDecodeStreamCommitCB stream_commit_cb;
LogicalDecodeStreamChangeCB stream_change_cb;
LogicalDecodeStreamMessageCB stream_message_cb;
+ LogicalDecodeStreamSequenceCB stream_sequence_cb;
LogicalDecodeStreamTruncateCB stream_truncate_cb;
} OutputPluginCallbacks;
diff --git a/src/include/replication/reorderbuffer.h b/src/include/replication/reorderbuffer.h
index f986101e50d..58a99b74060 100644
--- a/src/include/replication/reorderbuffer.h
+++ b/src/include/replication/reorderbuffer.h
@@ -74,6 +74,7 @@ typedef enum ReorderBufferChangeType
REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM,
REORDER_BUFFER_CHANGE_INTERNAL_SPEC_ABORT,
REORDER_BUFFER_CHANGE_TRUNCATE,
+ REORDER_BUFFER_CHANGE_SEQUENCE,
} ReorderBufferChangeType;
/* forward declaration */
@@ -167,6 +168,13 @@ typedef struct ReorderBufferChange
uint32 ninvalidations; /* Number of messages */
SharedInvalidationMessage *invalidations; /* invalidation message */
} inval;
+
+ /* Context data for Sequence changes */
+ struct
+ {
+ RelFileLocator locator;
+ ReorderBufferTupleBuf *tuple;
+ } sequence;
} data;
/*
@@ -392,6 +400,12 @@ typedef struct ReorderBufferTXN
*/
HTAB *toast_hash;
+ /*
+ * Sequence relfilenodes created in this transaction (also includes
+ * altered sequences, which assigns new relfilenode).
+ */
+ HTAB *sequences_hash;
+
/*
* non-hierarchical list of subtransactions that are *not* aborted. Only
* used in toplevel transactions.
@@ -470,6 +484,14 @@ typedef void (*ReorderBufferMessageCB) (ReorderBuffer *rb,
const char *prefix, Size sz,
const char *message);
+/* sequence callback signature */
+typedef void (*ReorderBufferSequenceCB) (ReorderBuffer *rb,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional,
+ int64 value);
+
/* begin prepare callback signature */
typedef void (*ReorderBufferBeginPrepareCB) (ReorderBuffer *rb,
ReorderBufferTXN *txn);
@@ -536,6 +558,14 @@ typedef void (*ReorderBufferStreamMessageCB) (
const char *prefix, Size sz,
const char *message);
+/* stream sequence callback signature */
+typedef void (*ReorderBufferStreamSequenceCB) (ReorderBuffer *rb,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional,
+ int64 value);
+
/* stream truncate callback signature */
typedef void (*ReorderBufferStreamTruncateCB) (
ReorderBuffer *rb,
@@ -592,6 +622,7 @@ struct ReorderBuffer
ReorderBufferApplyTruncateCB apply_truncate;
ReorderBufferCommitCB commit;
ReorderBufferMessageCB message;
+ ReorderBufferSequenceCB sequence;
/*
* Callbacks to be called when streaming a transaction at prepare time.
@@ -611,6 +642,7 @@ struct ReorderBuffer
ReorderBufferStreamCommitCB stream_commit;
ReorderBufferStreamChangeCB stream_change;
ReorderBufferStreamMessageCB stream_message;
+ ReorderBufferStreamSequenceCB stream_sequence;
ReorderBufferStreamTruncateCB stream_truncate;
/*
@@ -696,6 +728,10 @@ extern void ReorderBufferQueueMessage(ReorderBuffer *rb, TransactionId xid,
Snapshot snap, XLogRecPtr lsn,
bool transactional, const char *prefix,
Size message_size, const char *message);
+extern void ReorderBufferQueueSequence(ReorderBuffer *rb, TransactionId xid,
+ Snapshot snapshot, XLogRecPtr lsn, RepOriginId origin_id,
+ RelFileLocator locator, bool transactional,
+ ReorderBufferTupleBuf *tuplebuf);
extern void ReorderBufferCommit(ReorderBuffer *rb, TransactionId xid,
XLogRecPtr commit_lsn, XLogRecPtr end_lsn,
TimestampTz commit_time, RepOriginId origin_id, XLogRecPtr origin_lsn);
@@ -750,4 +786,10 @@ extern void ReorderBufferSetRestartPoint(ReorderBuffer *rb, XLogRecPtr ptr);
extern void StartupReorderBuffer(void);
+extern void ReorderBufferAddRelFileLocator(ReorderBuffer *rb, TransactionId xid,
+ XLogRecPtr lsn, RelFileLocator rlocator);
+extern bool ReorderBufferSequenceIsTransactional(ReorderBuffer *rb,
+ RelFileLocator locator,
+ TransactionId *xid);
+
#endif
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index d659adbfd6c..f4efc0192d5 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -2332,6 +2332,7 @@ ReorderBufferIterTXNState
ReorderBufferMessageCB
ReorderBufferPrepareCB
ReorderBufferRollbackPreparedCB
+ReorderBufferSequenceEnt
ReorderBufferStreamAbortCB
ReorderBufferStreamChangeCB
ReorderBufferStreamCommitCB
--
2.41.0
[text/x-patch] v20231203-0002-Add-decoding-of-sequences-to-test_decoding.patch (20.4K, ../../[email protected]/7-v20231203-0002-Add-decoding-of-sequences-to-test_decoding.patch)
download | inline diff:
From 3738cdf76da7ef0b52406297450cc5535d4bdd25 Mon Sep 17 00:00:00 2001
From: Tomas Vondra <[email protected]>
Date: Fri, 13 Jan 2023 20:32:44 +0100
Subject: [PATCH v20231203 2/4] Add decoding of sequences to test_decoding
Extends test_decoding with callbacks to decode sequences. This feature
is disabled by default (i.e. sequence changes are omitted from the
output), but may be enabled by passing 'include-sequences' option to the
test_decoding plugin.
Author: Tomas Vondra, Cary Huang
Reviewed-by: Ashutosh Bapat, Amit Kapila, Peter Eisentraut, Hannu
Krosing, Andres Freund, Zhijie Hou
Discussion: https://postgr.es/m/[email protected]
Discussion: https://postgr.es/m/[email protected]
---
contrib/test_decoding/Makefile | 3 +-
contrib/test_decoding/expected/sequence.out | 325 ++++++++++++++++++++
contrib/test_decoding/sql/sequence.sql | 119 +++++++
contrib/test_decoding/test_decoding.c | 87 ++++++
4 files changed, 533 insertions(+), 1 deletion(-)
create mode 100644 contrib/test_decoding/expected/sequence.out
create mode 100644 contrib/test_decoding/sql/sequence.sql
diff --git a/contrib/test_decoding/Makefile b/contrib/test_decoding/Makefile
index c7ce6037064..05c0c5a2f86 100644
--- a/contrib/test_decoding/Makefile
+++ b/contrib/test_decoding/Makefile
@@ -5,7 +5,8 @@ PGFILEDESC = "test_decoding - example of a logical decoding output plugin"
REGRESS = ddl xact rewrite toast permissions decoding_in_xact \
decoding_into_rel binary prepared replorigin time messages \
- spill slot truncate stream stats twophase twophase_stream
+ spill slot truncate stream stats twophase twophase_stream \
+ sequence
ISOLATION = mxact delayed_startup ondisk_startup concurrent_ddl_dml \
oldest_xmin snapshot_transfer subxact_without_top concurrent_stream \
twophase_snapshot slot_creation_error catalog_change_snapshot
diff --git a/contrib/test_decoding/expected/sequence.out b/contrib/test_decoding/expected/sequence.out
new file mode 100644
index 00000000000..98a8f25e5ad
--- /dev/null
+++ b/contrib/test_decoding/expected/sequence.out
@@ -0,0 +1,325 @@
+-- predictability
+SET synchronous_commit = on;
+SELECT 'init' FROM pg_create_logical_replication_slot('regression_slot', 'test_decoding');
+ ?column?
+----------
+ init
+(1 row)
+
+CREATE SEQUENCE test_sequence;
+-- test the sequence changes by several nextval() calls
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 1
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 2
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 4
+(1 row)
+
+-- test the sequence changes by several ALTER commands
+ALTER SEQUENCE test_sequence INCREMENT BY 10;
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 14
+(1 row)
+
+ALTER SEQUENCE test_sequence START WITH 3000;
+ALTER SEQUENCE test_sequence MAXVALUE 10000;
+ALTER SEQUENCE test_sequence RESTART WITH 4000;
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 4000
+(1 row)
+
+-- test the sequence changes by several setval() calls
+SELECT setval('test_sequence', 3500);
+ setval
+--------
+ 3500
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3510
+(1 row)
+
+SELECT setval('test_sequence', 3500, true);
+ setval
+--------
+ 3500
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3510
+(1 row)
+
+SELECT setval('test_sequence', 3500, false);
+ setval
+--------
+ 3500
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3500
+(1 row)
+
+-- show results and drop sequence
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '1');
+ data
+-------------------------------------------------------------
+ BEGIN
+ sequence public.test_sequence: transactional: 1 value: 1
+ COMMIT
+ sequence public.test_sequence: transactional: 0 value: 33
+ BEGIN
+ sequence public.test_sequence: transactional: 1 value: 4
+ COMMIT
+ sequence public.test_sequence: transactional: 0 value: 334
+ BEGIN
+ sequence public.test_sequence: transactional: 1 value: 46
+ COMMIT
+ BEGIN
+ sequence public.test_sequence: transactional: 1 value: 14
+ COMMIT
+ BEGIN
+ sequence public.test_sequence: transactional: 1 value: 4000
+ COMMIT
+ sequence public.test_sequence: transactional: 0 value: 4320
+ sequence public.test_sequence: transactional: 0 value: 3500
+ sequence public.test_sequence: transactional: 0 value: 3830
+ sequence public.test_sequence: transactional: 0 value: 3500
+ sequence public.test_sequence: transactional: 0 value: 3830
+ sequence public.test_sequence: transactional: 0 value: 3500
+ sequence public.test_sequence: transactional: 0 value: 3820
+(24 rows)
+
+DROP SEQUENCE test_sequence;
+-- rollback on sequence creation and update
+BEGIN;
+CREATE SEQUENCE test_sequence;
+CREATE TABLE test_table (a INT);
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 1
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 2
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3
+(1 row)
+
+SELECT setval('test_sequence', 3000);
+ setval
+--------
+ 3000
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3001
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3002
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3003
+(1 row)
+
+ALTER SEQUENCE test_sequence RESTART WITH 6000;
+INSERT INTO test_table VALUES( (SELECT nextval('test_sequence')) );
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 6001
+(1 row)
+
+ROLLBACK;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '1');
+ data
+------
+(0 rows)
+
+-- rollback on table creation with serial column
+BEGIN;
+CREATE TABLE test_table (a SERIAL, b INT);
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+INSERT INTO test_table (b) VALUES (300);
+SELECT setval('test_table_a_seq', 3000);
+ setval
+--------
+ 3000
+(1 row)
+
+INSERT INTO test_table (b) VALUES (400);
+INSERT INTO test_table (b) VALUES (500);
+INSERT INTO test_table (b) VALUES (600);
+ALTER SEQUENCE test_table_a_seq RESTART WITH 6000;
+INSERT INTO test_table (b) VALUES (700);
+INSERT INTO test_table (b) VALUES (800);
+INSERT INTO test_table (b) VALUES (900);
+ROLLBACK;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '1');
+ data
+------
+(0 rows)
+
+-- rollback on table with serial column
+CREATE TABLE test_table (a SERIAL, b INT);
+BEGIN;
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+INSERT INTO test_table (b) VALUES (300);
+SELECT setval('test_table_a_seq', 3000);
+ setval
+--------
+ 3000
+(1 row)
+
+INSERT INTO test_table (b) VALUES (400);
+INSERT INTO test_table (b) VALUES (500);
+INSERT INTO test_table (b) VALUES (600);
+ALTER SEQUENCE test_table_a_seq RESTART WITH 6000;
+INSERT INTO test_table (b) VALUES (700);
+INSERT INTO test_table (b) VALUES (800);
+INSERT INTO test_table (b) VALUES (900);
+ROLLBACK;
+-- check table and sequence values after rollback
+SELECT * from test_table_a_seq;
+ last_value | log_cnt | is_called
+------------+---------+-----------
+ 3003 | 30 | t
+(1 row)
+
+SELECT nextval('test_table_a_seq');
+ nextval
+---------
+ 3004
+(1 row)
+
+DROP TABLE test_table;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '1');
+ data
+----------------------------------------------------------------
+ BEGIN
+ sequence public.test_table_a_seq: transactional: 1 value: 1
+ COMMIT
+ sequence public.test_table_a_seq: transactional: 0 value: 33
+ sequence public.test_table_a_seq: transactional: 0 value: 3000
+ sequence public.test_table_a_seq: transactional: 0 value: 3033
+(6 rows)
+
+-- savepoint test on table with serial column
+BEGIN;
+CREATE TABLE test_table (a SERIAL, b INT);
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+SAVEPOINT a;
+INSERT INTO test_table (b) VALUES (300);
+ROLLBACK TO SAVEPOINT a;
+DROP TABLE test_table;
+COMMIT;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '1');
+ data
+--------------------------------------------------------------
+ BEGIN
+ sequence public.test_table_a_seq: transactional: 1 value: 1
+ sequence public.test_table_a_seq: transactional: 1 value: 33
+ table public.test_table: INSERT: a[integer]:1 b[integer]:100
+ table public.test_table: INSERT: a[integer]:2 b[integer]:200
+ COMMIT
+(6 rows)
+
+-- savepoint test on table with serial column
+BEGIN;
+CREATE SEQUENCE test_sequence;
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 1
+(1 row)
+
+SELECT setval('test_sequence', 3000);
+ setval
+--------
+ 3000
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3001
+(1 row)
+
+SAVEPOINT a;
+ALTER SEQUENCE test_sequence START WITH 7000;
+SELECT setval('test_sequence', 5000);
+ setval
+--------
+ 5000
+(1 row)
+
+ROLLBACK TO SAVEPOINT a;
+SELECT * FROM test_sequence;
+ last_value | log_cnt | is_called
+------------+---------+-----------
+ 3001 | 32 | t
+(1 row)
+
+DROP SEQUENCE test_sequence;
+COMMIT;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '1');
+ data
+-------------------------------------------------------------
+ BEGIN
+ sequence public.test_sequence: transactional: 1 value: 1
+ sequence public.test_sequence: transactional: 1 value: 33
+ sequence public.test_sequence: transactional: 1 value: 3000
+ sequence public.test_sequence: transactional: 1 value: 3033
+ COMMIT
+(6 rows)
+
+SELECT pg_drop_replication_slot('regression_slot');
+ pg_drop_replication_slot
+--------------------------
+
+(1 row)
+
diff --git a/contrib/test_decoding/sql/sequence.sql b/contrib/test_decoding/sql/sequence.sql
new file mode 100644
index 00000000000..c19a5d50747
--- /dev/null
+++ b/contrib/test_decoding/sql/sequence.sql
@@ -0,0 +1,119 @@
+-- predictability
+SET synchronous_commit = on;
+SELECT 'init' FROM pg_create_logical_replication_slot('regression_slot', 'test_decoding');
+
+CREATE SEQUENCE test_sequence;
+
+-- test the sequence changes by several nextval() calls
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+
+-- test the sequence changes by several ALTER commands
+ALTER SEQUENCE test_sequence INCREMENT BY 10;
+SELECT nextval('test_sequence');
+
+ALTER SEQUENCE test_sequence START WITH 3000;
+ALTER SEQUENCE test_sequence MAXVALUE 10000;
+ALTER SEQUENCE test_sequence RESTART WITH 4000;
+SELECT nextval('test_sequence');
+
+-- test the sequence changes by several setval() calls
+SELECT setval('test_sequence', 3500);
+SELECT nextval('test_sequence');
+SELECT setval('test_sequence', 3500, true);
+SELECT nextval('test_sequence');
+SELECT setval('test_sequence', 3500, false);
+SELECT nextval('test_sequence');
+
+-- show results and drop sequence
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '1');
+DROP SEQUENCE test_sequence;
+
+-- rollback on sequence creation and update
+BEGIN;
+CREATE SEQUENCE test_sequence;
+CREATE TABLE test_table (a INT);
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT setval('test_sequence', 3000);
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+ALTER SEQUENCE test_sequence RESTART WITH 6000;
+INSERT INTO test_table VALUES( (SELECT nextval('test_sequence')) );
+SELECT nextval('test_sequence');
+ROLLBACK;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '1');
+
+-- rollback on table creation with serial column
+BEGIN;
+CREATE TABLE test_table (a SERIAL, b INT);
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+INSERT INTO test_table (b) VALUES (300);
+SELECT setval('test_table_a_seq', 3000);
+INSERT INTO test_table (b) VALUES (400);
+INSERT INTO test_table (b) VALUES (500);
+INSERT INTO test_table (b) VALUES (600);
+ALTER SEQUENCE test_table_a_seq RESTART WITH 6000;
+INSERT INTO test_table (b) VALUES (700);
+INSERT INTO test_table (b) VALUES (800);
+INSERT INTO test_table (b) VALUES (900);
+ROLLBACK;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '1');
+
+-- rollback on table with serial column
+CREATE TABLE test_table (a SERIAL, b INT);
+
+BEGIN;
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+INSERT INTO test_table (b) VALUES (300);
+SELECT setval('test_table_a_seq', 3000);
+INSERT INTO test_table (b) VALUES (400);
+INSERT INTO test_table (b) VALUES (500);
+INSERT INTO test_table (b) VALUES (600);
+ALTER SEQUENCE test_table_a_seq RESTART WITH 6000;
+INSERT INTO test_table (b) VALUES (700);
+INSERT INTO test_table (b) VALUES (800);
+INSERT INTO test_table (b) VALUES (900);
+ROLLBACK;
+
+-- check table and sequence values after rollback
+SELECT * from test_table_a_seq;
+SELECT nextval('test_table_a_seq');
+
+DROP TABLE test_table;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '1');
+
+-- savepoint test on table with serial column
+BEGIN;
+CREATE TABLE test_table (a SERIAL, b INT);
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+SAVEPOINT a;
+INSERT INTO test_table (b) VALUES (300);
+ROLLBACK TO SAVEPOINT a;
+DROP TABLE test_table;
+COMMIT;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '1');
+
+-- savepoint test on table with serial column
+BEGIN;
+CREATE SEQUENCE test_sequence;
+SELECT nextval('test_sequence');
+SELECT setval('test_sequence', 3000);
+SELECT nextval('test_sequence');
+SAVEPOINT a;
+ALTER SEQUENCE test_sequence START WITH 7000;
+SELECT setval('test_sequence', 5000);
+ROLLBACK TO SAVEPOINT a;
+SELECT * FROM test_sequence;
+DROP SEQUENCE test_sequence;
+COMMIT;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '1');
+
+SELECT pg_drop_replication_slot('regression_slot');
diff --git a/contrib/test_decoding/test_decoding.c b/contrib/test_decoding/test_decoding.c
index 288fd0bb4ab..0a781edffc1 100644
--- a/contrib/test_decoding/test_decoding.c
+++ b/contrib/test_decoding/test_decoding.c
@@ -29,6 +29,7 @@ typedef struct
MemoryContext context;
bool include_xids;
bool include_timestamp;
+ bool include_sequences;
bool skip_empty_xacts;
bool only_local;
} TestDecodingData;
@@ -72,6 +73,10 @@ static void pg_decode_message(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn, XLogRecPtr lsn,
bool transactional, const char *prefix,
Size sz, const char *message);
+static void pg_decode_sequence(LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn, XLogRecPtr sequence_lsn,
+ Relation rel, bool transactional,
+ int64 value);
static bool pg_decode_filter_prepare(LogicalDecodingContext *ctx,
TransactionId xid,
const char *gid);
@@ -112,6 +117,10 @@ static void pg_decode_stream_message(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn, XLogRecPtr lsn,
bool transactional, const char *prefix,
Size sz, const char *message);
+static void pg_decode_stream_sequence(LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn, XLogRecPtr sequence_lsn,
+ Relation rel, bool transactional,
+ int64 value);
static void pg_decode_stream_truncate(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn,
int nrelations, Relation relations[],
@@ -135,6 +144,7 @@ _PG_output_plugin_init(OutputPluginCallbacks *cb)
cb->filter_by_origin_cb = pg_decode_filter;
cb->shutdown_cb = pg_decode_shutdown;
cb->message_cb = pg_decode_message;
+ cb->sequence_cb = pg_decode_sequence;
cb->filter_prepare_cb = pg_decode_filter_prepare;
cb->begin_prepare_cb = pg_decode_begin_prepare_txn;
cb->prepare_cb = pg_decode_prepare_txn;
@@ -147,6 +157,7 @@ _PG_output_plugin_init(OutputPluginCallbacks *cb)
cb->stream_commit_cb = pg_decode_stream_commit;
cb->stream_change_cb = pg_decode_stream_change;
cb->stream_message_cb = pg_decode_stream_message;
+ cb->stream_sequence_cb = pg_decode_stream_sequence;
cb->stream_truncate_cb = pg_decode_stream_truncate;
}
@@ -168,6 +179,7 @@ pg_decode_startup(LogicalDecodingContext *ctx, OutputPluginOptions *opt,
data->include_timestamp = false;
data->skip_empty_xacts = false;
data->only_local = false;
+ data->include_sequences = false;
ctx->output_plugin_private = data;
@@ -259,6 +271,17 @@ pg_decode_startup(LogicalDecodingContext *ctx, OutputPluginOptions *opt,
errmsg("could not parse value \"%s\" for parameter \"%s\"",
strVal(elem->arg), elem->defname)));
}
+ else if (strcmp(elem->defname, "include-sequences") == 0)
+ {
+
+ if (elem->arg == NULL)
+ data->include_sequences = true;
+ else if (!parse_bool(strVal(elem->arg), &data->include_sequences))
+ ereport(ERROR,
+ (errcode(ERRCODE_INVALID_PARAMETER_VALUE),
+ errmsg("could not parse value \"%s\" for parameter \"%s\"",
+ strVal(elem->arg), elem->defname)));
+ }
else
{
ereport(ERROR,
@@ -762,6 +785,38 @@ pg_decode_message(LogicalDecodingContext *ctx,
OutputPluginWrite(ctx, true);
}
+static void
+pg_decode_sequence(LogicalDecodingContext *ctx, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ bool transactional,
+ int64 value)
+{
+ TestDecodingData *data = ctx->output_plugin_private;
+ TestDecodingTxnData *txndata = txn->output_plugin_private;
+
+ if (!data->include_sequences)
+ return;
+
+ /* output BEGIN if we haven't yet, but only for the transactional case */
+ if (transactional)
+ {
+ if (data->skip_empty_xacts && !txndata->xact_wrote_changes)
+ {
+ pg_output_begin(ctx, data, txn, false);
+ }
+ txndata->xact_wrote_changes = true;
+ }
+
+ OutputPluginPrepareWrite(ctx, true);
+ appendStringInfoString(ctx->out, "sequence ");
+ appendStringInfoString(ctx->out,
+ quote_qualified_identifier(get_namespace_name(get_rel_namespace(RelationGetRelid(rel))),
+ RelationGetRelationName(rel)));
+ appendStringInfo(ctx->out, ": transactional: %d value: " INT64_FORMAT,
+ transactional, value);
+ OutputPluginWrite(ctx, true);
+}
+
static void
pg_decode_stream_start(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn)
@@ -974,6 +1029,38 @@ pg_decode_stream_message(LogicalDecodingContext *ctx,
OutputPluginWrite(ctx, true);
}
+static void
+pg_decode_stream_sequence(LogicalDecodingContext *ctx, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ bool transactional,
+ int64 value)
+{
+ TestDecodingData *data = ctx->output_plugin_private;
+ TestDecodingTxnData *txndata = txn->output_plugin_private;
+
+ if (!data->include_sequences)
+ return;
+
+ /* output BEGIN if we haven't yet, but only for the transactional case */
+ if (transactional)
+ {
+ if (data->skip_empty_xacts && !txndata->xact_wrote_changes)
+ {
+ pg_output_stream_start(ctx, data, txn, false);
+ }
+ txndata->xact_wrote_changes = true;
+ }
+
+ OutputPluginPrepareWrite(ctx, true);
+ appendStringInfoString(ctx->out, "streaming sequence ");
+ appendStringInfoString(ctx->out,
+ quote_qualified_identifier(get_namespace_name(get_rel_namespace(RelationGetRelid(rel))),
+ RelationGetRelationName(rel)));
+ appendStringInfo(ctx->out, ": transactional: %d value: " INT64_FORMAT,
+ transactional, value);
+ OutputPluginWrite(ctx, true);
+}
+
/*
* In streaming mode, we don't display the detailed information of Truncate.
* See pg_decode_stream_change.
--
2.41.0
[text/x-patch] v20231203-0003-Add-decoding-of-sequences-to-built-in-repl.patch (269.4K, ../../[email protected]/8-v20231203-0003-Add-decoding-of-sequences-to-built-in-repl.patch)
download | inline diff:
From fecf3e0cd332a227fce0635a16be509b8e6e2f6c Mon Sep 17 00:00:00 2001
From: Tomas Vondra <[email protected]>
Date: Sun, 26 Nov 2023 15:40:15 +0100
Subject: [PATCH v20231203 3/4] Add decoding of sequences to built-in
replication
This commit adds support for decoding of sequences to the built-in
replication (the infrastructure was added by commit 0da92dc530).
The syntax and behavior mostly mimics handling of tables, i.e. a
publication may be defined as FOR ALL SEQUENCES (replicating all
sequences in a database), FOR ALL SEQUENCES IN SCHEMA (replicating
all sequences in a particular schema) or individual sequences.
To publish sequence modifications, the publication has to include
'sequence' action. The protocol is extended with a new message,
describing sequence increments.
A new system view pg_publication_sequences lists all the sequences
added to a publication, both directly and indirectly. Various psql
commands (\d and \dRp) are improved to also display publications
including a given sequence, or sequences included in a publication.
Author: Tomas Vondra, Cary Huang, Masahiko Sawada
Reviewed-by: Ashutosh Bapat, Amit Kapila, Peter Eisentraut, Amit Kapila,
Hannu Krosing, Andres Freund, Petr Jelinek, Zhijie Hou
Discussion: https://postgr.es/m/[email protected]
Discussion: https://postgr.es/m/[email protected]
Discussion: https://postgr.es/m/CAD21AoDtKpdJcHOLjfPQ7TmpFqNB5__%3DQ_g1e8OBRrwT5LP-%3Dg%40mail.gmail.com
---
doc/src/sgml/catalogs.sgml | 10 +
doc/src/sgml/protocol.sgml | 80 ++
doc/src/sgml/ref/alter_publication.sgml | 25 +-
doc/src/sgml/ref/alter_subscription.sgml | 8 +-
doc/src/sgml/ref/create_publication.sgml | 51 +-
doc/src/sgml/system-views.sgml | 71 ++
src/backend/catalog/objectaddress.c | 45 +-
src/backend/catalog/pg_publication.c | 336 +++++-
src/backend/catalog/system_views.sql | 10 +
src/backend/commands/publicationcmds.c | 422 +++++--
src/backend/commands/sequence.c | 248 +++-
src/backend/commands/subscriptioncmds.c | 101 +-
src/backend/executor/execReplication.c | 4 +-
src/backend/parser/gram.y | 73 +-
src/backend/replication/logical/proto.c | 55 +
src/backend/replication/logical/tablesync.c | 135 ++-
src/backend/replication/logical/worker.c | 97 ++
src/backend/replication/pgoutput/pgoutput.c | 108 +-
src/backend/utils/cache/relcache.c | 28 +-
src/backend/utils/cache/syscache.c | 5 +-
src/bin/pg_dump/pg_dump.c | 64 +-
src/bin/pg_dump/pg_dump.h | 3 +
src/bin/pg_dump/t/002_pg_dump.pl | 47 +-
src/bin/psql/describe.c | 297 +++--
src/bin/psql/tab-complete.c | 39 +-
src/include/catalog/pg_proc.dat | 13 +
src/include/catalog/pg_publication.h | 26 +-
.../catalog/pg_publication_namespace.h | 10 +-
src/include/commands/sequence.h | 1 +
src/include/nodes/parsenodes.h | 10 +-
src/include/replication/logicalproto.h | 19 +-
src/test/regress/expected/object_address.out | 20 +-
src/test/regress/expected/psql.out | 6 +-
src/test/regress/expected/publication.out | 1055 +++++++++++++----
src/test/regress/expected/rules.out | 8 +
src/test/regress/sql/object_address.sql | 5 +-
src/test/regress/sql/publication.sql | 231 +++-
src/test/subscription/t/034_sequences.pl | 344 ++++++
src/tools/pgindent/typedefs.list | 1 +
39 files changed, 3589 insertions(+), 522 deletions(-)
create mode 100644 src/test/subscription/t/034_sequences.pl
diff --git a/doc/src/sgml/catalogs.sgml b/doc/src/sgml/catalogs.sgml
index 3ec7391ec5e..d61fa014316 100644
--- a/doc/src/sgml/catalogs.sgml
+++ b/doc/src/sgml/catalogs.sgml
@@ -6425,6 +6425,16 @@ SCRAM-SHA-256$<replaceable><iteration count></replaceable>:<replaceable>&l
Reference to schema
</para></entry>
</row>
+
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ <structfield>pntype</structfield> <type>char</type>
+ </para>
+ <para>
+ Determines which object type is included from this schema.
+ <literal>t</literal> for tables, <literal>s</literal> for sequences
+ </para></entry>
+ </row>
</tbody>
</tgroup>
</table>
diff --git a/doc/src/sgml/protocol.sgml b/doc/src/sgml/protocol.sgml
index af3f016f746..bfff15e0407 100644
--- a/doc/src/sgml/protocol.sgml
+++ b/doc/src/sgml/protocol.sgml
@@ -6376,6 +6376,86 @@ psql "dbname=postgres replication=database" -c "IDENTIFY_SYSTEM;"
</listitem>
</varlistentry>
+ <varlistentry id="protocol-logicalrep-message-formats-Sequence">
+ <term>Sequence</term>
+ <listitem>
+ <variablelist>
+ <varlistentry>
+ <term>Byte1('X')</term>
+ <listitem>
+ <para>
+ Identifies the message as a sequence message.
+ </para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term>Int32 (TransactionId)</term>
+ <listitem>
+ <para>
+ Xid of the transaction (only present for streamed transactions).
+ This field is available since protocol version 2.
+ </para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term>Int8 (0)</term>
+ <listitem>
+ <para>
+ Flags; currently unused.
+ </para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term>Int64 (XLogRecPtr)</term>
+ <listitem>
+ <para>
+ The LSN of the sequence increment.
+ </para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term>String</term>
+ <listitem>
+ <para>
+ Namespace (empty string for <literal>pg_catalog</literal>).
+ </para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term>String</term>
+ <listitem>
+ <para>
+ Sequence name.
+ </para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term>Int8</term>
+ <listitem>
+ <para>
+ 1 if the sequence update is transactional, 0 otherwise.
+ </para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term>Int64</term>
+ <listitem>
+ <para>
+ <structfield>value</structfield> value of the sequence.
+ </para>
+ </listitem>
+ </varlistentry>
+ </variablelist>
+ </listitem>
+ </varlistentry>
+
<varlistentry id="protocol-logicalrep-message-formats-Type">
<term>Type</term>
<listitem>
diff --git a/doc/src/sgml/ref/alter_publication.sgml b/doc/src/sgml/ref/alter_publication.sgml
index 44ae7e0e871..8caf2135fc7 100644
--- a/doc/src/sgml/ref/alter_publication.sgml
+++ b/doc/src/sgml/ref/alter_publication.sgml
@@ -31,7 +31,9 @@ ALTER PUBLICATION <replaceable class="parameter">name</replaceable> RENAME TO <r
<phrase>where <replaceable class="parameter">publication_object</replaceable> is one of:</phrase>
TABLE [ ONLY ] <replaceable class="parameter">table_name</replaceable> [ * ] [ ( <replaceable class="parameter">column_name</replaceable> [, ... ] ) ] [ WHERE ( <replaceable class="parameter">expression</replaceable> ) ] [, ... ]
+ SEQUENCE <replaceable class="parameter">sequence_name</replaceable> [, ... ]
TABLES IN SCHEMA { <replaceable class="parameter">schema_name</replaceable> | CURRENT_SCHEMA } [, ... ]
+ SEQUENCES IN SCHEMA { <replaceable class="parameter">schema_name</replaceable> | CURRENT_SCHEMA } [, ... ]
</synopsis>
</refsynopsisdiv>
@@ -44,13 +46,13 @@ ALTER PUBLICATION <replaceable class="parameter">name</replaceable> RENAME TO <r
</para>
<para>
- The first three variants change which tables/schemas are part of the
- publication. The <literal>SET</literal> clause will replace the list of
- tables/schemas in the publication with the specified list; the existing
- tables/schemas that were present in the publication will be removed. The
- <literal>ADD</literal> and <literal>DROP</literal> clauses will add and
- remove one or more tables/schemas from the publication. Note that adding
- tables/schemas to a publication that is already subscribed to will require an
+ The first three variants change which objects (tables, sequences or schemas)
+ are part of the publication. The <literal>SET</literal> clause will replace
+ the list of objects in the publication with the specified list; the existing
+ objects that were present in the publication will be removed.
+ The <literal>ADD</literal> and <literal>DROP</literal> clauses will add and
+ remove one or more objects from the publication. Note that adding objects
+ to a publication that is already subscribed to will require an
<link linkend="sql-altersubscription-params-refresh-publication">
<literal>ALTER SUBSCRIPTION ... REFRESH PUBLICATION</literal></link> action on the
subscribing side in order to become effective. Note also that
@@ -138,6 +140,15 @@ ALTER PUBLICATION <replaceable class="parameter">name</replaceable> RENAME TO <r
</listitem>
</varlistentry>
+ <varlistentry>
+ <term><replaceable class="parameter">sequence_name</replaceable></term>
+ <listitem>
+ <para>
+ Name of an existing sequence.
+ </para>
+ </listitem>
+ </varlistentry>
+
<varlistentry>
<term><replaceable class="parameter">schema_name</replaceable></term>
<listitem>
diff --git a/doc/src/sgml/ref/alter_subscription.sgml b/doc/src/sgml/ref/alter_subscription.sgml
index 6d36ff0dc90..479ec495896 100644
--- a/doc/src/sgml/ref/alter_subscription.sgml
+++ b/doc/src/sgml/ref/alter_subscription.sgml
@@ -152,8 +152,8 @@ ALTER SUBSCRIPTION <replaceable class="parameter">name</replaceable> RENAME TO <
<listitem>
<para>
Fetch missing table information from publisher. This will start
- replication of tables that were added to the subscribed-to publications
- since <link linkend="sql-createsubscription">
+ replication of tables and sequences that were added to the subscribed-to
+ publications since <link linkend="sql-createsubscription">
<command>CREATE SUBSCRIPTION</command></link> or
the last invocation of <command>REFRESH PUBLICATION</command>.
</para>
@@ -172,8 +172,8 @@ ALTER SUBSCRIPTION <replaceable class="parameter">name</replaceable> RENAME TO <
The default is <literal>true</literal>.
</para>
<para>
- Previously subscribed tables are not copied, even if a table's row
- filter <literal>WHERE</literal> clause has since been modified.
+ Previously subscribed tables and sequences are not copied, even if a
+ table's row filter <literal>WHERE</literal> clause has since been modified.
</para>
<para>
See <xref linkend="sql-createsubscription-notes"/> for details of
diff --git a/doc/src/sgml/ref/create_publication.sgml b/doc/src/sgml/ref/create_publication.sgml
index fd9c5deac95..ba578eb9aa4 100644
--- a/doc/src/sgml/ref/create_publication.sgml
+++ b/doc/src/sgml/ref/create_publication.sgml
@@ -22,14 +22,21 @@ PostgreSQL documentation
<refsynopsisdiv>
<synopsis>
CREATE PUBLICATION <replaceable class="parameter">name</replaceable>
- [ FOR ALL TABLES
+ [ FOR ALL <replaceable class="parameter">object_type</replaceable> [, ...]
| FOR <replaceable class="parameter">publication_object</replaceable> [, ... ] ]
[ WITH ( <replaceable class="parameter">publication_parameter</replaceable> [= <replaceable class="parameter">value</replaceable>] [, ... ] ) ]
+<phrase>where <replaceable class="parameter">object type</replaceable> is one of:</phrase>
+
+ TABLES
+ SEQUENCES
+
<phrase>where <replaceable class="parameter">publication_object</replaceable> is one of:</phrase>
TABLE [ ONLY ] <replaceable class="parameter">table_name</replaceable> [ * ] [ ( <replaceable class="parameter">column_name</replaceable> [, ... ] ) ] [ WHERE ( <replaceable class="parameter">expression</replaceable> ) ] [, ... ]
+ SEQUENCE <replaceable class="parameter">sequence_name</replaceable> [ * ] [, ... ]
TABLES IN SCHEMA { <replaceable class="parameter">schema_name</replaceable> | CURRENT_SCHEMA } [, ... ]
+ SEQUENCES IN SCHEMA { <replaceable class="parameter">schema_name</replaceable> | CURRENT_SCHEMA } [, ... ]
</synopsis>
</refsynopsisdiv>
@@ -117,22 +124,39 @@ CREATE PUBLICATION <replaceable class="parameter">name</replaceable>
</listitem>
</varlistentry>
+ <varlistentry id="sql-createpublication-for-sequence">
+ <term><literal>FOR SEQUENCE</literal></term>
+ <listitem>
+ <para>
+ Specifies a list of sequences to add to the publication.
+ </para>
+
+ <para>
+ Specifying a sequence that is part of a schema specified by <literal>FOR
+ ALL SEQUENCES IN SCHEMA</literal> is not supported.
+ </para>
+ </listitem>
+ </varlistentry>
+
<varlistentry id="sql-createpublication-params-for-all-tables">
<term><literal>FOR ALL TABLES</literal></term>
+ <term><literal>FOR ALL SEQUENCES</literal></term>
<listitem>
<para>
- Marks the publication as one that replicates changes for all tables in
- the database, including tables created in the future.
+ Marks the publication as one that replicates changes for all tables or
+ sequences in the database, including tables created in the future.
</para>
</listitem>
</varlistentry>
<varlistentry id="sql-createpublication-params-for-tables-in-schema">
<term><literal>FOR TABLES IN SCHEMA</literal></term>
+ <term><literal>FOR SEQUENCES IN SCHEMA</literal></term>
<listitem>
<para>
- Marks the publication as one that replicates changes for all tables in
- the specified list of schemas, including tables created in the future.
+ Marks the publication as one that replicates changes for all tables or
+ sequences in the specified list of schemas, including tables or sequences
+ created in the future.
</para>
<para>
@@ -240,10 +264,12 @@ CREATE PUBLICATION <replaceable class="parameter">name</replaceable>
<title>Notes</title>
<para>
- If <literal>FOR TABLE</literal>, <literal>FOR ALL TABLES</literal> or
- <literal>FOR TABLES IN SCHEMA</literal> are not specified, then the
- publication starts out with an empty set of tables. That is useful if
- tables or schemas are to be added later.
+ If <literal>FOR TABLE</literal>, <literal>FOR SEQUENCE</literal>,
+ <literal>FOR ALL TABLES</literal>, <literal>FOR ALL SEQUENCES</literal>,
+ <literal>FOR TABLES IN SCHEMA</literal> or <literal>FOR SEQUENCES IN SCHEMA</literal>
+ are not specified, then the publication starts out with an empty set
+ of tables/sequences. That is useful if tables, sequences or schemas
+ are to be added later.
</para>
<para>
@@ -258,9 +284,10 @@ CREATE PUBLICATION <replaceable class="parameter">name</replaceable>
</para>
<para>
- To add a table to a publication, the invoking user must have ownership
- rights on the table. The <command>FOR ALL TABLES</command> and
- <command>FOR TABLES IN SCHEMA</command> clauses require the invoking
+ To add a table or a sequence to a publication, the invoking user must
+ have ownership rights on the object. The <command>FOR ALL TABLES</command>,
+ <command>FOR ALL SEQUENCES</command>, <command>FOR TABLES IN SCHEMA</command>
+ and <command>FOR SEQUENCES IN SCHEMA</command> clauses require the invoking
user to be a superuser.
</para>
diff --git a/doc/src/sgml/system-views.sgml b/doc/src/sgml/system-views.sgml
index 0ef17456318..925105cd588 100644
--- a/doc/src/sgml/system-views.sgml
+++ b/doc/src/sgml/system-views.sgml
@@ -126,6 +126,11 @@
<entry>prepared transactions</entry>
</row>
+ <row>
+ <entry><link linkend="view-pg-publication-sequences"><structname>pg_publication_sequences</structname></link></entry>
+ <entry>publications and information of their associated sequences</entry>
+ </row>
+
<row>
<entry><link linkend="view-pg-publication-tables"><structname>pg_publication_tables</structname></link></entry>
<entry>publications and information of their associated tables</entry>
@@ -2138,6 +2143,72 @@ SELECT * FROM pg_locks pl LEFT JOIN pg_prepared_xacts ppx
</sect1>
+ <sect1 id="view-pg-publication-sequences">
+ <title><structname>pg_publication_sequences</structname></title>
+
+ <indexterm zone="view-pg-publication-sequences">
+ <primary>pg_publication_sequences</primary>
+ </indexterm>
+
+ <para>
+ The view <structname>pg_publication_sequences</structname> provides
+ information about the mapping between publications and information of
+ sequences they contain. Unlike the underlying catalog
+ <link linkend="catalog-pg-publication-rel"><structname>pg_publication_rel</structname></link>,
+ this view expands publications defined as <literal>FOR ALL SEQUENCES</literal>
+ and <literal>FOR SEQUENCES IN SCHEMA</literal>, so for such publications
+ there will be a row for each eligible sequence.
+ </para>
+
+ <table>
+ <title><structname>pg_publication_sequences</structname> Columns</title>
+ <tgroup cols="1">
+ <thead>
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ Column Type
+ </para>
+ <para>
+ Description
+ </para></entry>
+ </row>
+ </thead>
+
+ <tbody>
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ <structfield>pubname</structfield> <type>name</type>
+ (references <link linkend="catalog-pg-publication"><structname>pg_publication</structname></link>.<structfield>pubname</structfield>)
+ </para>
+ <para>
+ Name of publication
+ </para></entry>
+ </row>
+
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ <structfield>schemaname</structfield> <type>name</type>
+ (references <link linkend="catalog-pg-namespace"><structname>pg_namespace</structname></link>.<structfield>nspname</structfield>)
+ </para>
+ <para>
+ Name of schema containing sequence
+ </para></entry>
+ </row>
+
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ <structfield>sequencename</structfield> <type>name</type>
+ (references <link linkend="catalog-pg-class"><structname>pg_class</structname></link>.<structfield>relname</structfield>)
+ </para>
+ <para>
+ Name of sequence
+ </para></entry>
+ </row>
+ </tbody>
+ </tgroup>
+ </table>
+ </sect1>
+
<sect1 id="view-pg-publication-tables">
<title><structname>pg_publication_tables</structname></title>
diff --git a/src/backend/catalog/objectaddress.c b/src/backend/catalog/objectaddress.c
index bb4efcad202..c155cd57a70 100644
--- a/src/backend/catalog/objectaddress.c
+++ b/src/backend/catalog/objectaddress.c
@@ -1964,12 +1964,14 @@ get_object_address_publication_schema(List *object, bool missing_ok)
char *pubname;
char *schemaname;
Oid schemaid;
+ char *objtype;
ObjectAddressSet(address, PublicationNamespaceRelationId, InvalidOid);
/* Fetch schema name and publication name from input list */
schemaname = strVal(linitial(object));
pubname = strVal(lsecond(object));
+ objtype = strVal(lthird(object));
schemaid = get_namespace_oid(schemaname, missing_ok);
if (!OidIsValid(schemaid))
@@ -1982,10 +1984,12 @@ get_object_address_publication_schema(List *object, bool missing_ok)
/* Find the publication schema mapping in syscache */
address.objectId =
- GetSysCacheOid2(PUBLICATIONNAMESPACEMAP,
+ GetSysCacheOid3(PUBLICATIONNAMESPACEMAP,
Anum_pg_publication_namespace_oid,
ObjectIdGetDatum(schemaid),
- ObjectIdGetDatum(pub->oid));
+ ObjectIdGetDatum(pub->oid),
+ CharGetDatum(objtype[0]));
+
if (!OidIsValid(address.objectId) && !missing_ok)
ereport(ERROR,
(errcode(ERRCODE_UNDEFINED_OBJECT),
@@ -2260,7 +2264,6 @@ pg_get_object_address(PG_FUNCTION_ARGS)
*/
switch (type)
{
- case OBJECT_PUBLICATION_NAMESPACE:
case OBJECT_USER_MAPPING:
if (list_length(name) != 1)
ereport(ERROR,
@@ -2293,6 +2296,7 @@ pg_get_object_address(PG_FUNCTION_ARGS)
errmsg("name list length must be at least %d", 3)));
/* fall through to check args length */
/* FALLTHROUGH */
+ case OBJECT_PUBLICATION_NAMESPACE:
case OBJECT_OPERATOR:
if (list_length(args) != 2)
ereport(ERROR,
@@ -2365,6 +2369,8 @@ pg_get_object_address(PG_FUNCTION_ARGS)
objnode = (Node *) list_make2(name, linitial(args));
break;
case OBJECT_PUBLICATION_NAMESPACE:
+ objnode = (Node *) list_make3(linitial(name), linitial(args), lsecond(args));
+ break;
case OBJECT_USER_MAPPING:
objnode = (Node *) list_make2(linitial(name), linitial(args));
break;
@@ -2868,11 +2874,12 @@ get_catalog_object_by_oid(Relation catalog, AttrNumber oidcol, Oid objectId)
*
* Get publication name and schema name from the object address into pubname and
* nspname. Both pubname and nspname are palloc'd strings which will be freed by
- * the caller.
+ * the caller. The last parameter specifies which object type is included from
+ * the schema.
*/
static bool
getPublicationSchemaInfo(const ObjectAddress *object, bool missing_ok,
- char **pubname, char **nspname)
+ char **pubname, char **nspname, char **objtype)
{
HeapTuple tup;
Form_pg_publication_namespace pnform;
@@ -2908,6 +2915,14 @@ getPublicationSchemaInfo(const ObjectAddress *object, bool missing_ok,
return false;
}
+ /*
+ * The type is always a single character, but we need to pass it as a
+ * string, so allocate two charaters and set the first one. The second one
+ * is \0.
+ */
+ *objtype = palloc0(2);
+ *objtype[0] = pnform->pntype;
+
ReleaseSysCache(tup);
return true;
}
@@ -3978,15 +3993,17 @@ getObjectDescription(const ObjectAddress *object, bool missing_ok)
{
char *pubname;
char *nspname;
+ char *objtype;
if (!getPublicationSchemaInfo(object, missing_ok,
- &pubname, &nspname))
+ &pubname, &nspname, &objtype))
break;
- appendStringInfo(&buffer, _("publication of schema %s in publication %s"),
- nspname, pubname);
+ appendStringInfo(&buffer, _("publication of schema %s in publication %s type %s"),
+ nspname, pubname, objtype);
pfree(pubname);
pfree(nspname);
+ pfree(objtype);
break;
}
@@ -5832,18 +5849,24 @@ getObjectIdentityParts(const ObjectAddress *object,
{
char *pubname;
char *nspname;
+ char *objtype;
if (!getPublicationSchemaInfo(object, missing_ok, &pubname,
- &nspname))
+ &nspname, &objtype))
break;
- appendStringInfo(&buffer, "%s in publication %s",
- nspname, pubname);
+ appendStringInfo(&buffer, "%s in publication %s type %s",
+ nspname, pubname, objtype);
if (objargs)
*objargs = list_make1(pubname);
else
pfree(pubname);
+ if (objargs)
+ *objargs = lappend(*objargs, objtype);
+ else
+ pfree(objtype);
+
if (objname)
*objname = list_make1(nspname);
else
diff --git a/src/backend/catalog/pg_publication.c b/src/backend/catalog/pg_publication.c
index c488b6370b6..bbcbce61718 100644
--- a/src/backend/catalog/pg_publication.c
+++ b/src/backend/catalog/pg_publication.c
@@ -63,9 +63,10 @@ static void publication_translate_columns(Relation targetrel, List *columns,
static void
check_publication_add_relation(Relation targetrel)
{
- /* Must be a regular or partitioned table */
+ /* Must be a regular or partitioned table, or a sequence */
if (RelationGetForm(targetrel)->relkind != RELKIND_RELATION &&
- RelationGetForm(targetrel)->relkind != RELKIND_PARTITIONED_TABLE)
+ RelationGetForm(targetrel)->relkind != RELKIND_PARTITIONED_TABLE &&
+ RelationGetForm(targetrel)->relkind != RELKIND_SEQUENCE)
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("cannot add relation \"%s\" to publication",
@@ -142,7 +143,8 @@ static bool
is_publishable_class(Oid relid, Form_pg_class reltuple)
{
return (reltuple->relkind == RELKIND_RELATION ||
- reltuple->relkind == RELKIND_PARTITIONED_TABLE) &&
+ reltuple->relkind == RELKIND_PARTITIONED_TABLE ||
+ reltuple->relkind == RELKIND_SEQUENCE) &&
!IsCatalogRelationOid(relid) &&
reltuple->relpersistence == RELPERSISTENCE_PERMANENT &&
relid >= FirstNormalObjectId;
@@ -233,6 +235,52 @@ filter_partitions(List *table_infos)
}
}
+/*
+ * Check the character is a valid object type for schema publication.
+ *
+ * This recognizes either 't' for tables or 's' for sequences. Places that
+ * need to handle 'u' for unsupported relkinds need to do that explicitlyl
+ */
+static void
+AssertObjectTypeValid(char objectType)
+{
+#ifdef USE_ASSERT_CHECKING
+ Assert(objectType == PUB_OBJTYPE_SEQUENCE || objectType == PUB_OBJTYPE_TABLE);
+#endif
+}
+
+/*
+ * Determine object type matching a given a relkind value.
+ */
+char
+pub_get_object_type_for_relkind(char relkind)
+{
+ /* sequence maps directly to sequence relkind */
+ if (relkind == RELKIND_SEQUENCE)
+ return PUB_OBJTYPE_SEQUENCE;
+
+ /* for table, we match either regular or partitioned table */
+ if (relkind == RELKIND_RELATION ||
+ relkind == RELKIND_PARTITIONED_TABLE)
+ return PUB_OBJTYPE_TABLE;
+
+ return PUB_OBJTYPE_UNSUPPORTED;
+}
+
+/*
+ * Determine if publication object type matches the relkind.
+ *
+ * Returns true if the relation matches object type replicated by this schema,
+ * false otherwise.
+ */
+static bool
+pub_object_type_matches_relkind(char objectType, char relkind)
+{
+ AssertObjectTypeValid(objectType);
+
+ return (pub_get_object_type_for_relkind(relkind) == objectType);
+}
+
/*
* Returns true if any schema is associated with the publication, false if no
* schema is associated with the publication.
@@ -253,7 +301,7 @@ is_schema_publication(Oid pubid)
ObjectIdGetDatum(pubid));
scan = systable_beginscan(pubschsrel,
- PublicationNamespacePnnspidPnpubidIndexId,
+ PublicationNamespacePnnspidPnpubidPntypeIndexId,
true, NULL, 1, &scankey);
tup = systable_getnext(scan);
result = HeapTupleIsValid(tup);
@@ -339,7 +387,9 @@ GetTopMostAncestorInPublication(Oid puboid, List *ancestors, int *ancestor_level
}
else
{
- aschemaPubids = GetSchemaPublications(get_rel_namespace(ancestor));
+ /* we only search for ancestors of tables, so PUB_OBJTYPE_TABLE */
+ aschemaPubids = GetSchemaPublications(get_rel_namespace(ancestor),
+ PUB_OBJTYPE_TABLE);
if (list_member_oid(aschemaPubids, puboid))
{
topmost_relid = ancestor;
@@ -608,7 +658,7 @@ pub_collist_to_bitmapset(Bitmapset *columns, Datum pubcols, MemoryContext mcxt)
* Insert new publication / schema mapping.
*/
ObjectAddress
-publication_add_schema(Oid pubid, Oid schemaid, bool if_not_exists)
+publication_add_schema(Oid pubid, Oid schemaid, char objectType, bool if_not_exists)
{
Relation rel;
HeapTuple tup;
@@ -620,6 +670,8 @@ publication_add_schema(Oid pubid, Oid schemaid, bool if_not_exists)
ObjectAddress myself,
referenced;
+ AssertObjectTypeValid(objectType);
+
rel = table_open(PublicationNamespaceRelationId, RowExclusiveLock);
/*
@@ -627,9 +679,10 @@ publication_add_schema(Oid pubid, Oid schemaid, bool if_not_exists)
* duplicates, it's here just to provide nicer error message in common
* case. The real protection is the unique key on the catalog.
*/
- if (SearchSysCacheExists2(PUBLICATIONNAMESPACEMAP,
+ if (SearchSysCacheExists3(PUBLICATIONNAMESPACEMAP,
ObjectIdGetDatum(schemaid),
- ObjectIdGetDatum(pubid)))
+ ObjectIdGetDatum(pubid),
+ CharGetDatum(objectType)))
{
table_close(rel, RowExclusiveLock);
@@ -655,6 +708,8 @@ publication_add_schema(Oid pubid, Oid schemaid, bool if_not_exists)
ObjectIdGetDatum(pubid);
values[Anum_pg_publication_namespace_pnnspid - 1] =
ObjectIdGetDatum(schemaid);
+ values[Anum_pg_publication_namespace_pntype - 1] =
+ CharGetDatum(objectType);
tup = heap_form_tuple(RelationGetDescr(rel), values, nulls);
@@ -680,7 +735,7 @@ publication_add_schema(Oid pubid, Oid schemaid, bool if_not_exists)
* publication_add_relation for why we need to consider all the
* partitions.
*/
- schemaRels = GetSchemaPublicationRelations(schemaid,
+ schemaRels = GetSchemaPublicationRelations(schemaid, objectType,
PUBLICATION_PART_ALL);
InvalidatePublicationRels(schemaRels);
@@ -714,11 +769,14 @@ GetRelationPublications(Oid relid)
/*
* Gets list of relation oids for a publication.
*
- * This should only be used FOR TABLE publications, the FOR ALL TABLES
- * should use GetAllTablesPublicationRelations().
+ * This should only be used FOR TABLE / FOR SEQUENCE publications, the FOR
+ * ALL TABLES / SEQUENCES should use GetAllTablesPublicationRelations()
+ * and GetAllSequencesPublicationRelations().
+ *
+ * XXX pub_partopt only matters for tables, not sequences.
*/
List *
-GetPublicationRelations(Oid pubid, PublicationPartOpt pub_partopt)
+GetPublicationRelations(Oid pubid, char objectType, PublicationPartOpt pub_partopt)
{
List *result;
Relation pubrelsrel;
@@ -726,6 +784,8 @@ GetPublicationRelations(Oid pubid, PublicationPartOpt pub_partopt)
SysScanDesc scan;
HeapTuple tup;
+ AssertObjectTypeValid(objectType);
+
/* Find all publications associated with the relation. */
pubrelsrel = table_open(PublicationRelRelationId, AccessShareLock);
@@ -740,11 +800,29 @@ GetPublicationRelations(Oid pubid, PublicationPartOpt pub_partopt)
result = NIL;
while (HeapTupleIsValid(tup = systable_getnext(scan)))
{
+ char relkind;
Form_pg_publication_rel pubrel;
pubrel = (Form_pg_publication_rel) GETSTRUCT(tup);
- result = GetPubPartitionOptionRelations(result, pub_partopt,
- pubrel->prrelid);
+ relkind = get_rel_relkind(pubrel->prrelid);
+
+ /*
+ * If the relkind does not match the requested object type, ignore the
+ * relation. For example we might be interested only in sequences, so
+ * we ignore tables.
+ */
+ if (!pub_object_type_matches_relkind(objectType, relkind))
+ continue;
+
+ /*
+ * We don't have partitioned sequences, so just add them to the list.
+ * Otherwise consider adding all child relations, if requested.
+ */
+ if (relkind == RELKIND_SEQUENCE)
+ result = lappend_oid(result, pubrel->prrelid);
+ else
+ result = GetPubPartitionOptionRelations(result, pub_partopt,
+ pubrel->prrelid);
}
systable_endscan(scan);
@@ -794,6 +872,43 @@ GetAllTablesPublications(void)
return result;
}
+/*
+ * Gets list of publication oids for publications marked as FOR ALL SEQUENCES.
+ */
+List *
+GetAllSequencesPublications(void)
+{
+ List *result;
+ Relation rel;
+ ScanKeyData scankey;
+ SysScanDesc scan;
+ HeapTuple tup;
+
+ /* Find all publications that are marked as for all sequences. */
+ rel = table_open(PublicationRelationId, AccessShareLock);
+
+ ScanKeyInit(&scankey,
+ Anum_pg_publication_puballsequences,
+ BTEqualStrategyNumber, F_BOOLEQ,
+ BoolGetDatum(true));
+
+ scan = systable_beginscan(rel, InvalidOid, false,
+ NULL, 1, &scankey);
+
+ result = NIL;
+ while (HeapTupleIsValid(tup = systable_getnext(scan)))
+ {
+ Oid oid = ((Form_pg_publication) GETSTRUCT(tup))->oid;
+
+ result = lappend_oid(result, oid);
+ }
+
+ systable_endscan(scan);
+ table_close(rel, AccessShareLock);
+
+ return result;
+}
+
/*
* Gets list of all relation published by FOR ALL TABLES publication(s).
*
@@ -860,28 +975,38 @@ GetAllTablesPublicationRelations(bool pubviaroot)
/*
* Gets the list of schema oids for a publication.
*
- * This should only be used FOR TABLES IN SCHEMA publications.
+ * This should only be used FOR TABLES IN SCHEMA and FOR SEQUENCES IN SCHEMA
+ * publications.
+ *
+ * 'objectType' determines whether to get FOR TABLE or FOR SEQUENCES schemas
*/
List *
-GetPublicationSchemas(Oid pubid)
+GetPublicationSchemas(Oid pubid, char objectType)
{
List *result = NIL;
Relation pubschsrel;
- ScanKeyData scankey;
+ ScanKeyData scankey[2];
SysScanDesc scan;
HeapTuple tup;
+ AssertObjectTypeValid(objectType);
+
/* Find all schemas associated with the publication */
pubschsrel = table_open(PublicationNamespaceRelationId, AccessShareLock);
- ScanKeyInit(&scankey,
+ ScanKeyInit(&scankey[0],
Anum_pg_publication_namespace_pnpubid,
BTEqualStrategyNumber, F_OIDEQ,
ObjectIdGetDatum(pubid));
+ ScanKeyInit(&scankey[1],
+ Anum_pg_publication_namespace_pntype,
+ BTEqualStrategyNumber, F_CHAREQ,
+ CharGetDatum(objectType));
+
scan = systable_beginscan(pubschsrel,
- PublicationNamespacePnnspidPnpubidIndexId,
- true, NULL, 1, &scankey);
+ PublicationNamespacePnnspidPnpubidPntypeIndexId,
+ true, NULL, 2, scankey);
while (HeapTupleIsValid(tup = systable_getnext(scan)))
{
Form_pg_publication_namespace pubsch;
@@ -899,14 +1024,26 @@ GetPublicationSchemas(Oid pubid)
/*
* Gets the list of publication oids associated with a specified schema.
+ *
+ * objectType specifies whether we're looking for schemas including tables or
+ * sequences.
+ *
+ * Note: relcache calls this for all object types, not just tables and sequences.
+ * Which is why we handle the PUB_OBJTYPE_UNSUPPORTED object type too.
*/
List *
-GetSchemaPublications(Oid schemaid)
+GetSchemaPublications(Oid schemaid, char objectType)
{
List *result = NIL;
CatCList *pubschlist;
int i;
+ /* unsupported object type */
+ if (objectType == PUB_OBJTYPE_UNSUPPORTED)
+ return result;
+
+ AssertObjectTypeValid(objectType);
+
/* Find all publications associated with the schema */
pubschlist = SearchSysCacheList1(PUBLICATIONNAMESPACEMAP,
ObjectIdGetDatum(schemaid));
@@ -914,6 +1051,11 @@ GetSchemaPublications(Oid schemaid)
{
HeapTuple tup = &pubschlist->members[i]->tuple;
Oid pubid = ((Form_pg_publication_namespace) GETSTRUCT(tup))->pnpubid;
+ char pntype = ((Form_pg_publication_namespace) GETSTRUCT(tup))->pntype;
+
+ /* Skip schemas publishing a different object type. */
+ if (pntype != objectType)
+ continue;
result = lappend_oid(result, pubid);
}
@@ -925,9 +1067,13 @@ GetSchemaPublications(Oid schemaid)
/*
* Get the list of publishable relation oids for a specified schema.
+ *
+ * objectType specifies whether this is FOR ALL TABLES IN SCHEMA or FOR ALL
+ * SEQUENCES IN SCHEMA
*/
List *
-GetSchemaPublicationRelations(Oid schemaid, PublicationPartOpt pub_partopt)
+GetSchemaPublicationRelations(Oid schemaid, char objectType,
+ PublicationPartOpt pub_partopt)
{
Relation classRel;
ScanKeyData key[1];
@@ -936,6 +1082,7 @@ GetSchemaPublicationRelations(Oid schemaid, PublicationPartOpt pub_partopt)
List *result = NIL;
Assert(OidIsValid(schemaid));
+ AssertObjectTypeValid(objectType);
classRel = table_open(RelationRelationId, AccessShareLock);
@@ -956,9 +1103,17 @@ GetSchemaPublicationRelations(Oid schemaid, PublicationPartOpt pub_partopt)
continue;
relkind = get_rel_relkind(relid);
- if (relkind == RELKIND_RELATION)
- result = lappend_oid(result, relid);
- else if (relkind == RELKIND_PARTITIONED_TABLE)
+
+ /* Skip if the relkind does not match FOR ALL TABLES / SEQUENCES. */
+ if (!pub_object_type_matches_relkind(objectType, relkind))
+ continue;
+
+ /*
+ * If the object is a partitioned table, lookup all the child
+ * relations (if requested). Otherwise just add the object to the
+ * list.
+ */
+ if (relkind == RELKIND_PARTITIONED_TABLE)
{
List *partitionrels = NIL;
@@ -971,7 +1126,11 @@ GetSchemaPublicationRelations(Oid schemaid, PublicationPartOpt pub_partopt)
pub_partopt,
relForm->oid);
result = list_concat_unique_oid(result, partitionrels);
+ continue;
}
+
+ /* non-partitioned tables and sequences */
+ result = lappend_oid(result, relid);
}
table_endscan(scan);
@@ -980,28 +1139,68 @@ GetSchemaPublicationRelations(Oid schemaid, PublicationPartOpt pub_partopt)
}
/*
- * Gets the list of all relations published by FOR TABLES IN SCHEMA
- * publication.
+ * Gets the list of all relations published by FOR TABLES IN SCHEMA or
+ * FOR SEQUENCES IN SCHEMA publication.
*/
List *
-GetAllSchemaPublicationRelations(Oid pubid, PublicationPartOpt pub_partopt)
+GetAllSchemaPublicationRelations(Oid pubid, char objectType,
+ PublicationPartOpt pub_partopt)
{
List *result = NIL;
- List *pubschemalist = GetPublicationSchemas(pubid);
+ List *pubschemalist = GetPublicationSchemas(pubid, objectType);
ListCell *cell;
+ AssertObjectTypeValid(objectType);
+
foreach(cell, pubschemalist)
{
Oid schemaid = lfirst_oid(cell);
List *schemaRels = NIL;
- schemaRels = GetSchemaPublicationRelations(schemaid, pub_partopt);
+ schemaRels = GetSchemaPublicationRelations(schemaid, objectType,
+ pub_partopt);
result = list_concat(result, schemaRels);
}
return result;
}
+/*
+ * Gets list of all relation published by FOR ALL SEQUENCES publication(s).
+ */
+List *
+GetAllSequencesPublicationRelations(void)
+{
+ Relation classRel;
+ ScanKeyData key[1];
+ TableScanDesc scan;
+ HeapTuple tuple;
+ List *result = NIL;
+
+ classRel = table_open(RelationRelationId, AccessShareLock);
+
+ ScanKeyInit(&key[0],
+ Anum_pg_class_relkind,
+ BTEqualStrategyNumber, F_CHAREQ,
+ CharGetDatum(RELKIND_SEQUENCE));
+
+ scan = table_beginscan_catalog(classRel, 1, key);
+
+ while ((tuple = heap_getnext(scan, ForwardScanDirection)) != NULL)
+ {
+ Form_pg_class relForm = (Form_pg_class) GETSTRUCT(tuple);
+ Oid relid = relForm->oid;
+
+ if (is_publishable_class(relid, relForm))
+ result = lappend_oid(result, relid);
+ }
+
+ table_endscan(scan);
+
+ table_close(classRel, AccessShareLock);
+ return result;
+}
+
/*
* Get publication using oid
*
@@ -1024,10 +1223,12 @@ GetPublication(Oid pubid)
pub->oid = pubid;
pub->name = pstrdup(NameStr(pubform->pubname));
pub->alltables = pubform->puballtables;
+ pub->allsequences = pubform->puballsequences;
pub->pubactions.pubinsert = pubform->pubinsert;
pub->pubactions.pubupdate = pubform->pubupdate;
pub->pubactions.pubdelete = pubform->pubdelete;
pub->pubactions.pubtruncate = pubform->pubtruncate;
+ pub->pubactions.pubsequence = pubform->pubsequence;
pub->pubviaroot = pubform->pubviaroot;
ReleaseSysCache(tup);
@@ -1108,10 +1309,12 @@ pg_get_publication_tables(PG_FUNCTION_ARGS)
*schemarelids;
relids = GetPublicationRelations(pub_elem->oid,
+ PUB_OBJTYPE_TABLE,
pub_elem->pubviaroot ?
PUBLICATION_PART_ROOT :
PUBLICATION_PART_LEAF);
schemarelids = GetAllSchemaPublicationRelations(pub_elem->oid,
+ PUB_OBJTYPE_TABLE,
pub_elem->pubviaroot ?
PUBLICATION_PART_ROOT :
PUBLICATION_PART_LEAF);
@@ -1197,9 +1400,10 @@ pg_get_publication_tables(PG_FUNCTION_ARGS)
* FOR TABLES IN SCHEMA publications.
*/
if (!pub->alltables &&
- !SearchSysCacheExists2(PUBLICATIONNAMESPACEMAP,
+ !SearchSysCacheExists3(PUBLICATIONNAMESPACEMAP,
ObjectIdGetDatum(schemaid),
- ObjectIdGetDatum(pub->oid)))
+ ObjectIdGetDatum(pub->oid),
+ PUB_OBJTYPE_TABLE))
pubtuple = SearchSysCacheCopy2(PUBLICATIONRELMAP,
ObjectIdGetDatum(relid),
ObjectIdGetDatum(pub->oid));
@@ -1259,3 +1463,71 @@ pg_get_publication_tables(PG_FUNCTION_ARGS)
SRF_RETURN_DONE(funcctx);
}
+
+/*
+ * Returns Oids of sequences in a publication.
+ */
+Datum
+pg_get_publication_sequences(PG_FUNCTION_ARGS)
+{
+ FuncCallContext *funcctx;
+ char *pubname = text_to_cstring(PG_GETARG_TEXT_PP(0));
+ Publication *publication;
+ List *sequences;
+
+ /* stuff done only on the first call of the function */
+ if (SRF_IS_FIRSTCALL())
+ {
+ MemoryContext oldcontext;
+
+ /* create a function context for cross-call persistence */
+ funcctx = SRF_FIRSTCALL_INIT();
+
+ /* switch to memory context appropriate for multiple function calls */
+ oldcontext = MemoryContextSwitchTo(funcctx->multi_call_memory_ctx);
+
+ publication = GetPublicationByName(pubname, false);
+
+ /*
+ * Publications support partitioned tables, although all changes are
+ * replicated using leaf partition identity and schema, so we only
+ * need those.
+ */
+ if (publication->allsequences)
+ sequences = GetAllSequencesPublicationRelations();
+ else
+ {
+ List *relids,
+ *schemarelids;
+
+ relids = GetPublicationRelations(publication->oid,
+ PUB_OBJTYPE_SEQUENCE,
+ publication->pubviaroot ?
+ PUBLICATION_PART_ROOT :
+ PUBLICATION_PART_LEAF);
+ schemarelids = GetAllSchemaPublicationRelations(publication->oid,
+ PUB_OBJTYPE_SEQUENCE,
+ publication->pubviaroot ?
+ PUBLICATION_PART_ROOT :
+ PUBLICATION_PART_LEAF);
+ sequences = list_concat_unique_oid(relids, schemarelids);
+ }
+
+ funcctx->user_fctx = (void *) sequences;
+
+ MemoryContextSwitchTo(oldcontext);
+ }
+
+ /* stuff done on every call of the function */
+ funcctx = SRF_PERCALL_SETUP();
+ sequences = (List *) funcctx->user_fctx;
+
+ if (funcctx->call_cntr < list_length(sequences))
+ {
+ Oid relid = list_nth_oid(sequences, funcctx->call_cntr);
+
+ SRF_RETURN_NEXT(funcctx, ObjectIdGetDatum(relid));
+ }
+
+ SRF_RETURN_DONE(funcctx);
+}
diff --git a/src/backend/catalog/system_views.sql b/src/backend/catalog/system_views.sql
index 11d18ed9dd6..b24229d2703 100644
--- a/src/backend/catalog/system_views.sql
+++ b/src/backend/catalog/system_views.sql
@@ -401,6 +401,16 @@ CREATE VIEW pg_publication_tables AS
pg_class C JOIN pg_namespace N ON (N.oid = C.relnamespace)
WHERE C.oid = GPT.relid;
+CREATE VIEW pg_publication_sequences AS
+ SELECT
+ P.pubname AS pubname,
+ N.nspname AS schemaname,
+ C.relname AS sequencename
+ FROM pg_publication P,
+ LATERAL pg_get_publication_sequences(P.pubname) GPS,
+ pg_class C JOIN pg_namespace N ON (N.oid = C.relnamespace)
+ WHERE C.oid = GPS.relid;
+
CREATE VIEW pg_locks AS
SELECT * FROM pg_lock_status() AS L;
diff --git a/src/backend/commands/publicationcmds.c b/src/backend/commands/publicationcmds.c
index f4ba572697a..e7b2667f64a 100644
--- a/src/backend/commands/publicationcmds.c
+++ b/src/backend/commands/publicationcmds.c
@@ -16,6 +16,7 @@
#include "access/genam.h"
#include "access/htup_details.h"
+#include "access/relation.h"
#include "access/table.h"
#include "access/xact.h"
#include "catalog/catalog.h"
@@ -68,15 +69,17 @@ typedef struct rf_context
Oid parentid; /* relid of the parent relation */
} rf_context;
-static List *OpenTableList(List *tables);
-static void CloseTableList(List *rels);
+static List *OpenRelationList(List *rels, char objectType);
+static void CloseRelationList(List *rels);
static void LockSchemaList(List *schemalist);
-static void PublicationAddTables(Oid pubid, List *rels, bool if_not_exists,
- AlterPublicationStmt *stmt);
-static void PublicationDropTables(Oid pubid, List *rels, bool missing_ok);
-static void PublicationAddSchemas(Oid pubid, List *schemas, bool if_not_exists,
- AlterPublicationStmt *stmt);
-static void PublicationDropSchemas(Oid pubid, List *schemas, bool missing_ok);
+static void PublicationAddRelations(Oid pubid, List *rels, bool if_not_exists,
+ AlterPublicationStmt *stmt);
+static void PublicationDropRelations(Oid pubid, List *rels, bool missing_ok);
+static void PublicationAddSchemas(Oid pubid, List *schemas, char objectType,
+ bool if_not_exists, AlterPublicationStmt *stmt);
+static void PublicationDropSchemas(Oid pubid, List *schemas, char objectType,
+ bool missing_ok);
+
static void
@@ -97,6 +100,7 @@ parse_publication_options(ParseState *pstate,
pubactions->pubupdate = true;
pubactions->pubdelete = true;
pubactions->pubtruncate = true;
+ pubactions->pubsequence = true;
*publish_via_partition_root = false;
/* Parse options */
@@ -121,6 +125,7 @@ parse_publication_options(ParseState *pstate,
pubactions->pubupdate = false;
pubactions->pubdelete = false;
pubactions->pubtruncate = false;
+ pubactions->pubsequence = false;
*publish_given = true;
publish = defGetString(defel);
@@ -144,6 +149,8 @@ parse_publication_options(ParseState *pstate,
pubactions->pubdelete = true;
else if (strcmp(publish_opt, "truncate") == 0)
pubactions->pubtruncate = true;
+ else if (strcmp(publish_opt, "sequence") == 0)
+ pubactions->pubsequence = true;
else
ereport(ERROR,
(errcode(ERRCODE_SYNTAX_ERROR),
@@ -171,7 +178,8 @@ parse_publication_options(ParseState *pstate,
*/
static void
ObjectsInPublicationToOids(List *pubobjspec_list, ParseState *pstate,
- List **rels, List **schemas)
+ List **tables, List **sequences,
+ List **tables_schemas, List **sequences_schemas)
{
ListCell *cell;
PublicationObjSpec *pubobj;
@@ -189,13 +197,22 @@ ObjectsInPublicationToOids(List *pubobjspec_list, ParseState *pstate,
switch (pubobj->pubobjtype)
{
case PUBLICATIONOBJ_TABLE:
- *rels = lappend(*rels, pubobj->pubtable);
+ *tables = lappend(*tables, pubobj->pubtable);
+ break;
+ case PUBLICATIONOBJ_SEQUENCE:
+ *sequences = lappend(*sequences, pubobj->pubtable);
break;
case PUBLICATIONOBJ_TABLES_IN_SCHEMA:
schemaid = get_namespace_oid(pubobj->name, false);
/* Filter out duplicates if user specifies "sch1, sch1" */
- *schemas = list_append_unique_oid(*schemas, schemaid);
+ *tables_schemas = list_append_unique_oid(*tables_schemas, schemaid);
+ break;
+ case PUBLICATIONOBJ_SEQUENCES_IN_SCHEMA:
+ schemaid = get_namespace_oid(pubobj->name, false);
+
+ /* Filter out duplicates if user specifies "sch1, sch1" */
+ *sequences_schemas = list_append_unique_oid(*sequences_schemas, schemaid);
break;
case PUBLICATIONOBJ_TABLES_IN_CUR_SCHEMA:
search_path = fetch_search_path(false);
@@ -208,7 +225,20 @@ ObjectsInPublicationToOids(List *pubobjspec_list, ParseState *pstate,
list_free(search_path);
/* Filter out duplicates if user specifies "sch1, sch1" */
- *schemas = list_append_unique_oid(*schemas, schemaid);
+ *tables_schemas = list_append_unique_oid(*tables_schemas, schemaid);
+ break;
+ case PUBLICATIONOBJ_SEQUENCES_IN_CUR_SCHEMA:
+ search_path = fetch_search_path(false);
+ if (search_path == NIL) /* nothing valid in search_path? */
+ ereport(ERROR,
+ errcode(ERRCODE_UNDEFINED_SCHEMA),
+ errmsg("no schema has been selected for CURRENT_SCHEMA"));
+
+ schemaid = linitial_oid(search_path);
+ list_free(search_path);
+
+ /* Filter out duplicates if user specifies "sch1, sch1" */
+ *sequences_schemas = list_append_unique_oid(*sequences_schemas, schemaid);
break;
default:
/* shouldn't happen */
@@ -734,6 +764,7 @@ CheckPubRelationColumnList(char *pubname, List *tables,
ObjectAddress
CreatePublication(ParseState *pstate, CreatePublicationStmt *stmt)
{
+ ListCell *lc;
Relation rel;
ObjectAddress myself;
Oid puboid;
@@ -745,8 +776,27 @@ CreatePublication(ParseState *pstate, CreatePublicationStmt *stmt)
bool publish_via_partition_root_given;
bool publish_via_partition_root;
AclResult aclresult;
- List *relations = NIL;
- List *schemaidlist = NIL;
+ List *tables = NIL;
+ List *sequences = NIL;
+ List *tables_schemaidlist = NIL;
+ List *sequences_schemaidlist = NIL;
+
+ bool for_all_tables = false;
+ bool for_all_sequences = false;
+
+ /*
+ * Translate the list of object types (represented by strings) to bool
+ * flags.
+ */
+ foreach(lc, stmt->for_all_objects)
+ {
+ char *val = strVal(lfirst(lc));
+
+ if (strcmp(val, "tables") == 0)
+ for_all_tables = true;
+ else if (strcmp(val, "sequences") == 0)
+ for_all_sequences = true;
+ }
/* must have CREATE privilege on database */
aclresult = object_aclcheck(DatabaseRelationId, MyDatabaseId, GetUserId(), ACL_CREATE);
@@ -755,11 +805,17 @@ CreatePublication(ParseState *pstate, CreatePublicationStmt *stmt)
get_database_name(MyDatabaseId));
/* FOR ALL TABLES requires superuser */
- if (stmt->for_all_tables && !superuser())
+ if (for_all_tables && !superuser())
ereport(ERROR,
(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
errmsg("must be superuser to create FOR ALL TABLES publication")));
+ /* FOR ALL SEQUENCES requires superuser */
+ if (for_all_sequences && !superuser())
+ ereport(ERROR,
+ (errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
+ errmsg("must be superuser to create FOR ALL SEQUENCES publication")));
+
rel = table_open(PublicationRelationId, RowExclusiveLock);
/* Check if name is used */
@@ -789,7 +845,9 @@ CreatePublication(ParseState *pstate, CreatePublicationStmt *stmt)
Anum_pg_publication_oid);
values[Anum_pg_publication_oid - 1] = ObjectIdGetDatum(puboid);
values[Anum_pg_publication_puballtables - 1] =
- BoolGetDatum(stmt->for_all_tables);
+ BoolGetDatum(for_all_tables);
+ values[Anum_pg_publication_puballsequences - 1] =
+ BoolGetDatum(for_all_sequences);
values[Anum_pg_publication_pubinsert - 1] =
BoolGetDatum(pubactions.pubinsert);
values[Anum_pg_publication_pubupdate - 1] =
@@ -798,6 +856,8 @@ CreatePublication(ParseState *pstate, CreatePublicationStmt *stmt)
BoolGetDatum(pubactions.pubdelete);
values[Anum_pg_publication_pubtruncate - 1] =
BoolGetDatum(pubactions.pubtruncate);
+ values[Anum_pg_publication_pubsequence - 1] =
+ BoolGetDatum(pubactions.pubsequence);
values[Anum_pg_publication_pubviaroot - 1] =
BoolGetDatum(publish_via_partition_root);
@@ -815,46 +875,88 @@ CreatePublication(ParseState *pstate, CreatePublicationStmt *stmt)
CommandCounterIncrement();
/* Associate objects with the publication. */
- if (stmt->for_all_tables)
+ if (for_all_tables || for_all_sequences)
{
/* Invalidate relcache so that publication info is rebuilt. */
CacheInvalidateRelcacheAll();
}
- else
+
+ /*
+ * If the publication might have either tables or sequences (directly or
+ * through a schema), process that.
+ */
+ if (!for_all_tables || !for_all_sequences)
{
- ObjectsInPublicationToOids(stmt->pubobjects, pstate, &relations,
- &schemaidlist);
+ ObjectsInPublicationToOids(stmt->pubobjects, pstate,
+ &tables, &sequences,
+ &tables_schemaidlist,
+ &sequences_schemaidlist);
/* FOR TABLES IN SCHEMA requires superuser */
- if (schemaidlist != NIL && !superuser())
+ if (tables_schemaidlist != NIL && !superuser())
ereport(ERROR,
errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
errmsg("must be superuser to create FOR TABLES IN SCHEMA publication"));
- if (relations != NIL)
+ /* FOR SEQUENCES IN SCHEMA requires superuser */
+ if (sequences_schemaidlist != NIL && !superuser())
+ ereport(ERROR,
+ errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
+ errmsg("must be superuser to create FOR ALL SEQUENCES IN SCHEMA publication"));
+
+ /* tables added directly */
+ if (tables != NIL)
{
List *rels;
- rels = OpenTableList(relations);
+ rels = OpenRelationList(tables, PUB_OBJTYPE_TABLE);
+
TransformPubWhereClauses(rels, pstate->p_sourcetext,
publish_via_partition_root);
CheckPubRelationColumnList(stmt->pubname, rels,
- schemaidlist != NIL,
+ tables_schemaidlist != NIL,
publish_via_partition_root);
- PublicationAddTables(puboid, rels, true, NULL);
- CloseTableList(rels);
+ PublicationAddRelations(puboid, rels, true, NULL);
+ CloseRelationList(rels);
+ }
+
+ /* sequences added directly */
+ if (sequences != NIL)
+ {
+ List *rels;
+
+ rels = OpenRelationList(sequences, PUB_OBJTYPE_SEQUENCE);
+
+ PublicationAddRelations(puboid, rels, true, NULL);
+ CloseRelationList(rels);
}
- if (schemaidlist != NIL)
+ /* tables added through a schema */
+ if (tables_schemaidlist != NIL)
{
/*
* Schema lock is held until the publication is created to prevent
* concurrent schema deletion.
*/
- LockSchemaList(schemaidlist);
- PublicationAddSchemas(puboid, schemaidlist, true, NULL);
+ LockSchemaList(tables_schemaidlist);
+ PublicationAddSchemas(puboid,
+ tables_schemaidlist, PUB_OBJTYPE_TABLE,
+ true, NULL);
+ }
+
+ /* sequences added through a schema */
+ if (sequences_schemaidlist != NIL)
+ {
+ /*
+ * Schema lock is held until the publication is created to prevent
+ * concurrent schema deletion.
+ */
+ LockSchemaList(sequences_schemaidlist);
+ PublicationAddSchemas(puboid,
+ sequences_schemaidlist, PUB_OBJTYPE_SEQUENCE,
+ true, NULL);
}
}
@@ -917,6 +1019,7 @@ AlterPublicationOptions(ParseState *pstate, AlterPublicationStmt *stmt,
AccessShareLock);
root_relids = GetPublicationRelations(pubform->oid,
+ PUB_OBJTYPE_TABLE,
PUBLICATION_PART_ROOT);
foreach(lc, root_relids)
@@ -996,6 +1099,9 @@ AlterPublicationOptions(ParseState *pstate, AlterPublicationStmt *stmt,
values[Anum_pg_publication_pubtruncate - 1] = BoolGetDatum(pubactions.pubtruncate);
replaces[Anum_pg_publication_pubtruncate - 1] = true;
+
+ values[Anum_pg_publication_pubsequence - 1] = BoolGetDatum(pubactions.pubsequence);
+ replaces[Anum_pg_publication_pubsequence - 1] = true;
}
if (publish_via_partition_root_given)
@@ -1015,7 +1121,7 @@ AlterPublicationOptions(ParseState *pstate, AlterPublicationStmt *stmt,
pubform = (Form_pg_publication) GETSTRUCT(tup);
/* Invalidate the relcache. */
- if (pubform->puballtables)
+ if (pubform->puballtables || pubform->puballsequences)
{
CacheInvalidateRelcacheAll();
}
@@ -1031,6 +1137,7 @@ AlterPublicationOptions(ParseState *pstate, AlterPublicationStmt *stmt,
*/
if (root_relids == NIL)
relids = GetPublicationRelations(pubform->oid,
+ PUB_OBJTYPE_TABLE,
PUBLICATION_PART_ALL);
else
{
@@ -1044,7 +1151,20 @@ AlterPublicationOptions(ParseState *pstate, AlterPublicationStmt *stmt,
lfirst_oid(lc));
}
+ /* tables */
schemarelids = GetAllSchemaPublicationRelations(pubform->oid,
+ PUB_OBJTYPE_TABLE,
+ PUBLICATION_PART_ALL);
+ relids = list_concat_unique_oid(relids, schemarelids);
+
+ /* sequences */
+ relids = list_concat_unique_oid(relids,
+ GetPublicationRelations(pubform->oid,
+ PUB_OBJTYPE_SEQUENCE,
+ PUBLICATION_PART_ALL));
+
+ schemarelids = GetAllSchemaPublicationRelations(pubform->oid,
+ PUB_OBJTYPE_SEQUENCE,
PUBLICATION_PART_ALL);
relids = list_concat_unique_oid(relids, schemarelids);
@@ -1099,7 +1219,7 @@ AlterPublicationTables(AlterPublicationStmt *stmt, HeapTuple tup,
if (!tables && stmt->action != AP_SetObjects)
return;
- rels = OpenTableList(tables);
+ rels = OpenRelationList(tables, PUB_OBJTYPE_TABLE);
if (stmt->action == AP_AddObjects)
{
@@ -1110,13 +1230,14 @@ AlterPublicationTables(AlterPublicationStmt *stmt, HeapTuple tup,
CheckPubRelationColumnList(stmt->pubname, rels, publish_schema,
pubform->pubviaroot);
- PublicationAddTables(pubid, rels, false, stmt);
+ PublicationAddRelations(pubid, rels, false, stmt);
}
else if (stmt->action == AP_DropObjects)
- PublicationDropTables(pubid, rels, false);
+ PublicationDropRelations(pubid, rels, false);
else /* AP_SetObjects */
{
List *oldrelids = GetPublicationRelations(pubid,
+ PUB_OBJTYPE_TABLE,
PUBLICATION_PART_ROOT);
List *delrels = NIL;
ListCell *oldlc;
@@ -1233,18 +1354,18 @@ AlterPublicationTables(AlterPublicationStmt *stmt, HeapTuple tup,
}
/* And drop them. */
- PublicationDropTables(pubid, delrels, true);
+ PublicationDropRelations(pubid, delrels, true);
/*
* Don't bother calculating the difference for adding, we'll catch and
* skip existing ones when doing catalog update.
*/
- PublicationAddTables(pubid, rels, true, stmt);
+ PublicationAddRelations(pubid, rels, true, stmt);
- CloseTableList(delrels);
+ CloseRelationList(delrels);
}
- CloseTableList(rels);
+ CloseRelationList(rels);
}
/*
@@ -1254,7 +1375,8 @@ AlterPublicationTables(AlterPublicationStmt *stmt, HeapTuple tup,
*/
static void
AlterPublicationSchemas(AlterPublicationStmt *stmt,
- HeapTuple tup, List *schemaidlist)
+ HeapTuple tup, List *schemaidlist,
+ char objectType)
{
Form_pg_publication pubform = (Form_pg_publication) GETSTRUCT(tup);
@@ -1276,7 +1398,7 @@ AlterPublicationSchemas(AlterPublicationStmt *stmt,
ListCell *lc;
List *reloids;
- reloids = GetPublicationRelations(pubform->oid, PUBLICATION_PART_ROOT);
+ reloids = GetPublicationRelations(pubform->oid, objectType, PUBLICATION_PART_ROOT);
foreach(lc, reloids)
{
@@ -1303,13 +1425,13 @@ AlterPublicationSchemas(AlterPublicationStmt *stmt,
ReleaseSysCache(coltuple);
}
- PublicationAddSchemas(pubform->oid, schemaidlist, false, stmt);
+ PublicationAddSchemas(pubform->oid, schemaidlist, objectType, false, stmt);
}
else if (stmt->action == AP_DropObjects)
- PublicationDropSchemas(pubform->oid, schemaidlist, false);
+ PublicationDropSchemas(pubform->oid, schemaidlist, objectType, false);
else /* AP_SetObjects */
{
- List *oldschemaids = GetPublicationSchemas(pubform->oid);
+ List *oldschemaids = GetPublicationSchemas(pubform->oid, objectType);
List *delschemas = NIL;
/* Identify which schemas should be dropped */
@@ -1322,13 +1444,13 @@ AlterPublicationSchemas(AlterPublicationStmt *stmt,
LockSchemaList(delschemas);
/* And drop them */
- PublicationDropSchemas(pubform->oid, delschemas, true);
+ PublicationDropSchemas(pubform->oid, delschemas, objectType, true);
/*
* Don't bother calculating the difference for adding, we'll catch and
* skip existing ones when doing catalog update.
*/
- PublicationAddSchemas(pubform->oid, schemaidlist, true, stmt);
+ PublicationAddSchemas(pubform->oid, schemaidlist, objectType, true, stmt);
}
}
@@ -1338,12 +1460,13 @@ AlterPublicationSchemas(AlterPublicationStmt *stmt,
*/
static void
CheckAlterPublication(AlterPublicationStmt *stmt, HeapTuple tup,
- List *tables, List *schemaidlist)
+ List *tables, List *tables_schemaidlist,
+ List *sequences, List *sequences_schemaidlist)
{
Form_pg_publication pubform = (Form_pg_publication) GETSTRUCT(tup);
if ((stmt->action == AP_AddObjects || stmt->action == AP_SetObjects) &&
- schemaidlist && !superuser())
+ (tables_schemaidlist || sequences_schemaidlist) && !superuser())
ereport(ERROR,
(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
errmsg("must be superuser to add or set schemas")));
@@ -1352,13 +1475,24 @@ CheckAlterPublication(AlterPublicationStmt *stmt, HeapTuple tup,
* Check that user is allowed to manipulate the publication tables in
* schema
*/
- if (schemaidlist && pubform->puballtables)
+ if (tables_schemaidlist && pubform->puballtables)
ereport(ERROR,
(errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
errmsg("publication \"%s\" is defined as FOR ALL TABLES",
NameStr(pubform->pubname)),
errdetail("Schemas cannot be added to or dropped from FOR ALL TABLES publications.")));
+ /*
+ * Check that user is allowed to manipulate the publication sequences in
+ * schema
+ */
+ if (sequences_schemaidlist && pubform->puballsequences)
+ ereport(ERROR,
+ (errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
+ errmsg("publication \"%s\" is defined as FOR ALL SEQUENCES",
+ NameStr(pubform->pubname)),
+ errdetail("Sequences from schema cannot be added to, dropped from, or set on FOR ALL SEQUENCES publications.")));
+
/* Check that user is allowed to manipulate the publication tables. */
if (tables && pubform->puballtables)
ereport(ERROR,
@@ -1366,6 +1500,107 @@ CheckAlterPublication(AlterPublicationStmt *stmt, HeapTuple tup,
errmsg("publication \"%s\" is defined as FOR ALL TABLES",
NameStr(pubform->pubname)),
errdetail("Tables cannot be added to or dropped from FOR ALL TABLES publications.")));
+
+ /* Check that user is allowed to manipulate the publication sequences. */
+ if (sequences && pubform->puballsequences)
+ ereport(ERROR,
+ (errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
+ errmsg("publication \"%s\" is defined as FOR ALL SEQUENCES",
+ NameStr(pubform->pubname)),
+ errdetail("Sequences cannot be added to or dropped from FOR ALL SEQUENCES publications.")));
+}
+
+/*
+ * Add or remove table to/from publication.
+ */
+static void
+AlterPublicationSequences(AlterPublicationStmt *stmt, HeapTuple tup,
+ List *sequences)
+{
+ List *rels = NIL;
+ Form_pg_publication pubform = (Form_pg_publication) GETSTRUCT(tup);
+ Oid pubid = pubform->oid;
+
+ /*
+ * Nothing to do if no objects, except in SET: for that it is quite
+ * possible that user has not specified any tables in which case we need
+ * to remove all the existing tables.
+ */
+ if (!sequences && stmt->action != AP_SetObjects)
+ return;
+
+ rels = OpenRelationList(sequences, PUB_OBJTYPE_SEQUENCE);
+
+ if (stmt->action == AP_AddObjects)
+ {
+ PublicationAddRelations(pubid, rels, false, stmt);
+ }
+ else if (stmt->action == AP_DropObjects)
+ PublicationDropRelations(pubid, rels, false);
+ else /* AP_SetObjects */
+ {
+ List *oldrelids = GetPublicationRelations(pubid,
+ PUB_OBJTYPE_SEQUENCE,
+ PUBLICATION_PART_ROOT);
+ List *delrels = NIL;
+ ListCell *oldlc;
+
+ /*
+ * To recreate the relation list for the publication, look for
+ * existing relations that do not need to be dropped.
+ */
+ foreach(oldlc, oldrelids)
+ {
+ Oid oldrelid = lfirst_oid(oldlc);
+ ListCell *newlc;
+ PublicationRelInfo *oldrel;
+ bool found = false;
+
+ foreach(newlc, rels)
+ {
+ PublicationRelInfo *newpubrel;
+
+ newpubrel = (PublicationRelInfo *) lfirst(newlc);
+
+ /*
+ * Check if any of the new set of relations matches with the
+ * existing relations in the publication.
+ */
+ if (RelationGetRelid(newpubrel->relation) == oldrelid)
+ {
+ found = true;
+ break;
+ }
+ }
+
+ /*
+ * Add the non-matched relations to a list so that they can be
+ * dropped.
+ */
+ if (!found)
+ {
+ oldrel = palloc(sizeof(PublicationRelInfo));
+ oldrel->whereClause = NULL;
+ oldrel->columns = NIL;
+ oldrel->relation = table_open(oldrelid,
+ ShareUpdateExclusiveLock);
+ delrels = lappend(delrels, oldrel);
+ }
+ }
+
+ /* And drop them. */
+ PublicationDropRelations(pubid, delrels, true);
+
+ /*
+ * Don't bother calculating the difference for adding, we'll catch and
+ * skip existing ones when doing catalog update.
+ */
+ PublicationAddRelations(pubid, rels, true, stmt);
+
+ CloseRelationList(delrels);
+ }
+
+ CloseRelationList(rels);
}
/*
@@ -1403,14 +1638,20 @@ AlterPublication(ParseState *pstate, AlterPublicationStmt *stmt)
AlterPublicationOptions(pstate, stmt, rel, tup);
else
{
- List *relations = NIL;
- List *schemaidlist = NIL;
+ List *tables = NIL;
+ List *sequences = NIL;
+ List *tables_schemaidlist = NIL;
+ List *sequences_schemaidlist = NIL;
Oid pubid = pubform->oid;
- ObjectsInPublicationToOids(stmt->pubobjects, pstate, &relations,
- &schemaidlist);
+ ObjectsInPublicationToOids(stmt->pubobjects, pstate,
+ &tables, &sequences,
+ &tables_schemaidlist,
+ &sequences_schemaidlist);
- CheckAlterPublication(stmt, tup, relations, schemaidlist);
+ CheckAlterPublication(stmt, tup,
+ tables, tables_schemaidlist,
+ sequences, sequences_schemaidlist);
heap_freetuple(tup);
@@ -1431,9 +1672,16 @@ AlterPublication(ParseState *pstate, AlterPublicationStmt *stmt)
errmsg("publication \"%s\" does not exist",
stmt->pubname));
- AlterPublicationTables(stmt, tup, relations, pstate->p_sourcetext,
- schemaidlist != NIL);
- AlterPublicationSchemas(stmt, tup, schemaidlist);
+ AlterPublicationTables(stmt, tup, tables, pstate->p_sourcetext,
+ tables_schemaidlist != NIL);
+
+ AlterPublicationSchemas(stmt, tup, tables_schemaidlist,
+ PUB_OBJTYPE_TABLE);
+
+ AlterPublicationSequences(stmt, tup, sequences);
+
+ AlterPublicationSchemas(stmt, tup, sequences_schemaidlist,
+ PUB_OBJTYPE_SEQUENCE);
}
/* Cleanup. */
@@ -1501,7 +1749,7 @@ RemovePublicationById(Oid pubid)
pubform = (Form_pg_publication) GETSTRUCT(tup);
/* Invalidate relcache so that publication info is rebuilt. */
- if (pubform->puballtables)
+ if (pubform->puballtables || pubform->puballsequences)
CacheInvalidateRelcacheAll();
CatalogTupleDelete(rel, &tup->t_self);
@@ -1537,6 +1785,7 @@ RemovePublicationSchemaById(Oid psoid)
* partitions.
*/
schemaRels = GetSchemaPublicationRelations(pubsch->pnnspid,
+ pubsch->pntype,
PUBLICATION_PART_ALL);
InvalidatePublicationRels(schemaRels);
@@ -1553,10 +1802,10 @@ RemovePublicationSchemaById(Oid psoid)
* add them to a publication.
*/
static List *
-OpenTableList(List *tables)
+OpenRelationList(List *rels, char objectType)
{
List *relids = NIL;
- List *rels = NIL;
+ List *result = NIL;
ListCell *lc;
List *relids_with_rf = NIL;
List *relids_with_collist = NIL;
@@ -1564,19 +1813,35 @@ OpenTableList(List *tables)
/*
* Open, share-lock, and check all the explicitly-specified relations
*/
- foreach(lc, tables)
+ foreach(lc, rels)
{
PublicationTable *t = lfirst_node(PublicationTable, lc);
bool recurse = t->relation->inh;
Relation rel;
Oid myrelid;
PublicationRelInfo *pub_rel;
+ char myrelkind;
/* Allow query cancel in case this takes a long time */
CHECK_FOR_INTERRUPTS();
rel = table_openrv(t->relation, ShareUpdateExclusiveLock);
myrelid = RelationGetRelid(rel);
+ myrelkind = get_rel_relkind(myrelid);
+
+ /*
+ * Make sure the relkind matches the expected object type. This may
+ * happen e.g. when adding a sequence using ADD TABLE or a table using
+ * ADD SEQUENCE).
+ *
+ * XXX We let through unsupported object types (views etc.). Those
+ * will be caught later in check_publication_add_relation.
+ */
+ if (pub_get_object_type_for_relkind(myrelkind) != PUB_OBJTYPE_UNSUPPORTED &&
+ pub_get_object_type_for_relkind(myrelkind) != objectType)
+ ereport(ERROR,
+ errcode(ERRCODE_INVALID_PARAMETER_VALUE),
+ errmsg("object type does not match type expected by command"));
/*
* Filter out duplicates if user specifies "foo, foo".
@@ -1609,7 +1874,7 @@ OpenTableList(List *tables)
pub_rel->relation = rel;
pub_rel->whereClause = t->whereClause;
pub_rel->columns = t->columns;
- rels = lappend(rels, pub_rel);
+ result = lappend(result, pub_rel);
relids = lappend_oid(relids, myrelid);
if (t->whereClause)
@@ -1678,10 +1943,9 @@ OpenTableList(List *tables)
pub_rel->relation = rel;
/* child inherits WHERE clause from parent */
pub_rel->whereClause = t->whereClause;
-
/* child inherits column list from parent */
pub_rel->columns = t->columns;
- rels = lappend(rels, pub_rel);
+ result = lappend(result, pub_rel);
relids = lappend_oid(relids, childrelid);
if (t->whereClause)
@@ -1696,14 +1960,14 @@ OpenTableList(List *tables)
list_free(relids);
list_free(relids_with_rf);
- return rels;
+ return result;
}
/*
* Close all relations in the list.
*/
static void
-CloseTableList(List *rels)
+CloseRelationList(List *rels)
{
ListCell *lc;
@@ -1751,12 +2015,12 @@ LockSchemaList(List *schemalist)
* Add listed tables to the publication.
*/
static void
-PublicationAddTables(Oid pubid, List *rels, bool if_not_exists,
- AlterPublicationStmt *stmt)
+PublicationAddRelations(Oid pubid, List *rels, bool if_not_exists,
+ AlterPublicationStmt *stmt)
{
ListCell *lc;
- Assert(!stmt || !stmt->for_all_tables);
+ Assert(!stmt || !stmt->for_all_objects);
foreach(lc, rels)
{
@@ -1785,7 +2049,7 @@ PublicationAddTables(Oid pubid, List *rels, bool if_not_exists,
* Remove listed tables from the publication.
*/
static void
-PublicationDropTables(Oid pubid, List *rels, bool missing_ok)
+PublicationDropRelations(Oid pubid, List *rels, bool missing_ok)
{
ObjectAddress obj;
ListCell *lc;
@@ -1830,19 +2094,19 @@ PublicationDropTables(Oid pubid, List *rels, bool missing_ok)
* Add listed schemas to the publication.
*/
static void
-PublicationAddSchemas(Oid pubid, List *schemas, bool if_not_exists,
- AlterPublicationStmt *stmt)
+PublicationAddSchemas(Oid pubid, List *schemas, char objectType,
+ bool if_not_exists, AlterPublicationStmt *stmt)
{
ListCell *lc;
- Assert(!stmt || !stmt->for_all_tables);
+ Assert(!stmt || !stmt->for_all_objects);
foreach(lc, schemas)
{
Oid schemaid = lfirst_oid(lc);
ObjectAddress obj;
- obj = publication_add_schema(pubid, schemaid, if_not_exists);
+ obj = publication_add_schema(pubid, schemaid, objectType, if_not_exists);
if (stmt)
{
EventTriggerCollectSimpleCommand(obj, InvalidObjectAddress,
@@ -1858,7 +2122,7 @@ PublicationAddSchemas(Oid pubid, List *schemas, bool if_not_exists,
* Remove listed schemas from the publication.
*/
static void
-PublicationDropSchemas(Oid pubid, List *schemas, bool missing_ok)
+PublicationDropSchemas(Oid pubid, List *schemas, char objectType, bool missing_ok)
{
ObjectAddress obj;
ListCell *lc;
@@ -1868,10 +2132,11 @@ PublicationDropSchemas(Oid pubid, List *schemas, bool missing_ok)
{
Oid schemaid = lfirst_oid(lc);
- psid = GetSysCacheOid2(PUBLICATIONNAMESPACEMAP,
+ psid = GetSysCacheOid3(PUBLICATIONNAMESPACEMAP,
Anum_pg_publication_namespace_oid,
ObjectIdGetDatum(schemaid),
- ObjectIdGetDatum(pubid));
+ ObjectIdGetDatum(pubid),
+ CharGetDatum(objectType));
if (!OidIsValid(psid))
{
if (missing_ok)
@@ -1926,6 +2191,13 @@ AlterPublicationOwner_internal(Relation rel, HeapTuple tup, Oid newOwnerId)
NameStr(form->pubname)),
errhint("The owner of a FOR ALL TABLES publication must be a superuser.")));
+ if (form->puballsequences && !superuser_arg(newOwnerId))
+ ereport(ERROR,
+ (errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
+ errmsg("permission denied to change owner of publication \"%s\"",
+ NameStr(form->pubname)),
+ errhint("The owner of a FOR ALL SEQUENCES publication must be a superuser.")));
+
if (!superuser_arg(newOwnerId) && is_schema_publication(form->oid))
ereport(ERROR,
(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
diff --git a/src/backend/commands/sequence.c b/src/backend/commands/sequence.c
index 616b8685e39..d8404a6f75e 100644
--- a/src/backend/commands/sequence.c
+++ b/src/backend/commands/sequence.c
@@ -44,6 +44,7 @@
#include "utils/acl.h"
#include "utils/builtins.h"
#include "utils/lsyscache.h"
+#include "utils/pg_lsn.h"
#include "utils/resowner.h"
#include "utils/syscache.h"
#include "utils/varlena.h"
@@ -101,7 +102,8 @@ static Relation lock_and_open_sequence(SeqTable seq);
static void create_seq_hashtable(void);
static void init_sequence(Oid relid, SeqTable *p_elm, Relation *p_rel);
static Form_pg_sequence_data read_seq_tuple(Relation rel,
- Buffer *buf, HeapTuple seqdatatuple);
+ Buffer *buf, HeapTuple seqdatatuple,
+ XLogRecPtr *lsn);
static void init_params(ParseState *pstate, List *options, bool for_identity,
bool isInit,
Form_pg_sequence seqform,
@@ -284,7 +286,7 @@ ResetSequence(Oid seq_relid)
* indeed a sequence.
*/
init_sequence(seq_relid, &elm, &seq_rel);
- (void) read_seq_tuple(seq_rel, &buf, &seqdatatuple);
+ (void) read_seq_tuple(seq_rel, &buf, &seqdatatuple, NULL);
pgstuple = SearchSysCache1(SEQRELID, ObjectIdGetDatum(seq_relid));
if (!HeapTupleIsValid(pgstuple))
@@ -342,6 +344,166 @@ ResetSequence(Oid seq_relid)
relation_close(seq_rel, NoLock);
}
+/*
+ * Update the sequence state by modifying the existing sequence data row.
+ *
+ * This keeps the same relfilenode, so the behavior is non-transactional.
+ */
+static void
+SetSequence_non_transactional(Oid seqrelid, int64 value)
+{
+ SeqTable elm;
+ Relation seqrel;
+ Buffer buf;
+ HeapTupleData seqdatatuple;
+ Form_pg_sequence_data seq;
+
+ /* open and lock sequence */
+ init_sequence(seqrelid, &elm, &seqrel);
+
+ /* lock page' buffer and read tuple */
+ seq = read_seq_tuple(seqrel, &buf, &seqdatatuple, NULL);
+
+ /* check the comment above nextval_internal()'s equivalent call. */
+ if (RelationNeedsWAL(seqrel))
+ {
+ GetTopTransactionId();
+
+ if (XLogLogicalInfoActive())
+ GetCurrentTransactionId();
+ }
+
+ /* ready to change the on-disk (or really, in-buffer) tuple */
+ START_CRIT_SECTION();
+
+ seq->last_value = value;
+ seq->is_called = true;
+ seq->log_cnt = 0;
+
+ MarkBufferDirty(buf);
+
+ /* XLOG stuff */
+ if (RelationNeedsWAL(seqrel))
+ {
+ xl_seq_rec xlrec;
+ XLogRecPtr recptr;
+ Page page = BufferGetPage(buf);
+
+ XLogBeginInsert();
+ XLogRegisterBuffer(0, buf, REGBUF_WILL_INIT);
+
+ xlrec.locator = seqrel->rd_locator;
+
+ XLogRegisterData((char *) &xlrec, sizeof(xl_seq_rec));
+ XLogRegisterData((char *) seqdatatuple.t_data, seqdatatuple.t_len);
+
+ /* allow filtering by origin on a sequence update */
+ XLogSetRecordFlags(XLOG_INCLUDE_ORIGIN);
+
+ recptr = XLogInsert(RM_SEQ_ID, XLOG_SEQ_LOG);
+
+ PageSetLSN(page, recptr);
+ }
+
+ END_CRIT_SECTION();
+
+ UnlockReleaseBuffer(buf);
+
+ /* Clear local cache so that we don't think we have cached numbers */
+ /* Note that we do not change the currval() state */
+ elm->cached = elm->last;
+
+ relation_close(seqrel, NoLock);
+}
+
+/*
+ * Update the sequence state by creating a new relfilenode.
+ *
+ * This creates a new relfilenode, to allow transactional behavior.
+ */
+static void
+SetSequence_transactional(Oid seq_relid, int64 value)
+{
+ SeqTable elm;
+ Relation seqrel;
+ Buffer buf;
+ HeapTupleData seqdatatuple;
+ Form_pg_sequence_data seq;
+ HeapTuple tuple;
+
+ /* open and lock sequence */
+ init_sequence(seq_relid, &elm, &seqrel);
+
+ /* lock page' buffer and read tuple */
+ seq = read_seq_tuple(seqrel, &buf, &seqdatatuple, NULL);
+
+ /* Copy the existing sequence tuple. */
+ tuple = heap_copytuple(&seqdatatuple);
+
+ /* Now we're done with the old page */
+ UnlockReleaseBuffer(buf);
+
+ /*
+ * Modify the copied tuple to update the sequence state (similar to what
+ * ResetSequence does).
+ */
+ seq = (Form_pg_sequence_data) GETSTRUCT(tuple);
+ seq->last_value = value;
+ seq->is_called = true;
+ seq->log_cnt = 0;
+
+ /* make sure the relfilenode creation is associated with the XID */
+ if (XLogLogicalInfoActive())
+ GetCurrentTransactionId();
+
+ /*
+ * Create a new storage file for the sequence - this is needed for the
+ * transactional behavior.
+ */
+ RelationSetNewRelfilenumber(seqrel, seqrel->rd_rel->relpersistence);
+
+ /*
+ * Ensure sequence's relfrozenxid is at 0, since it won't contain any
+ * unfrozen XIDs. Same with relminmxid, since a sequence will never
+ * contain multixacts.
+ */
+ Assert(seqrel->rd_rel->relfrozenxid == InvalidTransactionId);
+ Assert(seqrel->rd_rel->relminmxid == InvalidMultiXactId);
+
+ /*
+ * Insert the modified tuple into the new storage file. This does all the
+ * necessary WAL-logging etc.
+ */
+ fill_seq_with_data(seqrel, tuple);
+
+ /* Clear local cache so that we don't think we have cached numbers */
+ /* Note that we do not change the currval() state */
+ elm->cached = elm->last;
+
+ relation_close(seqrel, NoLock);
+}
+
+/*
+ * Set a sequence to a specified internal state.
+ *
+ * The change is made transactionally, so that on failure of the current
+ * transaction, the sequence will be restored to its previous state.
+ * We do that by creating a whole new relfilenode for the sequence; so this
+ * works much like the rewriting forms of ALTER TABLE.
+ *
+ * Caller is assumed to have acquired AccessExclusiveLock on the sequence,
+ * which must not be released until end of transaction. Caller is also
+ * responsible for permissions checking.
+ */
+void
+SetSequence(Oid seq_relid, bool transactional, int64 value)
+{
+ if (transactional)
+ SetSequence_transactional(seq_relid, value);
+ else
+ SetSequence_non_transactional(seq_relid, value);
+}
+
/*
* Initialize a sequence's relation with the specified tuple as content
*
@@ -504,7 +666,7 @@ AlterSequence(ParseState *pstate, AlterSeqStmt *stmt)
seqform = (Form_pg_sequence) GETSTRUCT(seqtuple);
/* lock page buffer and read tuple into new sequence structure */
- (void) read_seq_tuple(seqrel, &buf, &datatuple);
+ (void) read_seq_tuple(seqrel, &buf, &datatuple, NULL);
/* copy the existing sequence data tuple, so it can be modified locally */
newdatatuple = heap_copytuple(&datatuple);
@@ -598,7 +760,7 @@ SequenceChangePersistence(Oid relid, char newrelpersistence)
GetCurrentTransactionId();
}
- (void) read_seq_tuple(seqrel, &buf, &seqdatatuple);
+ (void) read_seq_tuple(seqrel, &buf, &seqdatatuple, NULL);
RelationSetNewRelfilenumber(seqrel, newrelpersistence);
fill_seq_with_data(seqrel, &seqdatatuple);
UnlockReleaseBuffer(buf);
@@ -727,7 +889,7 @@ nextval_internal(Oid relid, bool check_permissions)
ReleaseSysCache(pgstuple);
/* lock page buffer and read tuple */
- seq = read_seq_tuple(seqrel, &buf, &seqdatatuple);
+ seq = read_seq_tuple(seqrel, &buf, &seqdatatuple, NULL);
page = BufferGetPage(buf);
elm->increment = incby;
@@ -1050,7 +1212,7 @@ do_setval(Oid relid, int64 next, bool iscalled)
PreventCommandIfParallelMode("setval()");
/* lock page buffer and read tuple */
- seq = read_seq_tuple(seqrel, &buf, &seqdatatuple);
+ seq = read_seq_tuple(seqrel, &buf, &seqdatatuple, NULL);
if ((next < minv) || (next > maxv))
ereport(ERROR,
@@ -1270,11 +1432,13 @@ init_sequence(Oid relid, SeqTable *p_elm, Relation *p_rel)
* *buf receives the reference to the pinned-and-ex-locked buffer
* *seqdatatuple receives the reference to the sequence tuple proper
* (this arg should point to a local variable of type HeapTupleData)
+ * *lsn receives LSN of the last sequence change (page LSN), optional
*
* Function's return value points to the data payload of the tuple
*/
static Form_pg_sequence_data
-read_seq_tuple(Relation rel, Buffer *buf, HeapTuple seqdatatuple)
+read_seq_tuple(Relation rel, Buffer *buf, HeapTuple seqdatatuple,
+ XLogRecPtr *lsn)
{
Page page;
ItemId lp;
@@ -1291,6 +1455,13 @@ read_seq_tuple(Relation rel, Buffer *buf, HeapTuple seqdatatuple)
elog(ERROR, "bad magic number in sequence \"%s\": %08X",
RelationGetRelationName(rel), sm->magic);
+ /*
+ * If the caller requested it, set the page LSN. This allows deciding
+ * which sequence changes are before/after the returned sequence state.
+ */
+ if (lsn)
+ *lsn = PageGetLSN(page);
+
lp = PageGetItemId(page, FirstOffsetNumber);
Assert(ItemIdIsNormal(lp));
@@ -1886,7 +2057,7 @@ pg_sequence_last_value(PG_FUNCTION_ARGS)
errmsg("permission denied for sequence %s",
RelationGetRelationName(seqrel))));
- seq = read_seq_tuple(seqrel, &buf, &seqtuple);
+ seq = read_seq_tuple(seqrel, &buf, &seqtuple, NULL);
is_called = seq->is_called;
result = seq->last_value;
@@ -1900,6 +2071,67 @@ pg_sequence_last_value(PG_FUNCTION_ARGS)
PG_RETURN_NULL();
}
+/*
+ * Return the current on-disk state of the sequence.
+ *
+ * Note: This is roughly equivalent to selecting the data from the sequence,
+ * except that it also returns the page LSN.
+ */
+Datum
+pg_sequence_state(PG_FUNCTION_ARGS)
+{
+ Oid relid = PG_GETARG_OID(0);
+ SeqTable elm;
+ Relation seqrel;
+ Buffer buf;
+ HeapTupleData seqtuple;
+ Form_pg_sequence_data seq;
+ Datum result;
+
+ int64 last_value;
+ int64 log_cnt;
+ bool is_called;
+ XLogRecPtr lsn;
+
+ TupleDesc tupdesc;
+ HeapTuple tuple;
+ Datum values[4];
+ bool nulls[4];
+
+ if (get_call_result_type(fcinfo, NULL, &tupdesc) != TYPEFUNC_COMPOSITE)
+ elog(ERROR, "return type must be a row type");
+
+ /* open and lock sequence */
+ init_sequence(relid, &elm, &seqrel);
+
+ if (pg_class_aclcheck(elm->relid, GetUserId(),
+ ACL_SELECT | ACL_USAGE) != ACLCHECK_OK)
+ ereport(ERROR,
+ (errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
+ errmsg("permission denied for sequence %s",
+ RelationGetRelationName(seqrel))));
+
+ seq = read_seq_tuple(seqrel, &buf, &seqtuple, &lsn);
+
+ is_called = seq->is_called;
+ last_value = seq->last_value;
+ log_cnt = seq->log_cnt;
+
+ UnlockReleaseBuffer(buf);
+ relation_close(seqrel, NoLock);
+
+ values[0] = LSNGetDatum(lsn);
+ values[1] = Int64GetDatum(last_value);
+ values[2] = Int64GetDatum(log_cnt);
+ values[3] = BoolGetDatum(is_called);
+
+ memset(nulls, 0, sizeof(nulls));
+
+ tuple = heap_form_tuple(tupdesc, values, nulls);
+ result = HeapTupleGetDatum(tuple);
+
+ PG_RETURN_DATUM(result);
+}
void
seq_redo(XLogReaderState *record)
diff --git a/src/backend/commands/subscriptioncmds.c b/src/backend/commands/subscriptioncmds.c
index edc82c11beb..0e06b36611f 100644
--- a/src/backend/commands/subscriptioncmds.c
+++ b/src/backend/commands/subscriptioncmds.c
@@ -100,6 +100,7 @@ typedef struct SubOpts
} SubOpts;
static List *fetch_table_list(WalReceiverConn *wrconn, List *publications);
+static List *fetch_sequence_list(WalReceiverConn *wrconn, List *publications);
static void check_publications_origin(WalReceiverConn *wrconn,
List *publications, bool copydata,
char *origin, Oid *subrel_local_oids,
@@ -730,9 +731,9 @@ CreateSubscription(ParseState *pstate, CreateSubscriptionStmt *stmt,
{
char *err;
WalReceiverConn *wrconn;
- List *tables;
+ List *relations;
ListCell *lc;
- char table_state;
+ char sync_state;
bool must_use_password;
/* Try to connect to the publisher. */
@@ -754,14 +755,17 @@ CreateSubscription(ParseState *pstate, CreateSubscriptionStmt *stmt,
* Set sync state based on if we were asked to do data copy or
* not.
*/
- table_state = opts.copy_data ? SUBREL_STATE_INIT : SUBREL_STATE_READY;
+ sync_state = opts.copy_data ? SUBREL_STATE_INIT : SUBREL_STATE_READY;
/*
- * Get the table list from publisher and build local table status
- * info.
+ * Get the table and sequence list from publisher and build local
+ * relation sync status info.
*/
- tables = fetch_table_list(wrconn, publications);
- foreach(lc, tables)
+ relations = fetch_table_list(wrconn, publications);
+ relations = list_concat(relations,
+ fetch_sequence_list(wrconn, publications));
+
+ foreach(lc, relations)
{
RangeVar *rv = (RangeVar *) lfirst(lc);
Oid relid;
@@ -772,7 +776,7 @@ CreateSubscription(ParseState *pstate, CreateSubscriptionStmt *stmt,
CheckSubscriptionRelkind(get_rel_relkind(relid),
rv->schemaname, rv->relname);
- AddSubscriptionRelState(subid, relid, table_state,
+ AddSubscriptionRelState(subid, relid, sync_state,
InvalidXLogRecPtr);
}
@@ -798,12 +802,12 @@ CreateSubscription(ParseState *pstate, CreateSubscriptionStmt *stmt,
*
* Note that if tables were specified but copy_data is false
* then it is safe to enable two_phase up-front because those
- * tables are already initially in READY state. When the
- * subscription has no tables, we leave the twophase state as
- * PENDING, to allow ALTER SUBSCRIPTION ... REFRESH
+ * relations are already initially in READY state. When the
+ * subscription has no relations, we leave the twophase state
+ * as PENDING, to allow ALTER SUBSCRIPTION ... REFRESH
* PUBLICATION to work.
*/
- if (opts.twophase && !opts.copy_data && tables != NIL)
+ if (opts.twophase && !opts.copy_data && relations != NIL)
twophase_enabled = true;
walrcv_create_slot(wrconn, opts.slot_name, false, twophase_enabled,
@@ -882,8 +886,10 @@ AlterSubscription_refresh(Subscription *sub, bool copy_data,
if (validate_publications)
check_publications(wrconn, validate_publications);
- /* Get the table list from publisher. */
+ /* Get the list of relations from publisher. */
pubrel_names = fetch_table_list(wrconn, sub->publications);
+ pubrel_names = list_concat(pubrel_names,
+ fetch_sequence_list(wrconn, sub->publications));
/* Get local table list. */
subrel_states = GetSubscriptionRelations(sub->oid, false);
@@ -2153,6 +2159,75 @@ fetch_table_list(WalReceiverConn *wrconn, List *publications)
return tablelist;
}
+/*
+ * Get the list of sequences which belong to specified publications on the
+ * publisher connection.
+ */
+static List *
+fetch_sequence_list(WalReceiverConn *wrconn, List *publications)
+{
+ WalRcvExecResult *res;
+ StringInfoData cmd;
+ TupleTableSlot *slot;
+ Oid tableRow[2] = {TEXTOID, TEXTOID};
+ ListCell *lc;
+ bool first;
+ List *tablelist = NIL;
+
+ Assert(list_length(publications) > 0);
+
+ initStringInfo(&cmd);
+ appendStringInfoString(&cmd, "SELECT DISTINCT s.schemaname, s.sequencename\n"
+ " FROM pg_catalog.pg_publication_sequences s\n"
+ " WHERE s.pubname IN (");
+ first = true;
+ foreach(lc, publications)
+ {
+ char *pubname = strVal(lfirst(lc));
+
+ if (first)
+ first = false;
+ else
+ appendStringInfoString(&cmd, ", ");
+
+ appendStringInfoString(&cmd, quote_literal_cstr(pubname));
+ }
+ appendStringInfoChar(&cmd, ')');
+
+ res = walrcv_exec(wrconn, cmd.data, 2, tableRow);
+ pfree(cmd.data);
+
+ if (res->status != WALRCV_OK_TUPLES)
+ ereport(ERROR,
+ (errmsg("could not receive list of replicated sequences from the publisher: %s",
+ res->err)));
+
+ /* Process sequences. */
+ slot = MakeSingleTupleTableSlot(res->tupledesc, &TTSOpsMinimalTuple);
+ while (tuplestore_gettupleslot(res->tuplestore, true, false, slot))
+ {
+ char *nspname;
+ char *relname;
+ bool isnull;
+ RangeVar *rv;
+
+ nspname = TextDatumGetCString(slot_getattr(slot, 1, &isnull));
+ Assert(!isnull);
+ relname = TextDatumGetCString(slot_getattr(slot, 2, &isnull));
+ Assert(!isnull);
+
+ rv = makeRangeVar(nspname, relname, -1);
+ tablelist = lappend(tablelist, rv);
+
+ ExecClearTuple(slot);
+ }
+ ExecDropSingleTupleTableSlot(slot);
+
+ walrcv_clear_result(res);
+
+ return tablelist;
+}
+
/*
* This is to report the connection failure while dropping replication slots.
* Here, we report the WARNING for all tablesync slots so that user can drop
diff --git a/src/backend/executor/execReplication.c b/src/backend/executor/execReplication.c
index 81f27042bc4..1d5639b47fc 100644
--- a/src/backend/executor/execReplication.c
+++ b/src/backend/executor/execReplication.c
@@ -745,7 +745,9 @@ void
CheckSubscriptionRelkind(char relkind, const char *nspname,
const char *relname)
{
- if (relkind != RELKIND_RELATION && relkind != RELKIND_PARTITIONED_TABLE)
+ if (relkind != RELKIND_RELATION &&
+ relkind != RELKIND_PARTITIONED_TABLE &&
+ relkind != RELKIND_SEQUENCE)
ereport(ERROR,
(errcode(ERRCODE_WRONG_OBJECT_TYPE),
errmsg("cannot use relation \"%s.%s\" as logical replication target",
diff --git a/src/backend/parser/gram.y b/src/backend/parser/gram.y
index d631ac89a91..700020aa273 100644
--- a/src/backend/parser/gram.y
+++ b/src/backend/parser/gram.y
@@ -452,7 +452,7 @@ static Node *makeRecursiveViewSelect(char *relname, List *aliases, Node *query);
transform_element_list transform_type_list
TriggerTransitions TriggerReferencing
vacuum_relation_list opt_vacuum_relation_list
- drop_option_list pub_obj_list
+ drop_option_list pub_obj_list pub_obj_type_list
%type <node> opt_routine_body
%type <groupclause> group_clause
@@ -586,6 +586,7 @@ static Node *makeRecursiveViewSelect(char *relname, List *aliases, Node *query);
%type <node> var_value zone_value
%type <rolespec> auth_ident RoleSpec opt_granted_by
%type <publicationobjectspec> PublicationObjSpec
+%type <node> pub_obj_type
%type <keyword> unreserved_keyword type_func_name_keyword
%type <keyword> col_name_keyword reserved_keyword
@@ -10427,12 +10428,16 @@ AlterOwnerStmt: ALTER AGGREGATE aggregate_with_argtypes OWNER TO RoleSpec
*
* CREATE PUBLICATION FOR ALL TABLES [WITH options]
*
+ * CREATE PUBLICATION FOR ALL SEQUENCES [WITH options]
+ *
* CREATE PUBLICATION FOR pub_obj [, ...] [WITH options]
*
* pub_obj is one of:
*
* TABLE table [, ...]
+ * SEQUENCE table [, ...]
* TABLES IN SCHEMA schema [, ...]
+ * SEQUENCES IN SCHEMA schema [, ...]
*
*****************************************************************************/
@@ -10445,13 +10450,13 @@ CreatePublicationStmt:
n->options = $4;
$$ = (Node *) n;
}
- | CREATE PUBLICATION name FOR ALL TABLES opt_definition
+ | CREATE PUBLICATION name FOR ALL pub_obj_type_list opt_definition
{
CreatePublicationStmt *n = makeNode(CreatePublicationStmt);
n->pubname = $3;
n->options = $7;
- n->for_all_tables = true;
+ n->for_all_objects = $6;
$$ = (Node *) n;
}
| CREATE PUBLICATION name FOR pub_obj_list opt_definition
@@ -10502,6 +10507,26 @@ PublicationObjSpec:
$$->pubobjtype = PUBLICATIONOBJ_TABLES_IN_CUR_SCHEMA;
$$->location = @4;
}
+ | SEQUENCE relation_expr
+ {
+ $$ = makeNode(PublicationObjSpec);
+ $$->pubobjtype = PUBLICATIONOBJ_SEQUENCE;
+ $$->pubtable = makeNode(PublicationTable);
+ $$->pubtable->relation = $2;
+ }
+ | SEQUENCES IN_P SCHEMA ColId
+ {
+ $$ = makeNode(PublicationObjSpec);
+ $$->pubobjtype = PUBLICATIONOBJ_SEQUENCES_IN_SCHEMA;
+ $$->name = $4;
+ $$->location = @4;
+ }
+ | SEQUENCES IN_P SCHEMA CURRENT_SCHEMA
+ {
+ $$ = makeNode(PublicationObjSpec);
+ $$->pubobjtype = PUBLICATIONOBJ_SEQUENCES_IN_CUR_SCHEMA;
+ $$->location = @4;
+ }
| ColId opt_column_list OptWhereClause
{
$$ = makeNode(PublicationObjSpec);
@@ -10563,6 +10588,19 @@ pub_obj_list: PublicationObjSpec
{ $$ = lappend($1, $3); }
;
+pub_obj_type: TABLES
+ { $$ = (Node *) makeString("tables"); }
+ | SEQUENCES
+ { $$ = (Node *) makeString("sequences"); }
+ ;
+
+pub_obj_type_list: pub_obj_type
+ { $$ = list_make1($1); }
+ | pub_obj_type_list ',' pub_obj_type
+ { $$ = lappend($1, $3); }
+ ;
+
+
/*****************************************************************************
*
* ALTER PUBLICATION name SET ( options )
@@ -10576,7 +10614,9 @@ pub_obj_list: PublicationObjSpec
* pub_obj is one of:
*
* TABLE table_name [, ...]
+ * SEQUENCE table_name [, ...]
* TABLES IN SCHEMA schema_name [, ...]
+ * SEQUENCES IN SCHEMA schema_name [, ...]
*
*****************************************************************************/
@@ -18936,7 +18976,8 @@ preprocess_pubobj_list(List *pubobjspec_list, core_yyscan_t yyscanner)
if (pubobj->pubobjtype == PUBLICATIONOBJ_CONTINUATION)
pubobj->pubobjtype = prevobjtype;
- if (pubobj->pubobjtype == PUBLICATIONOBJ_TABLE)
+ if (pubobj->pubobjtype == PUBLICATIONOBJ_TABLE ||
+ pubobj->pubobjtype == PUBLICATIONOBJ_SEQUENCE)
{
/* relation name or pubtable must be set for this type of object */
if (!pubobj->name && !pubobj->pubtable)
@@ -18987,6 +19028,30 @@ preprocess_pubobj_list(List *pubobjspec_list, core_yyscan_t yyscanner)
errmsg("invalid schema name"),
parser_errposition(pubobj->location));
}
+ else if (pubobj->pubobjtype == PUBLICATIONOBJ_SEQUENCES_IN_SCHEMA ||
+ pubobj->pubobjtype == PUBLICATIONOBJ_SEQUENCES_IN_CUR_SCHEMA)
+ {
+ /* WHERE clause is not allowed on a schema object */
+ if (pubobj->pubtable && pubobj->pubtable->whereClause)
+ ereport(ERROR,
+ errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("WHERE clause not allowed for schema"),
+ parser_errposition(pubobj->location));
+
+ /*
+ * We can distinguish between the different type of schema
+ * objects based on whether name and pubtable is set.
+ */
+ if (pubobj->name)
+ pubobj->pubobjtype = PUBLICATIONOBJ_SEQUENCES_IN_SCHEMA;
+ else if (!pubobj->name && !pubobj->pubtable)
+ pubobj->pubobjtype = PUBLICATIONOBJ_SEQUENCES_IN_CUR_SCHEMA;
+ else
+ ereport(ERROR,
+ errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("invalid schema name at or near"),
+ parser_errposition(pubobj->location));
+ }
prevobjtype = pubobj->pubobjtype;
}
diff --git a/src/backend/replication/logical/proto.c b/src/backend/replication/logical/proto.c
index aa471dccdf3..9d4fbe92ce0 100644
--- a/src/backend/replication/logical/proto.c
+++ b/src/backend/replication/logical/proto.c
@@ -663,6 +663,59 @@ logicalrep_write_message(StringInfo out, TransactionId xid, XLogRecPtr lsn,
pq_sendbytes(out, message, sz);
}
+/*
+ * Write SEQUENCE to stream
+ */
+void
+logicalrep_write_sequence(StringInfo out, Relation rel, TransactionId xid,
+ XLogRecPtr lsn, bool transactional,
+ int64 value)
+{
+ uint8 flags = 0;
+
+ pq_sendbyte(out, LOGICAL_REP_MSG_SEQUENCE);
+
+ /* transaction ID (if not valid, we're not streaming) */
+ if (TransactionIdIsValid(xid))
+ pq_sendint32(out, xid);
+
+ pq_sendint8(out, flags);
+
+ /* use Oid as relation identifier */
+ pq_sendint32(out, RelationGetRelid(rel));
+
+ /* write the LSN of the sequence change */
+ pq_sendint64(out, lsn);
+
+ /* write sequence info */
+ pq_sendint8(out, transactional);
+ pq_sendint64(out, value);
+}
+
+/*
+ * Read SEQUENCE from the stream.
+ */
+LogicalRepRelId
+logicalrep_read_sequence(StringInfo in, LogicalRepSequence *seqdata)
+{
+ LogicalRepRelId relid;
+
+ /* XXX skipping flags */
+ pq_getmsgint(in, 1);
+
+ /* read the relation id */
+ relid = pq_getmsgint(in, 4);
+
+ /* read the LSN of the sequence change */
+ seqdata->lsn = pq_getmsgint64(in);
+
+ /* info about the sequence change */
+ seqdata->transactional = pq_getmsgint(in, 1);
+ seqdata->value = pq_getmsgint64(in);
+
+ return relid;
+}
+
/*
* Write relation description to the output stream.
*/
@@ -1258,6 +1311,8 @@ logicalrep_message_type(LogicalRepMsgType action)
return "STREAM ABORT";
case LOGICAL_REP_MSG_STREAM_PREPARE:
return "STREAM PREPARE";
+ case LOGICAL_REP_MSG_SEQUENCE:
+ return "SEQUENCE";
}
/*
diff --git a/src/backend/replication/logical/tablesync.c b/src/backend/replication/logical/tablesync.c
index df3c42eb5de..f1312b2f69b 100644
--- a/src/backend/replication/logical/tablesync.c
+++ b/src/backend/replication/logical/tablesync.c
@@ -100,6 +100,7 @@
#include "catalog/pg_subscription_rel.h"
#include "catalog/pg_type.h"
#include "commands/copy.h"
+#include "commands/sequence.h"
#include "miscadmin.h"
#include "nodes/makefuncs.h"
#include "parser/parse_relation.h"
@@ -118,6 +119,7 @@
#include "utils/builtins.h"
#include "utils/lsyscache.h"
#include "utils/memutils.h"
+#include "utils/pg_lsn.h"
#include "utils/rls.h"
#include "utils/snapmgr.h"
#include "utils/syscache.h"
@@ -1229,6 +1231,106 @@ copy_table(Relation rel)
logicalrep_rel_close(relmapentry, NoLock);
}
+/*
+ * Fetch sequence data (current state) from the remote node, including the
+ * page LSN.
+ */
+static int64
+fetch_sequence_data(Oid remoteid, XLogRecPtr *lsn)
+{
+ WalRcvExecResult *res;
+ StringInfoData cmd;
+ TupleTableSlot *slot;
+ Oid tableRow[2] = {INT8OID, LSNOID};
+ int64 value = (Datum) 0;
+
+ initStringInfo(&cmd);
+ appendStringInfo(&cmd, "SELECT (last_value + log_cnt), page_lsn "
+ "FROM pg_sequence_state(%d)", remoteid);
+
+ res = walrcv_exec(LogRepWorkerWalRcvConn, cmd.data, 2, tableRow);
+ pfree(cmd.data);
+
+ if (res->status != WALRCV_OK_TUPLES)
+ ereport(ERROR,
+ (errmsg("could not receive list of replicated tables from the publisher: %s",
+ res->err)));
+
+ /* Process the sequence. */
+ slot = MakeSingleTupleTableSlot(res->tupledesc, &TTSOpsMinimalTuple);
+ while (tuplestore_gettupleslot(res->tuplestore, true, false, slot))
+ {
+ bool isnull;
+
+ value = DatumGetInt64(slot_getattr(slot, 1, &isnull));
+ Assert(!isnull);
+
+ *lsn = DatumGetInt64(slot_getattr(slot, 2, &isnull));
+ Assert(!isnull);
+ }
+
+ ExecDropSingleTupleTableSlot(slot);
+
+ walrcv_clear_result(res);
+
+ return value;
+}
+
+/*
+ * Copy existing data of a sequence from publisher.
+ *
+ * Caller is responsible for locking the local relation.
+ */
+static XLogRecPtr
+copy_sequence(Relation rel)
+{
+ LogicalRepRelMapEntry *relmapentry;
+ LogicalRepRelation lrel;
+ List *qual = NIL;
+ StringInfoData cmd;
+ int64 sequence_value;
+ XLogRecPtr lsn = InvalidXLogRecPtr;
+
+ /* Get the publisher relation info. */
+ fetch_remote_table_info(get_namespace_name(RelationGetNamespace(rel)),
+ RelationGetRelationName(rel), &lrel, &qual);
+
+ /* sequences don't have row filters */
+ Assert(!qual);
+
+ /* Put the relation into relmap. */
+ logicalrep_relmap_update(&lrel);
+
+ /* Map the publisher relation to local one. */
+ relmapentry = logicalrep_rel_open(lrel.remoteid, NoLock);
+ Assert(rel == relmapentry->localrel);
+
+ /* Start copy on the publisher. */
+ initStringInfo(&cmd);
+
+ Assert(lrel.relkind == RELKIND_SEQUENCE);
+
+ /*
+ * Logical replication of sequences is based on decoding WAL records,
+ * describing the "next" state of the sequence the current state in the
+ * relfilenode is yet to reach. But during the initial sync we read the
+ * current state, so we need to reconstruct the WAL record logged when we
+ * started the current batch of sequence values.
+ *
+ * Otherwise we might get duplicate values (on subscriber) if we failed
+ * over right after the sync.
+ */
+ sequence_value = fetch_sequence_data(lrel.remoteid, &lsn);
+
+ /* tablesync sets the sequences in non-transactional way */
+ SetSequence(RelationGetRelid(rel), false, sequence_value);
+
+ logicalrep_rel_close(relmapentry, NoLock);
+
+ /* return the LSN when the sequence state was set */
+ return lsn;
+}
+
/*
* Determine the tablesync slot name.
*
@@ -1277,6 +1379,7 @@ LogicalRepSyncTableStart(XLogRecPtr *origin_startpos)
UserContext ucxt;
bool must_use_password;
bool run_as_owner;
+ XLogRecPtr sequence_lsn = InvalidXLogRecPtr;
/* Check the state of the table synchronization. */
StartTransactionCommand();
@@ -1488,10 +1591,34 @@ LogicalRepSyncTableStart(XLogRecPtr *origin_startpos)
GetUserNameFromId(GetUserId(), true),
RelationGetRelationName(rel))));
- /* Now do the initial data copy */
- PushActiveSnapshot(GetTransactionSnapshot());
- copy_table(rel);
- PopActiveSnapshot();
+ /* Do the right action depending on the relation kind. */
+ if (get_rel_relkind(RelationGetRelid(rel)) == RELKIND_SEQUENCE)
+ {
+ /* Now do the initial sequence copy */
+ PushActiveSnapshot(GetTransactionSnapshot());
+ sequence_lsn = copy_sequence(rel);
+ PopActiveSnapshot();
+
+ /*
+ * Sequences are not consistent (in the MVCC sense) with respect to
+ * the replication slot, so the copy might have read a more recent
+ * state than origin_startpos. The sequence_lsn comes from page LSN
+ * (which is LSN of the last sequence change), so that's the right
+ * position where to start with the catchup apply.
+ *
+ * It might be before the slot, though (if the sequence was not used
+ * since between the slot creation and copy), so make sure the
+ * position does not move backwards.
+ */
+ *origin_startpos = Max(*origin_startpos, sequence_lsn);
+ }
+ else
+ {
+ /* Now do the initial data copy */
+ PushActiveSnapshot(GetTransactionSnapshot());
+ copy_table(rel);
+ PopActiveSnapshot();
+ }
res = walrcv_exec(LogRepWorkerWalRcvConn, "COMMIT", 0, NULL);
if (res->status != WALRCV_OK_COMMAND)
diff --git a/src/backend/replication/logical/worker.c b/src/backend/replication/logical/worker.c
index 21abf34ef7a..a5b52f60a5a 100644
--- a/src/backend/replication/logical/worker.c
+++ b/src/backend/replication/logical/worker.c
@@ -154,6 +154,7 @@
#include "catalog/pg_subscription.h"
#include "catalog/pg_subscription_rel.h"
#include "catalog/pg_tablespace.h"
+#include "commands/sequence.h"
#include "commands/tablecmds.h"
#include "commands/tablespace.h"
#include "commands/trigger.h"
@@ -1437,6 +1438,97 @@ apply_handle_origin(StringInfo s)
errmsg_internal("ORIGIN message sent out of order")));
}
+/*
+ * Handle SEQUENCE message.
+ */
+static void
+apply_handle_sequence(StringInfo s)
+{
+ LogicalRepSequence seq;
+ LogicalRepRelMapEntry *rel;
+ Oid relid;
+ bool already_in_transaction PG_USED_FOR_ASSERTS_ONLY;
+
+ if (is_skipping_changes() ||
+ handle_streamed_transaction(LOGICAL_REP_MSG_SEQUENCE, s))
+ return;
+
+ /*
+ * Remember if we're already in transaction (begin_replication step starts
+ * a transaction, so we can't use IsTransactionState() after that point.
+ */
+ already_in_transaction = IsTransactionState();
+
+ /*
+ * Make sure we're in a transaction (needed by SetSequence). For
+ * non-transactional updates we're guaranteed to start a new one, and
+ * we'll commit it at the end.
+ */
+ begin_replication_step();
+
+ relid = logicalrep_read_sequence(s, &seq);
+
+ /*
+ * Non-transactional sequence updates should not be part of a remote
+ * transaction. There should not be any running transaction.
+ */
+ Assert((!seq.transactional) || in_remote_transaction);
+ Assert(!(!seq.transactional && in_remote_transaction));
+ Assert(!(!seq.transactional && already_in_transaction));
+
+ /* lock the sequence in AccessExclusiveLock, as expected by SetSequence */
+ rel = logicalrep_rel_open(relid, AccessExclusiveLock);
+ if (!should_apply_changes_for_rel(rel))
+ {
+ /*
+ * The relation can't become interesting in the middle of the
+ * transaction so it's safe to unlock it.
+ */
+ logicalrep_rel_close(rel, AccessExclusiveLock);
+ end_replication_step();
+
+ /*
+ * Commit the per-stream transaction (we only do this when not in
+ * remote transaction, i.e. for non-transactional sequence updates.)
+ */
+ if (!in_remote_transaction)
+ CommitTransactionCommand();
+
+ return;
+ }
+
+ /*
+ * apply the sequence change
+ *
+ * XXX We don't need to reconstruct the value from last_value, log_cnt
+ * etc. because that happens on the publisher.
+ */
+ SetSequence(rel->localreloid, seq.transactional, seq.value);
+
+ logicalrep_rel_close(rel, NoLock);
+
+ end_replication_step();
+
+ /*
+ * Commit the per-stream transaction (we only do this when not in remote
+ * transaction, i.e. for non-transactional sequence updates.)
+ */
+ if (!in_remote_transaction)
+ {
+ CommitTransactionCommand();
+
+ /*
+ * Update origin state so we don't try applying this sequence change
+ * in case of crash. We only do this for non-trasactional changes,
+ * that are replayed directly.
+ *
+ * XXX We don't have replorigin_session_origin_timestamp, but we can
+ * just leave that set to 0.
+ */
+ replorigin_session_origin_lsn = seq.lsn;
+ }
+}
+
/*
* Initialize fileset (if not already done).
*
@@ -3347,6 +3439,10 @@ apply_dispatch(StringInfo s)
*/
break;
+ case LOGICAL_REP_MSG_SEQUENCE:
+ apply_handle_sequence(s);
+ return;
+
case LOGICAL_REP_MSG_STREAM_START:
apply_handle_stream_start(s);
break;
@@ -4365,6 +4461,7 @@ set_stream_options(WalRcvStreamOptions *options,
server_version = walrcv_server_version(LogRepWorkerWalRcvConn);
options->proto.logical.proto_version =
+ server_version >= 170000 ? LOGICALREP_PROTO_SEQUENCES_VERSION_NUM :
server_version >= 160000 ? LOGICALREP_PROTO_STREAM_PARALLEL_VERSION_NUM :
server_version >= 150000 ? LOGICALREP_PROTO_TWOPHASE_VERSION_NUM :
server_version >= 140000 ? LOGICALREP_PROTO_STREAM_VERSION_NUM :
diff --git a/src/backend/replication/pgoutput/pgoutput.c b/src/backend/replication/pgoutput/pgoutput.c
index f9ed1083df7..5082ab1f492 100644
--- a/src/backend/replication/pgoutput/pgoutput.c
+++ b/src/backend/replication/pgoutput/pgoutput.c
@@ -15,6 +15,7 @@
#include "access/tupconvert.h"
#include "catalog/partition.h"
#include "catalog/pg_publication.h"
+#include "catalog/pg_publication_namespace.h"
#include "catalog/pg_publication_rel.h"
#include "catalog/pg_subscription.h"
#include "commands/defrem.h"
@@ -55,6 +56,10 @@ static void pgoutput_message(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn, XLogRecPtr message_lsn,
bool transactional, const char *prefix,
Size sz, const char *message);
+static void pgoutput_sequence(LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn, XLogRecPtr sequence_lsn,
+ Relation relation, bool transactional,
+ int64 value);
static bool pgoutput_origin_filter(LogicalDecodingContext *ctx,
RepOriginId origin_id);
static void pgoutput_begin_prepare_txn(LogicalDecodingContext *ctx,
@@ -252,6 +257,7 @@ _PG_output_plugin_init(OutputPluginCallbacks *cb)
cb->change_cb = pgoutput_change;
cb->truncate_cb = pgoutput_truncate;
cb->message_cb = pgoutput_message;
+ cb->sequence_cb = pgoutput_sequence;
cb->commit_cb = pgoutput_commit_txn;
cb->begin_prepare_cb = pgoutput_begin_prepare_txn;
@@ -268,6 +274,7 @@ _PG_output_plugin_init(OutputPluginCallbacks *cb)
cb->stream_commit_cb = pgoutput_stream_commit;
cb->stream_change_cb = pgoutput_change;
cb->stream_message_cb = pgoutput_message;
+ cb->stream_sequence_cb = pgoutput_sequence;
cb->stream_truncate_cb = pgoutput_truncate;
/* transaction streaming - two-phase commit */
cb->stream_prepare_cb = pgoutput_stream_prepare_txn;
@@ -673,7 +680,7 @@ pgoutput_rollback_prepared_txn(LogicalDecodingContext *ctx,
*/
static void
maybe_send_schema(LogicalDecodingContext *ctx,
- ReorderBufferChange *change,
+ ReorderBufferTXN *txn,
Relation relation, RelationSyncEntry *relentry)
{
PGOutputData *data = (PGOutputData *) ctx->output_plugin_private;
@@ -690,10 +697,10 @@ maybe_send_schema(LogicalDecodingContext *ctx,
* the write methods will not include it.
*/
if (data->in_streaming)
- xid = change->txn->xid;
+ xid = txn->xid;
- if (rbtxn_is_subtxn(change->txn))
- topxid = rbtxn_get_toptxn(change->txn)->xid;
+ if (rbtxn_is_subtxn(txn))
+ topxid = rbtxn_get_toptxn(txn)->xid;
else
topxid = xid;
@@ -902,9 +909,10 @@ pgoutput_row_filter_init(PGOutputData *data, List *publications,
* (even if other publications have a row filter).
*/
if (!pub->alltables &&
- !SearchSysCacheExists2(PUBLICATIONNAMESPACEMAP,
+ !SearchSysCacheExists3(PUBLICATIONNAMESPACEMAP,
ObjectIdGetDatum(schemaid),
- ObjectIdGetDatum(pub->oid)))
+ ObjectIdGetDatum(pub->oid),
+ PUB_OBJTYPE_TABLE))
{
/*
* Check for the presence of a row filter in this publication.
@@ -1518,7 +1526,7 @@ pgoutput_change(LogicalDecodingContext *ctx, ReorderBufferTXN *txn,
* Schema should be sent using the original relation because it also sends
* the ancestor's relation.
*/
- maybe_send_schema(ctx, change, relation, relentry);
+ maybe_send_schema(ctx, txn, relation, relentry);
OutputPluginPrepareWrite(ctx, true);
@@ -1603,7 +1611,7 @@ pgoutput_truncate(LogicalDecodingContext *ctx, ReorderBufferTXN *txn,
if (txndata && !txndata->sent_begin_txn)
pgoutput_send_begin(ctx, txn);
- maybe_send_schema(ctx, change, relation, relentry);
+ maybe_send_schema(ctx, change->txn, relation, relentry);
}
if (nrelids > 0)
@@ -1663,6 +1671,70 @@ pgoutput_message(LogicalDecodingContext *ctx, ReorderBufferTXN *txn,
OutputPluginWrite(ctx, true);
}
+static void
+pgoutput_sequence(LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn, XLogRecPtr sequence_lsn,
+ Relation relation, bool transactional,
+ int64 value)
+{
+ PGOutputData *data = (PGOutputData *) ctx->output_plugin_private;
+ TransactionId xid = InvalidTransactionId;
+ RelationSyncEntry *relentry;
+
+ if (!is_publishable_relation(relation))
+ return;
+
+ /*
+ * If the negotiated protocol version does not support sequences, yet we
+ * found a sequence in the publication, error out.
+ */
+ if (data->protocol_version < LOGICALREP_PROTO_SEQUENCES_VERSION_NUM)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("protocol version does not support sequence replication")));
+
+ /*
+ * Remember the xid for the message in streaming mode. See
+ * pgoutput_change.
+ */
+ if (data->in_streaming)
+ xid = txn->xid;
+
+ relentry = get_rel_sync_entry(data, relation);
+
+ /*
+ * First check the sequence filter.
+ *
+ * We handle just REORDER_BUFFER_CHANGE_SEQUENCE here.
+ */
+ if (!relentry->pubactions.pubsequence)
+ return;
+
+ /*
+ * Output BEGIN if we haven't yet. Avoid for non-transactional sequence
+ * changes.
+ */
+ if (transactional)
+ {
+ PGOutputTxnData *txndata = (PGOutputTxnData *) txn->output_plugin_private;
+
+ /* Send BEGIN if we haven't yet */
+ if (txndata && !txndata->sent_begin_txn)
+ pgoutput_send_begin(ctx, txn);
+ }
+
+ maybe_send_schema(ctx, txn, relation, relentry);
+
+ OutputPluginPrepareWrite(ctx, true);
+ logicalrep_write_sequence(ctx->out,
+ relation,
+ xid,
+ sequence_lsn,
+ transactional,
+ value);
+ OutputPluginWrite(ctx, true);
+}
+
/*
* Return true if the data is associated with an origin and the user has
* requested the changes that don't have an origin, false otherwise.
@@ -1978,7 +2050,8 @@ get_rel_sync_entry(PGOutputData *data, Relation relation)
entry->schema_sent = false;
entry->streamed_txns = NIL;
entry->pubactions.pubinsert = entry->pubactions.pubupdate =
- entry->pubactions.pubdelete = entry->pubactions.pubtruncate = false;
+ entry->pubactions.pubdelete = entry->pubactions.pubtruncate =
+ entry->pubactions.pubsequence = false;
entry->new_slot = NULL;
entry->old_slot = NULL;
memset(entry->exprstate, 0, sizeof(entry->exprstate));
@@ -1993,18 +2066,19 @@ get_rel_sync_entry(PGOutputData *data, Relation relation)
{
Oid schemaId = get_rel_namespace(relid);
List *pubids = GetRelationPublications(relid);
+ char relkind = get_rel_relkind(relid);
+ char objectType = pub_get_object_type_for_relkind(relkind);
/*
* We don't acquire a lock on the namespace system table as we build
* the cache entry using a historic snapshot and all the later changes
* are absorbed while decoding WAL.
*/
- List *schemaPubids = GetSchemaPublications(schemaId);
+ List *schemaPubids = GetSchemaPublications(schemaId, objectType);
ListCell *lc;
Oid publish_as_relid = relid;
int publish_ancestor_level = 0;
bool am_partition = get_rel_relispartition(relid);
- char relkind = get_rel_relkind(relid);
List *rel_publications = NIL;
/* Reload publications if needed before use. */
@@ -2036,6 +2110,7 @@ get_rel_sync_entry(PGOutputData *data, Relation relation)
entry->pubactions.pubupdate = false;
entry->pubactions.pubdelete = false;
entry->pubactions.pubtruncate = false;
+ entry->pubactions.pubsequence = false;
/*
* Tuple slots cleanups. (Will be rebuilt later if needed).
@@ -2083,9 +2158,11 @@ get_rel_sync_entry(PGOutputData *data, Relation relation)
/*
* If this is a FOR ALL TABLES publication, pick the partition
- * root and set the ancestor level accordingly.
+ * root and set the ancestor level accordingly. If this is a FOR
+ * ALL SEQUENCES publication, we publish it too but we don't need
+ * to pick the partition root etc.
*/
- if (pub->alltables)
+ if (pub->alltables && (objectType == PUB_OBJTYPE_TABLE))
{
publish = true;
if (pub->pubviaroot && am_partition)
@@ -2096,6 +2173,10 @@ get_rel_sync_entry(PGOutputData *data, Relation relation)
ancestor_level = list_length(ancestors);
}
}
+ else if (pub->allsequences && (objectType == PUB_OBJTYPE_SEQUENCE))
+ {
+ publish = true;
+ }
if (!publish)
{
@@ -2149,6 +2230,7 @@ get_rel_sync_entry(PGOutputData *data, Relation relation)
entry->pubactions.pubupdate |= pub->pubactions.pubupdate;
entry->pubactions.pubdelete |= pub->pubactions.pubdelete;
entry->pubactions.pubtruncate |= pub->pubactions.pubtruncate;
+ entry->pubactions.pubsequence |= pub->pubactions.pubsequence;
/*
* We want to publish the changes as the top-most ancestor
diff --git a/src/backend/utils/cache/relcache.c b/src/backend/utils/cache/relcache.c
index b3faccbefe5..a797756b262 100644
--- a/src/backend/utils/cache/relcache.c
+++ b/src/backend/utils/cache/relcache.c
@@ -57,6 +57,7 @@
#include "catalog/pg_opclass.h"
#include "catalog/pg_proc.h"
#include "catalog/pg_publication.h"
+#include "catalog/pg_publication_namespace.h"
#include "catalog/pg_rewrite.h"
#include "catalog/pg_shseclabel.h"
#include "catalog/pg_statistic_ext.h"
@@ -5715,6 +5716,8 @@ RelationBuildPublicationDesc(Relation relation, PublicationDesc *pubdesc)
Oid schemaid;
List *ancestors = NIL;
Oid relid = RelationGetRelid(relation);
+ char relkind = relation->rd_rel->relkind;
+ char objType;
/*
* If not publishable, it publishes no actions. (pgoutput_change() will
@@ -5745,8 +5748,15 @@ RelationBuildPublicationDesc(Relation relation, PublicationDesc *pubdesc)
/* Fetch the publication membership info. */
puboids = GetRelationPublications(relid);
schemaid = RelationGetNamespace(relation);
- puboids = list_concat_unique_oid(puboids, GetSchemaPublications(schemaid));
+ objType = pub_get_object_type_for_relkind(relkind);
+ puboids = list_concat_unique_oid(puboids,
+ GetSchemaPublications(schemaid, objType));
+
+ /*
+ * If this is a partion (and thus a table), lookup all ancestors and track
+ * all publications them too.
+ */
if (relation->rd_rel->relispartition)
{
/* Add publications that the ancestors are in too. */
@@ -5758,12 +5768,23 @@ RelationBuildPublicationDesc(Relation relation, PublicationDesc *pubdesc)
puboids = list_concat_unique_oid(puboids,
GetRelationPublications(ancestor));
+
+ /* include all publications publishing schema of all ancestors */
schemaid = get_rel_namespace(ancestor);
puboids = list_concat_unique_oid(puboids,
- GetSchemaPublications(schemaid));
+ GetSchemaPublications(schemaid,
+ PUB_OBJTYPE_TABLE));
}
}
- puboids = list_concat_unique_oid(puboids, GetAllTablesPublications());
+
+ /*
+ * Consider also FOR ALL TABLES and FOR ALL SEQUENCES publications,
+ * depending on the relkind of the relation.
+ */
+ if (relation->rd_rel->relkind == RELKIND_SEQUENCE)
+ puboids = list_concat_unique_oid(puboids, GetAllSequencesPublications());
+ else
+ puboids = list_concat_unique_oid(puboids, GetAllTablesPublications());
foreach(lc, puboids)
{
@@ -5782,6 +5803,7 @@ RelationBuildPublicationDesc(Relation relation, PublicationDesc *pubdesc)
pubdesc->pubactions.pubupdate |= pubform->pubupdate;
pubdesc->pubactions.pubdelete |= pubform->pubdelete;
pubdesc->pubactions.pubtruncate |= pubform->pubtruncate;
+ pubdesc->pubactions.pubsequence |= pubform->pubsequence;
/*
* Check if all columns referenced in the filter expression are part
diff --git a/src/backend/utils/cache/syscache.c b/src/backend/utils/cache/syscache.c
index 8dbda0024f9..a080d588db2 100644
--- a/src/backend/utils/cache/syscache.c
+++ b/src/backend/utils/cache/syscache.c
@@ -450,9 +450,10 @@ static const struct cachedesc cacheinfo[] = {
},
[PUBLICATIONNAMESPACEMAP] = {
PublicationNamespaceRelationId,
- PublicationNamespacePnnspidPnpubidIndexId,
+ PublicationNamespacePnnspidPnpubidPntypeIndexId,
KEY(Anum_pg_publication_namespace_pnnspid,
- Anum_pg_publication_namespace_pnpubid),
+ Anum_pg_publication_namespace_pnpubid,
+ Anum_pg_publication_namespace_pntype),
64
},
[PUBLICATIONOID] = {
diff --git a/src/bin/pg_dump/pg_dump.c b/src/bin/pg_dump/pg_dump.c
index 8c0b5486b93..04048d8f629 100644
--- a/src/bin/pg_dump/pg_dump.c
+++ b/src/bin/pg_dump/pg_dump.c
@@ -4067,10 +4067,12 @@ getPublications(Archive *fout, int *numPublications)
int i_pubname;
int i_pubowner;
int i_puballtables;
+ int i_puballsequences;
int i_pubinsert;
int i_pubupdate;
int i_pubdelete;
int i_pubtruncate;
+ int i_pubsequence;
int i_pubviaroot;
int i,
ntups;
@@ -4086,23 +4088,29 @@ getPublications(Archive *fout, int *numPublications)
resetPQExpBuffer(query);
/* Get the publications. */
- if (fout->remoteVersion >= 130000)
+ if (fout->remoteVersion >= 170000)
+ appendPQExpBufferStr(query,
+ "SELECT p.tableoid, p.oid, p.pubname, "
+ "p.pubowner, "
+ "p.puballtables, p.puballsequences, p.pubinsert, p.pubupdate, p.pubdelete, p.pubtruncate, p.pubsequence, p.pubviaroot "
+ "FROM pg_publication p");
+ else if (fout->remoteVersion >= 130000)
appendPQExpBufferStr(query,
"SELECT p.tableoid, p.oid, p.pubname, "
"p.pubowner, "
- "p.puballtables, p.pubinsert, p.pubupdate, p.pubdelete, p.pubtruncate, p.pubviaroot "
+ "p.puballtables, false as p.puballsequences, p.pubinsert, p.pubupdate, p.pubdelete, p.pubtruncate, false as p.pubsequence, p.pubviaroot "
"FROM pg_publication p");
else if (fout->remoteVersion >= 110000)
appendPQExpBufferStr(query,
"SELECT p.tableoid, p.oid, p.pubname, "
"p.pubowner, "
- "p.puballtables, p.pubinsert, p.pubupdate, p.pubdelete, p.pubtruncate, false AS pubviaroot "
+ "p.puballtables, false as p.puballsequences, p.pubinsert, p.pubupdate, p.pubdelete, p.pubtruncate, false AS pubsequence, false AS pubviaroot "
"FROM pg_publication p");
else
appendPQExpBufferStr(query,
"SELECT p.tableoid, p.oid, p.pubname, "
"p.pubowner, "
- "p.puballtables, p.pubinsert, p.pubupdate, p.pubdelete, false AS pubtruncate, false AS pubviaroot "
+ "p.puballtables, false as p.puballsequences, p.pubinsert, p.pubupdate, p.pubdelete, false AS pubtruncate, false AS pubsequence, false AS pubviaroot "
"FROM pg_publication p");
res = ExecuteSqlQuery(fout, query->data, PGRES_TUPLES_OK);
@@ -4114,10 +4122,12 @@ getPublications(Archive *fout, int *numPublications)
i_pubname = PQfnumber(res, "pubname");
i_pubowner = PQfnumber(res, "pubowner");
i_puballtables = PQfnumber(res, "puballtables");
+ i_puballsequences = PQfnumber(res, "puballsequences");
i_pubinsert = PQfnumber(res, "pubinsert");
i_pubupdate = PQfnumber(res, "pubupdate");
i_pubdelete = PQfnumber(res, "pubdelete");
i_pubtruncate = PQfnumber(res, "pubtruncate");
+ i_pubsequence = PQfnumber(res, "pubsequence");
i_pubviaroot = PQfnumber(res, "pubviaroot");
pubinfo = pg_malloc(ntups * sizeof(PublicationInfo));
@@ -4133,6 +4143,8 @@ getPublications(Archive *fout, int *numPublications)
pubinfo[i].rolname = getRoleName(PQgetvalue(res, i, i_pubowner));
pubinfo[i].puballtables =
(strcmp(PQgetvalue(res, i, i_puballtables), "t") == 0);
+ pubinfo[i].puballsequences =
+ (strcmp(PQgetvalue(res, i, i_puballsequences), "t") == 0);
pubinfo[i].pubinsert =
(strcmp(PQgetvalue(res, i, i_pubinsert), "t") == 0);
pubinfo[i].pubupdate =
@@ -4141,6 +4153,8 @@ getPublications(Archive *fout, int *numPublications)
(strcmp(PQgetvalue(res, i, i_pubdelete), "t") == 0);
pubinfo[i].pubtruncate =
(strcmp(PQgetvalue(res, i, i_pubtruncate), "t") == 0);
+ pubinfo[i].pubsequence =
+ (strcmp(PQgetvalue(res, i, i_pubsequence), "t") == 0);
pubinfo[i].pubviaroot =
(strcmp(PQgetvalue(res, i, i_pubviaroot), "t") == 0);
@@ -4186,6 +4200,9 @@ dumpPublication(Archive *fout, const PublicationInfo *pubinfo)
if (pubinfo->puballtables)
appendPQExpBufferStr(query, " FOR ALL TABLES");
+ if (pubinfo->puballsequences)
+ appendPQExpBufferStr(query, " FOR ALL SEQUENCES");
+
appendPQExpBufferStr(query, " WITH (publish = '");
if (pubinfo->pubinsert)
{
@@ -4220,6 +4237,15 @@ dumpPublication(Archive *fout, const PublicationInfo *pubinfo)
first = false;
}
+ if (pubinfo->pubsequence)
+ {
+ if (!first)
+ appendPQExpBufferStr(query, ", ");
+
+ appendPQExpBufferStr(query, "sequence");
+ first = false;
+ }
+
appendPQExpBufferChar(query, '\'');
if (pubinfo->pubviaroot)
@@ -4266,6 +4292,7 @@ getPublicationNamespaces(Archive *fout)
int i_oid;
int i_pnpubid;
int i_pnnspid;
+ int i_pntype;
int i,
j,
ntups;
@@ -4277,7 +4304,7 @@ getPublicationNamespaces(Archive *fout)
/* Collect all publication membership info. */
appendPQExpBufferStr(query,
- "SELECT tableoid, oid, pnpubid, pnnspid "
+ "SELECT tableoid, oid, pnpubid, pnnspid, pntype "
"FROM pg_catalog.pg_publication_namespace");
res = ExecuteSqlQuery(fout, query->data, PGRES_TUPLES_OK);
@@ -4287,6 +4314,7 @@ getPublicationNamespaces(Archive *fout)
i_oid = PQfnumber(res, "oid");
i_pnpubid = PQfnumber(res, "pnpubid");
i_pnnspid = PQfnumber(res, "pnnspid");
+ i_pntype = PQfnumber(res, "pntype");
/* this allocation may be more than we need */
pubsinfo = pg_malloc(ntups * sizeof(PublicationSchemaInfo));
@@ -4296,6 +4324,7 @@ getPublicationNamespaces(Archive *fout)
{
Oid pnpubid = atooid(PQgetvalue(res, i, i_pnpubid));
Oid pnnspid = atooid(PQgetvalue(res, i, i_pnnspid));
+ char pntype = PQgetvalue(res, i, i_pntype)[0];
PublicationInfo *pubinfo;
NamespaceInfo *nspinfo;
@@ -4327,6 +4356,7 @@ getPublicationNamespaces(Archive *fout)
pubsinfo[j].dobj.name = nspinfo->dobj.name;
pubsinfo[j].publication = pubinfo;
pubsinfo[j].pubschema = nspinfo;
+ pubsinfo[j].pubtype = pntype;
/* Decide whether we want to dump it */
selectDumpablePublicationObject(&(pubsinfo[j].dobj), fout);
@@ -4492,7 +4522,11 @@ dumpPublicationNamespace(Archive *fout, const PublicationSchemaInfo *pubsinfo)
query = createPQExpBuffer();
appendPQExpBuffer(query, "ALTER PUBLICATION %s ", fmtId(pubinfo->dobj.name));
- appendPQExpBuffer(query, "ADD TABLES IN SCHEMA %s;\n", fmtId(schemainfo->dobj.name));
+
+ if (pubsinfo->pubtype == 't')
+ appendPQExpBuffer(query, "ADD TABLES IN SCHEMA %s;\n", fmtId(schemainfo->dobj.name));
+ else
+ appendPQExpBuffer(query, "ADD SEQUENCES IN SCHEMA %s;\n", fmtId(schemainfo->dobj.name));
/*
* There is no point in creating drop query as the drop is done by schema
@@ -4525,6 +4559,7 @@ dumpPublicationTable(Archive *fout, const PublicationRelInfo *pubrinfo)
TableInfo *tbinfo = pubrinfo->pubtable;
PQExpBuffer query;
char *tag;
+ char *description;
/* Do nothing in data-only dump */
if (dopt->dataOnly)
@@ -4534,8 +4569,19 @@ dumpPublicationTable(Archive *fout, const PublicationRelInfo *pubrinfo)
query = createPQExpBuffer();
- appendPQExpBuffer(query, "ALTER PUBLICATION %s ADD TABLE ONLY",
- fmtId(pubinfo->dobj.name));
+ if (tbinfo->relkind == RELKIND_SEQUENCE)
+ {
+ appendPQExpBuffer(query, "ALTER PUBLICATION %s ADD SEQUENCE",
+ fmtId(pubinfo->dobj.name));
+ description = "PUBLICATION SEQUENCE";
+ }
+ else
+ {
+ appendPQExpBuffer(query, "ALTER PUBLICATION %s ADD TABLE ONLY",
+ fmtId(pubinfo->dobj.name));
+ description = "PUBLICATION TABLE";
+ }
+
appendPQExpBuffer(query, " %s",
fmtQualifiedDumpable(tbinfo));
@@ -4565,7 +4611,7 @@ dumpPublicationTable(Archive *fout, const PublicationRelInfo *pubrinfo)
ARCHIVE_OPTS(.tag = tag,
.namespace = tbinfo->dobj.namespace->dobj.name,
.owner = pubinfo->rolname,
- .description = "PUBLICATION TABLE",
+ .description = description,
.section = SECTION_POST_DATA,
.createStmt = query->data));
diff --git a/src/bin/pg_dump/pg_dump.h b/src/bin/pg_dump/pg_dump.h
index 2fe3cbed9ac..f6593b0c499 100644
--- a/src/bin/pg_dump/pg_dump.h
+++ b/src/bin/pg_dump/pg_dump.h
@@ -622,10 +622,12 @@ typedef struct _PublicationInfo
DumpableObject dobj;
const char *rolname;
bool puballtables;
+ bool puballsequences;
bool pubinsert;
bool pubupdate;
bool pubdelete;
bool pubtruncate;
+ bool pubsequence;
bool pubviaroot;
} PublicationInfo;
@@ -651,6 +653,7 @@ typedef struct _PublicationSchemaInfo
DumpableObject dobj;
PublicationInfo *publication;
NamespaceInfo *pubschema;
+ char pubtype;
} PublicationSchemaInfo;
/*
diff --git a/src/bin/pg_dump/t/002_pg_dump.pl b/src/bin/pg_dump/t/002_pg_dump.pl
index eb3ec534b4d..1757dec8c65 100644
--- a/src/bin/pg_dump/t/002_pg_dump.pl
+++ b/src/bin/pg_dump/t/002_pg_dump.pl
@@ -2927,7 +2927,7 @@ my %tests = (
create_order => 50,
create_sql => 'CREATE PUBLICATION pub1;',
regexp => qr/^
- \QCREATE PUBLICATION pub1 WITH (publish = 'insert, update, delete, truncate');\E
+ \QCREATE PUBLICATION pub1 WITH (publish = 'insert, update, delete, truncate, sequence');\E
/xm,
like => { %full_runs, section_post_data => 1, },
},
@@ -2947,7 +2947,7 @@ my %tests = (
create_order => 50,
create_sql => 'CREATE PUBLICATION pub3;',
regexp => qr/^
- \QCREATE PUBLICATION pub3 WITH (publish = 'insert, update, delete, truncate');\E
+ \QCREATE PUBLICATION pub3 WITH (publish = 'insert, update, delete, truncate, sequence');\E
/xm,
like => { %full_runs, section_post_data => 1, },
},
@@ -2956,7 +2956,27 @@ my %tests = (
create_order => 50,
create_sql => 'CREATE PUBLICATION pub4;',
regexp => qr/^
- \QCREATE PUBLICATION pub4 WITH (publish = 'insert, update, delete, truncate');\E
+ \QCREATE PUBLICATION pub4 WITH (publish = 'insert, update, delete, truncate, sequence');\E
+ /xm,
+ like => { %full_runs, section_post_data => 1, },
+ },
+
+ 'CREATE PUBLICATION pub5' => {
+ create_order => 50,
+ create_sql => 'CREATE PUBLICATION pub5
+ FOR ALL SEQUENCES
+ WITH (publish = \'\');',
+ regexp => qr/^
+ \QCREATE PUBLICATION pub5 FOR ALL SEQUENCES WITH (publish = '');\E
+ /xm,
+ like => { %full_runs, section_post_data => 1, },
+ },
+
+ 'CREATE PUBLICATION pub6' => {
+ create_order => 50,
+ create_sql => 'CREATE PUBLICATION pub6;',
+ regexp => qr/^
+ \QCREATE PUBLICATION pub6 WITH (publish = 'insert, update, delete, truncate, sequence');\E
/xm,
like => { %full_runs, section_post_data => 1, },
},
@@ -3102,6 +3122,27 @@ my %tests = (
unlike => { exclude_dump_test_schema => 1, },
},
+ 'ALTER PUBLICATION pub3 ADD SEQUENCES IN SCHEMA dump_test' => {
+ create_order => 51,
+ create_sql =>
+ 'ALTER PUBLICATION pub3 ADD SEQUENCES IN SCHEMA dump_test;',
+ regexp => qr/^
+ \QALTER PUBLICATION pub3 ADD SEQUENCES IN SCHEMA dump_test;\E
+ /xm,
+ like => { %full_runs, section_post_data => 1, },
+ unlike => { exclude_dump_test_schema => 1, },
+ },
+
+ 'ALTER PUBLICATION pub3 ADD SEQUENCES IN SCHEMA public' => {
+ create_order => 52,
+ create_sql =>
+ 'ALTER PUBLICATION pub3 ADD SEQUENCES IN SCHEMA public;',
+ regexp => qr/^
+ \QALTER PUBLICATION pub3 ADD SEQUENCES IN SCHEMA public;\E
+ /xm,
+ like => { %full_runs, section_post_data => 1, },
+ },
+
'CREATE SCHEMA public' => {
regexp => qr/^CREATE SCHEMA public;/m,
diff --git a/src/bin/psql/describe.c b/src/bin/psql/describe.c
index 5077e7b3581..e0dc4dcfccd 100644
--- a/src/bin/psql/describe.c
+++ b/src/bin/psql/describe.c
@@ -1711,28 +1711,19 @@ describeOneTableDetails(const char *schemaname,
if (tableinfo.relkind == RELKIND_SEQUENCE)
{
PGresult *result = NULL;
- printQueryOpt myopt = pset.popt;
- char *footers[2] = {NULL, NULL};
if (pset.sversion >= 100000)
{
printfPQExpBuffer(&buf,
- "SELECT pg_catalog.format_type(seqtypid, NULL) AS \"%s\",\n"
- " seqstart AS \"%s\",\n"
- " seqmin AS \"%s\",\n"
- " seqmax AS \"%s\",\n"
- " seqincrement AS \"%s\",\n"
- " CASE WHEN seqcycle THEN '%s' ELSE '%s' END AS \"%s\",\n"
- " seqcache AS \"%s\"\n",
- gettext_noop("Type"),
- gettext_noop("Start"),
- gettext_noop("Minimum"),
- gettext_noop("Maximum"),
- gettext_noop("Increment"),
+ "SELECT pg_catalog.format_type(seqtypid, NULL),\n"
+ " seqstart,\n"
+ " seqmin,\n"
+ " seqmax,\n"
+ " seqincrement,\n"
+ " CASE WHEN seqcycle THEN '%s' ELSE '%s' END,\n"
+ " seqcache\n",
gettext_noop("yes"),
- gettext_noop("no"),
- gettext_noop("Cycles?"),
- gettext_noop("Cache"));
+ gettext_noop("no"));
appendPQExpBuffer(&buf,
"FROM pg_catalog.pg_sequence\n"
"WHERE seqrelid = '%s';",
@@ -1741,22 +1732,15 @@ describeOneTableDetails(const char *schemaname,
else
{
printfPQExpBuffer(&buf,
- "SELECT 'bigint' AS \"%s\",\n"
- " start_value AS \"%s\",\n"
- " min_value AS \"%s\",\n"
- " max_value AS \"%s\",\n"
- " increment_by AS \"%s\",\n"
- " CASE WHEN is_cycled THEN '%s' ELSE '%s' END AS \"%s\",\n"
- " cache_value AS \"%s\"\n",
- gettext_noop("Type"),
- gettext_noop("Start"),
- gettext_noop("Minimum"),
- gettext_noop("Maximum"),
- gettext_noop("Increment"),
+ "SELECT 'bigint',\n"
+ " start_value,\n"
+ " min_value,\n"
+ " max_value,\n"
+ " increment_by,\n"
+ " CASE WHEN is_cycled THEN '%s' ELSE '%s' END,\n"
+ " cache_value\n",
gettext_noop("yes"),
- gettext_noop("no"),
- gettext_noop("Cycles?"),
- gettext_noop("Cache"));
+ gettext_noop("no"));
appendPQExpBuffer(&buf, "FROM %s", fmtId(schemaname));
/* must be separate because fmtId isn't reentrant */
appendPQExpBuffer(&buf, ".%s;", fmtId(relationname));
@@ -1766,6 +1750,59 @@ describeOneTableDetails(const char *schemaname,
if (!res)
goto error_return;
+ numrows = PQntuples(res);
+
+ /*
+ * XXX reset to use expanded output for sequences (maybe we should
+ * keep this disabled, just like for tables?)
+ */
+ myopt.expanded = pset.popt.topt.expanded;
+
+ printTableInit(&cont, &myopt, title.data, 7, numrows);
+ printTableInitialized = true;
+
+ if (tableinfo.relpersistence == 'u')
+ printfPQExpBuffer(&title, _("Unlogged sequence \"%s.%s\""),
+ schemaname, relationname);
+ else
+ printfPQExpBuffer(&title, _("Sequence \"%s.%s\""),
+ schemaname, relationname);
+
+ printTableAddHeader(&cont, gettext_noop("Type"), true, 'l');
+ printTableAddHeader(&cont, gettext_noop("Start"), true, 'r');
+ printTableAddHeader(&cont, gettext_noop("Minimum"), true, 'r');
+ printTableAddHeader(&cont, gettext_noop("Maximum"), true, 'r');
+ printTableAddHeader(&cont, gettext_noop("Increment"), true, 'r');
+ printTableAddHeader(&cont, gettext_noop("Cycles?"), true, 'l');
+ printTableAddHeader(&cont, gettext_noop("Cache"), true, 'r');
+
+ /* Generate table cells to be printed */
+ for (i = 0; i < numrows; i++)
+ {
+ /* Type */
+ printTableAddCell(&cont, PQgetvalue(res, i, 0), false, false);
+
+ /* Start */
+ printTableAddCell(&cont, PQgetvalue(res, i, 1), false, false);
+
+ /* Minimum */
+ printTableAddCell(&cont, PQgetvalue(res, i, 2), false, false);
+
+ /* Maximum */
+ printTableAddCell(&cont, PQgetvalue(res, i, 3), false, false);
+
+ /* Increment */
+ printTableAddCell(&cont, PQgetvalue(res, i, 4), false, false);
+
+ /* Cycles? */
+ printTableAddCell(&cont, PQgetvalue(res, i, 5), false, false);
+
+ /* Cache */
+ printTableAddCell(&cont, PQgetvalue(res, i, 6), false, false);
+ }
+
+ /* Footer information about a sequence */
+
/* Get the column that owns this sequence */
printfPQExpBuffer(&buf, "SELECT pg_catalog.quote_ident(nspname) || '.' ||"
"\n pg_catalog.quote_ident(relname) || '.' ||"
@@ -1797,32 +1834,63 @@ describeOneTableDetails(const char *schemaname,
switch (PQgetvalue(result, 0, 1)[0])
{
case 'a':
- footers[0] = psprintf(_("Owned by: %s"),
- PQgetvalue(result, 0, 0));
+ printTableAddFooter(&cont,
+ psprintf(_("Owned by: %s"),
+ PQgetvalue(result, 0, 0)));
break;
case 'i':
- footers[0] = psprintf(_("Sequence for identity column: %s"),
- PQgetvalue(result, 0, 0));
+ printTableAddFooter(&cont,
+ psprintf(_("Sequence for identity column: %s"),
+ PQgetvalue(result, 0, 0)));
break;
}
}
PQclear(result);
- if (tableinfo.relpersistence == 'u')
- printfPQExpBuffer(&title, _("Unlogged sequence \"%s.%s\""),
- schemaname, relationname);
- else
- printfPQExpBuffer(&title, _("Sequence \"%s.%s\""),
- schemaname, relationname);
+ /* print any publications */
+ if (pset.sversion >= 170000)
+ {
+ int tuples = 0;
- myopt.footers = footers;
- myopt.topt.default_footer = false;
- myopt.title = title.data;
- myopt.translate_header = true;
+ printfPQExpBuffer(&buf,
+ "SELECT pubname\n"
+ "FROM pg_catalog.pg_publication p\n"
+ " JOIN pg_catalog.pg_publication_namespace pn ON p.oid = pn.pnpubid\n"
+ " JOIN pg_catalog.pg_class pc ON pc.relnamespace = pn.pnnspid\n"
+ "WHERE pc.oid ='%s' and pn.pntype = 's' and pg_catalog.pg_relation_is_publishable('%s')\n"
+ "UNION\n"
+ "SELECT pubname\n"
+ "FROM pg_catalog.pg_publication p\n"
+ " JOIN pg_catalog.pg_publication_rel pr ON p.oid = pr.prpubid\n"
+ "WHERE pr.prrelid = '%s'\n"
+ "UNION\n"
+ "SELECT pubname\n"
+ "FROM pg_catalog.pg_publication p\n"
+ "WHERE p.puballsequences AND pg_catalog.pg_relation_is_publishable('%s')\n"
+ "ORDER BY 1;",
+ oid, oid, oid, oid);
- printQuery(res, &myopt, pset.queryFout, false, pset.logfile);
+ result = PSQLexec(buf.data);
+ if (!result)
+ goto error_return;
+ else
+ tuples = PQntuples(result);
+
+ if (tuples > 0)
+ printTableAddFooter(&cont, _("Publications:"));
+
+ /* Might be an empty set - that's ok */
+ for (i = 0; i < tuples; i++)
+ {
+ printfPQExpBuffer(&buf, " \"%s\"",
+ PQgetvalue(result, i, 0));
+
+ printTableAddFooter(&cont, buf.data);
+ }
+ PQclear(result);
+ }
- free(footers[0]);
+ printTable(&cont, pset.queryFout, false, pset.logfile);
retval = true;
goto error_return; /* not an error, just return early */
@@ -2049,6 +2117,11 @@ describeOneTableDetails(const char *schemaname,
for (i = 0; i < cols; i++)
printTableAddHeader(&cont, headers[i], true, 'l');
+ res = PSQLexec(buf.data);
+ if (!res)
+ goto error_return;
+ numrows = PQntuples(res);
+
/* Generate table cells to be printed */
for (i = 0; i < numrows; i++)
{
@@ -2974,7 +3047,7 @@ describeOneTableDetails(const char *schemaname,
"FROM pg_catalog.pg_publication p\n"
" JOIN pg_catalog.pg_publication_namespace pn ON p.oid = pn.pnpubid\n"
" JOIN pg_catalog.pg_class pc ON pc.relnamespace = pn.pnnspid\n"
- "WHERE pc.oid ='%s' and pg_catalog.pg_relation_is_publishable('%s')\n"
+ "WHERE pc.oid ='%s' and pn.pntype = 't' and pg_catalog.pg_relation_is_publishable('%s')\n"
"UNION\n"
"SELECT pubname\n"
" , pg_get_expr(pr.prqual, c.oid)\n"
@@ -5111,7 +5184,7 @@ listSchemas(const char *pattern, bool verbose, bool showSystem)
int i;
printfPQExpBuffer(&buf,
- "SELECT pubname \n"
+ "SELECT pubname, (CASE WHEN pntype = 't' THEN 'tables' ELSE 'sequences' END) AS pubtype\n"
"FROM pg_catalog.pg_publication p\n"
" JOIN pg_catalog.pg_publication_namespace pn ON p.oid = pn.pnpubid\n"
" JOIN pg_catalog.pg_namespace n ON n.oid = pn.pnnspid \n"
@@ -5137,8 +5210,9 @@ listSchemas(const char *pattern, bool verbose, bool showSystem)
/* Might be an empty set - that's ok */
for (i = 0; i < pub_schema_tuples; i++)
{
- printfPQExpBuffer(&buf, " \"%s\"",
- PQgetvalue(result, i, 0));
+ printfPQExpBuffer(&buf, " \"%s\" (%s)",
+ PQgetvalue(result, i, 0),
+ PQgetvalue(result, i, 1));
footers[i + 1] = pg_strdup(buf.data);
}
@@ -6254,7 +6328,7 @@ listPublications(const char *pattern)
PQExpBufferData buf;
PGresult *res;
printQueryOpt myopt = pset.popt;
- static const bool translate_columns[] = {false, false, false, false, false, false, false, false};
+ static const bool translate_columns[] = {false, false, false, false, false, false, false, false, false, false};
if (pset.sversion < 100000)
{
@@ -6268,23 +6342,45 @@ listPublications(const char *pattern)
initPQExpBuffer(&buf);
- printfPQExpBuffer(&buf,
- "SELECT pubname AS \"%s\",\n"
- " pg_catalog.pg_get_userbyid(pubowner) AS \"%s\",\n"
- " puballtables AS \"%s\",\n"
- " pubinsert AS \"%s\",\n"
- " pubupdate AS \"%s\",\n"
- " pubdelete AS \"%s\"",
- gettext_noop("Name"),
- gettext_noop("Owner"),
- gettext_noop("All tables"),
- gettext_noop("Inserts"),
- gettext_noop("Updates"),
- gettext_noop("Deletes"));
+ if (pset.sversion >= 170000)
+ printfPQExpBuffer(&buf,
+ "SELECT pubname AS \"%s\",\n"
+ " pg_catalog.pg_get_userbyid(pubowner) AS \"%s\",\n"
+ " puballtables AS \"%s\",\n"
+ " puballsequences AS \"%s\",\n"
+ " pubinsert AS \"%s\",\n"
+ " pubupdate AS \"%s\",\n"
+ " pubdelete AS \"%s\"",
+ gettext_noop("Name"),
+ gettext_noop("Owner"),
+ gettext_noop("All tables"),
+ gettext_noop("All sequences"),
+ gettext_noop("Inserts"),
+ gettext_noop("Updates"),
+ gettext_noop("Deletes"));
+ else
+ printfPQExpBuffer(&buf,
+ "SELECT pubname AS \"%s\",\n"
+ " pg_catalog.pg_get_userbyid(pubowner) AS \"%s\",\n"
+ " puballtables AS \"%s\",\n"
+ " pubinsert AS \"%s\",\n"
+ " pubupdate AS \"%s\",\n"
+ " pubdelete AS \"%s\"",
+ gettext_noop("Name"),
+ gettext_noop("Owner"),
+ gettext_noop("All tables"),
+ gettext_noop("Inserts"),
+ gettext_noop("Updates"),
+ gettext_noop("Deletes"));
+
if (pset.sversion >= 110000)
appendPQExpBuffer(&buf,
",\n pubtruncate AS \"%s\"",
gettext_noop("Truncates"));
+ if (pset.sversion >= 170000)
+ appendPQExpBuffer(&buf,
+ ",\n pubsequence AS \"%s\"",
+ gettext_noop("Sequences"));
if (pset.sversion >= 130000)
appendPQExpBuffer(&buf,
",\n pubviaroot AS \"%s\"",
@@ -6378,6 +6474,7 @@ describePublications(const char *pattern)
PGresult *res;
bool has_pubtruncate;
bool has_pubviaroot;
+ bool has_pubsequence;
PQExpBufferData title;
printTableContent cont;
@@ -6394,6 +6491,7 @@ describePublications(const char *pattern)
has_pubtruncate = (pset.sversion >= 110000);
has_pubviaroot = (pset.sversion >= 130000);
+ has_pubsequence = (pset.sversion >= 170000);
initPQExpBuffer(&buf);
@@ -6407,6 +6505,10 @@ describePublications(const char *pattern)
if (has_pubviaroot)
appendPQExpBufferStr(&buf,
", pubviaroot");
+ if (has_pubsequence)
+ appendPQExpBufferStr(&buf,
+ ", puballsequences, pubsequence");
+
appendPQExpBufferStr(&buf,
"\nFROM pg_catalog.pg_publication\n");
@@ -6452,6 +6554,7 @@ describePublications(const char *pattern)
char *pubid = PQgetvalue(res, i, 0);
char *pubname = PQgetvalue(res, i, 1);
bool puballtables = strcmp(PQgetvalue(res, i, 3), "t") == 0;
+ bool puballsequences = strcmp(PQgetvalue(res, i, 9), "t") == 0;
printTableOpt myopt = pset.popt.topt;
if (has_pubtruncate)
@@ -6459,29 +6562,43 @@ describePublications(const char *pattern)
if (has_pubviaroot)
ncols++;
+ /* sequences have two extra columns (puballsequences, pubsequences) */
+ if (has_pubsequence)
+ ncols += 2;
+
initPQExpBuffer(&title);
printfPQExpBuffer(&title, _("Publication %s"), pubname);
printTableInit(&cont, &myopt, title.data, ncols, nrows);
printTableAddHeader(&cont, gettext_noop("Owner"), true, align);
printTableAddHeader(&cont, gettext_noop("All tables"), true, align);
+ if (has_pubsequence)
+ printTableAddHeader(&cont, gettext_noop("All sequences"), true, align);
printTableAddHeader(&cont, gettext_noop("Inserts"), true, align);
printTableAddHeader(&cont, gettext_noop("Updates"), true, align);
printTableAddHeader(&cont, gettext_noop("Deletes"), true, align);
if (has_pubtruncate)
printTableAddHeader(&cont, gettext_noop("Truncates"), true, align);
+ if (has_pubsequence)
+ printTableAddHeader(&cont, gettext_noop("Sequences"), true, align);
if (has_pubviaroot)
printTableAddHeader(&cont, gettext_noop("Via root"), true, align);
- printTableAddCell(&cont, PQgetvalue(res, i, 2), false, false);
- printTableAddCell(&cont, PQgetvalue(res, i, 3), false, false);
- printTableAddCell(&cont, PQgetvalue(res, i, 4), false, false);
- printTableAddCell(&cont, PQgetvalue(res, i, 5), false, false);
- printTableAddCell(&cont, PQgetvalue(res, i, 6), false, false);
+ printTableAddCell(&cont, PQgetvalue(res, i, 2), false, false); /* owner */
+ printTableAddCell(&cont, PQgetvalue(res, i, 3), false, false); /* all tables */
+
+ if (has_pubsequence)
+ printTableAddCell(&cont, PQgetvalue(res, i, 9), false, false); /* all sequences */
+
+ printTableAddCell(&cont, PQgetvalue(res, i, 4), false, false); /* insert */
+ printTableAddCell(&cont, PQgetvalue(res, i, 5), false, false); /* update */
+ printTableAddCell(&cont, PQgetvalue(res, i, 6), false, false); /* delete */
if (has_pubtruncate)
- printTableAddCell(&cont, PQgetvalue(res, i, 7), false, false);
+ printTableAddCell(&cont, PQgetvalue(res, i, 7), false, false); /* truncate */
+ if (has_pubsequence)
+ printTableAddCell(&cont, PQgetvalue(res, i, 10), false, false); /* sequence */
if (has_pubviaroot)
- printTableAddCell(&cont, PQgetvalue(res, i, 8), false, false);
+ printTableAddCell(&cont, PQgetvalue(res, i, 8), false, false); /* via root */
if (!puballtables)
{
@@ -6512,6 +6629,7 @@ describePublications(const char *pattern)
"WHERE c.relnamespace = n.oid\n"
" AND c.oid = pr.prrelid\n"
" AND pr.prpubid = '%s'\n"
+ " AND c.relkind != 'S'\n" /* exclude sequences */
"ORDER BY 1,2", pubid);
if (!addFooterToPublicationDesc(&buf, _("Tables:"), false, &cont))
goto error_return;
@@ -6523,7 +6641,7 @@ describePublications(const char *pattern)
"SELECT n.nspname\n"
"FROM pg_catalog.pg_namespace n\n"
" JOIN pg_catalog.pg_publication_namespace pn ON n.oid = pn.pnnspid\n"
- "WHERE pn.pnpubid = '%s'\n"
+ "WHERE pn.pnpubid = '%s' AND pn.pntype = 't'\n"
"ORDER BY 1", pubid);
if (!addFooterToPublicationDesc(&buf, _("Tables from schemas:"),
true, &cont))
@@ -6531,6 +6649,37 @@ describePublications(const char *pattern)
}
}
+ if (!puballsequences)
+ {
+ /* Get the sequences for the specified publication */
+ printfPQExpBuffer(&buf,
+ "SELECT n.nspname, c.relname, NULL, NULL\n"
+ "FROM pg_catalog.pg_class c,\n"
+ " pg_catalog.pg_namespace n,\n"
+ " pg_catalog.pg_publication_rel pr\n"
+ "WHERE c.relnamespace = n.oid\n"
+ " AND c.oid = pr.prrelid\n"
+ " AND pr.prpubid = '%s'\n"
+ " AND c.relkind = 'S'\n" /* only sequences */
+ "ORDER BY 1,2", pubid);
+ if (!addFooterToPublicationDesc(&buf, "Sequences:", false, &cont))
+ goto error_return;
+
+ if (pset.sversion >= 150000)
+ {
+ /* Get the schemas for the specified publication */
+ printfPQExpBuffer(&buf,
+ "SELECT n.nspname\n"
+ "FROM pg_catalog.pg_namespace n\n"
+ " JOIN pg_catalog.pg_publication_namespace pn ON n.oid = pn.pnnspid\n"
+ "WHERE pn.pnpubid = '%s' AND pn.pntype = 's'\n"
+ "ORDER BY 1", pubid);
+ if (!addFooterToPublicationDesc(&buf, "Sequences from schemas:",
+ true, &cont))
+ goto error_return;
+ }
+ }
+
printTable(&cont, pset.queryFout, false, pset.logfile);
printTableCleanup(&cont);
diff --git a/src/bin/psql/tab-complete.c b/src/bin/psql/tab-complete.c
index 049801186c3..b0545e9e917 100644
--- a/src/bin/psql/tab-complete.c
+++ b/src/bin/psql/tab-complete.c
@@ -1890,11 +1890,15 @@ psql_completion(const char *text, int start, int end)
COMPLETE_WITH("ADD", "DROP", "OWNER TO", "RENAME TO", "SET");
/* ALTER PUBLICATION <name> ADD */
else if (Matches("ALTER", "PUBLICATION", MatchAny, "ADD"))
- COMPLETE_WITH("TABLES IN SCHEMA", "TABLE");
+ COMPLETE_WITH("TABLES IN SCHEMA", "TABLE", "SEQUENCES IN SCHEMA", "SEQUENCE");
else if (Matches("ALTER", "PUBLICATION", MatchAny, "ADD|SET", "TABLE") ||
(HeadMatches("ALTER", "PUBLICATION", MatchAny, "ADD|SET", "TABLE") &&
ends_with(prev_wd, ',')))
COMPLETE_WITH_SCHEMA_QUERY(Query_for_list_of_tables);
+ else if (Matches("ALTER", "PUBLICATION", MatchAny, "ADD|SET", "SEQUENCE") ||
+ (HeadMatches("ALTER", "PUBLICATION", MatchAny, "ADD|SET", "SEQUENCE") &&
+ ends_with(prev_wd, ',')))
+ COMPLETE_WITH_SCHEMA_QUERY(Query_for_list_of_sequences);
/*
* "ALTER PUBLICATION <name> SET TABLE <name> WHERE (" - complete with
@@ -1914,11 +1918,11 @@ psql_completion(const char *text, int start, int end)
COMPLETE_WITH(",");
/* ALTER PUBLICATION <name> DROP */
else if (Matches("ALTER", "PUBLICATION", MatchAny, "DROP"))
- COMPLETE_WITH("TABLES IN SCHEMA", "TABLE");
+ COMPLETE_WITH("TABLES IN SCHEMA", "TABLE", "SEQUENCES IN SCHEMA", "SEQUENCE");
/* ALTER PUBLICATION <name> SET */
else if (Matches("ALTER", "PUBLICATION", MatchAny, "SET"))
- COMPLETE_WITH("(", "TABLES IN SCHEMA", "TABLE");
- else if (Matches("ALTER", "PUBLICATION", MatchAny, "ADD|DROP|SET", "TABLES", "IN", "SCHEMA"))
+ COMPLETE_WITH("(", "TABLES IN SCHEMA", "TABLE", "SEQUENCES IN SCHEMA", "SEQUENCE");
+ else if (Matches("ALTER", "PUBLICATION", MatchAny, "ADD|DROP|SET", "TABLES|SEQUENCES", "IN", "SCHEMA"))
COMPLETE_WITH_QUERY_PLUS(Query_for_list_of_schemas
" AND nspname NOT LIKE E'pg\\\\_%%'",
"CURRENT_SCHEMA");
@@ -3123,24 +3127,31 @@ psql_completion(const char *text, int start, int end)
/* CREATE PUBLICATION */
else if (Matches("CREATE", "PUBLICATION", MatchAny))
- COMPLETE_WITH("FOR TABLE", "FOR ALL TABLES", "FOR TABLES IN SCHEMA", "WITH (");
+ COMPLETE_WITH("FOR TABLE", "FOR ALL TABLES", "FOR TABLES IN SCHEMA", "FOR SEQUENCE", "FOR ALL SEQUENCES", "FOR SEQUENCES IN SCHEMA", "WITH (");
else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR"))
- COMPLETE_WITH("TABLE", "ALL TABLES", "TABLES IN SCHEMA");
+ COMPLETE_WITH("TABLE", "ALL TABLES", "TABLES IN SCHEMA", "SEQUENCE", "ALL SEQUENCES", "SEQUENCES IN SCHEMA");
else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "ALL"))
- COMPLETE_WITH("TABLES");
- else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "ALL", "TABLES"))
+ COMPLETE_WITH("TABLES", "SEQUENCES");
+ else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "ALL", "TABLES|SEQUENCES"))
COMPLETE_WITH("WITH (");
- else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "TABLES"))
+ else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "TABLES|SEQUENCES"))
COMPLETE_WITH("IN SCHEMA");
- else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "TABLE", MatchAny) && !ends_with(prev_wd, ','))
+ else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "TABLE|SEQUENCE", MatchAny) && !ends_with(prev_wd, ','))
COMPLETE_WITH("WHERE (", "WITH (");
/* Complete "CREATE PUBLICATION <name> FOR TABLE" with "<table>, ..." */
else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "TABLE"))
COMPLETE_WITH_SCHEMA_QUERY(Query_for_list_of_tables);
/*
- * "CREATE PUBLICATION <name> FOR TABLE <name> WHERE (" - complete with
- * table attributes
+ * Complete "CREATE PUBLICATION <name> FOR SEQUENCE" with "<sequence>,
+ * ..."
+ */
+ else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "SEQUENCE"))
+ COMPLETE_WITH_SCHEMA_QUERY(Query_for_list_of_sequences);
+
+ /*
+ * "CREATE PUBLICATION <name> FOR TABLE|SEQUENCE <name> WHERE (" -
+ * complete with table attributes
*/
else if (HeadMatches("CREATE", "PUBLICATION", MatchAny) && TailMatches("WHERE"))
COMPLETE_WITH("(");
@@ -3152,11 +3163,11 @@ psql_completion(const char *text, int start, int end)
/*
* Complete "CREATE PUBLICATION <name> FOR TABLES IN SCHEMA <schema>, ..."
*/
- else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "TABLES", "IN", "SCHEMA"))
+ else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "TABLES|SEQUENCES", "IN", "SCHEMA"))
COMPLETE_WITH_QUERY_PLUS(Query_for_list_of_schemas
" AND nspname NOT LIKE E'pg\\\\_%%'",
"CURRENT_SCHEMA");
- else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "TABLES", "IN", "SCHEMA", MatchAny) && (!ends_with(prev_wd, ',')))
+ else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "TABLES|SEQUENCES", "IN", "SCHEMA", MatchAny) && (!ends_with(prev_wd, ',')))
COMPLETE_WITH("WITH (");
/* Complete "CREATE PUBLICATION <name> [...] WITH" */
else if (HeadMatches("CREATE", "PUBLICATION") && TailMatches("WITH", "("))
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index fb58dee3bcd..f4ae0bd4fb1 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -3329,6 +3329,14 @@
proname => 'pg_sequence_last_value', provolatile => 'v', proparallel => 'u',
prorettype => 'int8', proargtypes => 'regclass',
prosrc => 'pg_sequence_last_value' },
+{ oid => '6312',
+ descr => 'current on-disk sequence state',
+ proname => 'pg_sequence_state', provolatile => 'v',
+ prorettype => 'record', proargtypes => 'regclass',
+ proallargtypes => '{regclass,pg_lsn,int8,int8,bool}',
+ proargmodes => '{i,o,o,o,o}',
+ proargnames => '{seq_oid,page_lsn,last_value,log_cnt,is_called}',
+ prosrc => 'pg_sequence_state' },
{ oid => '275', descr => 'return the next oid for a system table',
proname => 'pg_nextoid', provolatile => 'v', proparallel => 'u',
@@ -11873,6 +11881,11 @@
proargmodes => '{v,o,o,o,o}',
proargnames => '{pubname,pubid,relid,attrs,qual}',
prosrc => 'pg_get_publication_tables' },
+{ oid => '8000', descr => 'get OIDs of sequences in a publication',
+ proname => 'pg_get_publication_sequences', prorows => '1000', proretset => 't',
+ provolatile => 's', prorettype => 'oid', proargtypes => 'text',
+ proallargtypes => '{text,oid}', proargmodes => '{i,o}',
+ proargnames => '{pubname,relid}', prosrc => 'pg_get_publication_sequences' },
{ oid => '6121',
descr => 'returns whether a relation can be part of a publication',
proname => 'pg_relation_is_publishable', provolatile => 's',
diff --git a/src/include/catalog/pg_publication.h b/src/include/catalog/pg_publication.h
index d929c9a1880..a26e4e51ca2 100644
--- a/src/include/catalog/pg_publication.h
+++ b/src/include/catalog/pg_publication.h
@@ -40,6 +40,12 @@ CATALOG(pg_publication,6104,PublicationRelationId)
*/
bool puballtables;
+ /*
+ * indicates that this is special publication which should encompass all
+ * sequences in the database (except for the unlogged and temp ones)
+ */
+ bool puballsequences;
+
/* true if inserts are published */
bool pubinsert;
@@ -52,6 +58,9 @@ CATALOG(pg_publication,6104,PublicationRelationId)
/* true if truncates are published */
bool pubtruncate;
+ /* true if sequences are published */
+ bool pubsequence;
+
/* true if partition changes are published using root schema */
bool pubviaroot;
} FormData_pg_publication;
@@ -72,6 +81,7 @@ typedef struct PublicationActions
bool pubupdate;
bool pubdelete;
bool pubtruncate;
+ bool pubsequence;
} PublicationActions;
typedef struct PublicationDesc
@@ -99,6 +109,7 @@ typedef struct Publication
Oid oid;
char *name;
bool alltables;
+ bool allsequences;
bool pubviaroot;
PublicationActions pubactions;
} Publication;
@@ -130,14 +141,15 @@ typedef enum PublicationPartOpt
PUBLICATION_PART_ALL,
} PublicationPartOpt;
-extern List *GetPublicationRelations(Oid pubid, PublicationPartOpt pub_partopt);
+extern List *GetPublicationRelations(Oid pubid, char objectType,
+ PublicationPartOpt pub_partopt);
extern List *GetAllTablesPublications(void);
extern List *GetAllTablesPublicationRelations(bool pubviaroot);
-extern List *GetPublicationSchemas(Oid pubid);
-extern List *GetSchemaPublications(Oid schemaid);
-extern List *GetSchemaPublicationRelations(Oid schemaid,
+extern List *GetPublicationSchemas(Oid pubid, char objectType);
+extern List *GetSchemaPublications(Oid schemaid, char objectType);
+extern List *GetSchemaPublicationRelations(Oid schemaid, char objectType,
PublicationPartOpt pub_partopt);
-extern List *GetAllSchemaPublicationRelations(Oid pubid,
+extern List *GetAllSchemaPublicationRelations(Oid puboid, char objectType,
PublicationPartOpt pub_partopt);
extern List *GetPubPartitionOptionRelations(List *result,
PublicationPartOpt pub_partopt,
@@ -145,11 +157,15 @@ extern List *GetPubPartitionOptionRelations(List *result,
extern Oid GetTopMostAncestorInPublication(Oid puboid, List *ancestors,
int *ancestor_level);
+extern List *GetAllSequencesPublications(void);
+extern List *GetAllSequencesPublicationRelations(void);
+
extern bool is_publishable_relation(Relation rel);
extern bool is_schema_publication(Oid pubid);
extern ObjectAddress publication_add_relation(Oid pubid, PublicationRelInfo *pri,
bool if_not_exists);
extern ObjectAddress publication_add_schema(Oid pubid, Oid schemaid,
+ char objectType,
bool if_not_exists);
extern Bitmapset *pub_collist_to_bitmapset(Bitmapset *columns, Datum pubcols,
diff --git a/src/include/catalog/pg_publication_namespace.h b/src/include/catalog/pg_publication_namespace.h
index b361c8ab5e0..f58e670ba7b 100644
--- a/src/include/catalog/pg_publication_namespace.h
+++ b/src/include/catalog/pg_publication_namespace.h
@@ -32,6 +32,7 @@ CATALOG(pg_publication_namespace,6237,PublicationNamespaceRelationId)
Oid oid; /* oid */
Oid pnpubid BKI_LOOKUP(pg_publication); /* Oid of the publication */
Oid pnnspid BKI_LOOKUP(pg_namespace); /* Oid of the schema */
+ char pntype; /* object type to include */
} FormData_pg_publication_namespace;
/* ----------------
@@ -42,6 +43,13 @@ CATALOG(pg_publication_namespace,6237,PublicationNamespaceRelationId)
typedef FormData_pg_publication_namespace *Form_pg_publication_namespace;
DECLARE_UNIQUE_INDEX_PKEY(pg_publication_namespace_oid_index, 6238, PublicationNamespaceObjectIndexId, pg_publication_namespace, btree(oid oid_ops));
-DECLARE_UNIQUE_INDEX(pg_publication_namespace_pnnspid_pnpubid_index, 6239, PublicationNamespacePnnspidPnpubidIndexId, pg_publication_namespace, btree(pnnspid oid_ops, pnpubid oid_ops));
+DECLARE_UNIQUE_INDEX(pg_publication_namespace_pnnspid_pnpubid_pntype_index, 8903, PublicationNamespacePnnspidPnpubidPntypeIndexId, pg_publication_namespace, btree(pnnspid oid_ops, pnpubid oid_ops, pntype char_ops));
+
+/* object type to include from a schema, maps to relkind */
+#define PUB_OBJTYPE_TABLE 't' /* table (regular or partitioned) */
+#define PUB_OBJTYPE_SEQUENCE 's' /* sequence object */
+#define PUB_OBJTYPE_UNSUPPORTED 'u' /* used for non-replicated types */
+
+extern char pub_get_object_type_for_relkind(char relkind);
#endif /* PG_PUBLICATION_NAMESPACE_H */
diff --git a/src/include/commands/sequence.h b/src/include/commands/sequence.h
index 7db7b3da7bc..fe3c813d993 100644
--- a/src/include/commands/sequence.h
+++ b/src/include/commands/sequence.h
@@ -60,6 +60,7 @@ extern ObjectAddress AlterSequence(ParseState *pstate, AlterSeqStmt *stmt);
extern void SequenceChangePersistence(Oid relid, char newrelpersistence);
extern void DeleteSequenceTuple(Oid relid);
extern void ResetSequence(Oid seq_relid);
+extern void SetSequence(Oid seq_relid, bool transactional, int64 value);
extern void ResetSequenceCaches(void);
extern void seq_redo(XLogReaderState *record);
diff --git a/src/include/nodes/parsenodes.h b/src/include/nodes/parsenodes.h
index e494309da8d..356f8724ec9 100644
--- a/src/include/nodes/parsenodes.h
+++ b/src/include/nodes/parsenodes.h
@@ -3990,6 +3990,10 @@ typedef enum PublicationObjSpecType
PUBLICATIONOBJ_TABLES_IN_SCHEMA, /* All tables in schema */
PUBLICATIONOBJ_TABLES_IN_CUR_SCHEMA, /* All tables in first element of
* search_path */
+ PUBLICATIONOBJ_SEQUENCE, /* Sequence type */
+ PUBLICATIONOBJ_SEQUENCES_IN_SCHEMA, /* Sequences in schema type */
+ PUBLICATIONOBJ_SEQUENCES_IN_CUR_SCHEMA, /* Get the first element of
+ * search_path */
PUBLICATIONOBJ_CONTINUATION, /* Continuation of previous type */
} PublicationObjSpecType;
@@ -4008,7 +4012,8 @@ typedef struct CreatePublicationStmt
char *pubname; /* Name of the publication */
List *options; /* List of DefElem nodes */
List *pubobjects; /* Optional list of publication objects */
- bool for_all_tables; /* Special publication for all tables in db */
+ List *for_all_objects; /* Special publication for all objects in
+ * db */
} CreatePublicationStmt;
typedef enum AlterPublicationAction
@@ -4031,7 +4036,8 @@ typedef struct AlterPublicationStmt
* objects.
*/
List *pubobjects; /* Optional list of publication objects */
- bool for_all_tables; /* Special publication for all tables in db */
+ List *for_all_objects; /* Special publication for all objects in
+ * db */
AlterPublicationAction action; /* What action to perform with the given
* objects */
} AlterPublicationStmt;
diff --git a/src/include/replication/logicalproto.h b/src/include/replication/logicalproto.h
index 4bb04c4eac8..f6bbb74d780 100644
--- a/src/include/replication/logicalproto.h
+++ b/src/include/replication/logicalproto.h
@@ -42,7 +42,8 @@
#define LOGICALREP_PROTO_STREAM_VERSION_NUM 2
#define LOGICALREP_PROTO_TWOPHASE_VERSION_NUM 3
#define LOGICALREP_PROTO_STREAM_PARALLEL_VERSION_NUM 4
-#define LOGICALREP_PROTO_MAX_VERSION_NUM LOGICALREP_PROTO_STREAM_PARALLEL_VERSION_NUM
+#define LOGICALREP_PROTO_SEQUENCES_VERSION_NUM 5
+#define LOGICALREP_PROTO_MAX_VERSION_NUM LOGICALREP_PROTO_SEQUENCES_VERSION_NUM
/*
* Logical message types
@@ -66,6 +67,7 @@ typedef enum LogicalRepMsgType
LOGICAL_REP_MSG_RELATION = 'R',
LOGICAL_REP_MSG_TYPE = 'Y',
LOGICAL_REP_MSG_MESSAGE = 'M',
+ LOGICAL_REP_MSG_SEQUENCE = 'Q',
LOGICAL_REP_MSG_BEGIN_PREPARE = 'b',
LOGICAL_REP_MSG_PREPARE = 'P',
LOGICAL_REP_MSG_COMMIT_PREPARED = 'K',
@@ -123,6 +125,17 @@ typedef struct LogicalRepTyp
char *typname; /* name of the remote type */
} LogicalRepTyp;
+/* Sequence info */
+typedef struct LogicalRepSequence
+{
+ XLogRecPtr lsn; /* LSN of the sequence change */
+ Oid remoteid; /* unique id of the remote sequence */
+ char *nspname; /* schema name of remote sequence */
+ char *seqname; /* name of the remote sequence */
+ bool transactional;
+ int64 value;
+} LogicalRepSequence;
+
/* Transaction info */
typedef struct LogicalRepBeginData
{
@@ -246,6 +259,10 @@ extern List *logicalrep_read_truncate(StringInfo in,
bool *cascade, bool *restart_seqs);
extern void logicalrep_write_message(StringInfo out, TransactionId xid, XLogRecPtr lsn,
bool transactional, const char *prefix, Size sz, const char *message);
+extern void logicalrep_write_sequence(StringInfo out, Relation rel,
+ TransactionId xid, XLogRecPtr lsn,
+ bool transactional, int64 value);
+extern LogicalRepRelId logicalrep_read_sequence(StringInfo in, LogicalRepSequence *seqdata);
extern void logicalrep_write_rel(StringInfo out, TransactionId xid,
Relation rel, Bitmapset *columns);
extern LogicalRepRelation *logicalrep_read_rel(StringInfo in);
diff --git a/src/test/regress/expected/object_address.out b/src/test/regress/expected/object_address.out
index fc42d418bf1..9aec2612d01 100644
--- a/src/test/regress/expected/object_address.out
+++ b/src/test/regress/expected/object_address.out
@@ -47,6 +47,7 @@ CREATE TRANSFORM FOR int LANGUAGE SQL (
SET client_min_messages = 'ERROR';
CREATE PUBLICATION addr_pub FOR TABLE addr_nsp.gentable;
CREATE PUBLICATION addr_pub_schema FOR TABLES IN SCHEMA addr_nsp;
+CREATE PUBLICATION addr_pub_schema2 FOR SEQUENCES IN SCHEMA addr_nsp;
RESET client_min_messages;
CREATE SUBSCRIPTION regress_addr_sub CONNECTION '' PUBLICATION bar WITH (connect = false, slot_name = NONE);
WARNING: subscription was created, but is not connected
@@ -315,12 +316,12 @@ WARNING: error for function of access method,{addr_nsp,zwei},{}: name list leng
WARNING: error for function of access method,{addr_nsp,zwei},{integer}: name list length must be at least 3
WARNING: error for function of access method,{eins,zwei,drei},{}: argument list length must be exactly 2
WARNING: error for function of access method,{eins,zwei,drei},{integer}: argument list length must be exactly 2
-WARNING: error for publication namespace,{eins},{}: argument list length must be exactly 1
-WARNING: error for publication namespace,{eins},{integer}: schema "eins" does not exist
-WARNING: error for publication namespace,{addr_nsp,zwei},{}: name list length must be exactly 1
-WARNING: error for publication namespace,{addr_nsp,zwei},{integer}: name list length must be exactly 1
-WARNING: error for publication namespace,{eins,zwei,drei},{}: name list length must be exactly 1
-WARNING: error for publication namespace,{eins,zwei,drei},{integer}: name list length must be exactly 1
+WARNING: error for publication namespace,{eins},{}: argument list length must be exactly 2
+WARNING: error for publication namespace,{eins},{integer}: argument list length must be exactly 2
+WARNING: error for publication namespace,{addr_nsp,zwei},{}: argument list length must be exactly 2
+WARNING: error for publication namespace,{addr_nsp,zwei},{integer}: argument list length must be exactly 2
+WARNING: error for publication namespace,{eins,zwei,drei},{}: argument list length must be exactly 2
+WARNING: error for publication namespace,{eins,zwei,drei},{integer}: argument list length must be exactly 2
WARNING: error for publication relation,{eins},{}: argument list length must be exactly 1
WARNING: error for publication relation,{eins},{integer}: relation "eins" does not exist
WARNING: error for publication relation,{addr_nsp,zwei},{}: argument list length must be exactly 1
@@ -441,7 +442,8 @@ WITH objects (type, name, args) AS (VALUES
('transform', '{int}', '{sql}'),
('access method', '{btree}', '{}'),
('publication', '{addr_pub}', '{}'),
- ('publication namespace', '{addr_nsp}', '{addr_pub_schema}'),
+ ('publication namespace', '{addr_nsp}', '{addr_pub_schema, t}'),
+ ('publication namespace', '{addr_nsp}', '{addr_pub_schema2, s}'),
('publication relation', '{addr_nsp, gentable}', '{addr_pub}'),
('subscription', '{regress_addr_sub}', '{}'),
('statistics object', '{addr_nsp, gentable_stat}', '{}')
@@ -504,7 +506,8 @@ text search template|addr_nsp|addr_ts_temp|addr_nsp.addr_ts_temp|t
subscription|NULL|regress_addr_sub|regress_addr_sub|t
publication|NULL|addr_pub|addr_pub|t
publication relation|NULL|NULL|addr_nsp.gentable in publication addr_pub|t
-publication namespace|NULL|NULL|addr_nsp in publication addr_pub_schema|t
+publication namespace|NULL|NULL|addr_nsp in publication addr_pub_schema type t|t
+publication namespace|NULL|NULL|addr_nsp in publication addr_pub_schema2 type s|t
---
--- Cleanup resources
---
@@ -516,6 +519,7 @@ drop cascades to server integer
drop cascades to user mapping for regress_addr_user on server integer
DROP PUBLICATION addr_pub;
DROP PUBLICATION addr_pub_schema;
+DROP PUBLICATION addr_pub_schema2;
DROP SUBSCRIPTION regress_addr_sub;
DROP SCHEMA addr_nsp CASCADE;
NOTICE: drop cascades to 14 other objects
diff --git a/src/test/regress/expected/psql.out b/src/test/regress/expected/psql.out
index 13e4f6db7ba..bab5c27422c 100644
--- a/src/test/regress/expected/psql.out
+++ b/src/test/regress/expected/psql.out
@@ -6240,9 +6240,9 @@ List of schemas
(0 rows)
\dRp "no.such.publication"
- List of publications
- Name | Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
-------+-------+------------+---------+---------+---------+-----------+----------
+ List of publications
+ Name | Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+------+-------+------------+---------------+---------+---------+---------+-----------+-----------+----------
(0 rows)
\dRs "no.such.subscription"
diff --git a/src/test/regress/expected/publication.out b/src/test/regress/expected/publication.out
index 16361a91f9f..2f5c72c9d2b 100644
--- a/src/test/regress/expected/publication.out
+++ b/src/test/regress/expected/publication.out
@@ -30,20 +30,20 @@ ERROR: conflicting or redundant options
LINE 1: ...ub_xxx WITH (publish_via_partition_root = 'true', publish_vi...
^
\dRp
- List of publications
- Name | Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------+--------------------------+------------+---------+---------+---------+-----------+----------
- testpib_ins_trunct | regress_publication_user | f | t | f | f | f | f
- testpub_default | regress_publication_user | f | f | t | f | f | f
+ List of publications
+ Name | Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ testpib_ins_trunct | regress_publication_user | f | f | t | f | f | f | f | f
+ testpub_default | regress_publication_user | f | f | f | t | f | f | f | f
(2 rows)
-ALTER PUBLICATION testpub_default SET (publish = 'insert, update, delete');
+ALTER PUBLICATION testpub_default SET (publish = 'insert, update, delete, sequence');
\dRp
- List of publications
- Name | Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------+--------------------------+------------+---------+---------+---------+-----------+----------
- testpib_ins_trunct | regress_publication_user | f | t | f | f | f | f
- testpub_default | regress_publication_user | f | t | t | t | f | f
+ List of publications
+ Name | Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ testpib_ins_trunct | regress_publication_user | f | f | t | f | f | f | f | f
+ testpub_default | regress_publication_user | f | f | t | t | t | f | t | f
(2 rows)
--- adding tables
@@ -61,6 +61,9 @@ CREATE TABLE testpub_tbl2 (id serial primary key, data text);
ALTER PUBLICATION testpub_foralltables ADD TABLE testpub_tbl2;
ERROR: publication "testpub_foralltables" is defined as FOR ALL TABLES
DETAIL: Tables cannot be added to or dropped from FOR ALL TABLES publications.
+-- fail - can't add a table using ADD SEQUENCE command
+ALTER PUBLICATION testpub_foralltables ADD SEQUENCE testpub_tbl2;
+ERROR: object type does not match type expected by command
-- fail - can't drop from all tables publication
ALTER PUBLICATION testpub_foralltables DROP TABLE testpub_tbl2;
ERROR: publication "testpub_foralltables" is defined as FOR ALL TABLES
@@ -87,10 +90,10 @@ RESET client_min_messages;
-- should be able to add schema to 'FOR TABLE' publication
ALTER PUBLICATION testpub_fortable ADD TABLES IN SCHEMA pub_test;
\dRp+ testpub_fortable
- Publication testpub_fortable
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_fortable
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"public.testpub_tbl1"
Tables from schemas:
@@ -99,20 +102,20 @@ Tables from schemas:
-- should be able to drop schema from 'FOR TABLE' publication
ALTER PUBLICATION testpub_fortable DROP TABLES IN SCHEMA pub_test;
\dRp+ testpub_fortable
- Publication testpub_fortable
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_fortable
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"public.testpub_tbl1"
-- should be able to set schema to 'FOR TABLE' publication
ALTER PUBLICATION testpub_fortable SET TABLES IN SCHEMA pub_test;
\dRp+ testpub_fortable
- Publication testpub_fortable
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_fortable
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test"
@@ -123,10 +126,10 @@ CREATE PUBLICATION testpub_forschema FOR TABLES IN SCHEMA pub_test;
CREATE PUBLICATION testpub_for_tbl_schema FOR TABLES IN SCHEMA pub_test, TABLE pub_test.testpub_nopk;
RESET client_min_messages;
\dRp+ testpub_for_tbl_schema
- Publication testpub_for_tbl_schema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_for_tbl_schema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"pub_test.testpub_nopk"
Tables from schemas:
@@ -144,10 +147,10 @@ LINE 1: ...CATION testpub_parsertst FOR TABLES IN SCHEMA foo, test.foo;
-- should be able to add a table of the same schema to the schema publication
ALTER PUBLICATION testpub_forschema ADD TABLE pub_test.testpub_nopk;
\dRp+ testpub_forschema
- Publication testpub_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"pub_test.testpub_nopk"
Tables from schemas:
@@ -156,10 +159,10 @@ Tables from schemas:
-- should be able to drop the table
ALTER PUBLICATION testpub_forschema DROP TABLE pub_test.testpub_nopk;
\dRp+ testpub_forschema
- Publication testpub_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test"
@@ -170,10 +173,10 @@ ERROR: relation "testpub_nopk" is not part of the publication
-- should be able to set table to schema publication
ALTER PUBLICATION testpub_forschema SET TABLE pub_test.testpub_nopk;
\dRp+ testpub_forschema
- Publication testpub_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"pub_test.testpub_nopk"
@@ -197,10 +200,10 @@ Not-null constraints:
"testpub_tbl2_id_not_null" NOT NULL "id"
\dRp+ testpub_foralltables
- Publication testpub_foralltables
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | t | t | t | f | f | f
+ Publication testpub_foralltables
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | t | f | t | t | f | f | f | f
(1 row)
DROP TABLE testpub_tbl2;
@@ -212,24 +215,525 @@ CREATE PUBLICATION testpub3 FOR TABLE testpub_tbl3;
CREATE PUBLICATION testpub4 FOR TABLE ONLY testpub_tbl3;
RESET client_min_messages;
\dRp+ testpub3
- Publication testpub3
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub3
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"public.testpub_tbl3"
"public.testpub_tbl3a"
\dRp+ testpub4
- Publication testpub4
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub4
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"public.testpub_tbl3"
DROP TABLE testpub_tbl3, testpub_tbl3a;
DROP PUBLICATION testpub3, testpub4;
+--- adding sequences
+CREATE SEQUENCE testpub_seq0;
+CREATE SEQUENCE pub_test.testpub_seq1;
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_forallsequences FOR ALL SEQUENCES WITH (publish = 'sequence');
+RESET client_min_messages;
+ALTER PUBLICATION testpub_forallsequences SET (publish = 'insert, sequence');
+CREATE SEQUENCE testpub_seq2;
+-- fail - can't add to for all sequences publication
+ALTER PUBLICATION testpub_forallsequences ADD SEQUENCE testpub_seq2;
+ERROR: publication "testpub_forallsequences" is defined as FOR ALL SEQUENCES
+DETAIL: Sequences cannot be added to or dropped from FOR ALL SEQUENCES publications.
+-- fail - can't drop from all sequences publication
+ALTER PUBLICATION testpub_forallsequences DROP SEQUENCE testpub_seq2;
+ERROR: publication "testpub_forallsequences" is defined as FOR ALL SEQUENCES
+DETAIL: Sequences cannot be added to or dropped from FOR ALL SEQUENCES publications.
+-- fail - can't add to for all sequences publication
+ALTER PUBLICATION testpub_forallsequences SET SEQUENCE pub_test.testpub_seq1;
+ERROR: publication "testpub_forallsequences" is defined as FOR ALL SEQUENCES
+DETAIL: Sequences cannot be added to or dropped from FOR ALL SEQUENCES publications.
+-- fail - can't add schema to 'FOR ALL SEQUENCES' publication
+ALTER PUBLICATION testpub_forallsequences ADD SEQUENCES IN SCHEMA pub_test;
+ERROR: publication "testpub_forallsequences" is defined as FOR ALL SEQUENCES
+DETAIL: Sequences from schema cannot be added to, dropped from, or set on FOR ALL SEQUENCES publications.
+-- fail - can't drop schema from 'FOR ALL SEQUENCES' publication
+ALTER PUBLICATION testpub_forallsequences DROP SEQUENCES IN SCHEMA pub_test;
+ERROR: publication "testpub_forallsequences" is defined as FOR ALL SEQUENCES
+DETAIL: Sequences from schema cannot be added to, dropped from, or set on FOR ALL SEQUENCES publications.
+-- fail - can't set schema to 'FOR ALL SEQUENCES' publication
+ALTER PUBLICATION testpub_forallsequences SET SEQUENCES IN SCHEMA pub_test;
+ERROR: publication "testpub_forallsequences" is defined as FOR ALL SEQUENCES
+DETAIL: Sequences from schema cannot be added to, dropped from, or set on FOR ALL SEQUENCES publications.
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_forsequence FOR SEQUENCE testpub_seq0;
+RESET client_min_messages;
+-- should be able to add schema to 'FOR SEQUENCE' publication
+ALTER PUBLICATION testpub_forsequence ADD SEQUENCES IN SCHEMA pub_test;
+\dRp+ testpub_forsequence
+ Publication testpub_forsequence
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Sequences:
+ "public.testpub_seq0"
+Sequences from schemas:
+ "pub_test"
+
+-- add sequence from the schema we already added
+ALTER PUBLICATION testpub_forsequence ADD SEQUENCE pub_test.testpub_seq1;
+-- fail - can't add sequence using ADD TABLE command
+ALTER PUBLICATION testpub_forsequence ADD TABLE pub_test.testpub_seq1;
+ERROR: object type does not match type expected by command
+-- should be able to drop schema from 'FOR SEQUENCE' publication
+ALTER PUBLICATION testpub_forsequence DROP SEQUENCES IN SCHEMA pub_test;
+\dRp+ testpub_forsequence
+ Publication testpub_forsequence
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Sequences:
+ "pub_test.testpub_seq1"
+ "public.testpub_seq0"
+
+-- should be able to set schema to 'FOR SEQUENCE' publication
+ALTER PUBLICATION testpub_forsequence SET SEQUENCES IN SCHEMA pub_test;
+\dRp+ testpub_forsequence
+ Publication testpub_forsequence
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Sequences from schemas:
+ "pub_test"
+
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_forschema FOR SEQUENCES IN SCHEMA pub_test;
+RESET client_min_messages;
+-- should be able to set sequence to schema publication
+ALTER PUBLICATION testpub_forschema SET SEQUENCE pub_test.testpub_seq1;
+\dRp+ testpub_forschema
+ Publication testpub_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Sequences:
+ "pub_test.testpub_seq1"
+
+SELECT pubname, puballtables, puballsequences FROM pg_publication WHERE pubname = 'testpub_forallsequences';
+ pubname | puballtables | puballsequences
+-------------------------+--------------+-----------------
+ testpub_forallsequences | f | t
+(1 row)
+
+\d+ pub_test.testpub_seq1
+ Sequence "pub_test.testpub_seq1"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+Publications:
+ "testpub_forallsequences"
+ "testpub_forschema"
+ "testpub_forsequence"
+
+\dRp+ testpub_forallsequences
+ Publication testpub_forallsequences
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | t | t | f | f | f | t | f
+(1 row)
+
+DROP SEQUENCE testpub_seq0, pub_test.testpub_seq1, testpub_seq2;
+DROP PUBLICATION testpub_forallsequences, testpub_forsequence, testpub_forschema;
+-- publication testing multiple sequences at the same time
+CREATE SEQUENCE testpub_seq1;
+CREATE SEQUENCE testpub_seq2;
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_multi FOR SEQUENCE testpub_seq1, testpub_seq2;
+RESET client_min_messages;
+\dRp+ testpub_multi
+ Publication testpub_multi
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Sequences:
+ "public.testpub_seq1"
+ "public.testpub_seq2"
+
+DROP PUBLICATION testpub_multi;
+DROP SEQUENCE testpub_seq1;
+DROP SEQUENCE testpub_seq2;
+-- Publication mixing tables and sequences
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_mix;
+RESET client_min_messages;
+CREATE SEQUENCE testpub_seq1;
+CREATE SEQUENCE pub_test.testpub_seq2;
+ALTER PUBLICATION testpub_mix ADD SEQUENCE testpub_seq1, TABLE testpub_tbl1;
+\dRp+ testpub_mix
+ Publication testpub_mix
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Tables:
+ "public.testpub_tbl1"
+Sequences:
+ "public.testpub_seq1"
+
+ALTER PUBLICATION testpub_mix ADD SEQUENCES IN SCHEMA pub_test, TABLES IN SCHEMA pub_test;
+\dRp+ testpub_mix
+ Publication testpub_mix
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Tables:
+ "public.testpub_tbl1"
+Tables from schemas:
+ "pub_test"
+Sequences:
+ "public.testpub_seq1"
+Sequences from schemas:
+ "pub_test"
+
+ALTER PUBLICATION testpub_mix DROP SEQUENCES IN SCHEMA pub_test;
+\dRp+ testpub_mix
+ Publication testpub_mix
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Tables:
+ "public.testpub_tbl1"
+Tables from schemas:
+ "pub_test"
+Sequences:
+ "public.testpub_seq1"
+
+ALTER PUBLICATION testpub_mix DROP TABLES IN SCHEMA pub_test;
+\dRp+ testpub_mix
+ Publication testpub_mix
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Tables:
+ "public.testpub_tbl1"
+Sequences:
+ "public.testpub_seq1"
+
+DROP PUBLICATION testpub_mix;
+DROP SEQUENCE testpub_seq1;
+DROP SEQUENCE pub_test.testpub_seq2;
+-- make sure we replicate only the correct relation type
+CREATE SCHEMA pub_test1;
+CREATE SEQUENCE pub_test1.test_seq1;
+CREATE TABLE pub_test1.test_tbl1 (a int primary key, b int);
+CREATE SCHEMA pub_test2;
+CREATE SEQUENCE pub_test2.test_seq2;
+CREATE TABLE pub_test2.test_tbl2 (a int primary key, b int);
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_schemas;
+RESET client_min_messages;
+-- add tables from one schema, sequences from the other
+ALTER PUBLICATION testpub_schemas ADD TABLES IN SCHEMA pub_test2;
+ALTER PUBLICATION testpub_schemas ADD SEQUENCES IN SCHEMA pub_test1;
+\dRp+ testpub_schemas
+ Publication testpub_schemas
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Tables from schemas:
+ "pub_test2"
+Sequences from schemas:
+ "pub_test1"
+
+\dn+ pub_test1
+ List of schemas
+ Name | Owner | Access privileges | Description
+-----------+--------------------------+-------------------+-------------
+ pub_test1 | regress_publication_user | |
+Publications:
+ "testpub_schemas" (sequences)
+
+\dn+ pub_test2
+ List of schemas
+ Name | Owner | Access privileges | Description
+-----------+--------------------------+-------------------+-------------
+ pub_test2 | regress_publication_user | |
+Publications:
+ "testpub_schemas" (tables)
+
+\d+ pub_test1.test_seq1;
+ Sequence "pub_test1.test_seq1"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test1.test_tbl1;
+ Table "pub_test1.test_tbl1"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl1_pkey" PRIMARY KEY, btree (a)
+
+\d+ pub_test2.test_seq2;
+ Sequence "pub_test2.test_seq2"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+
+\d+ pub_test2.test_tbl2;
+ Table "pub_test2.test_tbl2"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl2_pkey" PRIMARY KEY, btree (a)
+Publications:
+ "testpub_schemas"
+
+-- add the other object type from each schema
+ALTER PUBLICATION testpub_schemas ADD TABLES IN SCHEMA pub_test1;
+ALTER PUBLICATION testpub_schemas ADD SEQUENCES IN SCHEMA pub_test2;
+\dRp+ testpub_schemas
+ Publication testpub_schemas
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Tables from schemas:
+ "pub_test1"
+ "pub_test2"
+Sequences from schemas:
+ "pub_test1"
+ "pub_test2"
+
+\dn+ pub_test1
+ List of schemas
+ Name | Owner | Access privileges | Description
+-----------+--------------------------+-------------------+-------------
+ pub_test1 | regress_publication_user | |
+Publications:
+ "testpub_schemas" (sequences)
+ "testpub_schemas" (tables)
+
+\dn+ pub_test2
+ List of schemas
+ Name | Owner | Access privileges | Description
+-----------+--------------------------+-------------------+-------------
+ pub_test2 | regress_publication_user | |
+Publications:
+ "testpub_schemas" (tables)
+ "testpub_schemas" (sequences)
+
+\d+ pub_test1.test_seq1;
+ Sequence "pub_test1.test_seq1"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test1.test_tbl1;
+ Table "pub_test1.test_tbl1"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl1_pkey" PRIMARY KEY, btree (a)
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test2.test_seq2;
+ Sequence "pub_test2.test_seq2"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test2.test_tbl2;
+ Table "pub_test2.test_tbl2"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl2_pkey" PRIMARY KEY, btree (a)
+Publications:
+ "testpub_schemas"
+
+-- now drop the object type added first
+ALTER PUBLICATION testpub_schemas DROP TABLES IN SCHEMA pub_test2;
+ALTER PUBLICATION testpub_schemas DROP SEQUENCES IN SCHEMA pub_test1;
+\dRp+ testpub_schemas
+ Publication testpub_schemas
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Tables from schemas:
+ "pub_test1"
+Sequences from schemas:
+ "pub_test2"
+
+\dn+ pub_test1
+ List of schemas
+ Name | Owner | Access privileges | Description
+-----------+--------------------------+-------------------+-------------
+ pub_test1 | regress_publication_user | |
+Publications:
+ "testpub_schemas" (tables)
+
+\dn+ pub_test2
+ List of schemas
+ Name | Owner | Access privileges | Description
+-----------+--------------------------+-------------------+-------------
+ pub_test2 | regress_publication_user | |
+Publications:
+ "testpub_schemas" (sequences)
+
+\d+ pub_test1.test_seq1;
+ Sequence "pub_test1.test_seq1"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+
+\d+ pub_test1.test_tbl1;
+ Table "pub_test1.test_tbl1"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl1_pkey" PRIMARY KEY, btree (a)
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test2.test_seq2;
+ Sequence "pub_test2.test_seq2"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test2.test_tbl2;
+ Table "pub_test2.test_tbl2"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl2_pkey" PRIMARY KEY, btree (a)
+
+-- add a different schema (not including the already published sequences)
+ALTER PUBLICATION testpub_schemas ADD TABLE pub_test2.test_tbl2;
+ALTER PUBLICATION testpub_schemas ADD SEQUENCE pub_test1.test_seq1;
+\dRp+ testpub_schemas
+ Publication testpub_schemas
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Tables:
+ "pub_test2.test_tbl2"
+Tables from schemas:
+ "pub_test1"
+Sequences:
+ "pub_test1.test_seq1"
+Sequences from schemas:
+ "pub_test2"
+
+\d+ pub_test1.test_seq1;
+ Sequence "pub_test1.test_seq1"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test1.test_tbl1;
+ Table "pub_test1.test_tbl1"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl1_pkey" PRIMARY KEY, btree (a)
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test2.test_seq2;
+ Sequence "pub_test2.test_seq2"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test2.test_tbl2;
+ Table "pub_test2.test_tbl2"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl2_pkey" PRIMARY KEY, btree (a)
+Publications:
+ "testpub_schemas"
+
+-- now drop the explicitly added objects again
+ALTER PUBLICATION testpub_schemas DROP TABLE pub_test2.test_tbl2;
+ALTER PUBLICATION testpub_schemas DROP SEQUENCE pub_test1.test_seq1;
+\dRp+ testpub_schemas
+ Publication testpub_schemas
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Tables from schemas:
+ "pub_test1"
+Sequences from schemas:
+ "pub_test2"
+
+\d+ pub_test1.test_seq1;
+ Sequence "pub_test1.test_seq1"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+
+\d+ pub_test1.test_tbl1;
+ Table "pub_test1.test_tbl1"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl1_pkey" PRIMARY KEY, btree (a)
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test2.test_seq2;
+ Sequence "pub_test2.test_seq2"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test2.test_tbl2;
+ Table "pub_test2.test_tbl2"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl2_pkey" PRIMARY KEY, btree (a)
+
+DROP PUBLICATION testpub_schemas;
+DROP TABLE pub_test1.test_tbl1, pub_test2.test_tbl2;
+DROP SEQUENCE pub_test1.test_seq1, pub_test2.test_seq2;
+DROP SCHEMA pub_test1, pub_test2;
-- Tests for partitioned tables
SET client_min_messages = 'ERROR';
CREATE PUBLICATION testpub_forparted;
@@ -245,10 +749,10 @@ UPDATE testpub_parted1 SET a = 1;
-- only parent is listed as being in publication, not the partition
ALTER PUBLICATION testpub_forparted ADD TABLE testpub_parted;
\dRp+ testpub_forparted
- Publication testpub_forparted
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_forparted
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"public.testpub_parted"
@@ -263,10 +767,10 @@ ALTER TABLE testpub_parted DETACH PARTITION testpub_parted1;
UPDATE testpub_parted1 SET a = 1;
ALTER PUBLICATION testpub_forparted SET (publish_via_partition_root = true);
\dRp+ testpub_forparted
- Publication testpub_forparted
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | t
+ Publication testpub_forparted
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | t
Tables:
"public.testpub_parted"
@@ -295,10 +799,10 @@ SET client_min_messages = 'ERROR';
CREATE PUBLICATION testpub5 FOR TABLE testpub_rf_tbl1, testpub_rf_tbl2 WHERE (c <> 'test' AND d < 5) WITH (publish = 'insert');
RESET client_min_messages;
\dRp+ testpub5
- Publication testpub5
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | f | f | f | f
+ Publication testpub5
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | f | f | f | f | f
Tables:
"public.testpub_rf_tbl1"
"public.testpub_rf_tbl2" WHERE ((c <> 'test'::text) AND (d < 5))
@@ -311,10 +815,10 @@ Tables:
ALTER PUBLICATION testpub5 ADD TABLE testpub_rf_tbl3 WHERE (e > 1000 AND e < 2000);
\dRp+ testpub5
- Publication testpub5
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | f | f | f | f
+ Publication testpub5
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | f | f | f | f | f
Tables:
"public.testpub_rf_tbl1"
"public.testpub_rf_tbl2" WHERE ((c <> 'test'::text) AND (d < 5))
@@ -330,10 +834,10 @@ Publications:
ALTER PUBLICATION testpub5 DROP TABLE testpub_rf_tbl2;
\dRp+ testpub5
- Publication testpub5
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | f | f | f | f
+ Publication testpub5
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | f | f | f | f | f
Tables:
"public.testpub_rf_tbl1"
"public.testpub_rf_tbl3" WHERE ((e > 1000) AND (e < 2000))
@@ -341,10 +845,10 @@ Tables:
-- remove testpub_rf_tbl1 and add testpub_rf_tbl3 again (another WHERE expression)
ALTER PUBLICATION testpub5 SET TABLE testpub_rf_tbl3 WHERE (e > 300 AND e < 500);
\dRp+ testpub5
- Publication testpub5
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | f | f | f | f
+ Publication testpub5
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | f | f | f | f | f
Tables:
"public.testpub_rf_tbl3" WHERE ((e > 300) AND (e < 500))
@@ -377,10 +881,10 @@ SET client_min_messages = 'ERROR';
CREATE PUBLICATION testpub_syntax1 FOR TABLE testpub_rf_tbl1, ONLY testpub_rf_tbl3 WHERE (e < 999) WITH (publish = 'insert');
RESET client_min_messages;
\dRp+ testpub_syntax1
- Publication testpub_syntax1
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | f | f | f | f
+ Publication testpub_syntax1
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | f | f | f | f | f
Tables:
"public.testpub_rf_tbl1"
"public.testpub_rf_tbl3" WHERE (e < 999)
@@ -390,10 +894,10 @@ SET client_min_messages = 'ERROR';
CREATE PUBLICATION testpub_syntax2 FOR TABLE testpub_rf_tbl1, testpub_rf_schema1.testpub_rf_tbl5 WHERE (h < 999) WITH (publish = 'insert');
RESET client_min_messages;
\dRp+ testpub_syntax2
- Publication testpub_syntax2
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | f | f | f | f
+ Publication testpub_syntax2
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | f | f | f | f | f
Tables:
"public.testpub_rf_tbl1"
"testpub_rf_schema1.testpub_rf_tbl5" WHERE (h < 999)
@@ -508,10 +1012,10 @@ CREATE PUBLICATION testpub6 FOR TABLES IN SCHEMA testpub_rf_schema2;
ALTER PUBLICATION testpub6 SET TABLES IN SCHEMA testpub_rf_schema2, TABLE testpub_rf_schema2.testpub_rf_tbl6 WHERE (i < 99);
RESET client_min_messages;
\dRp+ testpub6
- Publication testpub6
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub6
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"testpub_rf_schema2.testpub_rf_tbl6" WHERE (i < 99)
Tables from schemas:
@@ -725,10 +1229,10 @@ CREATE PUBLICATION testpub_table_ins WITH (publish = 'insert, truncate');
RESET client_min_messages;
ALTER PUBLICATION testpub_table_ins ADD TABLE testpub_tbl5 (a); -- ok
\dRp+ testpub_table_ins
- Publication testpub_table_ins
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | f | f | t | f
+ Publication testpub_table_ins
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | f | f | t | f | f
Tables:
"public.testpub_tbl5" (a)
@@ -902,10 +1406,10 @@ CREATE TABLE testpub_tbl_both_filters (a int, b int, c int, PRIMARY KEY (a,c));
ALTER TABLE testpub_tbl_both_filters REPLICA IDENTITY USING INDEX testpub_tbl_both_filters_pkey;
ALTER PUBLICATION testpub_both_filters ADD TABLE testpub_tbl_both_filters (a,c) WHERE (c != 1);
\dRp+ testpub_both_filters
- Publication testpub_both_filters
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_both_filters
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"public.testpub_tbl_both_filters" (a, c) WHERE (c <> 1)
@@ -1110,10 +1614,10 @@ ERROR: relation "testpub_tbl1" is already member of publication "testpub_fortbl
CREATE PUBLICATION testpub_fortbl FOR TABLE testpub_tbl1;
ERROR: publication "testpub_fortbl" already exists
\dRp+ testpub_fortbl
- Publication testpub_fortbl
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_fortbl
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"pub_test.testpub_nopk"
"public.testpub_tbl1"
@@ -1153,10 +1657,10 @@ Not-null constraints:
"testpub_tbl1_id_not_null" NOT NULL "id"
\dRp+ testpub_default
- Publication testpub_default
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | f | f
+ Publication testpub_default
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | f | t | f
Tables:
"pub_test.testpub_nopk"
"public.testpub_tbl1"
@@ -1236,10 +1740,10 @@ REVOKE CREATE ON DATABASE regression FROM regress_publication_user2;
DROP TABLE testpub_parted;
DROP TABLE testpub_tbl1;
\dRp+ testpub_default
- Publication testpub_default
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | f | f
+ Publication testpub_default
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | f | t | f
(1 row)
-- fail - must be owner of publication
@@ -1249,20 +1753,20 @@ ERROR: must be owner of publication testpub_default
RESET ROLE;
ALTER PUBLICATION testpub_default RENAME TO testpub_foo;
\dRp testpub_foo
- List of publications
- Name | Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
--------------+--------------------------+------------+---------+---------+---------+-----------+----------
- testpub_foo | regress_publication_user | f | t | t | t | f | f
+ List of publications
+ Name | Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+-------------+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ testpub_foo | regress_publication_user | f | f | t | t | t | f | t | f
(1 row)
-- rename back to keep the rest simple
ALTER PUBLICATION testpub_foo RENAME TO testpub_default;
ALTER PUBLICATION testpub_default OWNER TO regress_publication_user2;
\dRp testpub_default
- List of publications
- Name | Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
------------------+---------------------------+------------+---------+---------+---------+-----------+----------
- testpub_default | regress_publication_user2 | f | t | t | t | f | f
+ List of publications
+ Name | Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+-----------------+---------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ testpub_default | regress_publication_user2 | f | f | t | t | t | f | t | f
(1 row)
-- adding schemas and tables
@@ -1278,19 +1782,19 @@ CREATE TABLE "CURRENT_SCHEMA"."CURRENT_SCHEMA"(id int);
SET client_min_messages = 'ERROR';
CREATE PUBLICATION testpub1_forschema FOR TABLES IN SCHEMA pub_test1;
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
CREATE PUBLICATION testpub2_forschema FOR TABLES IN SCHEMA pub_test1, pub_test2, pub_test3;
\dRp+ testpub2_forschema
- Publication testpub2_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub2_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
"pub_test2"
@@ -1304,44 +1808,44 @@ CREATE PUBLICATION testpub6_forschema FOR TABLES IN SCHEMA "CURRENT_SCHEMA", CUR
CREATE PUBLICATION testpub_fortable FOR TABLE "CURRENT_SCHEMA"."CURRENT_SCHEMA";
RESET client_min_messages;
\dRp+ testpub3_forschema
- Publication testpub3_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub3_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"public"
\dRp+ testpub4_forschema
- Publication testpub4_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub4_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"CURRENT_SCHEMA"
\dRp+ testpub5_forschema
- Publication testpub5_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub5_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"CURRENT_SCHEMA"
"public"
\dRp+ testpub6_forschema
- Publication testpub6_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub6_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"CURRENT_SCHEMA"
"public"
\dRp+ testpub_fortable
- Publication testpub_fortable
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_fortable
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"CURRENT_SCHEMA.CURRENT_SCHEMA"
@@ -1375,10 +1879,10 @@ ERROR: schema "testpub_view" does not exist
-- dropping the schema should reflect the change in publication
DROP SCHEMA pub_test3;
\dRp+ testpub2_forschema
- Publication testpub2_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub2_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
"pub_test2"
@@ -1386,20 +1890,20 @@ Tables from schemas:
-- renaming the schema should reflect the change in publication
ALTER SCHEMA pub_test1 RENAME to pub_test1_renamed;
\dRp+ testpub2_forschema
- Publication testpub2_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub2_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1_renamed"
"pub_test2"
ALTER SCHEMA pub_test1_renamed RENAME to pub_test1;
\dRp+ testpub2_forschema
- Publication testpub2_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub2_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
"pub_test2"
@@ -1407,10 +1911,10 @@ Tables from schemas:
-- alter publication add schema
ALTER PUBLICATION testpub1_forschema ADD TABLES IN SCHEMA pub_test2;
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
"pub_test2"
@@ -1419,10 +1923,10 @@ Tables from schemas:
ALTER PUBLICATION testpub1_forschema ADD TABLES IN SCHEMA non_existent_schema;
ERROR: schema "non_existent_schema" does not exist
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
"pub_test2"
@@ -1431,10 +1935,10 @@ Tables from schemas:
ALTER PUBLICATION testpub1_forschema ADD TABLES IN SCHEMA pub_test1;
ERROR: schema "pub_test1" is already member of publication "testpub1_forschema"
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
"pub_test2"
@@ -1442,10 +1946,10 @@ Tables from schemas:
-- alter publication drop schema
ALTER PUBLICATION testpub1_forschema DROP TABLES IN SCHEMA pub_test2;
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
@@ -1453,10 +1957,10 @@ Tables from schemas:
ALTER PUBLICATION testpub1_forschema DROP TABLES IN SCHEMA pub_test2;
ERROR: tables from schema "pub_test2" are not part of the publication
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
@@ -1464,29 +1968,29 @@ Tables from schemas:
ALTER PUBLICATION testpub1_forschema DROP TABLES IN SCHEMA non_existent_schema;
ERROR: schema "non_existent_schema" does not exist
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
-- drop all schemas
ALTER PUBLICATION testpub1_forschema DROP TABLES IN SCHEMA pub_test1;
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
(1 row)
-- alter publication set multiple schema
ALTER PUBLICATION testpub1_forschema SET TABLES IN SCHEMA pub_test1, pub_test2;
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
"pub_test2"
@@ -1495,10 +1999,10 @@ Tables from schemas:
ALTER PUBLICATION testpub1_forschema SET TABLES IN SCHEMA non_existent_schema;
ERROR: schema "non_existent_schema" does not exist
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
"pub_test2"
@@ -1507,10 +2011,10 @@ Tables from schemas:
-- removing the duplicate schemas
ALTER PUBLICATION testpub1_forschema SET TABLES IN SCHEMA pub_test1, pub_test1;
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
@@ -1589,18 +2093,18 @@ SET client_min_messages = 'ERROR';
CREATE PUBLICATION testpub3_forschema;
RESET client_min_messages;
\dRp+ testpub3_forschema
- Publication testpub3_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub3_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
(1 row)
ALTER PUBLICATION testpub3_forschema SET TABLES IN SCHEMA pub_test1;
\dRp+ testpub3_forschema
- Publication testpub3_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub3_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
@@ -1610,20 +2114,20 @@ CREATE PUBLICATION testpub_forschema_fortable FOR TABLES IN SCHEMA pub_test1, TA
CREATE PUBLICATION testpub_fortable_forschema FOR TABLE pub_test2.tbl1, TABLES IN SCHEMA pub_test1;
RESET client_min_messages;
\dRp+ testpub_forschema_fortable
- Publication testpub_forschema_fortable
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_forschema_fortable
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"pub_test2.tbl1"
Tables from schemas:
"pub_test1"
\dRp+ testpub_fortable_forschema
- Publication testpub_fortable_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_fortable_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"pub_test2.tbl1"
Tables from schemas:
@@ -1667,40 +2171,85 @@ CREATE SCHEMA sch1;
CREATE SCHEMA sch2;
CREATE TABLE sch1.tbl1 (a int) PARTITION BY RANGE(a);
CREATE TABLE sch2.tbl1_part1 PARTITION OF sch1.tbl1 FOR VALUES FROM (1) to (10);
+CREATE SEQUENCE sch1.seq1;
+CREATE SEQUENCE sch2.seq2;
-- Schema publication that does not include the schema that has the parent table
CREATE PUBLICATION pub FOR TABLES IN SCHEMA sch2 WITH (PUBLISH_VIA_PARTITION_ROOT=1);
+ALTER PUBLICATION pub ADD SEQUENCES IN SCHEMA sch2;
SELECT * FROM pg_publication_tables;
pubname | schemaname | tablename | attnames | rowfilter
---------+------------+------------+----------+-----------
pub | sch2 | tbl1_part1 | {a} |
(1 row)
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch2 | seq2
+(1 row)
+
DROP PUBLICATION pub;
-- Table publication that does not include the parent table
CREATE PUBLICATION pub FOR TABLE sch2.tbl1_part1 WITH (PUBLISH_VIA_PARTITION_ROOT=1);
+ALTER PUBLICATION pub ADD SEQUENCE sch2.seq2;
SELECT * FROM pg_publication_tables;
pubname | schemaname | tablename | attnames | rowfilter
---------+------------+------------+----------+-----------
pub | sch2 | tbl1_part1 | {a} |
(1 row)
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch2 | seq2
+(1 row)
+
-- Table publication that includes both the parent table and the child table
ALTER PUBLICATION pub ADD TABLE sch1.tbl1;
+ALTER PUBLICATION pub ADD SEQUENCE sch1.seq1;
SELECT * FROM pg_publication_tables;
pubname | schemaname | tablename | attnames | rowfilter
---------+------------+-----------+----------+-----------
pub | sch1 | tbl1 | {a} |
(1 row)
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch1 | seq1
+ pub | sch2 | seq2
+(2 rows)
+
DROP PUBLICATION pub;
-- Schema publication that does not include the schema that has the parent table
CREATE PUBLICATION pub FOR TABLES IN SCHEMA sch2 WITH (PUBLISH_VIA_PARTITION_ROOT=0);
+ALTER PUBLICATION pub ADD SEQUENCE sch1.seq1;
SELECT * FROM pg_publication_tables;
pubname | schemaname | tablename | attnames | rowfilter
---------+------------+------------+----------+-----------
pub | sch2 | tbl1_part1 | {a} |
(1 row)
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch1 | seq1
+(1 row)
+
+DROP PUBLICATION pub;
+-- Sequence publication
+CREATE PUBLICATION pub FOR SEQUENCE sch2.seq2;
+SELECT * FROM pg_publication_tables;
+ pubname | schemaname | tablename | attnames | rowfilter
+---------+------------+-----------+----------+-----------
+(0 rows)
+
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch2 | seq2
+(1 row)
+
DROP PUBLICATION pub;
-- Table publication that does not include the parent table
CREATE PUBLICATION pub FOR TABLE sch2.tbl1_part1 WITH (PUBLISH_VIA_PARTITION_ROOT=0);
@@ -1710,14 +2259,26 @@ SELECT * FROM pg_publication_tables;
pub | sch2 | tbl1_part1 | {a} |
(1 row)
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+(0 rows)
+
-- Table publication that includes both the parent table and the child table
ALTER PUBLICATION pub ADD TABLE sch1.tbl1;
+ALTER PUBLICATION pub ADD SEQUENCES IN SCHEMA sch2;
SELECT * FROM pg_publication_tables;
pubname | schemaname | tablename | attnames | rowfilter
---------+------------+------------+----------+-----------
pub | sch2 | tbl1_part1 | {a} |
(1 row)
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch2 | seq2
+(1 row)
+
DROP PUBLICATION pub;
DROP TABLE sch2.tbl1_part1;
DROP TABLE sch1.tbl1;
@@ -1733,9 +2294,81 @@ SELECT * FROM pg_publication_tables;
pub | sch1 | tbl1 | {a} |
(1 row)
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+(0 rows)
+
+DROP PUBLICATION pub;
+-- Schema publication
+CREATE PUBLICATION pub FOR SEQUENCE sch2.seq2;
+SELECT * FROM pg_publication_tables;
+ pubname | schemaname | tablename | attnames | rowfilter
+---------+------------+-----------+----------+-----------
+(0 rows)
+
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch2 | seq2
+(1 row)
+
+DROP PUBLICATION pub;
+-- Sequence publication
+CREATE PUBLICATION pub FOR SEQUENCES IN SCHEMA sch2;
+SELECT * FROM pg_publication_tables;
+ pubname | schemaname | tablename | attnames | rowfilter
+---------+------------+-----------+----------+-----------
+(0 rows)
+
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch2 | seq2
+(1 row)
+
+ALTER PUBLICATION pub ADD SEQUENCE sch1.seq1;
+SELECT * FROM pg_publication_tables;
+ pubname | schemaname | tablename | attnames | rowfilter
+---------+------------+-----------+----------+-----------
+(0 rows)
+
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch1 | seq1
+ pub | sch2 | seq2
+(2 rows)
+
+ALTER PUBLICATION pub DROP SEQUENCE sch1.seq1;
+SELECT * FROM pg_publication_tables;
+ pubname | schemaname | tablename | attnames | rowfilter
+---------+------------+-----------+----------+-----------
+(0 rows)
+
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch2 | seq2
+(1 row)
+
+ALTER PUBLICATION pub ADD SEQUENCES IN SCHEMA sch1;
+SELECT * FROM pg_publication_tables;
+ pubname | schemaname | tablename | attnames | rowfilter
+---------+------------+-----------+----------+-----------
+(0 rows)
+
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch1 | seq1
+ pub | sch2 | seq2
+(2 rows)
+
RESET client_min_messages;
DROP PUBLICATION pub;
DROP TABLE sch1.tbl1;
+DROP SEQUENCE sch1.seq1, sch2.seq2;
DROP SCHEMA sch1 cascade;
DROP SCHEMA sch2 cascade;
RESET SESSION AUTHORIZATION;
diff --git a/src/test/regress/expected/rules.out b/src/test/regress/expected/rules.out
index 05070393b99..c62dead92a7 100644
--- a/src/test/regress/expected/rules.out
+++ b/src/test/regress/expected/rules.out
@@ -1441,6 +1441,14 @@ pg_prepared_xacts| SELECT p.transaction,
FROM ((pg_prepared_xact() p(transaction, gid, prepared, ownerid, dbid)
LEFT JOIN pg_authid u ON ((p.ownerid = u.oid)))
LEFT JOIN pg_database d ON ((p.dbid = d.oid)));
+pg_publication_sequences| SELECT p.pubname,
+ n.nspname AS schemaname,
+ c.relname AS sequencename
+ FROM pg_publication p,
+ LATERAL pg_get_publication_sequences((p.pubname)::text) gps(relid),
+ (pg_class c
+ JOIN pg_namespace n ON ((n.oid = c.relnamespace)))
+ WHERE (c.oid = gps.relid);
pg_publication_tables| SELECT p.pubname,
n.nspname AS schemaname,
c.relname AS tablename,
diff --git a/src/test/regress/sql/object_address.sql b/src/test/regress/sql/object_address.sql
index 1a6c61f49d5..5c6cdeb05df 100644
--- a/src/test/regress/sql/object_address.sql
+++ b/src/test/regress/sql/object_address.sql
@@ -50,6 +50,7 @@ CREATE TRANSFORM FOR int LANGUAGE SQL (
SET client_min_messages = 'ERROR';
CREATE PUBLICATION addr_pub FOR TABLE addr_nsp.gentable;
CREATE PUBLICATION addr_pub_schema FOR TABLES IN SCHEMA addr_nsp;
+CREATE PUBLICATION addr_pub_schema2 FOR SEQUENCES IN SCHEMA addr_nsp;
RESET client_min_messages;
CREATE SUBSCRIPTION regress_addr_sub CONNECTION '' PUBLICATION bar WITH (connect = false, slot_name = NONE);
CREATE STATISTICS addr_nsp.gentable_stat ON a, b FROM addr_nsp.gentable;
@@ -206,7 +207,8 @@ WITH objects (type, name, args) AS (VALUES
('transform', '{int}', '{sql}'),
('access method', '{btree}', '{}'),
('publication', '{addr_pub}', '{}'),
- ('publication namespace', '{addr_nsp}', '{addr_pub_schema}'),
+ ('publication namespace', '{addr_nsp}', '{addr_pub_schema, t}'),
+ ('publication namespace', '{addr_nsp}', '{addr_pub_schema2, s}'),
('publication relation', '{addr_nsp, gentable}', '{addr_pub}'),
('subscription', '{regress_addr_sub}', '{}'),
('statistics object', '{addr_nsp, gentable_stat}', '{}')
@@ -227,6 +229,7 @@ ORDER BY addr1.classid, addr1.objid, addr1.objsubid;
DROP FOREIGN DATA WRAPPER addr_fdw CASCADE;
DROP PUBLICATION addr_pub;
DROP PUBLICATION addr_pub_schema;
+DROP PUBLICATION addr_pub_schema2;
DROP SUBSCRIPTION regress_addr_sub;
DROP SCHEMA addr_nsp CASCADE;
diff --git a/src/test/regress/sql/publication.sql b/src/test/regress/sql/publication.sql
index d5051a5e746..016b58c67ec 100644
--- a/src/test/regress/sql/publication.sql
+++ b/src/test/regress/sql/publication.sql
@@ -27,7 +27,7 @@ CREATE PUBLICATION testpub_xxx WITH (publish_via_partition_root = 'true', publis
\dRp
-ALTER PUBLICATION testpub_default SET (publish = 'insert, update, delete');
+ALTER PUBLICATION testpub_default SET (publish = 'insert, update, delete, sequence');
\dRp
@@ -46,6 +46,8 @@ ALTER PUBLICATION testpub_foralltables SET (publish = 'insert, update');
CREATE TABLE testpub_tbl2 (id serial primary key, data text);
-- fail - can't add to for all tables publication
ALTER PUBLICATION testpub_foralltables ADD TABLE testpub_tbl2;
+-- fail - can't add a table using ADD SEQUENCE command
+ALTER PUBLICATION testpub_foralltables ADD SEQUENCE testpub_tbl2;
-- fail - can't drop from all tables publication
ALTER PUBLICATION testpub_foralltables DROP TABLE testpub_tbl2;
-- fail - can't add to for all tables publication
@@ -117,6 +119,188 @@ RESET client_min_messages;
DROP TABLE testpub_tbl3, testpub_tbl3a;
DROP PUBLICATION testpub3, testpub4;
+--- adding sequences
+CREATE SEQUENCE testpub_seq0;
+CREATE SEQUENCE pub_test.testpub_seq1;
+
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_forallsequences FOR ALL SEQUENCES WITH (publish = 'sequence');
+RESET client_min_messages;
+ALTER PUBLICATION testpub_forallsequences SET (publish = 'insert, sequence');
+
+CREATE SEQUENCE testpub_seq2;
+-- fail - can't add to for all sequences publication
+ALTER PUBLICATION testpub_forallsequences ADD SEQUENCE testpub_seq2;
+-- fail - can't drop from all sequences publication
+ALTER PUBLICATION testpub_forallsequences DROP SEQUENCE testpub_seq2;
+-- fail - can't add to for all sequences publication
+ALTER PUBLICATION testpub_forallsequences SET SEQUENCE pub_test.testpub_seq1;
+
+-- fail - can't add schema to 'FOR ALL SEQUENCES' publication
+ALTER PUBLICATION testpub_forallsequences ADD SEQUENCES IN SCHEMA pub_test;
+-- fail - can't drop schema from 'FOR ALL SEQUENCES' publication
+ALTER PUBLICATION testpub_forallsequences DROP SEQUENCES IN SCHEMA pub_test;
+-- fail - can't set schema to 'FOR ALL SEQUENCES' publication
+ALTER PUBLICATION testpub_forallsequences SET SEQUENCES IN SCHEMA pub_test;
+
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_forsequence FOR SEQUENCE testpub_seq0;
+RESET client_min_messages;
+-- should be able to add schema to 'FOR SEQUENCE' publication
+ALTER PUBLICATION testpub_forsequence ADD SEQUENCES IN SCHEMA pub_test;
+\dRp+ testpub_forsequence
+-- add sequence from the schema we already added
+ALTER PUBLICATION testpub_forsequence ADD SEQUENCE pub_test.testpub_seq1;
+-- fail - can't add sequence using ADD TABLE command
+ALTER PUBLICATION testpub_forsequence ADD TABLE pub_test.testpub_seq1;
+-- should be able to drop schema from 'FOR SEQUENCE' publication
+ALTER PUBLICATION testpub_forsequence DROP SEQUENCES IN SCHEMA pub_test;
+\dRp+ testpub_forsequence
+-- should be able to set schema to 'FOR SEQUENCE' publication
+ALTER PUBLICATION testpub_forsequence SET SEQUENCES IN SCHEMA pub_test;
+\dRp+ testpub_forsequence
+
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_forschema FOR SEQUENCES IN SCHEMA pub_test;
+RESET client_min_messages;
+-- should be able to set sequence to schema publication
+ALTER PUBLICATION testpub_forschema SET SEQUENCE pub_test.testpub_seq1;
+\dRp+ testpub_forschema
+
+SELECT pubname, puballtables, puballsequences FROM pg_publication WHERE pubname = 'testpub_forallsequences';
+\d+ pub_test.testpub_seq1
+\dRp+ testpub_forallsequences
+DROP SEQUENCE testpub_seq0, pub_test.testpub_seq1, testpub_seq2;
+DROP PUBLICATION testpub_forallsequences, testpub_forsequence, testpub_forschema;
+
+
+-- publication testing multiple sequences at the same time
+CREATE SEQUENCE testpub_seq1;
+CREATE SEQUENCE testpub_seq2;
+
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_multi FOR SEQUENCE testpub_seq1, testpub_seq2;
+RESET client_min_messages;
+
+\dRp+ testpub_multi
+
+DROP PUBLICATION testpub_multi;
+DROP SEQUENCE testpub_seq1;
+DROP SEQUENCE testpub_seq2;
+
+
+-- Publication mixing tables and sequences
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_mix;
+RESET client_min_messages;
+
+CREATE SEQUENCE testpub_seq1;
+CREATE SEQUENCE pub_test.testpub_seq2;
+
+ALTER PUBLICATION testpub_mix ADD SEQUENCE testpub_seq1, TABLE testpub_tbl1;
+\dRp+ testpub_mix
+
+ALTER PUBLICATION testpub_mix ADD SEQUENCES IN SCHEMA pub_test, TABLES IN SCHEMA pub_test;
+\dRp+ testpub_mix
+
+ALTER PUBLICATION testpub_mix DROP SEQUENCES IN SCHEMA pub_test;
+\dRp+ testpub_mix
+
+ALTER PUBLICATION testpub_mix DROP TABLES IN SCHEMA pub_test;
+\dRp+ testpub_mix
+
+DROP PUBLICATION testpub_mix;
+DROP SEQUENCE testpub_seq1;
+DROP SEQUENCE pub_test.testpub_seq2;
+
+
+-- make sure we replicate only the correct relation type
+CREATE SCHEMA pub_test1;
+CREATE SEQUENCE pub_test1.test_seq1;
+CREATE TABLE pub_test1.test_tbl1 (a int primary key, b int);
+
+CREATE SCHEMA pub_test2;
+CREATE SEQUENCE pub_test2.test_seq2;
+CREATE TABLE pub_test2.test_tbl2 (a int primary key, b int);
+
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_schemas;
+RESET client_min_messages;
+
+-- add tables from one schema, sequences from the other
+ALTER PUBLICATION testpub_schemas ADD TABLES IN SCHEMA pub_test2;
+ALTER PUBLICATION testpub_schemas ADD SEQUENCES IN SCHEMA pub_test1;
+
+\dRp+ testpub_schemas
+
+\dn+ pub_test1
+\dn+ pub_test2
+
+\d+ pub_test1.test_seq1;
+\d+ pub_test1.test_tbl1;
+
+\d+ pub_test2.test_seq2;
+\d+ pub_test2.test_tbl2;
+
+-- add the other object type from each schema
+ALTER PUBLICATION testpub_schemas ADD TABLES IN SCHEMA pub_test1;
+ALTER PUBLICATION testpub_schemas ADD SEQUENCES IN SCHEMA pub_test2;
+
+\dRp+ testpub_schemas
+
+\dn+ pub_test1
+\dn+ pub_test2
+
+\d+ pub_test1.test_seq1;
+\d+ pub_test1.test_tbl1;
+
+\d+ pub_test2.test_seq2;
+\d+ pub_test2.test_tbl2;
+
+-- now drop the object type added first
+ALTER PUBLICATION testpub_schemas DROP TABLES IN SCHEMA pub_test2;
+ALTER PUBLICATION testpub_schemas DROP SEQUENCES IN SCHEMA pub_test1;
+
+\dRp+ testpub_schemas
+
+\dn+ pub_test1
+\dn+ pub_test2
+
+\d+ pub_test1.test_seq1;
+\d+ pub_test1.test_tbl1;
+
+\d+ pub_test2.test_seq2;
+\d+ pub_test2.test_tbl2;
+
+-- add a different schema (not including the already published sequences)
+ALTER PUBLICATION testpub_schemas ADD TABLE pub_test2.test_tbl2;
+ALTER PUBLICATION testpub_schemas ADD SEQUENCE pub_test1.test_seq1;
+
+\dRp+ testpub_schemas
+
+\d+ pub_test1.test_seq1;
+\d+ pub_test1.test_tbl1;
+
+\d+ pub_test2.test_seq2;
+\d+ pub_test2.test_tbl2;
+
+-- now drop the explicitly added objects again
+ALTER PUBLICATION testpub_schemas DROP TABLE pub_test2.test_tbl2;
+ALTER PUBLICATION testpub_schemas DROP SEQUENCE pub_test1.test_seq1;
+
+\dRp+ testpub_schemas
+
+\d+ pub_test1.test_seq1;
+\d+ pub_test1.test_tbl1;
+
+\d+ pub_test2.test_seq2;
+\d+ pub_test2.test_tbl2;
+
+DROP PUBLICATION testpub_schemas;
+DROP TABLE pub_test1.test_tbl1, pub_test2.test_tbl2;
+DROP SEQUENCE pub_test1.test_seq1, pub_test2.test_seq2;
+DROP SCHEMA pub_test1, pub_test2;
+
-- Tests for partitioned tables
SET client_min_messages = 'ERROR';
CREATE PUBLICATION testpub_forparted;
@@ -1052,32 +1236,51 @@ CREATE SCHEMA sch1;
CREATE SCHEMA sch2;
CREATE TABLE sch1.tbl1 (a int) PARTITION BY RANGE(a);
CREATE TABLE sch2.tbl1_part1 PARTITION OF sch1.tbl1 FOR VALUES FROM (1) to (10);
+CREATE SEQUENCE sch1.seq1;
+CREATE SEQUENCE sch2.seq2;
-- Schema publication that does not include the schema that has the parent table
CREATE PUBLICATION pub FOR TABLES IN SCHEMA sch2 WITH (PUBLISH_VIA_PARTITION_ROOT=1);
+ALTER PUBLICATION pub ADD SEQUENCES IN SCHEMA sch2;
SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
DROP PUBLICATION pub;
-- Table publication that does not include the parent table
CREATE PUBLICATION pub FOR TABLE sch2.tbl1_part1 WITH (PUBLISH_VIA_PARTITION_ROOT=1);
+ALTER PUBLICATION pub ADD SEQUENCE sch2.seq2;
SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
-- Table publication that includes both the parent table and the child table
ALTER PUBLICATION pub ADD TABLE sch1.tbl1;
+ALTER PUBLICATION pub ADD SEQUENCE sch1.seq1;
SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
DROP PUBLICATION pub;
-- Schema publication that does not include the schema that has the parent table
CREATE PUBLICATION pub FOR TABLES IN SCHEMA sch2 WITH (PUBLISH_VIA_PARTITION_ROOT=0);
+ALTER PUBLICATION pub ADD SEQUENCE sch1.seq1;
+SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
+
+DROP PUBLICATION pub;
+-- Sequence publication
+CREATE PUBLICATION pub FOR SEQUENCE sch2.seq2;
SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
DROP PUBLICATION pub;
-- Table publication that does not include the parent table
CREATE PUBLICATION pub FOR TABLE sch2.tbl1_part1 WITH (PUBLISH_VIA_PARTITION_ROOT=0);
SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
-- Table publication that includes both the parent table and the child table
ALTER PUBLICATION pub ADD TABLE sch1.tbl1;
+ALTER PUBLICATION pub ADD SEQUENCES IN SCHEMA sch2;
SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
DROP PUBLICATION pub;
DROP TABLE sch2.tbl1_part1;
@@ -1090,10 +1293,36 @@ CREATE TABLE sch1.tbl1_part3 (a int) PARTITION BY RANGE(a);
ALTER TABLE sch1.tbl1 ATTACH PARTITION sch1.tbl1_part3 FOR VALUES FROM (20) to (30);
CREATE PUBLICATION pub FOR TABLES IN SCHEMA sch1 WITH (PUBLISH_VIA_PARTITION_ROOT=1);
SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
+
+DROP PUBLICATION pub;
+-- Schema publication
+CREATE PUBLICATION pub FOR SEQUENCE sch2.seq2;
+SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
+
+DROP PUBLICATION pub;
+-- Sequence publication
+CREATE PUBLICATION pub FOR SEQUENCES IN SCHEMA sch2;
+SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
+
+ALTER PUBLICATION pub ADD SEQUENCE sch1.seq1;
+SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
+
+ALTER PUBLICATION pub DROP SEQUENCE sch1.seq1;
+SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
+
+ALTER PUBLICATION pub ADD SEQUENCES IN SCHEMA sch1;
+SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
RESET client_min_messages;
DROP PUBLICATION pub;
DROP TABLE sch1.tbl1;
+DROP SEQUENCE sch1.seq1, sch2.seq2;
DROP SCHEMA sch1 cascade;
DROP SCHEMA sch2 cascade;
diff --git a/src/test/subscription/t/034_sequences.pl b/src/test/subscription/t/034_sequences.pl
new file mode 100644
index 00000000000..2e9317642fc
--- /dev/null
+++ b/src/test/subscription/t/034_sequences.pl
@@ -0,0 +1,344 @@
+
+# Copyright (c) 2021, PostgreSQL Global Development Group
+
+# This tests that sequences are replicated correctly by logical replication
+use strict;
+use warnings;
+use PostgreSQL::Test::Cluster;
+use PostgreSQL::Test::Utils;
+use Test::More;
+
+# Initialize publisher node
+my $node_publisher = PostgreSQL::Test::Cluster->new('publisher');
+$node_publisher->init(allows_streaming => 'logical');
+$node_publisher->start;
+
+# Create subscriber node
+my $node_subscriber = PostgreSQL::Test::Cluster->new('subscriber');
+$node_subscriber->init(allows_streaming => 'logical');
+$node_subscriber->start;
+
+# Create some preexisting content on publisher
+my $ddl = qq(
+ CREATE TABLE seq_test (v BIGINT);
+ CREATE SEQUENCE s;
+);
+
+# Setup structure on the publisher
+$node_publisher->safe_psql('postgres', $ddl);
+
+# Create some the same structure on subscriber, and an extra sequence that
+# we'll create on the publisher later
+$ddl = qq(
+ CREATE TABLE seq_test (v BIGINT);
+ CREATE SEQUENCE s;
+ CREATE SEQUENCE s2;
+);
+
+$node_subscriber->safe_psql('postgres', $ddl);
+
+# Setup logical replication
+my $publisher_connstr = $node_publisher->connstr . ' dbname=postgres';
+$node_publisher->safe_psql('postgres',
+ "CREATE PUBLICATION seq_pub");
+
+$node_publisher->safe_psql('postgres',
+ "ALTER PUBLICATION seq_pub ADD SEQUENCE s");
+
+# function that alters the sequence in a completed subtransaction
+$node_publisher->safe_psql('postgres',
+ "CREATE FUNCTION regress_alter_sequence(s regclass, v integer) RETURNS void as \$\$
+BEGIN
+EXECUTE format('alter sequence %s restart %s', s, v);
+EXCEPTION WHEN others THEN NULL;
+END; \$\$ LANGUAGE plpgsql");
+
+$node_subscriber->safe_psql('postgres',
+ "CREATE SUBSCRIPTION seq_sub CONNECTION '$publisher_connstr' PUBLICATION seq_pub"
+);
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Wait for initial sync to finish as well
+my $synced_query =
+ "SELECT count(1) = 0 FROM pg_subscription_rel WHERE srsubstate NOT IN ('s', 'r');";
+$node_subscriber->poll_query_until('postgres', $synced_query)
+ or die "Timed out while waiting for subscriber to synchronize data";
+
+# Insert initial test data
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ -- generate a number of values using the sequence
+ INSERT INTO seq_test SELECT nextval('s') FROM generate_series(1,100);
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+my $result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s;
+));
+
+is( $result, '132|0|t',
+ 'initial test data replicated');
+
+
+# advance the sequence in a rolled-back transaction - the rollback
+# does not wait for the replication, so we could see any intermediate state
+# so do something else after the test, to ensure we wait for everything
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ INSERT INTO seq_test SELECT nextval('s') FROM generate_series(1,100);
+ ROLLBACK;
+ INSERT INTO seq_test VALUES (-1);
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s;
+));
+
+is( $result, '231|0|t',
+ 'advance sequence in rolled-back transaction');
+
+
+# create a new sequence and roll it back - should not be replicated, due to
+# the transactional behavior
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ CREATE SEQUENCE s2;
+ ALTER PUBLICATION seq_pub ADD SEQUENCE s2;
+ INSERT INTO seq_test SELECT nextval('s2') FROM generate_series(1,100);
+ ROLLBACK;
+));
+
+# Refresh publication after sequence is added to publication
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ ALTER SUBSCRIPTION seq_sub REFRESH PUBLICATION
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s2;
+));
+
+is( $result, '1|0|f',
+ 'create new sequence and roll it back');
+
+
+# create a new sequence, advance it in a rolled-back transaction, but commit
+# the create - the advance should be replicated nevertheless
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ ALTER SEQUENCE s RESTART WITH 1000;
+ SAVEPOINT sp1;
+ INSERT INTO seq_test SELECT nextval('s') FROM generate_series(1,100);
+ ROLLBACK TO sp1;
+ COMMIT;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Wait for sync of the second sequence we just added to finish
+$synced_query =
+ "SELECT count(1) = 0 FROM pg_subscription_rel WHERE srsubstate NOT IN ('s', 'r');";
+$node_subscriber->poll_query_until('postgres', $synced_query)
+ or die "Timed out while waiting for subscriber to synchronize data";
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s;
+));
+
+is( $result, '1131|0|t',
+ 'create sequence, advance it in rolled-back transaction, but commit the create');
+
+
+# advance the new sequence in a transaction, and roll it back - the rollback
+# does not wait for the replication, so we could see any intermediate state
+# so do something else after the test, to ensure we wait for everything
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ INSERT INTO seq_test SELECT nextval('s') FROM generate_series(1,100);
+ ROLLBACK;
+ INSERT INTO seq_test VALUES (-1);
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s;
+));
+
+is( $result, '1230|0|t',
+ 'advance the new sequence in a transaction and roll it back');
+
+
+# advance the sequence in a subtransaction - the subtransaction gets rolled
+# back, but commit the main one - the changes should still be replicated
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ SAVEPOINT s1;
+ INSERT INTO seq_test SELECT nextval('s') FROM generate_series(1,100);
+ ROLLBACK TO s1;
+ COMMIT;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s;
+));
+
+is( $result, '1329|0|t',
+ 'advance sequence in a subtransaction');
+
+
+# alter the sequence twice in the same transaction
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ ALTER SEQUENCE s RESTART 1000;
+ SAVEPOINT s1;
+ ALTER SEQUENCE s RESTART 2000;
+ INSERT INTO seq_test SELECT nextval('s') FROM generate_series(1,40);
+ COMMIT;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s;
+));
+
+is( $result, '2065|0|t',
+ 'alter sequence twice in a subtransaction');
+
+
+# alter the sequence twice in the same transaction, rollback the second change
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ ALTER SEQUENCE s RESTART 1000;
+ SAVEPOINT s1;
+ ALTER SEQUENCE s RESTART 2000;
+ ROLLBACK TO s1;
+ INSERT INTO seq_test SELECT nextval('s') FROM generate_series(1,40);
+ COMMIT;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s;
+));
+
+is( $result, '1065|0|t',
+ 'alter sequence twice in a subtransaction');
+
+
+# alter the sequence in a subtransaction that commits, followed by advancing
+# the sequence in a subtransaction that gets rolled back, but then the main
+# one commits - only the second set of changes should be replicated, as it
+# belongs to the alter, which was in the committed part
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ SELECT regress_alter_sequence('s', 1000); -- reset to 1000
+ INSERT INTO seq_test SELECT nextval('s') FROM generate_series(1,100);
+ SAVEPOINT s1;
+ SELECT regress_alter_sequence('s', 2000); -- reset to 2000
+ INSERT INTO seq_test SELECT nextval('s') FROM generate_series(1,100);
+ ROLLBACK TO s1;
+ INSERT INTO seq_test SELECT nextval('s') FROM generate_series(1,100);
+ COMMIT;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s;
+));
+
+is( $result, '1230|0|t',
+ 'alter sequence in a subtransaction');
+
+
+# alter the sequence in a subtransaction that commits, followed by advancing
+# the sequence in another subtransactio. the changes should be replicated, as
+# it belongs to the second alter, and both alters commit
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ SELECT regress_alter_sequence('s', 1000); -- reset to 1000
+ SAVEPOINT s1;
+ SELECT regress_alter_sequence('s', 2000); -- reset to 2000
+ INSERT INTO seq_test SELECT nextval('s') FROM generate_series(1,100);
+ COMMIT;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s;
+));
+
+is( $result, '2131|0|t',
+ 'alter sequence in a subtransaction');
+
+
+# alter the sequence in a subtransaction that commits, followed by advancing
+# the sequence in a subtransaction that gets rolled back, but then the main
+# one commits - the changes should still be replicated, because they belong
+# to the alter, which was in the committed part
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ SELECT regress_alter_sequence('s', 1000); -- reset to 1000
+ INSERT INTO seq_test SELECT nextval('s') FROM generate_series(1,100);
+ SAVEPOINT s1;
+ INSERT INTO seq_test SELECT nextval('s') FROM generate_series(1,100);
+ ROLLBACK TO s1;
+ INSERT INTO seq_test SELECT nextval('s') FROM generate_series(1,100);
+ COMMIT;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s;
+));
+
+is( $result, '1329|0|t',
+ 'alter sequence in a subtransaction');
+
+
+done_testing();
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index f4efc0192d5..800f7bba3c4 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -1503,6 +1503,7 @@ LogicalRepRelId
LogicalRepRelMapEntry
LogicalRepRelation
LogicalRepRollbackPreparedTxnData
+LogicalRepSequence
LogicalRepStreamAbortData
LogicalRepTupleData
LogicalRepTyp
--
2.41.0
[text/x-patch] v20231203-0004-global-hash-table.patch (12.1K, ../../[email protected]/9-v20231203-0004-global-hash-table.patch)
download | inline diff:
From e646d7f07f402475df57ac1ce946bad4d6606290 Mon Sep 17 00:00:00 2001
From: Tomas Vondra <[email protected]>
Date: Sat, 2 Dec 2023 10:28:26 +0100
Subject: [PATCH v20231203 4/4] global hash table
---
.../replication/logical/reorderbuffer.c | 177 +++++++-----------
src/include/replication/reorderbuffer.h | 6 +
2 files changed, 74 insertions(+), 109 deletions(-)
diff --git a/src/backend/replication/logical/reorderbuffer.c b/src/backend/replication/logical/reorderbuffer.c
index bb16ca4f7b9..f7d9a26cfcb 100644
--- a/src/backend/replication/logical/reorderbuffer.c
+++ b/src/backend/replication/logical/reorderbuffer.c
@@ -415,6 +415,14 @@ ReorderBufferAllocate(void)
dlist_init(&buffer->txns_by_base_snapshot_lsn);
dclist_init(&buffer->catchange_txns);
+ /* Create hash of sequences, mapping relfilelocator to transaction. */
+ hash_ctl.keysize = sizeof(RelFileLocator);
+ hash_ctl.entrysize = sizeof(ReorderBufferSequenceEnt);
+ hash_ctl.hcxt = buffer->context;
+
+ buffer->sequences_hash = hash_create("ReorderBufferSequenceHash", 32, &hash_ctl,
+ HASH_ELEM | HASH_BLOBS | HASH_CONTEXT);
+
/*
* Ensure there's no stale data from prior uses of this slot, in case some
* prior exit avoided calling ReorderBufferFree. Failure to do this can
@@ -1015,45 +1023,27 @@ ReorderBufferSequenceIsTransactional(ReorderBuffer *rb,
RelFileLocator rlocator,
TransactionId *xid)
{
+ ReorderBufferSequenceEnt *entry;
bool found = false;
- dlist_iter iter;
AssertCheckSequences(rb);
+ entry = hash_search(rb->sequences_hash,
+ (void *) &rlocator,
+ HASH_FIND,
+ &found);
+
/*
- * Walk all top-level transactions (some of which may be subxacts, except
- * that we haven't processed the assignments yet), and check if any of
- * them created the relfilenode.
+ * If we found an entry with matching relfilenode, we're done - we
+ * have to treat the sequence change as transactional, and replay it
+ * in the same (sub)transaction just like any other change.
+ *
+ * Optionally set XID of the (sub)xact that created the relfilenode.
*/
- dlist_foreach(iter, &rb->toplevel_by_lsn)
+ if (found)
{
- ReorderBufferSequenceEnt *entry;
- ReorderBufferTXN *txn = dlist_container(ReorderBufferTXN, node,
- iter.cur);
-
- /* transaction has no relfilenodes at all */
- if (!txn->sequences_hash)
- continue;
-
- entry = hash_search(txn->sequences_hash,
- (void *) &rlocator,
- HASH_FIND,
- &found);
-
- /*
- * If we found an entry with matching relfilenode, we're done - we
- * have to treat the sequence change as transactional, and replay it
- * in the same (sub)transaction just like any other change.
- *
- * Optionally set XID of the (sub)xact that created the relfilenode.
- */
- if (found)
- {
- if (xid)
- *xid = entry->txn->xid;
-
- break;
- }
+ if (xid)
+ *xid = entry->txn->xid;
}
return found;
@@ -1068,27 +1058,22 @@ ReorderBufferSequenceIsTransactional(ReorderBuffer *rb,
* the aborted one.
*/
static void
-ReorderBufferSequenceCleanup(ReorderBufferTXN *txn)
+ReorderBufferSequenceCleanup(ReorderBuffer *rb, ReorderBufferTXN *txn)
{
HASH_SEQ_STATUS scan_status;
ReorderBufferSequenceEnt *ent;
- /* Bail out if not a subxact, or if there are no entries. */
- if (!rbtxn_is_known_subxact(txn))
- return;
-
+ /* Bail out if there are no sequence entries for this xact. */
if (!txn->sequences_hash)
return;
/*
- * Scan the top-level transaction hash and remove the entries from it. If
- * we have entries for subxact, the top-level hash must have been
- * initialized.
+ * Scan the global transaction hash and remove the entries from it.
*/
hash_seq_init(&scan_status, txn->sequences_hash);
while ((ent = (ReorderBufferSequenceEnt *) hash_seq_search(&scan_status)) != NULL)
{
- if (hash_search(txn->toptxn->sequences_hash,
+ if (hash_search(rb->sequences_hash,
(void *) &ent->rlocator,
HASH_REMOVE, NULL) == NULL)
elog(ERROR, "hash table corrupted");
@@ -1286,28 +1271,23 @@ ReorderBufferAddRelFileLocator(ReorderBuffer *rb, TransactionId xid,
txn = ReorderBufferTXNByXid(rb, xid, true, NULL, lsn, true);
/*
- * First add it to the top-level transaction, but make sure it links to
+ * First add it to the global hash table, and make sure it links to
* the correct subtransaction (so that we add later changes to it).
*/
- if (txn->toptxn)
- {
- /* make sure the hash table is initialized */
- ReorderBufferTXNSequencesInit(rb, txn->toptxn);
+ entry = hash_search(rb->sequences_hash,
+ (void *) &rlocator,
+ HASH_ENTER,
+ &found);
- /* search the lookup table */
- entry = hash_search(txn->toptxn->sequences_hash,
- (void *) &rlocator,
- HASH_ENTER,
- &found);
+ /*
+ * We've just decoded creation of the relfilenode, so if we found it
+ * in the hash table, something is wrong.
+ */
+ Assert(!found);
- /*
- * We've just decoded creation of the relfilenode, so if we found it
- * in the hash table, something is wrong.
- */
- Assert(!found);
+ entry->txn = txn;
- entry->txn = txn;
- }
+ /* now add it to the hash table in the transaction itself */
/* make sure the hash table is initialized */
ReorderBufferTXNSequencesInit(rb, txn);
@@ -1417,7 +1397,6 @@ AssertCheckSequences(ReorderBuffer *rb)
{
#ifdef USE_ASSERT_CHECKING
LogicalDecodingContext *ctx = rb->private_data;
- dlist_iter iter;
/*
* Skip the verification if we don't reach the LSN at which we start
@@ -1433,57 +1412,53 @@ AssertCheckSequences(ReorderBuffer *rb)
return;
/*
- * Make sure the relfilenodes from subxacts are properly recorded in the
- * top-level transaction hash table.
+ * Make sure the relfilenodes from all xacts are properly recorded in the
+ * global hash table of sequences.
*/
- dlist_foreach(iter, &rb->toplevel_by_lsn)
{
+ HASH_SEQ_STATUS hash_seq;
+ ReorderBufferTXNByIdEnt *ent;
int nentries = 0,
nsubentries = 0;
- dlist_iter subiter;
- ReorderBufferTXN *txn = dlist_container(ReorderBufferTXN, node,
- iter.cur);
/*
- * We don't skip top-level transactions without relfilenodes, because
- * there might be a subtransaction with some, and we want to detect
+ * We do this check even if the global hash table is empty, because
+ * there might be a transaction with some, and we want to detect
* such cases too.
*/
- if (txn->sequences_hash)
- nentries = hash_get_num_entries(txn->sequences_hash);
+ nentries = hash_get_num_entries(rb->sequences_hash);
- /* walk all subxacts, and check the hash table in each one */
- dlist_foreach(subiter, &txn->subtxns)
+ /* walk all xacts, and check the hash table in each one */
+ hash_seq_init(&hash_seq, rb->by_txn);
+ while ((ent = hash_seq_search(&hash_seq)) != NULL)
{
- HASH_SEQ_STATUS scan_status;
+ ReorderBufferTXN *txn = ent->txn;
ReorderBufferSequenceEnt *entry;
-
- ReorderBufferTXN *subtxn = dlist_container(ReorderBufferTXN, node,
- subiter.cur);
+ HASH_SEQ_STATUS scan_status;
/*
* If this subxact has no relfilenodes, skip it. We'll do the
- * check in the opposite direction (that all top-level
- * relfilenodes are in the correct subxact) later.
+ * check in the opposite direction (that all relfilenodes are
+ * also in hash for the correct xact) later.
*/
- if (!subtxn->sequences_hash)
+ if (!txn->sequences_hash)
continue;
- /* add number of relfilenodes in this subxact */
- nsubentries += hash_get_num_entries(subtxn->sequences_hash);
+ /* account for relfilenodes in this xact */
+ nsubentries += hash_get_num_entries(txn->sequences_hash);
/*
* Check that all subxact relfilenodes are in the top-level txn
* too, and are pointing to this subtransaction.
*/
- hash_seq_init(&scan_status, subtxn->sequences_hash);
+ hash_seq_init(&scan_status, txn->sequences_hash);
while ((entry = (ReorderBufferSequenceEnt *) hash_seq_search(&scan_status)) != NULL)
{
bool found = false;
ReorderBufferSequenceEnt *subentry;
- /* search for the same relfilenode in the top-level txn */
- subentry = hash_search(txn->sequences_hash,
+ /* search for the same relfilenode in the global hash table */
+ subentry = hash_search(rb->sequences_hash,
(void *) &entry->rlocator,
HASH_FIND,
&found);
@@ -1498,7 +1473,7 @@ AssertCheckSequences(ReorderBuffer *rb)
* The entry has to point to the subxact - there's no subxact
* "below" this one to which the relfilenode could belong.
*/
- Assert(subentry->txn == subtxn);
+ Assert(subentry->txn == txn);
}
}
@@ -1508,32 +1483,25 @@ AssertCheckSequences(ReorderBuffer *rb)
* directly, so this needs to be inequality.
*/
Assert(nentries >= nsubentries);
+ }
+ {
/*
* Now do the check in the opposite direction - check that every entry
- * in the top-level txn (except those pointing to the top-level txn
- * itself) point to one of the subxacts, and there's an entry in the
- * subxact hash.
+ * in the global hash table point to one of the running xacts, and
+ * there's an entry in the transaction hash table.
*/
- if (txn->sequences_hash)
+ if (rb->sequences_hash)
{
HASH_SEQ_STATUS scan_status;
ReorderBufferSequenceEnt *entry;
- hash_seq_init(&scan_status, txn->sequences_hash);
+ hash_seq_init(&scan_status, rb->sequences_hash);
while ((entry = (ReorderBufferSequenceEnt *) hash_seq_search(&scan_status)) != NULL)
{
bool found = false;
ReorderBufferSequenceEnt *subentry;
- /* Skip entries for the top-level txn itself. */
- if (entry->txn == txn)
- continue;
-
- /* Is it a subxact of this txn? */
- Assert(rbtxn_is_known_subxact(entry->txn));
- Assert(entry->txn->toptxn == txn);
-
/*
* Search for the same relfilenode in the subxact (it should
* be initialized, as we expect it to contain the
@@ -2181,6 +2149,9 @@ ReorderBufferCleanupTXN(ReorderBuffer *rb, ReorderBufferTXN *txn)
if (rbtxn_is_serialized(txn))
ReorderBufferRestoreCleanup(rb, txn);
+ /* remove the sequence relfilenode from the global hash table */
+ ReorderBufferSequenceCleanup(rb, txn);
+
/* deallocate */
ReorderBufferReturnTXN(rb, txn);
}
@@ -3519,9 +3490,6 @@ ReorderBufferFinishPrepared(ReorderBuffer *rb, TransactionId xid,
ReorderBufferExecuteInvalidations(txn->ninvalidations,
txn->invalidations);
- /* cleanup: remove sequence relfilenodes from the top-level txn */
- ReorderBufferSequenceCleanup(txn);
-
ReorderBufferCleanupTXN(rb, txn);
}
@@ -3571,9 +3539,6 @@ ReorderBufferAbort(ReorderBuffer *rb, TransactionId xid, XLogRecPtr lsn,
/* cosmetic... */
txn->final_lsn = lsn;
- /* remove sequence relfilenodes from the top-level txn */
- ReorderBufferSequenceCleanup(txn);
-
/* remove potential on-disk data, and deallocate */
ReorderBufferCleanupTXN(rb, txn);
@@ -3613,9 +3578,6 @@ ReorderBufferAbortOld(ReorderBuffer *rb, TransactionId oldestRunningXid)
if (rbtxn_is_streamed(txn))
rb->stream_abort(rb, txn, InvalidXLogRecPtr);
- /* remove sequence relfilenodes from the top-level txn */
- ReorderBufferSequenceCleanup(txn);
-
/* remove potential on-disk data, and deallocate this tx */
ReorderBufferCleanupTXN(rb, txn);
@@ -3668,9 +3630,6 @@ ReorderBufferForget(ReorderBuffer *rb, TransactionId xid, XLogRecPtr lsn)
else
Assert(txn->ninvalidations == 0);
- /* remove sequence relfilenodes from a top-level txn */
- ReorderBufferSequenceCleanup(txn);
-
/* remove potential on-disk data, and deallocate */
ReorderBufferCleanupTXN(rb, txn);
}
diff --git a/src/include/replication/reorderbuffer.h b/src/include/replication/reorderbuffer.h
index 58a99b74060..ad3a56d8f0f 100644
--- a/src/include/replication/reorderbuffer.h
+++ b/src/include/replication/reorderbuffer.h
@@ -704,6 +704,12 @@ struct ReorderBuffer
*/
int64 totalTxns; /* total number of transactions sent */
int64 totalBytes; /* total amount of data decoded */
+
+ /*
+ * Sequence relfilenodes created in any transaction (also includes
+ * altered sequences, which assigns new relfilenode).
+ */
+ HTAB *sequences_hash;
};
--
2.41.0
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2023-12-03 18:26 Tomas Vondra <[email protected]>
parent: Tomas Vondra <[email protected]>
2 siblings, 1 reply; 127+ messages in thread
From: Tomas Vondra @ 2023-12-03 18:26 UTC (permalink / raw)
To: Hayato Kuroda (Fujitsu) <[email protected]>; Amit Kapila <[email protected]>; +Cc: Zhijie Hou (Fujitsu) <[email protected]>; Ashutosh Bapat <[email protected]>; PostgreSQL Hackers <[email protected]>; Masahiko Sawada <[email protected]>; Peter Eisentraut <[email protected]>; Dilip Kumar <[email protected]>
On 12/3/23 18:52, Tomas Vondra wrote:
> ...
>
> Some time ago I floated the idea of maybe "queuing" the sequence changes
> and only replay them on the next commit, somehow. But we did ran into
> problems with which snapshot to use, that I didn't know how to solve.
> Maybe we should try again. The idea is we'd queue the non-transactional
> changes somewhere (can't be in the transaction, because we must keep
> them even if it aborts), and then "inject" them into the next commit.
> That'd mean we wouldn't do the separate start/abort for each change.
>
Another idea is that maybe we could somehow inform ReorderBuffer whether
the output plugin even is interested in sequences. That'd help with
cases where we don't even want/need to replicate sequences, e.g. because
the publication does not specify (publish=sequence).
What happens now in that case is we call ReorderBufferQueueSequence(),
it does the whole dance with starting/aborting the transaction, calls
rb->sequence() which just does "meh" and doesn't do anything. Maybe we
could just short-circuit this by asking the output plugin somehow.
In an extreme case the plugin may not even specify the sequence
callbacks, and we're still doing all of this.
regards
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2023-12-05 12:17 Amit Kapila <[email protected]>
parent: Tomas Vondra <[email protected]>
2 siblings, 0 replies; 127+ messages in thread
From: Amit Kapila @ 2023-12-05 12:17 UTC (permalink / raw)
To: Tomas Vondra <[email protected]>; +Cc: Hayato Kuroda (Fujitsu) <[email protected]>; Zhijie Hou (Fujitsu) <[email protected]>; Ashutosh Bapat <[email protected]>; PostgreSQL Hackers <[email protected]>; Masahiko Sawada <[email protected]>; Peter Eisentraut <[email protected]>; Dilip Kumar <[email protected]>
On Sun, Dec 3, 2023 at 11:22 PM Tomas Vondra
<[email protected]> wrote:
>
> Thanks for the script. Are you also measuring the time it takes to
> decode this using test_decoding?
>
> FWIW I did more comprehensive suite of tests over the weekend, with a
> couple more variations. I'm attaching the updated scripts, running it
> should be as simple as
>
> ./run.sh BRANCH TRANSACTIONS RUNS
>
> so perhaps
>
> ./run.sh master 1000 3
>
> to do 3 runs with 1000 transactions per client. And it'll run a bunch of
> combinations hard-coded in the script, and write the timings into a CSV
> file (with "master" in each row).
>
> I did this on two machines (i5 with 4 cores, xeon with 16/32 cores). I
> did this with current master, the basic patch (without the 0002 part),
> and then with the optimized approach (single global hash table, see the
> 0004 part). That's what master / patched / optimized in the results is.
>
> Interestingly enough, the i5 handled this much faster, it seems to be
> better in single-core tasks. The xeon is still running, so the results
> for "optimized" only have one run (out of 3), but shouldn't change much.
>
> Attached is also a table summarizing this, and visualizing the timing
> change (vs. master) in the last couple columns. Green is "faster" than
> master (but we don't really expect that), and "red" means slower than
> master (the more red, the slower).
>
> There results are grouped by script (see the attached .tgz), with either
> 32 or 96 clients (which does affect the timing, but not between master
> and patch). Some executions have no pg_sleep() calls, some have 0.001
> wait (but that doesn't seem to make much difference).
>
> Overall, I'd group the results into about three groups:
>
> 1) good cases [nextval, nextval-40, nextval-abort]
>
> These are cases that slow down a bit, but the slowdown is mostly within
> reasonable bounds (we're making the decoding to do more stuff, so it'd
> be a bit silly to require that extra work to make no impact). And I do
> think this is reasonable, because this is pretty much an extreme / worst
> case behavior. People don't really do just nextval() calls, without
> doing anything else. Not to mention doing aborts for 100% transactions.
>
> So in practice this is going to be within noise (and in those cases the
> results even show speedup, which seems a bit surprising). It's somewhat
> dependent on CPU too - on xeon there's hardly any regression.
>
>
> 2) nextval-40-abort
>
> Here the slowdown is clear, but I'd argue it generally falls in the same
> group as (1). Yes, I'd be happier if it didn't behave like this, but if
> someone can show me a practical workload affected by this ...
>
>
> 3) irrelevant cases [all the alters taking insane amounts of time]
>
> I absolutely refuse to care about these extreme cases where decoding
> 100k transactions takes 5-10 minutes (on i5), or up to 30 minutes (on
> xeon). If this was a problem for some practical workload, we'd have
> already heard about it I guess. And even if there was such workload, it
> wouldn't be up to this patch to fix that. There's clearly something
> misbehaving in the snapshot builder.
>
>
> I was hopeful the global hash table would be an improvement, but that
> doesn't seem to be the case. I haven't done much profiling yet, but I'd
> guess most of the overhead is due to ReorderBufferQueueSequence()
> starting and aborting a transaction in the non-transactinal case. Which
> is unfortunate, but I don't know if there's a way to optimize that.
>
Before discussing the alternative ideas you shared, let me try to
clarify my understanding so that we are on the same page. I see two
observations based on the testing and discussion we had (a) for
non-transactional cases, the overhead observed is mainly due to
starting/aborting a transaction for each change; (b) for transactional
cases, we see overhead due to traversing all the top-level txns and
check the hash table for each one to find whether change is
transactional.
Am, I missing something?
--
With Regards,
Amit Kapila.
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2023-12-06 05:42 Dilip Kumar <[email protected]>
parent: Tomas Vondra <[email protected]>
2 siblings, 2 replies; 127+ messages in thread
From: Dilip Kumar @ 2023-12-06 05:42 UTC (permalink / raw)
To: Tomas Vondra <[email protected]>; +Cc: Hayato Kuroda (Fujitsu) <[email protected]>; Amit Kapila <[email protected]>; Zhijie Hou (Fujitsu) <[email protected]>; Ashutosh Bapat <[email protected]>; PostgreSQL Hackers <[email protected]>; Masahiko Sawada <[email protected]>; Peter Eisentraut <[email protected]>
On Sun, Dec 3, 2023 at 11:22 PM Tomas Vondra
<[email protected]> wrote:
>
> Some time ago I floated the idea of maybe "queuing" the sequence changes
> and only replay them on the next commit, somehow. But we did ran into
> problems with which snapshot to use, that I didn't know how to solve.
> Maybe we should try again. The idea is we'd queue the non-transactional
> changes somewhere (can't be in the transaction, because we must keep
> them even if it aborts), and then "inject" them into the next commit.
> That'd mean we wouldn't do the separate start/abort for each change.
Why can't we use the same concept of
SnapBuildDistributeNewCatalogSnapshot(), I mean we keep queuing the
non-transactional changes (have some base snapshot before the first
change), and whenever there is any catalog change, queue new snapshot
change also in the queue of the non-transactional sequence change so
that while sending it to downstream whenever it is necessary we will
change the historic snapshot?
--
Regards,
Dilip Kumar
EnterpriseDB: http://www.enterprisedb.com
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2023-12-06 09:05 Dilip Kumar <[email protected]>
parent: Dilip Kumar <[email protected]>
1 sibling, 1 reply; 127+ messages in thread
From: Dilip Kumar @ 2023-12-06 09:05 UTC (permalink / raw)
To: Tomas Vondra <[email protected]>; +Cc: Hayato Kuroda (Fujitsu) <[email protected]>; Amit Kapila <[email protected]>; Zhijie Hou (Fujitsu) <[email protected]>; Ashutosh Bapat <[email protected]>; PostgreSQL Hackers <[email protected]>; Masahiko Sawada <[email protected]>; Peter Eisentraut <[email protected]>
On Wed, Dec 6, 2023 at 11:12 AM Dilip Kumar <[email protected]> wrote:
>
> On Sun, Dec 3, 2023 at 11:22 PM Tomas Vondra
> <[email protected]> wrote:
> >
I was also wondering what happens if the sequence changes are
transactional but somehow the snap builder state changes to
SNAPBUILD_FULL_SNAPSHOT in between processing of the smgr_decode() and
the seq_decode() which means RelFileLocator will not be added to the
hash table and during the seq_decode() we will consider the change as
non-transactional. I haven't fully analyzed that what is the real
problem in this case but have we considered this case? what happens if
the transaction having both ALTER SEQUENCE and nextval() gets aborted
but the nextva() has been considered as non-transactional because
smgr_decode() changes were not processed because snap builder state
was not yet SNAPBUILD_FULL_SNAPSHOT.
--
Regards,
Dilip Kumar
EnterpriseDB: http://www.enterprisedb.com
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2023-12-06 10:05 Amit Kapila <[email protected]>
parent: Dilip Kumar <[email protected]>
1 sibling, 1 reply; 127+ messages in thread
From: Amit Kapila @ 2023-12-06 10:05 UTC (permalink / raw)
To: Dilip Kumar <[email protected]>; +Cc: Tomas Vondra <[email protected]>; Hayato Kuroda (Fujitsu) <[email protected]>; Zhijie Hou (Fujitsu) <[email protected]>; Ashutosh Bapat <[email protected]>; PostgreSQL Hackers <[email protected]>; Masahiko Sawada <[email protected]>; Peter Eisentraut <[email protected]>
On Wed, Dec 6, 2023 at 11:12 AM Dilip Kumar <[email protected]> wrote:
>
> On Sun, Dec 3, 2023 at 11:22 PM Tomas Vondra
> <[email protected]> wrote:
> >
>
> > Some time ago I floated the idea of maybe "queuing" the sequence changes
> > and only replay them on the next commit, somehow. But we did ran into
> > problems with which snapshot to use, that I didn't know how to solve.
> > Maybe we should try again. The idea is we'd queue the non-transactional
> > changes somewhere (can't be in the transaction, because we must keep
> > them even if it aborts), and then "inject" them into the next commit.
> > That'd mean we wouldn't do the separate start/abort for each change.
>
> Why can't we use the same concept of
> SnapBuildDistributeNewCatalogSnapshot(), I mean we keep queuing the
> non-transactional changes (have some base snapshot before the first
> change), and whenever there is any catalog change, queue new snapshot
> change also in the queue of the non-transactional sequence change so
> that while sending it to downstream whenever it is necessary we will
> change the historic snapshot?
>
Oh, do you mean maintain different historic snapshots and then switch
based on the change we are processing? I guess the other thing we need
to consider is the order of processing the changes if we maintain
separate queues that need to be processed.
--
With Regards,
Amit Kapila.
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2023-12-06 10:19 Amit Kapila <[email protected]>
parent: Tomas Vondra <[email protected]>
0 siblings, 1 reply; 127+ messages in thread
From: Amit Kapila @ 2023-12-06 10:19 UTC (permalink / raw)
To: Tomas Vondra <[email protected]>; +Cc: Hayato Kuroda (Fujitsu) <[email protected]>; Zhijie Hou (Fujitsu) <[email protected]>; Ashutosh Bapat <[email protected]>; PostgreSQL Hackers <[email protected]>; Masahiko Sawada <[email protected]>; Peter Eisentraut <[email protected]>; Dilip Kumar <[email protected]>
On Sun, Dec 3, 2023 at 11:56 PM Tomas Vondra
<[email protected]> wrote:
>
> On 12/3/23 18:52, Tomas Vondra wrote:
> > ...
> >
>
> Another idea is that maybe we could somehow inform ReorderBuffer whether
> the output plugin even is interested in sequences. That'd help with
> cases where we don't even want/need to replicate sequences, e.g. because
> the publication does not specify (publish=sequence).
>
> What happens now in that case is we call ReorderBufferQueueSequence(),
> it does the whole dance with starting/aborting the transaction, calls
> rb->sequence() which just does "meh" and doesn't do anything. Maybe we
> could just short-circuit this by asking the output plugin somehow.
>
> In an extreme case the plugin may not even specify the sequence
> callbacks, and we're still doing all of this.
>
We could explore this but I guess it won't solve the problem we are
facing in cases where all sequences are published and plugin has
specified the sequence callbacks. I think it would add some overhead
of this check in positive cases where we decide to anyway do send the
changes.
--
With Regards,
Amit Kapila.
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2023-12-06 11:05 Dilip Kumar <[email protected]>
parent: Amit Kapila <[email protected]>
0 siblings, 1 reply; 127+ messages in thread
From: Dilip Kumar @ 2023-12-06 11:05 UTC (permalink / raw)
To: Amit Kapila <[email protected]>; +Cc: Tomas Vondra <[email protected]>; Hayato Kuroda (Fujitsu) <[email protected]>; Zhijie Hou (Fujitsu) <[email protected]>; Ashutosh Bapat <[email protected]>; PostgreSQL Hackers <[email protected]>; Masahiko Sawada <[email protected]>; Peter Eisentraut <[email protected]>
On Wed, Dec 6, 2023 at 3:36 PM Amit Kapila <[email protected]> wrote:
>
> > Why can't we use the same concept of
> > SnapBuildDistributeNewCatalogSnapshot(), I mean we keep queuing the
> > non-transactional changes (have some base snapshot before the first
> > change), and whenever there is any catalog change, queue new snapshot
> > change also in the queue of the non-transactional sequence change so
> > that while sending it to downstream whenever it is necessary we will
> > change the historic snapshot?
> >
>
> Oh, do you mean maintain different historic snapshots and then switch
> based on the change we are processing? I guess the other thing we need
> to consider is the order of processing the changes if we maintain
> separate queues that need to be processed.
I mean we will not specifically maintain the historic changes, but if
there is any catalog change where we are pushing the snapshot to all
the transaction's change queue, at the same time we will push this
snapshot in the non-transactional sequence queue as well. I am not
sure what is the problem with the ordering? because we will be
queueing all non-transactional sequence changes in a separate queue in
the order they arrive and as soon as we process the next commit we
will process all the non-transactional changes at that time. Do you
see issue with that?
--
Regards,
Dilip Kumar
EnterpriseDB: http://www.enterprisedb.com
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2023-12-06 13:39 Tomas Vondra <[email protected]>
parent: Dilip Kumar <[email protected]>
0 siblings, 1 reply; 127+ messages in thread
From: Tomas Vondra @ 2023-12-06 13:39 UTC (permalink / raw)
To: Dilip Kumar <[email protected]>; +Cc: Hayato Kuroda (Fujitsu) <[email protected]>; Amit Kapila <[email protected]>; Zhijie Hou (Fujitsu) <[email protected]>; Ashutosh Bapat <[email protected]>; PostgreSQL Hackers <[email protected]>; Masahiko Sawada <[email protected]>; Peter Eisentraut <[email protected]>
On 12/6/23 10:05, Dilip Kumar wrote:
> On Wed, Dec 6, 2023 at 11:12 AM Dilip Kumar <[email protected]> wrote:
>>
>> On Sun, Dec 3, 2023 at 11:22 PM Tomas Vondra
>> <[email protected]> wrote:
>>>
>
> I was also wondering what happens if the sequence changes are
> transactional but somehow the snap builder state changes to
> SNAPBUILD_FULL_SNAPSHOT in between processing of the smgr_decode() and
> the seq_decode() which means RelFileLocator will not be added to the
> hash table and during the seq_decode() we will consider the change as
> non-transactional. I haven't fully analyzed that what is the real
> problem in this case but have we considered this case? what happens if
> the transaction having both ALTER SEQUENCE and nextval() gets aborted
> but the nextva() has been considered as non-transactional because
> smgr_decode() changes were not processed because snap builder state
> was not yet SNAPBUILD_FULL_SNAPSHOT.
>
Yes, if something like this happens, that'd be a problem:
1) decoding starts, with
SnapBuildCurrentState(builder) < SNAPBUILD_FULL_SNAPSHOT
2) transaction that creates a new refilenode gets decoded, but we skip
it because we don't have the correct snapshot
3) snapshot changes to SNAPBUILD_FULL_SNAPSHOT
4) we decode sequence change from nextval() for the sequence
This would lead to us attempting to apply sequence change for a
relfilenode that's not visible yet (and may even get aborted).
But can this even happen? Can we start decoding in the middle of a
transaction? How come this wouldn't affect e.g. XLOG_HEAP2_NEW_CID,
which is also skipped until SNAPBUILD_FULL_SNAPSHOT. Or logical
messages, where we also call the output plugin in non-transactional cases.
regards
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2023-12-06 13:47 Tomas Vondra <[email protected]>
parent: Dilip Kumar <[email protected]>
0 siblings, 2 replies; 127+ messages in thread
From: Tomas Vondra @ 2023-12-06 13:47 UTC (permalink / raw)
To: Dilip Kumar <[email protected]>; Amit Kapila <[email protected]>; +Cc: Hayato Kuroda (Fujitsu) <[email protected]>; Zhijie Hou (Fujitsu) <[email protected]>; Ashutosh Bapat <[email protected]>; PostgreSQL Hackers <[email protected]>; Masahiko Sawada <[email protected]>; Peter Eisentraut <[email protected]>
On 12/6/23 12:05, Dilip Kumar wrote:
> On Wed, Dec 6, 2023 at 3:36 PM Amit Kapila <[email protected]> wrote:
>>
>>> Why can't we use the same concept of
>>> SnapBuildDistributeNewCatalogSnapshot(), I mean we keep queuing the
>>> non-transactional changes (have some base snapshot before the first
>>> change), and whenever there is any catalog change, queue new snapshot
>>> change also in the queue of the non-transactional sequence change so
>>> that while sending it to downstream whenever it is necessary we will
>>> change the historic snapshot?
>>>
>>
>> Oh, do you mean maintain different historic snapshots and then switch
>> based on the change we are processing? I guess the other thing we need
>> to consider is the order of processing the changes if we maintain
>> separate queues that need to be processed.
>
> I mean we will not specifically maintain the historic changes, but if
> there is any catalog change where we are pushing the snapshot to all
> the transaction's change queue, at the same time we will push this
> snapshot in the non-transactional sequence queue as well. I am not
> sure what is the problem with the ordering? because we will be
> queueing all non-transactional sequence changes in a separate queue in
> the order they arrive and as soon as we process the next commit we
> will process all the non-transactional changes at that time. Do you
> see issue with that?
>
Isn't this (in principle) the idea of queuing the non-transactional
changes and then applying them on the next commit? Yes, I didn't get
very far with that, but I got stuck exactly on tracking which snapshot
to use, so if there's a way to do that, that'd fix my issue.
Also, would this mean we don't need to track the relfilenodes, if we're
able to query the catalog? Would we be able to check if the relfilenode
was created by the current xact?
regards
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2023-12-06 13:50 Tomas Vondra <[email protected]>
parent: Amit Kapila <[email protected]>
0 siblings, 1 reply; 127+ messages in thread
From: Tomas Vondra @ 2023-12-06 13:50 UTC (permalink / raw)
To: Amit Kapila <[email protected]>; +Cc: Hayato Kuroda (Fujitsu) <[email protected]>; Zhijie Hou (Fujitsu) <[email protected]>; Ashutosh Bapat <[email protected]>; PostgreSQL Hackers <[email protected]>; Masahiko Sawada <[email protected]>; Peter Eisentraut <[email protected]>; Dilip Kumar <[email protected]>
On 12/6/23 11:19, Amit Kapila wrote:
> On Sun, Dec 3, 2023 at 11:56 PM Tomas Vondra
> <[email protected]> wrote:
>>
>> On 12/3/23 18:52, Tomas Vondra wrote:
>>> ...
>>>
>>
>> Another idea is that maybe we could somehow inform ReorderBuffer whether
>> the output plugin even is interested in sequences. That'd help with
>> cases where we don't even want/need to replicate sequences, e.g. because
>> the publication does not specify (publish=sequence).
>>
>> What happens now in that case is we call ReorderBufferQueueSequence(),
>> it does the whole dance with starting/aborting the transaction, calls
>> rb->sequence() which just does "meh" and doesn't do anything. Maybe we
>> could just short-circuit this by asking the output plugin somehow.
>>
>> In an extreme case the plugin may not even specify the sequence
>> callbacks, and we're still doing all of this.
>>
>
> We could explore this but I guess it won't solve the problem we are
> facing in cases where all sequences are published and plugin has
> specified the sequence callbacks. I think it would add some overhead
> of this check in positive cases where we decide to anyway do send the
> changes.
Well, the idea is the check would be very simple (essentially just a
boolean flag somewhere), so not really measurable.
And if the plugin requests decoding sequences, I guess it's natural it
may have a bit of overhead. It needs to do more things, after all. It
needs to be acceptable, ofc.
regards
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2023-12-07 04:03 Amit Kapila <[email protected]>
parent: Tomas Vondra <[email protected]>
0 siblings, 0 replies; 127+ messages in thread
From: Amit Kapila @ 2023-12-07 04:03 UTC (permalink / raw)
To: Tomas Vondra <[email protected]>; +Cc: Hayato Kuroda (Fujitsu) <[email protected]>; Zhijie Hou (Fujitsu) <[email protected]>; Ashutosh Bapat <[email protected]>; PostgreSQL Hackers <[email protected]>; Masahiko Sawada <[email protected]>; Peter Eisentraut <[email protected]>; Dilip Kumar <[email protected]>
On Wed, Dec 6, 2023 at 7:20 PM Tomas Vondra
<[email protected]> wrote:
>
> On 12/6/23 11:19, Amit Kapila wrote:
> > On Sun, Dec 3, 2023 at 11:56 PM Tomas Vondra
> > <[email protected]> wrote:
> >>
> >> On 12/3/23 18:52, Tomas Vondra wrote:
> >>> ...
> >>>
> >>
> >> Another idea is that maybe we could somehow inform ReorderBuffer whether
> >> the output plugin even is interested in sequences. That'd help with
> >> cases where we don't even want/need to replicate sequences, e.g. because
> >> the publication does not specify (publish=sequence).
> >>
> >> What happens now in that case is we call ReorderBufferQueueSequence(),
> >> it does the whole dance with starting/aborting the transaction, calls
> >> rb->sequence() which just does "meh" and doesn't do anything. Maybe we
> >> could just short-circuit this by asking the output plugin somehow.
> >>
> >> In an extreme case the plugin may not even specify the sequence
> >> callbacks, and we're still doing all of this.
> >>
> >
> > We could explore this but I guess it won't solve the problem we are
> > facing in cases where all sequences are published and plugin has
> > specified the sequence callbacks. I think it would add some overhead
> > of this check in positive cases where we decide to anyway do send the
> > changes.
>
> Well, the idea is the check would be very simple (essentially just a
> boolean flag somewhere), so not really measurable.
>
> And if the plugin requests decoding sequences, I guess it's natural it
> may have a bit of overhead. It needs to do more things, after all. It
> needs to be acceptable, ofc.
>
I agree with you that if it can be done cheaply or without a
measurable overhead then it would be a good idea and can serve other
purposes as well. For example, see discussion [1]. I had more of what
the patch in email [1] is doing where it needs to start/stop xact and
do so relcache access etc. which seems can add some overhead if done
for each change, though I haven't measured so can't be sure.
[1] - https://www.postgresql.org/message-id/CAGfChW5Qo2SrjJ7rU9YYtZbRaWv6v-Z8MJn%3DdQNx4uCSqDEOHA%40mail.g...
--
With Regards,
Amit Kapila.
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2023-12-07 04:35 Dilip Kumar <[email protected]>
parent: Tomas Vondra <[email protected]>
1 sibling, 0 replies; 127+ messages in thread
From: Dilip Kumar @ 2023-12-07 04:35 UTC (permalink / raw)
To: Tomas Vondra <[email protected]>; +Cc: Amit Kapila <[email protected]>; Hayato Kuroda (Fujitsu) <[email protected]>; Zhijie Hou (Fujitsu) <[email protected]>; Ashutosh Bapat <[email protected]>; PostgreSQL Hackers <[email protected]>; Masahiko Sawada <[email protected]>; Peter Eisentraut <[email protected]>
On Wed, Dec 6, 2023 at 7:17 PM Tomas Vondra
<[email protected]> wrote:
>
> On 12/6/23 12:05, Dilip Kumar wrote:
> > On Wed, Dec 6, 2023 at 3:36 PM Amit Kapila <[email protected]> wrote:
> >>
> >>> Why can't we use the same concept of
> >>> SnapBuildDistributeNewCatalogSnapshot(), I mean we keep queuing the
> >>> non-transactional changes (have some base snapshot before the first
> >>> change), and whenever there is any catalog change, queue new snapshot
> >>> change also in the queue of the non-transactional sequence change so
> >>> that while sending it to downstream whenever it is necessary we will
> >>> change the historic snapshot?
> >>>
> >>
> >> Oh, do you mean maintain different historic snapshots and then switch
> >> based on the change we are processing? I guess the other thing we need
> >> to consider is the order of processing the changes if we maintain
> >> separate queues that need to be processed.
> >
> > I mean we will not specifically maintain the historic changes, but if
> > there is any catalog change where we are pushing the snapshot to all
> > the transaction's change queue, at the same time we will push this
> > snapshot in the non-transactional sequence queue as well. I am not
> > sure what is the problem with the ordering? because we will be
> > queueing all non-transactional sequence changes in a separate queue in
> > the order they arrive and as soon as we process the next commit we
> > will process all the non-transactional changes at that time. Do you
> > see issue with that?
> >
>
> Isn't this (in principle) the idea of queuing the non-transactional
> changes and then applying them on the next commit?
Yes, it is.
Yes, I didn't get
> very far with that, but I got stuck exactly on tracking which snapshot
> to use, so if there's a way to do that, that'd fix my issue.
Thinking more about the snapshot issue do we need to even bother about
changing the snapshot at all while streaming the non-transactional
sequence changes or we can send all the non-transactional changes with
a single snapshot? So mainly snapshot logically gets changed due to
these 2 events case1: When any transaction gets committed which has
done catalog operation (this changes the global snapshot) and case2:
When within a transaction, there is some catalog change (this just
updates the 'curcid' in the base snapshot of the transaction).
Now, if we are thinking that we are streaming all the
non-transactional sequence changes right before the next commit then
we are not bothered about the (case1) at all because all changes we
have queues so far are before this commit. And if we come to a
(case2), if we are performing any catalog change on the sequence then
the following changes on the same sequence will be considered
transactional and if the changes are just on some other catalog (not
relevant to our sequence operation) then also we should not be worried
about command_id change because visibility of catalog lookup for our
sequence will be unaffected by this.
In short, I am trying to say that we can safely queue the
non-transactional sequence changes and stream them based on the
snapshot we got when we decode the first change, and as long as we are
planning to stream just before the next commit (or next in-progress
stream), we don't ever need to update the snapshot.
> Also, would this mean we don't need to track the relfilenodes, if we're
> able to query the catalog? Would we be able to check if the relfilenode
> was created by the current xact?
I think by querying the catalog and checking the xmin we should be
able to figure that out, but isn't that costlier than looking up the
relfilenode in hash? Because just for identifying whether the changes
are transactional or non-transactional you would have to query the
catalog, that means for each change before we decide whether we add to
the transaction's change queue or non-transactional change queue we
will have to query the catalog i.e. you will have to start/stop the
transaction?
--
Regards,
Dilip Kumar
EnterpriseDB: http://www.enterprisedb.com
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2023-12-07 04:56 Amit Kapila <[email protected]>
parent: Tomas Vondra <[email protected]>
1 sibling, 1 reply; 127+ messages in thread
From: Amit Kapila @ 2023-12-07 04:56 UTC (permalink / raw)
To: Tomas Vondra <[email protected]>; +Cc: Dilip Kumar <[email protected]>; Hayato Kuroda (Fujitsu) <[email protected]>; Zhijie Hou (Fujitsu) <[email protected]>; Ashutosh Bapat <[email protected]>; PostgreSQL Hackers <[email protected]>; Masahiko Sawada <[email protected]>; Peter Eisentraut <[email protected]>
On Wed, Dec 6, 2023 at 7:17 PM Tomas Vondra
<[email protected]> wrote:
>
> On 12/6/23 12:05, Dilip Kumar wrote:
> > On Wed, Dec 6, 2023 at 3:36 PM Amit Kapila <[email protected]> wrote:
> >>
> >>> Why can't we use the same concept of
> >>> SnapBuildDistributeNewCatalogSnapshot(), I mean we keep queuing the
> >>> non-transactional changes (have some base snapshot before the first
> >>> change), and whenever there is any catalog change, queue new snapshot
> >>> change also in the queue of the non-transactional sequence change so
> >>> that while sending it to downstream whenever it is necessary we will
> >>> change the historic snapshot?
> >>>
> >>
> >> Oh, do you mean maintain different historic snapshots and then switch
> >> based on the change we are processing? I guess the other thing we need
> >> to consider is the order of processing the changes if we maintain
> >> separate queues that need to be processed.
> >
> > I mean we will not specifically maintain the historic changes, but if
> > there is any catalog change where we are pushing the snapshot to all
> > the transaction's change queue, at the same time we will push this
> > snapshot in the non-transactional sequence queue as well. I am not
> > sure what is the problem with the ordering?
> >
Currently, we set up the historic snapshot before starting a
transaction to process the change and then adapt the updates to it
while processing the changes for the transaction. Now, while
processing this new queue of non-transactional sequence messages, we
probably need a separate snapshot and updates to it. So, either we
need some sort of switching between snapshots or do it in different
transactions.
> > because we will be
> > queueing all non-transactional sequence changes in a separate queue in
> > the order they arrive and as soon as we process the next commit we
> > will process all the non-transactional changes at that time. Do you
> > see issue with that?
> >
>
> Isn't this (in principle) the idea of queuing the non-transactional
> changes and then applying them on the next commit? Yes, I didn't get
> very far with that, but I got stuck exactly on tracking which snapshot
> to use, so if there's a way to do that, that'd fix my issue.
>
> Also, would this mean we don't need to track the relfilenodes, if we're
> able to query the catalog? Would we be able to check if the relfilenode
> was created by the current xact?
>
I thought this new mechanism was for processing a queue of
non-transactional sequence changes. The tracking of relfilenodes is to
distinguish between transactional and non-transactional messages, so I
think we probably still need that.
--
With Regards,
Amit Kapila.
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2023-12-07 05:11 Dilip Kumar <[email protected]>
parent: Tomas Vondra <[email protected]>
0 siblings, 1 reply; 127+ messages in thread
From: Dilip Kumar @ 2023-12-07 05:11 UTC (permalink / raw)
To: Tomas Vondra <[email protected]>; +Cc: Hayato Kuroda (Fujitsu) <[email protected]>; Amit Kapila <[email protected]>; Zhijie Hou (Fujitsu) <[email protected]>; Ashutosh Bapat <[email protected]>; PostgreSQL Hackers <[email protected]>; Masahiko Sawada <[email protected]>; Peter Eisentraut <[email protected]>
On Wed, Dec 6, 2023 at 7:09 PM Tomas Vondra
<[email protected]> wrote:
>
> Yes, if something like this happens, that'd be a problem:
>
> 1) decoding starts, with
>
> SnapBuildCurrentState(builder) < SNAPBUILD_FULL_SNAPSHOT
>
> 2) transaction that creates a new refilenode gets decoded, but we skip
> it because we don't have the correct snapshot
>
> 3) snapshot changes to SNAPBUILD_FULL_SNAPSHOT
>
> 4) we decode sequence change from nextval() for the sequence
>
> This would lead to us attempting to apply sequence change for a
> relfilenode that's not visible yet (and may even get aborted).
>
> But can this even happen? Can we start decoding in the middle of a
> transaction? How come this wouldn't affect e.g. XLOG_HEAP2_NEW_CID,
> which is also skipped until SNAPBUILD_FULL_SNAPSHOT. Or logical
> messages, where we also call the output plugin in non-transactional cases.
It's not a problem for logical messages because whether the message is
transaction or non-transactional is decided while WAL logs the message
itself. But here our problem starts with deciding whether the change
is transactional vs non-transactional, because if we insert the
'relfilenode' in hash then the subsequent sequence change in the same
transaction would be considered transactional otherwise
non-transactional. And XLOG_HEAP2_NEW_CID is just for changing the
snapshot->curcid which will only affect the catalog visibility of the
upcoming operation in the same transaction, but that's not an issue
because if some of the changes of this transaction are seen when
snapbuild state < SNAPBUILD_FULL_SNAPSHOT then this transaction has to
get committed before the state change to SNAPBUILD_CONSISTENT_SNAPSHOT
i.e. the commit LSN of this transaction is going to be <
start_decoding_at.
--
Regards,
Dilip Kumar
EnterpriseDB: http://www.enterprisedb.com
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2023-12-12 10:01 Tomas Vondra <[email protected]>
parent: Amit Kapila <[email protected]>
0 siblings, 1 reply; 127+ messages in thread
From: Tomas Vondra @ 2023-12-12 10:01 UTC (permalink / raw)
To: Amit Kapila <[email protected]>; +Cc: Dilip Kumar <[email protected]>; Hayato Kuroda (Fujitsu) <[email protected]>; Zhijie Hou (Fujitsu) <[email protected]>; Ashutosh Bapat <[email protected]>; PostgreSQL Hackers <[email protected]>; Masahiko Sawada <[email protected]>; Peter Eisentraut <[email protected]>
Hi,
There's been a lot discussed over the past month or so, and it's become
difficult to get a good idea what's the current state - what issues
remain to be solved, what's unrelated to this patch, and how to move if
forward. Long-running threads tend to be confusing, so I had a short
call with Amit to discuss the current state yesterday, and to make sure
we're on the same page. I believe it was very helpful, and I've promised
to post a short summary of the call - issues, what we agreed seems like
a path forward, etc.
Obviously, I might have misunderstood something, in which case Amit can
correct me. And I'd certainly welcome opinions from others.
In general, we discussed three areas - desirability of the feature,
correctness and performance. I believe a brief summary of the agreement
would be this:
- desirability of the feature: Random IDs (UUIDs etc.) are likely a much
better solution for distributed (esp. active-active) systems. But there
are important use cases that are likely to keep using regular sequences
(online upgrades of single-node instances, existing systems, ...).
- correctness: There's one possible correctness issue, when the snapshot
changes to FULL between record creating a sequence relfilenode and that
sequence advancing. This needs to be verified/reproduced, and fixed.
- performance issues: We've agreed the case with a lot of aborts (when
DecodeCommit consumes a lot of CPU) is unrelated to this patch. We've
discussed whether the overhead with many sequence changes (nextval-40)
is acceptable, and/or how to improve it.
Next, I'll go over these points in more details, with my understanding
of what the challenges are, possible solutions etc. Most of this was
discussed/agreed on the call, but some are ideas I had only after the
call when writing this summary.
1) desirability of the feature
Firstly, do we actually want/need this feature? I believe that's very
much a question of what use cases we're targeting.
If we only focus on distributed databases (particularly those with
multiple active nodes), then we probably agree that the right solution
is to not use sequences (~generators of incrementing values) but UUIDs
or similar random identifiers (better not call them sequences, there's
not much sequential about it). The huge advantage is this does not
require replicating any state between the nodes, so logical decoding can
simply ignore them and replicate just the generated values. I don't
think there's any argument about that. If I as building such distributed
system, I'd certainly use such random IDs.
The question is what to do about the other use cases - online upgrades
relying on logical decoding, failovers to logical replicas, and so on.
Or what to do about existing systems that can't be easily changed to use
different/random identifiers. Those are not really distributed systems
and therefore don't quite need random IDs.
Furthermore, it's not like random IDs have no drawbacks - UUIDv4 can
easily lead to massive write amplification, for example. There are
variants like UUIDv7 reducing the impact, but there's other stuff.
My takeaway from this is there's still value in having this feature.
2) correctness
The only correctness issue I'm aware of is the question what happens
when the snapshot switches to SNAPBUILD_FULL_SNAPSHOT between decoding
the relfilenode creation and the sequence increment, pointed out by
Dilip in [1].
If this happens (and while I don't have a reproducer, I also don't have
a very clear idea why it couldn't happen), it breaks how the patch
decides between transactional and non-transactional sequence changes.
So this seems like a fatal flaw - it definitely needs to be solved. I
don't have a good idea how to do that, unfortunately. The problem is the
dependency on an earlier record, and that this needs to be evaluated
immediately (in the decode phase). Logical messages don't have the same
issue because the "transactional" flag does not depend on earlier stuff,
and other records are not interpreted until apply/commit, when we know
everything relevant was decoded.
I don't know what the solution is. Either we find a way to make sure not
to lose/skip the smgr record, or we need to rethink how we determine the
transactional flag (perhaps even try again adding it to the WAL record,
but we didn't find a way to do that earlier).
3) performance issues
We have discussed two cases - "ddl-abort" and "nextval-40".
The "ddl-abort" is when the workload does a lot of DDL and then aborts
them, leading to profiles dominated by DecodeCommit. The agreement here
is that while this is a valid issue and we should try fixing it, it's
unrelated to this patch. The issue exists even on master. So in the
context of this patch we can ignore this issue.
The "nextval-40" applies to workloads doing a lot of regular sequence
changes. We only decode/apply changes written to WAL, and that happens
only for every 32 increments or so. The test was with a very simple
transaction (just sequence advanced to write WAL + 1-row insert), which
means it's pretty much a worst case impact. For larger transactions,
it's going to be hardly measurable. Also, this only measured decoding,
not apply (which also will make this less significant).
Most of the overhead comes from ReorderBufferQueueSequence() starting
and then aborting a transaction, per the profile in [2]. This only
happens in the non-transactional case, but we expect that in regular
Anyway, let's say we want to mitigate this overhead. I think there are
three ways to do that:
a) find a way to not have to apply sequence changes immediately, but
queue them until the next commit
This would give a chance to combine multiple sequence changes into a
single "replay change", reducing the overhead. There's a couple problems
with this, though. Firstly, it can't help OLTP workloads because the
transactions are short so sequence changes are unlikely to combine. It's
also, not clear how expensive this be - could it be expensive enough to
outweigh the benefits?
All of this is assuming it can be implemented, we don't have such patch
yet. I was speculating about something like this earlier, but I haven't
managed to make that work. Doesn't mean it's impossible, ofc.
b) provide a way for the output plugin to skip sequence decoding early
The way the decoding is coded now, ReorderBufferQueueSequence does all
the expensive dance even if the output plugin does not implement the
sequence callbacks.
Maybe we should have a way to allow skipping all of this early, right at
the beginning of ReorderBufferQueueSequence (and thus before we even try
to start/abort the transaction).
Ofc, this is not a perfect solution either - it won't help workloads
that actually need/want sequence decoding but the workload is such that
the decoding has significant overhead, or with plugins that choose to
support decoding sequences in genera. For example the built-in output
plugin would certainly support sequences - and the overhead would still
be there (even if no sequences are added to the publication).
b) instruct people to increase the sequence cache from 32 to 1024
This would reduce the number of WAL messages that need to be decoded and
replayed, reducing the overhead proportionally. Of course, this also
means the sequence will "jump forward" more in case of crash or failover
to the logical replica, but I think that's acceptable tradeoff. People
should not expect sequences to be gap-less anyway.
Considering nextval-40 is pretty much a worst-case behavior, I think
this might actually be an acceptable solution/workaround.
regards
[1]
https://www.postgresql.org/message-id/CAFiTN-vAx-Y%2B19ROKOcWnGf7ix2VOTUebpzteaGw9XQyCAeK6g%40mail.g...
[2]
https://www.postgresql.org/message-id/0bc34f71-7745-dc16-d765-5ba1f0776a3f%40enterprisedb.com
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2023-12-13 12:56 Amit Kapila <[email protected]>
parent: Dilip Kumar <[email protected]>
0 siblings, 2 replies; 127+ messages in thread
From: Amit Kapila @ 2023-12-13 12:56 UTC (permalink / raw)
To: Dilip Kumar <[email protected]>; +Cc: Tomas Vondra <[email protected]>; Hayato Kuroda (Fujitsu) <[email protected]>; Zhijie Hou (Fujitsu) <[email protected]>; Ashutosh Bapat <[email protected]>; PostgreSQL Hackers <[email protected]>; Masahiko Sawada <[email protected]>; Peter Eisentraut <[email protected]>
On Thu, Dec 7, 2023 at 10:41 AM Dilip Kumar <[email protected]> wrote:
>
> On Wed, Dec 6, 2023 at 7:09 PM Tomas Vondra
> <[email protected]> wrote:
> >
> > Yes, if something like this happens, that'd be a problem:
> >
> > 1) decoding starts, with
> >
> > SnapBuildCurrentState(builder) < SNAPBUILD_FULL_SNAPSHOT
> >
> > 2) transaction that creates a new refilenode gets decoded, but we skip
> > it because we don't have the correct snapshot
> >
> > 3) snapshot changes to SNAPBUILD_FULL_SNAPSHOT
> >
> > 4) we decode sequence change from nextval() for the sequence
> >
> > This would lead to us attempting to apply sequence change for a
> > relfilenode that's not visible yet (and may even get aborted).
> >
> > But can this even happen? Can we start decoding in the middle of a
> > transaction? How come this wouldn't affect e.g. XLOG_HEAP2_NEW_CID,
> > which is also skipped until SNAPBUILD_FULL_SNAPSHOT. Or logical
> > messages, where we also call the output plugin in non-transactional cases.
>
> It's not a problem for logical messages because whether the message is
> transaction or non-transactional is decided while WAL logs the message
> itself. But here our problem starts with deciding whether the change
> is transactional vs non-transactional, because if we insert the
> 'relfilenode' in hash then the subsequent sequence change in the same
> transaction would be considered transactional otherwise
> non-transactional.
>
It is correct that we can make a wrong decision about whether a change
is transactional or non-transactional when sequence DDL happens before
the SNAPBUILD_FULL_SNAPSHOT state and the sequence operation happens
after that state. However, one thing to note here is that we won't try
to stream such a change because for non-transactional cases we don't
proceed unless the snapshot is in a consistent state. Now, if the
decision had been correct then we would probably have queued the
sequence change and discarded at commit.
One thing that we deviate here is that for non-sequence transactional
cases (including logical messages), we immediately start queuing the
changes as soon as we reach SNAPBUILD_FULL_SNAPSHOT state (provided
SnapBuildProcessChange() returns true which is quite possible) and
take final decision at commit/prepare/abort time. However, that won't
be the case for sequences because of the dependency of determining
transactional cases on one of the prior records. Now, I am not
completely sure at this stage if such a deviation can cause any
problem and or whether we are okay to have such a deviation for
sequences.
--
With Regards,
Amit Kapila.
^ permalink raw reply [nested|flat] 127+ messages in thread
* RE: logical decoding and replication of sequences, take 2
@ 2023-12-14 03:44 Hayato Kuroda (Fujitsu) <[email protected]>
parent: Amit Kapila <[email protected]>
1 sibling, 0 replies; 127+ messages in thread
From: Hayato Kuroda (Fujitsu) @ 2023-12-14 03:44 UTC (permalink / raw)
To: 'Amit Kapila' <[email protected]>; Dilip Kumar <[email protected]>; PostgreSQL Hackers <[email protected]>; Tomas Vondra <[email protected]>; +Cc: Zhijie Hou (Fujitsu) <[email protected]>; Ashutosh Bapat <[email protected]>; Masahiko Sawada <[email protected]>; Peter Eisentraut <[email protected]>
Dear hackers,
> It is correct that we can make a wrong decision about whether a change
> is transactional or non-transactional when sequence DDL happens before
> the SNAPBUILD_FULL_SNAPSHOT state and the sequence operation happens
> after that state.
I found a workload which decoder distinguish wrongly.
# Prerequisite
Apply an attached patch for inspecting the sequence status. It can be applied atop v20231203 patch set.
Also, a table and a sequence must be defined:
```
CREATE TABLE foo (var int);
CREATE SEQUENCE s;
```
# Workload
Then, you can execute concurrent transactions from three clients like below:
Client-1
BEGIN;
INSERT INTO foo VALUES (1);
Client-2
SELECT pg_create_logical_replication_slot('slot', 'test_decoding');
Client-3
BEGIN;
ALTER SEQUENCE s MAXVALUE 5000;
COMMIT;
SAVEPOINT s1;
SELECT setval('s', 2000);
ROLLBACK;
SELECT pg_logical_slot_get_changes('slot', 'test_decoding');
# Result and analysis
At first, below lines would be output on the log. This meant that WAL records
for ALTER SEQUENCE were decoded but skipped because the snapshot had been building.
```
...
LOG: logical decoding found initial starting point at 0/154D238
DETAIL: Waiting for transactions (approximately 1) older than 741 to end.
STATEMENT: SELECT * FROM pg_create_logical_replication_slot('slot', 'test_decoding');
LOG: XXX: smgr_decode. snapshot is SNAPBUILD_BUILDING_SNAPSHOT
STATEMENT: SELECT * FROM pg_create_logical_replication_slot('slot', 'test_decoding');
LOG: XXX: skipped
STATEMENT: SELECT * FROM pg_create_logical_replication_slot('slot', 'test_decoding');
LOG: XXX: seq_decode. snapshot is SNAPBUILD_BUILDING_SNAPSHOT
STATEMENT: SELECT * FROM pg_create_logical_replication_slot('slot', 'test_decoding');
LOG: XXX: skipped
...
```
Note that above `seq_decode...` line was not output via `setval()`, it was done
by ALTER SEQUENCE statement. Below is a call stack for inserting WAL.
```
XLogInsert(RM_SEQ_ID, XLOG_SEQ_LOG);
fill_seq_fork_with_data
fill_seq_with_data
AlterSequence
```
Then, subsequent lines would say like them. This means that the snapshot becomes
FULL and `setval()` is regarded non-transactional wrongly.
```
LOG: logical decoding found initial consistent point at 0/154D658
DETAIL: Waiting for transactions (approximately 1) older than 742 to end.
STATEMENT: SELECT * FROM pg_create_logical_replication_slot('slot', 'test_decoding');
LOG: XXX: seq_decode. snapshot is SNAPBUILD_FULL_SNAPSHOT
STATEMENT: SELECT * FROM pg_create_logical_replication_slot('slot', 'test_decoding');
LOG: XXX: the sequence is non-transactional
STATEMENT: SELECT * FROM pg_create_logical_replication_slot('slot', 'test_decoding');
LOG: XXX: not consistent: skipped
```
The change would be discarded because the snapshot has not been CONSISTENT yet
by the below part. If it has been transactional, we would have queued this
change though the transaction will be skipped at commit.
```
else if (!transactional &&
(SnapBuildCurrentState(builder) != SNAPBUILD_CONSISTENT ||
SnapBuildXactNeedsSkip(builder, buf->origptr)))
return;
```
But anyway, we could find a case which we can make a wrong decision. This example
is lucky - does not output wrongly, but I'm not sure all the case like that.
Best Regards,
Hayato Kuroda
FUJITSU LIMITED
diff --git a/src/backend/replication/logical/decode.c b/src/backend/replication/logical/decode.c
index d48d88081f..73e38cafd8 100644
--- a/src/backend/replication/logical/decode.c
+++ b/src/backend/replication/logical/decode.c
@@ -1397,12 +1397,17 @@ seq_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
ReorderBufferProcessXid(ctx->reorder, XLogRecGetXid(r), buf->origptr);
+ elog(LOG, "XXX: seq_decode. snapshot is %s", SnapBuildIdentify(builder));
+
/*
* If we don't have snapshot or we are just fast-forwarding, there is no
* point in decoding sequences.
*/
if (SnapBuildCurrentState(builder) < SNAPBUILD_FULL_SNAPSHOT)
+ {
+ elog(LOG, "XXX: skipped");
return;
+ }
/* only interested in our database */
XLogRecGetBlockTag(r, 0, &target_locator, NULL, NULL);
@@ -1437,14 +1442,22 @@ seq_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
target_locator,
NULL);
+ elog(LOG, "XXX: the sequence is %s", transactional ? "transactional" : "non-transactional");
+
/* Skip the change if already processed (per the snapshot). */
if (transactional &&
!SnapBuildProcessChange(builder, xid, buf->origptr))
+ {
+ elog(LOG, "XXX: already decoded: skipped");
return;
+ }
else if (!transactional &&
(SnapBuildCurrentState(builder) != SNAPBUILD_CONSISTENT ||
SnapBuildXactNeedsSkip(builder, buf->origptr)))
+ {
+ elog(LOG, "XXX: not consistent: skipped");
return;
+ }
/*
* We also skip decoding in fast_forward mode. This check must be last
@@ -1527,13 +1540,18 @@ smgr_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
ReorderBufferProcessXid(ctx->reorder, XLogRecGetXid(r), buf->origptr);
+ elog(LOG, "XXX: smgr_decode. snapshot is %s", SnapBuildIdentify(builder));
+
/*
* If we don't have snapshot or we are just fast-forwarding, there is no
* point in decoding relfilenode information.
*/
if (SnapBuildCurrentState(builder) < SNAPBUILD_FULL_SNAPSHOT ||
ctx->fast_forward)
+ {
+ elog(LOG, "XXX: skipped");
return;
+ }
/* only interested in our database */
xlrec = (xl_smgr_create *) XLogRecGetData(r);
diff --git a/src/backend/replication/logical/snapbuild.c b/src/backend/replication/logical/snapbuild.c
index fec190a8b2..efb2afb09d 100644
--- a/src/backend/replication/logical/snapbuild.c
+++ b/src/backend/replication/logical/snapbuild.c
@@ -2135,3 +2135,27 @@ CheckPointSnapBuild(void)
}
FreeDir(snap_dir);
}
+
+char *
+SnapBuildIdentify(SnapBuild *builder)
+{
+ char *ret = NULL;
+
+ switch (builder->state)
+ {
+ case SNAPBUILD_START:
+ ret = "SNAPBUILD_START";
+ break;
+ case SNAPBUILD_BUILDING_SNAPSHOT:
+ ret = "SNAPBUILD_BUILDING_SNAPSHOT";
+ break;
+ case SNAPBUILD_FULL_SNAPSHOT:
+ ret = "SNAPBUILD_FULL_SNAPSHOT";
+ break;
+ case SNAPBUILD_CONSISTENT:
+ ret = "SNAPBUILD_CONSISTENT";
+ break;
+ }
+
+ return ret;
+}
diff --git a/src/include/replication/snapbuild.h b/src/include/replication/snapbuild.h
index 63f50a13d6..bf257f5f5a 100644
--- a/src/include/replication/snapbuild.h
+++ b/src/include/replication/snapbuild.h
@@ -91,4 +91,6 @@ extern void SnapBuildProcessRunningXacts(SnapBuild *builder, XLogRecPtr lsn,
struct xl_running_xacts *running);
extern void SnapBuildSerializationPoint(SnapBuild *builder, XLogRecPtr lsn);
+extern char *SnapBuildIdentify(SnapBuild *builder);
+
#endif /* SNAPBUILD_H */
Attachments:
[text/plain] add_elog.txt (3.3K, ../../TY3PR01MB98899FA0082AE9918886E150F58CA@TY3PR01MB9889.jpnprd01.prod.outlook.com/2-add_elog.txt)
download | inline diff:
diff --git a/src/backend/replication/logical/decode.c b/src/backend/replication/logical/decode.c
index d48d88081f..73e38cafd8 100644
--- a/src/backend/replication/logical/decode.c
+++ b/src/backend/replication/logical/decode.c
@@ -1397,12 +1397,17 @@ seq_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
ReorderBufferProcessXid(ctx->reorder, XLogRecGetXid(r), buf->origptr);
+ elog(LOG, "XXX: seq_decode. snapshot is %s", SnapBuildIdentify(builder));
+
/*
* If we don't have snapshot or we are just fast-forwarding, there is no
* point in decoding sequences.
*/
if (SnapBuildCurrentState(builder) < SNAPBUILD_FULL_SNAPSHOT)
+ {
+ elog(LOG, "XXX: skipped");
return;
+ }
/* only interested in our database */
XLogRecGetBlockTag(r, 0, &target_locator, NULL, NULL);
@@ -1437,14 +1442,22 @@ seq_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
target_locator,
NULL);
+ elog(LOG, "XXX: the sequence is %s", transactional ? "transactional" : "non-transactional");
+
/* Skip the change if already processed (per the snapshot). */
if (transactional &&
!SnapBuildProcessChange(builder, xid, buf->origptr))
+ {
+ elog(LOG, "XXX: already decoded: skipped");
return;
+ }
else if (!transactional &&
(SnapBuildCurrentState(builder) != SNAPBUILD_CONSISTENT ||
SnapBuildXactNeedsSkip(builder, buf->origptr)))
+ {
+ elog(LOG, "XXX: not consistent: skipped");
return;
+ }
/*
* We also skip decoding in fast_forward mode. This check must be last
@@ -1527,13 +1540,18 @@ smgr_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
ReorderBufferProcessXid(ctx->reorder, XLogRecGetXid(r), buf->origptr);
+ elog(LOG, "XXX: smgr_decode. snapshot is %s", SnapBuildIdentify(builder));
+
/*
* If we don't have snapshot or we are just fast-forwarding, there is no
* point in decoding relfilenode information.
*/
if (SnapBuildCurrentState(builder) < SNAPBUILD_FULL_SNAPSHOT ||
ctx->fast_forward)
+ {
+ elog(LOG, "XXX: skipped");
return;
+ }
/* only interested in our database */
xlrec = (xl_smgr_create *) XLogRecGetData(r);
diff --git a/src/backend/replication/logical/snapbuild.c b/src/backend/replication/logical/snapbuild.c
index fec190a8b2..efb2afb09d 100644
--- a/src/backend/replication/logical/snapbuild.c
+++ b/src/backend/replication/logical/snapbuild.c
@@ -2135,3 +2135,27 @@ CheckPointSnapBuild(void)
}
FreeDir(snap_dir);
}
+
+char *
+SnapBuildIdentify(SnapBuild *builder)
+{
+ char *ret = NULL;
+
+ switch (builder->state)
+ {
+ case SNAPBUILD_START:
+ ret = "SNAPBUILD_START";
+ break;
+ case SNAPBUILD_BUILDING_SNAPSHOT:
+ ret = "SNAPBUILD_BUILDING_SNAPSHOT";
+ break;
+ case SNAPBUILD_FULL_SNAPSHOT:
+ ret = "SNAPBUILD_FULL_SNAPSHOT";
+ break;
+ case SNAPBUILD_CONSISTENT:
+ ret = "SNAPBUILD_CONSISTENT";
+ break;
+ }
+
+ return ret;
+}
diff --git a/src/include/replication/snapbuild.h b/src/include/replication/snapbuild.h
index 63f50a13d6..bf257f5f5a 100644
--- a/src/include/replication/snapbuild.h
+++ b/src/include/replication/snapbuild.h
@@ -91,4 +91,6 @@ extern void SnapBuildProcessRunningXacts(SnapBuild *builder, XLogRecPtr lsn,
struct xl_running_xacts *running);
extern void SnapBuildSerializationPoint(SnapBuild *builder, XLogRecPtr lsn);
+extern char *SnapBuildIdentify(SnapBuild *builder);
+
#endif /* SNAPBUILD_H */
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2023-12-14 05:22 Dilip Kumar <[email protected]>
parent: Amit Kapila <[email protected]>
1 sibling, 1 reply; 127+ messages in thread
From: Dilip Kumar @ 2023-12-14 05:22 UTC (permalink / raw)
To: Amit Kapila <[email protected]>; +Cc: Tomas Vondra <[email protected]>; Hayato Kuroda (Fujitsu) <[email protected]>; Zhijie Hou (Fujitsu) <[email protected]>; Ashutosh Bapat <[email protected]>; PostgreSQL Hackers <[email protected]>; Masahiko Sawada <[email protected]>; Peter Eisentraut <[email protected]>
On Wed, Dec 13, 2023 at 6:26 PM Amit Kapila <[email protected]> wrote:
>
> > > But can this even happen? Can we start decoding in the middle of a
> > > transaction? How come this wouldn't affect e.g. XLOG_HEAP2_NEW_CID,
> > > which is also skipped until SNAPBUILD_FULL_SNAPSHOT. Or logical
> > > messages, where we also call the output plugin in non-transactional cases.
> >
> > It's not a problem for logical messages because whether the message is
> > transaction or non-transactional is decided while WAL logs the message
> > itself. But here our problem starts with deciding whether the change
> > is transactional vs non-transactional, because if we insert the
> > 'relfilenode' in hash then the subsequent sequence change in the same
> > transaction would be considered transactional otherwise
> > non-transactional.
> >
>
> It is correct that we can make a wrong decision about whether a change
> is transactional or non-transactional when sequence DDL happens before
> the SNAPBUILD_FULL_SNAPSHOT state and the sequence operation happens
> after that state. However, one thing to note here is that we won't try
> to stream such a change because for non-transactional cases we don't
> proceed unless the snapshot is in a consistent state. Now, if the
> decision had been correct then we would probably have queued the
> sequence change and discarded at commit.
>
> One thing that we deviate here is that for non-sequence transactional
> cases (including logical messages), we immediately start queuing the
> changes as soon as we reach SNAPBUILD_FULL_SNAPSHOT state (provided
> SnapBuildProcessChange() returns true which is quite possible) and
> take final decision at commit/prepare/abort time. However, that won't
> be the case for sequences because of the dependency of determining
> transactional cases on one of the prior records. Now, I am not
> completely sure at this stage if such a deviation can cause any
> problem and or whether we are okay to have such a deviation for
> sequences.
Okay, so this particular scenario that I raised is somehow saved, I
mean although we are considering transactional sequence operation as
non-transactional we also know that if some of the changes for a
transaction are skipped because the snapshot was not FULL that means
that transaction can not be streamed because that transaction has to
be committed before snapshot become CONSISTENT (based on the snapshot
state change machinery). Ideally based on the same logic that the
snapshot is not consistent the non-transactional sequence changes are
also skipped. But the only thing that makes me a bit uncomfortable is
that even though the result is not wrong we have made some wrong
intermediate decisions i.e. considered transactional change as
non-transactions.
One solution to this issue is that, even if the snapshot state does
not reach FULL just add the sequence relids to the hash, I mean that
hash is only maintained for deciding whether the sequence is changed
in that transaction or not. So no adding such relids to hash seems
like a root cause of the issue. Honestly, I haven't analyzed this
idea in detail about how easy it would be to add only these changes to
the hash and what are the other dependencies, but this seems like a
worthwhile direction IMHO.
--
Regards,
Dilip Kumar
EnterpriseDB: http://www.enterprisedb.com
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2023-12-14 07:01 Ashutosh Bapat <[email protected]>
parent: Dilip Kumar <[email protected]>
0 siblings, 2 replies; 127+ messages in thread
From: Ashutosh Bapat @ 2023-12-14 07:01 UTC (permalink / raw)
To: Dilip Kumar <[email protected]>; +Cc: Amit Kapila <[email protected]>; Tomas Vondra <[email protected]>; Hayato Kuroda (Fujitsu) <[email protected]>; Zhijie Hou (Fujitsu) <[email protected]>; PostgreSQL Hackers <[email protected]>; Masahiko Sawada <[email protected]>; Peter Eisentraut <[email protected]>
On Thu, Dec 14, 2023 at 10:53 AM Dilip Kumar <[email protected]> wrote:
> > >
> >
> > It is correct that we can make a wrong decision about whether a change
> > is transactional or non-transactional when sequence DDL happens before
> > the SNAPBUILD_FULL_SNAPSHOT state and the sequence operation happens
> > after that state. However, one thing to note here is that we won't try
> > to stream such a change because for non-transactional cases we don't
> > proceed unless the snapshot is in a consistent state. Now, if the
> > decision had been correct then we would probably have queued the
> > sequence change and discarded at commit.
> >
> > One thing that we deviate here is that for non-sequence transactional
> > cases (including logical messages), we immediately start queuing the
> > changes as soon as we reach SNAPBUILD_FULL_SNAPSHOT state (provided
> > SnapBuildProcessChange() returns true which is quite possible) and
> > take final decision at commit/prepare/abort time. However, that won't
> > be the case for sequences because of the dependency of determining
> > transactional cases on one of the prior records. Now, I am not
> > completely sure at this stage if such a deviation can cause any
> > problem and or whether we are okay to have such a deviation for
> > sequences.
>
> Okay, so this particular scenario that I raised is somehow saved, I
> mean although we are considering transactional sequence operation as
> non-transactional we also know that if some of the changes for a
> transaction are skipped because the snapshot was not FULL that means
> that transaction can not be streamed because that transaction has to
> be committed before snapshot become CONSISTENT (based on the snapshot
> state change machinery). Ideally based on the same logic that the
> snapshot is not consistent the non-transactional sequence changes are
> also skipped. But the only thing that makes me a bit uncomfortable is
> that even though the result is not wrong we have made some wrong
> intermediate decisions i.e. considered transactional change as
> non-transactions.
>
> One solution to this issue is that, even if the snapshot state does
> not reach FULL just add the sequence relids to the hash, I mean that
> hash is only maintained for deciding whether the sequence is changed
> in that transaction or not. So no adding such relids to hash seems
> like a root cause of the issue. Honestly, I haven't analyzed this
> idea in detail about how easy it would be to add only these changes to
> the hash and what are the other dependencies, but this seems like a
> worthwhile direction IMHO.
I also thought about the same solution. I tried this solution as the
attached patch on top of Hayato's diagnostic changes. Following log
messages are seen in server error log. Those indicate that the
sequence change was correctly deemed as a transactional change (line
2023-12-14 12:14:55.591 IST [321229] LOG: XXX: the sequence is
transactional).
2023-12-14 12:12:50.550 IST [321229] ERROR: relation
"pg_replication_slot" does not exist at character 15
2023-12-14 12:12:50.550 IST [321229] STATEMENT: select * from
pg_replication_slot;
2023-12-14 12:12:57.289 IST [321229] LOG: logical decoding found
initial starting point at 0/1598D50
2023-12-14 12:12:57.289 IST [321229] DETAIL: Waiting for transactions
(approximately 1) older than 759 to end.
2023-12-14 12:12:57.289 IST [321229] STATEMENT: SELECT
pg_create_logical_replication_slot('slot', 'test_decoding');
2023-12-14 12:13:49.551 IST [321229] LOG: XXX: smgr_decode. snapshot
is SNAPBUILD_BUILDING_SNAPSHOT
2023-12-14 12:13:49.551 IST [321229] STATEMENT: SELECT
pg_create_logical_replication_slot('slot', 'test_decoding');
2023-12-14 12:13:49.551 IST [321229] LOG: XXX: seq_decode. snapshot is
SNAPBUILD_BUILDING_SNAPSHOT
2023-12-14 12:13:49.551 IST [321229] STATEMENT: SELECT
pg_create_logical_replication_slot('slot', 'test_decoding');
2023-12-14 12:13:49.551 IST [321229] LOG: XXX: skipped
2023-12-14 12:13:49.551 IST [321229] STATEMENT: SELECT
pg_create_logical_replication_slot('slot', 'test_decoding');
2023-12-14 12:13:49.552 IST [321229] LOG: logical decoding found
initial consistent point at 0/1599170
2023-12-14 12:13:49.552 IST [321229] DETAIL: Waiting for transactions
(approximately 1) older than 760 to end.
2023-12-14 12:13:49.552 IST [321229] STATEMENT: SELECT
pg_create_logical_replication_slot('slot', 'test_decoding');
2023-12-14 12:14:55.591 IST [321229] LOG: XXX: seq_decode. snapshot is
SNAPBUILD_FULL_SNAPSHOT
2023-12-14 12:14:55.591 IST [321230] STATEMENT: SELECT
pg_create_logical_replication_slot('slot', 'test_decoding');
2023-12-14 12:14:55.591 IST [321229] LOG: XXX: the sequence is transactional
2023-12-14 12:14:55.591 IST [321229] STATEMENT: SELECT
pg_create_logical_replication_slot('slot', 'test_decoding');
2023-12-14 12:14:55.813 IST [321229] LOG: logical decoding found
consistent point at 0/15992E8
2023-12-14 12:14:55.813 IST [321229] DETAIL: There are no running transactions.
2023-12-14 12:14:55.813 IST [321229] STATEMENT: SELECT
pg_create_logical_replication_slot('slot', 'test_decoding');
It looks like the solution works. But this is the only place where we
process a change before SNAPSHOT reaches FULL. But this is also the
only record which affects a decision to queue/not a following change.
So it should be ok. The sequence_hash'es as separate for each
transaction and they are cleaned when processing COMMIT record. So I
think we don't have any side effects of adding relfilenode to sequence
hash even though snapshot is not FULL.
As a side note
1. the prologue of ReorderBufferSequenceCleanup() mentions only abort,
but this function will be called for COMMIT as well. Prologue needs to
be fixed.
2. Now that sequence hashes are per transaction, do we need
ReoderBufferTXN in ReorderBufferSequenceEnt?
--
Best Wishes,
Ashutosh Bapat
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2023-12-14 07:07 Dilip Kumar <[email protected]>
parent: Ashutosh Bapat <[email protected]>
1 sibling, 0 replies; 127+ messages in thread
From: Dilip Kumar @ 2023-12-14 07:07 UTC (permalink / raw)
To: Ashutosh Bapat <[email protected]>; +Cc: Amit Kapila <[email protected]>; Tomas Vondra <[email protected]>; Hayato Kuroda (Fujitsu) <[email protected]>; Zhijie Hou (Fujitsu) <[email protected]>; PostgreSQL Hackers <[email protected]>; Masahiko Sawada <[email protected]>; Peter Eisentraut <[email protected]>
On Thu, Dec 14, 2023 at 12:31 PM Ashutosh Bapat
<[email protected]> wrote:
>
> On Thu, Dec 14, 2023 at 10:53 AM Dilip Kumar <[email protected]> wrote:
>
> > > >
> > >
> > > It is correct that we can make a wrong decision about whether a change
> > > is transactional or non-transactional when sequence DDL happens before
> > > the SNAPBUILD_FULL_SNAPSHOT state and the sequence operation happens
> > > after that state. However, one thing to note here is that we won't try
> > > to stream such a change because for non-transactional cases we don't
> > > proceed unless the snapshot is in a consistent state. Now, if the
> > > decision had been correct then we would probably have queued the
> > > sequence change and discarded at commit.
> > >
> > > One thing that we deviate here is that for non-sequence transactional
> > > cases (including logical messages), we immediately start queuing the
> > > changes as soon as we reach SNAPBUILD_FULL_SNAPSHOT state (provided
> > > SnapBuildProcessChange() returns true which is quite possible) and
> > > take final decision at commit/prepare/abort time. However, that won't
> > > be the case for sequences because of the dependency of determining
> > > transactional cases on one of the prior records. Now, I am not
> > > completely sure at this stage if such a deviation can cause any
> > > problem and or whether we are okay to have such a deviation for
> > > sequences.
> >
> > Okay, so this particular scenario that I raised is somehow saved, I
> > mean although we are considering transactional sequence operation as
> > non-transactional we also know that if some of the changes for a
> > transaction are skipped because the snapshot was not FULL that means
> > that transaction can not be streamed because that transaction has to
> > be committed before snapshot become CONSISTENT (based on the snapshot
> > state change machinery). Ideally based on the same logic that the
> > snapshot is not consistent the non-transactional sequence changes are
> > also skipped. But the only thing that makes me a bit uncomfortable is
> > that even though the result is not wrong we have made some wrong
> > intermediate decisions i.e. considered transactional change as
> > non-transactions.
> >
> > One solution to this issue is that, even if the snapshot state does
> > not reach FULL just add the sequence relids to the hash, I mean that
> > hash is only maintained for deciding whether the sequence is changed
> > in that transaction or not. So no adding such relids to hash seems
> > like a root cause of the issue. Honestly, I haven't analyzed this
> > idea in detail about how easy it would be to add only these changes to
> > the hash and what are the other dependencies, but this seems like a
> > worthwhile direction IMHO.
>
> I also thought about the same solution. I tried this solution as the
> attached patch on top of Hayato's diagnostic changes.
I think you forgot to attach the patch.
--
Regards,
Dilip Kumar
EnterpriseDB: http://www.enterprisedb.com
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2023-12-14 09:06 Amit Kapila <[email protected]>
parent: Ashutosh Bapat <[email protected]>
1 sibling, 1 reply; 127+ messages in thread
From: Amit Kapila @ 2023-12-14 09:06 UTC (permalink / raw)
To: Ashutosh Bapat <[email protected]>; +Cc: Dilip Kumar <[email protected]>; Tomas Vondra <[email protected]>; Hayato Kuroda (Fujitsu) <[email protected]>; Zhijie Hou (Fujitsu) <[email protected]>; PostgreSQL Hackers <[email protected]>; Masahiko Sawada <[email protected]>; Peter Eisentraut <[email protected]>
On Thu, Dec 14, 2023 at 12:31 PM Ashutosh Bapat
<[email protected]> wrote:
>
> On Thu, Dec 14, 2023 at 10:53 AM Dilip Kumar <[email protected]> wrote:
>
> > > >
> > >
> > > It is correct that we can make a wrong decision about whether a change
> > > is transactional or non-transactional when sequence DDL happens before
> > > the SNAPBUILD_FULL_SNAPSHOT state and the sequence operation happens
> > > after that state. However, one thing to note here is that we won't try
> > > to stream such a change because for non-transactional cases we don't
> > > proceed unless the snapshot is in a consistent state. Now, if the
> > > decision had been correct then we would probably have queued the
> > > sequence change and discarded at commit.
> > >
> > > One thing that we deviate here is that for non-sequence transactional
> > > cases (including logical messages), we immediately start queuing the
> > > changes as soon as we reach SNAPBUILD_FULL_SNAPSHOT state (provided
> > > SnapBuildProcessChange() returns true which is quite possible) and
> > > take final decision at commit/prepare/abort time. However, that won't
> > > be the case for sequences because of the dependency of determining
> > > transactional cases on one of the prior records. Now, I am not
> > > completely sure at this stage if such a deviation can cause any
> > > problem and or whether we are okay to have such a deviation for
> > > sequences.
> >
> > Okay, so this particular scenario that I raised is somehow saved, I
> > mean although we are considering transactional sequence operation as
> > non-transactional we also know that if some of the changes for a
> > transaction are skipped because the snapshot was not FULL that means
> > that transaction can not be streamed because that transaction has to
> > be committed before snapshot become CONSISTENT (based on the snapshot
> > state change machinery). Ideally based on the same logic that the
> > snapshot is not consistent the non-transactional sequence changes are
> > also skipped. But the only thing that makes me a bit uncomfortable is
> > that even though the result is not wrong we have made some wrong
> > intermediate decisions i.e. considered transactional change as
> > non-transactions.
> >
> > One solution to this issue is that, even if the snapshot state does
> > not reach FULL just add the sequence relids to the hash, I mean that
> > hash is only maintained for deciding whether the sequence is changed
> > in that transaction or not. So no adding such relids to hash seems
> > like a root cause of the issue. Honestly, I haven't analyzed this
> > idea in detail about how easy it would be to add only these changes to
> > the hash and what are the other dependencies, but this seems like a
> > worthwhile direction IMHO.
>
>
...
> It looks like the solution works. But this is the only place where we
> process a change before SNAPSHOT reaches FULL. But this is also the
> only record which affects a decision to queue/not a following change.
> So it should be ok. The sequence_hash'es as separate for each
> transaction and they are cleaned when processing COMMIT record.
>
>
It looks like the solution works. But this is the only place where we
process a change before SNAPSHOT reaches FULL. But this is also the
only record which affects a decision to queue/not a following change.
So it should be ok. The sequence_hash'es as separate for each
transaction and they are cleaned when processing COMMIT record.
>
But it is possible that even commit or abort also happens before the
snapshot reaches full state in which case the hash table will have
stale or invalid (for aborts) entries. That will probably be cleaned
at a later point by running_xact records. Now, I think in theory, it
is possible that the same RelFileLocator can again be allocated before
we clean up the existing entry which can probably confuse the system.
It might or might not be a problem in practice but I think the more
assumptions we add for sequences, the more difficult it will become to
ensure its correctness.
--
With Regards,
Amit Kapila.
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2023-12-14 09:14 Ashutosh Bapat <[email protected]>
parent: Amit Kapila <[email protected]>
0 siblings, 1 reply; 127+ messages in thread
From: Ashutosh Bapat @ 2023-12-14 09:14 UTC (permalink / raw)
To: Amit Kapila <[email protected]>; Dilip Kumar <[email protected]>; +Cc: Tomas Vondra <[email protected]>; Hayato Kuroda (Fujitsu) <[email protected]>; Zhijie Hou (Fujitsu) <[email protected]>; PostgreSQL Hackers <[email protected]>; Masahiko Sawada <[email protected]>; Peter Eisentraut <[email protected]>
On Thu, Dec 14, 2023 at 12:37 PM Dilip Kumar <[email protected]> wrote:
>
> I think you forgot to attach the patch.
Sorry. Here it is.
On Thu, Dec 14, 2023 at 2:36 PM Amit Kapila <[email protected]> wrote:
>
> >
> It looks like the solution works. But this is the only place where we
> process a change before SNAPSHOT reaches FULL. But this is also the
> only record which affects a decision to queue/not a following change.
> So it should be ok. The sequence_hash'es as separate for each
> transaction and they are cleaned when processing COMMIT record.
> >
>
> But it is possible that even commit or abort also happens before the
> snapshot reaches full state in which case the hash table will have
> stale or invalid (for aborts) entries. That will probably be cleaned
> at a later point by running_xact records.
Why would cleaning wait till running_xact records? Won't txn entry
itself be removed when processing commit/abort record? At the same the
sequence hash will be cleaned as well.
> Now, I think in theory, it
> is possible that the same RelFileLocator can again be allocated before
> we clean up the existing entry which can probably confuse the system.
How? The transaction allocating the first time would be cleaned before
it happens the second time. So shouldn't matter.
--
Best Wishes,
Ashutosh Bapat
diff --git a/src/backend/replication/logical/decode.c b/src/backend/replication/logical/decode.c
index 73e38cafd8..8e2ebbd8e0 100644
--- a/src/backend/replication/logical/decode.c
+++ b/src/backend/replication/logical/decode.c
@@ -1543,10 +1543,15 @@ smgr_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
elog(LOG, "XXX: smgr_decode. snapshot is %s", SnapBuildIdentify(builder));
/*
- * If we don't have snapshot or we are just fast-forwarding, there is no
- * point in decoding relfilenode information.
+ * If we are not building snapshot yet or we are just fast-forwarding, there
+ * is no point in decoding relfilenode information. If the sequence
+ * associated with relfilenode here is changed in the same transaction but
+ * after snapshot was built, the relfilenode needs to be present in the
+ * sequence hash table so that the change can be deemed as transactional.
+ * Otherwise it will not be queued. Hence we process this change even before
+ * we have built snapshot.
*/
- if (SnapBuildCurrentState(builder) < SNAPBUILD_FULL_SNAPSHOT ||
+ if (SnapBuildCurrentState(builder) < SNAPBUILD_BUILDING_SNAPSHOT ||
ctx->fast_forward)
{
elog(LOG, "XXX: skipped");
Attachments:
[text/plain] add_relfilenode_before_full.txt (1.2K, ../../CAExHW5uRGwr0hAC5Gr4EbA4Q=Odjk+8nFTmdsfB1Hk-9r-81DA@mail.gmail.com/2-add_relfilenode_before_full.txt)
download | inline diff:
diff --git a/src/backend/replication/logical/decode.c b/src/backend/replication/logical/decode.c
index 73e38cafd8..8e2ebbd8e0 100644
--- a/src/backend/replication/logical/decode.c
+++ b/src/backend/replication/logical/decode.c
@@ -1543,10 +1543,15 @@ smgr_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
elog(LOG, "XXX: smgr_decode. snapshot is %s", SnapBuildIdentify(builder));
/*
- * If we don't have snapshot or we are just fast-forwarding, there is no
- * point in decoding relfilenode information.
+ * If we are not building snapshot yet or we are just fast-forwarding, there
+ * is no point in decoding relfilenode information. If the sequence
+ * associated with relfilenode here is changed in the same transaction but
+ * after snapshot was built, the relfilenode needs to be present in the
+ * sequence hash table so that the change can be deemed as transactional.
+ * Otherwise it will not be queued. Hence we process this change even before
+ * we have built snapshot.
*/
- if (SnapBuildCurrentState(builder) < SNAPBUILD_FULL_SNAPSHOT ||
+ if (SnapBuildCurrentState(builder) < SNAPBUILD_BUILDING_SNAPSHOT ||
ctx->fast_forward)
{
elog(LOG, "XXX: skipped");
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2023-12-14 09:20 Amit Kapila <[email protected]>
parent: Ashutosh Bapat <[email protected]>
0 siblings, 1 reply; 127+ messages in thread
From: Amit Kapila @ 2023-12-14 09:20 UTC (permalink / raw)
To: Ashutosh Bapat <[email protected]>; +Cc: Dilip Kumar <[email protected]>; Tomas Vondra <[email protected]>; Hayato Kuroda (Fujitsu) <[email protected]>; Zhijie Hou (Fujitsu) <[email protected]>; PostgreSQL Hackers <[email protected]>; Masahiko Sawada <[email protected]>; Peter Eisentraut <[email protected]>
On Thu, Dec 14, 2023 at 2:45 PM Ashutosh Bapat
<[email protected]> wrote:
>
> On Thu, Dec 14, 2023 at 12:37 PM Dilip Kumar <[email protected]> wrote:
> >
> > I think you forgot to attach the patch.
>
> Sorry. Here it is.
>
> On Thu, Dec 14, 2023 at 2:36 PM Amit Kapila <[email protected]> wrote:
> >
> > >
> > It looks like the solution works. But this is the only place where we
> > process a change before SNAPSHOT reaches FULL. But this is also the
> > only record which affects a decision to queue/not a following change.
> > So it should be ok. The sequence_hash'es as separate for each
> > transaction and they are cleaned when processing COMMIT record.
> > >
> >
> > But it is possible that even commit or abort also happens before the
> > snapshot reaches full state in which case the hash table will have
> > stale or invalid (for aborts) entries. That will probably be cleaned
> > at a later point by running_xact records.
>
> Why would cleaning wait till running_xact records? Won't txn entry
> itself be removed when processing commit/abort record? At the same the
> sequence hash will be cleaned as well.
>
> > Now, I think in theory, it
> > is possible that the same RelFileLocator can again be allocated before
> > we clean up the existing entry which can probably confuse the system.
>
> How? The transaction allocating the first time would be cleaned before
> it happens the second time. So shouldn't matter.
>
It can only be cleaned if we process it but xact_decode won't allow us
to process it and I don't think it would be a good idea to add another
hack for sequences here. See below code:
xact_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
{
SnapBuild *builder = ctx->snapshot_builder;
ReorderBuffer *reorder = ctx->reorder;
XLogReaderState *r = buf->record;
uint8 info = XLogRecGetInfo(r) & XLOG_XACT_OPMASK;
/*
* If the snapshot isn't yet fully built, we cannot decode anything, so
* bail out.
*/
if (SnapBuildCurrentState(builder) < SNAPBUILD_FULL_SNAPSHOT)
return;
--
With Regards,
Amit Kapila.
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2023-12-14 15:44 Ashutosh Bapat <[email protected]>
parent: Amit Kapila <[email protected]>
0 siblings, 1 reply; 127+ messages in thread
From: Ashutosh Bapat @ 2023-12-14 15:44 UTC (permalink / raw)
To: Amit Kapila <[email protected]>; +Cc: Dilip Kumar <[email protected]>; Tomas Vondra <[email protected]>; Hayato Kuroda (Fujitsu) <[email protected]>; Zhijie Hou (Fujitsu) <[email protected]>; PostgreSQL Hackers <[email protected]>; Masahiko Sawada <[email protected]>; Peter Eisentraut <[email protected]>
On Thu, Dec 14, 2023 at 2:51 PM Amit Kapila <[email protected]> wrote:
>
> On Thu, Dec 14, 2023 at 2:45 PM Ashutosh Bapat
> <[email protected]> wrote:
> >
> > On Thu, Dec 14, 2023 at 12:37 PM Dilip Kumar <[email protected]> wrote:
> > >
> > > I think you forgot to attach the patch.
> >
> > Sorry. Here it is.
> >
> > On Thu, Dec 14, 2023 at 2:36 PM Amit Kapila <[email protected]> wrote:
> > >
> > > >
> > > It looks like the solution works. But this is the only place where we
> > > process a change before SNAPSHOT reaches FULL. But this is also the
> > > only record which affects a decision to queue/not a following change.
> > > So it should be ok. The sequence_hash'es as separate for each
> > > transaction and they are cleaned when processing COMMIT record.
> > > >
> > >
> > > But it is possible that even commit or abort also happens before the
> > > snapshot reaches full state in which case the hash table will have
> > > stale or invalid (for aborts) entries. That will probably be cleaned
> > > at a later point by running_xact records.
> >
> > Why would cleaning wait till running_xact records? Won't txn entry
> > itself be removed when processing commit/abort record? At the same the
> > sequence hash will be cleaned as well.
> >
> > > Now, I think in theory, it
> > > is possible that the same RelFileLocator can again be allocated before
> > > we clean up the existing entry which can probably confuse the system.
> >
> > How? The transaction allocating the first time would be cleaned before
> > it happens the second time. So shouldn't matter.
> >
>
> It can only be cleaned if we process it but xact_decode won't allow us
> to process it and I don't think it would be a good idea to add another
> hack for sequences here. See below code:
>
> xact_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
> {
> SnapBuild *builder = ctx->snapshot_builder;
> ReorderBuffer *reorder = ctx->reorder;
> XLogReaderState *r = buf->record;
> uint8 info = XLogRecGetInfo(r) & XLOG_XACT_OPMASK;
>
> /*
> * If the snapshot isn't yet fully built, we cannot decode anything, so
> * bail out.
> */
> if (SnapBuildCurrentState(builder) < SNAPBUILD_FULL_SNAPSHOT)
> return;
That may be true for a transaction which is decoded, but I think all
the transactions which are added to ReorderBuffer should be cleaned up
once they have been processed irrespective of whether they are
decoded/sent downstream or not. In this case I see the sequence hash
being cleaned up for the sequence related transaction in Hayato's
reproducer. See attached patch with a diagnostic change and the output
below (notice sequence cleanup called on transaction 767).
2023-12-14 21:06:36.756 IST [386957] LOG: logical decoding found
initial starting point at 0/15B2F68
2023-12-14 21:06:36.756 IST [386957] DETAIL: Waiting for transactions
(approximately 1) older than 767 to end.
2023-12-14 21:06:36.756 IST [386957] STATEMENT: SELECT
pg_create_logical_replication_slot('slot', 'test_decoding');
2023-12-14 21:07:05.679 IST [386957] LOG: XXX: smgr_decode. snapshot
is SNAPBUILD_BUILDING_SNAPSHOT
2023-12-14 21:07:05.679 IST [386957] STATEMENT: SELECT
pg_create_logical_replication_slot('slot', 'test_decoding');
2023-12-14 21:07:05.679 IST [386957] LOG: XXX: seq_decode. snapshot
is SNAPBUILD_BUILDING_SNAPSHOT
2023-12-14 21:07:05.679 IST [386957] STATEMENT: SELECT
pg_create_logical_replication_slot('slot', 'test_decoding');
2023-12-14 21:07:05.679 IST [386957] LOG: XXX: skipped
2023-12-14 21:07:05.679 IST [386957] STATEMENT: SELECT
pg_create_logical_replication_slot('slot', 'test_decoding');
2023-12-14 21:07:05.710 IST [386957] LOG: logical decoding found
initial consistent point at 0/15B3388
2023-12-14 21:07:05.710 IST [386957] DETAIL: Waiting for transactions
(approximately 1) older than 768 to end.
2023-12-14 21:07:05.710 IST [386957] STATEMENT: SELECT
pg_create_logical_replication_slot('slot', 'test_decoding');
2023-12-14 21:07:39.292 IST [386298] LOG: checkpoint starting: time
2023-12-14 21:07:40.919 IST [386957] LOG: XXX: seq_decode. snapshot
is SNAPBUILD_FULL_SNAPSHOT
2023-12-14 21:07:40.919 IST [386957] STATEMENT: SELECT
pg_create_logical_replication_slot('slot', 'test_decoding');
2023-12-14 21:07:40.919 IST [386957] LOG: XXX: the sequence is transactional
2023-12-14 21:07:40.919 IST [386957] STATEMENT: SELECT
pg_create_logical_replication_slot('slot', 'test_decoding');
2023-12-14 21:07:40.919 IST [386957] LOG: sequence cleanup called on
transaction 767
2023-12-14 21:07:40.919 IST [386957] STATEMENT: SELECT
pg_create_logical_replication_slot('slot', 'test_decoding');
2023-12-14 21:07:40.919 IST [386957] LOG: logical decoding found
consistent point at 0/15B3518
2023-12-14 21:07:40.919 IST [386957] DETAIL: There are no running transactions.
2023-12-14 21:07:40.919 IST [386957] STATEMENT: SELECT
pg_create_logical_replication_slot('slot', 'test_decoding');
We see similar output when pg_logical_slot_get_changes() is called.
I haven't found the code path from where the sequence cleanup gets
called. But it's being called. Am I missing something?
--
Best Wishes,
Ashutosh Bapat
diff --git a/src/backend/replication/logical/decode.c b/src/backend/replication/logical/decode.c
index 73e38cafd8..8e2ebbd8e0 100644
--- a/src/backend/replication/logical/decode.c
+++ b/src/backend/replication/logical/decode.c
@@ -1543,10 +1543,15 @@ smgr_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
elog(LOG, "XXX: smgr_decode. snapshot is %s", SnapBuildIdentify(builder));
/*
- * If we don't have snapshot or we are just fast-forwarding, there is no
- * point in decoding relfilenode information.
+ * If we are not building snapshot yet or we are just fast-forwarding, there
+ * is no point in decoding relfilenode information. If the sequence
+ * associated with relfilenode here is changed in the same transaction but
+ * after snapshot was built, the relfilenode needs to be present in the
+ * sequence hash table so that the change can be deemed as transactional.
+ * Otherwise it will not be queued. Hence we process this change even before
+ * we have built snapshot.
*/
- if (SnapBuildCurrentState(builder) < SNAPBUILD_FULL_SNAPSHOT ||
+ if (SnapBuildCurrentState(builder) < SNAPBUILD_BUILDING_SNAPSHOT ||
ctx->fast_forward)
{
elog(LOG, "XXX: skipped");
diff --git a/src/backend/replication/logical/reorderbuffer.c b/src/backend/replication/logical/reorderbuffer.c
index f7d9a26cfc..25a742fef0 100644
--- a/src/backend/replication/logical/reorderbuffer.c
+++ b/src/backend/replication/logical/reorderbuffer.c
@@ -1078,6 +1078,8 @@ ReorderBufferSequenceCleanup(ReorderBuffer *rb, ReorderBufferTXN *txn)
HASH_REMOVE, NULL) == NULL)
elog(ERROR, "hash table corrupted");
}
+
+ elog(LOG, "sequence cleanup called on transaction %d", txn->xid);
}
/*
Attachments:
[text/plain] sequence_clean_elog.txt (1.7K, ../../CAExHW5tR+rmrarU1m0DCmhq0e7d00XDbdhcHJV+iGkcPTfo8hg@mail.gmail.com/2-sequence_clean_elog.txt)
download | inline diff:
diff --git a/src/backend/replication/logical/decode.c b/src/backend/replication/logical/decode.c
index 73e38cafd8..8e2ebbd8e0 100644
--- a/src/backend/replication/logical/decode.c
+++ b/src/backend/replication/logical/decode.c
@@ -1543,10 +1543,15 @@ smgr_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
elog(LOG, "XXX: smgr_decode. snapshot is %s", SnapBuildIdentify(builder));
/*
- * If we don't have snapshot or we are just fast-forwarding, there is no
- * point in decoding relfilenode information.
+ * If we are not building snapshot yet or we are just fast-forwarding, there
+ * is no point in decoding relfilenode information. If the sequence
+ * associated with relfilenode here is changed in the same transaction but
+ * after snapshot was built, the relfilenode needs to be present in the
+ * sequence hash table so that the change can be deemed as transactional.
+ * Otherwise it will not be queued. Hence we process this change even before
+ * we have built snapshot.
*/
- if (SnapBuildCurrentState(builder) < SNAPBUILD_FULL_SNAPSHOT ||
+ if (SnapBuildCurrentState(builder) < SNAPBUILD_BUILDING_SNAPSHOT ||
ctx->fast_forward)
{
elog(LOG, "XXX: skipped");
diff --git a/src/backend/replication/logical/reorderbuffer.c b/src/backend/replication/logical/reorderbuffer.c
index f7d9a26cfc..25a742fef0 100644
--- a/src/backend/replication/logical/reorderbuffer.c
+++ b/src/backend/replication/logical/reorderbuffer.c
@@ -1078,6 +1078,8 @@ ReorderBufferSequenceCleanup(ReorderBuffer *rb, ReorderBufferTXN *txn)
HASH_REMOVE, NULL) == NULL)
elog(ERROR, "hash table corrupted");
}
+
+ elog(LOG, "sequence cleanup called on transaction %d", txn->xid);
}
/*
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2023-12-15 02:33 Amit Kapila <[email protected]>
parent: Ashutosh Bapat <[email protected]>
0 siblings, 1 reply; 127+ messages in thread
From: Amit Kapila @ 2023-12-15 02:33 UTC (permalink / raw)
To: Ashutosh Bapat <[email protected]>; +Cc: Dilip Kumar <[email protected]>; Tomas Vondra <[email protected]>; Hayato Kuroda (Fujitsu) <[email protected]>; Zhijie Hou (Fujitsu) <[email protected]>; PostgreSQL Hackers <[email protected]>; Masahiko Sawada <[email protected]>; Peter Eisentraut <[email protected]>
On Thu, Dec 14, 2023 at 9:14 PM Ashutosh Bapat
<[email protected]> wrote:
>
> On Thu, Dec 14, 2023 at 2:51 PM Amit Kapila <[email protected]> wrote:
> >
> > It can only be cleaned if we process it but xact_decode won't allow us
> > to process it and I don't think it would be a good idea to add another
> > hack for sequences here. See below code:
> >
> > xact_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
> > {
> > SnapBuild *builder = ctx->snapshot_builder;
> > ReorderBuffer *reorder = ctx->reorder;
> > XLogReaderState *r = buf->record;
> > uint8 info = XLogRecGetInfo(r) & XLOG_XACT_OPMASK;
> >
> > /*
> > * If the snapshot isn't yet fully built, we cannot decode anything, so
> > * bail out.
> > */
> > if (SnapBuildCurrentState(builder) < SNAPBUILD_FULL_SNAPSHOT)
> > return;
>
> That may be true for a transaction which is decoded, but I think all
> the transactions which are added to ReorderBuffer should be cleaned up
> once they have been processed irrespective of whether they are
> decoded/sent downstream or not. In this case I see the sequence hash
> being cleaned up for the sequence related transaction in Hayato's
> reproducer.
>
It was because the test you are using was not designed to show the
problem I mentioned. In this case, the rollback was after a full
snapshot state was reached.
--
With Regards,
Amit Kapila.
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2023-12-19 12:54 Christophe Pettus <[email protected]>
parent: Tomas Vondra <[email protected]>
0 siblings, 1 reply; 127+ messages in thread
From: Christophe Pettus @ 2023-12-19 12:54 UTC (permalink / raw)
To: Tomas Vondra <[email protected]>; +Cc: Amit Kapila <[email protected]>; Dilip Kumar <[email protected]>; Hayato Kuroda (Fujitsu) <[email protected]>; Zhijie Hou (Fujitsu) <[email protected]>; Ashutosh Bapat <[email protected]>; PostgreSQL Hackers <[email protected]>; Masahiko Sawada <[email protected]>; Peter Eisentraut <[email protected]>
Hi,
I wanted to hop in here on one particular issue:
> On Dec 12, 2023, at 02:01, Tomas Vondra <[email protected]> wrote:
> - desirability of the feature: Random IDs (UUIDs etc.) are likely a much
> better solution for distributed (esp. active-active) systems. But there
> are important use cases that are likely to keep using regular sequences
> (online upgrades of single-node instances, existing systems, ...).
+1.
Right now, the lack of sequence replication is a rather large foot-gun on logical replication upgrades. Copying the sequences over during the cutover period is doable, of course, but:
(a) There's no out-of-the-box tooling that does it, so everyone has to write some scripts just for that one function.
(b) It's one more thing that extends the cutover window.
I don't think it is a good idea to make it mandatory: for example, there's a strong use case for replicating a table but not a sequence associated with it. But it's definitely a missing feature in logical replication.
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2023-12-21 13:17 Tomas Vondra <[email protected]>
parent: Christophe Pettus <[email protected]>
0 siblings, 1 reply; 127+ messages in thread
From: Tomas Vondra @ 2023-12-21 13:17 UTC (permalink / raw)
To: Christophe Pettus <[email protected]>; +Cc: Amit Kapila <[email protected]>; Dilip Kumar <[email protected]>; Hayato Kuroda (Fujitsu) <[email protected]>; Zhijie Hou (Fujitsu) <[email protected]>; Ashutosh Bapat <[email protected]>; PostgreSQL Hackers <[email protected]>; Masahiko Sawada <[email protected]>; Peter Eisentraut <[email protected]>
On 12/19/23 13:54, Christophe Pettus wrote:
> Hi,
>
> I wanted to hop in here on one particular issue:
>
>> On Dec 12, 2023, at 02:01, Tomas Vondra <[email protected]> wrote:
>> - desirability of the feature: Random IDs (UUIDs etc.) are likely a much
>> better solution for distributed (esp. active-active) systems. But there
>> are important use cases that are likely to keep using regular sequences
>> (online upgrades of single-node instances, existing systems, ...).
>
> +1.
>
> Right now, the lack of sequence replication is a rather large
> foot-gun on logical replication upgrades. Copying the sequences
> over during the cutover period is doable, of course, but:
>
> (a) There's no out-of-the-box tooling that does it, so everyone has
> to write some scripts just for that one function.
>
> (b) It's one more thing that extends the cutover window.
>
I agree it's an annoying gap for this use case. But if this is the only
use cases, maybe a better solution would be to provide such tooling
instead of adding it to the logical decoding?
It might seem a bit strange if most data is copied by replication
directly, while sequences need special handling, ofc.
> I don't think it is a good idea to make it mandatory: for example,
> there's a strong use case for replicating a table but not a sequence
> associated with it. But it's definitely a missing feature in
> logical replication.
I don't think the plan was to make replication of sequences mandatory,
certainly not with the built-in replication. If you don't add sequences
to the publication, the sequence changes will be skipped.
But it still needs to be part of the decoding, which adds overhead for
all logical decoding uses, even if the sequence changes end up being
discarded. That's somewhat annoying, especially considering sequences
are fairly common part of the WAL stream.
regards
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2023-12-21 14:04 Tomas Vondra <[email protected]>
parent: Amit Kapila <[email protected]>
0 siblings, 1 reply; 127+ messages in thread
From: Tomas Vondra @ 2023-12-21 14:04 UTC (permalink / raw)
To: Amit Kapila <[email protected]>; Ashutosh Bapat <[email protected]>; +Cc: Dilip Kumar <[email protected]>; Hayato Kuroda (Fujitsu) <[email protected]>; Zhijie Hou (Fujitsu) <[email protected]>; PostgreSQL Hackers <[email protected]>; Masahiko Sawada <[email protected]>; Peter Eisentraut <[email protected]>
On 12/15/23 03:33, Amit Kapila wrote:
> On Thu, Dec 14, 2023 at 9:14 PM Ashutosh Bapat
> <[email protected]> wrote:
>>
>> On Thu, Dec 14, 2023 at 2:51 PM Amit Kapila <[email protected]> wrote:
>>>
>>> It can only be cleaned if we process it but xact_decode won't allow us
>>> to process it and I don't think it would be a good idea to add another
>>> hack for sequences here. See below code:
>>>
>>> xact_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
>>> {
>>> SnapBuild *builder = ctx->snapshot_builder;
>>> ReorderBuffer *reorder = ctx->reorder;
>>> XLogReaderState *r = buf->record;
>>> uint8 info = XLogRecGetInfo(r) & XLOG_XACT_OPMASK;
>>>
>>> /*
>>> * If the snapshot isn't yet fully built, we cannot decode anything, so
>>> * bail out.
>>> */
>>> if (SnapBuildCurrentState(builder) < SNAPBUILD_FULL_SNAPSHOT)
>>> return;
>>
>> That may be true for a transaction which is decoded, but I think all
>> the transactions which are added to ReorderBuffer should be cleaned up
>> once they have been processed irrespective of whether they are
>> decoded/sent downstream or not. In this case I see the sequence hash
>> being cleaned up for the sequence related transaction in Hayato's
>> reproducer.
>>
>
> It was because the test you are using was not designed to show the
> problem I mentioned. In this case, the rollback was after a full
> snapshot state was reached.
>
Right, I haven't tried to reproduce this, but it very much looks like we
the entry would not be removed if the xact aborts/commits before the
snapshot reaches FULL state.
I suppose one way to deal with this would be to first check if an entry
for the same relfilenode exists. If it does, the original transaction
must have terminated, but we haven't cleaned it up yet - in which case
we can just "move" the relfilenode to the new one.
However, can't that happen even with full snapshots? I mean, let's say a
transaction creates a relfilenode and terminates without writing an
abort record (surely that's possible, right?). And then another xact
comes and generates the same relfilenode (presumably that's unlikely,
but perhaps possible?). Aren't we in pretty much the same situation,
until the next RUNNING_XACTS cleans up the hash table?
I think tracking all relfilenodes would fix the original issue (with
treating some changes as transactional), and the tweak that "moves" the
relfilenode to the new xact would fix this other issue too.
That being said, I feel a bit uneasy about it, for similar reasons as
Amit. If we start processing records before full snapshot, that seems
like moving the assumptions a bit. For example it means we'd create
ReorderBufferTXN entries for cases that'd have skipped before. OTOH this
is (or should be) only a very temporary period while starting the
replication, I believe.
regards
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2024-01-11 16:26 Tomas Vondra <[email protected]>
parent: Tomas Vondra <[email protected]>
0 siblings, 1 reply; 127+ messages in thread
From: Tomas Vondra @ 2024-01-11 16:26 UTC (permalink / raw)
To: Amit Kapila <[email protected]>; Ashutosh Bapat <[email protected]>; +Cc: Dilip Kumar <[email protected]>; Hayato Kuroda (Fujitsu) <[email protected]>; Zhijie Hou (Fujitsu) <[email protected]>; PostgreSQL Hackers <[email protected]>; Masahiko Sawada <[email protected]>; Peter Eisentraut <[email protected]>
Hi,
Here's new version of this patch series. It rebases the 2023/12/03
version, and there's a couple improvements to address the performance
and correctness questions.
Since the 2023/12/03 version was posted, there were a couple off-list
discussions with several people - with Amit, as mentioned in [1], and
then also internally and at pgconf.eu.
My personal (very brief) takeaway from these discussions is this:
1) desirability: We want a built-in way to handle sequences in logical
replication. I think everyone agrees this is not a way to do distributed
sequences in an active-active setups, but that there are other use cases
that need this feature - typically upgrades / logical failover.
Multiple approaches were discussed (support in logical replication or a
separate tool to be executed on the logical replica). Both might work,
people usually end up with some sort of custom tool anyway. But it's
cumbersome, and the consensus seems the logical rep feature is better.
2) performance: There was concern about the performance impact, and that
it affects everyone, including those who don't replicate sequences (as
the overhead is mostly incurred before calls to output plugin etc.).
I do agree with this, but I don't think sequences can be decoded in a
much cheaper way. There was a proposal [2] that maybe we could batch the
non-transactional sequences changes in the "next" transaction, and
distribute them similarly to SnapBuildDistributeNewCatalogSnapshot()
distributes catalog snapshots.
But I doubt that'd actually work. Or more precisely - if we can make the
code work, I think it would not solve the issue for some common cases.
Consider for example a case with many concurrent top-level transactions,
making this quite expensive. And I'd bet sequence changes are far more
common than catalog changes.
However, I think we ultimately agreed that the overhead is acceptable if
it only applies to use cases that actually need to decode sequences. So
if there was a way to skip sequence decoding when not necessary, that
would work. Unfortunately, that can't be based on simply checking which
callbacks are defined by the output plugin, because e.g. pgoutput needs
to handle both cases (so the callbacks need to be defined). Nor it can
be determined based on what's included in the publication (as that's not
available that early).
The agreement was that the best way is to have a CREATE SUBSCRIPTION
option that would instruct the upstream to decode sequences. By default
this option is 'off' (because that's the no-overhead case), but it can
be enabled for each subscription.
This is what 0005 implements, and interestingly enough, this is what an
earlier version [3] from 2023/04/02 did.
This means that if you add a sequence to the publication, but leave
"sequences=off" in CREATE SUBSCRIPTION, the sequence won't be replicated
after all. That may seems a bit surprising, and I don't like it, but I
don't think there's a better way to do this.
3) correctness: The last point is about making "transactional" flag
correct when the snapshot state changes mid-transaction, originally
pointed out by Dilip [4]. Per [5] this however happens to work
correctly, because while we identify the change as 'non-transactional'
(which is incorrect), we immediately throw it again (so we don't try to
apply it, which would error-out).
One option would be to document/describe this in the comments, per 0006.
This means that when ReorderBufferSequenceIsTransactional() returns
true, it's correct. But if it returns 'false', it means 'maybe'. I agree
it seems a bit strange, but with the extra comments I think it's OK. It
simply means that if we get transactional=false incorrectly, we're
guaranteed to not process it. Maybe we could rename the function to make
this clear from the name.
The other solution proposed in the thread [6] was to always decode the
relfilenode, and add it to the hash table. 0007 does this, and it works.
But I agree this seems possibly worse than 0006 - it means we may be
adding entries to the hash table, and it's not clear when exactly we'll
clean them up etc. It'd be the only place processing stuff before the
snapshots reaches FULL.
I personally would go with 0006, i.e. just explaining why doing it this
way is correct.
regards
[1]
https://www.postgresql.org/message-id/12822961-b7de-9d59-dd27-2e3dc3980c7e%40enterprisedb.com
[2]
https://www.postgresql.org/message-id/CAFiTN-vm3-bGfm-uJdzRLERMHozW8xjZHu4rdmtWR-rP-SJYMQ%40mail.gma...
[3]
https://www.postgresql.org/message-id/1f96b282-cb90-8302-cee8-7b3f5576a31c%40enterprisedb.com
[4]
https://www.postgresql.org/message-id/CAFiTN-vAx-Y%2B19ROKOcWnGf7ix2VOTUebpzteaGw9XQyCAeK6g%40mail.g...
[5]
https://www.postgresql.org/message-id/CAA4eK1LFise9iN%2BNN%3Dagrk4prR1qD%2BebvzNjKAWUog2%2Bhy3HxQ%40...
[6]
https://www.postgresql.org/message-id/CAFiTN-sYpyUBabxopJysqH3DAp4OZUCTi6m_qtgt8d32vDcWSA%40mail.gma...
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
Attachments:
[text/x-patch] v20240111-0001-Logical-decoding-of-sequences.patch (69.0K, ../../[email protected]/2-v20240111-0001-Logical-decoding-of-sequences.patch)
download | inline diff:
From d9d9d23f2e056785507ab51f1eae37ae563b1579 Mon Sep 17 00:00:00 2001
From: Tomas Vondra <[email protected]>
Date: Thu, 2 Nov 2023 15:03:13 +0100
Subject: [PATCH v20240111 1/7] Logical decoding of sequences
Extend the logical decoding to allow decoding sequence changes.
Each sequence change is either transactional or non-transactional,
depending on what whether the sequence relfilenode is visible outside
the transaction.
When a transaction creates a new relfilenode to the sequence (usually
because of CREATE SEQUENCE or ALTER SEQUENCE), all subsequent sequence
changes are treated as transactional - queued just like other changes in
the transaction, and either replayed at commit time or discarded if the
transaction aborts.
All other sequence changes, performed on sequences / relfilenodes
created by earlier transactions, are non-transactional. These changes
are not queued but replayed immediately after decoding. This also means
these changes are not subject to abort.
This mixed behavior is necessary, and is consistent with how sequences
behave in general - calling nextval(s) makes the change visible to
concurrent transactions, and can't be rolled back. But if DDL is
performed on a sequence, changes become transactional - invisible to
concurrent transactions until commit, and subject to abort.
To identify changes that need transactional handling, each transaction
tracks relfilenodes it created in a hash table. Then when a change is
decoded, this hash table is used to check if the relfilenode is new.
For each relfilenode we track the XID of (sub)transaction that created
it, which is needed for cleanup at the transaction end. We don't need to
check the XID to decide if an increment is transactional - if we find a
match in the hash table, it has to be the same transaction.
This requires only minor changes to WAL-logging. We need to ensure the
sequence record is always associated with the subxact XID - until now
the change might have XID 0 if it was the first change in a subxact. But
the sequence might have been created in the same top-level transaction,
in which case it still needs transactional treatment. So we ensure the
XID is assigned when WAL-logging sequence changes, if wal_level=logical.
A patch adding decoding of sequences was originally submitted by Cary
Huang. This commit reworks many important aspects (e.g. the WAL logging
and transactional/non-transactional handling). However, the original
patch and reviews were very useful.
Author: Tomas Vondra, Cary Huang
Reviewed-by: Ashutosh Bapat, Amit Kapila, Peter Eisentraut, Hannu
Krosing, Andres Freund, Zhijie Hou, Peter Smith
Discussion: https://postgr.es/m/[email protected]
Discussion: https://postgr.es/m/[email protected]
Discussion: https://postgr.es/m/[email protected]
---
doc/src/sgml/logicaldecoding.sgml | 62 +-
src/backend/commands/sequence.c | 89 +++
src/backend/replication/logical/decode.c | 228 +++++-
src/backend/replication/logical/logical.c | 100 ++-
.../replication/logical/reorderbuffer.c | 738 +++++++++++++++++-
src/include/access/rmgrlist.h | 4 +-
src/include/replication/decode.h | 2 +
src/include/replication/logical.h | 5 +
src/include/replication/output_plugin.h | 23 +
src/include/replication/reorderbuffer.h | 42 +
src/tools/pgindent/typedefs.list | 1 +
11 files changed, 1282 insertions(+), 12 deletions(-)
diff --git a/doc/src/sgml/logicaldecoding.sgml b/doc/src/sgml/logicaldecoding.sgml
index cd152d4ced9..90e99a95128 100644
--- a/doc/src/sgml/logicaldecoding.sgml
+++ b/doc/src/sgml/logicaldecoding.sgml
@@ -488,6 +488,7 @@ typedef struct OutputPluginCallbacks
LogicalDecodeTruncateCB truncate_cb;
LogicalDecodeCommitCB commit_cb;
LogicalDecodeMessageCB message_cb;
+ LogicalDecodeSequenceCB sequence_cb;
LogicalDecodeFilterByOriginCB filter_by_origin_cb;
LogicalDecodeShutdownCB shutdown_cb;
LogicalDecodeFilterPrepareCB filter_prepare_cb;
@@ -502,6 +503,7 @@ typedef struct OutputPluginCallbacks
LogicalDecodeStreamCommitCB stream_commit_cb;
LogicalDecodeStreamChangeCB stream_change_cb;
LogicalDecodeStreamMessageCB stream_message_cb;
+ LogicalDecodeStreamSequenceCB stream_sequence_cb;
LogicalDecodeStreamTruncateCB stream_truncate_cb;
} OutputPluginCallbacks;
@@ -510,10 +512,13 @@ typedef void (*LogicalOutputPluginInit) (struct OutputPluginCallbacks *cb);
The <function>begin_cb</function>, <function>change_cb</function>
and <function>commit_cb</function> callbacks are required,
while <function>startup_cb</function>, <function>truncate_cb</function>,
- <function>message_cb</function>, <function>filter_by_origin_cb</function>,
+ <function>message_cb</function>, <function>sequence_cb</function>,
+ <function>filter_by_origin_cb</function>,
and <function>shutdown_cb</function> are optional.
If <function>truncate_cb</function> is not set but a
<command>TRUNCATE</command> is to be decoded, the action will be ignored.
+ Similarly, if <function>sequence_cb</function> is not set and a sequence
+ change is to be decoded, the action will be ignored.
</para>
<para>
@@ -521,7 +526,8 @@ typedef void (*LogicalOutputPluginInit) (struct OutputPluginCallbacks *cb);
in-progress transactions. The <function>stream_start_cb</function>,
<function>stream_stop_cb</function>, <function>stream_abort_cb</function>,
<function>stream_commit_cb</function>, and <function>stream_change_cb</function>
- are required, while <function>stream_message_cb</function> and
+ are required, while <function>stream_message_cb</function>,
+ <function>stream_sequence_cb</function> and
<function>stream_truncate_cb</function> are optional. The
<function>stream_prepare_cb</function> is also required if the output
plugin also support two-phase commits.
@@ -840,6 +846,35 @@ typedef void (*LogicalDecodeMessageCB) (struct LogicalDecodingContext *ctx,
</para>
</sect3>
+ <sect3 id="logicaldecoding-output-plugin-sequence">
+ <title>Sequence Callback</title>
+
+ <para>
+ The optional <function>sequence_cb</function> callback is called for
+ actions that update a sequence value.
+<programlisting>
+typedef void (*LogicalDecodeSequenceCB) (struct LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional,
+ int64 value);
+</programlisting>
+ The <parameter>txn</parameter> parameter contains meta information about
+ the transaction the sequence change is part of. Note however that for
+ non-transactional updates, the transaction may be NULL, depending on
+ if the transaction already has an XID assigned.
+ The <parameter>sequence_lsn</parameter> has the WAL location of the
+ sequence update. <parameter>transactional</parameter> determines whether
+ the sequence has to be replayed as part of the transaction or directly.
+
+ The <parameter>value</parameter> parameter describes the sequence change.
+ Note that this may not be the value obtained by the process updating the
+ process, but the future sequence value written to WAL (typically about
+ 32 values ahead).
+ </para>
+ </sect3>
+
<sect3 id="logicaldecoding-output-plugin-filter-prepare">
<title>Prepare Filter Callback</title>
@@ -1051,6 +1086,24 @@ typedef void (*LogicalDecodeStreamMessageCB) (struct LogicalDecodingContext *ctx
</para>
</sect3>
+ <sect3 id="logicaldecoding-output-plugin-stream-sequence">
+ <title>Stream Sequence Callback</title>
+ <para>
+ The optional <function>stream_sequence_cb</function> callback is called
+ for actions that change a sequence in a block of streamed changes
+ (demarcated by <function>stream_start_cb</function> and
+ <function>stream_stop_cb</function> calls).
+<programlisting>
+typedef void (*LogicalDecodeStreamSequenceCB) (struct LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional,
+ int64 value);
+</programlisting>
+ </para>
+ </sect3>
+
<sect3 id="logicaldecoding-output-plugin-stream-truncate">
<title>Stream Truncate Callback</title>
<para>
@@ -1230,8 +1283,9 @@ OutputPluginWrite(ctx, true);
in-progress transactions. There are multiple required streaming callbacks
(<function>stream_start_cb</function>, <function>stream_stop_cb</function>,
<function>stream_abort_cb</function>, <function>stream_commit_cb</function>
- and <function>stream_change_cb</function>) and two optional callbacks
- (<function>stream_message_cb</function> and <function>stream_truncate_cb</function>).
+ and <function>stream_change_cb</function>) and three optional callbacks
+ (<function>stream_message_cb</function>, <function>stream_sequence_cb</function>,
+ and <function>stream_truncate_cb</function>).
Also, if streaming of two-phase commands is to be supported, then additional
callbacks must be provided. (See <xref linkend="logicaldecoding-two-phase-commits"/>
for details).
diff --git a/src/backend/commands/sequence.c b/src/backend/commands/sequence.c
index f7803744a5b..af27d96ecd7 100644
--- a/src/backend/commands/sequence.c
+++ b/src/backend/commands/sequence.c
@@ -204,6 +204,14 @@ DefineSequence(ParseState *pstate, CreateSeqStmt *seq)
stmt->tablespacename = NULL;
stmt->if_not_exists = seq->if_not_exists;
+ /*
+ * Make sure the relfilenode creation (in DefineRelation) is associated
+ * with the XID, so that decoding of sequences can handle this as a
+ * transactional change.
+ */
+ if (XLogLogicalInfoActive())
+ GetCurrentTransactionId();
+
address = DefineRelation(stmt, RELKIND_SEQUENCE, seq->ownerId, NULL, NULL);
seqoid = address.objectId;
Assert(seqoid != InvalidOid);
@@ -302,6 +310,13 @@ ResetSequence(Oid seq_relid)
seq->is_called = false;
seq->log_cnt = 0;
+ /*
+ * Make sure the relfilenode creation is associated with the XID, so that
+ * decoding of sequences can handle this as a transactional change.
+ */
+ if (XLogLogicalInfoActive())
+ GetCurrentTransactionId();
+
/*
* Create a new storage file for the sequence.
*/
@@ -392,8 +407,19 @@ fill_seq_fork_with_data(Relation rel, HeapTuple tuple, ForkNumber forkNum)
/* check the comment above nextval_internal()'s equivalent call. */
if (RelationNeedsWAL(rel))
+ {
GetTopTransactionId();
+ /*
+ * Make sure the subtransaction has a XID assigned, so that the
+ * sequence change WAL record is properly associated with it. This
+ * matters for sequences created/altered in a transaction - the
+ * subsequent changes need to be handled as transactional.
+ */
+ if (XLogLogicalInfoActive())
+ GetCurrentTransactionId();
+ }
+
START_CRIT_SECTION();
MarkBufferDirty(buf);
@@ -417,6 +443,9 @@ fill_seq_fork_with_data(Relation rel, HeapTuple tuple, ForkNumber forkNum)
XLogRegisterData((char *) &xlrec, sizeof(xl_seq_rec));
XLogRegisterData((char *) tuple->t_data, tuple->t_len);
+ /* allow filtering by origin on a sequence update */
+ XLogSetRecordFlags(XLOG_INCLUDE_ORIGIN);
+
recptr = XLogInsert(RM_SEQ_ID, XLOG_SEQ_LOG);
PageSetLSN(page, recptr);
@@ -499,6 +528,14 @@ AlterSequence(ParseState *pstate, AlterSeqStmt *stmt)
if (RelationNeedsWAL(seqrel))
GetTopTransactionId();
+ /*
+ * Make sure the relfilenode creation is associated with the XID, so
+ * that decoding of sequences can handle this as a transactional
+ * change.
+ */
+ if (XLogLogicalInfoActive())
+ GetCurrentTransactionId();
+
/*
* Create a new storage file for the sequence, making the state
* changes transactional.
@@ -548,8 +585,19 @@ SequenceChangePersistence(Oid relid, char newrelpersistence)
/* check the comment above nextval_internal()'s equivalent call. */
if (RelationNeedsWAL(seqrel))
+ {
GetTopTransactionId();
+ /*
+ * Make sure the subtransaction has a XID assigned, so that the
+ * sequence change WAL record is properly associated with it. This
+ * matters for sequences created/altered in a transaction - the
+ * subsequent changes need to be handled as transactional.
+ */
+ if (XLogLogicalInfoActive())
+ GetCurrentTransactionId();
+ }
+
(void) read_seq_tuple(seqrel, &buf, &seqdatatuple);
RelationSetNewRelfilenumber(seqrel, newrelpersistence);
fill_seq_with_data(seqrel, &seqdatatuple);
@@ -792,10 +840,34 @@ nextval_internal(Oid relid, bool check_permissions)
* It's sufficient to ensure the toplevel transaction has an xid, no need
* to assign xids subxacts, that'll already trigger an appropriate wait.
* (Have to do that here, so we're outside the critical section)
+ *
+ * We have to ensure we have a proper XID, which will be included in the
+ * XLOG record by XLogRecordAssemble. Otherwise the first nextval() in a
+ * subxact (without any preceding changes) would get XID 0, and it would
+ * then be impossible to decide which top xact it belongs to. It'd also
+ * trigger assert in DecodeSequence. We only do that with
+ * wal_level=logical, though.
+ *
+ * XXX This might seem unnecessary, because if there's no XID the xact
+ * couldn't have done anything important yet, e.g. it could not have
+ * created a sequence. But that's incorrect, because of subxacts. The
+ * current subtransaction might not have done anything yet (thus no XID),
+ * but an earlier one might have created the sequence.
*/
if (logit && RelationNeedsWAL(seqrel))
+ {
GetTopTransactionId();
+ /*
+ * Make sure the subtransaction has a XID assigned, so that the
+ * sequence change WAL record is properly associated with it. This
+ * matters for sequences created/altered in a transaction - the
+ * subsequent changes need to be handled as transactional.
+ */
+ if (XLogLogicalInfoActive())
+ GetCurrentTransactionId();
+ }
+
/* ready to change the on-disk (or really, in-buffer) tuple */
START_CRIT_SECTION();
@@ -835,6 +907,9 @@ nextval_internal(Oid relid, bool check_permissions)
XLogRegisterData((char *) &xlrec, sizeof(xl_seq_rec));
XLogRegisterData((char *) seqdatatuple.t_data, seqdatatuple.t_len);
+ /* allow filtering by origin on a sequence update */
+ XLogSetRecordFlags(XLOG_INCLUDE_ORIGIN);
+
recptr = XLogInsert(RM_SEQ_ID, XLOG_SEQ_LOG);
PageSetLSN(page, recptr);
@@ -996,8 +1071,19 @@ do_setval(Oid relid, int64 next, bool iscalled)
/* check the comment above nextval_internal()'s equivalent call. */
if (RelationNeedsWAL(seqrel))
+ {
GetTopTransactionId();
+ /*
+ * Make sure the subtransaction has a XID assigned, so that the
+ * sequence change WAL record is properly associated with it. This
+ * matters for sequences created/altered in a transaction - the
+ * subsequent changes need to be handled as transactional.
+ */
+ if (XLogLogicalInfoActive())
+ GetCurrentTransactionId();
+ }
+
/* ready to change the on-disk (or really, in-buffer) tuple */
START_CRIT_SECTION();
@@ -1021,6 +1107,9 @@ do_setval(Oid relid, int64 next, bool iscalled)
XLogRegisterData((char *) &xlrec, sizeof(xl_seq_rec));
XLogRegisterData((char *) seqdatatuple.t_data, seqdatatuple.t_len);
+ /* allow filtering by origin on a sequence update */
+ XLogSetRecordFlags(XLOG_INCLUDE_ORIGIN);
+
recptr = XLogInsert(RM_SEQ_ID, XLOG_SEQ_LOG);
PageSetLSN(page, recptr);
diff --git a/src/backend/replication/logical/decode.c b/src/backend/replication/logical/decode.c
index b3f8f908d18..3462c545bb7 100644
--- a/src/backend/replication/logical/decode.c
+++ b/src/backend/replication/logical/decode.c
@@ -35,6 +35,7 @@
#include "access/xlogrecord.h"
#include "access/xlogutils.h"
#include "catalog/pg_control.h"
+#include "catalog/storage_xlog.h"
#include "replication/decode.h"
#include "replication/logical.h"
#include "replication/message.h"
@@ -42,6 +43,7 @@
#include "replication/reorderbuffer.h"
#include "replication/snapbuild.h"
#include "storage/standby.h"
+#include "commands/sequence.h"
/* individual record(group)'s handlers */
static void DecodeInsert(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
@@ -63,6 +65,7 @@ static void DecodePrepare(LogicalDecodingContext *ctx, XLogRecordBuffer *buf,
/* common function to decode tuples */
static void DecodeXLogTuple(char *data, Size len, ReorderBufferTupleBuf *tuple);
+static void DecodeSeqTuple(char *data, Size len, ReorderBufferTupleBuf *tuple);
/* helper functions for decoding transactions */
static inline bool FilterPrepare(LogicalDecodingContext *ctx,
@@ -629,7 +632,7 @@ logicalmsg_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
* We need to set processing_required flag to notify the message's
* existence to the caller. Usually, the flag is set when either the
* COMMIT or ABORT records are decoded, but this must be turned on
- * here because the non-transactional logical message is decoded
+ * here because the non-transactional sequence change is decoded
* without waiting for these records.
*/
if (!message->transactional)
@@ -1320,3 +1323,226 @@ DecodeTXNNeedSkip(LogicalDecodingContext *ctx, XLogRecordBuffer *buf,
return false;
}
+
+/*
+ * DecodeSeqTuple
+ * decode tuple describing the sequence change
+ *
+ * Sequences are represented as a table with a single row, which gets updated by
+ * nextval() etc. The tuple is stored in WAL right after the xl_seq_rec, so we
+ * simply copy it into the tuplebuf (similar to seq_redo).
+ */
+static void
+DecodeSeqTuple(char *data, Size len, ReorderBufferTupleBuf *tuple)
+{
+ int datalen = len - sizeof(xl_seq_rec) - SizeofHeapTupleHeader;
+
+ Assert(datalen >= 0);
+
+ tuple->tuple.t_len = datalen + SizeofHeapTupleHeader;
+
+ ItemPointerSetInvalid(&tuple->tuple.t_self);
+
+ tuple->tuple.t_tableOid = InvalidOid;
+
+ memcpy(((char *) tuple->tuple.t_data),
+ data + sizeof(xl_seq_rec),
+ SizeofHeapTupleHeader + datalen);
+}
+
+/*
+ * Handle sequence decode
+ *
+ * Decoding sequences is a bit tricky, because while most sequence actions
+ * are non-transactional (immediately visible and not subject to rollback),
+ * some need to be handled as transactional (e.g. for sequences created in
+ * a transaction are invisible until that transaction commits).
+ *
+ * By default, a sequence change is non-transactional - we must not queue
+ * it in a transaction as other changes, because the transaction might get
+ * rolled back and we'd discard the change. But that'd be incorrect, as the
+ * change is already visible to other processes accessing the sequence. If
+ * we discard the change, the downstream will not be notified about the
+ * change, and may fall behind.
+ *
+ * On the other hand, the sequence may be created in a transaction. In this
+ * case we *have to* queue it in the transaction just like other changes,
+ * because we don't want to process changes for sequences that may be unknown
+ * outside the transaction - the downstream might get confused about which
+ * sequence it's related to etc.
+ */
+void
+seq_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
+{
+ SnapBuild *builder = ctx->snapshot_builder;
+ ReorderBufferTupleBuf *tuplebuf;
+ RelFileLocator target_locator;
+ XLogReaderState *r = buf->record;
+ char *tupledata = NULL;
+ Size tuplelen;
+ Size datalen = 0;
+ TransactionId xid = XLogRecGetXid(r);
+ uint8 info = XLogRecGetInfo(buf->record) & ~XLR_INFO_MASK;
+ Snapshot snapshot = NULL;
+ RepOriginId origin_id = XLogRecGetOrigin(r);
+ bool transactional;
+
+ /* ignore sequences when the plugin does not have the callbacks */
+ if (!ctx->sequences)
+ return;
+
+ /* only decode changes flagged with XLOG_SEQ_LOG */
+ if (info != XLOG_SEQ_LOG)
+ elog(ERROR, "unexpected RM_SEQ_ID record type: %u", info);
+
+ ReorderBufferProcessXid(ctx->reorder, XLogRecGetXid(r), buf->origptr);
+
+ /*
+ * If we don't have snapshot or we are just fast-forwarding, there is no
+ * point in decoding sequences.
+ */
+ if (SnapBuildCurrentState(builder) < SNAPBUILD_FULL_SNAPSHOT)
+ return;
+
+ /* only interested in our database */
+ XLogRecGetBlockTag(r, 0, &target_locator, NULL, NULL);
+ if (target_locator.dbOid != ctx->slot->data.database)
+ return;
+
+ /* output plugin doesn't look for this origin, no need to queue */
+ if (FilterByOrigin(ctx, XLogRecGetOrigin(r)))
+ return;
+
+ /* time to decode the sequence data from the tuple */
+ tupledata = XLogRecGetData(r);
+ datalen = XLogRecGetDataLen(r);
+ tuplelen = datalen - SizeOfHeapHeader - sizeof(xl_seq_rec);
+
+ /* the sequence should not have changed without data */
+ if (!datalen || !tupledata)
+ elog(ERROR, "sequence decode missing tuple data");
+
+ tuplebuf = ReorderBufferGetTupleBuf(ctx->reorder, tuplelen);
+ DecodeSeqTuple(tupledata, datalen, tuplebuf);
+
+ /*
+ * Should we handle the sequence change as transactional or not?
+ *
+ * If the relfilenode was created in the current transaction, treat the
+ * change as transactional and queue it. Otherwise it needs to be treated
+ * as non-transactional, in which case we just send it to the plugin right
+ * away.
+ */
+ transactional = ReorderBufferSequenceIsTransactional(ctx->reorder,
+ target_locator,
+ NULL);
+
+ /* Skip the change if already processed (per the snapshot). */
+ if (transactional &&
+ !SnapBuildProcessChange(builder, xid, buf->origptr))
+ return;
+ else if (!transactional &&
+ (SnapBuildCurrentState(builder) != SNAPBUILD_CONSISTENT ||
+ SnapBuildXactNeedsSkip(builder, buf->origptr)))
+ return;
+
+ /*
+ * We also skip decoding in fast_forward mode. This check must be last
+ * because we don't want to set the processing_required flag unless we
+ * have a decodable sequence change.
+ */
+ if (ctx->fast_forward)
+ {
+ /*
+ * We need to set processing_required flag to notify the sequence
+ * change existence to the caller. Usually, the flag is set when
+ * either the COMMIT or ABORT records are decoded, but this must be
+ * turned on here because the non-transactional logical message is
+ * decoded without waiting for these records.
+ */
+ if (!transactional)
+ ctx->processing_required = true;
+
+ return;
+ }
+
+ /*
+ * If this is a non-transactional change, get the snapshot we're expected
+ * to use. We only get here when the snapshot is consistent, and the
+ * change is not meant to be skipped.
+ *
+ * For transactional changes we don't need a snapshot, we'll use the
+ * regular snapshot maintained by ReorderBuffer. We just leave it NULL.
+ */
+ if (!transactional)
+ snapshot = SnapBuildGetOrBuildSnapshot(builder);
+
+ /* Queue the change (or send immediately if not transactional). */
+ ReorderBufferQueueSequence(ctx->reorder, xid, snapshot, buf->endptr,
+ origin_id, target_locator, transactional,
+ tuplebuf);
+}
+
+/*
+ * Decode relfilenode change
+ *
+ * Decode SMGR records, so that we can later decide which sequence changes
+ * need to be treated as transactional. Most sequence changes are going to
+ * be non-transactional (applied as if outside the decoded transaction, not
+ * subject to rollback, etc.). Changes for sequences created/altered in the
+ * transaction need to be handled as transactional (i.e. applied as part of
+ * the decoded transaction, same as all other changes).
+ *
+ * To decide which of those cases is it we decode XLOG_SMGR_CREATE records
+ * and track relfilenodes created in each (sub)transaction. Changes for
+ * these relfilenodes are then queued and treated as transactional, while
+ * remaining changes are treated as non-transactional.
+ *
+ * We only care about XLOG_SMGR_CREATE records for "our" database (logical
+ * decoding is restricted to a single database), and we do the filtering
+ * and skipping, as appropriate.
+ */
+void
+smgr_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
+{
+ SnapBuild *builder = ctx->snapshot_builder;
+ XLogReaderState *r = buf->record;
+ TransactionId xid = XLogRecGetXid(r);
+ uint8 info = XLogRecGetInfo(buf->record) & ~XLR_INFO_MASK;
+ xl_smgr_create *xlrec;
+
+ /*
+ * Bail out when not decoding sequences, which is currently the only case
+ * when we need to know about relfilenodes created in a transaction.
+ */
+ if (!ctx->sequences)
+ return;
+
+ /*
+ * We only care about XLOG_SMGR_CREATE, because that's what determines if
+ * the following sequence changes are transactional.
+ */
+ if (info != XLOG_SMGR_CREATE)
+ return;
+
+ ReorderBufferProcessXid(ctx->reorder, XLogRecGetXid(r), buf->origptr);
+
+ /*
+ * If we don't have snapshot or we are just fast-forwarding, there is no
+ * point in decoding relfilenode information.
+ */
+ if (SnapBuildCurrentState(builder) < SNAPBUILD_FULL_SNAPSHOT ||
+ ctx->fast_forward)
+ return;
+
+ /* only interested in our database */
+ xlrec = (xl_smgr_create *) XLogRecGetData(r);
+ if (xlrec->rlocator.dbOid != ctx->slot->data.database)
+ return;
+
+ /* output plugin doesn't look for this origin, no need to queue */
+ if (FilterByOrigin(ctx, XLogRecGetOrigin(r)))
+ return;
+
+ ReorderBufferAddRelFileLocator(ctx->reorder, xid, buf->endptr, xlrec->rlocator);
+}
diff --git a/src/backend/replication/logical/logical.c b/src/backend/replication/logical/logical.c
index ca09c683f11..8cadc0e9e6f 100644
--- a/src/backend/replication/logical/logical.c
+++ b/src/backend/replication/logical/logical.c
@@ -75,6 +75,9 @@ static void truncate_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
static void message_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
XLogRecPtr message_lsn, bool transactional,
const char *prefix, Size message_size, const char *message);
+static void sequence_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ bool transactional, int64 value);
/* streaming callbacks */
static void stream_start_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
@@ -92,6 +95,9 @@ static void stream_change_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn
static void stream_message_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
XLogRecPtr message_lsn, bool transactional,
const char *prefix, Size message_size, const char *message);
+static void stream_sequence_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ bool transactional, int64 value);
static void stream_truncate_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
int nrelations, Relation relations[], ReorderBufferChange *change);
@@ -223,13 +229,14 @@ StartupDecodingContext(List *output_plugin_options,
ctx->reorder->apply_truncate = truncate_cb_wrapper;
ctx->reorder->commit = commit_cb_wrapper;
ctx->reorder->message = message_cb_wrapper;
+ ctx->reorder->sequence = sequence_cb_wrapper;
/*
* To support streaming, we require start/stop/abort/commit/change
- * callbacks. The message and truncate callbacks are optional, similar to
- * regular output plugins. We however enable streaming when at least one
- * of the methods is enabled so that we can easily identify missing
- * methods.
+ * callbacks. The message, sequence and truncate callbacks are optional,
+ * similar to regular output plugins. We however enable streaming when at
+ * least one of the methods is enabled so that we can easily identify
+ * missing methods.
*
* We decide it here, but only check it later in the wrappers.
*/
@@ -239,6 +246,7 @@ StartupDecodingContext(List *output_plugin_options,
(ctx->callbacks.stream_commit_cb != NULL) ||
(ctx->callbacks.stream_change_cb != NULL) ||
(ctx->callbacks.stream_message_cb != NULL) ||
+ (ctx->callbacks.stream_sequence_cb != NULL) ||
(ctx->callbacks.stream_truncate_cb != NULL);
/*
@@ -256,6 +264,7 @@ StartupDecodingContext(List *output_plugin_options,
ctx->reorder->stream_commit = stream_commit_cb_wrapper;
ctx->reorder->stream_change = stream_change_cb_wrapper;
ctx->reorder->stream_message = stream_message_cb_wrapper;
+ ctx->reorder->stream_sequence = stream_sequence_cb_wrapper;
ctx->reorder->stream_truncate = stream_truncate_cb_wrapper;
@@ -289,6 +298,13 @@ StartupDecodingContext(List *output_plugin_options,
*/
ctx->reorder->update_progress_txn = update_progress_txn_cb_wrapper;
+ /*
+ * To support logical decoding of sequences, we require the sequence
+ * callback. We decide it here, but only check it later in the wrappers.
+ */
+ ctx->sequences = ((ctx->callbacks.sequence_cb != NULL) ||
+ (ctx->callbacks.stream_sequence_cb != NULL));
+
ctx->out = makeStringInfo();
ctx->prepare_write = prepare_write;
ctx->write = do_write;
@@ -1262,6 +1278,42 @@ message_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
error_context_stack = errcallback.previous;
}
+static void
+sequence_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel, bool transactional,
+ int64 value)
+{
+ LogicalDecodingContext *ctx = cache->private_data;
+ LogicalErrorCallbackState state;
+ ErrorContextCallback errcallback;
+
+ Assert(!ctx->fast_forward);
+
+ if (ctx->callbacks.sequence_cb == NULL)
+ return;
+
+ /* Push callback + info on the error context stack */
+ state.ctx = ctx;
+ state.callback_name = "sequence";
+ state.report_location = sequence_lsn;
+ errcallback.callback = output_plugin_error_callback;
+ errcallback.arg = (void *) &state;
+ errcallback.previous = error_context_stack;
+ error_context_stack = &errcallback;
+
+ /* set output state */
+ ctx->accept_writes = true;
+ ctx->write_xid = txn != NULL ? txn->xid : InvalidTransactionId;
+ ctx->write_location = sequence_lsn;
+
+ /* do the actual work: call callback */
+ ctx->callbacks.sequence_cb(ctx, txn, sequence_lsn, rel,
+ transactional, value);
+
+ /* Pop the error context stack */
+ error_context_stack = errcallback.previous;
+}
+
static void
stream_start_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
XLogRecPtr first_lsn)
@@ -1577,6 +1629,46 @@ stream_message_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
error_context_stack = errcallback.previous;
}
+static void
+stream_sequence_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ bool transactional, int64 value)
+{
+ LogicalDecodingContext *ctx = cache->private_data;
+ LogicalErrorCallbackState state;
+ ErrorContextCallback errcallback;
+
+ Assert(!ctx->fast_forward);
+
+ /* We're only supposed to call this when streaming is supported. */
+ Assert(ctx->streaming);
+
+ /* this callback is optional */
+ if (ctx->callbacks.stream_sequence_cb == NULL)
+ return;
+
+ /* Push callback + info on the error context stack */
+ state.ctx = ctx;
+ state.callback_name = "stream_sequence";
+ state.report_location = sequence_lsn;
+ errcallback.callback = output_plugin_error_callback;
+ errcallback.arg = (void *) &state;
+ errcallback.previous = error_context_stack;
+ error_context_stack = &errcallback;
+
+ /* set output state */
+ ctx->accept_writes = true;
+ ctx->write_xid = txn != NULL ? txn->xid : InvalidTransactionId;
+ ctx->write_location = sequence_lsn;
+
+ /* do the actual work: call callback */
+ ctx->callbacks.sequence_cb(ctx, txn, sequence_lsn, rel,
+ transactional, value);
+
+ /* Pop the error context stack */
+ error_context_stack = errcallback.previous;
+}
+
static void
stream_truncate_cb_wrapper(ReorderBuffer *cache, ReorderBufferTXN *txn,
int nrelations, Relation relations[],
diff --git a/src/backend/replication/logical/reorderbuffer.c b/src/backend/replication/logical/reorderbuffer.c
index d1334ffb550..6ac91068799 100644
--- a/src/backend/replication/logical/reorderbuffer.c
+++ b/src/backend/replication/logical/reorderbuffer.c
@@ -77,6 +77,41 @@
* a bit more memory to the oldest subtransactions, because it's likely
* they are the source for the next sequence of changes.
*
+ * When decoding sequences, we differentiate between transactional and
+ * non-transactional changes. Sequence change is "non-transactional" if
+ * it was visible to outside transactions immediately, and would not be
+ * undone by a rollback. Most sequence changes are non-transactional, as
+ * the changes performed by nextval() are visible and can't be undone.
+ * Non-transactional changes are replayed immediately, as if outside any
+ * transaction (and thus also not subject to rollback etc.).
+ *
+ * However, if a transaction modifies the sequence in a way that creates
+ * a new relfilenode (e.g. by running ALTER SEQUENCE), we consider the
+ * subsequent changes "transactional" and queue them in the transaction
+ * that performed them, just like other changes. If that transaction rolls
+ * back, these sequence changes are discarded too (together with the
+ * new relfilenode).
+ *
+ * This mixed behavior is necessary - sequences are non-transactional
+ * (e.g. ROLLBACK does not undo that the sequence moved forward). But
+ * changes on new relfilenodes need to be handled as transactional.
+ *
+ * To decide if a sequence change is transactional, we maintain a hash
+ * table of relfilenodes created in each (sub)transaction, along with
+ * the XID of the (sub)transaction that created the relfilenode. The
+ * entries from subtransactions are copied to the top-level transaction
+ * to make checks cheaper. The hash table gets cleaned up when the
+ * transaction completes (commit/abort).
+ *
+ * The XID of the subxact that created the relfilenode is necessary, as
+ * that's where the sequence changes need to be queued - if this subxact
+ * rolls back, we want to discard the changes tied to this relfilenode
+ * (even if done in some other subtransaction).
+ *
+ * The XID may be valid even for non-transactional sequences - we simply
+ * keep the XID decoded from WAL, it's up to the reorderbuffer to decide
+ * if the change is transactional.
+ *
* -------------------------------------------------------------------------
*/
#include "postgres.h"
@@ -91,6 +126,7 @@
#include "access/xact.h"
#include "access/xlog_internal.h"
#include "catalog/catalog.h"
+#include "commands/sequence.h"
#include "lib/binaryheap.h"
#include "miscadmin.h"
#include "pgstat.h"
@@ -116,6 +152,13 @@ typedef struct ReorderBufferTXNByIdEnt
ReorderBufferTXN *txn;
} ReorderBufferTXNByIdEnt;
+/* entry for hash table we use to track sequences created in running xacts */
+typedef struct ReorderBufferSequenceEnt
+{
+ RelFileLocator rlocator;
+ ReorderBufferTXN *txn;
+} ReorderBufferSequenceEnt;
+
/* data structures for (relfilelocator, ctid) => (cmin, cmax) mapping */
typedef struct ReorderBufferTupleCidKey
{
@@ -225,6 +268,7 @@ static void ReorderBufferTransferSnapToParent(ReorderBufferTXN *txn,
ReorderBufferTXN *subtxn);
static void AssertTXNLsnOrder(ReorderBuffer *rb);
+static void AssertCheckSequences(ReorderBuffer *rb);
/* ---------------------------------------
* support functions for lsn-order iterating over the ->changes of a
@@ -456,12 +500,48 @@ ReorderBufferReturnTXN(ReorderBuffer *rb, ReorderBufferTXN *txn)
txn->invalidations = NULL;
}
+ if (txn->sequences_hash != NULL)
+ {
+ hash_destroy(txn->sequences_hash);
+ txn->sequences_hash = NULL;
+ }
+
/* Reset the toast hash */
ReorderBufferToastReset(rb, txn);
pfree(txn);
}
+/*
+ * Initialize hash table of relfilenodes created by the transaction.
+ *
+ * Each entry maps the relfilenode to the (sub)transaction that created the
+ * relfilenode - which is also the transaction the sequence change needs to
+ * be part of (in transactional case).
+ *
+ * We don't do this in ReorderBufferGetTXN because that'd allocate the hash
+ * for all transactions, and we expect new relfilenodes to be fairly rare.
+ * So only do that when adding the first entry.
+ */
+static void
+ReorderBufferTXNSequencesInit(ReorderBuffer *rb, ReorderBufferTXN *txn)
+{
+ HASHCTL hash_ctl;
+
+ /* bail out if already initialized */
+ if (txn->sequences_hash)
+ return;
+
+ /* hash table of sequences, mapping relfilelocator to transaction */
+ hash_ctl.keysize = sizeof(RelFileLocator);
+ hash_ctl.entrysize = sizeof(ReorderBufferSequenceEnt);
+ hash_ctl.hcxt = rb->context;
+
+ /* we expect relfilenodes to be created only rarely, so 32 seems enough */
+ txn->sequences_hash = hash_create("ReorderBufferTXNSequenceHash", 32, &hash_ctl,
+ HASH_ELEM | HASH_BLOBS | HASH_CONTEXT);
+}
+
/*
* Get a fresh ReorderBufferChange.
*/
@@ -536,6 +616,13 @@ ReorderBufferReturnChange(ReorderBuffer *rb, ReorderBufferChange *change,
change->data.truncate.relids = NULL;
}
break;
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
+ if (change->data.sequence.tuple)
+ {
+ ReorderBufferReturnTupleBuf(rb, change->data.sequence.tuple);
+ change->data.sequence.tuple = NULL;
+ }
+ break;
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM:
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_ABORT:
case REORDER_BUFFER_CHANGE_INTERNAL_COMMAND_ID:
@@ -804,7 +891,8 @@ ReorderBufferQueueChange(ReorderBuffer *rb, TransactionId xid, XLogRecPtr lsn,
change->action == REORDER_BUFFER_CHANGE_DELETE ||
change->action == REORDER_BUFFER_CHANGE_INTERNAL_SPEC_INSERT ||
change->action == REORDER_BUFFER_CHANGE_TRUNCATE ||
- change->action == REORDER_BUFFER_CHANGE_MESSAGE)
+ change->action == REORDER_BUFFER_CHANGE_MESSAGE ||
+ change->action == REORDER_BUFFER_CHANGE_SEQUENCE)
{
ReorderBufferTXN *toptxn = rbtxn_get_toptxn(txn);
@@ -895,6 +983,352 @@ ReorderBufferQueueMessage(ReorderBuffer *rb, TransactionId xid,
}
}
+/*
+ * Treat the sequence change as transactional?
+ *
+ * To decide if a sequence change should be handled as transactional or applied
+ * immediately, we need to decide if the relfilenode was created by a running
+ * transaction. Each transaction has a hash table of such relfilenodes.
+ *
+ * A naive approach would be to just loop through all transactions and check
+ * each of them, but there may be many (easily thousands of) subtransactions,
+ * and the check happens for each sequence change. So this could be very costly.
+ *
+ * To limit the number of transactions we need to check, we transfer the
+ * relfilenodes to the top-level xact, which allows us to check just the
+ * top-level xacts (and there should be a very limited number of those). We
+ * expect relfilenode creation to be a quite rare thing, so this should not
+ * waste too much memory. OTOH sequence changes are very common, so making
+ * the check cheaper seems like a good trade off.
+ *
+ * Returns true for transactional changes, false otherwise.
+ *
+ * Optionaly (when the "xid" parameter is set) returns XID of the (sub)xact
+ * to use for queueing transactional changes.
+ *
+ * XXX As an optimization, maybe we should try searching the current xact (or
+ * it's top-level xact) first. If the change is transactional, where we'd
+ * find the match.
+ */
+bool
+ReorderBufferSequenceIsTransactional(ReorderBuffer *rb,
+ RelFileLocator rlocator,
+ TransactionId *xid)
+{
+ bool found = false;
+ dlist_iter iter;
+
+ AssertCheckSequences(rb);
+
+ /*
+ * Walk all top-level transactions (some of which may be subxacts, except
+ * that we haven't processed the assignments yet), and check if any of
+ * them created the relfilenode.
+ */
+ dlist_foreach(iter, &rb->toplevel_by_lsn)
+ {
+ ReorderBufferSequenceEnt *entry;
+ ReorderBufferTXN *txn = dlist_container(ReorderBufferTXN, node,
+ iter.cur);
+
+ /* transaction has no relfilenodes at all */
+ if (!txn->sequences_hash)
+ continue;
+
+ entry = hash_search(txn->sequences_hash,
+ (void *) &rlocator,
+ HASH_FIND,
+ &found);
+
+ /*
+ * If we found an entry with matching relfilenode, we're done - we
+ * have to treat the sequence change as transactional, and replay it
+ * in the same (sub)transaction just like any other change.
+ *
+ * Optionally set XID of the (sub)xact that created the relfilenode.
+ */
+ if (found)
+ {
+ if (xid)
+ *xid = entry->txn->xid;
+
+ break;
+ }
+ }
+
+ return found;
+}
+
+/*
+ * Cleanup sequences after a subtransaction got committed/aborted.
+ *
+ * The hash table will get destroyed in ReorderBufferReturnTXN, so we don't
+ * need to worry about that. But the entries were copied to the parent xact,
+ * and that's still being decoded - we make sure to remove the entries from
+ * the committed/aborted one.
+ */
+static void
+ReorderBufferSequenceCleanup(ReorderBufferTXN *txn)
+{
+ HASH_SEQ_STATUS scan_status;
+ ReorderBufferSequenceEnt *ent;
+
+ /* Bail out if not a subxact, or if there are no entries. */
+ if (!rbtxn_is_known_subxact(txn))
+ return;
+
+ if (!txn->sequences_hash)
+ return;
+
+ /*
+ * Scan the top-level transaction hash and remove the entries from it. If
+ * we have entries for subxact, the top-level hash must have been
+ * initialized.
+ */
+ hash_seq_init(&scan_status, txn->sequences_hash);
+ while ((ent = (ReorderBufferSequenceEnt *) hash_seq_search(&scan_status)) != NULL)
+ {
+ if (hash_search(txn->toptxn->sequences_hash,
+ (void *) &ent->rlocator,
+ HASH_REMOVE, NULL) == NULL)
+ elog(ERROR, "hash table corrupted");
+ }
+}
+
+/*
+ * A transactional sequence change is queued to be processed upon commit
+ * and a non-transactional change gets processed immediately.
+ *
+ * A sequence update may be both transactional and non-transactional. When
+ * created in a running transaction, treat it as transactional and queue
+ * the change in it. Otherwise treat it as non-transactional, so that we
+ * don't forget the change in case of a rollback.
+ */
+void
+ReorderBufferQueueSequence(ReorderBuffer *rb, TransactionId xid,
+ Snapshot snapshot, XLogRecPtr lsn, RepOriginId origin_id,
+ RelFileLocator rlocator, bool transactional,
+ ReorderBufferTupleBuf *tuplebuf)
+{
+ AssertCheckSequences(rb);
+
+ /*
+ * Change needs to be handled as transactional, because the sequence was
+ * created in a transaction that is still seen as running. In that case
+ * all the changes need to be queued in that transaction, we must not send
+ * the changes to the downstream until the transaction commits.
+ *
+ * There's a bit of a trouble with subtransactions - we can't queue it
+ * into the subxact, because it might be rolled back and we'd lose the
+ * change (and that'd be wrong, as sequences are not transactional). But
+ * we also can't queue it into the top-level transaction, because the
+ * relfilenode creation is transactional - if that subxact gets aborted,
+ * we must throw away the changes too.
+ *
+ * So we need to queue it into the same (sub)xact that created the new
+ * sequence relfilenode, which is why we have the XID in the hash table.
+ */
+ if (transactional)
+ {
+ MemoryContext oldcontext;
+ ReorderBufferChange *change;
+
+ /* allocate and queue the transactional sequence change */
+ oldcontext = MemoryContextSwitchTo(rb->context);
+
+ change = ReorderBufferGetChange(rb);
+
+ change->action = REORDER_BUFFER_CHANGE_SEQUENCE;
+ change->origin_id = origin_id;
+
+ memcpy(&change->data.sequence.locator, &rlocator, sizeof(RelFileLocator));
+
+ change->data.sequence.tuple = tuplebuf;
+
+ /* lookup the XID for transaction that created the relfilenode */
+ ReorderBufferSequenceIsTransactional(rb, rlocator, &xid);
+
+ /* the XID should be valid for a transactional change */
+ Assert(TransactionIdIsValid(xid));
+
+ /* add it to the same (sub)xact that created that relfilenode */
+ ReorderBufferQueueChange(rb, xid, lsn, change, false);
+
+ MemoryContextSwitchTo(oldcontext);
+ }
+ else
+ {
+ /*
+ * This change is for a sequence that was not created in any running
+ * transaction, so we treat it as non-transactional and just send it
+ * to the output plugin directly.
+ */
+ ReorderBufferTXN *txn = NULL;
+ volatile Snapshot snapshot_now = snapshot;
+ bool using_subtxn;
+
+ /* non-transactional changes require a valid snapshot and XID */
+ Assert(snapshot_now);
+ Assert(TransactionIdIsValid(xid));
+
+ /* Make sure the sequence is not in any of the hash tables */
+ Assert(!ReorderBufferSequenceIsTransactional(rb, rlocator, NULL));
+
+ txn = ReorderBufferTXNByXid(rb, xid, true, NULL, lsn, true);
+
+ /* setup snapshot to allow catalog access */
+ SetupHistoricSnapshot(snapshot_now, NULL);
+
+ /*
+ * Decoding needs access to syscaches et al., which in turn use
+ * heavyweight locks and such. Thus we need to have enough state
+ * around to keep track of those. The easiest way is to simply use a
+ * transaction internally. That also allows us to easily enforce that
+ * nothing writes to the database by checking for xid assignments.
+ *
+ * When we're called via the SQL SRF there's already a transaction
+ * started, so start an explicit subtransaction there.
+ */
+ using_subtxn = IsTransactionOrTransactionBlock();
+
+ PG_TRY();
+ {
+ Relation relation;
+ HeapTuple tuple;
+ Form_pg_sequence_data seq;
+ Oid reloid;
+ int64 value;
+
+ if (using_subtxn)
+ BeginInternalSubTransaction("sequence");
+ else
+ StartTransactionCommand();
+
+ reloid = RelidByRelfilenumber(rlocator.spcOid, rlocator.relNumber);
+
+ if (reloid == InvalidOid)
+ elog(ERROR, "could not map filenode \"%s\" to relation OID",
+ relpathperm(rlocator,
+ MAIN_FORKNUM));
+
+ relation = RelationIdGetRelation(reloid);
+
+ if (!RelationIsValid(relation))
+ elog(ERROR, "could not open relation with OID %u (for filenode \"%s\")",
+ reloid,
+ relpathperm(rlocator, MAIN_FORKNUM));
+
+ tuple = &tuplebuf->tuple;
+ seq = (Form_pg_sequence_data) GETSTRUCT(tuple);
+
+ /*
+ * Calculate the first value of the next batch (at which point we
+ * generate and decode another WAL record).
+ */
+ value = seq->last_value;
+ value += (seq->is_called) ? seq->log_cnt : 0;
+
+ rb->sequence(rb, txn, lsn, relation, transactional, value);
+
+ RelationClose(relation);
+
+ TeardownHistoricSnapshot(false);
+
+ AbortCurrentTransaction();
+
+ if (using_subtxn)
+ RollbackAndReleaseCurrentSubTransaction();
+ }
+ PG_CATCH();
+ {
+ TeardownHistoricSnapshot(true);
+
+ AbortCurrentTransaction();
+
+ if (using_subtxn)
+ RollbackAndReleaseCurrentSubTransaction();
+
+ PG_RE_THROW();
+ }
+ PG_END_TRY();
+ }
+}
+
+/*
+ * ReorderBufferAddRelFileLocator
+ * Add newly created relfilenode to the hash table for transaction.
+ *
+ * If the transaction is already known to be a subtransaction, we add the same
+ * entry into the parent transaction (but it still points at the subxact, so
+ * that we know where to queue changes or what to discard in case of an abort).
+ */
+void
+ReorderBufferAddRelFileLocator(ReorderBuffer *rb, TransactionId xid,
+ XLogRecPtr lsn, RelFileLocator rlocator)
+{
+ bool found;
+ ReorderBufferSequenceEnt *entry;
+ ReorderBufferTXN *txn;
+
+ AssertCheckSequences(rb);
+
+ /*
+ * We only care about sequence relfilenodes for now, and those always have
+ * a XID. So if there's no XID, don't bother adding them to the hash.
+ */
+ if (!TransactionIdIsValid(xid))
+ return;
+
+ /*
+ * This might be the first change decoded for this transaction, so make
+ * sure we create it if needed.
+ */
+ txn = ReorderBufferTXNByXid(rb, xid, true, NULL, lsn, true);
+
+ /*
+ * First add it to the top-level transaction, but make sure it links to
+ * the correct subtransaction (so that we add later changes to it).
+ */
+ if (txn->toptxn)
+ {
+ /* make sure the hash table is initialized */
+ ReorderBufferTXNSequencesInit(rb, txn->toptxn);
+
+ /* search the lookup table */
+ entry = hash_search(txn->toptxn->sequences_hash,
+ (void *) &rlocator,
+ HASH_ENTER,
+ &found);
+
+ /*
+ * We've just decoded creation of the relfilenode, so if we found it
+ * in the hash table, something is wrong.
+ */
+ Assert(!found);
+
+ entry->txn = txn;
+ }
+
+ /* make sure the hash table is initialized */
+ ReorderBufferTXNSequencesInit(rb, txn);
+
+ /* search the lookup table */
+ entry = hash_search(txn->sequences_hash,
+ (void *) &rlocator,
+ HASH_ENTER,
+ &found);
+
+ /*
+ * We've just decoded creation of the relfilenode, so if we found it in
+ * the hash table, something is wrong.
+ */
+ Assert(!found);
+
+ entry->txn = txn;
+
+ AssertCheckSequences(rb);
+}
+
/*
* AssertTXNLsnOrder
* Verify LSN ordering of transaction lists in the reorderbuffer
@@ -969,6 +1403,161 @@ AssertTXNLsnOrder(ReorderBuffer *rb)
#endif
}
+/*
+ * AssertCheckSequences
+ * Verify consistency of hash tables tracking new relfilenodes.
+ *
+ * We check that the hash table in the top-level transaction is consistent with
+ * respect to the hash tables in it's subxacts. That is, each entry has to be
+ * either from the top-level xact or one of it's subtransactions. Likewise, each
+ * entry in subxact hash table has to have a match in the top-level hash table.
+ */
+static void
+AssertCheckSequences(ReorderBuffer *rb)
+{
+#ifdef USE_ASSERT_CHECKING
+ LogicalDecodingContext *ctx = rb->private_data;
+ dlist_iter iter;
+
+ /*
+ * Skip the verification if we don't reach the LSN at which we start
+ * decoding the contents of transactions yet because until we reach the
+ * LSN, we could have transactions that don't have the association between
+ * the top-level transaction and subtransaction yet and consequently have
+ * the same LSN. We don't guarantee this association until we try to
+ * decode the actual contents of transaction. The ordering of the records
+ * prior to the start_decoding_at LSN should have been checked before the
+ * restart.
+ */
+ if (SnapBuildXactNeedsSkip(ctx->snapshot_builder, ctx->reader->EndRecPtr))
+ return;
+
+ /*
+ * Make sure the relfilenodes from subxacts are properly recorded in the
+ * top-level transaction hash table.
+ */
+ dlist_foreach(iter, &rb->toplevel_by_lsn)
+ {
+ int nentries = 0,
+ nsubentries = 0;
+ dlist_iter subiter;
+ ReorderBufferTXN *txn = dlist_container(ReorderBufferTXN, node,
+ iter.cur);
+
+ /*
+ * We don't skip top-level transactions without relfilenodes, because
+ * there might be a subtransaction with some, and we want to detect
+ * such cases too.
+ */
+ if (txn->sequences_hash)
+ nentries = hash_get_num_entries(txn->sequences_hash);
+
+ /* walk all subxacts, and check the hash table in each one */
+ dlist_foreach(subiter, &txn->subtxns)
+ {
+ HASH_SEQ_STATUS scan_status;
+ ReorderBufferSequenceEnt *entry;
+
+ ReorderBufferTXN *subtxn = dlist_container(ReorderBufferTXN, node,
+ subiter.cur);
+
+ /*
+ * If this subxact has no relfilenodes, skip it. We'll do the
+ * check in the opposite direction (that all top-level
+ * relfilenodes are in the correct subxact) later.
+ */
+ if (!subtxn->sequences_hash)
+ continue;
+
+ /* add number of relfilenodes in this subxact */
+ nsubentries += hash_get_num_entries(subtxn->sequences_hash);
+
+ /*
+ * Check that all subxact relfilenodes are in the top-level txn
+ * too, and are pointing to this subtransaction.
+ */
+ hash_seq_init(&scan_status, subtxn->sequences_hash);
+ while ((entry = (ReorderBufferSequenceEnt *) hash_seq_search(&scan_status)) != NULL)
+ {
+ bool found = false;
+ ReorderBufferSequenceEnt *subentry;
+
+ /* search for the same relfilenode in the top-level txn */
+ subentry = hash_search(txn->sequences_hash,
+ (void *) &entry->rlocator,
+ HASH_FIND,
+ &found);
+
+ /*
+ * The top-level txn hash should have the relfilenode too, and
+ * it should point to this subxact.
+ */
+ Assert(found);
+
+ /*
+ * The entry has to point to the subxact - there's no subxact
+ * "below" this one to which the relfilenode could belong.
+ */
+ Assert(subentry->txn == subtxn);
+ }
+ }
+
+ /*
+ * We shouldn't have more relfilenodes in subtransactions than in the
+ * top-level one. There might be relfilenodes in the top-level one
+ * directly, so this needs to be inequality.
+ */
+ Assert(nentries >= nsubentries);
+
+ /*
+ * Now do the check in the opposite direction - check that every entry
+ * in the top-level txn (except those pointing to the top-level txn
+ * itself) point to one of the subxacts, and there's an entry in the
+ * subxact hash.
+ */
+ if (txn->sequences_hash)
+ {
+ HASH_SEQ_STATUS scan_status;
+ ReorderBufferSequenceEnt *entry;
+
+ hash_seq_init(&scan_status, txn->sequences_hash);
+ while ((entry = (ReorderBufferSequenceEnt *) hash_seq_search(&scan_status)) != NULL)
+ {
+ bool found = false;
+ ReorderBufferSequenceEnt *subentry;
+
+ /* Skip entries for the top-level txn itself. */
+ if (entry->txn == txn)
+ continue;
+
+ /* Is it a subxact of this txn? */
+ Assert(rbtxn_is_known_subxact(entry->txn));
+ Assert(entry->txn->toptxn == txn);
+
+ /*
+ * Search for the same relfilenode in the subxact (it should
+ * be initialized, as we expect it to contain the
+ * relfilenode).
+ */
+ subentry = hash_search(entry->txn->sequences_hash,
+ (void *) &entry->rlocator,
+ HASH_FIND,
+ &found);
+
+ Assert(found);
+
+ /*
+ * Check the txn pointer in the top-level hash table entry is
+ * consistent with the subxact hash table (we already checked
+ * the subxact entry points to the subxact itself).
+ */
+ Assert(subentry->txn = entry->txn);
+ }
+ }
+ }
+#endif
+}
+
/*
* AssertChangeLsnOrder
*
@@ -1094,6 +1683,9 @@ ReorderBufferAssignChild(ReorderBuffer *rb, TransactionId xid,
subtxn->toplevel_xid = xid;
Assert(subtxn->nsubtxns == 0);
+ /* There should be no sequence relfilenodes in the subxact yet. */
+ Assert(!subtxn->sequences_hash);
+
/* set the reference to top-level transaction */
subtxn->toptxn = txn;
@@ -1997,6 +2589,35 @@ ReorderBufferApplyMessage(ReorderBuffer *rb, ReorderBufferTXN *txn,
change->data.msg.message);
}
+/*
+ * Helper function for ReorderBufferProcessTXN for applying sequences.
+ */
+static inline void
+ReorderBufferApplySequence(ReorderBuffer *rb, ReorderBufferTXN *txn,
+ Relation relation, ReorderBufferChange *change,
+ bool streaming)
+{
+ HeapTuple tuple;
+ Form_pg_sequence_data seq;
+ int64 value;
+
+ tuple = &change->data.sequence.tuple->tuple;
+ seq = (Form_pg_sequence_data) GETSTRUCT(tuple);
+
+ value = seq->last_value;
+ value += (seq->is_called) ? seq->log_cnt : 0;
+
+ /*
+ * When called from ReorderBufferApplySequence, we're applying changes
+ * accumulated in a ReorderBufferTXN, so all those are transactional
+ * changes of sequences.
+ */
+ if (streaming)
+ rb->stream_sequence(rb, txn, change->lsn, relation, true, value);
+ else
+ rb->sequence(rb, txn, change->lsn, relation, true, value);
+}
+
/*
* Function to store the command id and snapshot at the end of the current
* stream so that we can reuse the same while sending the next stream.
@@ -2442,6 +3063,31 @@ ReorderBufferProcessTXN(ReorderBuffer *rb, ReorderBufferTXN *txn,
case REORDER_BUFFER_CHANGE_INTERNAL_TUPLECID:
elog(ERROR, "tuplecid value in changequeue");
break;
+
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
+ Assert(snapshot_now);
+
+ reloid = RelidByRelfilenumber(change->data.sequence.locator.spcOid,
+ change->data.sequence.locator.relNumber);
+
+ if (reloid == InvalidOid)
+ elog(ERROR, "could not map filenode \"%s\" to relation OID",
+ relpathperm(change->data.sequence.locator,
+ MAIN_FORKNUM));
+
+ relation = RelationIdGetRelation(reloid);
+
+ if (!RelationIsValid(relation))
+ elog(ERROR, "could not open relation with OID %u (for filenode \"%s\")",
+ reloid,
+ relpathperm(change->data.sequence.locator,
+ MAIN_FORKNUM));
+
+ if (RelationIsLogicallyLogged(relation))
+ ReorderBufferApplySequence(rb, txn, relation, change, streaming);
+
+ RelationClose(relation);
+ break;
}
/*
@@ -2872,6 +3518,10 @@ ReorderBufferFinishPrepared(ReorderBuffer *rb, TransactionId xid,
/* cleanup: make sure there's no cache pollution */
ReorderBufferExecuteInvalidations(txn->ninvalidations,
txn->invalidations);
+
+ /* cleanup: remove sequence relfilenodes from the top-level txn */
+ ReorderBufferSequenceCleanup(txn);
+
ReorderBufferCleanupTXN(rb, txn);
}
@@ -2921,8 +3571,13 @@ ReorderBufferAbort(ReorderBuffer *rb, TransactionId xid, XLogRecPtr lsn,
/* cosmetic... */
txn->final_lsn = lsn;
+ /* remove sequence relfilenodes from the top-level txn */
+ ReorderBufferSequenceCleanup(txn);
+
/* remove potential on-disk data, and deallocate */
ReorderBufferCleanupTXN(rb, txn);
+
+ AssertCheckSequences(rb);
}
/*
@@ -2958,8 +3613,13 @@ ReorderBufferAbortOld(ReorderBuffer *rb, TransactionId oldestRunningXid)
if (rbtxn_is_streamed(txn))
rb->stream_abort(rb, txn, InvalidXLogRecPtr);
+ /* remove sequence relfilenodes from the top-level txn */
+ ReorderBufferSequenceCleanup(txn);
+
/* remove potential on-disk data, and deallocate this tx */
ReorderBufferCleanupTXN(rb, txn);
+
+ AssertCheckSequences(rb);
}
else
return;
@@ -3008,6 +3668,9 @@ ReorderBufferForget(ReorderBuffer *rb, TransactionId xid, XLogRecPtr lsn)
else
Assert(txn->ninvalidations == 0);
+ /* remove sequence relfilenodes from a top-level txn */
+ ReorderBufferSequenceCleanup(txn);
+
/* remove potential on-disk data, and deallocate */
ReorderBufferCleanupTXN(rb, txn);
}
@@ -3909,6 +4572,39 @@ ReorderBufferSerializeChange(ReorderBuffer *rb, ReorderBufferTXN *txn,
memcpy(data, change->data.truncate.relids, size);
data += size;
+ break;
+ }
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
+ {
+ char *data;
+ ReorderBufferTupleBuf *tup;
+ Size len = 0;
+
+ tup = change->data.sequence.tuple;
+
+ if (tup)
+ {
+ sz += sizeof(HeapTupleData);
+ len = tup->tuple.t_len;
+ sz += len;
+ }
+
+ /* make sure we have enough space */
+ ReorderBufferSerializeReserve(rb, sz);
+
+ data = ((char *) rb->outbuf) + sizeof(ReorderBufferDiskChange);
+ /* might have been reallocated above */
+ ondisk = (ReorderBufferDiskChange *) rb->outbuf;
+
+ if (len)
+ {
+ memcpy(data, &tup->tuple, sizeof(HeapTupleData));
+ data += sizeof(HeapTupleData);
+
+ memcpy(data, tup->tuple.t_data, len);
+ data += len;
+ }
+
break;
}
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM:
@@ -4173,6 +4869,22 @@ ReorderBufferChangeSize(ReorderBufferChange *change)
{
sz += sizeof(Oid) * change->data.truncate.nrelids;
+ break;
+ }
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
+ {
+ ReorderBufferTupleBuf *tup;
+ Size len = 0;
+
+ tup = change->data.sequence.tuple;
+
+ if (tup)
+ {
+ sz += sizeof(HeapTupleData);
+ len = tup->tuple.t_len;
+ sz += len;
+ }
+
break;
}
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM:
@@ -4476,6 +5188,30 @@ ReorderBufferRestoreChange(ReorderBuffer *rb, ReorderBufferTXN *txn,
break;
}
+
+ case REORDER_BUFFER_CHANGE_SEQUENCE:
+ if (change->data.sequence.tuple)
+ {
+ uint32 tuplelen = ((HeapTuple) data)->t_len;
+
+ change->data.sequence.tuple =
+ ReorderBufferGetTupleBuf(rb, tuplelen - SizeofHeapTupleHeader);
+
+ /* restore ->tuple */
+ memcpy(&change->data.sequence.tuple->tuple, data,
+ sizeof(HeapTupleData));
+ data += sizeof(HeapTupleData);
+
+ /* reset t_data pointer into the new tuplebuf */
+ change->data.sequence.tuple->tuple.t_data =
+ ReorderBufferTupleBufData(change->data.sequence.tuple);
+
+ /* restore tuple data itself */
+ memcpy(change->data.sequence.tuple->tuple.t_data, data, tuplelen);
+ data += tuplelen;
+ }
+ break;
+
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM:
case REORDER_BUFFER_CHANGE_INTERNAL_SPEC_ABORT:
case REORDER_BUFFER_CHANGE_INTERNAL_COMMAND_ID:
diff --git a/src/include/access/rmgrlist.h b/src/include/access/rmgrlist.h
index 78e6b908c6e..c6d902fc077 100644
--- a/src/include/access/rmgrlist.h
+++ b/src/include/access/rmgrlist.h
@@ -27,7 +27,7 @@
/* symbol name, textual name, redo, desc, identify, startup, cleanup, mask, decode */
PG_RMGR(RM_XLOG_ID, "XLOG", xlog_redo, xlog_desc, xlog_identify, NULL, NULL, NULL, xlog_decode)
PG_RMGR(RM_XACT_ID, "Transaction", xact_redo, xact_desc, xact_identify, NULL, NULL, NULL, xact_decode)
-PG_RMGR(RM_SMGR_ID, "Storage", smgr_redo, smgr_desc, smgr_identify, NULL, NULL, NULL, NULL)
+PG_RMGR(RM_SMGR_ID, "Storage", smgr_redo, smgr_desc, smgr_identify, NULL, NULL, NULL, smgr_decode)
PG_RMGR(RM_CLOG_ID, "CLOG", clog_redo, clog_desc, clog_identify, NULL, NULL, NULL, NULL)
PG_RMGR(RM_DBASE_ID, "Database", dbase_redo, dbase_desc, dbase_identify, NULL, NULL, NULL, NULL)
PG_RMGR(RM_TBLSPC_ID, "Tablespace", tblspc_redo, tblspc_desc, tblspc_identify, NULL, NULL, NULL, NULL)
@@ -40,7 +40,7 @@ PG_RMGR(RM_BTREE_ID, "Btree", btree_redo, btree_desc, btree_identify, btree_xlog
PG_RMGR(RM_HASH_ID, "Hash", hash_redo, hash_desc, hash_identify, NULL, NULL, hash_mask, NULL)
PG_RMGR(RM_GIN_ID, "Gin", gin_redo, gin_desc, gin_identify, gin_xlog_startup, gin_xlog_cleanup, gin_mask, NULL)
PG_RMGR(RM_GIST_ID, "Gist", gist_redo, gist_desc, gist_identify, gist_xlog_startup, gist_xlog_cleanup, gist_mask, NULL)
-PG_RMGR(RM_SEQ_ID, "Sequence", seq_redo, seq_desc, seq_identify, NULL, NULL, seq_mask, NULL)
+PG_RMGR(RM_SEQ_ID, "Sequence", seq_redo, seq_desc, seq_identify, NULL, NULL, seq_mask, seq_decode)
PG_RMGR(RM_SPGIST_ID, "SPGist", spg_redo, spg_desc, spg_identify, spg_xlog_startup, spg_xlog_cleanup, spg_mask, NULL)
PG_RMGR(RM_BRIN_ID, "BRIN", brin_redo, brin_desc, brin_identify, NULL, NULL, brin_mask, NULL)
PG_RMGR(RM_COMMIT_TS_ID, "CommitTs", commit_ts_redo, commit_ts_desc, commit_ts_identify, NULL, NULL, NULL, NULL)
diff --git a/src/include/replication/decode.h b/src/include/replication/decode.h
index 17a5b5e95ab..3de02ee4a02 100644
--- a/src/include/replication/decode.h
+++ b/src/include/replication/decode.h
@@ -27,6 +27,8 @@ extern void heap2_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
extern void xact_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
extern void standby_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
extern void logicalmsg_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
+extern void seq_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
+extern void smgr_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf);
extern void LogicalDecodingProcessRecord(LogicalDecodingContext *ctx,
XLogReaderState *record);
diff --git a/src/include/replication/logical.h b/src/include/replication/logical.h
index dc2df4ce92a..4db252f06d4 100644
--- a/src/include/replication/logical.h
+++ b/src/include/replication/logical.h
@@ -90,6 +90,11 @@ typedef struct LogicalDecodingContext
*/
bool twophase;
+ /*
+ * Does the output pluging support decoding of sequence changes?
+ */
+ bool sequences;
+
/*
* Is two-phase option given by output plugin?
*
diff --git a/src/include/replication/output_plugin.h b/src/include/replication/output_plugin.h
index 44988ebdd89..2ae25c6f4fe 100644
--- a/src/include/replication/output_plugin.h
+++ b/src/include/replication/output_plugin.h
@@ -90,6 +90,16 @@ typedef void (*LogicalDecodeMessageCB) (struct LogicalDecodingContext *ctx,
Size message_size,
const char *message);
+/*
+ * Called for the generic logical decoding sequences.
+ */
+typedef void (*LogicalDecodeSequenceCB) (struct LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional,
+ int64 value);
+
/*
* Filter changes by origin.
*/
@@ -201,6 +211,17 @@ typedef void (*LogicalDecodeStreamMessageCB) (struct LogicalDecodingContext *ctx
Size message_size,
const char *message);
+/*
+ * Called for the streaming generic logical decoding sequences from in-progress
+ * transactions.
+ */
+typedef void (*LogicalDecodeStreamSequenceCB) (struct LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional,
+ int64 value);
+
/*
* Callback for streaming truncates from in-progress transactions.
*/
@@ -221,6 +242,7 @@ typedef struct OutputPluginCallbacks
LogicalDecodeTruncateCB truncate_cb;
LogicalDecodeCommitCB commit_cb;
LogicalDecodeMessageCB message_cb;
+ LogicalDecodeSequenceCB sequence_cb;
LogicalDecodeFilterByOriginCB filter_by_origin_cb;
LogicalDecodeShutdownCB shutdown_cb;
@@ -239,6 +261,7 @@ typedef struct OutputPluginCallbacks
LogicalDecodeStreamCommitCB stream_commit_cb;
LogicalDecodeStreamChangeCB stream_change_cb;
LogicalDecodeStreamMessageCB stream_message_cb;
+ LogicalDecodeStreamSequenceCB stream_sequence_cb;
LogicalDecodeStreamTruncateCB stream_truncate_cb;
} OutputPluginCallbacks;
diff --git a/src/include/replication/reorderbuffer.h b/src/include/replication/reorderbuffer.h
index 3e232c6c27f..0134133517a 100644
--- a/src/include/replication/reorderbuffer.h
+++ b/src/include/replication/reorderbuffer.h
@@ -74,6 +74,7 @@ typedef enum ReorderBufferChangeType
REORDER_BUFFER_CHANGE_INTERNAL_SPEC_CONFIRM,
REORDER_BUFFER_CHANGE_INTERNAL_SPEC_ABORT,
REORDER_BUFFER_CHANGE_TRUNCATE,
+ REORDER_BUFFER_CHANGE_SEQUENCE,
} ReorderBufferChangeType;
/* forward declaration */
@@ -167,6 +168,13 @@ typedef struct ReorderBufferChange
uint32 ninvalidations; /* Number of messages */
SharedInvalidationMessage *invalidations; /* invalidation message */
} inval;
+
+ /* Context data for Sequence changes */
+ struct
+ {
+ RelFileLocator locator;
+ ReorderBufferTupleBuf *tuple;
+ } sequence;
} data;
/*
@@ -392,6 +400,12 @@ typedef struct ReorderBufferTXN
*/
HTAB *toast_hash;
+ /*
+ * Sequence relfilenodes created in this transaction (also includes
+ * altered sequences, which assigns new relfilenode).
+ */
+ HTAB *sequences_hash;
+
/*
* non-hierarchical list of subtransactions that are *not* aborted. Only
* used in toplevel transactions.
@@ -470,6 +484,14 @@ typedef void (*ReorderBufferMessageCB) (ReorderBuffer *rb,
const char *prefix, Size sz,
const char *message);
+/* sequence callback signature */
+typedef void (*ReorderBufferSequenceCB) (ReorderBuffer *rb,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional,
+ int64 value);
+
/* begin prepare callback signature */
typedef void (*ReorderBufferBeginPrepareCB) (ReorderBuffer *rb,
ReorderBufferTXN *txn);
@@ -536,6 +558,14 @@ typedef void (*ReorderBufferStreamMessageCB) (
const char *prefix, Size sz,
const char *message);
+/* stream sequence callback signature */
+typedef void (*ReorderBufferStreamSequenceCB) (ReorderBuffer *rb,
+ ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn,
+ Relation rel,
+ bool transactional,
+ int64 value);
+
/* stream truncate callback signature */
typedef void (*ReorderBufferStreamTruncateCB) (
ReorderBuffer *rb,
@@ -592,6 +622,7 @@ struct ReorderBuffer
ReorderBufferApplyTruncateCB apply_truncate;
ReorderBufferCommitCB commit;
ReorderBufferMessageCB message;
+ ReorderBufferSequenceCB sequence;
/*
* Callbacks to be called when streaming a transaction at prepare time.
@@ -611,6 +642,7 @@ struct ReorderBuffer
ReorderBufferStreamCommitCB stream_commit;
ReorderBufferStreamChangeCB stream_change;
ReorderBufferStreamMessageCB stream_message;
+ ReorderBufferStreamSequenceCB stream_sequence;
ReorderBufferStreamTruncateCB stream_truncate;
/*
@@ -696,6 +728,10 @@ extern void ReorderBufferQueueMessage(ReorderBuffer *rb, TransactionId xid,
Snapshot snap, XLogRecPtr lsn,
bool transactional, const char *prefix,
Size message_size, const char *message);
+extern void ReorderBufferQueueSequence(ReorderBuffer *rb, TransactionId xid,
+ Snapshot snapshot, XLogRecPtr lsn, RepOriginId origin_id,
+ RelFileLocator locator, bool transactional,
+ ReorderBufferTupleBuf *tuplebuf);
extern void ReorderBufferCommit(ReorderBuffer *rb, TransactionId xid,
XLogRecPtr commit_lsn, XLogRecPtr end_lsn,
TimestampTz commit_time, RepOriginId origin_id, XLogRecPtr origin_lsn);
@@ -750,4 +786,10 @@ extern void ReorderBufferSetRestartPoint(ReorderBuffer *rb, XLogRecPtr ptr);
extern void StartupReorderBuffer(void);
+extern void ReorderBufferAddRelFileLocator(ReorderBuffer *rb, TransactionId xid,
+ XLogRecPtr lsn, RelFileLocator rlocator);
+extern bool ReorderBufferSequenceIsTransactional(ReorderBuffer *rb,
+ RelFileLocator locator,
+ TransactionId *xid);
+
#endif
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 5fd46b7bd1f..b632fbd7cf3 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -2336,6 +2336,7 @@ ReorderBufferIterTXNState
ReorderBufferMessageCB
ReorderBufferPrepareCB
ReorderBufferRollbackPreparedCB
+ReorderBufferSequenceEnt
ReorderBufferStreamAbortCB
ReorderBufferStreamChangeCB
ReorderBufferStreamCommitCB
--
2.43.0
[text/x-patch] v20240111-0002-Add-decoding-of-sequences-to-test_decoding.patch (20.4K, ../../[email protected]/3-v20240111-0002-Add-decoding-of-sequences-to-test_decoding.patch)
download | inline diff:
From 965576ff4a4fbc94309aa08a6ab0ea5e074df08a Mon Sep 17 00:00:00 2001
From: Tomas Vondra <[email protected]>
Date: Fri, 13 Jan 2023 20:32:44 +0100
Subject: [PATCH v20240111 2/7] Add decoding of sequences to test_decoding
Extends test_decoding with callbacks to decode sequences. This feature
is disabled by default (i.e. sequence changes are omitted from the
output), but may be enabled by passing 'include-sequences' option to the
test_decoding plugin.
Author: Tomas Vondra, Cary Huang
Reviewed-by: Ashutosh Bapat, Amit Kapila, Peter Eisentraut, Hannu
Krosing, Andres Freund, Zhijie Hou
Discussion: https://postgr.es/m/[email protected]
Discussion: https://postgr.es/m/[email protected]
---
contrib/test_decoding/Makefile | 3 +-
contrib/test_decoding/expected/sequence.out | 325 ++++++++++++++++++++
contrib/test_decoding/sql/sequence.sql | 119 +++++++
contrib/test_decoding/test_decoding.c | 87 ++++++
4 files changed, 533 insertions(+), 1 deletion(-)
create mode 100644 contrib/test_decoding/expected/sequence.out
create mode 100644 contrib/test_decoding/sql/sequence.sql
diff --git a/contrib/test_decoding/Makefile b/contrib/test_decoding/Makefile
index c7ce6037064..05c0c5a2f86 100644
--- a/contrib/test_decoding/Makefile
+++ b/contrib/test_decoding/Makefile
@@ -5,7 +5,8 @@ PGFILEDESC = "test_decoding - example of a logical decoding output plugin"
REGRESS = ddl xact rewrite toast permissions decoding_in_xact \
decoding_into_rel binary prepared replorigin time messages \
- spill slot truncate stream stats twophase twophase_stream
+ spill slot truncate stream stats twophase twophase_stream \
+ sequence
ISOLATION = mxact delayed_startup ondisk_startup concurrent_ddl_dml \
oldest_xmin snapshot_transfer subxact_without_top concurrent_stream \
twophase_snapshot slot_creation_error catalog_change_snapshot
diff --git a/contrib/test_decoding/expected/sequence.out b/contrib/test_decoding/expected/sequence.out
new file mode 100644
index 00000000000..98a8f25e5ad
--- /dev/null
+++ b/contrib/test_decoding/expected/sequence.out
@@ -0,0 +1,325 @@
+-- predictability
+SET synchronous_commit = on;
+SELECT 'init' FROM pg_create_logical_replication_slot('regression_slot', 'test_decoding');
+ ?column?
+----------
+ init
+(1 row)
+
+CREATE SEQUENCE test_sequence;
+-- test the sequence changes by several nextval() calls
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 1
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 2
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 4
+(1 row)
+
+-- test the sequence changes by several ALTER commands
+ALTER SEQUENCE test_sequence INCREMENT BY 10;
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 14
+(1 row)
+
+ALTER SEQUENCE test_sequence START WITH 3000;
+ALTER SEQUENCE test_sequence MAXVALUE 10000;
+ALTER SEQUENCE test_sequence RESTART WITH 4000;
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 4000
+(1 row)
+
+-- test the sequence changes by several setval() calls
+SELECT setval('test_sequence', 3500);
+ setval
+--------
+ 3500
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3510
+(1 row)
+
+SELECT setval('test_sequence', 3500, true);
+ setval
+--------
+ 3500
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3510
+(1 row)
+
+SELECT setval('test_sequence', 3500, false);
+ setval
+--------
+ 3500
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3500
+(1 row)
+
+-- show results and drop sequence
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '1');
+ data
+-------------------------------------------------------------
+ BEGIN
+ sequence public.test_sequence: transactional: 1 value: 1
+ COMMIT
+ sequence public.test_sequence: transactional: 0 value: 33
+ BEGIN
+ sequence public.test_sequence: transactional: 1 value: 4
+ COMMIT
+ sequence public.test_sequence: transactional: 0 value: 334
+ BEGIN
+ sequence public.test_sequence: transactional: 1 value: 46
+ COMMIT
+ BEGIN
+ sequence public.test_sequence: transactional: 1 value: 14
+ COMMIT
+ BEGIN
+ sequence public.test_sequence: transactional: 1 value: 4000
+ COMMIT
+ sequence public.test_sequence: transactional: 0 value: 4320
+ sequence public.test_sequence: transactional: 0 value: 3500
+ sequence public.test_sequence: transactional: 0 value: 3830
+ sequence public.test_sequence: transactional: 0 value: 3500
+ sequence public.test_sequence: transactional: 0 value: 3830
+ sequence public.test_sequence: transactional: 0 value: 3500
+ sequence public.test_sequence: transactional: 0 value: 3820
+(24 rows)
+
+DROP SEQUENCE test_sequence;
+-- rollback on sequence creation and update
+BEGIN;
+CREATE SEQUENCE test_sequence;
+CREATE TABLE test_table (a INT);
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 1
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 2
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3
+(1 row)
+
+SELECT setval('test_sequence', 3000);
+ setval
+--------
+ 3000
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3001
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3002
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3003
+(1 row)
+
+ALTER SEQUENCE test_sequence RESTART WITH 6000;
+INSERT INTO test_table VALUES( (SELECT nextval('test_sequence')) );
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 6001
+(1 row)
+
+ROLLBACK;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '1');
+ data
+------
+(0 rows)
+
+-- rollback on table creation with serial column
+BEGIN;
+CREATE TABLE test_table (a SERIAL, b INT);
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+INSERT INTO test_table (b) VALUES (300);
+SELECT setval('test_table_a_seq', 3000);
+ setval
+--------
+ 3000
+(1 row)
+
+INSERT INTO test_table (b) VALUES (400);
+INSERT INTO test_table (b) VALUES (500);
+INSERT INTO test_table (b) VALUES (600);
+ALTER SEQUENCE test_table_a_seq RESTART WITH 6000;
+INSERT INTO test_table (b) VALUES (700);
+INSERT INTO test_table (b) VALUES (800);
+INSERT INTO test_table (b) VALUES (900);
+ROLLBACK;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '1');
+ data
+------
+(0 rows)
+
+-- rollback on table with serial column
+CREATE TABLE test_table (a SERIAL, b INT);
+BEGIN;
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+INSERT INTO test_table (b) VALUES (300);
+SELECT setval('test_table_a_seq', 3000);
+ setval
+--------
+ 3000
+(1 row)
+
+INSERT INTO test_table (b) VALUES (400);
+INSERT INTO test_table (b) VALUES (500);
+INSERT INTO test_table (b) VALUES (600);
+ALTER SEQUENCE test_table_a_seq RESTART WITH 6000;
+INSERT INTO test_table (b) VALUES (700);
+INSERT INTO test_table (b) VALUES (800);
+INSERT INTO test_table (b) VALUES (900);
+ROLLBACK;
+-- check table and sequence values after rollback
+SELECT * from test_table_a_seq;
+ last_value | log_cnt | is_called
+------------+---------+-----------
+ 3003 | 30 | t
+(1 row)
+
+SELECT nextval('test_table_a_seq');
+ nextval
+---------
+ 3004
+(1 row)
+
+DROP TABLE test_table;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '1');
+ data
+----------------------------------------------------------------
+ BEGIN
+ sequence public.test_table_a_seq: transactional: 1 value: 1
+ COMMIT
+ sequence public.test_table_a_seq: transactional: 0 value: 33
+ sequence public.test_table_a_seq: transactional: 0 value: 3000
+ sequence public.test_table_a_seq: transactional: 0 value: 3033
+(6 rows)
+
+-- savepoint test on table with serial column
+BEGIN;
+CREATE TABLE test_table (a SERIAL, b INT);
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+SAVEPOINT a;
+INSERT INTO test_table (b) VALUES (300);
+ROLLBACK TO SAVEPOINT a;
+DROP TABLE test_table;
+COMMIT;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '1');
+ data
+--------------------------------------------------------------
+ BEGIN
+ sequence public.test_table_a_seq: transactional: 1 value: 1
+ sequence public.test_table_a_seq: transactional: 1 value: 33
+ table public.test_table: INSERT: a[integer]:1 b[integer]:100
+ table public.test_table: INSERT: a[integer]:2 b[integer]:200
+ COMMIT
+(6 rows)
+
+-- savepoint test on table with serial column
+BEGIN;
+CREATE SEQUENCE test_sequence;
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 1
+(1 row)
+
+SELECT setval('test_sequence', 3000);
+ setval
+--------
+ 3000
+(1 row)
+
+SELECT nextval('test_sequence');
+ nextval
+---------
+ 3001
+(1 row)
+
+SAVEPOINT a;
+ALTER SEQUENCE test_sequence START WITH 7000;
+SELECT setval('test_sequence', 5000);
+ setval
+--------
+ 5000
+(1 row)
+
+ROLLBACK TO SAVEPOINT a;
+SELECT * FROM test_sequence;
+ last_value | log_cnt | is_called
+------------+---------+-----------
+ 3001 | 32 | t
+(1 row)
+
+DROP SEQUENCE test_sequence;
+COMMIT;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '1');
+ data
+-------------------------------------------------------------
+ BEGIN
+ sequence public.test_sequence: transactional: 1 value: 1
+ sequence public.test_sequence: transactional: 1 value: 33
+ sequence public.test_sequence: transactional: 1 value: 3000
+ sequence public.test_sequence: transactional: 1 value: 3033
+ COMMIT
+(6 rows)
+
+SELECT pg_drop_replication_slot('regression_slot');
+ pg_drop_replication_slot
+--------------------------
+
+(1 row)
+
diff --git a/contrib/test_decoding/sql/sequence.sql b/contrib/test_decoding/sql/sequence.sql
new file mode 100644
index 00000000000..c19a5d50747
--- /dev/null
+++ b/contrib/test_decoding/sql/sequence.sql
@@ -0,0 +1,119 @@
+-- predictability
+SET synchronous_commit = on;
+SELECT 'init' FROM pg_create_logical_replication_slot('regression_slot', 'test_decoding');
+
+CREATE SEQUENCE test_sequence;
+
+-- test the sequence changes by several nextval() calls
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+
+-- test the sequence changes by several ALTER commands
+ALTER SEQUENCE test_sequence INCREMENT BY 10;
+SELECT nextval('test_sequence');
+
+ALTER SEQUENCE test_sequence START WITH 3000;
+ALTER SEQUENCE test_sequence MAXVALUE 10000;
+ALTER SEQUENCE test_sequence RESTART WITH 4000;
+SELECT nextval('test_sequence');
+
+-- test the sequence changes by several setval() calls
+SELECT setval('test_sequence', 3500);
+SELECT nextval('test_sequence');
+SELECT setval('test_sequence', 3500, true);
+SELECT nextval('test_sequence');
+SELECT setval('test_sequence', 3500, false);
+SELECT nextval('test_sequence');
+
+-- show results and drop sequence
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '1');
+DROP SEQUENCE test_sequence;
+
+-- rollback on sequence creation and update
+BEGIN;
+CREATE SEQUENCE test_sequence;
+CREATE TABLE test_table (a INT);
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT setval('test_sequence', 3000);
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+SELECT nextval('test_sequence');
+ALTER SEQUENCE test_sequence RESTART WITH 6000;
+INSERT INTO test_table VALUES( (SELECT nextval('test_sequence')) );
+SELECT nextval('test_sequence');
+ROLLBACK;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '1');
+
+-- rollback on table creation with serial column
+BEGIN;
+CREATE TABLE test_table (a SERIAL, b INT);
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+INSERT INTO test_table (b) VALUES (300);
+SELECT setval('test_table_a_seq', 3000);
+INSERT INTO test_table (b) VALUES (400);
+INSERT INTO test_table (b) VALUES (500);
+INSERT INTO test_table (b) VALUES (600);
+ALTER SEQUENCE test_table_a_seq RESTART WITH 6000;
+INSERT INTO test_table (b) VALUES (700);
+INSERT INTO test_table (b) VALUES (800);
+INSERT INTO test_table (b) VALUES (900);
+ROLLBACK;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '1');
+
+-- rollback on table with serial column
+CREATE TABLE test_table (a SERIAL, b INT);
+
+BEGIN;
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+INSERT INTO test_table (b) VALUES (300);
+SELECT setval('test_table_a_seq', 3000);
+INSERT INTO test_table (b) VALUES (400);
+INSERT INTO test_table (b) VALUES (500);
+INSERT INTO test_table (b) VALUES (600);
+ALTER SEQUENCE test_table_a_seq RESTART WITH 6000;
+INSERT INTO test_table (b) VALUES (700);
+INSERT INTO test_table (b) VALUES (800);
+INSERT INTO test_table (b) VALUES (900);
+ROLLBACK;
+
+-- check table and sequence values after rollback
+SELECT * from test_table_a_seq;
+SELECT nextval('test_table_a_seq');
+
+DROP TABLE test_table;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '1');
+
+-- savepoint test on table with serial column
+BEGIN;
+CREATE TABLE test_table (a SERIAL, b INT);
+INSERT INTO test_table (b) VALUES (100);
+INSERT INTO test_table (b) VALUES (200);
+SAVEPOINT a;
+INSERT INTO test_table (b) VALUES (300);
+ROLLBACK TO SAVEPOINT a;
+DROP TABLE test_table;
+COMMIT;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '1');
+
+-- savepoint test on table with serial column
+BEGIN;
+CREATE SEQUENCE test_sequence;
+SELECT nextval('test_sequence');
+SELECT setval('test_sequence', 3000);
+SELECT nextval('test_sequence');
+SAVEPOINT a;
+ALTER SEQUENCE test_sequence START WITH 7000;
+SELECT setval('test_sequence', 5000);
+ROLLBACK TO SAVEPOINT a;
+SELECT * FROM test_sequence;
+DROP SEQUENCE test_sequence;
+COMMIT;
+SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1', 'include-sequences', '1');
+
+SELECT pg_drop_replication_slot('regression_slot');
diff --git a/contrib/test_decoding/test_decoding.c b/contrib/test_decoding/test_decoding.c
index 4f4f51a89c1..200caad6863 100644
--- a/contrib/test_decoding/test_decoding.c
+++ b/contrib/test_decoding/test_decoding.c
@@ -29,6 +29,7 @@ typedef struct
MemoryContext context;
bool include_xids;
bool include_timestamp;
+ bool include_sequences;
bool skip_empty_xacts;
bool only_local;
} TestDecodingData;
@@ -72,6 +73,10 @@ static void pg_decode_message(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn, XLogRecPtr lsn,
bool transactional, const char *prefix,
Size sz, const char *message);
+static void pg_decode_sequence(LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn, XLogRecPtr sequence_lsn,
+ Relation rel, bool transactional,
+ int64 value);
static bool pg_decode_filter_prepare(LogicalDecodingContext *ctx,
TransactionId xid,
const char *gid);
@@ -112,6 +117,10 @@ static void pg_decode_stream_message(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn, XLogRecPtr lsn,
bool transactional, const char *prefix,
Size sz, const char *message);
+static void pg_decode_stream_sequence(LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn, XLogRecPtr sequence_lsn,
+ Relation rel, bool transactional,
+ int64 value);
static void pg_decode_stream_truncate(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn,
int nrelations, Relation relations[],
@@ -135,6 +144,7 @@ _PG_output_plugin_init(OutputPluginCallbacks *cb)
cb->filter_by_origin_cb = pg_decode_filter;
cb->shutdown_cb = pg_decode_shutdown;
cb->message_cb = pg_decode_message;
+ cb->sequence_cb = pg_decode_sequence;
cb->filter_prepare_cb = pg_decode_filter_prepare;
cb->begin_prepare_cb = pg_decode_begin_prepare_txn;
cb->prepare_cb = pg_decode_prepare_txn;
@@ -147,6 +157,7 @@ _PG_output_plugin_init(OutputPluginCallbacks *cb)
cb->stream_commit_cb = pg_decode_stream_commit;
cb->stream_change_cb = pg_decode_stream_change;
cb->stream_message_cb = pg_decode_stream_message;
+ cb->stream_sequence_cb = pg_decode_stream_sequence;
cb->stream_truncate_cb = pg_decode_stream_truncate;
}
@@ -168,6 +179,7 @@ pg_decode_startup(LogicalDecodingContext *ctx, OutputPluginOptions *opt,
data->include_timestamp = false;
data->skip_empty_xacts = false;
data->only_local = false;
+ data->include_sequences = false;
ctx->output_plugin_private = data;
@@ -259,6 +271,17 @@ pg_decode_startup(LogicalDecodingContext *ctx, OutputPluginOptions *opt,
errmsg("could not parse value \"%s\" for parameter \"%s\"",
strVal(elem->arg), elem->defname)));
}
+ else if (strcmp(elem->defname, "include-sequences") == 0)
+ {
+
+ if (elem->arg == NULL)
+ data->include_sequences = true;
+ else if (!parse_bool(strVal(elem->arg), &data->include_sequences))
+ ereport(ERROR,
+ (errcode(ERRCODE_INVALID_PARAMETER_VALUE),
+ errmsg("could not parse value \"%s\" for parameter \"%s\"",
+ strVal(elem->arg), elem->defname)));
+ }
else
{
ereport(ERROR,
@@ -762,6 +785,38 @@ pg_decode_message(LogicalDecodingContext *ctx,
OutputPluginWrite(ctx, true);
}
+static void
+pg_decode_sequence(LogicalDecodingContext *ctx, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ bool transactional,
+ int64 value)
+{
+ TestDecodingData *data = ctx->output_plugin_private;
+ TestDecodingTxnData *txndata = txn->output_plugin_private;
+
+ if (!data->include_sequences)
+ return;
+
+ /* output BEGIN if we haven't yet, but only for the transactional case */
+ if (transactional)
+ {
+ if (data->skip_empty_xacts && !txndata->xact_wrote_changes)
+ {
+ pg_output_begin(ctx, data, txn, false);
+ }
+ txndata->xact_wrote_changes = true;
+ }
+
+ OutputPluginPrepareWrite(ctx, true);
+ appendStringInfoString(ctx->out, "sequence ");
+ appendStringInfoString(ctx->out,
+ quote_qualified_identifier(get_namespace_name(get_rel_namespace(RelationGetRelid(rel))),
+ RelationGetRelationName(rel)));
+ appendStringInfo(ctx->out, ": transactional: %d value: " INT64_FORMAT,
+ transactional, value);
+ OutputPluginWrite(ctx, true);
+}
+
static void
pg_decode_stream_start(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn)
@@ -974,6 +1029,38 @@ pg_decode_stream_message(LogicalDecodingContext *ctx,
OutputPluginWrite(ctx, true);
}
+static void
+pg_decode_stream_sequence(LogicalDecodingContext *ctx, ReorderBufferTXN *txn,
+ XLogRecPtr sequence_lsn, Relation rel,
+ bool transactional,
+ int64 value)
+{
+ TestDecodingData *data = ctx->output_plugin_private;
+ TestDecodingTxnData *txndata = txn->output_plugin_private;
+
+ if (!data->include_sequences)
+ return;
+
+ /* output BEGIN if we haven't yet, but only for the transactional case */
+ if (transactional)
+ {
+ if (data->skip_empty_xacts && !txndata->xact_wrote_changes)
+ {
+ pg_output_stream_start(ctx, data, txn, false);
+ }
+ txndata->xact_wrote_changes = true;
+ }
+
+ OutputPluginPrepareWrite(ctx, true);
+ appendStringInfoString(ctx->out, "streaming sequence ");
+ appendStringInfoString(ctx->out,
+ quote_qualified_identifier(get_namespace_name(get_rel_namespace(RelationGetRelid(rel))),
+ RelationGetRelationName(rel)));
+ appendStringInfo(ctx->out, ": transactional: %d value: " INT64_FORMAT,
+ transactional, value);
+ OutputPluginWrite(ctx, true);
+}
+
/*
* In streaming mode, we don't display the detailed information of Truncate.
* See pg_decode_stream_change.
--
2.43.0
[text/x-patch] v20240111-0003-Add-decoding-of-sequences-to-built-in-repl.patch (269.4K, ../../[email protected]/4-v20240111-0003-Add-decoding-of-sequences-to-built-in-repl.patch)
download | inline diff:
From 46c1e9e697245435a01fa1d86e375071d7f7f42f Mon Sep 17 00:00:00 2001
From: Tomas Vondra <[email protected]>
Date: Wed, 10 Jan 2024 16:08:48 +0100
Subject: [PATCH v20240111 3/7] Add decoding of sequences to built-in
replication
This commit adds support for decoding of sequences to the built-in
replication (the infrastructure was added by commit 0da92dc530).
The syntax and behavior mostly mimics handling of tables, i.e. a
publication may be defined as FOR ALL SEQUENCES (replicating all
sequences in a database), FOR ALL SEQUENCES IN SCHEMA (replicating
all sequences in a particular schema) or individual sequences.
To publish sequence modifications, the publication has to include
'sequence' action. The protocol is extended with a new message,
describing sequence increments.
A new system view pg_publication_sequences lists all the sequences
added to a publication, both directly and indirectly. Various psql
commands (\d and \dRp) are improved to also display publications
including a given sequence, or sequences included in a publication.
Author: Tomas Vondra, Cary Huang, Masahiko Sawada
Reviewed-by: Ashutosh Bapat, Amit Kapila, Peter Eisentraut, Amit Kapila,
Hannu Krosing, Andres Freund, Petr Jelinek, Zhijie Hou
Discussion: https://postgr.es/m/[email protected]
Discussion: https://postgr.es/m/[email protected]
Discussion: https://postgr.es/m/CAD21AoDtKpdJcHOLjfPQ7TmpFqNB5__%3DQ_g1e8OBRrwT5LP-%3Dg%40mail.gmail.com
---
doc/src/sgml/catalogs.sgml | 10 +
doc/src/sgml/protocol.sgml | 80 ++
doc/src/sgml/ref/alter_publication.sgml | 25 +-
doc/src/sgml/ref/alter_subscription.sgml | 8 +-
doc/src/sgml/ref/create_publication.sgml | 51 +-
doc/src/sgml/system-views.sgml | 71 ++
src/backend/catalog/objectaddress.c | 45 +-
src/backend/catalog/pg_publication.c | 336 +++++-
src/backend/catalog/system_views.sql | 10 +
src/backend/commands/publicationcmds.c | 422 +++++--
src/backend/commands/sequence.c | 248 +++-
src/backend/commands/subscriptioncmds.c | 101 +-
src/backend/executor/execReplication.c | 4 +-
src/backend/parser/gram.y | 73 +-
src/backend/replication/logical/proto.c | 55 +
src/backend/replication/logical/tablesync.c | 135 ++-
src/backend/replication/logical/worker.c | 97 ++
src/backend/replication/pgoutput/pgoutput.c | 108 +-
src/backend/utils/cache/relcache.c | 28 +-
src/backend/utils/cache/syscache.c | 5 +-
src/bin/pg_dump/pg_dump.c | 64 +-
src/bin/pg_dump/pg_dump.h | 3 +
src/bin/pg_dump/t/002_pg_dump.pl | 47 +-
src/bin/psql/describe.c | 297 +++--
src/bin/psql/tab-complete.c | 39 +-
src/include/catalog/pg_proc.dat | 13 +
src/include/catalog/pg_publication.h | 26 +-
.../catalog/pg_publication_namespace.h | 10 +-
src/include/commands/sequence.h | 1 +
src/include/nodes/parsenodes.h | 10 +-
src/include/replication/logicalproto.h | 19 +-
src/test/regress/expected/object_address.out | 20 +-
src/test/regress/expected/psql.out | 6 +-
src/test/regress/expected/publication.out | 1055 +++++++++++++----
src/test/regress/expected/rules.out | 8 +
src/test/regress/sql/object_address.sql | 5 +-
src/test/regress/sql/publication.sql | 231 +++-
src/test/subscription/t/034_sequences.pl | 344 ++++++
src/tools/pgindent/typedefs.list | 1 +
39 files changed, 3589 insertions(+), 522 deletions(-)
create mode 100644 src/test/subscription/t/034_sequences.pl
diff --git a/doc/src/sgml/catalogs.sgml b/doc/src/sgml/catalogs.sgml
index 3ec7391ec5e..d61fa014316 100644
--- a/doc/src/sgml/catalogs.sgml
+++ b/doc/src/sgml/catalogs.sgml
@@ -6425,6 +6425,16 @@ SCRAM-SHA-256$<replaceable><iteration count></replaceable>:<replaceable>&l
Reference to schema
</para></entry>
</row>
+
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ <structfield>pntype</structfield> <type>char</type>
+ </para>
+ <para>
+ Determines which object type is included from this schema.
+ <literal>t</literal> for tables, <literal>s</literal> for sequences
+ </para></entry>
+ </row>
</tbody>
</tgroup>
</table>
diff --git a/doc/src/sgml/protocol.sgml b/doc/src/sgml/protocol.sgml
index 6c3e8a631d7..e2e0decf2ed 100644
--- a/doc/src/sgml/protocol.sgml
+++ b/doc/src/sgml/protocol.sgml
@@ -6477,6 +6477,86 @@ psql "dbname=postgres replication=database" -c "IDENTIFY_SYSTEM;"
</listitem>
</varlistentry>
+ <varlistentry id="protocol-logicalrep-message-formats-Sequence">
+ <term>Sequence</term>
+ <listitem>
+ <variablelist>
+ <varlistentry>
+ <term>Byte1('X')</term>
+ <listitem>
+ <para>
+ Identifies the message as a sequence message.
+ </para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term>Int32 (TransactionId)</term>
+ <listitem>
+ <para>
+ Xid of the transaction (only present for streamed transactions).
+ This field is available since protocol version 2.
+ </para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term>Int8 (0)</term>
+ <listitem>
+ <para>
+ Flags; currently unused.
+ </para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term>Int64 (XLogRecPtr)</term>
+ <listitem>
+ <para>
+ The LSN of the sequence increment.
+ </para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term>String</term>
+ <listitem>
+ <para>
+ Namespace (empty string for <literal>pg_catalog</literal>).
+ </para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term>String</term>
+ <listitem>
+ <para>
+ Sequence name.
+ </para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term>Int8</term>
+ <listitem>
+ <para>
+ 1 if the sequence update is transactional, 0 otherwise.
+ </para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term>Int64</term>
+ <listitem>
+ <para>
+ <structfield>value</structfield> value of the sequence.
+ </para>
+ </listitem>
+ </varlistentry>
+ </variablelist>
+ </listitem>
+ </varlistentry>
+
<varlistentry id="protocol-logicalrep-message-formats-Type">
<term>Type</term>
<listitem>
diff --git a/doc/src/sgml/ref/alter_publication.sgml b/doc/src/sgml/ref/alter_publication.sgml
index 44ae7e0e871..8caf2135fc7 100644
--- a/doc/src/sgml/ref/alter_publication.sgml
+++ b/doc/src/sgml/ref/alter_publication.sgml
@@ -31,7 +31,9 @@ ALTER PUBLICATION <replaceable class="parameter">name</replaceable> RENAME TO <r
<phrase>where <replaceable class="parameter">publication_object</replaceable> is one of:</phrase>
TABLE [ ONLY ] <replaceable class="parameter">table_name</replaceable> [ * ] [ ( <replaceable class="parameter">column_name</replaceable> [, ... ] ) ] [ WHERE ( <replaceable class="parameter">expression</replaceable> ) ] [, ... ]
+ SEQUENCE <replaceable class="parameter">sequence_name</replaceable> [, ... ]
TABLES IN SCHEMA { <replaceable class="parameter">schema_name</replaceable> | CURRENT_SCHEMA } [, ... ]
+ SEQUENCES IN SCHEMA { <replaceable class="parameter">schema_name</replaceable> | CURRENT_SCHEMA } [, ... ]
</synopsis>
</refsynopsisdiv>
@@ -44,13 +46,13 @@ ALTER PUBLICATION <replaceable class="parameter">name</replaceable> RENAME TO <r
</para>
<para>
- The first three variants change which tables/schemas are part of the
- publication. The <literal>SET</literal> clause will replace the list of
- tables/schemas in the publication with the specified list; the existing
- tables/schemas that were present in the publication will be removed. The
- <literal>ADD</literal> and <literal>DROP</literal> clauses will add and
- remove one or more tables/schemas from the publication. Note that adding
- tables/schemas to a publication that is already subscribed to will require an
+ The first three variants change which objects (tables, sequences or schemas)
+ are part of the publication. The <literal>SET</literal> clause will replace
+ the list of objects in the publication with the specified list; the existing
+ objects that were present in the publication will be removed.
+ The <literal>ADD</literal> and <literal>DROP</literal> clauses will add and
+ remove one or more objects from the publication. Note that adding objects
+ to a publication that is already subscribed to will require an
<link linkend="sql-altersubscription-params-refresh-publication">
<literal>ALTER SUBSCRIPTION ... REFRESH PUBLICATION</literal></link> action on the
subscribing side in order to become effective. Note also that
@@ -138,6 +140,15 @@ ALTER PUBLICATION <replaceable class="parameter">name</replaceable> RENAME TO <r
</listitem>
</varlistentry>
+ <varlistentry>
+ <term><replaceable class="parameter">sequence_name</replaceable></term>
+ <listitem>
+ <para>
+ Name of an existing sequence.
+ </para>
+ </listitem>
+ </varlistentry>
+
<varlistentry>
<term><replaceable class="parameter">schema_name</replaceable></term>
<listitem>
diff --git a/doc/src/sgml/ref/alter_subscription.sgml b/doc/src/sgml/ref/alter_subscription.sgml
index 6d36ff0dc90..479ec495896 100644
--- a/doc/src/sgml/ref/alter_subscription.sgml
+++ b/doc/src/sgml/ref/alter_subscription.sgml
@@ -152,8 +152,8 @@ ALTER SUBSCRIPTION <replaceable class="parameter">name</replaceable> RENAME TO <
<listitem>
<para>
Fetch missing table information from publisher. This will start
- replication of tables that were added to the subscribed-to publications
- since <link linkend="sql-createsubscription">
+ replication of tables and sequences that were added to the subscribed-to
+ publications since <link linkend="sql-createsubscription">
<command>CREATE SUBSCRIPTION</command></link> or
the last invocation of <command>REFRESH PUBLICATION</command>.
</para>
@@ -172,8 +172,8 @@ ALTER SUBSCRIPTION <replaceable class="parameter">name</replaceable> RENAME TO <
The default is <literal>true</literal>.
</para>
<para>
- Previously subscribed tables are not copied, even if a table's row
- filter <literal>WHERE</literal> clause has since been modified.
+ Previously subscribed tables and sequences are not copied, even if a
+ table's row filter <literal>WHERE</literal> clause has since been modified.
</para>
<para>
See <xref linkend="sql-createsubscription-notes"/> for details of
diff --git a/doc/src/sgml/ref/create_publication.sgml b/doc/src/sgml/ref/create_publication.sgml
index fd9c5deac95..ba578eb9aa4 100644
--- a/doc/src/sgml/ref/create_publication.sgml
+++ b/doc/src/sgml/ref/create_publication.sgml
@@ -22,14 +22,21 @@ PostgreSQL documentation
<refsynopsisdiv>
<synopsis>
CREATE PUBLICATION <replaceable class="parameter">name</replaceable>
- [ FOR ALL TABLES
+ [ FOR ALL <replaceable class="parameter">object_type</replaceable> [, ...]
| FOR <replaceable class="parameter">publication_object</replaceable> [, ... ] ]
[ WITH ( <replaceable class="parameter">publication_parameter</replaceable> [= <replaceable class="parameter">value</replaceable>] [, ... ] ) ]
+<phrase>where <replaceable class="parameter">object type</replaceable> is one of:</phrase>
+
+ TABLES
+ SEQUENCES
+
<phrase>where <replaceable class="parameter">publication_object</replaceable> is one of:</phrase>
TABLE [ ONLY ] <replaceable class="parameter">table_name</replaceable> [ * ] [ ( <replaceable class="parameter">column_name</replaceable> [, ... ] ) ] [ WHERE ( <replaceable class="parameter">expression</replaceable> ) ] [, ... ]
+ SEQUENCE <replaceable class="parameter">sequence_name</replaceable> [ * ] [, ... ]
TABLES IN SCHEMA { <replaceable class="parameter">schema_name</replaceable> | CURRENT_SCHEMA } [, ... ]
+ SEQUENCES IN SCHEMA { <replaceable class="parameter">schema_name</replaceable> | CURRENT_SCHEMA } [, ... ]
</synopsis>
</refsynopsisdiv>
@@ -117,22 +124,39 @@ CREATE PUBLICATION <replaceable class="parameter">name</replaceable>
</listitem>
</varlistentry>
+ <varlistentry id="sql-createpublication-for-sequence">
+ <term><literal>FOR SEQUENCE</literal></term>
+ <listitem>
+ <para>
+ Specifies a list of sequences to add to the publication.
+ </para>
+
+ <para>
+ Specifying a sequence that is part of a schema specified by <literal>FOR
+ ALL SEQUENCES IN SCHEMA</literal> is not supported.
+ </para>
+ </listitem>
+ </varlistentry>
+
<varlistentry id="sql-createpublication-params-for-all-tables">
<term><literal>FOR ALL TABLES</literal></term>
+ <term><literal>FOR ALL SEQUENCES</literal></term>
<listitem>
<para>
- Marks the publication as one that replicates changes for all tables in
- the database, including tables created in the future.
+ Marks the publication as one that replicates changes for all tables or
+ sequences in the database, including tables created in the future.
</para>
</listitem>
</varlistentry>
<varlistentry id="sql-createpublication-params-for-tables-in-schema">
<term><literal>FOR TABLES IN SCHEMA</literal></term>
+ <term><literal>FOR SEQUENCES IN SCHEMA</literal></term>
<listitem>
<para>
- Marks the publication as one that replicates changes for all tables in
- the specified list of schemas, including tables created in the future.
+ Marks the publication as one that replicates changes for all tables or
+ sequences in the specified list of schemas, including tables or sequences
+ created in the future.
</para>
<para>
@@ -240,10 +264,12 @@ CREATE PUBLICATION <replaceable class="parameter">name</replaceable>
<title>Notes</title>
<para>
- If <literal>FOR TABLE</literal>, <literal>FOR ALL TABLES</literal> or
- <literal>FOR TABLES IN SCHEMA</literal> are not specified, then the
- publication starts out with an empty set of tables. That is useful if
- tables or schemas are to be added later.
+ If <literal>FOR TABLE</literal>, <literal>FOR SEQUENCE</literal>,
+ <literal>FOR ALL TABLES</literal>, <literal>FOR ALL SEQUENCES</literal>,
+ <literal>FOR TABLES IN SCHEMA</literal> or <literal>FOR SEQUENCES IN SCHEMA</literal>
+ are not specified, then the publication starts out with an empty set
+ of tables/sequences. That is useful if tables, sequences or schemas
+ are to be added later.
</para>
<para>
@@ -258,9 +284,10 @@ CREATE PUBLICATION <replaceable class="parameter">name</replaceable>
</para>
<para>
- To add a table to a publication, the invoking user must have ownership
- rights on the table. The <command>FOR ALL TABLES</command> and
- <command>FOR TABLES IN SCHEMA</command> clauses require the invoking
+ To add a table or a sequence to a publication, the invoking user must
+ have ownership rights on the object. The <command>FOR ALL TABLES</command>,
+ <command>FOR ALL SEQUENCES</command>, <command>FOR TABLES IN SCHEMA</command>
+ and <command>FOR SEQUENCES IN SCHEMA</command> clauses require the invoking
user to be a superuser.
</para>
diff --git a/doc/src/sgml/system-views.sgml b/doc/src/sgml/system-views.sgml
index 72d01fc624c..d2f91173917 100644
--- a/doc/src/sgml/system-views.sgml
+++ b/doc/src/sgml/system-views.sgml
@@ -126,6 +126,11 @@
<entry>prepared transactions</entry>
</row>
+ <row>
+ <entry><link linkend="view-pg-publication-sequences"><structname>pg_publication_sequences</structname></link></entry>
+ <entry>publications and information of their associated sequences</entry>
+ </row>
+
<row>
<entry><link linkend="view-pg-publication-tables"><structname>pg_publication_tables</structname></link></entry>
<entry>publications and information of their associated tables</entry>
@@ -2138,6 +2143,72 @@ SELECT * FROM pg_locks pl LEFT JOIN pg_prepared_xacts ppx
</sect1>
+ <sect1 id="view-pg-publication-sequences">
+ <title><structname>pg_publication_sequences</structname></title>
+
+ <indexterm zone="view-pg-publication-sequences">
+ <primary>pg_publication_sequences</primary>
+ </indexterm>
+
+ <para>
+ The view <structname>pg_publication_sequences</structname> provides
+ information about the mapping between publications and information of
+ sequences they contain. Unlike the underlying catalog
+ <link linkend="catalog-pg-publication-rel"><structname>pg_publication_rel</structname></link>,
+ this view expands publications defined as <literal>FOR ALL SEQUENCES</literal>
+ and <literal>FOR SEQUENCES IN SCHEMA</literal>, so for such publications
+ there will be a row for each eligible sequence.
+ </para>
+
+ <table>
+ <title><structname>pg_publication_sequences</structname> Columns</title>
+ <tgroup cols="1">
+ <thead>
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ Column Type
+ </para>
+ <para>
+ Description
+ </para></entry>
+ </row>
+ </thead>
+
+ <tbody>
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ <structfield>pubname</structfield> <type>name</type>
+ (references <link linkend="catalog-pg-publication"><structname>pg_publication</structname></link>.<structfield>pubname</structfield>)
+ </para>
+ <para>
+ Name of publication
+ </para></entry>
+ </row>
+
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ <structfield>schemaname</structfield> <type>name</type>
+ (references <link linkend="catalog-pg-namespace"><structname>pg_namespace</structname></link>.<structfield>nspname</structfield>)
+ </para>
+ <para>
+ Name of schema containing sequence
+ </para></entry>
+ </row>
+
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ <structfield>sequencename</structfield> <type>name</type>
+ (references <link linkend="catalog-pg-class"><structname>pg_class</structname></link>.<structfield>relname</structfield>)
+ </para>
+ <para>
+ Name of sequence
+ </para></entry>
+ </row>
+ </tbody>
+ </tgroup>
+ </table>
+ </sect1>
+
<sect1 id="view-pg-publication-tables">
<title><structname>pg_publication_tables</structname></title>
diff --git a/src/backend/catalog/objectaddress.c b/src/backend/catalog/objectaddress.c
index 8bb2924b9cd..66796bf10eb 100644
--- a/src/backend/catalog/objectaddress.c
+++ b/src/backend/catalog/objectaddress.c
@@ -1964,12 +1964,14 @@ get_object_address_publication_schema(List *object, bool missing_ok)
char *pubname;
char *schemaname;
Oid schemaid;
+ char *objtype;
ObjectAddressSet(address, PublicationNamespaceRelationId, InvalidOid);
/* Fetch schema name and publication name from input list */
schemaname = strVal(linitial(object));
pubname = strVal(lsecond(object));
+ objtype = strVal(lthird(object));
schemaid = get_namespace_oid(schemaname, missing_ok);
if (!OidIsValid(schemaid))
@@ -1982,10 +1984,12 @@ get_object_address_publication_schema(List *object, bool missing_ok)
/* Find the publication schema mapping in syscache */
address.objectId =
- GetSysCacheOid2(PUBLICATIONNAMESPACEMAP,
+ GetSysCacheOid3(PUBLICATIONNAMESPACEMAP,
Anum_pg_publication_namespace_oid,
ObjectIdGetDatum(schemaid),
- ObjectIdGetDatum(pub->oid));
+ ObjectIdGetDatum(pub->oid),
+ CharGetDatum(objtype[0]));
+
if (!OidIsValid(address.objectId) && !missing_ok)
ereport(ERROR,
(errcode(ERRCODE_UNDEFINED_OBJECT),
@@ -2260,7 +2264,6 @@ pg_get_object_address(PG_FUNCTION_ARGS)
*/
switch (type)
{
- case OBJECT_PUBLICATION_NAMESPACE:
case OBJECT_USER_MAPPING:
if (list_length(name) != 1)
ereport(ERROR,
@@ -2293,6 +2296,7 @@ pg_get_object_address(PG_FUNCTION_ARGS)
errmsg("name list length must be at least %d", 3)));
/* fall through to check args length */
/* FALLTHROUGH */
+ case OBJECT_PUBLICATION_NAMESPACE:
case OBJECT_OPERATOR:
if (list_length(args) != 2)
ereport(ERROR,
@@ -2365,6 +2369,8 @@ pg_get_object_address(PG_FUNCTION_ARGS)
objnode = (Node *) list_make2(name, linitial(args));
break;
case OBJECT_PUBLICATION_NAMESPACE:
+ objnode = (Node *) list_make3(linitial(name), linitial(args), lsecond(args));
+ break;
case OBJECT_USER_MAPPING:
objnode = (Node *) list_make2(linitial(name), linitial(args));
break;
@@ -2868,11 +2874,12 @@ get_catalog_object_by_oid(Relation catalog, AttrNumber oidcol, Oid objectId)
*
* Get publication name and schema name from the object address into pubname and
* nspname. Both pubname and nspname are palloc'd strings which will be freed by
- * the caller.
+ * the caller. The last parameter specifies which object type is included from
+ * the schema.
*/
static bool
getPublicationSchemaInfo(const ObjectAddress *object, bool missing_ok,
- char **pubname, char **nspname)
+ char **pubname, char **nspname, char **objtype)
{
HeapTuple tup;
Form_pg_publication_namespace pnform;
@@ -2908,6 +2915,14 @@ getPublicationSchemaInfo(const ObjectAddress *object, bool missing_ok,
return false;
}
+ /*
+ * The type is always a single character, but we need to pass it as a
+ * string, so allocate two charaters and set the first one. The second one
+ * is \0.
+ */
+ *objtype = palloc0(2);
+ *objtype[0] = pnform->pntype;
+
ReleaseSysCache(tup);
return true;
}
@@ -3978,15 +3993,17 @@ getObjectDescription(const ObjectAddress *object, bool missing_ok)
{
char *pubname;
char *nspname;
+ char *objtype;
if (!getPublicationSchemaInfo(object, missing_ok,
- &pubname, &nspname))
+ &pubname, &nspname, &objtype))
break;
- appendStringInfo(&buffer, _("publication of schema %s in publication %s"),
- nspname, pubname);
+ appendStringInfo(&buffer, _("publication of schema %s in publication %s type %s"),
+ nspname, pubname, objtype);
pfree(pubname);
pfree(nspname);
+ pfree(objtype);
break;
}
@@ -5832,18 +5849,24 @@ getObjectIdentityParts(const ObjectAddress *object,
{
char *pubname;
char *nspname;
+ char *objtype;
if (!getPublicationSchemaInfo(object, missing_ok, &pubname,
- &nspname))
+ &nspname, &objtype))
break;
- appendStringInfo(&buffer, "%s in publication %s",
- nspname, pubname);
+ appendStringInfo(&buffer, "%s in publication %s type %s",
+ nspname, pubname, objtype);
if (objargs)
*objargs = list_make1(pubname);
else
pfree(pubname);
+ if (objargs)
+ *objargs = lappend(*objargs, objtype);
+ else
+ pfree(objtype);
+
if (objname)
*objname = list_make1(nspname);
else
diff --git a/src/backend/catalog/pg_publication.c b/src/backend/catalog/pg_publication.c
index b98b0ce0aef..ac5684f3f09 100644
--- a/src/backend/catalog/pg_publication.c
+++ b/src/backend/catalog/pg_publication.c
@@ -63,9 +63,10 @@ static void publication_translate_columns(Relation targetrel, List *columns,
static void
check_publication_add_relation(Relation targetrel)
{
- /* Must be a regular or partitioned table */
+ /* Must be a regular or partitioned table, or a sequence */
if (RelationGetForm(targetrel)->relkind != RELKIND_RELATION &&
- RelationGetForm(targetrel)->relkind != RELKIND_PARTITIONED_TABLE)
+ RelationGetForm(targetrel)->relkind != RELKIND_PARTITIONED_TABLE &&
+ RelationGetForm(targetrel)->relkind != RELKIND_SEQUENCE)
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("cannot add relation \"%s\" to publication",
@@ -142,7 +143,8 @@ static bool
is_publishable_class(Oid relid, Form_pg_class reltuple)
{
return (reltuple->relkind == RELKIND_RELATION ||
- reltuple->relkind == RELKIND_PARTITIONED_TABLE) &&
+ reltuple->relkind == RELKIND_PARTITIONED_TABLE ||
+ reltuple->relkind == RELKIND_SEQUENCE) &&
!IsCatalogRelationOid(relid) &&
reltuple->relpersistence == RELPERSISTENCE_PERMANENT &&
relid >= FirstNormalObjectId;
@@ -233,6 +235,52 @@ filter_partitions(List *table_infos)
}
}
+/*
+ * Check the character is a valid object type for schema publication.
+ *
+ * This recognizes either 't' for tables or 's' for sequences. Places that
+ * need to handle 'u' for unsupported relkinds need to do that explicitlyl
+ */
+static void
+AssertObjectTypeValid(char objectType)
+{
+#ifdef USE_ASSERT_CHECKING
+ Assert(objectType == PUB_OBJTYPE_SEQUENCE || objectType == PUB_OBJTYPE_TABLE);
+#endif
+}
+
+/*
+ * Determine object type matching a given a relkind value.
+ */
+char
+pub_get_object_type_for_relkind(char relkind)
+{
+ /* sequence maps directly to sequence relkind */
+ if (relkind == RELKIND_SEQUENCE)
+ return PUB_OBJTYPE_SEQUENCE;
+
+ /* for table, we match either regular or partitioned table */
+ if (relkind == RELKIND_RELATION ||
+ relkind == RELKIND_PARTITIONED_TABLE)
+ return PUB_OBJTYPE_TABLE;
+
+ return PUB_OBJTYPE_UNSUPPORTED;
+}
+
+/*
+ * Determine if publication object type matches the relkind.
+ *
+ * Returns true if the relation matches object type replicated by this schema,
+ * false otherwise.
+ */
+static bool
+pub_object_type_matches_relkind(char objectType, char relkind)
+{
+ AssertObjectTypeValid(objectType);
+
+ return (pub_get_object_type_for_relkind(relkind) == objectType);
+}
+
/*
* Returns true if any schema is associated with the publication, false if no
* schema is associated with the publication.
@@ -253,7 +301,7 @@ is_schema_publication(Oid pubid)
ObjectIdGetDatum(pubid));
scan = systable_beginscan(pubschsrel,
- PublicationNamespacePnnspidPnpubidIndexId,
+ PublicationNamespacePnnspidPnpubidPntypeIndexId,
true, NULL, 1, &scankey);
tup = systable_getnext(scan);
result = HeapTupleIsValid(tup);
@@ -339,7 +387,9 @@ GetTopMostAncestorInPublication(Oid puboid, List *ancestors, int *ancestor_level
}
else
{
- aschemaPubids = GetSchemaPublications(get_rel_namespace(ancestor));
+ /* we only search for ancestors of tables, so PUB_OBJTYPE_TABLE */
+ aschemaPubids = GetSchemaPublications(get_rel_namespace(ancestor),
+ PUB_OBJTYPE_TABLE);
if (list_member_oid(aschemaPubids, puboid))
{
topmost_relid = ancestor;
@@ -608,7 +658,7 @@ pub_collist_to_bitmapset(Bitmapset *columns, Datum pubcols, MemoryContext mcxt)
* Insert new publication / schema mapping.
*/
ObjectAddress
-publication_add_schema(Oid pubid, Oid schemaid, bool if_not_exists)
+publication_add_schema(Oid pubid, Oid schemaid, char objectType, bool if_not_exists)
{
Relation rel;
HeapTuple tup;
@@ -620,6 +670,8 @@ publication_add_schema(Oid pubid, Oid schemaid, bool if_not_exists)
ObjectAddress myself,
referenced;
+ AssertObjectTypeValid(objectType);
+
rel = table_open(PublicationNamespaceRelationId, RowExclusiveLock);
/*
@@ -627,9 +679,10 @@ publication_add_schema(Oid pubid, Oid schemaid, bool if_not_exists)
* duplicates, it's here just to provide nicer error message in common
* case. The real protection is the unique key on the catalog.
*/
- if (SearchSysCacheExists2(PUBLICATIONNAMESPACEMAP,
+ if (SearchSysCacheExists3(PUBLICATIONNAMESPACEMAP,
ObjectIdGetDatum(schemaid),
- ObjectIdGetDatum(pubid)))
+ ObjectIdGetDatum(pubid),
+ CharGetDatum(objectType)))
{
table_close(rel, RowExclusiveLock);
@@ -655,6 +708,8 @@ publication_add_schema(Oid pubid, Oid schemaid, bool if_not_exists)
ObjectIdGetDatum(pubid);
values[Anum_pg_publication_namespace_pnnspid - 1] =
ObjectIdGetDatum(schemaid);
+ values[Anum_pg_publication_namespace_pntype - 1] =
+ CharGetDatum(objectType);
tup = heap_form_tuple(RelationGetDescr(rel), values, nulls);
@@ -680,7 +735,7 @@ publication_add_schema(Oid pubid, Oid schemaid, bool if_not_exists)
* publication_add_relation for why we need to consider all the
* partitions.
*/
- schemaRels = GetSchemaPublicationRelations(schemaid,
+ schemaRels = GetSchemaPublicationRelations(schemaid, objectType,
PUBLICATION_PART_ALL);
InvalidatePublicationRels(schemaRels);
@@ -714,11 +769,14 @@ GetRelationPublications(Oid relid)
/*
* Gets list of relation oids for a publication.
*
- * This should only be used FOR TABLE publications, the FOR ALL TABLES
- * should use GetAllTablesPublicationRelations().
+ * This should only be used FOR TABLE / FOR SEQUENCE publications, the FOR
+ * ALL TABLES / SEQUENCES should use GetAllTablesPublicationRelations()
+ * and GetAllSequencesPublicationRelations().
+ *
+ * XXX pub_partopt only matters for tables, not sequences.
*/
List *
-GetPublicationRelations(Oid pubid, PublicationPartOpt pub_partopt)
+GetPublicationRelations(Oid pubid, char objectType, PublicationPartOpt pub_partopt)
{
List *result;
Relation pubrelsrel;
@@ -726,6 +784,8 @@ GetPublicationRelations(Oid pubid, PublicationPartOpt pub_partopt)
SysScanDesc scan;
HeapTuple tup;
+ AssertObjectTypeValid(objectType);
+
/* Find all publications associated with the relation. */
pubrelsrel = table_open(PublicationRelRelationId, AccessShareLock);
@@ -740,11 +800,29 @@ GetPublicationRelations(Oid pubid, PublicationPartOpt pub_partopt)
result = NIL;
while (HeapTupleIsValid(tup = systable_getnext(scan)))
{
+ char relkind;
Form_pg_publication_rel pubrel;
pubrel = (Form_pg_publication_rel) GETSTRUCT(tup);
- result = GetPubPartitionOptionRelations(result, pub_partopt,
- pubrel->prrelid);
+ relkind = get_rel_relkind(pubrel->prrelid);
+
+ /*
+ * If the relkind does not match the requested object type, ignore the
+ * relation. For example we might be interested only in sequences, so
+ * we ignore tables.
+ */
+ if (!pub_object_type_matches_relkind(objectType, relkind))
+ continue;
+
+ /*
+ * We don't have partitioned sequences, so just add them to the list.
+ * Otherwise consider adding all child relations, if requested.
+ */
+ if (relkind == RELKIND_SEQUENCE)
+ result = lappend_oid(result, pubrel->prrelid);
+ else
+ result = GetPubPartitionOptionRelations(result, pub_partopt,
+ pubrel->prrelid);
}
systable_endscan(scan);
@@ -794,6 +872,43 @@ GetAllTablesPublications(void)
return result;
}
+/*
+ * Gets list of publication oids for publications marked as FOR ALL SEQUENCES.
+ */
+List *
+GetAllSequencesPublications(void)
+{
+ List *result;
+ Relation rel;
+ ScanKeyData scankey;
+ SysScanDesc scan;
+ HeapTuple tup;
+
+ /* Find all publications that are marked as for all sequences. */
+ rel = table_open(PublicationRelationId, AccessShareLock);
+
+ ScanKeyInit(&scankey,
+ Anum_pg_publication_puballsequences,
+ BTEqualStrategyNumber, F_BOOLEQ,
+ BoolGetDatum(true));
+
+ scan = systable_beginscan(rel, InvalidOid, false,
+ NULL, 1, &scankey);
+
+ result = NIL;
+ while (HeapTupleIsValid(tup = systable_getnext(scan)))
+ {
+ Oid oid = ((Form_pg_publication) GETSTRUCT(tup))->oid;
+
+ result = lappend_oid(result, oid);
+ }
+
+ systable_endscan(scan);
+ table_close(rel, AccessShareLock);
+
+ return result;
+}
+
/*
* Gets list of all relation published by FOR ALL TABLES publication(s).
*
@@ -860,28 +975,38 @@ GetAllTablesPublicationRelations(bool pubviaroot)
/*
* Gets the list of schema oids for a publication.
*
- * This should only be used FOR TABLES IN SCHEMA publications.
+ * This should only be used FOR TABLES IN SCHEMA and FOR SEQUENCES IN SCHEMA
+ * publications.
+ *
+ * 'objectType' determines whether to get FOR TABLE or FOR SEQUENCES schemas
*/
List *
-GetPublicationSchemas(Oid pubid)
+GetPublicationSchemas(Oid pubid, char objectType)
{
List *result = NIL;
Relation pubschsrel;
- ScanKeyData scankey;
+ ScanKeyData scankey[2];
SysScanDesc scan;
HeapTuple tup;
+ AssertObjectTypeValid(objectType);
+
/* Find all schemas associated with the publication */
pubschsrel = table_open(PublicationNamespaceRelationId, AccessShareLock);
- ScanKeyInit(&scankey,
+ ScanKeyInit(&scankey[0],
Anum_pg_publication_namespace_pnpubid,
BTEqualStrategyNumber, F_OIDEQ,
ObjectIdGetDatum(pubid));
+ ScanKeyInit(&scankey[1],
+ Anum_pg_publication_namespace_pntype,
+ BTEqualStrategyNumber, F_CHAREQ,
+ CharGetDatum(objectType));
+
scan = systable_beginscan(pubschsrel,
- PublicationNamespacePnnspidPnpubidIndexId,
- true, NULL, 1, &scankey);
+ PublicationNamespacePnnspidPnpubidPntypeIndexId,
+ true, NULL, 2, scankey);
while (HeapTupleIsValid(tup = systable_getnext(scan)))
{
Form_pg_publication_namespace pubsch;
@@ -899,14 +1024,26 @@ GetPublicationSchemas(Oid pubid)
/*
* Gets the list of publication oids associated with a specified schema.
+ *
+ * objectType specifies whether we're looking for schemas including tables or
+ * sequences.
+ *
+ * Note: relcache calls this for all object types, not just tables and sequences.
+ * Which is why we handle the PUB_OBJTYPE_UNSUPPORTED object type too.
*/
List *
-GetSchemaPublications(Oid schemaid)
+GetSchemaPublications(Oid schemaid, char objectType)
{
List *result = NIL;
CatCList *pubschlist;
int i;
+ /* unsupported object type */
+ if (objectType == PUB_OBJTYPE_UNSUPPORTED)
+ return result;
+
+ AssertObjectTypeValid(objectType);
+
/* Find all publications associated with the schema */
pubschlist = SearchSysCacheList1(PUBLICATIONNAMESPACEMAP,
ObjectIdGetDatum(schemaid));
@@ -914,6 +1051,11 @@ GetSchemaPublications(Oid schemaid)
{
HeapTuple tup = &pubschlist->members[i]->tuple;
Oid pubid = ((Form_pg_publication_namespace) GETSTRUCT(tup))->pnpubid;
+ char pntype = ((Form_pg_publication_namespace) GETSTRUCT(tup))->pntype;
+
+ /* Skip schemas publishing a different object type. */
+ if (pntype != objectType)
+ continue;
result = lappend_oid(result, pubid);
}
@@ -925,9 +1067,13 @@ GetSchemaPublications(Oid schemaid)
/*
* Get the list of publishable relation oids for a specified schema.
+ *
+ * objectType specifies whether this is FOR ALL TABLES IN SCHEMA or FOR ALL
+ * SEQUENCES IN SCHEMA
*/
List *
-GetSchemaPublicationRelations(Oid schemaid, PublicationPartOpt pub_partopt)
+GetSchemaPublicationRelations(Oid schemaid, char objectType,
+ PublicationPartOpt pub_partopt)
{
Relation classRel;
ScanKeyData key[1];
@@ -936,6 +1082,7 @@ GetSchemaPublicationRelations(Oid schemaid, PublicationPartOpt pub_partopt)
List *result = NIL;
Assert(OidIsValid(schemaid));
+ AssertObjectTypeValid(objectType);
classRel = table_open(RelationRelationId, AccessShareLock);
@@ -956,9 +1103,17 @@ GetSchemaPublicationRelations(Oid schemaid, PublicationPartOpt pub_partopt)
continue;
relkind = get_rel_relkind(relid);
- if (relkind == RELKIND_RELATION)
- result = lappend_oid(result, relid);
- else if (relkind == RELKIND_PARTITIONED_TABLE)
+
+ /* Skip if the relkind does not match FOR ALL TABLES / SEQUENCES. */
+ if (!pub_object_type_matches_relkind(objectType, relkind))
+ continue;
+
+ /*
+ * If the object is a partitioned table, lookup all the child
+ * relations (if requested). Otherwise just add the object to the
+ * list.
+ */
+ if (relkind == RELKIND_PARTITIONED_TABLE)
{
List *partitionrels = NIL;
@@ -971,7 +1126,11 @@ GetSchemaPublicationRelations(Oid schemaid, PublicationPartOpt pub_partopt)
pub_partopt,
relForm->oid);
result = list_concat_unique_oid(result, partitionrels);
+ continue;
}
+
+ /* non-partitioned tables and sequences */
+ result = lappend_oid(result, relid);
}
table_endscan(scan);
@@ -980,28 +1139,68 @@ GetSchemaPublicationRelations(Oid schemaid, PublicationPartOpt pub_partopt)
}
/*
- * Gets the list of all relations published by FOR TABLES IN SCHEMA
- * publication.
+ * Gets the list of all relations published by FOR TABLES IN SCHEMA or
+ * FOR SEQUENCES IN SCHEMA publication.
*/
List *
-GetAllSchemaPublicationRelations(Oid pubid, PublicationPartOpt pub_partopt)
+GetAllSchemaPublicationRelations(Oid pubid, char objectType,
+ PublicationPartOpt pub_partopt)
{
List *result = NIL;
- List *pubschemalist = GetPublicationSchemas(pubid);
+ List *pubschemalist = GetPublicationSchemas(pubid, objectType);
ListCell *cell;
+ AssertObjectTypeValid(objectType);
+
foreach(cell, pubschemalist)
{
Oid schemaid = lfirst_oid(cell);
List *schemaRels = NIL;
- schemaRels = GetSchemaPublicationRelations(schemaid, pub_partopt);
+ schemaRels = GetSchemaPublicationRelations(schemaid, objectType,
+ pub_partopt);
result = list_concat(result, schemaRels);
}
return result;
}
+/*
+ * Gets list of all relation published by FOR ALL SEQUENCES publication(s).
+ */
+List *
+GetAllSequencesPublicationRelations(void)
+{
+ Relation classRel;
+ ScanKeyData key[1];
+ TableScanDesc scan;
+ HeapTuple tuple;
+ List *result = NIL;
+
+ classRel = table_open(RelationRelationId, AccessShareLock);
+
+ ScanKeyInit(&key[0],
+ Anum_pg_class_relkind,
+ BTEqualStrategyNumber, F_CHAREQ,
+ CharGetDatum(RELKIND_SEQUENCE));
+
+ scan = table_beginscan_catalog(classRel, 1, key);
+
+ while ((tuple = heap_getnext(scan, ForwardScanDirection)) != NULL)
+ {
+ Form_pg_class relForm = (Form_pg_class) GETSTRUCT(tuple);
+ Oid relid = relForm->oid;
+
+ if (is_publishable_class(relid, relForm))
+ result = lappend_oid(result, relid);
+ }
+
+ table_endscan(scan);
+
+ table_close(classRel, AccessShareLock);
+ return result;
+}
+
/*
* Get publication using oid
*
@@ -1024,10 +1223,12 @@ GetPublication(Oid pubid)
pub->oid = pubid;
pub->name = pstrdup(NameStr(pubform->pubname));
pub->alltables = pubform->puballtables;
+ pub->allsequences = pubform->puballsequences;
pub->pubactions.pubinsert = pubform->pubinsert;
pub->pubactions.pubupdate = pubform->pubupdate;
pub->pubactions.pubdelete = pubform->pubdelete;
pub->pubactions.pubtruncate = pubform->pubtruncate;
+ pub->pubactions.pubsequence = pubform->pubsequence;
pub->pubviaroot = pubform->pubviaroot;
ReleaseSysCache(tup);
@@ -1108,10 +1309,12 @@ pg_get_publication_tables(PG_FUNCTION_ARGS)
*schemarelids;
relids = GetPublicationRelations(pub_elem->oid,
+ PUB_OBJTYPE_TABLE,
pub_elem->pubviaroot ?
PUBLICATION_PART_ROOT :
PUBLICATION_PART_LEAF);
schemarelids = GetAllSchemaPublicationRelations(pub_elem->oid,
+ PUB_OBJTYPE_TABLE,
pub_elem->pubviaroot ?
PUBLICATION_PART_ROOT :
PUBLICATION_PART_LEAF);
@@ -1197,9 +1400,10 @@ pg_get_publication_tables(PG_FUNCTION_ARGS)
* FOR TABLES IN SCHEMA publications.
*/
if (!pub->alltables &&
- !SearchSysCacheExists2(PUBLICATIONNAMESPACEMAP,
+ !SearchSysCacheExists3(PUBLICATIONNAMESPACEMAP,
ObjectIdGetDatum(schemaid),
- ObjectIdGetDatum(pub->oid)))
+ ObjectIdGetDatum(pub->oid),
+ PUB_OBJTYPE_TABLE))
pubtuple = SearchSysCacheCopy2(PUBLICATIONRELMAP,
ObjectIdGetDatum(relid),
ObjectIdGetDatum(pub->oid));
@@ -1259,3 +1463,71 @@ pg_get_publication_tables(PG_FUNCTION_ARGS)
SRF_RETURN_DONE(funcctx);
}
+
+/*
+ * Returns Oids of sequences in a publication.
+ */
+Datum
+pg_get_publication_sequences(PG_FUNCTION_ARGS)
+{
+ FuncCallContext *funcctx;
+ char *pubname = text_to_cstring(PG_GETARG_TEXT_PP(0));
+ Publication *publication;
+ List *sequences;
+
+ /* stuff done only on the first call of the function */
+ if (SRF_IS_FIRSTCALL())
+ {
+ MemoryContext oldcontext;
+
+ /* create a function context for cross-call persistence */
+ funcctx = SRF_FIRSTCALL_INIT();
+
+ /* switch to memory context appropriate for multiple function calls */
+ oldcontext = MemoryContextSwitchTo(funcctx->multi_call_memory_ctx);
+
+ publication = GetPublicationByName(pubname, false);
+
+ /*
+ * Publications support partitioned tables, although all changes are
+ * replicated using leaf partition identity and schema, so we only
+ * need those.
+ */
+ if (publication->allsequences)
+ sequences = GetAllSequencesPublicationRelations();
+ else
+ {
+ List *relids,
+ *schemarelids;
+
+ relids = GetPublicationRelations(publication->oid,
+ PUB_OBJTYPE_SEQUENCE,
+ publication->pubviaroot ?
+ PUBLICATION_PART_ROOT :
+ PUBLICATION_PART_LEAF);
+ schemarelids = GetAllSchemaPublicationRelations(publication->oid,
+ PUB_OBJTYPE_SEQUENCE,
+ publication->pubviaroot ?
+ PUBLICATION_PART_ROOT :
+ PUBLICATION_PART_LEAF);
+ sequences = list_concat_unique_oid(relids, schemarelids);
+ }
+
+ funcctx->user_fctx = (void *) sequences;
+
+ MemoryContextSwitchTo(oldcontext);
+ }
+
+ /* stuff done on every call of the function */
+ funcctx = SRF_PERCALL_SETUP();
+ sequences = (List *) funcctx->user_fctx;
+
+ if (funcctx->call_cntr < list_length(sequences))
+ {
+ Oid relid = list_nth_oid(sequences, funcctx->call_cntr);
+
+ SRF_RETURN_NEXT(funcctx, ObjectIdGetDatum(relid));
+ }
+
+ SRF_RETURN_DONE(funcctx);
+}
diff --git a/src/backend/catalog/system_views.sql b/src/backend/catalog/system_views.sql
index e43e36f5ac6..dfcfe494d4f 100644
--- a/src/backend/catalog/system_views.sql
+++ b/src/backend/catalog/system_views.sql
@@ -401,6 +401,16 @@ CREATE VIEW pg_publication_tables AS
pg_class C JOIN pg_namespace N ON (N.oid = C.relnamespace)
WHERE C.oid = GPT.relid;
+CREATE VIEW pg_publication_sequences AS
+ SELECT
+ P.pubname AS pubname,
+ N.nspname AS schemaname,
+ C.relname AS sequencename
+ FROM pg_publication P,
+ LATERAL pg_get_publication_sequences(P.pubname) GPS,
+ pg_class C JOIN pg_namespace N ON (N.oid = C.relnamespace)
+ WHERE C.oid = GPS.relid;
+
CREATE VIEW pg_locks AS
SELECT * FROM pg_lock_status() AS L;
diff --git a/src/backend/commands/publicationcmds.c b/src/backend/commands/publicationcmds.c
index 88db8258d71..3dd2c5a5d75 100644
--- a/src/backend/commands/publicationcmds.c
+++ b/src/backend/commands/publicationcmds.c
@@ -16,6 +16,7 @@
#include "access/genam.h"
#include "access/htup_details.h"
+#include "access/relation.h"
#include "access/table.h"
#include "access/xact.h"
#include "catalog/catalog.h"
@@ -68,15 +69,17 @@ typedef struct rf_context
Oid parentid; /* relid of the parent relation */
} rf_context;
-static List *OpenTableList(List *tables);
-static void CloseTableList(List *rels);
+static List *OpenRelationList(List *rels, char objectType);
+static void CloseRelationList(List *rels);
static void LockSchemaList(List *schemalist);
-static void PublicationAddTables(Oid pubid, List *rels, bool if_not_exists,
- AlterPublicationStmt *stmt);
-static void PublicationDropTables(Oid pubid, List *rels, bool missing_ok);
-static void PublicationAddSchemas(Oid pubid, List *schemas, bool if_not_exists,
- AlterPublicationStmt *stmt);
-static void PublicationDropSchemas(Oid pubid, List *schemas, bool missing_ok);
+static void PublicationAddRelations(Oid pubid, List *rels, bool if_not_exists,
+ AlterPublicationStmt *stmt);
+static void PublicationDropRelations(Oid pubid, List *rels, bool missing_ok);
+static void PublicationAddSchemas(Oid pubid, List *schemas, char objectType,
+ bool if_not_exists, AlterPublicationStmt *stmt);
+static void PublicationDropSchemas(Oid pubid, List *schemas, char objectType,
+ bool missing_ok);
+
static void
@@ -97,6 +100,7 @@ parse_publication_options(ParseState *pstate,
pubactions->pubupdate = true;
pubactions->pubdelete = true;
pubactions->pubtruncate = true;
+ pubactions->pubsequence = true;
*publish_via_partition_root = false;
/* Parse options */
@@ -121,6 +125,7 @@ parse_publication_options(ParseState *pstate,
pubactions->pubupdate = false;
pubactions->pubdelete = false;
pubactions->pubtruncate = false;
+ pubactions->pubsequence = false;
*publish_given = true;
publish = defGetString(defel);
@@ -144,6 +149,8 @@ parse_publication_options(ParseState *pstate,
pubactions->pubdelete = true;
else if (strcmp(publish_opt, "truncate") == 0)
pubactions->pubtruncate = true;
+ else if (strcmp(publish_opt, "sequence") == 0)
+ pubactions->pubsequence = true;
else
ereport(ERROR,
(errcode(ERRCODE_SYNTAX_ERROR),
@@ -171,7 +178,8 @@ parse_publication_options(ParseState *pstate,
*/
static void
ObjectsInPublicationToOids(List *pubobjspec_list, ParseState *pstate,
- List **rels, List **schemas)
+ List **tables, List **sequences,
+ List **tables_schemas, List **sequences_schemas)
{
ListCell *cell;
PublicationObjSpec *pubobj;
@@ -189,13 +197,22 @@ ObjectsInPublicationToOids(List *pubobjspec_list, ParseState *pstate,
switch (pubobj->pubobjtype)
{
case PUBLICATIONOBJ_TABLE:
- *rels = lappend(*rels, pubobj->pubtable);
+ *tables = lappend(*tables, pubobj->pubtable);
+ break;
+ case PUBLICATIONOBJ_SEQUENCE:
+ *sequences = lappend(*sequences, pubobj->pubtable);
break;
case PUBLICATIONOBJ_TABLES_IN_SCHEMA:
schemaid = get_namespace_oid(pubobj->name, false);
/* Filter out duplicates if user specifies "sch1, sch1" */
- *schemas = list_append_unique_oid(*schemas, schemaid);
+ *tables_schemas = list_append_unique_oid(*tables_schemas, schemaid);
+ break;
+ case PUBLICATIONOBJ_SEQUENCES_IN_SCHEMA:
+ schemaid = get_namespace_oid(pubobj->name, false);
+
+ /* Filter out duplicates if user specifies "sch1, sch1" */
+ *sequences_schemas = list_append_unique_oid(*sequences_schemas, schemaid);
break;
case PUBLICATIONOBJ_TABLES_IN_CUR_SCHEMA:
search_path = fetch_search_path(false);
@@ -208,7 +225,20 @@ ObjectsInPublicationToOids(List *pubobjspec_list, ParseState *pstate,
list_free(search_path);
/* Filter out duplicates if user specifies "sch1, sch1" */
- *schemas = list_append_unique_oid(*schemas, schemaid);
+ *tables_schemas = list_append_unique_oid(*tables_schemas, schemaid);
+ break;
+ case PUBLICATIONOBJ_SEQUENCES_IN_CUR_SCHEMA:
+ search_path = fetch_search_path(false);
+ if (search_path == NIL) /* nothing valid in search_path? */
+ ereport(ERROR,
+ errcode(ERRCODE_UNDEFINED_SCHEMA),
+ errmsg("no schema has been selected for CURRENT_SCHEMA"));
+
+ schemaid = linitial_oid(search_path);
+ list_free(search_path);
+
+ /* Filter out duplicates if user specifies "sch1, sch1" */
+ *sequences_schemas = list_append_unique_oid(*sequences_schemas, schemaid);
break;
default:
/* shouldn't happen */
@@ -734,6 +764,7 @@ CheckPubRelationColumnList(char *pubname, List *tables,
ObjectAddress
CreatePublication(ParseState *pstate, CreatePublicationStmt *stmt)
{
+ ListCell *lc;
Relation rel;
ObjectAddress myself;
Oid puboid;
@@ -745,8 +776,27 @@ CreatePublication(ParseState *pstate, CreatePublicationStmt *stmt)
bool publish_via_partition_root_given;
bool publish_via_partition_root;
AclResult aclresult;
- List *relations = NIL;
- List *schemaidlist = NIL;
+ List *tables = NIL;
+ List *sequences = NIL;
+ List *tables_schemaidlist = NIL;
+ List *sequences_schemaidlist = NIL;
+
+ bool for_all_tables = false;
+ bool for_all_sequences = false;
+
+ /*
+ * Translate the list of object types (represented by strings) to bool
+ * flags.
+ */
+ foreach(lc, stmt->for_all_objects)
+ {
+ char *val = strVal(lfirst(lc));
+
+ if (strcmp(val, "tables") == 0)
+ for_all_tables = true;
+ else if (strcmp(val, "sequences") == 0)
+ for_all_sequences = true;
+ }
/* must have CREATE privilege on database */
aclresult = object_aclcheck(DatabaseRelationId, MyDatabaseId, GetUserId(), ACL_CREATE);
@@ -755,11 +805,17 @@ CreatePublication(ParseState *pstate, CreatePublicationStmt *stmt)
get_database_name(MyDatabaseId));
/* FOR ALL TABLES requires superuser */
- if (stmt->for_all_tables && !superuser())
+ if (for_all_tables && !superuser())
ereport(ERROR,
(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
errmsg("must be superuser to create FOR ALL TABLES publication")));
+ /* FOR ALL SEQUENCES requires superuser */
+ if (for_all_sequences && !superuser())
+ ereport(ERROR,
+ (errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
+ errmsg("must be superuser to create FOR ALL SEQUENCES publication")));
+
rel = table_open(PublicationRelationId, RowExclusiveLock);
/* Check if name is used */
@@ -789,7 +845,9 @@ CreatePublication(ParseState *pstate, CreatePublicationStmt *stmt)
Anum_pg_publication_oid);
values[Anum_pg_publication_oid - 1] = ObjectIdGetDatum(puboid);
values[Anum_pg_publication_puballtables - 1] =
- BoolGetDatum(stmt->for_all_tables);
+ BoolGetDatum(for_all_tables);
+ values[Anum_pg_publication_puballsequences - 1] =
+ BoolGetDatum(for_all_sequences);
values[Anum_pg_publication_pubinsert - 1] =
BoolGetDatum(pubactions.pubinsert);
values[Anum_pg_publication_pubupdate - 1] =
@@ -798,6 +856,8 @@ CreatePublication(ParseState *pstate, CreatePublicationStmt *stmt)
BoolGetDatum(pubactions.pubdelete);
values[Anum_pg_publication_pubtruncate - 1] =
BoolGetDatum(pubactions.pubtruncate);
+ values[Anum_pg_publication_pubsequence - 1] =
+ BoolGetDatum(pubactions.pubsequence);
values[Anum_pg_publication_pubviaroot - 1] =
BoolGetDatum(publish_via_partition_root);
@@ -815,46 +875,88 @@ CreatePublication(ParseState *pstate, CreatePublicationStmt *stmt)
CommandCounterIncrement();
/* Associate objects with the publication. */
- if (stmt->for_all_tables)
+ if (for_all_tables || for_all_sequences)
{
/* Invalidate relcache so that publication info is rebuilt. */
CacheInvalidateRelcacheAll();
}
- else
+
+ /*
+ * If the publication might have either tables or sequences (directly or
+ * through a schema), process that.
+ */
+ if (!for_all_tables || !for_all_sequences)
{
- ObjectsInPublicationToOids(stmt->pubobjects, pstate, &relations,
- &schemaidlist);
+ ObjectsInPublicationToOids(stmt->pubobjects, pstate,
+ &tables, &sequences,
+ &tables_schemaidlist,
+ &sequences_schemaidlist);
/* FOR TABLES IN SCHEMA requires superuser */
- if (schemaidlist != NIL && !superuser())
+ if (tables_schemaidlist != NIL && !superuser())
ereport(ERROR,
errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
errmsg("must be superuser to create FOR TABLES IN SCHEMA publication"));
- if (relations != NIL)
+ /* FOR SEQUENCES IN SCHEMA requires superuser */
+ if (sequences_schemaidlist != NIL && !superuser())
+ ereport(ERROR,
+ errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
+ errmsg("must be superuser to create FOR ALL SEQUENCES IN SCHEMA publication"));
+
+ /* tables added directly */
+ if (tables != NIL)
{
List *rels;
- rels = OpenTableList(relations);
+ rels = OpenRelationList(tables, PUB_OBJTYPE_TABLE);
+
TransformPubWhereClauses(rels, pstate->p_sourcetext,
publish_via_partition_root);
CheckPubRelationColumnList(stmt->pubname, rels,
- schemaidlist != NIL,
+ tables_schemaidlist != NIL,
publish_via_partition_root);
- PublicationAddTables(puboid, rels, true, NULL);
- CloseTableList(rels);
+ PublicationAddRelations(puboid, rels, true, NULL);
+ CloseRelationList(rels);
+ }
+
+ /* sequences added directly */
+ if (sequences != NIL)
+ {
+ List *rels;
+
+ rels = OpenRelationList(sequences, PUB_OBJTYPE_SEQUENCE);
+
+ PublicationAddRelations(puboid, rels, true, NULL);
+ CloseRelationList(rels);
}
- if (schemaidlist != NIL)
+ /* tables added through a schema */
+ if (tables_schemaidlist != NIL)
{
/*
* Schema lock is held until the publication is created to prevent
* concurrent schema deletion.
*/
- LockSchemaList(schemaidlist);
- PublicationAddSchemas(puboid, schemaidlist, true, NULL);
+ LockSchemaList(tables_schemaidlist);
+ PublicationAddSchemas(puboid,
+ tables_schemaidlist, PUB_OBJTYPE_TABLE,
+ true, NULL);
+ }
+
+ /* sequences added through a schema */
+ if (sequences_schemaidlist != NIL)
+ {
+ /*
+ * Schema lock is held until the publication is created to prevent
+ * concurrent schema deletion.
+ */
+ LockSchemaList(sequences_schemaidlist);
+ PublicationAddSchemas(puboid,
+ sequences_schemaidlist, PUB_OBJTYPE_SEQUENCE,
+ true, NULL);
}
}
@@ -917,6 +1019,7 @@ AlterPublicationOptions(ParseState *pstate, AlterPublicationStmt *stmt,
AccessShareLock);
root_relids = GetPublicationRelations(pubform->oid,
+ PUB_OBJTYPE_TABLE,
PUBLICATION_PART_ROOT);
foreach(lc, root_relids)
@@ -996,6 +1099,9 @@ AlterPublicationOptions(ParseState *pstate, AlterPublicationStmt *stmt,
values[Anum_pg_publication_pubtruncate - 1] = BoolGetDatum(pubactions.pubtruncate);
replaces[Anum_pg_publication_pubtruncate - 1] = true;
+
+ values[Anum_pg_publication_pubsequence - 1] = BoolGetDatum(pubactions.pubsequence);
+ replaces[Anum_pg_publication_pubsequence - 1] = true;
}
if (publish_via_partition_root_given)
@@ -1015,7 +1121,7 @@ AlterPublicationOptions(ParseState *pstate, AlterPublicationStmt *stmt,
pubform = (Form_pg_publication) GETSTRUCT(tup);
/* Invalidate the relcache. */
- if (pubform->puballtables)
+ if (pubform->puballtables || pubform->puballsequences)
{
CacheInvalidateRelcacheAll();
}
@@ -1031,6 +1137,7 @@ AlterPublicationOptions(ParseState *pstate, AlterPublicationStmt *stmt,
*/
if (root_relids == NIL)
relids = GetPublicationRelations(pubform->oid,
+ PUB_OBJTYPE_TABLE,
PUBLICATION_PART_ALL);
else
{
@@ -1044,7 +1151,20 @@ AlterPublicationOptions(ParseState *pstate, AlterPublicationStmt *stmt,
lfirst_oid(lc));
}
+ /* tables */
schemarelids = GetAllSchemaPublicationRelations(pubform->oid,
+ PUB_OBJTYPE_TABLE,
+ PUBLICATION_PART_ALL);
+ relids = list_concat_unique_oid(relids, schemarelids);
+
+ /* sequences */
+ relids = list_concat_unique_oid(relids,
+ GetPublicationRelations(pubform->oid,
+ PUB_OBJTYPE_SEQUENCE,
+ PUBLICATION_PART_ALL));
+
+ schemarelids = GetAllSchemaPublicationRelations(pubform->oid,
+ PUB_OBJTYPE_SEQUENCE,
PUBLICATION_PART_ALL);
relids = list_concat_unique_oid(relids, schemarelids);
@@ -1099,7 +1219,7 @@ AlterPublicationTables(AlterPublicationStmt *stmt, HeapTuple tup,
if (!tables && stmt->action != AP_SetObjects)
return;
- rels = OpenTableList(tables);
+ rels = OpenRelationList(tables, PUB_OBJTYPE_TABLE);
if (stmt->action == AP_AddObjects)
{
@@ -1110,13 +1230,14 @@ AlterPublicationTables(AlterPublicationStmt *stmt, HeapTuple tup,
CheckPubRelationColumnList(stmt->pubname, rels, publish_schema,
pubform->pubviaroot);
- PublicationAddTables(pubid, rels, false, stmt);
+ PublicationAddRelations(pubid, rels, false, stmt);
}
else if (stmt->action == AP_DropObjects)
- PublicationDropTables(pubid, rels, false);
+ PublicationDropRelations(pubid, rels, false);
else /* AP_SetObjects */
{
List *oldrelids = GetPublicationRelations(pubid,
+ PUB_OBJTYPE_TABLE,
PUBLICATION_PART_ROOT);
List *delrels = NIL;
ListCell *oldlc;
@@ -1233,18 +1354,18 @@ AlterPublicationTables(AlterPublicationStmt *stmt, HeapTuple tup,
}
/* And drop them. */
- PublicationDropTables(pubid, delrels, true);
+ PublicationDropRelations(pubid, delrels, true);
/*
* Don't bother calculating the difference for adding, we'll catch and
* skip existing ones when doing catalog update.
*/
- PublicationAddTables(pubid, rels, true, stmt);
+ PublicationAddRelations(pubid, rels, true, stmt);
- CloseTableList(delrels);
+ CloseRelationList(delrels);
}
- CloseTableList(rels);
+ CloseRelationList(rels);
}
/*
@@ -1254,7 +1375,8 @@ AlterPublicationTables(AlterPublicationStmt *stmt, HeapTuple tup,
*/
static void
AlterPublicationSchemas(AlterPublicationStmt *stmt,
- HeapTuple tup, List *schemaidlist)
+ HeapTuple tup, List *schemaidlist,
+ char objectType)
{
Form_pg_publication pubform = (Form_pg_publication) GETSTRUCT(tup);
@@ -1276,7 +1398,7 @@ AlterPublicationSchemas(AlterPublicationStmt *stmt,
ListCell *lc;
List *reloids;
- reloids = GetPublicationRelations(pubform->oid, PUBLICATION_PART_ROOT);
+ reloids = GetPublicationRelations(pubform->oid, objectType, PUBLICATION_PART_ROOT);
foreach(lc, reloids)
{
@@ -1303,13 +1425,13 @@ AlterPublicationSchemas(AlterPublicationStmt *stmt,
ReleaseSysCache(coltuple);
}
- PublicationAddSchemas(pubform->oid, schemaidlist, false, stmt);
+ PublicationAddSchemas(pubform->oid, schemaidlist, objectType, false, stmt);
}
else if (stmt->action == AP_DropObjects)
- PublicationDropSchemas(pubform->oid, schemaidlist, false);
+ PublicationDropSchemas(pubform->oid, schemaidlist, objectType, false);
else /* AP_SetObjects */
{
- List *oldschemaids = GetPublicationSchemas(pubform->oid);
+ List *oldschemaids = GetPublicationSchemas(pubform->oid, objectType);
List *delschemas = NIL;
/* Identify which schemas should be dropped */
@@ -1322,13 +1444,13 @@ AlterPublicationSchemas(AlterPublicationStmt *stmt,
LockSchemaList(delschemas);
/* And drop them */
- PublicationDropSchemas(pubform->oid, delschemas, true);
+ PublicationDropSchemas(pubform->oid, delschemas, objectType, true);
/*
* Don't bother calculating the difference for adding, we'll catch and
* skip existing ones when doing catalog update.
*/
- PublicationAddSchemas(pubform->oid, schemaidlist, true, stmt);
+ PublicationAddSchemas(pubform->oid, schemaidlist, objectType, true, stmt);
}
}
@@ -1338,12 +1460,13 @@ AlterPublicationSchemas(AlterPublicationStmt *stmt,
*/
static void
CheckAlterPublication(AlterPublicationStmt *stmt, HeapTuple tup,
- List *tables, List *schemaidlist)
+ List *tables, List *tables_schemaidlist,
+ List *sequences, List *sequences_schemaidlist)
{
Form_pg_publication pubform = (Form_pg_publication) GETSTRUCT(tup);
if ((stmt->action == AP_AddObjects || stmt->action == AP_SetObjects) &&
- schemaidlist && !superuser())
+ (tables_schemaidlist || sequences_schemaidlist) && !superuser())
ereport(ERROR,
(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
errmsg("must be superuser to add or set schemas")));
@@ -1352,13 +1475,24 @@ CheckAlterPublication(AlterPublicationStmt *stmt, HeapTuple tup,
* Check that user is allowed to manipulate the publication tables in
* schema
*/
- if (schemaidlist && pubform->puballtables)
+ if (tables_schemaidlist && pubform->puballtables)
ereport(ERROR,
(errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
errmsg("publication \"%s\" is defined as FOR ALL TABLES",
NameStr(pubform->pubname)),
errdetail("Schemas cannot be added to or dropped from FOR ALL TABLES publications.")));
+ /*
+ * Check that user is allowed to manipulate the publication sequences in
+ * schema
+ */
+ if (sequences_schemaidlist && pubform->puballsequences)
+ ereport(ERROR,
+ (errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
+ errmsg("publication \"%s\" is defined as FOR ALL SEQUENCES",
+ NameStr(pubform->pubname)),
+ errdetail("Sequences from schema cannot be added to, dropped from, or set on FOR ALL SEQUENCES publications.")));
+
/* Check that user is allowed to manipulate the publication tables. */
if (tables && pubform->puballtables)
ereport(ERROR,
@@ -1366,6 +1500,107 @@ CheckAlterPublication(AlterPublicationStmt *stmt, HeapTuple tup,
errmsg("publication \"%s\" is defined as FOR ALL TABLES",
NameStr(pubform->pubname)),
errdetail("Tables cannot be added to or dropped from FOR ALL TABLES publications.")));
+
+ /* Check that user is allowed to manipulate the publication sequences. */
+ if (sequences && pubform->puballsequences)
+ ereport(ERROR,
+ (errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
+ errmsg("publication \"%s\" is defined as FOR ALL SEQUENCES",
+ NameStr(pubform->pubname)),
+ errdetail("Sequences cannot be added to or dropped from FOR ALL SEQUENCES publications.")));
+}
+
+/*
+ * Add or remove table to/from publication.
+ */
+static void
+AlterPublicationSequences(AlterPublicationStmt *stmt, HeapTuple tup,
+ List *sequences)
+{
+ List *rels = NIL;
+ Form_pg_publication pubform = (Form_pg_publication) GETSTRUCT(tup);
+ Oid pubid = pubform->oid;
+
+ /*
+ * Nothing to do if no objects, except in SET: for that it is quite
+ * possible that user has not specified any tables in which case we need
+ * to remove all the existing tables.
+ */
+ if (!sequences && stmt->action != AP_SetObjects)
+ return;
+
+ rels = OpenRelationList(sequences, PUB_OBJTYPE_SEQUENCE);
+
+ if (stmt->action == AP_AddObjects)
+ {
+ PublicationAddRelations(pubid, rels, false, stmt);
+ }
+ else if (stmt->action == AP_DropObjects)
+ PublicationDropRelations(pubid, rels, false);
+ else /* AP_SetObjects */
+ {
+ List *oldrelids = GetPublicationRelations(pubid,
+ PUB_OBJTYPE_SEQUENCE,
+ PUBLICATION_PART_ROOT);
+ List *delrels = NIL;
+ ListCell *oldlc;
+
+ /*
+ * To recreate the relation list for the publication, look for
+ * existing relations that do not need to be dropped.
+ */
+ foreach(oldlc, oldrelids)
+ {
+ Oid oldrelid = lfirst_oid(oldlc);
+ ListCell *newlc;
+ PublicationRelInfo *oldrel;
+ bool found = false;
+
+ foreach(newlc, rels)
+ {
+ PublicationRelInfo *newpubrel;
+
+ newpubrel = (PublicationRelInfo *) lfirst(newlc);
+
+ /*
+ * Check if any of the new set of relations matches with the
+ * existing relations in the publication.
+ */
+ if (RelationGetRelid(newpubrel->relation) == oldrelid)
+ {
+ found = true;
+ break;
+ }
+ }
+
+ /*
+ * Add the non-matched relations to a list so that they can be
+ * dropped.
+ */
+ if (!found)
+ {
+ oldrel = palloc(sizeof(PublicationRelInfo));
+ oldrel->whereClause = NULL;
+ oldrel->columns = NIL;
+ oldrel->relation = table_open(oldrelid,
+ ShareUpdateExclusiveLock);
+ delrels = lappend(delrels, oldrel);
+ }
+ }
+
+ /* And drop them. */
+ PublicationDropRelations(pubid, delrels, true);
+
+ /*
+ * Don't bother calculating the difference for adding, we'll catch and
+ * skip existing ones when doing catalog update.
+ */
+ PublicationAddRelations(pubid, rels, true, stmt);
+
+ CloseRelationList(delrels);
+ }
+
+ CloseRelationList(rels);
}
/*
@@ -1403,14 +1638,20 @@ AlterPublication(ParseState *pstate, AlterPublicationStmt *stmt)
AlterPublicationOptions(pstate, stmt, rel, tup);
else
{
- List *relations = NIL;
- List *schemaidlist = NIL;
+ List *tables = NIL;
+ List *sequences = NIL;
+ List *tables_schemaidlist = NIL;
+ List *sequences_schemaidlist = NIL;
Oid pubid = pubform->oid;
- ObjectsInPublicationToOids(stmt->pubobjects, pstate, &relations,
- &schemaidlist);
+ ObjectsInPublicationToOids(stmt->pubobjects, pstate,
+ &tables, &sequences,
+ &tables_schemaidlist,
+ &sequences_schemaidlist);
- CheckAlterPublication(stmt, tup, relations, schemaidlist);
+ CheckAlterPublication(stmt, tup,
+ tables, tables_schemaidlist,
+ sequences, sequences_schemaidlist);
heap_freetuple(tup);
@@ -1431,9 +1672,16 @@ AlterPublication(ParseState *pstate, AlterPublicationStmt *stmt)
errmsg("publication \"%s\" does not exist",
stmt->pubname));
- AlterPublicationTables(stmt, tup, relations, pstate->p_sourcetext,
- schemaidlist != NIL);
- AlterPublicationSchemas(stmt, tup, schemaidlist);
+ AlterPublicationTables(stmt, tup, tables, pstate->p_sourcetext,
+ tables_schemaidlist != NIL);
+
+ AlterPublicationSchemas(stmt, tup, tables_schemaidlist,
+ PUB_OBJTYPE_TABLE);
+
+ AlterPublicationSequences(stmt, tup, sequences);
+
+ AlterPublicationSchemas(stmt, tup, sequences_schemaidlist,
+ PUB_OBJTYPE_SEQUENCE);
}
/* Cleanup. */
@@ -1501,7 +1749,7 @@ RemovePublicationById(Oid pubid)
pubform = (Form_pg_publication) GETSTRUCT(tup);
/* Invalidate relcache so that publication info is rebuilt. */
- if (pubform->puballtables)
+ if (pubform->puballtables || pubform->puballsequences)
CacheInvalidateRelcacheAll();
CatalogTupleDelete(rel, &tup->t_self);
@@ -1537,6 +1785,7 @@ RemovePublicationSchemaById(Oid psoid)
* partitions.
*/
schemaRels = GetSchemaPublicationRelations(pubsch->pnnspid,
+ pubsch->pntype,
PUBLICATION_PART_ALL);
InvalidatePublicationRels(schemaRels);
@@ -1553,10 +1802,10 @@ RemovePublicationSchemaById(Oid psoid)
* add them to a publication.
*/
static List *
-OpenTableList(List *tables)
+OpenRelationList(List *rels, char objectType)
{
List *relids = NIL;
- List *rels = NIL;
+ List *result = NIL;
ListCell *lc;
List *relids_with_rf = NIL;
List *relids_with_collist = NIL;
@@ -1564,19 +1813,35 @@ OpenTableList(List *tables)
/*
* Open, share-lock, and check all the explicitly-specified relations
*/
- foreach(lc, tables)
+ foreach(lc, rels)
{
PublicationTable *t = lfirst_node(PublicationTable, lc);
bool recurse = t->relation->inh;
Relation rel;
Oid myrelid;
PublicationRelInfo *pub_rel;
+ char myrelkind;
/* Allow query cancel in case this takes a long time */
CHECK_FOR_INTERRUPTS();
rel = table_openrv(t->relation, ShareUpdateExclusiveLock);
myrelid = RelationGetRelid(rel);
+ myrelkind = get_rel_relkind(myrelid);
+
+ /*
+ * Make sure the relkind matches the expected object type. This may
+ * happen e.g. when adding a sequence using ADD TABLE or a table using
+ * ADD SEQUENCE).
+ *
+ * XXX We let through unsupported object types (views etc.). Those
+ * will be caught later in check_publication_add_relation.
+ */
+ if (pub_get_object_type_for_relkind(myrelkind) != PUB_OBJTYPE_UNSUPPORTED &&
+ pub_get_object_type_for_relkind(myrelkind) != objectType)
+ ereport(ERROR,
+ errcode(ERRCODE_INVALID_PARAMETER_VALUE),
+ errmsg("object type does not match type expected by command"));
/*
* Filter out duplicates if user specifies "foo, foo".
@@ -1609,7 +1874,7 @@ OpenTableList(List *tables)
pub_rel->relation = rel;
pub_rel->whereClause = t->whereClause;
pub_rel->columns = t->columns;
- rels = lappend(rels, pub_rel);
+ result = lappend(result, pub_rel);
relids = lappend_oid(relids, myrelid);
if (t->whereClause)
@@ -1678,10 +1943,9 @@ OpenTableList(List *tables)
pub_rel->relation = rel;
/* child inherits WHERE clause from parent */
pub_rel->whereClause = t->whereClause;
-
/* child inherits column list from parent */
pub_rel->columns = t->columns;
- rels = lappend(rels, pub_rel);
+ result = lappend(result, pub_rel);
relids = lappend_oid(relids, childrelid);
if (t->whereClause)
@@ -1696,14 +1960,14 @@ OpenTableList(List *tables)
list_free(relids);
list_free(relids_with_rf);
- return rels;
+ return result;
}
/*
* Close all relations in the list.
*/
static void
-CloseTableList(List *rels)
+CloseRelationList(List *rels)
{
ListCell *lc;
@@ -1751,12 +2015,12 @@ LockSchemaList(List *schemalist)
* Add listed tables to the publication.
*/
static void
-PublicationAddTables(Oid pubid, List *rels, bool if_not_exists,
- AlterPublicationStmt *stmt)
+PublicationAddRelations(Oid pubid, List *rels, bool if_not_exists,
+ AlterPublicationStmt *stmt)
{
ListCell *lc;
- Assert(!stmt || !stmt->for_all_tables);
+ Assert(!stmt || !stmt->for_all_objects);
foreach(lc, rels)
{
@@ -1785,7 +2049,7 @@ PublicationAddTables(Oid pubid, List *rels, bool if_not_exists,
* Remove listed tables from the publication.
*/
static void
-PublicationDropTables(Oid pubid, List *rels, bool missing_ok)
+PublicationDropRelations(Oid pubid, List *rels, bool missing_ok)
{
ObjectAddress obj;
ListCell *lc;
@@ -1830,19 +2094,19 @@ PublicationDropTables(Oid pubid, List *rels, bool missing_ok)
* Add listed schemas to the publication.
*/
static void
-PublicationAddSchemas(Oid pubid, List *schemas, bool if_not_exists,
- AlterPublicationStmt *stmt)
+PublicationAddSchemas(Oid pubid, List *schemas, char objectType,
+ bool if_not_exists, AlterPublicationStmt *stmt)
{
ListCell *lc;
- Assert(!stmt || !stmt->for_all_tables);
+ Assert(!stmt || !stmt->for_all_objects);
foreach(lc, schemas)
{
Oid schemaid = lfirst_oid(lc);
ObjectAddress obj;
- obj = publication_add_schema(pubid, schemaid, if_not_exists);
+ obj = publication_add_schema(pubid, schemaid, objectType, if_not_exists);
if (stmt)
{
EventTriggerCollectSimpleCommand(obj, InvalidObjectAddress,
@@ -1858,7 +2122,7 @@ PublicationAddSchemas(Oid pubid, List *schemas, bool if_not_exists,
* Remove listed schemas from the publication.
*/
static void
-PublicationDropSchemas(Oid pubid, List *schemas, bool missing_ok)
+PublicationDropSchemas(Oid pubid, List *schemas, char objectType, bool missing_ok)
{
ObjectAddress obj;
ListCell *lc;
@@ -1868,10 +2132,11 @@ PublicationDropSchemas(Oid pubid, List *schemas, bool missing_ok)
{
Oid schemaid = lfirst_oid(lc);
- psid = GetSysCacheOid2(PUBLICATIONNAMESPACEMAP,
+ psid = GetSysCacheOid3(PUBLICATIONNAMESPACEMAP,
Anum_pg_publication_namespace_oid,
ObjectIdGetDatum(schemaid),
- ObjectIdGetDatum(pubid));
+ ObjectIdGetDatum(pubid),
+ CharGetDatum(objectType));
if (!OidIsValid(psid))
{
if (missing_ok)
@@ -1926,6 +2191,13 @@ AlterPublicationOwner_internal(Relation rel, HeapTuple tup, Oid newOwnerId)
NameStr(form->pubname)),
errhint("The owner of a FOR ALL TABLES publication must be a superuser.")));
+ if (form->puballsequences && !superuser_arg(newOwnerId))
+ ereport(ERROR,
+ (errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
+ errmsg("permission denied to change owner of publication \"%s\"",
+ NameStr(form->pubname)),
+ errhint("The owner of a FOR ALL SEQUENCES publication must be a superuser.")));
+
if (!superuser_arg(newOwnerId) && is_schema_publication(form->oid))
ereport(ERROR,
(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
diff --git a/src/backend/commands/sequence.c b/src/backend/commands/sequence.c
index af27d96ecd7..f3d1f3ec391 100644
--- a/src/backend/commands/sequence.c
+++ b/src/backend/commands/sequence.c
@@ -44,6 +44,7 @@
#include "utils/acl.h"
#include "utils/builtins.h"
#include "utils/lsyscache.h"
+#include "utils/pg_lsn.h"
#include "utils/resowner.h"
#include "utils/syscache.h"
#include "utils/varlena.h"
@@ -101,7 +102,8 @@ static Relation lock_and_open_sequence(SeqTable seq);
static void create_seq_hashtable(void);
static void init_sequence(Oid relid, SeqTable *p_elm, Relation *p_rel);
static Form_pg_sequence_data read_seq_tuple(Relation rel,
- Buffer *buf, HeapTuple seqdatatuple);
+ Buffer *buf, HeapTuple seqdatatuple,
+ XLogRecPtr *lsn);
static void init_params(ParseState *pstate, List *options, bool for_identity,
bool isInit,
Form_pg_sequence seqform,
@@ -284,7 +286,7 @@ ResetSequence(Oid seq_relid)
* indeed a sequence.
*/
init_sequence(seq_relid, &elm, &seq_rel);
- (void) read_seq_tuple(seq_rel, &buf, &seqdatatuple);
+ (void) read_seq_tuple(seq_rel, &buf, &seqdatatuple, NULL);
pgstuple = SearchSysCache1(SEQRELID, ObjectIdGetDatum(seq_relid));
if (!HeapTupleIsValid(pgstuple))
@@ -342,6 +344,166 @@ ResetSequence(Oid seq_relid)
relation_close(seq_rel, NoLock);
}
+/*
+ * Update the sequence state by modifying the existing sequence data row.
+ *
+ * This keeps the same relfilenode, so the behavior is non-transactional.
+ */
+static void
+SetSequence_non_transactional(Oid seqrelid, int64 value)
+{
+ SeqTable elm;
+ Relation seqrel;
+ Buffer buf;
+ HeapTupleData seqdatatuple;
+ Form_pg_sequence_data seq;
+
+ /* open and lock sequence */
+ init_sequence(seqrelid, &elm, &seqrel);
+
+ /* lock page' buffer and read tuple */
+ seq = read_seq_tuple(seqrel, &buf, &seqdatatuple, NULL);
+
+ /* check the comment above nextval_internal()'s equivalent call. */
+ if (RelationNeedsWAL(seqrel))
+ {
+ GetTopTransactionId();
+
+ if (XLogLogicalInfoActive())
+ GetCurrentTransactionId();
+ }
+
+ /* ready to change the on-disk (or really, in-buffer) tuple */
+ START_CRIT_SECTION();
+
+ seq->last_value = value;
+ seq->is_called = true;
+ seq->log_cnt = 0;
+
+ MarkBufferDirty(buf);
+
+ /* XLOG stuff */
+ if (RelationNeedsWAL(seqrel))
+ {
+ xl_seq_rec xlrec;
+ XLogRecPtr recptr;
+ Page page = BufferGetPage(buf);
+
+ XLogBeginInsert();
+ XLogRegisterBuffer(0, buf, REGBUF_WILL_INIT);
+
+ xlrec.locator = seqrel->rd_locator;
+
+ XLogRegisterData((char *) &xlrec, sizeof(xl_seq_rec));
+ XLogRegisterData((char *) seqdatatuple.t_data, seqdatatuple.t_len);
+
+ /* allow filtering by origin on a sequence update */
+ XLogSetRecordFlags(XLOG_INCLUDE_ORIGIN);
+
+ recptr = XLogInsert(RM_SEQ_ID, XLOG_SEQ_LOG);
+
+ PageSetLSN(page, recptr);
+ }
+
+ END_CRIT_SECTION();
+
+ UnlockReleaseBuffer(buf);
+
+ /* Clear local cache so that we don't think we have cached numbers */
+ /* Note that we do not change the currval() state */
+ elm->cached = elm->last;
+
+ relation_close(seqrel, NoLock);
+}
+
+/*
+ * Update the sequence state by creating a new relfilenode.
+ *
+ * This creates a new relfilenode, to allow transactional behavior.
+ */
+static void
+SetSequence_transactional(Oid seq_relid, int64 value)
+{
+ SeqTable elm;
+ Relation seqrel;
+ Buffer buf;
+ HeapTupleData seqdatatuple;
+ Form_pg_sequence_data seq;
+ HeapTuple tuple;
+
+ /* open and lock sequence */
+ init_sequence(seq_relid, &elm, &seqrel);
+
+ /* lock page' buffer and read tuple */
+ seq = read_seq_tuple(seqrel, &buf, &seqdatatuple, NULL);
+
+ /* Copy the existing sequence tuple. */
+ tuple = heap_copytuple(&seqdatatuple);
+
+ /* Now we're done with the old page */
+ UnlockReleaseBuffer(buf);
+
+ /*
+ * Modify the copied tuple to update the sequence state (similar to what
+ * ResetSequence does).
+ */
+ seq = (Form_pg_sequence_data) GETSTRUCT(tuple);
+ seq->last_value = value;
+ seq->is_called = true;
+ seq->log_cnt = 0;
+
+ /* make sure the relfilenode creation is associated with the XID */
+ if (XLogLogicalInfoActive())
+ GetCurrentTransactionId();
+
+ /*
+ * Create a new storage file for the sequence - this is needed for the
+ * transactional behavior.
+ */
+ RelationSetNewRelfilenumber(seqrel, seqrel->rd_rel->relpersistence);
+
+ /*
+ * Ensure sequence's relfrozenxid is at 0, since it won't contain any
+ * unfrozen XIDs. Same with relminmxid, since a sequence will never
+ * contain multixacts.
+ */
+ Assert(seqrel->rd_rel->relfrozenxid == InvalidTransactionId);
+ Assert(seqrel->rd_rel->relminmxid == InvalidMultiXactId);
+
+ /*
+ * Insert the modified tuple into the new storage file. This does all the
+ * necessary WAL-logging etc.
+ */
+ fill_seq_with_data(seqrel, tuple);
+
+ /* Clear local cache so that we don't think we have cached numbers */
+ /* Note that we do not change the currval() state */
+ elm->cached = elm->last;
+
+ relation_close(seqrel, NoLock);
+}
+
+/*
+ * Set a sequence to a specified internal state.
+ *
+ * The change is made transactionally, so that on failure of the current
+ * transaction, the sequence will be restored to its previous state.
+ * We do that by creating a whole new relfilenode for the sequence; so this
+ * works much like the rewriting forms of ALTER TABLE.
+ *
+ * Caller is assumed to have acquired AccessExclusiveLock on the sequence,
+ * which must not be released until end of transaction. Caller is also
+ * responsible for permissions checking.
+ */
+void
+SetSequence(Oid seq_relid, bool transactional, int64 value)
+{
+ if (transactional)
+ SetSequence_transactional(seq_relid, value);
+ else
+ SetSequence_non_transactional(seq_relid, value);
+}
+
/*
* Initialize a sequence's relation with the specified tuple as content
*
@@ -504,7 +666,7 @@ AlterSequence(ParseState *pstate, AlterSeqStmt *stmt)
seqform = (Form_pg_sequence) GETSTRUCT(seqtuple);
/* lock page buffer and read tuple into new sequence structure */
- (void) read_seq_tuple(seqrel, &buf, &datatuple);
+ (void) read_seq_tuple(seqrel, &buf, &datatuple, NULL);
/* copy the existing sequence data tuple, so it can be modified locally */
newdatatuple = heap_copytuple(&datatuple);
@@ -598,7 +760,7 @@ SequenceChangePersistence(Oid relid, char newrelpersistence)
GetCurrentTransactionId();
}
- (void) read_seq_tuple(seqrel, &buf, &seqdatatuple);
+ (void) read_seq_tuple(seqrel, &buf, &seqdatatuple, NULL);
RelationSetNewRelfilenumber(seqrel, newrelpersistence);
fill_seq_with_data(seqrel, &seqdatatuple);
UnlockReleaseBuffer(buf);
@@ -727,7 +889,7 @@ nextval_internal(Oid relid, bool check_permissions)
ReleaseSysCache(pgstuple);
/* lock page buffer and read tuple */
- seq = read_seq_tuple(seqrel, &buf, &seqdatatuple);
+ seq = read_seq_tuple(seqrel, &buf, &seqdatatuple, NULL);
page = BufferGetPage(buf);
elm->increment = incby;
@@ -1050,7 +1212,7 @@ do_setval(Oid relid, int64 next, bool iscalled)
PreventCommandIfParallelMode("setval()");
/* lock page buffer and read tuple */
- seq = read_seq_tuple(seqrel, &buf, &seqdatatuple);
+ seq = read_seq_tuple(seqrel, &buf, &seqdatatuple, NULL);
if ((next < minv) || (next > maxv))
ereport(ERROR,
@@ -1270,11 +1432,13 @@ init_sequence(Oid relid, SeqTable *p_elm, Relation *p_rel)
* *buf receives the reference to the pinned-and-ex-locked buffer
* *seqdatatuple receives the reference to the sequence tuple proper
* (this arg should point to a local variable of type HeapTupleData)
+ * *lsn receives LSN of the last sequence change (page LSN), optional
*
* Function's return value points to the data payload of the tuple
*/
static Form_pg_sequence_data
-read_seq_tuple(Relation rel, Buffer *buf, HeapTuple seqdatatuple)
+read_seq_tuple(Relation rel, Buffer *buf, HeapTuple seqdatatuple,
+ XLogRecPtr *lsn)
{
Page page;
ItemId lp;
@@ -1291,6 +1455,13 @@ read_seq_tuple(Relation rel, Buffer *buf, HeapTuple seqdatatuple)
elog(ERROR, "bad magic number in sequence \"%s\": %08X",
RelationGetRelationName(rel), sm->magic);
+ /*
+ * If the caller requested it, set the page LSN. This allows deciding
+ * which sequence changes are before/after the returned sequence state.
+ */
+ if (lsn)
+ *lsn = PageGetLSN(page);
+
lp = PageGetItemId(page, FirstOffsetNumber);
Assert(ItemIdIsNormal(lp));
@@ -1886,7 +2057,7 @@ pg_sequence_last_value(PG_FUNCTION_ARGS)
errmsg("permission denied for sequence %s",
RelationGetRelationName(seqrel))));
- seq = read_seq_tuple(seqrel, &buf, &seqtuple);
+ seq = read_seq_tuple(seqrel, &buf, &seqtuple, NULL);
is_called = seq->is_called;
result = seq->last_value;
@@ -1900,6 +2071,67 @@ pg_sequence_last_value(PG_FUNCTION_ARGS)
PG_RETURN_NULL();
}
+/*
+ * Return the current on-disk state of the sequence.
+ *
+ * Note: This is roughly equivalent to selecting the data from the sequence,
+ * except that it also returns the page LSN.
+ */
+Datum
+pg_sequence_state(PG_FUNCTION_ARGS)
+{
+ Oid relid = PG_GETARG_OID(0);
+ SeqTable elm;
+ Relation seqrel;
+ Buffer buf;
+ HeapTupleData seqtuple;
+ Form_pg_sequence_data seq;
+ Datum result;
+
+ int64 last_value;
+ int64 log_cnt;
+ bool is_called;
+ XLogRecPtr lsn;
+
+ TupleDesc tupdesc;
+ HeapTuple tuple;
+ Datum values[4];
+ bool nulls[4];
+
+ if (get_call_result_type(fcinfo, NULL, &tupdesc) != TYPEFUNC_COMPOSITE)
+ elog(ERROR, "return type must be a row type");
+
+ /* open and lock sequence */
+ init_sequence(relid, &elm, &seqrel);
+
+ if (pg_class_aclcheck(elm->relid, GetUserId(),
+ ACL_SELECT | ACL_USAGE) != ACLCHECK_OK)
+ ereport(ERROR,
+ (errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
+ errmsg("permission denied for sequence %s",
+ RelationGetRelationName(seqrel))));
+
+ seq = read_seq_tuple(seqrel, &buf, &seqtuple, &lsn);
+
+ is_called = seq->is_called;
+ last_value = seq->last_value;
+ log_cnt = seq->log_cnt;
+
+ UnlockReleaseBuffer(buf);
+ relation_close(seqrel, NoLock);
+
+ values[0] = LSNGetDatum(lsn);
+ values[1] = Int64GetDatum(last_value);
+ values[2] = Int64GetDatum(log_cnt);
+ values[3] = BoolGetDatum(is_called);
+
+ memset(nulls, 0, sizeof(nulls));
+
+ tuple = heap_form_tuple(tupdesc, values, nulls);
+ result = HeapTupleGetDatum(tuple);
+
+ PG_RETURN_DATUM(result);
+}
void
seq_redo(XLogReaderState *record)
diff --git a/src/backend/commands/subscriptioncmds.c b/src/backend/commands/subscriptioncmds.c
index 75e6cd8ae3c..144609c1b9f 100644
--- a/src/backend/commands/subscriptioncmds.c
+++ b/src/backend/commands/subscriptioncmds.c
@@ -100,6 +100,7 @@ typedef struct SubOpts
} SubOpts;
static List *fetch_table_list(WalReceiverConn *wrconn, List *publications);
+static List *fetch_sequence_list(WalReceiverConn *wrconn, List *publications);
static void check_publications_origin(WalReceiverConn *wrconn,
List *publications, bool copydata,
char *origin, Oid *subrel_local_oids,
@@ -730,9 +731,9 @@ CreateSubscription(ParseState *pstate, CreateSubscriptionStmt *stmt,
{
char *err;
WalReceiverConn *wrconn;
- List *tables;
+ List *relations;
ListCell *lc;
- char table_state;
+ char sync_state;
bool must_use_password;
/* Try to connect to the publisher. */
@@ -754,14 +755,17 @@ CreateSubscription(ParseState *pstate, CreateSubscriptionStmt *stmt,
* Set sync state based on if we were asked to do data copy or
* not.
*/
- table_state = opts.copy_data ? SUBREL_STATE_INIT : SUBREL_STATE_READY;
+ sync_state = opts.copy_data ? SUBREL_STATE_INIT : SUBREL_STATE_READY;
/*
- * Get the table list from publisher and build local table status
- * info.
+ * Get the table and sequence list from publisher and build local
+ * relation sync status info.
*/
- tables = fetch_table_list(wrconn, publications);
- foreach(lc, tables)
+ relations = fetch_table_list(wrconn, publications);
+ relations = list_concat(relations,
+ fetch_sequence_list(wrconn, publications));
+
+ foreach(lc, relations)
{
RangeVar *rv = (RangeVar *) lfirst(lc);
Oid relid;
@@ -772,7 +776,7 @@ CreateSubscription(ParseState *pstate, CreateSubscriptionStmt *stmt,
CheckSubscriptionRelkind(get_rel_relkind(relid),
rv->schemaname, rv->relname);
- AddSubscriptionRelState(subid, relid, table_state,
+ AddSubscriptionRelState(subid, relid, sync_state,
InvalidXLogRecPtr, true);
}
@@ -798,12 +802,12 @@ CreateSubscription(ParseState *pstate, CreateSubscriptionStmt *stmt,
*
* Note that if tables were specified but copy_data is false
* then it is safe to enable two_phase up-front because those
- * tables are already initially in READY state. When the
- * subscription has no tables, we leave the twophase state as
- * PENDING, to allow ALTER SUBSCRIPTION ... REFRESH
+ * relations are already initially in READY state. When the
+ * subscription has no relations, we leave the twophase state
+ * as PENDING, to allow ALTER SUBSCRIPTION ... REFRESH
* PUBLICATION to work.
*/
- if (opts.twophase && !opts.copy_data && tables != NIL)
+ if (opts.twophase && !opts.copy_data && relations != NIL)
twophase_enabled = true;
walrcv_create_slot(wrconn, opts.slot_name, false, twophase_enabled,
@@ -882,8 +886,10 @@ AlterSubscription_refresh(Subscription *sub, bool copy_data,
if (validate_publications)
check_publications(wrconn, validate_publications);
- /* Get the table list from publisher. */
+ /* Get the list of relations from publisher. */
pubrel_names = fetch_table_list(wrconn, sub->publications);
+ pubrel_names = list_concat(pubrel_names,
+ fetch_sequence_list(wrconn, sub->publications));
/* Get local table list. */
subrel_states = GetSubscriptionRelations(sub->oid, false);
@@ -2153,6 +2159,75 @@ fetch_table_list(WalReceiverConn *wrconn, List *publications)
return tablelist;
}
+/*
+ * Get the list of sequences which belong to specified publications on the
+ * publisher connection.
+ */
+static List *
+fetch_sequence_list(WalReceiverConn *wrconn, List *publications)
+{
+ WalRcvExecResult *res;
+ StringInfoData cmd;
+ TupleTableSlot *slot;
+ Oid tableRow[2] = {TEXTOID, TEXTOID};
+ ListCell *lc;
+ bool first;
+ List *tablelist = NIL;
+
+ Assert(list_length(publications) > 0);
+
+ initStringInfo(&cmd);
+ appendStringInfoString(&cmd, "SELECT DISTINCT s.schemaname, s.sequencename\n"
+ " FROM pg_catalog.pg_publication_sequences s\n"
+ " WHERE s.pubname IN (");
+ first = true;
+ foreach(lc, publications)
+ {
+ char *pubname = strVal(lfirst(lc));
+
+ if (first)
+ first = false;
+ else
+ appendStringInfoString(&cmd, ", ");
+
+ appendStringInfoString(&cmd, quote_literal_cstr(pubname));
+ }
+ appendStringInfoChar(&cmd, ')');
+
+ res = walrcv_exec(wrconn, cmd.data, 2, tableRow);
+ pfree(cmd.data);
+
+ if (res->status != WALRCV_OK_TUPLES)
+ ereport(ERROR,
+ (errmsg("could not receive list of replicated sequences from the publisher: %s",
+ res->err)));
+
+ /* Process sequences. */
+ slot = MakeSingleTupleTableSlot(res->tupledesc, &TTSOpsMinimalTuple);
+ while (tuplestore_gettupleslot(res->tuplestore, true, false, slot))
+ {
+ char *nspname;
+ char *relname;
+ bool isnull;
+ RangeVar *rv;
+
+ nspname = TextDatumGetCString(slot_getattr(slot, 1, &isnull));
+ Assert(!isnull);
+ relname = TextDatumGetCString(slot_getattr(slot, 2, &isnull));
+ Assert(!isnull);
+
+ rv = makeRangeVar(nspname, relname, -1);
+ tablelist = lappend(tablelist, rv);
+
+ ExecClearTuple(slot);
+ }
+ ExecDropSingleTupleTableSlot(slot);
+
+ walrcv_clear_result(res);
+
+ return tablelist;
+}
+
/*
* This is to report the connection failure while dropping replication slots.
* Here, we report the WARNING for all tablesync slots so that user can drop
diff --git a/src/backend/executor/execReplication.c b/src/backend/executor/execReplication.c
index ff8a7f77f5d..debe92f96b5 100644
--- a/src/backend/executor/execReplication.c
+++ b/src/backend/executor/execReplication.c
@@ -745,7 +745,9 @@ void
CheckSubscriptionRelkind(char relkind, const char *nspname,
const char *relname)
{
- if (relkind != RELKIND_RELATION && relkind != RELKIND_PARTITIONED_TABLE)
+ if (relkind != RELKIND_RELATION &&
+ relkind != RELKIND_PARTITIONED_TABLE &&
+ relkind != RELKIND_SEQUENCE)
ereport(ERROR,
(errcode(ERRCODE_WRONG_OBJECT_TYPE),
errmsg("cannot use relation \"%s.%s\" as logical replication target",
diff --git a/src/backend/parser/gram.y b/src/backend/parser/gram.y
index 6b88096e8e1..037bfc6b6f0 100644
--- a/src/backend/parser/gram.y
+++ b/src/backend/parser/gram.y
@@ -452,7 +452,7 @@ static Node *makeRecursiveViewSelect(char *relname, List *aliases, Node *query);
transform_element_list transform_type_list
TriggerTransitions TriggerReferencing
vacuum_relation_list opt_vacuum_relation_list
- drop_option_list pub_obj_list
+ drop_option_list pub_obj_list pub_obj_type_list
%type <node> opt_routine_body
%type <groupclause> group_clause
@@ -586,6 +586,7 @@ static Node *makeRecursiveViewSelect(char *relname, List *aliases, Node *query);
%type <node> var_value zone_value
%type <rolespec> auth_ident RoleSpec opt_granted_by
%type <publicationobjectspec> PublicationObjSpec
+%type <node> pub_obj_type
%type <keyword> unreserved_keyword type_func_name_keyword
%type <keyword> col_name_keyword reserved_keyword
@@ -10437,12 +10438,16 @@ AlterOwnerStmt: ALTER AGGREGATE aggregate_with_argtypes OWNER TO RoleSpec
*
* CREATE PUBLICATION FOR ALL TABLES [WITH options]
*
+ * CREATE PUBLICATION FOR ALL SEQUENCES [WITH options]
+ *
* CREATE PUBLICATION FOR pub_obj [, ...] [WITH options]
*
* pub_obj is one of:
*
* TABLE table [, ...]
+ * SEQUENCE table [, ...]
* TABLES IN SCHEMA schema [, ...]
+ * SEQUENCES IN SCHEMA schema [, ...]
*
*****************************************************************************/
@@ -10455,13 +10460,13 @@ CreatePublicationStmt:
n->options = $4;
$$ = (Node *) n;
}
- | CREATE PUBLICATION name FOR ALL TABLES opt_definition
+ | CREATE PUBLICATION name FOR ALL pub_obj_type_list opt_definition
{
CreatePublicationStmt *n = makeNode(CreatePublicationStmt);
n->pubname = $3;
n->options = $7;
- n->for_all_tables = true;
+ n->for_all_objects = $6;
$$ = (Node *) n;
}
| CREATE PUBLICATION name FOR pub_obj_list opt_definition
@@ -10512,6 +10517,26 @@ PublicationObjSpec:
$$->pubobjtype = PUBLICATIONOBJ_TABLES_IN_CUR_SCHEMA;
$$->location = @4;
}
+ | SEQUENCE relation_expr
+ {
+ $$ = makeNode(PublicationObjSpec);
+ $$->pubobjtype = PUBLICATIONOBJ_SEQUENCE;
+ $$->pubtable = makeNode(PublicationTable);
+ $$->pubtable->relation = $2;
+ }
+ | SEQUENCES IN_P SCHEMA ColId
+ {
+ $$ = makeNode(PublicationObjSpec);
+ $$->pubobjtype = PUBLICATIONOBJ_SEQUENCES_IN_SCHEMA;
+ $$->name = $4;
+ $$->location = @4;
+ }
+ | SEQUENCES IN_P SCHEMA CURRENT_SCHEMA
+ {
+ $$ = makeNode(PublicationObjSpec);
+ $$->pubobjtype = PUBLICATIONOBJ_SEQUENCES_IN_CUR_SCHEMA;
+ $$->location = @4;
+ }
| ColId opt_column_list OptWhereClause
{
$$ = makeNode(PublicationObjSpec);
@@ -10573,6 +10598,19 @@ pub_obj_list: PublicationObjSpec
{ $$ = lappend($1, $3); }
;
+pub_obj_type: TABLES
+ { $$ = (Node *) makeString("tables"); }
+ | SEQUENCES
+ { $$ = (Node *) makeString("sequences"); }
+ ;
+
+pub_obj_type_list: pub_obj_type
+ { $$ = list_make1($1); }
+ | pub_obj_type_list ',' pub_obj_type
+ { $$ = lappend($1, $3); }
+ ;
+
+
/*****************************************************************************
*
* ALTER PUBLICATION name SET ( options )
@@ -10586,7 +10624,9 @@ pub_obj_list: PublicationObjSpec
* pub_obj is one of:
*
* TABLE table_name [, ...]
+ * SEQUENCE table_name [, ...]
* TABLES IN SCHEMA schema_name [, ...]
+ * SEQUENCES IN SCHEMA schema_name [, ...]
*
*****************************************************************************/
@@ -18963,7 +19003,8 @@ preprocess_pubobj_list(List *pubobjspec_list, core_yyscan_t yyscanner)
if (pubobj->pubobjtype == PUBLICATIONOBJ_CONTINUATION)
pubobj->pubobjtype = prevobjtype;
- if (pubobj->pubobjtype == PUBLICATIONOBJ_TABLE)
+ if (pubobj->pubobjtype == PUBLICATIONOBJ_TABLE ||
+ pubobj->pubobjtype == PUBLICATIONOBJ_SEQUENCE)
{
/* relation name or pubtable must be set for this type of object */
if (!pubobj->name && !pubobj->pubtable)
@@ -19014,6 +19055,30 @@ preprocess_pubobj_list(List *pubobjspec_list, core_yyscan_t yyscanner)
errmsg("invalid schema name"),
parser_errposition(pubobj->location));
}
+ else if (pubobj->pubobjtype == PUBLICATIONOBJ_SEQUENCES_IN_SCHEMA ||
+ pubobj->pubobjtype == PUBLICATIONOBJ_SEQUENCES_IN_CUR_SCHEMA)
+ {
+ /* WHERE clause is not allowed on a schema object */
+ if (pubobj->pubtable && pubobj->pubtable->whereClause)
+ ereport(ERROR,
+ errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("WHERE clause not allowed for schema"),
+ parser_errposition(pubobj->location));
+
+ /*
+ * We can distinguish between the different type of schema
+ * objects based on whether name and pubtable is set.
+ */
+ if (pubobj->name)
+ pubobj->pubobjtype = PUBLICATIONOBJ_SEQUENCES_IN_SCHEMA;
+ else if (!pubobj->name && !pubobj->pubtable)
+ pubobj->pubobjtype = PUBLICATIONOBJ_SEQUENCES_IN_CUR_SCHEMA;
+ else
+ ereport(ERROR,
+ errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("invalid schema name at or near"),
+ parser_errposition(pubobj->location));
+ }
prevobjtype = pubobj->pubobjtype;
}
diff --git a/src/backend/replication/logical/proto.c b/src/backend/replication/logical/proto.c
index f273005c15d..a98296f667d 100644
--- a/src/backend/replication/logical/proto.c
+++ b/src/backend/replication/logical/proto.c
@@ -663,6 +663,59 @@ logicalrep_write_message(StringInfo out, TransactionId xid, XLogRecPtr lsn,
pq_sendbytes(out, message, sz);
}
+/*
+ * Write SEQUENCE to stream
+ */
+void
+logicalrep_write_sequence(StringInfo out, Relation rel, TransactionId xid,
+ XLogRecPtr lsn, bool transactional,
+ int64 value)
+{
+ uint8 flags = 0;
+
+ pq_sendbyte(out, LOGICAL_REP_MSG_SEQUENCE);
+
+ /* transaction ID (if not valid, we're not streaming) */
+ if (TransactionIdIsValid(xid))
+ pq_sendint32(out, xid);
+
+ pq_sendint8(out, flags);
+
+ /* use Oid as relation identifier */
+ pq_sendint32(out, RelationGetRelid(rel));
+
+ /* write the LSN of the sequence change */
+ pq_sendint64(out, lsn);
+
+ /* write sequence info */
+ pq_sendint8(out, transactional);
+ pq_sendint64(out, value);
+}
+
+/*
+ * Read SEQUENCE from the stream.
+ */
+LogicalRepRelId
+logicalrep_read_sequence(StringInfo in, LogicalRepSequence *seqdata)
+{
+ LogicalRepRelId relid;
+
+ /* XXX skipping flags */
+ pq_getmsgint(in, 1);
+
+ /* read the relation id */
+ relid = pq_getmsgint(in, 4);
+
+ /* read the LSN of the sequence change */
+ seqdata->lsn = pq_getmsgint64(in);
+
+ /* info about the sequence change */
+ seqdata->transactional = pq_getmsgint(in, 1);
+ seqdata->value = pq_getmsgint64(in);
+
+ return relid;
+}
+
/*
* Write relation description to the output stream.
*/
@@ -1258,6 +1311,8 @@ logicalrep_message_type(LogicalRepMsgType action)
return "STREAM ABORT";
case LOGICAL_REP_MSG_STREAM_PREPARE:
return "STREAM PREPARE";
+ case LOGICAL_REP_MSG_SEQUENCE:
+ return "SEQUENCE";
}
/*
diff --git a/src/backend/replication/logical/tablesync.c b/src/backend/replication/logical/tablesync.c
index 06d5b3df33a..47590e2d59e 100644
--- a/src/backend/replication/logical/tablesync.c
+++ b/src/backend/replication/logical/tablesync.c
@@ -100,6 +100,7 @@
#include "catalog/pg_subscription_rel.h"
#include "catalog/pg_type.h"
#include "commands/copy.h"
+#include "commands/sequence.h"
#include "miscadmin.h"
#include "nodes/makefuncs.h"
#include "parser/parse_relation.h"
@@ -118,6 +119,7 @@
#include "utils/builtins.h"
#include "utils/lsyscache.h"
#include "utils/memutils.h"
+#include "utils/pg_lsn.h"
#include "utils/rls.h"
#include "utils/snapmgr.h"
#include "utils/syscache.h"
@@ -1239,6 +1241,106 @@ copy_table(Relation rel)
logicalrep_rel_close(relmapentry, NoLock);
}
+/*
+ * Fetch sequence data (current state) from the remote node, including the
+ * page LSN.
+ */
+static int64
+fetch_sequence_data(Oid remoteid, XLogRecPtr *lsn)
+{
+ WalRcvExecResult *res;
+ StringInfoData cmd;
+ TupleTableSlot *slot;
+ Oid tableRow[2] = {INT8OID, LSNOID};
+ int64 value = (Datum) 0;
+
+ initStringInfo(&cmd);
+ appendStringInfo(&cmd, "SELECT (last_value + log_cnt), page_lsn "
+ "FROM pg_sequence_state(%d)", remoteid);
+
+ res = walrcv_exec(LogRepWorkerWalRcvConn, cmd.data, 2, tableRow);
+ pfree(cmd.data);
+
+ if (res->status != WALRCV_OK_TUPLES)
+ ereport(ERROR,
+ (errmsg("could not receive list of replicated tables from the publisher: %s",
+ res->err)));
+
+ /* Process the sequence. */
+ slot = MakeSingleTupleTableSlot(res->tupledesc, &TTSOpsMinimalTuple);
+ while (tuplestore_gettupleslot(res->tuplestore, true, false, slot))
+ {
+ bool isnull;
+
+ value = DatumGetInt64(slot_getattr(slot, 1, &isnull));
+ Assert(!isnull);
+
+ *lsn = DatumGetInt64(slot_getattr(slot, 2, &isnull));
+ Assert(!isnull);
+ }
+
+ ExecDropSingleTupleTableSlot(slot);
+
+ walrcv_clear_result(res);
+
+ return value;
+}
+
+/*
+ * Copy existing data of a sequence from publisher.
+ *
+ * Caller is responsible for locking the local relation.
+ */
+static XLogRecPtr
+copy_sequence(Relation rel)
+{
+ LogicalRepRelMapEntry *relmapentry;
+ LogicalRepRelation lrel;
+ List *qual = NIL;
+ StringInfoData cmd;
+ int64 sequence_value;
+ XLogRecPtr lsn = InvalidXLogRecPtr;
+
+ /* Get the publisher relation info. */
+ fetch_remote_table_info(get_namespace_name(RelationGetNamespace(rel)),
+ RelationGetRelationName(rel), &lrel, &qual);
+
+ /* sequences don't have row filters */
+ Assert(!qual);
+
+ /* Put the relation into relmap. */
+ logicalrep_relmap_update(&lrel);
+
+ /* Map the publisher relation to local one. */
+ relmapentry = logicalrep_rel_open(lrel.remoteid, NoLock);
+ Assert(rel == relmapentry->localrel);
+
+ /* Start copy on the publisher. */
+ initStringInfo(&cmd);
+
+ Assert(lrel.relkind == RELKIND_SEQUENCE);
+
+ /*
+ * Logical replication of sequences is based on decoding WAL records,
+ * describing the "next" state of the sequence the current state in the
+ * relfilenode is yet to reach. But during the initial sync we read the
+ * current state, so we need to reconstruct the WAL record logged when we
+ * started the current batch of sequence values.
+ *
+ * Otherwise we might get duplicate values (on subscriber) if we failed
+ * over right after the sync.
+ */
+ sequence_value = fetch_sequence_data(lrel.remoteid, &lsn);
+
+ /* tablesync sets the sequences in non-transactional way */
+ SetSequence(RelationGetRelid(rel), false, sequence_value);
+
+ logicalrep_rel_close(relmapentry, NoLock);
+
+ /* return the LSN when the sequence state was set */
+ return lsn;
+}
+
/*
* Determine the tablesync slot name.
*
@@ -1287,6 +1389,7 @@ LogicalRepSyncTableStart(XLogRecPtr *origin_startpos)
UserContext ucxt;
bool must_use_password;
bool run_as_owner;
+ XLogRecPtr sequence_lsn = InvalidXLogRecPtr;
/* Check the state of the table synchronization. */
StartTransactionCommand();
@@ -1498,10 +1601,34 @@ LogicalRepSyncTableStart(XLogRecPtr *origin_startpos)
GetUserNameFromId(GetUserId(), true),
RelationGetRelationName(rel))));
- /* Now do the initial data copy */
- PushActiveSnapshot(GetTransactionSnapshot());
- copy_table(rel);
- PopActiveSnapshot();
+ /* Do the right action depending on the relation kind. */
+ if (get_rel_relkind(RelationGetRelid(rel)) == RELKIND_SEQUENCE)
+ {
+ /* Now do the initial sequence copy */
+ PushActiveSnapshot(GetTransactionSnapshot());
+ sequence_lsn = copy_sequence(rel);
+ PopActiveSnapshot();
+
+ /*
+ * Sequences are not consistent (in the MVCC sense) with respect to
+ * the replication slot, so the copy might have read a more recent
+ * state than origin_startpos. The sequence_lsn comes from page LSN
+ * (which is LSN of the last sequence change), so that's the right
+ * position where to start with the catchup apply.
+ *
+ * It might be before the slot, though (if the sequence was not used
+ * since between the slot creation and copy), so make sure the
+ * position does not move backwards.
+ */
+ *origin_startpos = Max(*origin_startpos, sequence_lsn);
+ }
+ else
+ {
+ /* Now do the initial data copy */
+ PushActiveSnapshot(GetTransactionSnapshot());
+ copy_table(rel);
+ PopActiveSnapshot();
+ }
res = walrcv_exec(LogRepWorkerWalRcvConn, "COMMIT", 0, NULL);
if (res->status != WALRCV_OK_COMMAND)
diff --git a/src/backend/replication/logical/worker.c b/src/backend/replication/logical/worker.c
index 911835c5cb3..01fe9527b6b 100644
--- a/src/backend/replication/logical/worker.c
+++ b/src/backend/replication/logical/worker.c
@@ -154,6 +154,7 @@
#include "catalog/pg_subscription.h"
#include "catalog/pg_subscription_rel.h"
#include "catalog/pg_tablespace.h"
+#include "commands/sequence.h"
#include "commands/tablecmds.h"
#include "commands/tablespace.h"
#include "commands/trigger.h"
@@ -1437,6 +1438,97 @@ apply_handle_origin(StringInfo s)
errmsg_internal("ORIGIN message sent out of order")));
}
+/*
+ * Handle SEQUENCE message.
+ */
+static void
+apply_handle_sequence(StringInfo s)
+{
+ LogicalRepSequence seq;
+ LogicalRepRelMapEntry *rel;
+ Oid relid;
+ bool already_in_transaction PG_USED_FOR_ASSERTS_ONLY;
+
+ if (is_skipping_changes() ||
+ handle_streamed_transaction(LOGICAL_REP_MSG_SEQUENCE, s))
+ return;
+
+ /*
+ * Remember if we're already in transaction (begin_replication step starts
+ * a transaction, so we can't use IsTransactionState() after that point.
+ */
+ already_in_transaction = IsTransactionState();
+
+ /*
+ * Make sure we're in a transaction (needed by SetSequence). For
+ * non-transactional updates we're guaranteed to start a new one, and
+ * we'll commit it at the end.
+ */
+ begin_replication_step();
+
+ relid = logicalrep_read_sequence(s, &seq);
+
+ /*
+ * Non-transactional sequence updates should not be part of a remote
+ * transaction. There should not be any running transaction.
+ */
+ Assert((!seq.transactional) || in_remote_transaction);
+ Assert(!(!seq.transactional && in_remote_transaction));
+ Assert(!(!seq.transactional && already_in_transaction));
+
+ /* lock the sequence in AccessExclusiveLock, as expected by SetSequence */
+ rel = logicalrep_rel_open(relid, AccessExclusiveLock);
+ if (!should_apply_changes_for_rel(rel))
+ {
+ /*
+ * The relation can't become interesting in the middle of the
+ * transaction so it's safe to unlock it.
+ */
+ logicalrep_rel_close(rel, AccessExclusiveLock);
+ end_replication_step();
+
+ /*
+ * Commit the per-stream transaction (we only do this when not in
+ * remote transaction, i.e. for non-transactional sequence updates.)
+ */
+ if (!in_remote_transaction)
+ CommitTransactionCommand();
+
+ return;
+ }
+
+ /*
+ * apply the sequence change
+ *
+ * XXX We don't need to reconstruct the value from last_value, log_cnt
+ * etc. because that happens on the publisher.
+ */
+ SetSequence(rel->localreloid, seq.transactional, seq.value);
+
+ logicalrep_rel_close(rel, NoLock);
+
+ end_replication_step();
+
+ /*
+ * Commit the per-stream transaction (we only do this when not in remote
+ * transaction, i.e. for non-transactional sequence updates.)
+ */
+ if (!in_remote_transaction)
+ {
+ CommitTransactionCommand();
+
+ /*
+ * Update origin state so we don't try applying this sequence change
+ * in case of crash. We only do this for non-trasactional changes,
+ * that are replayed directly.
+ *
+ * XXX We don't have replorigin_session_origin_timestamp, but we can
+ * just leave that set to 0.
+ */
+ replorigin_session_origin_lsn = seq.lsn;
+ }
+}
+
/*
* Initialize fileset (if not already done).
*
@@ -3347,6 +3439,10 @@ apply_dispatch(StringInfo s)
*/
break;
+ case LOGICAL_REP_MSG_SEQUENCE:
+ apply_handle_sequence(s);
+ return;
+
case LOGICAL_REP_MSG_STREAM_START:
apply_handle_stream_start(s);
break;
@@ -4365,6 +4461,7 @@ set_stream_options(WalRcvStreamOptions *options,
server_version = walrcv_server_version(LogRepWorkerWalRcvConn);
options->proto.logical.proto_version =
+ server_version >= 170000 ? LOGICALREP_PROTO_SEQUENCES_VERSION_NUM :
server_version >= 160000 ? LOGICALREP_PROTO_STREAM_PARALLEL_VERSION_NUM :
server_version >= 150000 ? LOGICALREP_PROTO_TWOPHASE_VERSION_NUM :
server_version >= 140000 ? LOGICALREP_PROTO_STREAM_VERSION_NUM :
diff --git a/src/backend/replication/pgoutput/pgoutput.c b/src/backend/replication/pgoutput/pgoutput.c
index 425238187f6..8b8f81c5f8e 100644
--- a/src/backend/replication/pgoutput/pgoutput.c
+++ b/src/backend/replication/pgoutput/pgoutput.c
@@ -15,6 +15,7 @@
#include "access/tupconvert.h"
#include "catalog/partition.h"
#include "catalog/pg_publication.h"
+#include "catalog/pg_publication_namespace.h"
#include "catalog/pg_publication_rel.h"
#include "catalog/pg_subscription.h"
#include "commands/defrem.h"
@@ -55,6 +56,10 @@ static void pgoutput_message(LogicalDecodingContext *ctx,
ReorderBufferTXN *txn, XLogRecPtr message_lsn,
bool transactional, const char *prefix,
Size sz, const char *message);
+static void pgoutput_sequence(LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn, XLogRecPtr sequence_lsn,
+ Relation relation, bool transactional,
+ int64 value);
static bool pgoutput_origin_filter(LogicalDecodingContext *ctx,
RepOriginId origin_id);
static void pgoutput_begin_prepare_txn(LogicalDecodingContext *ctx,
@@ -252,6 +257,7 @@ _PG_output_plugin_init(OutputPluginCallbacks *cb)
cb->change_cb = pgoutput_change;
cb->truncate_cb = pgoutput_truncate;
cb->message_cb = pgoutput_message;
+ cb->sequence_cb = pgoutput_sequence;
cb->commit_cb = pgoutput_commit_txn;
cb->begin_prepare_cb = pgoutput_begin_prepare_txn;
@@ -268,6 +274,7 @@ _PG_output_plugin_init(OutputPluginCallbacks *cb)
cb->stream_commit_cb = pgoutput_stream_commit;
cb->stream_change_cb = pgoutput_change;
cb->stream_message_cb = pgoutput_message;
+ cb->stream_sequence_cb = pgoutput_sequence;
cb->stream_truncate_cb = pgoutput_truncate;
/* transaction streaming - two-phase commit */
cb->stream_prepare_cb = pgoutput_stream_prepare_txn;
@@ -678,7 +685,7 @@ pgoutput_rollback_prepared_txn(LogicalDecodingContext *ctx,
*/
static void
maybe_send_schema(LogicalDecodingContext *ctx,
- ReorderBufferChange *change,
+ ReorderBufferTXN *txn,
Relation relation, RelationSyncEntry *relentry)
{
PGOutputData *data = (PGOutputData *) ctx->output_plugin_private;
@@ -695,10 +702,10 @@ maybe_send_schema(LogicalDecodingContext *ctx,
* the write methods will not include it.
*/
if (data->in_streaming)
- xid = change->txn->xid;
+ xid = txn->xid;
- if (rbtxn_is_subtxn(change->txn))
- topxid = rbtxn_get_toptxn(change->txn)->xid;
+ if (rbtxn_is_subtxn(txn))
+ topxid = rbtxn_get_toptxn(txn)->xid;
else
topxid = xid;
@@ -907,9 +914,10 @@ pgoutput_row_filter_init(PGOutputData *data, List *publications,
* (even if other publications have a row filter).
*/
if (!pub->alltables &&
- !SearchSysCacheExists2(PUBLICATIONNAMESPACEMAP,
+ !SearchSysCacheExists3(PUBLICATIONNAMESPACEMAP,
ObjectIdGetDatum(schemaid),
- ObjectIdGetDatum(pub->oid)))
+ ObjectIdGetDatum(pub->oid),
+ PUB_OBJTYPE_TABLE))
{
/*
* Check for the presence of a row filter in this publication.
@@ -1523,7 +1531,7 @@ pgoutput_change(LogicalDecodingContext *ctx, ReorderBufferTXN *txn,
* Schema should be sent using the original relation because it also sends
* the ancestor's relation.
*/
- maybe_send_schema(ctx, change, relation, relentry);
+ maybe_send_schema(ctx, txn, relation, relentry);
OutputPluginPrepareWrite(ctx, true);
@@ -1608,7 +1616,7 @@ pgoutput_truncate(LogicalDecodingContext *ctx, ReorderBufferTXN *txn,
if (txndata && !txndata->sent_begin_txn)
pgoutput_send_begin(ctx, txn);
- maybe_send_schema(ctx, change, relation, relentry);
+ maybe_send_schema(ctx, change->txn, relation, relentry);
}
if (nrelids > 0)
@@ -1668,6 +1676,70 @@ pgoutput_message(LogicalDecodingContext *ctx, ReorderBufferTXN *txn,
OutputPluginWrite(ctx, true);
}
+static void
+pgoutput_sequence(LogicalDecodingContext *ctx,
+ ReorderBufferTXN *txn, XLogRecPtr sequence_lsn,
+ Relation relation, bool transactional,
+ int64 value)
+{
+ PGOutputData *data = (PGOutputData *) ctx->output_plugin_private;
+ TransactionId xid = InvalidTransactionId;
+ RelationSyncEntry *relentry;
+
+ if (!is_publishable_relation(relation))
+ return;
+
+ /*
+ * If the negotiated protocol version does not support sequences, yet we
+ * found a sequence in the publication, error out.
+ */
+ if (data->protocol_version < LOGICALREP_PROTO_SEQUENCES_VERSION_NUM)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("protocol version does not support sequence replication")));
+
+ /*
+ * Remember the xid for the message in streaming mode. See
+ * pgoutput_change.
+ */
+ if (data->in_streaming)
+ xid = txn->xid;
+
+ relentry = get_rel_sync_entry(data, relation);
+
+ /*
+ * First check the sequence filter.
+ *
+ * We handle just REORDER_BUFFER_CHANGE_SEQUENCE here.
+ */
+ if (!relentry->pubactions.pubsequence)
+ return;
+
+ /*
+ * Output BEGIN if we haven't yet. Avoid for non-transactional sequence
+ * changes.
+ */
+ if (transactional)
+ {
+ PGOutputTxnData *txndata = (PGOutputTxnData *) txn->output_plugin_private;
+
+ /* Send BEGIN if we haven't yet */
+ if (txndata && !txndata->sent_begin_txn)
+ pgoutput_send_begin(ctx, txn);
+ }
+
+ maybe_send_schema(ctx, txn, relation, relentry);
+
+ OutputPluginPrepareWrite(ctx, true);
+ logicalrep_write_sequence(ctx->out,
+ relation,
+ xid,
+ sequence_lsn,
+ transactional,
+ value);
+ OutputPluginWrite(ctx, true);
+}
+
/*
* Return true if the data is associated with an origin and the user has
* requested the changes that don't have an origin, false otherwise.
@@ -1983,7 +2055,8 @@ get_rel_sync_entry(PGOutputData *data, Relation relation)
entry->schema_sent = false;
entry->streamed_txns = NIL;
entry->pubactions.pubinsert = entry->pubactions.pubupdate =
- entry->pubactions.pubdelete = entry->pubactions.pubtruncate = false;
+ entry->pubactions.pubdelete = entry->pubactions.pubtruncate =
+ entry->pubactions.pubsequence = false;
entry->new_slot = NULL;
entry->old_slot = NULL;
memset(entry->exprstate, 0, sizeof(entry->exprstate));
@@ -1998,18 +2071,19 @@ get_rel_sync_entry(PGOutputData *data, Relation relation)
{
Oid schemaId = get_rel_namespace(relid);
List *pubids = GetRelationPublications(relid);
+ char relkind = get_rel_relkind(relid);
+ char objectType = pub_get_object_type_for_relkind(relkind);
/*
* We don't acquire a lock on the namespace system table as we build
* the cache entry using a historic snapshot and all the later changes
* are absorbed while decoding WAL.
*/
- List *schemaPubids = GetSchemaPublications(schemaId);
+ List *schemaPubids = GetSchemaPublications(schemaId, objectType);
ListCell *lc;
Oid publish_as_relid = relid;
int publish_ancestor_level = 0;
bool am_partition = get_rel_relispartition(relid);
- char relkind = get_rel_relkind(relid);
List *rel_publications = NIL;
/* Reload publications if needed before use. */
@@ -2041,6 +2115,7 @@ get_rel_sync_entry(PGOutputData *data, Relation relation)
entry->pubactions.pubupdate = false;
entry->pubactions.pubdelete = false;
entry->pubactions.pubtruncate = false;
+ entry->pubactions.pubsequence = false;
/*
* Tuple slots cleanups. (Will be rebuilt later if needed).
@@ -2088,9 +2163,11 @@ get_rel_sync_entry(PGOutputData *data, Relation relation)
/*
* If this is a FOR ALL TABLES publication, pick the partition
- * root and set the ancestor level accordingly.
+ * root and set the ancestor level accordingly. If this is a FOR
+ * ALL SEQUENCES publication, we publish it too but we don't need
+ * to pick the partition root etc.
*/
- if (pub->alltables)
+ if (pub->alltables && (objectType == PUB_OBJTYPE_TABLE))
{
publish = true;
if (pub->pubviaroot && am_partition)
@@ -2101,6 +2178,10 @@ get_rel_sync_entry(PGOutputData *data, Relation relation)
ancestor_level = list_length(ancestors);
}
}
+ else if (pub->allsequences && (objectType == PUB_OBJTYPE_SEQUENCE))
+ {
+ publish = true;
+ }
if (!publish)
{
@@ -2154,6 +2235,7 @@ get_rel_sync_entry(PGOutputData *data, Relation relation)
entry->pubactions.pubupdate |= pub->pubactions.pubupdate;
entry->pubactions.pubdelete |= pub->pubactions.pubdelete;
entry->pubactions.pubtruncate |= pub->pubactions.pubtruncate;
+ entry->pubactions.pubsequence |= pub->pubactions.pubsequence;
/*
* We want to publish the changes as the top-most ancestor
diff --git a/src/backend/utils/cache/relcache.c b/src/backend/utils/cache/relcache.c
index 20273f0be16..1bc77b47600 100644
--- a/src/backend/utils/cache/relcache.c
+++ b/src/backend/utils/cache/relcache.c
@@ -57,6 +57,7 @@
#include "catalog/pg_opclass.h"
#include "catalog/pg_proc.h"
#include "catalog/pg_publication.h"
+#include "catalog/pg_publication_namespace.h"
#include "catalog/pg_rewrite.h"
#include "catalog/pg_shseclabel.h"
#include "catalog/pg_statistic_ext.h"
@@ -5715,6 +5716,8 @@ RelationBuildPublicationDesc(Relation relation, PublicationDesc *pubdesc)
Oid schemaid;
List *ancestors = NIL;
Oid relid = RelationGetRelid(relation);
+ char relkind = relation->rd_rel->relkind;
+ char objType;
/*
* If not publishable, it publishes no actions. (pgoutput_change() will
@@ -5745,8 +5748,15 @@ RelationBuildPublicationDesc(Relation relation, PublicationDesc *pubdesc)
/* Fetch the publication membership info. */
puboids = GetRelationPublications(relid);
schemaid = RelationGetNamespace(relation);
- puboids = list_concat_unique_oid(puboids, GetSchemaPublications(schemaid));
+ objType = pub_get_object_type_for_relkind(relkind);
+ puboids = list_concat_unique_oid(puboids,
+ GetSchemaPublications(schemaid, objType));
+
+ /*
+ * If this is a partion (and thus a table), lookup all ancestors and track
+ * all publications them too.
+ */
if (relation->rd_rel->relispartition)
{
/* Add publications that the ancestors are in too. */
@@ -5758,12 +5768,23 @@ RelationBuildPublicationDesc(Relation relation, PublicationDesc *pubdesc)
puboids = list_concat_unique_oid(puboids,
GetRelationPublications(ancestor));
+
+ /* include all publications publishing schema of all ancestors */
schemaid = get_rel_namespace(ancestor);
puboids = list_concat_unique_oid(puboids,
- GetSchemaPublications(schemaid));
+ GetSchemaPublications(schemaid,
+ PUB_OBJTYPE_TABLE));
}
}
- puboids = list_concat_unique_oid(puboids, GetAllTablesPublications());
+
+ /*
+ * Consider also FOR ALL TABLES and FOR ALL SEQUENCES publications,
+ * depending on the relkind of the relation.
+ */
+ if (relation->rd_rel->relkind == RELKIND_SEQUENCE)
+ puboids = list_concat_unique_oid(puboids, GetAllSequencesPublications());
+ else
+ puboids = list_concat_unique_oid(puboids, GetAllTablesPublications());
foreach(lc, puboids)
{
@@ -5782,6 +5803,7 @@ RelationBuildPublicationDesc(Relation relation, PublicationDesc *pubdesc)
pubdesc->pubactions.pubupdate |= pubform->pubupdate;
pubdesc->pubactions.pubdelete |= pubform->pubdelete;
pubdesc->pubactions.pubtruncate |= pubform->pubtruncate;
+ pubdesc->pubactions.pubsequence |= pubform->pubsequence;
/*
* Check if all columns referenced in the filter expression are part
diff --git a/src/backend/utils/cache/syscache.c b/src/backend/utils/cache/syscache.c
index c10a7b3d0d2..ab3f3f191c2 100644
--- a/src/backend/utils/cache/syscache.c
+++ b/src/backend/utils/cache/syscache.c
@@ -450,9 +450,10 @@ static const struct cachedesc cacheinfo[] = {
},
[PUBLICATIONNAMESPACEMAP] = {
PublicationNamespaceRelationId,
- PublicationNamespacePnnspidPnpubidIndexId,
+ PublicationNamespacePnnspidPnpubidPntypeIndexId,
KEY(Anum_pg_publication_namespace_pnnspid,
- Anum_pg_publication_namespace_pnpubid),
+ Anum_pg_publication_namespace_pnpubid,
+ Anum_pg_publication_namespace_pntype),
64
},
[PUBLICATIONOID] = {
diff --git a/src/bin/pg_dump/pg_dump.c b/src/bin/pg_dump/pg_dump.c
index 22d1e6cf922..808ff4befc6 100644
--- a/src/bin/pg_dump/pg_dump.c
+++ b/src/bin/pg_dump/pg_dump.c
@@ -4088,10 +4088,12 @@ getPublications(Archive *fout, int *numPublications)
int i_pubname;
int i_pubowner;
int i_puballtables;
+ int i_puballsequences;
int i_pubinsert;
int i_pubupdate;
int i_pubdelete;
int i_pubtruncate;
+ int i_pubsequence;
int i_pubviaroot;
int i,
ntups;
@@ -4107,23 +4109,29 @@ getPublications(Archive *fout, int *numPublications)
resetPQExpBuffer(query);
/* Get the publications. */
- if (fout->remoteVersion >= 130000)
+ if (fout->remoteVersion >= 170000)
+ appendPQExpBufferStr(query,
+ "SELECT p.tableoid, p.oid, p.pubname, "
+ "p.pubowner, "
+ "p.puballtables, p.puballsequences, p.pubinsert, p.pubupdate, p.pubdelete, p.pubtruncate, p.pubsequence, p.pubviaroot "
+ "FROM pg_publication p");
+ else if (fout->remoteVersion >= 130000)
appendPQExpBufferStr(query,
"SELECT p.tableoid, p.oid, p.pubname, "
"p.pubowner, "
- "p.puballtables, p.pubinsert, p.pubupdate, p.pubdelete, p.pubtruncate, p.pubviaroot "
+ "p.puballtables, false as p.puballsequences, p.pubinsert, p.pubupdate, p.pubdelete, p.pubtruncate, false as p.pubsequence, p.pubviaroot "
"FROM pg_publication p");
else if (fout->remoteVersion >= 110000)
appendPQExpBufferStr(query,
"SELECT p.tableoid, p.oid, p.pubname, "
"p.pubowner, "
- "p.puballtables, p.pubinsert, p.pubupdate, p.pubdelete, p.pubtruncate, false AS pubviaroot "
+ "p.puballtables, false as p.puballsequences, p.pubinsert, p.pubupdate, p.pubdelete, p.pubtruncate, false AS pubsequence, false AS pubviaroot "
"FROM pg_publication p");
else
appendPQExpBufferStr(query,
"SELECT p.tableoid, p.oid, p.pubname, "
"p.pubowner, "
- "p.puballtables, p.pubinsert, p.pubupdate, p.pubdelete, false AS pubtruncate, false AS pubviaroot "
+ "p.puballtables, false as p.puballsequences, p.pubinsert, p.pubupdate, p.pubdelete, false AS pubtruncate, false AS pubsequence, false AS pubviaroot "
"FROM pg_publication p");
res = ExecuteSqlQuery(fout, query->data, PGRES_TUPLES_OK);
@@ -4135,10 +4143,12 @@ getPublications(Archive *fout, int *numPublications)
i_pubname = PQfnumber(res, "pubname");
i_pubowner = PQfnumber(res, "pubowner");
i_puballtables = PQfnumber(res, "puballtables");
+ i_puballsequences = PQfnumber(res, "puballsequences");
i_pubinsert = PQfnumber(res, "pubinsert");
i_pubupdate = PQfnumber(res, "pubupdate");
i_pubdelete = PQfnumber(res, "pubdelete");
i_pubtruncate = PQfnumber(res, "pubtruncate");
+ i_pubsequence = PQfnumber(res, "pubsequence");
i_pubviaroot = PQfnumber(res, "pubviaroot");
pubinfo = pg_malloc(ntups * sizeof(PublicationInfo));
@@ -4154,6 +4164,8 @@ getPublications(Archive *fout, int *numPublications)
pubinfo[i].rolname = getRoleName(PQgetvalue(res, i, i_pubowner));
pubinfo[i].puballtables =
(strcmp(PQgetvalue(res, i, i_puballtables), "t") == 0);
+ pubinfo[i].puballsequences =
+ (strcmp(PQgetvalue(res, i, i_puballsequences), "t") == 0);
pubinfo[i].pubinsert =
(strcmp(PQgetvalue(res, i, i_pubinsert), "t") == 0);
pubinfo[i].pubupdate =
@@ -4162,6 +4174,8 @@ getPublications(Archive *fout, int *numPublications)
(strcmp(PQgetvalue(res, i, i_pubdelete), "t") == 0);
pubinfo[i].pubtruncate =
(strcmp(PQgetvalue(res, i, i_pubtruncate), "t") == 0);
+ pubinfo[i].pubsequence =
+ (strcmp(PQgetvalue(res, i, i_pubsequence), "t") == 0);
pubinfo[i].pubviaroot =
(strcmp(PQgetvalue(res, i, i_pubviaroot), "t") == 0);
@@ -4207,6 +4221,9 @@ dumpPublication(Archive *fout, const PublicationInfo *pubinfo)
if (pubinfo->puballtables)
appendPQExpBufferStr(query, " FOR ALL TABLES");
+ if (pubinfo->puballsequences)
+ appendPQExpBufferStr(query, " FOR ALL SEQUENCES");
+
appendPQExpBufferStr(query, " WITH (publish = '");
if (pubinfo->pubinsert)
{
@@ -4241,6 +4258,15 @@ dumpPublication(Archive *fout, const PublicationInfo *pubinfo)
first = false;
}
+ if (pubinfo->pubsequence)
+ {
+ if (!first)
+ appendPQExpBufferStr(query, ", ");
+
+ appendPQExpBufferStr(query, "sequence");
+ first = false;
+ }
+
appendPQExpBufferChar(query, '\'');
if (pubinfo->pubviaroot)
@@ -4287,6 +4313,7 @@ getPublicationNamespaces(Archive *fout)
int i_oid;
int i_pnpubid;
int i_pnnspid;
+ int i_pntype;
int i,
j,
ntups;
@@ -4298,7 +4325,7 @@ getPublicationNamespaces(Archive *fout)
/* Collect all publication membership info. */
appendPQExpBufferStr(query,
- "SELECT tableoid, oid, pnpubid, pnnspid "
+ "SELECT tableoid, oid, pnpubid, pnnspid, pntype "
"FROM pg_catalog.pg_publication_namespace");
res = ExecuteSqlQuery(fout, query->data, PGRES_TUPLES_OK);
@@ -4308,6 +4335,7 @@ getPublicationNamespaces(Archive *fout)
i_oid = PQfnumber(res, "oid");
i_pnpubid = PQfnumber(res, "pnpubid");
i_pnnspid = PQfnumber(res, "pnnspid");
+ i_pntype = PQfnumber(res, "pntype");
/* this allocation may be more than we need */
pubsinfo = pg_malloc(ntups * sizeof(PublicationSchemaInfo));
@@ -4317,6 +4345,7 @@ getPublicationNamespaces(Archive *fout)
{
Oid pnpubid = atooid(PQgetvalue(res, i, i_pnpubid));
Oid pnnspid = atooid(PQgetvalue(res, i, i_pnnspid));
+ char pntype = PQgetvalue(res, i, i_pntype)[0];
PublicationInfo *pubinfo;
NamespaceInfo *nspinfo;
@@ -4348,6 +4377,7 @@ getPublicationNamespaces(Archive *fout)
pubsinfo[j].dobj.name = nspinfo->dobj.name;
pubsinfo[j].publication = pubinfo;
pubsinfo[j].pubschema = nspinfo;
+ pubsinfo[j].pubtype = pntype;
/* Decide whether we want to dump it */
selectDumpablePublicationObject(&(pubsinfo[j].dobj), fout);
@@ -4513,7 +4543,11 @@ dumpPublicationNamespace(Archive *fout, const PublicationSchemaInfo *pubsinfo)
query = createPQExpBuffer();
appendPQExpBuffer(query, "ALTER PUBLICATION %s ", fmtId(pubinfo->dobj.name));
- appendPQExpBuffer(query, "ADD TABLES IN SCHEMA %s;\n", fmtId(schemainfo->dobj.name));
+
+ if (pubsinfo->pubtype == 't')
+ appendPQExpBuffer(query, "ADD TABLES IN SCHEMA %s;\n", fmtId(schemainfo->dobj.name));
+ else
+ appendPQExpBuffer(query, "ADD SEQUENCES IN SCHEMA %s;\n", fmtId(schemainfo->dobj.name));
/*
* There is no point in creating drop query as the drop is done by schema
@@ -4546,6 +4580,7 @@ dumpPublicationTable(Archive *fout, const PublicationRelInfo *pubrinfo)
TableInfo *tbinfo = pubrinfo->pubtable;
PQExpBuffer query;
char *tag;
+ char *description;
/* Do nothing in data-only dump */
if (dopt->dataOnly)
@@ -4555,8 +4590,19 @@ dumpPublicationTable(Archive *fout, const PublicationRelInfo *pubrinfo)
query = createPQExpBuffer();
- appendPQExpBuffer(query, "ALTER PUBLICATION %s ADD TABLE ONLY",
- fmtId(pubinfo->dobj.name));
+ if (tbinfo->relkind == RELKIND_SEQUENCE)
+ {
+ appendPQExpBuffer(query, "ALTER PUBLICATION %s ADD SEQUENCE",
+ fmtId(pubinfo->dobj.name));
+ description = "PUBLICATION SEQUENCE";
+ }
+ else
+ {
+ appendPQExpBuffer(query, "ALTER PUBLICATION %s ADD TABLE ONLY",
+ fmtId(pubinfo->dobj.name));
+ description = "PUBLICATION TABLE";
+ }
+
appendPQExpBuffer(query, " %s",
fmtQualifiedDumpable(tbinfo));
@@ -4586,7 +4632,7 @@ dumpPublicationTable(Archive *fout, const PublicationRelInfo *pubrinfo)
ARCHIVE_OPTS(.tag = tag,
.namespace = tbinfo->dobj.namespace->dobj.name,
.owner = pubinfo->rolname,
- .description = "PUBLICATION TABLE",
+ .description = description,
.section = SECTION_POST_DATA,
.createStmt = query->data));
diff --git a/src/bin/pg_dump/pg_dump.h b/src/bin/pg_dump/pg_dump.h
index 9a34347cfc7..c87114593e7 100644
--- a/src/bin/pg_dump/pg_dump.h
+++ b/src/bin/pg_dump/pg_dump.h
@@ -624,10 +624,12 @@ typedef struct _PublicationInfo
DumpableObject dobj;
const char *rolname;
bool puballtables;
+ bool puballsequences;
bool pubinsert;
bool pubupdate;
bool pubdelete;
bool pubtruncate;
+ bool pubsequence;
bool pubviaroot;
} PublicationInfo;
@@ -653,6 +655,7 @@ typedef struct _PublicationSchemaInfo
DumpableObject dobj;
PublicationInfo *publication;
NamespaceInfo *pubschema;
+ char pubtype;
} PublicationSchemaInfo;
/*
diff --git a/src/bin/pg_dump/t/002_pg_dump.pl b/src/bin/pg_dump/t/002_pg_dump.pl
index 3912dbf481a..ecced19c4a1 100644
--- a/src/bin/pg_dump/t/002_pg_dump.pl
+++ b/src/bin/pg_dump/t/002_pg_dump.pl
@@ -2927,7 +2927,7 @@ my %tests = (
create_order => 50,
create_sql => 'CREATE PUBLICATION pub1;',
regexp => qr/^
- \QCREATE PUBLICATION pub1 WITH (publish = 'insert, update, delete, truncate');\E
+ \QCREATE PUBLICATION pub1 WITH (publish = 'insert, update, delete, truncate, sequence');\E
/xm,
like => { %full_runs, section_post_data => 1, },
},
@@ -2947,7 +2947,7 @@ my %tests = (
create_order => 50,
create_sql => 'CREATE PUBLICATION pub3;',
regexp => qr/^
- \QCREATE PUBLICATION pub3 WITH (publish = 'insert, update, delete, truncate');\E
+ \QCREATE PUBLICATION pub3 WITH (publish = 'insert, update, delete, truncate, sequence');\E
/xm,
like => { %full_runs, section_post_data => 1, },
},
@@ -2956,7 +2956,27 @@ my %tests = (
create_order => 50,
create_sql => 'CREATE PUBLICATION pub4;',
regexp => qr/^
- \QCREATE PUBLICATION pub4 WITH (publish = 'insert, update, delete, truncate');\E
+ \QCREATE PUBLICATION pub4 WITH (publish = 'insert, update, delete, truncate, sequence');\E
+ /xm,
+ like => { %full_runs, section_post_data => 1, },
+ },
+
+ 'CREATE PUBLICATION pub5' => {
+ create_order => 50,
+ create_sql => 'CREATE PUBLICATION pub5
+ FOR ALL SEQUENCES
+ WITH (publish = \'\');',
+ regexp => qr/^
+ \QCREATE PUBLICATION pub5 FOR ALL SEQUENCES WITH (publish = '');\E
+ /xm,
+ like => { %full_runs, section_post_data => 1, },
+ },
+
+ 'CREATE PUBLICATION pub6' => {
+ create_order => 50,
+ create_sql => 'CREATE PUBLICATION pub6;',
+ regexp => qr/^
+ \QCREATE PUBLICATION pub6 WITH (publish = 'insert, update, delete, truncate, sequence');\E
/xm,
like => { %full_runs, section_post_data => 1, },
},
@@ -3102,6 +3122,27 @@ my %tests = (
unlike => { exclude_dump_test_schema => 1, },
},
+ 'ALTER PUBLICATION pub3 ADD SEQUENCES IN SCHEMA dump_test' => {
+ create_order => 51,
+ create_sql =>
+ 'ALTER PUBLICATION pub3 ADD SEQUENCES IN SCHEMA dump_test;',
+ regexp => qr/^
+ \QALTER PUBLICATION pub3 ADD SEQUENCES IN SCHEMA dump_test;\E
+ /xm,
+ like => { %full_runs, section_post_data => 1, },
+ unlike => { exclude_dump_test_schema => 1, },
+ },
+
+ 'ALTER PUBLICATION pub3 ADD SEQUENCES IN SCHEMA public' => {
+ create_order => 52,
+ create_sql =>
+ 'ALTER PUBLICATION pub3 ADD SEQUENCES IN SCHEMA public;',
+ regexp => qr/^
+ \QALTER PUBLICATION pub3 ADD SEQUENCES IN SCHEMA public;\E
+ /xm,
+ like => { %full_runs, section_post_data => 1, },
+ },
+
'CREATE SCHEMA public' => {
regexp => qr/^CREATE SCHEMA public;/m,
diff --git a/src/bin/psql/describe.c b/src/bin/psql/describe.c
index 37f95163201..79fc50180b1 100644
--- a/src/bin/psql/describe.c
+++ b/src/bin/psql/describe.c
@@ -1711,28 +1711,19 @@ describeOneTableDetails(const char *schemaname,
if (tableinfo.relkind == RELKIND_SEQUENCE)
{
PGresult *result = NULL;
- printQueryOpt myopt = pset.popt;
- char *footers[2] = {NULL, NULL};
if (pset.sversion >= 100000)
{
printfPQExpBuffer(&buf,
- "SELECT pg_catalog.format_type(seqtypid, NULL) AS \"%s\",\n"
- " seqstart AS \"%s\",\n"
- " seqmin AS \"%s\",\n"
- " seqmax AS \"%s\",\n"
- " seqincrement AS \"%s\",\n"
- " CASE WHEN seqcycle THEN '%s' ELSE '%s' END AS \"%s\",\n"
- " seqcache AS \"%s\"\n",
- gettext_noop("Type"),
- gettext_noop("Start"),
- gettext_noop("Minimum"),
- gettext_noop("Maximum"),
- gettext_noop("Increment"),
+ "SELECT pg_catalog.format_type(seqtypid, NULL),\n"
+ " seqstart,\n"
+ " seqmin,\n"
+ " seqmax,\n"
+ " seqincrement,\n"
+ " CASE WHEN seqcycle THEN '%s' ELSE '%s' END,\n"
+ " seqcache\n",
gettext_noop("yes"),
- gettext_noop("no"),
- gettext_noop("Cycles?"),
- gettext_noop("Cache"));
+ gettext_noop("no"));
appendPQExpBuffer(&buf,
"FROM pg_catalog.pg_sequence\n"
"WHERE seqrelid = '%s';",
@@ -1741,22 +1732,15 @@ describeOneTableDetails(const char *schemaname,
else
{
printfPQExpBuffer(&buf,
- "SELECT 'bigint' AS \"%s\",\n"
- " start_value AS \"%s\",\n"
- " min_value AS \"%s\",\n"
- " max_value AS \"%s\",\n"
- " increment_by AS \"%s\",\n"
- " CASE WHEN is_cycled THEN '%s' ELSE '%s' END AS \"%s\",\n"
- " cache_value AS \"%s\"\n",
- gettext_noop("Type"),
- gettext_noop("Start"),
- gettext_noop("Minimum"),
- gettext_noop("Maximum"),
- gettext_noop("Increment"),
+ "SELECT 'bigint',\n"
+ " start_value,\n"
+ " min_value,\n"
+ " max_value,\n"
+ " increment_by,\n"
+ " CASE WHEN is_cycled THEN '%s' ELSE '%s' END,\n"
+ " cache_value\n",
gettext_noop("yes"),
- gettext_noop("no"),
- gettext_noop("Cycles?"),
- gettext_noop("Cache"));
+ gettext_noop("no"));
appendPQExpBuffer(&buf, "FROM %s", fmtId(schemaname));
/* must be separate because fmtId isn't reentrant */
appendPQExpBuffer(&buf, ".%s;", fmtId(relationname));
@@ -1766,6 +1750,59 @@ describeOneTableDetails(const char *schemaname,
if (!res)
goto error_return;
+ numrows = PQntuples(res);
+
+ /*
+ * XXX reset to use expanded output for sequences (maybe we should
+ * keep this disabled, just like for tables?)
+ */
+ myopt.expanded = pset.popt.topt.expanded;
+
+ printTableInit(&cont, &myopt, title.data, 7, numrows);
+ printTableInitialized = true;
+
+ if (tableinfo.relpersistence == 'u')
+ printfPQExpBuffer(&title, _("Unlogged sequence \"%s.%s\""),
+ schemaname, relationname);
+ else
+ printfPQExpBuffer(&title, _("Sequence \"%s.%s\""),
+ schemaname, relationname);
+
+ printTableAddHeader(&cont, gettext_noop("Type"), true, 'l');
+ printTableAddHeader(&cont, gettext_noop("Start"), true, 'r');
+ printTableAddHeader(&cont, gettext_noop("Minimum"), true, 'r');
+ printTableAddHeader(&cont, gettext_noop("Maximum"), true, 'r');
+ printTableAddHeader(&cont, gettext_noop("Increment"), true, 'r');
+ printTableAddHeader(&cont, gettext_noop("Cycles?"), true, 'l');
+ printTableAddHeader(&cont, gettext_noop("Cache"), true, 'r');
+
+ /* Generate table cells to be printed */
+ for (i = 0; i < numrows; i++)
+ {
+ /* Type */
+ printTableAddCell(&cont, PQgetvalue(res, i, 0), false, false);
+
+ /* Start */
+ printTableAddCell(&cont, PQgetvalue(res, i, 1), false, false);
+
+ /* Minimum */
+ printTableAddCell(&cont, PQgetvalue(res, i, 2), false, false);
+
+ /* Maximum */
+ printTableAddCell(&cont, PQgetvalue(res, i, 3), false, false);
+
+ /* Increment */
+ printTableAddCell(&cont, PQgetvalue(res, i, 4), false, false);
+
+ /* Cycles? */
+ printTableAddCell(&cont, PQgetvalue(res, i, 5), false, false);
+
+ /* Cache */
+ printTableAddCell(&cont, PQgetvalue(res, i, 6), false, false);
+ }
+
+ /* Footer information about a sequence */
+
/* Get the column that owns this sequence */
printfPQExpBuffer(&buf, "SELECT pg_catalog.quote_ident(nspname) || '.' ||"
"\n pg_catalog.quote_ident(relname) || '.' ||"
@@ -1797,32 +1834,63 @@ describeOneTableDetails(const char *schemaname,
switch (PQgetvalue(result, 0, 1)[0])
{
case 'a':
- footers[0] = psprintf(_("Owned by: %s"),
- PQgetvalue(result, 0, 0));
+ printTableAddFooter(&cont,
+ psprintf(_("Owned by: %s"),
+ PQgetvalue(result, 0, 0)));
break;
case 'i':
- footers[0] = psprintf(_("Sequence for identity column: %s"),
- PQgetvalue(result, 0, 0));
+ printTableAddFooter(&cont,
+ psprintf(_("Sequence for identity column: %s"),
+ PQgetvalue(result, 0, 0)));
break;
}
}
PQclear(result);
- if (tableinfo.relpersistence == 'u')
- printfPQExpBuffer(&title, _("Unlogged sequence \"%s.%s\""),
- schemaname, relationname);
- else
- printfPQExpBuffer(&title, _("Sequence \"%s.%s\""),
- schemaname, relationname);
+ /* print any publications */
+ if (pset.sversion >= 170000)
+ {
+ int tuples = 0;
- myopt.footers = footers;
- myopt.topt.default_footer = false;
- myopt.title = title.data;
- myopt.translate_header = true;
+ printfPQExpBuffer(&buf,
+ "SELECT pubname\n"
+ "FROM pg_catalog.pg_publication p\n"
+ " JOIN pg_catalog.pg_publication_namespace pn ON p.oid = pn.pnpubid\n"
+ " JOIN pg_catalog.pg_class pc ON pc.relnamespace = pn.pnnspid\n"
+ "WHERE pc.oid ='%s' and pn.pntype = 's' and pg_catalog.pg_relation_is_publishable('%s')\n"
+ "UNION\n"
+ "SELECT pubname\n"
+ "FROM pg_catalog.pg_publication p\n"
+ " JOIN pg_catalog.pg_publication_rel pr ON p.oid = pr.prpubid\n"
+ "WHERE pr.prrelid = '%s'\n"
+ "UNION\n"
+ "SELECT pubname\n"
+ "FROM pg_catalog.pg_publication p\n"
+ "WHERE p.puballsequences AND pg_catalog.pg_relation_is_publishable('%s')\n"
+ "ORDER BY 1;",
+ oid, oid, oid, oid);
- printQuery(res, &myopt, pset.queryFout, false, pset.logfile);
+ result = PSQLexec(buf.data);
+ if (!result)
+ goto error_return;
+ else
+ tuples = PQntuples(result);
+
+ if (tuples > 0)
+ printTableAddFooter(&cont, _("Publications:"));
+
+ /* Might be an empty set - that's ok */
+ for (i = 0; i < tuples; i++)
+ {
+ printfPQExpBuffer(&buf, " \"%s\"",
+ PQgetvalue(result, i, 0));
+
+ printTableAddFooter(&cont, buf.data);
+ }
+ PQclear(result);
+ }
- free(footers[0]);
+ printTable(&cont, pset.queryFout, false, pset.logfile);
retval = true;
goto error_return; /* not an error, just return early */
@@ -2049,6 +2117,11 @@ describeOneTableDetails(const char *schemaname,
for (i = 0; i < cols; i++)
printTableAddHeader(&cont, headers[i], true, 'l');
+ res = PSQLexec(buf.data);
+ if (!res)
+ goto error_return;
+ numrows = PQntuples(res);
+
/* Generate table cells to be printed */
for (i = 0; i < numrows; i++)
{
@@ -2974,7 +3047,7 @@ describeOneTableDetails(const char *schemaname,
"FROM pg_catalog.pg_publication p\n"
" JOIN pg_catalog.pg_publication_namespace pn ON p.oid = pn.pnpubid\n"
" JOIN pg_catalog.pg_class pc ON pc.relnamespace = pn.pnnspid\n"
- "WHERE pc.oid ='%s' and pg_catalog.pg_relation_is_publishable('%s')\n"
+ "WHERE pc.oid ='%s' and pn.pntype = 't' and pg_catalog.pg_relation_is_publishable('%s')\n"
"UNION\n"
"SELECT pubname\n"
" , pg_get_expr(pr.prqual, c.oid)\n"
@@ -5111,7 +5184,7 @@ listSchemas(const char *pattern, bool verbose, bool showSystem)
int i;
printfPQExpBuffer(&buf,
- "SELECT pubname \n"
+ "SELECT pubname, (CASE WHEN pntype = 't' THEN 'tables' ELSE 'sequences' END) AS pubtype\n"
"FROM pg_catalog.pg_publication p\n"
" JOIN pg_catalog.pg_publication_namespace pn ON p.oid = pn.pnpubid\n"
" JOIN pg_catalog.pg_namespace n ON n.oid = pn.pnnspid \n"
@@ -5137,8 +5210,9 @@ listSchemas(const char *pattern, bool verbose, bool showSystem)
/* Might be an empty set - that's ok */
for (i = 0; i < pub_schema_tuples; i++)
{
- printfPQExpBuffer(&buf, " \"%s\"",
- PQgetvalue(result, i, 0));
+ printfPQExpBuffer(&buf, " \"%s\" (%s)",
+ PQgetvalue(result, i, 0),
+ PQgetvalue(result, i, 1));
footers[i + 1] = pg_strdup(buf.data);
}
@@ -6254,7 +6328,7 @@ listPublications(const char *pattern)
PQExpBufferData buf;
PGresult *res;
printQueryOpt myopt = pset.popt;
- static const bool translate_columns[] = {false, false, false, false, false, false, false, false};
+ static const bool translate_columns[] = {false, false, false, false, false, false, false, false, false, false};
if (pset.sversion < 100000)
{
@@ -6268,23 +6342,45 @@ listPublications(const char *pattern)
initPQExpBuffer(&buf);
- printfPQExpBuffer(&buf,
- "SELECT pubname AS \"%s\",\n"
- " pg_catalog.pg_get_userbyid(pubowner) AS \"%s\",\n"
- " puballtables AS \"%s\",\n"
- " pubinsert AS \"%s\",\n"
- " pubupdate AS \"%s\",\n"
- " pubdelete AS \"%s\"",
- gettext_noop("Name"),
- gettext_noop("Owner"),
- gettext_noop("All tables"),
- gettext_noop("Inserts"),
- gettext_noop("Updates"),
- gettext_noop("Deletes"));
+ if (pset.sversion >= 170000)
+ printfPQExpBuffer(&buf,
+ "SELECT pubname AS \"%s\",\n"
+ " pg_catalog.pg_get_userbyid(pubowner) AS \"%s\",\n"
+ " puballtables AS \"%s\",\n"
+ " puballsequences AS \"%s\",\n"
+ " pubinsert AS \"%s\",\n"
+ " pubupdate AS \"%s\",\n"
+ " pubdelete AS \"%s\"",
+ gettext_noop("Name"),
+ gettext_noop("Owner"),
+ gettext_noop("All tables"),
+ gettext_noop("All sequences"),
+ gettext_noop("Inserts"),
+ gettext_noop("Updates"),
+ gettext_noop("Deletes"));
+ else
+ printfPQExpBuffer(&buf,
+ "SELECT pubname AS \"%s\",\n"
+ " pg_catalog.pg_get_userbyid(pubowner) AS \"%s\",\n"
+ " puballtables AS \"%s\",\n"
+ " pubinsert AS \"%s\",\n"
+ " pubupdate AS \"%s\",\n"
+ " pubdelete AS \"%s\"",
+ gettext_noop("Name"),
+ gettext_noop("Owner"),
+ gettext_noop("All tables"),
+ gettext_noop("Inserts"),
+ gettext_noop("Updates"),
+ gettext_noop("Deletes"));
+
if (pset.sversion >= 110000)
appendPQExpBuffer(&buf,
",\n pubtruncate AS \"%s\"",
gettext_noop("Truncates"));
+ if (pset.sversion >= 170000)
+ appendPQExpBuffer(&buf,
+ ",\n pubsequence AS \"%s\"",
+ gettext_noop("Sequences"));
if (pset.sversion >= 130000)
appendPQExpBuffer(&buf,
",\n pubviaroot AS \"%s\"",
@@ -6378,6 +6474,7 @@ describePublications(const char *pattern)
PGresult *res;
bool has_pubtruncate;
bool has_pubviaroot;
+ bool has_pubsequence;
PQExpBufferData title;
printTableContent cont;
@@ -6394,6 +6491,7 @@ describePublications(const char *pattern)
has_pubtruncate = (pset.sversion >= 110000);
has_pubviaroot = (pset.sversion >= 130000);
+ has_pubsequence = (pset.sversion >= 170000);
initPQExpBuffer(&buf);
@@ -6407,6 +6505,10 @@ describePublications(const char *pattern)
if (has_pubviaroot)
appendPQExpBufferStr(&buf,
", pubviaroot");
+ if (has_pubsequence)
+ appendPQExpBufferStr(&buf,
+ ", puballsequences, pubsequence");
+
appendPQExpBufferStr(&buf,
"\nFROM pg_catalog.pg_publication\n");
@@ -6452,6 +6554,7 @@ describePublications(const char *pattern)
char *pubid = PQgetvalue(res, i, 0);
char *pubname = PQgetvalue(res, i, 1);
bool puballtables = strcmp(PQgetvalue(res, i, 3), "t") == 0;
+ bool puballsequences = strcmp(PQgetvalue(res, i, 9), "t") == 0;
printTableOpt myopt = pset.popt.topt;
if (has_pubtruncate)
@@ -6459,29 +6562,43 @@ describePublications(const char *pattern)
if (has_pubviaroot)
ncols++;
+ /* sequences have two extra columns (puballsequences, pubsequences) */
+ if (has_pubsequence)
+ ncols += 2;
+
initPQExpBuffer(&title);
printfPQExpBuffer(&title, _("Publication %s"), pubname);
printTableInit(&cont, &myopt, title.data, ncols, nrows);
printTableAddHeader(&cont, gettext_noop("Owner"), true, align);
printTableAddHeader(&cont, gettext_noop("All tables"), true, align);
+ if (has_pubsequence)
+ printTableAddHeader(&cont, gettext_noop("All sequences"), true, align);
printTableAddHeader(&cont, gettext_noop("Inserts"), true, align);
printTableAddHeader(&cont, gettext_noop("Updates"), true, align);
printTableAddHeader(&cont, gettext_noop("Deletes"), true, align);
if (has_pubtruncate)
printTableAddHeader(&cont, gettext_noop("Truncates"), true, align);
+ if (has_pubsequence)
+ printTableAddHeader(&cont, gettext_noop("Sequences"), true, align);
if (has_pubviaroot)
printTableAddHeader(&cont, gettext_noop("Via root"), true, align);
- printTableAddCell(&cont, PQgetvalue(res, i, 2), false, false);
- printTableAddCell(&cont, PQgetvalue(res, i, 3), false, false);
- printTableAddCell(&cont, PQgetvalue(res, i, 4), false, false);
- printTableAddCell(&cont, PQgetvalue(res, i, 5), false, false);
- printTableAddCell(&cont, PQgetvalue(res, i, 6), false, false);
+ printTableAddCell(&cont, PQgetvalue(res, i, 2), false, false); /* owner */
+ printTableAddCell(&cont, PQgetvalue(res, i, 3), false, false); /* all tables */
+
+ if (has_pubsequence)
+ printTableAddCell(&cont, PQgetvalue(res, i, 9), false, false); /* all sequences */
+
+ printTableAddCell(&cont, PQgetvalue(res, i, 4), false, false); /* insert */
+ printTableAddCell(&cont, PQgetvalue(res, i, 5), false, false); /* update */
+ printTableAddCell(&cont, PQgetvalue(res, i, 6), false, false); /* delete */
if (has_pubtruncate)
- printTableAddCell(&cont, PQgetvalue(res, i, 7), false, false);
+ printTableAddCell(&cont, PQgetvalue(res, i, 7), false, false); /* truncate */
+ if (has_pubsequence)
+ printTableAddCell(&cont, PQgetvalue(res, i, 10), false, false); /* sequence */
if (has_pubviaroot)
- printTableAddCell(&cont, PQgetvalue(res, i, 8), false, false);
+ printTableAddCell(&cont, PQgetvalue(res, i, 8), false, false); /* via root */
if (!puballtables)
{
@@ -6512,6 +6629,7 @@ describePublications(const char *pattern)
"WHERE c.relnamespace = n.oid\n"
" AND c.oid = pr.prrelid\n"
" AND pr.prpubid = '%s'\n"
+ " AND c.relkind != 'S'\n" /* exclude sequences */
"ORDER BY 1,2", pubid);
if (!addFooterToPublicationDesc(&buf, _("Tables:"), false, &cont))
goto error_return;
@@ -6523,7 +6641,7 @@ describePublications(const char *pattern)
"SELECT n.nspname\n"
"FROM pg_catalog.pg_namespace n\n"
" JOIN pg_catalog.pg_publication_namespace pn ON n.oid = pn.pnnspid\n"
- "WHERE pn.pnpubid = '%s'\n"
+ "WHERE pn.pnpubid = '%s' AND pn.pntype = 't'\n"
"ORDER BY 1", pubid);
if (!addFooterToPublicationDesc(&buf, _("Tables from schemas:"),
true, &cont))
@@ -6531,6 +6649,37 @@ describePublications(const char *pattern)
}
}
+ if (!puballsequences)
+ {
+ /* Get the sequences for the specified publication */
+ printfPQExpBuffer(&buf,
+ "SELECT n.nspname, c.relname, NULL, NULL\n"
+ "FROM pg_catalog.pg_class c,\n"
+ " pg_catalog.pg_namespace n,\n"
+ " pg_catalog.pg_publication_rel pr\n"
+ "WHERE c.relnamespace = n.oid\n"
+ " AND c.oid = pr.prrelid\n"
+ " AND pr.prpubid = '%s'\n"
+ " AND c.relkind = 'S'\n" /* only sequences */
+ "ORDER BY 1,2", pubid);
+ if (!addFooterToPublicationDesc(&buf, "Sequences:", false, &cont))
+ goto error_return;
+
+ if (pset.sversion >= 150000)
+ {
+ /* Get the schemas for the specified publication */
+ printfPQExpBuffer(&buf,
+ "SELECT n.nspname\n"
+ "FROM pg_catalog.pg_namespace n\n"
+ " JOIN pg_catalog.pg_publication_namespace pn ON n.oid = pn.pnnspid\n"
+ "WHERE pn.pnpubid = '%s' AND pn.pntype = 's'\n"
+ "ORDER BY 1", pubid);
+ if (!addFooterToPublicationDesc(&buf, "Sequences from schemas:",
+ true, &cont))
+ goto error_return;
+ }
+ }
+
printTable(&cont, pset.queryFout, false, pset.logfile);
printTableCleanup(&cont);
diff --git a/src/bin/psql/tab-complete.c b/src/bin/psql/tab-complete.c
index 09914165e42..cd1bd4ebf81 100644
--- a/src/bin/psql/tab-complete.c
+++ b/src/bin/psql/tab-complete.c
@@ -1890,11 +1890,15 @@ psql_completion(const char *text, int start, int end)
COMPLETE_WITH("ADD", "DROP", "OWNER TO", "RENAME TO", "SET");
/* ALTER PUBLICATION <name> ADD */
else if (Matches("ALTER", "PUBLICATION", MatchAny, "ADD"))
- COMPLETE_WITH("TABLES IN SCHEMA", "TABLE");
+ COMPLETE_WITH("TABLES IN SCHEMA", "TABLE", "SEQUENCES IN SCHEMA", "SEQUENCE");
else if (Matches("ALTER", "PUBLICATION", MatchAny, "ADD|SET", "TABLE") ||
(HeadMatches("ALTER", "PUBLICATION", MatchAny, "ADD|SET", "TABLE") &&
ends_with(prev_wd, ',')))
COMPLETE_WITH_SCHEMA_QUERY(Query_for_list_of_tables);
+ else if (Matches("ALTER", "PUBLICATION", MatchAny, "ADD|SET", "SEQUENCE") ||
+ (HeadMatches("ALTER", "PUBLICATION", MatchAny, "ADD|SET", "SEQUENCE") &&
+ ends_with(prev_wd, ',')))
+ COMPLETE_WITH_SCHEMA_QUERY(Query_for_list_of_sequences);
/*
* "ALTER PUBLICATION <name> SET TABLE <name> WHERE (" - complete with
@@ -1914,11 +1918,11 @@ psql_completion(const char *text, int start, int end)
COMPLETE_WITH(",");
/* ALTER PUBLICATION <name> DROP */
else if (Matches("ALTER", "PUBLICATION", MatchAny, "DROP"))
- COMPLETE_WITH("TABLES IN SCHEMA", "TABLE");
+ COMPLETE_WITH("TABLES IN SCHEMA", "TABLE", "SEQUENCES IN SCHEMA", "SEQUENCE");
/* ALTER PUBLICATION <name> SET */
else if (Matches("ALTER", "PUBLICATION", MatchAny, "SET"))
- COMPLETE_WITH("(", "TABLES IN SCHEMA", "TABLE");
- else if (Matches("ALTER", "PUBLICATION", MatchAny, "ADD|DROP|SET", "TABLES", "IN", "SCHEMA"))
+ COMPLETE_WITH("(", "TABLES IN SCHEMA", "TABLE", "SEQUENCES IN SCHEMA", "SEQUENCE");
+ else if (Matches("ALTER", "PUBLICATION", MatchAny, "ADD|DROP|SET", "TABLES|SEQUENCES", "IN", "SCHEMA"))
COMPLETE_WITH_QUERY_PLUS(Query_for_list_of_schemas
" AND nspname NOT LIKE E'pg\\\\_%%'",
"CURRENT_SCHEMA");
@@ -3131,24 +3135,31 @@ psql_completion(const char *text, int start, int end)
/* CREATE PUBLICATION */
else if (Matches("CREATE", "PUBLICATION", MatchAny))
- COMPLETE_WITH("FOR TABLE", "FOR ALL TABLES", "FOR TABLES IN SCHEMA", "WITH (");
+ COMPLETE_WITH("FOR TABLE", "FOR ALL TABLES", "FOR TABLES IN SCHEMA", "FOR SEQUENCE", "FOR ALL SEQUENCES", "FOR SEQUENCES IN SCHEMA", "WITH (");
else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR"))
- COMPLETE_WITH("TABLE", "ALL TABLES", "TABLES IN SCHEMA");
+ COMPLETE_WITH("TABLE", "ALL TABLES", "TABLES IN SCHEMA", "SEQUENCE", "ALL SEQUENCES", "SEQUENCES IN SCHEMA");
else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "ALL"))
- COMPLETE_WITH("TABLES");
- else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "ALL", "TABLES"))
+ COMPLETE_WITH("TABLES", "SEQUENCES");
+ else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "ALL", "TABLES|SEQUENCES"))
COMPLETE_WITH("WITH (");
- else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "TABLES"))
+ else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "TABLES|SEQUENCES"))
COMPLETE_WITH("IN SCHEMA");
- else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "TABLE", MatchAny) && !ends_with(prev_wd, ','))
+ else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "TABLE|SEQUENCE", MatchAny) && !ends_with(prev_wd, ','))
COMPLETE_WITH("WHERE (", "WITH (");
/* Complete "CREATE PUBLICATION <name> FOR TABLE" with "<table>, ..." */
else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "TABLE"))
COMPLETE_WITH_SCHEMA_QUERY(Query_for_list_of_tables);
/*
- * "CREATE PUBLICATION <name> FOR TABLE <name> WHERE (" - complete with
- * table attributes
+ * Complete "CREATE PUBLICATION <name> FOR SEQUENCE" with "<sequence>,
+ * ..."
+ */
+ else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "SEQUENCE"))
+ COMPLETE_WITH_SCHEMA_QUERY(Query_for_list_of_sequences);
+
+ /*
+ * "CREATE PUBLICATION <name> FOR TABLE|SEQUENCE <name> WHERE (" -
+ * complete with table attributes
*/
else if (HeadMatches("CREATE", "PUBLICATION", MatchAny) && TailMatches("WHERE"))
COMPLETE_WITH("(");
@@ -3160,11 +3171,11 @@ psql_completion(const char *text, int start, int end)
/*
* Complete "CREATE PUBLICATION <name> FOR TABLES IN SCHEMA <schema>, ..."
*/
- else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "TABLES", "IN", "SCHEMA"))
+ else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "TABLES|SEQUENCES", "IN", "SCHEMA"))
COMPLETE_WITH_QUERY_PLUS(Query_for_list_of_schemas
" AND nspname NOT LIKE E'pg\\\\_%%'",
"CURRENT_SCHEMA");
- else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "TABLES", "IN", "SCHEMA", MatchAny) && (!ends_with(prev_wd, ',')))
+ else if (Matches("CREATE", "PUBLICATION", MatchAny, "FOR", "TABLES|SEQUENCES", "IN", "SCHEMA", MatchAny) && (!ends_with(prev_wd, ',')))
COMPLETE_WITH("WITH (");
/* Complete "CREATE PUBLICATION <name> [...] WITH" */
else if (HeadMatches("CREATE", "PUBLICATION") && TailMatches("WITH", "("))
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 7979392776d..79953b3b865 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -3329,6 +3329,14 @@
proname => 'pg_sequence_last_value', provolatile => 'v', proparallel => 'u',
prorettype => 'int8', proargtypes => 'regclass',
prosrc => 'pg_sequence_last_value' },
+{ oid => '6312',
+ descr => 'current on-disk sequence state',
+ proname => 'pg_sequence_state', provolatile => 'v',
+ prorettype => 'record', proargtypes => 'regclass',
+ proallargtypes => '{regclass,pg_lsn,int8,int8,bool}',
+ proargmodes => '{i,o,o,o,o}',
+ proargnames => '{seq_oid,page_lsn,last_value,log_cnt,is_called}',
+ prosrc => 'pg_sequence_state' },
{ oid => '275', descr => 'return the next oid for a system table',
proname => 'pg_nextoid', provolatile => 'v', proparallel => 'u',
@@ -11897,6 +11905,11 @@
proargmodes => '{v,o,o,o,o}',
proargnames => '{pubname,pubid,relid,attrs,qual}',
prosrc => 'pg_get_publication_tables' },
+{ oid => '8000', descr => 'get OIDs of sequences in a publication',
+ proname => 'pg_get_publication_sequences', prorows => '1000', proretset => 't',
+ provolatile => 's', prorettype => 'oid', proargtypes => 'text',
+ proallargtypes => '{text,oid}', proargmodes => '{i,o}',
+ proargnames => '{pubname,relid}', prosrc => 'pg_get_publication_sequences' },
{ oid => '6121',
descr => 'returns whether a relation can be part of a publication',
proname => 'pg_relation_is_publishable', provolatile => 's',
diff --git a/src/include/catalog/pg_publication.h b/src/include/catalog/pg_publication.h
index 6e9bbe80ffc..a4ffeb20026 100644
--- a/src/include/catalog/pg_publication.h
+++ b/src/include/catalog/pg_publication.h
@@ -40,6 +40,12 @@ CATALOG(pg_publication,6104,PublicationRelationId)
*/
bool puballtables;
+ /*
+ * indicates that this is special publication which should encompass all
+ * sequences in the database (except for the unlogged and temp ones)
+ */
+ bool puballsequences;
+
/* true if inserts are published */
bool pubinsert;
@@ -52,6 +58,9 @@ CATALOG(pg_publication,6104,PublicationRelationId)
/* true if truncates are published */
bool pubtruncate;
+ /* true if sequences are published */
+ bool pubsequence;
+
/* true if partition changes are published using root schema */
bool pubviaroot;
} FormData_pg_publication;
@@ -72,6 +81,7 @@ typedef struct PublicationActions
bool pubupdate;
bool pubdelete;
bool pubtruncate;
+ bool pubsequence;
} PublicationActions;
typedef struct PublicationDesc
@@ -99,6 +109,7 @@ typedef struct Publication
Oid oid;
char *name;
bool alltables;
+ bool allsequences;
bool pubviaroot;
PublicationActions pubactions;
} Publication;
@@ -130,14 +141,15 @@ typedef enum PublicationPartOpt
PUBLICATION_PART_ALL,
} PublicationPartOpt;
-extern List *GetPublicationRelations(Oid pubid, PublicationPartOpt pub_partopt);
+extern List *GetPublicationRelations(Oid pubid, char objectType,
+ PublicationPartOpt pub_partopt);
extern List *GetAllTablesPublications(void);
extern List *GetAllTablesPublicationRelations(bool pubviaroot);
-extern List *GetPublicationSchemas(Oid pubid);
-extern List *GetSchemaPublications(Oid schemaid);
-extern List *GetSchemaPublicationRelations(Oid schemaid,
+extern List *GetPublicationSchemas(Oid pubid, char objectType);
+extern List *GetSchemaPublications(Oid schemaid, char objectType);
+extern List *GetSchemaPublicationRelations(Oid schemaid, char objectType,
PublicationPartOpt pub_partopt);
-extern List *GetAllSchemaPublicationRelations(Oid pubid,
+extern List *GetAllSchemaPublicationRelations(Oid puboid, char objectType,
PublicationPartOpt pub_partopt);
extern List *GetPubPartitionOptionRelations(List *result,
PublicationPartOpt pub_partopt,
@@ -145,11 +157,15 @@ extern List *GetPubPartitionOptionRelations(List *result,
extern Oid GetTopMostAncestorInPublication(Oid puboid, List *ancestors,
int *ancestor_level);
+extern List *GetAllSequencesPublications(void);
+extern List *GetAllSequencesPublicationRelations(void);
+
extern bool is_publishable_relation(Relation rel);
extern bool is_schema_publication(Oid pubid);
extern ObjectAddress publication_add_relation(Oid pubid, PublicationRelInfo *pri,
bool if_not_exists);
extern ObjectAddress publication_add_schema(Oid pubid, Oid schemaid,
+ char objectType,
bool if_not_exists);
extern Bitmapset *pub_collist_to_bitmapset(Bitmapset *columns, Datum pubcols,
diff --git a/src/include/catalog/pg_publication_namespace.h b/src/include/catalog/pg_publication_namespace.h
index 92aa58c1ed0..30f599b4f5a 100644
--- a/src/include/catalog/pg_publication_namespace.h
+++ b/src/include/catalog/pg_publication_namespace.h
@@ -32,6 +32,7 @@ CATALOG(pg_publication_namespace,6237,PublicationNamespaceRelationId)
Oid oid; /* oid */
Oid pnpubid BKI_LOOKUP(pg_publication); /* Oid of the publication */
Oid pnnspid BKI_LOOKUP(pg_namespace); /* Oid of the schema */
+ char pntype; /* object type to include */
} FormData_pg_publication_namespace;
/* ----------------
@@ -42,6 +43,13 @@ CATALOG(pg_publication_namespace,6237,PublicationNamespaceRelationId)
typedef FormData_pg_publication_namespace *Form_pg_publication_namespace;
DECLARE_UNIQUE_INDEX_PKEY(pg_publication_namespace_oid_index, 6238, PublicationNamespaceObjectIndexId, pg_publication_namespace, btree(oid oid_ops));
-DECLARE_UNIQUE_INDEX(pg_publication_namespace_pnnspid_pnpubid_index, 6239, PublicationNamespacePnnspidPnpubidIndexId, pg_publication_namespace, btree(pnnspid oid_ops, pnpubid oid_ops));
+DECLARE_UNIQUE_INDEX(pg_publication_namespace_pnnspid_pnpubid_pntype_index, 8903, PublicationNamespacePnnspidPnpubidPntypeIndexId, pg_publication_namespace, btree(pnnspid oid_ops, pnpubid oid_ops, pntype char_ops));
+
+/* object type to include from a schema, maps to relkind */
+#define PUB_OBJTYPE_TABLE 't' /* table (regular or partitioned) */
+#define PUB_OBJTYPE_SEQUENCE 's' /* sequence object */
+#define PUB_OBJTYPE_UNSUPPORTED 'u' /* used for non-replicated types */
+
+extern char pub_get_object_type_for_relkind(char relkind);
#endif /* PG_PUBLICATION_NAMESPACE_H */
diff --git a/src/include/commands/sequence.h b/src/include/commands/sequence.h
index e88cbee3b56..03a4fe1173c 100644
--- a/src/include/commands/sequence.h
+++ b/src/include/commands/sequence.h
@@ -60,6 +60,7 @@ extern ObjectAddress AlterSequence(ParseState *pstate, AlterSeqStmt *stmt);
extern void SequenceChangePersistence(Oid relid, char newrelpersistence);
extern void DeleteSequenceTuple(Oid relid);
extern void ResetSequence(Oid seq_relid);
+extern void SetSequence(Oid seq_relid, bool transactional, int64 value);
extern void ResetSequenceCaches(void);
extern void seq_redo(XLogReaderState *record);
diff --git a/src/include/nodes/parsenodes.h b/src/include/nodes/parsenodes.h
index b3181f34aee..e9fa10ea9ea 100644
--- a/src/include/nodes/parsenodes.h
+++ b/src/include/nodes/parsenodes.h
@@ -3991,6 +3991,10 @@ typedef enum PublicationObjSpecType
PUBLICATIONOBJ_TABLES_IN_SCHEMA, /* All tables in schema */
PUBLICATIONOBJ_TABLES_IN_CUR_SCHEMA, /* All tables in first element of
* search_path */
+ PUBLICATIONOBJ_SEQUENCE, /* Sequence type */
+ PUBLICATIONOBJ_SEQUENCES_IN_SCHEMA, /* Sequences in schema type */
+ PUBLICATIONOBJ_SEQUENCES_IN_CUR_SCHEMA, /* Get the first element of
+ * search_path */
PUBLICATIONOBJ_CONTINUATION, /* Continuation of previous type */
} PublicationObjSpecType;
@@ -4009,7 +4013,8 @@ typedef struct CreatePublicationStmt
char *pubname; /* Name of the publication */
List *options; /* List of DefElem nodes */
List *pubobjects; /* Optional list of publication objects */
- bool for_all_tables; /* Special publication for all tables in db */
+ List *for_all_objects; /* Special publication for all objects in
+ * db */
} CreatePublicationStmt;
typedef enum AlterPublicationAction
@@ -4032,7 +4037,8 @@ typedef struct AlterPublicationStmt
* objects.
*/
List *pubobjects; /* Optional list of publication objects */
- bool for_all_tables; /* Special publication for all tables in db */
+ List *for_all_objects; /* Special publication for all objects in
+ * db */
AlterPublicationAction action; /* What action to perform with the given
* objects */
} AlterPublicationStmt;
diff --git a/src/include/replication/logicalproto.h b/src/include/replication/logicalproto.h
index c409638a2ef..082b5364dd3 100644
--- a/src/include/replication/logicalproto.h
+++ b/src/include/replication/logicalproto.h
@@ -42,7 +42,8 @@
#define LOGICALREP_PROTO_STREAM_VERSION_NUM 2
#define LOGICALREP_PROTO_TWOPHASE_VERSION_NUM 3
#define LOGICALREP_PROTO_STREAM_PARALLEL_VERSION_NUM 4
-#define LOGICALREP_PROTO_MAX_VERSION_NUM LOGICALREP_PROTO_STREAM_PARALLEL_VERSION_NUM
+#define LOGICALREP_PROTO_SEQUENCES_VERSION_NUM 5
+#define LOGICALREP_PROTO_MAX_VERSION_NUM LOGICALREP_PROTO_SEQUENCES_VERSION_NUM
/*
* Logical message types
@@ -66,6 +67,7 @@ typedef enum LogicalRepMsgType
LOGICAL_REP_MSG_RELATION = 'R',
LOGICAL_REP_MSG_TYPE = 'Y',
LOGICAL_REP_MSG_MESSAGE = 'M',
+ LOGICAL_REP_MSG_SEQUENCE = 'Q',
LOGICAL_REP_MSG_BEGIN_PREPARE = 'b',
LOGICAL_REP_MSG_PREPARE = 'P',
LOGICAL_REP_MSG_COMMIT_PREPARED = 'K',
@@ -123,6 +125,17 @@ typedef struct LogicalRepTyp
char *typname; /* name of the remote type */
} LogicalRepTyp;
+/* Sequence info */
+typedef struct LogicalRepSequence
+{
+ XLogRecPtr lsn; /* LSN of the sequence change */
+ Oid remoteid; /* unique id of the remote sequence */
+ char *nspname; /* schema name of remote sequence */
+ char *seqname; /* name of the remote sequence */
+ bool transactional;
+ int64 value;
+} LogicalRepSequence;
+
/* Transaction info */
typedef struct LogicalRepBeginData
{
@@ -246,6 +259,10 @@ extern List *logicalrep_read_truncate(StringInfo in,
bool *cascade, bool *restart_seqs);
extern void logicalrep_write_message(StringInfo out, TransactionId xid, XLogRecPtr lsn,
bool transactional, const char *prefix, Size sz, const char *message);
+extern void logicalrep_write_sequence(StringInfo out, Relation rel,
+ TransactionId xid, XLogRecPtr lsn,
+ bool transactional, int64 value);
+extern LogicalRepRelId logicalrep_read_sequence(StringInfo in, LogicalRepSequence *seqdata);
extern void logicalrep_write_rel(StringInfo out, TransactionId xid,
Relation rel, Bitmapset *columns);
extern LogicalRepRelation *logicalrep_read_rel(StringInfo in);
diff --git a/src/test/regress/expected/object_address.out b/src/test/regress/expected/object_address.out
index fc42d418bf1..9aec2612d01 100644
--- a/src/test/regress/expected/object_address.out
+++ b/src/test/regress/expected/object_address.out
@@ -47,6 +47,7 @@ CREATE TRANSFORM FOR int LANGUAGE SQL (
SET client_min_messages = 'ERROR';
CREATE PUBLICATION addr_pub FOR TABLE addr_nsp.gentable;
CREATE PUBLICATION addr_pub_schema FOR TABLES IN SCHEMA addr_nsp;
+CREATE PUBLICATION addr_pub_schema2 FOR SEQUENCES IN SCHEMA addr_nsp;
RESET client_min_messages;
CREATE SUBSCRIPTION regress_addr_sub CONNECTION '' PUBLICATION bar WITH (connect = false, slot_name = NONE);
WARNING: subscription was created, but is not connected
@@ -315,12 +316,12 @@ WARNING: error for function of access method,{addr_nsp,zwei},{}: name list leng
WARNING: error for function of access method,{addr_nsp,zwei},{integer}: name list length must be at least 3
WARNING: error for function of access method,{eins,zwei,drei},{}: argument list length must be exactly 2
WARNING: error for function of access method,{eins,zwei,drei},{integer}: argument list length must be exactly 2
-WARNING: error for publication namespace,{eins},{}: argument list length must be exactly 1
-WARNING: error for publication namespace,{eins},{integer}: schema "eins" does not exist
-WARNING: error for publication namespace,{addr_nsp,zwei},{}: name list length must be exactly 1
-WARNING: error for publication namespace,{addr_nsp,zwei},{integer}: name list length must be exactly 1
-WARNING: error for publication namespace,{eins,zwei,drei},{}: name list length must be exactly 1
-WARNING: error for publication namespace,{eins,zwei,drei},{integer}: name list length must be exactly 1
+WARNING: error for publication namespace,{eins},{}: argument list length must be exactly 2
+WARNING: error for publication namespace,{eins},{integer}: argument list length must be exactly 2
+WARNING: error for publication namespace,{addr_nsp,zwei},{}: argument list length must be exactly 2
+WARNING: error for publication namespace,{addr_nsp,zwei},{integer}: argument list length must be exactly 2
+WARNING: error for publication namespace,{eins,zwei,drei},{}: argument list length must be exactly 2
+WARNING: error for publication namespace,{eins,zwei,drei},{integer}: argument list length must be exactly 2
WARNING: error for publication relation,{eins},{}: argument list length must be exactly 1
WARNING: error for publication relation,{eins},{integer}: relation "eins" does not exist
WARNING: error for publication relation,{addr_nsp,zwei},{}: argument list length must be exactly 1
@@ -441,7 +442,8 @@ WITH objects (type, name, args) AS (VALUES
('transform', '{int}', '{sql}'),
('access method', '{btree}', '{}'),
('publication', '{addr_pub}', '{}'),
- ('publication namespace', '{addr_nsp}', '{addr_pub_schema}'),
+ ('publication namespace', '{addr_nsp}', '{addr_pub_schema, t}'),
+ ('publication namespace', '{addr_nsp}', '{addr_pub_schema2, s}'),
('publication relation', '{addr_nsp, gentable}', '{addr_pub}'),
('subscription', '{regress_addr_sub}', '{}'),
('statistics object', '{addr_nsp, gentable_stat}', '{}')
@@ -504,7 +506,8 @@ text search template|addr_nsp|addr_ts_temp|addr_nsp.addr_ts_temp|t
subscription|NULL|regress_addr_sub|regress_addr_sub|t
publication|NULL|addr_pub|addr_pub|t
publication relation|NULL|NULL|addr_nsp.gentable in publication addr_pub|t
-publication namespace|NULL|NULL|addr_nsp in publication addr_pub_schema|t
+publication namespace|NULL|NULL|addr_nsp in publication addr_pub_schema type t|t
+publication namespace|NULL|NULL|addr_nsp in publication addr_pub_schema2 type s|t
---
--- Cleanup resources
---
@@ -516,6 +519,7 @@ drop cascades to server integer
drop cascades to user mapping for regress_addr_user on server integer
DROP PUBLICATION addr_pub;
DROP PUBLICATION addr_pub_schema;
+DROP PUBLICATION addr_pub_schema2;
DROP SUBSCRIPTION regress_addr_sub;
DROP SCHEMA addr_nsp CASCADE;
NOTICE: drop cascades to 14 other objects
diff --git a/src/test/regress/expected/psql.out b/src/test/regress/expected/psql.out
index 4f3fd464207..e1d0fda06fd 100644
--- a/src/test/regress/expected/psql.out
+++ b/src/test/regress/expected/psql.out
@@ -6244,9 +6244,9 @@ List of schemas
(0 rows)
\dRp "no.such.publication"
- List of publications
- Name | Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
-------+-------+------------+---------+---------+---------+-----------+----------
+ List of publications
+ Name | Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+------+-------+------------+---------------+---------+---------+---------+-----------+-----------+----------
(0 rows)
\dRs "no.such.subscription"
diff --git a/src/test/regress/expected/publication.out b/src/test/regress/expected/publication.out
index 16361a91f9f..2f5c72c9d2b 100644
--- a/src/test/regress/expected/publication.out
+++ b/src/test/regress/expected/publication.out
@@ -30,20 +30,20 @@ ERROR: conflicting or redundant options
LINE 1: ...ub_xxx WITH (publish_via_partition_root = 'true', publish_vi...
^
\dRp
- List of publications
- Name | Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------+--------------------------+------------+---------+---------+---------+-----------+----------
- testpib_ins_trunct | regress_publication_user | f | t | f | f | f | f
- testpub_default | regress_publication_user | f | f | t | f | f | f
+ List of publications
+ Name | Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ testpib_ins_trunct | regress_publication_user | f | f | t | f | f | f | f | f
+ testpub_default | regress_publication_user | f | f | f | t | f | f | f | f
(2 rows)
-ALTER PUBLICATION testpub_default SET (publish = 'insert, update, delete');
+ALTER PUBLICATION testpub_default SET (publish = 'insert, update, delete, sequence');
\dRp
- List of publications
- Name | Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------+--------------------------+------------+---------+---------+---------+-----------+----------
- testpib_ins_trunct | regress_publication_user | f | t | f | f | f | f
- testpub_default | regress_publication_user | f | t | t | t | f | f
+ List of publications
+ Name | Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ testpib_ins_trunct | regress_publication_user | f | f | t | f | f | f | f | f
+ testpub_default | regress_publication_user | f | f | t | t | t | f | t | f
(2 rows)
--- adding tables
@@ -61,6 +61,9 @@ CREATE TABLE testpub_tbl2 (id serial primary key, data text);
ALTER PUBLICATION testpub_foralltables ADD TABLE testpub_tbl2;
ERROR: publication "testpub_foralltables" is defined as FOR ALL TABLES
DETAIL: Tables cannot be added to or dropped from FOR ALL TABLES publications.
+-- fail - can't add a table using ADD SEQUENCE command
+ALTER PUBLICATION testpub_foralltables ADD SEQUENCE testpub_tbl2;
+ERROR: object type does not match type expected by command
-- fail - can't drop from all tables publication
ALTER PUBLICATION testpub_foralltables DROP TABLE testpub_tbl2;
ERROR: publication "testpub_foralltables" is defined as FOR ALL TABLES
@@ -87,10 +90,10 @@ RESET client_min_messages;
-- should be able to add schema to 'FOR TABLE' publication
ALTER PUBLICATION testpub_fortable ADD TABLES IN SCHEMA pub_test;
\dRp+ testpub_fortable
- Publication testpub_fortable
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_fortable
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"public.testpub_tbl1"
Tables from schemas:
@@ -99,20 +102,20 @@ Tables from schemas:
-- should be able to drop schema from 'FOR TABLE' publication
ALTER PUBLICATION testpub_fortable DROP TABLES IN SCHEMA pub_test;
\dRp+ testpub_fortable
- Publication testpub_fortable
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_fortable
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"public.testpub_tbl1"
-- should be able to set schema to 'FOR TABLE' publication
ALTER PUBLICATION testpub_fortable SET TABLES IN SCHEMA pub_test;
\dRp+ testpub_fortable
- Publication testpub_fortable
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_fortable
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test"
@@ -123,10 +126,10 @@ CREATE PUBLICATION testpub_forschema FOR TABLES IN SCHEMA pub_test;
CREATE PUBLICATION testpub_for_tbl_schema FOR TABLES IN SCHEMA pub_test, TABLE pub_test.testpub_nopk;
RESET client_min_messages;
\dRp+ testpub_for_tbl_schema
- Publication testpub_for_tbl_schema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_for_tbl_schema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"pub_test.testpub_nopk"
Tables from schemas:
@@ -144,10 +147,10 @@ LINE 1: ...CATION testpub_parsertst FOR TABLES IN SCHEMA foo, test.foo;
-- should be able to add a table of the same schema to the schema publication
ALTER PUBLICATION testpub_forschema ADD TABLE pub_test.testpub_nopk;
\dRp+ testpub_forschema
- Publication testpub_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"pub_test.testpub_nopk"
Tables from schemas:
@@ -156,10 +159,10 @@ Tables from schemas:
-- should be able to drop the table
ALTER PUBLICATION testpub_forschema DROP TABLE pub_test.testpub_nopk;
\dRp+ testpub_forschema
- Publication testpub_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test"
@@ -170,10 +173,10 @@ ERROR: relation "testpub_nopk" is not part of the publication
-- should be able to set table to schema publication
ALTER PUBLICATION testpub_forschema SET TABLE pub_test.testpub_nopk;
\dRp+ testpub_forschema
- Publication testpub_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"pub_test.testpub_nopk"
@@ -197,10 +200,10 @@ Not-null constraints:
"testpub_tbl2_id_not_null" NOT NULL "id"
\dRp+ testpub_foralltables
- Publication testpub_foralltables
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | t | t | t | f | f | f
+ Publication testpub_foralltables
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | t | f | t | t | f | f | f | f
(1 row)
DROP TABLE testpub_tbl2;
@@ -212,24 +215,525 @@ CREATE PUBLICATION testpub3 FOR TABLE testpub_tbl3;
CREATE PUBLICATION testpub4 FOR TABLE ONLY testpub_tbl3;
RESET client_min_messages;
\dRp+ testpub3
- Publication testpub3
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub3
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"public.testpub_tbl3"
"public.testpub_tbl3a"
\dRp+ testpub4
- Publication testpub4
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub4
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"public.testpub_tbl3"
DROP TABLE testpub_tbl3, testpub_tbl3a;
DROP PUBLICATION testpub3, testpub4;
+--- adding sequences
+CREATE SEQUENCE testpub_seq0;
+CREATE SEQUENCE pub_test.testpub_seq1;
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_forallsequences FOR ALL SEQUENCES WITH (publish = 'sequence');
+RESET client_min_messages;
+ALTER PUBLICATION testpub_forallsequences SET (publish = 'insert, sequence');
+CREATE SEQUENCE testpub_seq2;
+-- fail - can't add to for all sequences publication
+ALTER PUBLICATION testpub_forallsequences ADD SEQUENCE testpub_seq2;
+ERROR: publication "testpub_forallsequences" is defined as FOR ALL SEQUENCES
+DETAIL: Sequences cannot be added to or dropped from FOR ALL SEQUENCES publications.
+-- fail - can't drop from all sequences publication
+ALTER PUBLICATION testpub_forallsequences DROP SEQUENCE testpub_seq2;
+ERROR: publication "testpub_forallsequences" is defined as FOR ALL SEQUENCES
+DETAIL: Sequences cannot be added to or dropped from FOR ALL SEQUENCES publications.
+-- fail - can't add to for all sequences publication
+ALTER PUBLICATION testpub_forallsequences SET SEQUENCE pub_test.testpub_seq1;
+ERROR: publication "testpub_forallsequences" is defined as FOR ALL SEQUENCES
+DETAIL: Sequences cannot be added to or dropped from FOR ALL SEQUENCES publications.
+-- fail - can't add schema to 'FOR ALL SEQUENCES' publication
+ALTER PUBLICATION testpub_forallsequences ADD SEQUENCES IN SCHEMA pub_test;
+ERROR: publication "testpub_forallsequences" is defined as FOR ALL SEQUENCES
+DETAIL: Sequences from schema cannot be added to, dropped from, or set on FOR ALL SEQUENCES publications.
+-- fail - can't drop schema from 'FOR ALL SEQUENCES' publication
+ALTER PUBLICATION testpub_forallsequences DROP SEQUENCES IN SCHEMA pub_test;
+ERROR: publication "testpub_forallsequences" is defined as FOR ALL SEQUENCES
+DETAIL: Sequences from schema cannot be added to, dropped from, or set on FOR ALL SEQUENCES publications.
+-- fail - can't set schema to 'FOR ALL SEQUENCES' publication
+ALTER PUBLICATION testpub_forallsequences SET SEQUENCES IN SCHEMA pub_test;
+ERROR: publication "testpub_forallsequences" is defined as FOR ALL SEQUENCES
+DETAIL: Sequences from schema cannot be added to, dropped from, or set on FOR ALL SEQUENCES publications.
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_forsequence FOR SEQUENCE testpub_seq0;
+RESET client_min_messages;
+-- should be able to add schema to 'FOR SEQUENCE' publication
+ALTER PUBLICATION testpub_forsequence ADD SEQUENCES IN SCHEMA pub_test;
+\dRp+ testpub_forsequence
+ Publication testpub_forsequence
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Sequences:
+ "public.testpub_seq0"
+Sequences from schemas:
+ "pub_test"
+
+-- add sequence from the schema we already added
+ALTER PUBLICATION testpub_forsequence ADD SEQUENCE pub_test.testpub_seq1;
+-- fail - can't add sequence using ADD TABLE command
+ALTER PUBLICATION testpub_forsequence ADD TABLE pub_test.testpub_seq1;
+ERROR: object type does not match type expected by command
+-- should be able to drop schema from 'FOR SEQUENCE' publication
+ALTER PUBLICATION testpub_forsequence DROP SEQUENCES IN SCHEMA pub_test;
+\dRp+ testpub_forsequence
+ Publication testpub_forsequence
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Sequences:
+ "pub_test.testpub_seq1"
+ "public.testpub_seq0"
+
+-- should be able to set schema to 'FOR SEQUENCE' publication
+ALTER PUBLICATION testpub_forsequence SET SEQUENCES IN SCHEMA pub_test;
+\dRp+ testpub_forsequence
+ Publication testpub_forsequence
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Sequences from schemas:
+ "pub_test"
+
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_forschema FOR SEQUENCES IN SCHEMA pub_test;
+RESET client_min_messages;
+-- should be able to set sequence to schema publication
+ALTER PUBLICATION testpub_forschema SET SEQUENCE pub_test.testpub_seq1;
+\dRp+ testpub_forschema
+ Publication testpub_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Sequences:
+ "pub_test.testpub_seq1"
+
+SELECT pubname, puballtables, puballsequences FROM pg_publication WHERE pubname = 'testpub_forallsequences';
+ pubname | puballtables | puballsequences
+-------------------------+--------------+-----------------
+ testpub_forallsequences | f | t
+(1 row)
+
+\d+ pub_test.testpub_seq1
+ Sequence "pub_test.testpub_seq1"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+Publications:
+ "testpub_forallsequences"
+ "testpub_forschema"
+ "testpub_forsequence"
+
+\dRp+ testpub_forallsequences
+ Publication testpub_forallsequences
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | t | t | f | f | f | t | f
+(1 row)
+
+DROP SEQUENCE testpub_seq0, pub_test.testpub_seq1, testpub_seq2;
+DROP PUBLICATION testpub_forallsequences, testpub_forsequence, testpub_forschema;
+-- publication testing multiple sequences at the same time
+CREATE SEQUENCE testpub_seq1;
+CREATE SEQUENCE testpub_seq2;
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_multi FOR SEQUENCE testpub_seq1, testpub_seq2;
+RESET client_min_messages;
+\dRp+ testpub_multi
+ Publication testpub_multi
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Sequences:
+ "public.testpub_seq1"
+ "public.testpub_seq2"
+
+DROP PUBLICATION testpub_multi;
+DROP SEQUENCE testpub_seq1;
+DROP SEQUENCE testpub_seq2;
+-- Publication mixing tables and sequences
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_mix;
+RESET client_min_messages;
+CREATE SEQUENCE testpub_seq1;
+CREATE SEQUENCE pub_test.testpub_seq2;
+ALTER PUBLICATION testpub_mix ADD SEQUENCE testpub_seq1, TABLE testpub_tbl1;
+\dRp+ testpub_mix
+ Publication testpub_mix
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Tables:
+ "public.testpub_tbl1"
+Sequences:
+ "public.testpub_seq1"
+
+ALTER PUBLICATION testpub_mix ADD SEQUENCES IN SCHEMA pub_test, TABLES IN SCHEMA pub_test;
+\dRp+ testpub_mix
+ Publication testpub_mix
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Tables:
+ "public.testpub_tbl1"
+Tables from schemas:
+ "pub_test"
+Sequences:
+ "public.testpub_seq1"
+Sequences from schemas:
+ "pub_test"
+
+ALTER PUBLICATION testpub_mix DROP SEQUENCES IN SCHEMA pub_test;
+\dRp+ testpub_mix
+ Publication testpub_mix
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Tables:
+ "public.testpub_tbl1"
+Tables from schemas:
+ "pub_test"
+Sequences:
+ "public.testpub_seq1"
+
+ALTER PUBLICATION testpub_mix DROP TABLES IN SCHEMA pub_test;
+\dRp+ testpub_mix
+ Publication testpub_mix
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Tables:
+ "public.testpub_tbl1"
+Sequences:
+ "public.testpub_seq1"
+
+DROP PUBLICATION testpub_mix;
+DROP SEQUENCE testpub_seq1;
+DROP SEQUENCE pub_test.testpub_seq2;
+-- make sure we replicate only the correct relation type
+CREATE SCHEMA pub_test1;
+CREATE SEQUENCE pub_test1.test_seq1;
+CREATE TABLE pub_test1.test_tbl1 (a int primary key, b int);
+CREATE SCHEMA pub_test2;
+CREATE SEQUENCE pub_test2.test_seq2;
+CREATE TABLE pub_test2.test_tbl2 (a int primary key, b int);
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_schemas;
+RESET client_min_messages;
+-- add tables from one schema, sequences from the other
+ALTER PUBLICATION testpub_schemas ADD TABLES IN SCHEMA pub_test2;
+ALTER PUBLICATION testpub_schemas ADD SEQUENCES IN SCHEMA pub_test1;
+\dRp+ testpub_schemas
+ Publication testpub_schemas
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Tables from schemas:
+ "pub_test2"
+Sequences from schemas:
+ "pub_test1"
+
+\dn+ pub_test1
+ List of schemas
+ Name | Owner | Access privileges | Description
+-----------+--------------------------+-------------------+-------------
+ pub_test1 | regress_publication_user | |
+Publications:
+ "testpub_schemas" (sequences)
+
+\dn+ pub_test2
+ List of schemas
+ Name | Owner | Access privileges | Description
+-----------+--------------------------+-------------------+-------------
+ pub_test2 | regress_publication_user | |
+Publications:
+ "testpub_schemas" (tables)
+
+\d+ pub_test1.test_seq1;
+ Sequence "pub_test1.test_seq1"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test1.test_tbl1;
+ Table "pub_test1.test_tbl1"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl1_pkey" PRIMARY KEY, btree (a)
+
+\d+ pub_test2.test_seq2;
+ Sequence "pub_test2.test_seq2"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+
+\d+ pub_test2.test_tbl2;
+ Table "pub_test2.test_tbl2"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl2_pkey" PRIMARY KEY, btree (a)
+Publications:
+ "testpub_schemas"
+
+-- add the other object type from each schema
+ALTER PUBLICATION testpub_schemas ADD TABLES IN SCHEMA pub_test1;
+ALTER PUBLICATION testpub_schemas ADD SEQUENCES IN SCHEMA pub_test2;
+\dRp+ testpub_schemas
+ Publication testpub_schemas
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Tables from schemas:
+ "pub_test1"
+ "pub_test2"
+Sequences from schemas:
+ "pub_test1"
+ "pub_test2"
+
+\dn+ pub_test1
+ List of schemas
+ Name | Owner | Access privileges | Description
+-----------+--------------------------+-------------------+-------------
+ pub_test1 | regress_publication_user | |
+Publications:
+ "testpub_schemas" (sequences)
+ "testpub_schemas" (tables)
+
+\dn+ pub_test2
+ List of schemas
+ Name | Owner | Access privileges | Description
+-----------+--------------------------+-------------------+-------------
+ pub_test2 | regress_publication_user | |
+Publications:
+ "testpub_schemas" (tables)
+ "testpub_schemas" (sequences)
+
+\d+ pub_test1.test_seq1;
+ Sequence "pub_test1.test_seq1"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test1.test_tbl1;
+ Table "pub_test1.test_tbl1"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl1_pkey" PRIMARY KEY, btree (a)
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test2.test_seq2;
+ Sequence "pub_test2.test_seq2"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test2.test_tbl2;
+ Table "pub_test2.test_tbl2"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl2_pkey" PRIMARY KEY, btree (a)
+Publications:
+ "testpub_schemas"
+
+-- now drop the object type added first
+ALTER PUBLICATION testpub_schemas DROP TABLES IN SCHEMA pub_test2;
+ALTER PUBLICATION testpub_schemas DROP SEQUENCES IN SCHEMA pub_test1;
+\dRp+ testpub_schemas
+ Publication testpub_schemas
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Tables from schemas:
+ "pub_test1"
+Sequences from schemas:
+ "pub_test2"
+
+\dn+ pub_test1
+ List of schemas
+ Name | Owner | Access privileges | Description
+-----------+--------------------------+-------------------+-------------
+ pub_test1 | regress_publication_user | |
+Publications:
+ "testpub_schemas" (tables)
+
+\dn+ pub_test2
+ List of schemas
+ Name | Owner | Access privileges | Description
+-----------+--------------------------+-------------------+-------------
+ pub_test2 | regress_publication_user | |
+Publications:
+ "testpub_schemas" (sequences)
+
+\d+ pub_test1.test_seq1;
+ Sequence "pub_test1.test_seq1"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+
+\d+ pub_test1.test_tbl1;
+ Table "pub_test1.test_tbl1"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl1_pkey" PRIMARY KEY, btree (a)
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test2.test_seq2;
+ Sequence "pub_test2.test_seq2"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test2.test_tbl2;
+ Table "pub_test2.test_tbl2"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl2_pkey" PRIMARY KEY, btree (a)
+
+-- add a different schema (not including the already published sequences)
+ALTER PUBLICATION testpub_schemas ADD TABLE pub_test2.test_tbl2;
+ALTER PUBLICATION testpub_schemas ADD SEQUENCE pub_test1.test_seq1;
+\dRp+ testpub_schemas
+ Publication testpub_schemas
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Tables:
+ "pub_test2.test_tbl2"
+Tables from schemas:
+ "pub_test1"
+Sequences:
+ "pub_test1.test_seq1"
+Sequences from schemas:
+ "pub_test2"
+
+\d+ pub_test1.test_seq1;
+ Sequence "pub_test1.test_seq1"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test1.test_tbl1;
+ Table "pub_test1.test_tbl1"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl1_pkey" PRIMARY KEY, btree (a)
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test2.test_seq2;
+ Sequence "pub_test2.test_seq2"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test2.test_tbl2;
+ Table "pub_test2.test_tbl2"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl2_pkey" PRIMARY KEY, btree (a)
+Publications:
+ "testpub_schemas"
+
+-- now drop the explicitly added objects again
+ALTER PUBLICATION testpub_schemas DROP TABLE pub_test2.test_tbl2;
+ALTER PUBLICATION testpub_schemas DROP SEQUENCE pub_test1.test_seq1;
+\dRp+ testpub_schemas
+ Publication testpub_schemas
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
+Tables from schemas:
+ "pub_test1"
+Sequences from schemas:
+ "pub_test2"
+
+\d+ pub_test1.test_seq1;
+ Sequence "pub_test1.test_seq1"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+
+\d+ pub_test1.test_tbl1;
+ Table "pub_test1.test_tbl1"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl1_pkey" PRIMARY KEY, btree (a)
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test2.test_seq2;
+ Sequence "pub_test2.test_seq2"
+ Type | Start | Minimum | Maximum | Increment | Cycles? | Cache
+--------+-------+---------+---------------------+-----------+---------+-------
+ bigint | 1 | 1 | 9223372036854775807 | 1 | no | 1
+Publications:
+ "testpub_schemas"
+
+\d+ pub_test2.test_tbl2;
+ Table "pub_test2.test_tbl2"
+ Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
+--------+---------+-----------+----------+---------+---------+--------------+-------------
+ a | integer | | not null | | plain | |
+ b | integer | | | | plain | |
+Indexes:
+ "test_tbl2_pkey" PRIMARY KEY, btree (a)
+
+DROP PUBLICATION testpub_schemas;
+DROP TABLE pub_test1.test_tbl1, pub_test2.test_tbl2;
+DROP SEQUENCE pub_test1.test_seq1, pub_test2.test_seq2;
+DROP SCHEMA pub_test1, pub_test2;
-- Tests for partitioned tables
SET client_min_messages = 'ERROR';
CREATE PUBLICATION testpub_forparted;
@@ -245,10 +749,10 @@ UPDATE testpub_parted1 SET a = 1;
-- only parent is listed as being in publication, not the partition
ALTER PUBLICATION testpub_forparted ADD TABLE testpub_parted;
\dRp+ testpub_forparted
- Publication testpub_forparted
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_forparted
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"public.testpub_parted"
@@ -263,10 +767,10 @@ ALTER TABLE testpub_parted DETACH PARTITION testpub_parted1;
UPDATE testpub_parted1 SET a = 1;
ALTER PUBLICATION testpub_forparted SET (publish_via_partition_root = true);
\dRp+ testpub_forparted
- Publication testpub_forparted
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | t
+ Publication testpub_forparted
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | t
Tables:
"public.testpub_parted"
@@ -295,10 +799,10 @@ SET client_min_messages = 'ERROR';
CREATE PUBLICATION testpub5 FOR TABLE testpub_rf_tbl1, testpub_rf_tbl2 WHERE (c <> 'test' AND d < 5) WITH (publish = 'insert');
RESET client_min_messages;
\dRp+ testpub5
- Publication testpub5
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | f | f | f | f
+ Publication testpub5
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | f | f | f | f | f
Tables:
"public.testpub_rf_tbl1"
"public.testpub_rf_tbl2" WHERE ((c <> 'test'::text) AND (d < 5))
@@ -311,10 +815,10 @@ Tables:
ALTER PUBLICATION testpub5 ADD TABLE testpub_rf_tbl3 WHERE (e > 1000 AND e < 2000);
\dRp+ testpub5
- Publication testpub5
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | f | f | f | f
+ Publication testpub5
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | f | f | f | f | f
Tables:
"public.testpub_rf_tbl1"
"public.testpub_rf_tbl2" WHERE ((c <> 'test'::text) AND (d < 5))
@@ -330,10 +834,10 @@ Publications:
ALTER PUBLICATION testpub5 DROP TABLE testpub_rf_tbl2;
\dRp+ testpub5
- Publication testpub5
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | f | f | f | f
+ Publication testpub5
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | f | f | f | f | f
Tables:
"public.testpub_rf_tbl1"
"public.testpub_rf_tbl3" WHERE ((e > 1000) AND (e < 2000))
@@ -341,10 +845,10 @@ Tables:
-- remove testpub_rf_tbl1 and add testpub_rf_tbl3 again (another WHERE expression)
ALTER PUBLICATION testpub5 SET TABLE testpub_rf_tbl3 WHERE (e > 300 AND e < 500);
\dRp+ testpub5
- Publication testpub5
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | f | f | f | f
+ Publication testpub5
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | f | f | f | f | f
Tables:
"public.testpub_rf_tbl3" WHERE ((e > 300) AND (e < 500))
@@ -377,10 +881,10 @@ SET client_min_messages = 'ERROR';
CREATE PUBLICATION testpub_syntax1 FOR TABLE testpub_rf_tbl1, ONLY testpub_rf_tbl3 WHERE (e < 999) WITH (publish = 'insert');
RESET client_min_messages;
\dRp+ testpub_syntax1
- Publication testpub_syntax1
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | f | f | f | f
+ Publication testpub_syntax1
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | f | f | f | f | f
Tables:
"public.testpub_rf_tbl1"
"public.testpub_rf_tbl3" WHERE (e < 999)
@@ -390,10 +894,10 @@ SET client_min_messages = 'ERROR';
CREATE PUBLICATION testpub_syntax2 FOR TABLE testpub_rf_tbl1, testpub_rf_schema1.testpub_rf_tbl5 WHERE (h < 999) WITH (publish = 'insert');
RESET client_min_messages;
\dRp+ testpub_syntax2
- Publication testpub_syntax2
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | f | f | f | f
+ Publication testpub_syntax2
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | f | f | f | f | f
Tables:
"public.testpub_rf_tbl1"
"testpub_rf_schema1.testpub_rf_tbl5" WHERE (h < 999)
@@ -508,10 +1012,10 @@ CREATE PUBLICATION testpub6 FOR TABLES IN SCHEMA testpub_rf_schema2;
ALTER PUBLICATION testpub6 SET TABLES IN SCHEMA testpub_rf_schema2, TABLE testpub_rf_schema2.testpub_rf_tbl6 WHERE (i < 99);
RESET client_min_messages;
\dRp+ testpub6
- Publication testpub6
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub6
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"testpub_rf_schema2.testpub_rf_tbl6" WHERE (i < 99)
Tables from schemas:
@@ -725,10 +1229,10 @@ CREATE PUBLICATION testpub_table_ins WITH (publish = 'insert, truncate');
RESET client_min_messages;
ALTER PUBLICATION testpub_table_ins ADD TABLE testpub_tbl5 (a); -- ok
\dRp+ testpub_table_ins
- Publication testpub_table_ins
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | f | f | t | f
+ Publication testpub_table_ins
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | f | f | t | f | f
Tables:
"public.testpub_tbl5" (a)
@@ -902,10 +1406,10 @@ CREATE TABLE testpub_tbl_both_filters (a int, b int, c int, PRIMARY KEY (a,c));
ALTER TABLE testpub_tbl_both_filters REPLICA IDENTITY USING INDEX testpub_tbl_both_filters_pkey;
ALTER PUBLICATION testpub_both_filters ADD TABLE testpub_tbl_both_filters (a,c) WHERE (c != 1);
\dRp+ testpub_both_filters
- Publication testpub_both_filters
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_both_filters
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"public.testpub_tbl_both_filters" (a, c) WHERE (c <> 1)
@@ -1110,10 +1614,10 @@ ERROR: relation "testpub_tbl1" is already member of publication "testpub_fortbl
CREATE PUBLICATION testpub_fortbl FOR TABLE testpub_tbl1;
ERROR: publication "testpub_fortbl" already exists
\dRp+ testpub_fortbl
- Publication testpub_fortbl
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_fortbl
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"pub_test.testpub_nopk"
"public.testpub_tbl1"
@@ -1153,10 +1657,10 @@ Not-null constraints:
"testpub_tbl1_id_not_null" NOT NULL "id"
\dRp+ testpub_default
- Publication testpub_default
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | f | f
+ Publication testpub_default
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | f | t | f
Tables:
"pub_test.testpub_nopk"
"public.testpub_tbl1"
@@ -1236,10 +1740,10 @@ REVOKE CREATE ON DATABASE regression FROM regress_publication_user2;
DROP TABLE testpub_parted;
DROP TABLE testpub_tbl1;
\dRp+ testpub_default
- Publication testpub_default
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | f | f
+ Publication testpub_default
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | f | t | f
(1 row)
-- fail - must be owner of publication
@@ -1249,20 +1753,20 @@ ERROR: must be owner of publication testpub_default
RESET ROLE;
ALTER PUBLICATION testpub_default RENAME TO testpub_foo;
\dRp testpub_foo
- List of publications
- Name | Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
--------------+--------------------------+------------+---------+---------+---------+-----------+----------
- testpub_foo | regress_publication_user | f | t | t | t | f | f
+ List of publications
+ Name | Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+-------------+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ testpub_foo | regress_publication_user | f | f | t | t | t | f | t | f
(1 row)
-- rename back to keep the rest simple
ALTER PUBLICATION testpub_foo RENAME TO testpub_default;
ALTER PUBLICATION testpub_default OWNER TO regress_publication_user2;
\dRp testpub_default
- List of publications
- Name | Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
------------------+---------------------------+------------+---------+---------+---------+-----------+----------
- testpub_default | regress_publication_user2 | f | t | t | t | f | f
+ List of publications
+ Name | Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+-----------------+---------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ testpub_default | regress_publication_user2 | f | f | t | t | t | f | t | f
(1 row)
-- adding schemas and tables
@@ -1278,19 +1782,19 @@ CREATE TABLE "CURRENT_SCHEMA"."CURRENT_SCHEMA"(id int);
SET client_min_messages = 'ERROR';
CREATE PUBLICATION testpub1_forschema FOR TABLES IN SCHEMA pub_test1;
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
CREATE PUBLICATION testpub2_forschema FOR TABLES IN SCHEMA pub_test1, pub_test2, pub_test3;
\dRp+ testpub2_forschema
- Publication testpub2_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub2_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
"pub_test2"
@@ -1304,44 +1808,44 @@ CREATE PUBLICATION testpub6_forschema FOR TABLES IN SCHEMA "CURRENT_SCHEMA", CUR
CREATE PUBLICATION testpub_fortable FOR TABLE "CURRENT_SCHEMA"."CURRENT_SCHEMA";
RESET client_min_messages;
\dRp+ testpub3_forschema
- Publication testpub3_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub3_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"public"
\dRp+ testpub4_forschema
- Publication testpub4_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub4_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"CURRENT_SCHEMA"
\dRp+ testpub5_forschema
- Publication testpub5_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub5_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"CURRENT_SCHEMA"
"public"
\dRp+ testpub6_forschema
- Publication testpub6_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub6_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"CURRENT_SCHEMA"
"public"
\dRp+ testpub_fortable
- Publication testpub_fortable
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_fortable
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"CURRENT_SCHEMA.CURRENT_SCHEMA"
@@ -1375,10 +1879,10 @@ ERROR: schema "testpub_view" does not exist
-- dropping the schema should reflect the change in publication
DROP SCHEMA pub_test3;
\dRp+ testpub2_forschema
- Publication testpub2_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub2_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
"pub_test2"
@@ -1386,20 +1890,20 @@ Tables from schemas:
-- renaming the schema should reflect the change in publication
ALTER SCHEMA pub_test1 RENAME to pub_test1_renamed;
\dRp+ testpub2_forschema
- Publication testpub2_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub2_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1_renamed"
"pub_test2"
ALTER SCHEMA pub_test1_renamed RENAME to pub_test1;
\dRp+ testpub2_forschema
- Publication testpub2_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub2_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
"pub_test2"
@@ -1407,10 +1911,10 @@ Tables from schemas:
-- alter publication add schema
ALTER PUBLICATION testpub1_forschema ADD TABLES IN SCHEMA pub_test2;
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
"pub_test2"
@@ -1419,10 +1923,10 @@ Tables from schemas:
ALTER PUBLICATION testpub1_forschema ADD TABLES IN SCHEMA non_existent_schema;
ERROR: schema "non_existent_schema" does not exist
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
"pub_test2"
@@ -1431,10 +1935,10 @@ Tables from schemas:
ALTER PUBLICATION testpub1_forschema ADD TABLES IN SCHEMA pub_test1;
ERROR: schema "pub_test1" is already member of publication "testpub1_forschema"
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
"pub_test2"
@@ -1442,10 +1946,10 @@ Tables from schemas:
-- alter publication drop schema
ALTER PUBLICATION testpub1_forschema DROP TABLES IN SCHEMA pub_test2;
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
@@ -1453,10 +1957,10 @@ Tables from schemas:
ALTER PUBLICATION testpub1_forschema DROP TABLES IN SCHEMA pub_test2;
ERROR: tables from schema "pub_test2" are not part of the publication
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
@@ -1464,29 +1968,29 @@ Tables from schemas:
ALTER PUBLICATION testpub1_forschema DROP TABLES IN SCHEMA non_existent_schema;
ERROR: schema "non_existent_schema" does not exist
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
-- drop all schemas
ALTER PUBLICATION testpub1_forschema DROP TABLES IN SCHEMA pub_test1;
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
(1 row)
-- alter publication set multiple schema
ALTER PUBLICATION testpub1_forschema SET TABLES IN SCHEMA pub_test1, pub_test2;
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
"pub_test2"
@@ -1495,10 +1999,10 @@ Tables from schemas:
ALTER PUBLICATION testpub1_forschema SET TABLES IN SCHEMA non_existent_schema;
ERROR: schema "non_existent_schema" does not exist
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
"pub_test2"
@@ -1507,10 +2011,10 @@ Tables from schemas:
-- removing the duplicate schemas
ALTER PUBLICATION testpub1_forschema SET TABLES IN SCHEMA pub_test1, pub_test1;
\dRp+ testpub1_forschema
- Publication testpub1_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub1_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
@@ -1589,18 +2093,18 @@ SET client_min_messages = 'ERROR';
CREATE PUBLICATION testpub3_forschema;
RESET client_min_messages;
\dRp+ testpub3_forschema
- Publication testpub3_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub3_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
(1 row)
ALTER PUBLICATION testpub3_forschema SET TABLES IN SCHEMA pub_test1;
\dRp+ testpub3_forschema
- Publication testpub3_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub3_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables from schemas:
"pub_test1"
@@ -1610,20 +2114,20 @@ CREATE PUBLICATION testpub_forschema_fortable FOR TABLES IN SCHEMA pub_test1, TA
CREATE PUBLICATION testpub_fortable_forschema FOR TABLE pub_test2.tbl1, TABLES IN SCHEMA pub_test1;
RESET client_min_messages;
\dRp+ testpub_forschema_fortable
- Publication testpub_forschema_fortable
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_forschema_fortable
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"pub_test2.tbl1"
Tables from schemas:
"pub_test1"
\dRp+ testpub_fortable_forschema
- Publication testpub_fortable_forschema
- Owner | All tables | Inserts | Updates | Deletes | Truncates | Via root
---------------------------+------------+---------+---------+---------+-----------+----------
- regress_publication_user | f | t | t | t | t | f
+ Publication testpub_fortable_forschema
+ Owner | All tables | All sequences | Inserts | Updates | Deletes | Truncates | Sequences | Via root
+--------------------------+------------+---------------+---------+---------+---------+-----------+-----------+----------
+ regress_publication_user | f | f | t | t | t | t | t | f
Tables:
"pub_test2.tbl1"
Tables from schemas:
@@ -1667,40 +2171,85 @@ CREATE SCHEMA sch1;
CREATE SCHEMA sch2;
CREATE TABLE sch1.tbl1 (a int) PARTITION BY RANGE(a);
CREATE TABLE sch2.tbl1_part1 PARTITION OF sch1.tbl1 FOR VALUES FROM (1) to (10);
+CREATE SEQUENCE sch1.seq1;
+CREATE SEQUENCE sch2.seq2;
-- Schema publication that does not include the schema that has the parent table
CREATE PUBLICATION pub FOR TABLES IN SCHEMA sch2 WITH (PUBLISH_VIA_PARTITION_ROOT=1);
+ALTER PUBLICATION pub ADD SEQUENCES IN SCHEMA sch2;
SELECT * FROM pg_publication_tables;
pubname | schemaname | tablename | attnames | rowfilter
---------+------------+------------+----------+-----------
pub | sch2 | tbl1_part1 | {a} |
(1 row)
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch2 | seq2
+(1 row)
+
DROP PUBLICATION pub;
-- Table publication that does not include the parent table
CREATE PUBLICATION pub FOR TABLE sch2.tbl1_part1 WITH (PUBLISH_VIA_PARTITION_ROOT=1);
+ALTER PUBLICATION pub ADD SEQUENCE sch2.seq2;
SELECT * FROM pg_publication_tables;
pubname | schemaname | tablename | attnames | rowfilter
---------+------------+------------+----------+-----------
pub | sch2 | tbl1_part1 | {a} |
(1 row)
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch2 | seq2
+(1 row)
+
-- Table publication that includes both the parent table and the child table
ALTER PUBLICATION pub ADD TABLE sch1.tbl1;
+ALTER PUBLICATION pub ADD SEQUENCE sch1.seq1;
SELECT * FROM pg_publication_tables;
pubname | schemaname | tablename | attnames | rowfilter
---------+------------+-----------+----------+-----------
pub | sch1 | tbl1 | {a} |
(1 row)
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch1 | seq1
+ pub | sch2 | seq2
+(2 rows)
+
DROP PUBLICATION pub;
-- Schema publication that does not include the schema that has the parent table
CREATE PUBLICATION pub FOR TABLES IN SCHEMA sch2 WITH (PUBLISH_VIA_PARTITION_ROOT=0);
+ALTER PUBLICATION pub ADD SEQUENCE sch1.seq1;
SELECT * FROM pg_publication_tables;
pubname | schemaname | tablename | attnames | rowfilter
---------+------------+------------+----------+-----------
pub | sch2 | tbl1_part1 | {a} |
(1 row)
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch1 | seq1
+(1 row)
+
+DROP PUBLICATION pub;
+-- Sequence publication
+CREATE PUBLICATION pub FOR SEQUENCE sch2.seq2;
+SELECT * FROM pg_publication_tables;
+ pubname | schemaname | tablename | attnames | rowfilter
+---------+------------+-----------+----------+-----------
+(0 rows)
+
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch2 | seq2
+(1 row)
+
DROP PUBLICATION pub;
-- Table publication that does not include the parent table
CREATE PUBLICATION pub FOR TABLE sch2.tbl1_part1 WITH (PUBLISH_VIA_PARTITION_ROOT=0);
@@ -1710,14 +2259,26 @@ SELECT * FROM pg_publication_tables;
pub | sch2 | tbl1_part1 | {a} |
(1 row)
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+(0 rows)
+
-- Table publication that includes both the parent table and the child table
ALTER PUBLICATION pub ADD TABLE sch1.tbl1;
+ALTER PUBLICATION pub ADD SEQUENCES IN SCHEMA sch2;
SELECT * FROM pg_publication_tables;
pubname | schemaname | tablename | attnames | rowfilter
---------+------------+------------+----------+-----------
pub | sch2 | tbl1_part1 | {a} |
(1 row)
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch2 | seq2
+(1 row)
+
DROP PUBLICATION pub;
DROP TABLE sch2.tbl1_part1;
DROP TABLE sch1.tbl1;
@@ -1733,9 +2294,81 @@ SELECT * FROM pg_publication_tables;
pub | sch1 | tbl1 | {a} |
(1 row)
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+(0 rows)
+
+DROP PUBLICATION pub;
+-- Schema publication
+CREATE PUBLICATION pub FOR SEQUENCE sch2.seq2;
+SELECT * FROM pg_publication_tables;
+ pubname | schemaname | tablename | attnames | rowfilter
+---------+------------+-----------+----------+-----------
+(0 rows)
+
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch2 | seq2
+(1 row)
+
+DROP PUBLICATION pub;
+-- Sequence publication
+CREATE PUBLICATION pub FOR SEQUENCES IN SCHEMA sch2;
+SELECT * FROM pg_publication_tables;
+ pubname | schemaname | tablename | attnames | rowfilter
+---------+------------+-----------+----------+-----------
+(0 rows)
+
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch2 | seq2
+(1 row)
+
+ALTER PUBLICATION pub ADD SEQUENCE sch1.seq1;
+SELECT * FROM pg_publication_tables;
+ pubname | schemaname | tablename | attnames | rowfilter
+---------+------------+-----------+----------+-----------
+(0 rows)
+
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch1 | seq1
+ pub | sch2 | seq2
+(2 rows)
+
+ALTER PUBLICATION pub DROP SEQUENCE sch1.seq1;
+SELECT * FROM pg_publication_tables;
+ pubname | schemaname | tablename | attnames | rowfilter
+---------+------------+-----------+----------+-----------
+(0 rows)
+
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch2 | seq2
+(1 row)
+
+ALTER PUBLICATION pub ADD SEQUENCES IN SCHEMA sch1;
+SELECT * FROM pg_publication_tables;
+ pubname | schemaname | tablename | attnames | rowfilter
+---------+------------+-----------+----------+-----------
+(0 rows)
+
+SELECT * FROM pg_publication_sequences;
+ pubname | schemaname | sequencename
+---------+------------+--------------
+ pub | sch1 | seq1
+ pub | sch2 | seq2
+(2 rows)
+
RESET client_min_messages;
DROP PUBLICATION pub;
DROP TABLE sch1.tbl1;
+DROP SEQUENCE sch1.seq1, sch2.seq2;
DROP SCHEMA sch1 cascade;
DROP SCHEMA sch2 cascade;
RESET SESSION AUTHORIZATION;
diff --git a/src/test/regress/expected/rules.out b/src/test/regress/expected/rules.out
index d878a971df9..a03e6d93780 100644
--- a/src/test/regress/expected/rules.out
+++ b/src/test/regress/expected/rules.out
@@ -1441,6 +1441,14 @@ pg_prepared_xacts| SELECT p.transaction,
FROM ((pg_prepared_xact() p(transaction, gid, prepared, ownerid, dbid)
LEFT JOIN pg_authid u ON ((p.ownerid = u.oid)))
LEFT JOIN pg_database d ON ((p.dbid = d.oid)));
+pg_publication_sequences| SELECT p.pubname,
+ n.nspname AS schemaname,
+ c.relname AS sequencename
+ FROM pg_publication p,
+ LATERAL pg_get_publication_sequences((p.pubname)::text) gps(relid),
+ (pg_class c
+ JOIN pg_namespace n ON ((n.oid = c.relnamespace)))
+ WHERE (c.oid = gps.relid);
pg_publication_tables| SELECT p.pubname,
n.nspname AS schemaname,
c.relname AS tablename,
diff --git a/src/test/regress/sql/object_address.sql b/src/test/regress/sql/object_address.sql
index 1a6c61f49d5..5c6cdeb05df 100644
--- a/src/test/regress/sql/object_address.sql
+++ b/src/test/regress/sql/object_address.sql
@@ -50,6 +50,7 @@ CREATE TRANSFORM FOR int LANGUAGE SQL (
SET client_min_messages = 'ERROR';
CREATE PUBLICATION addr_pub FOR TABLE addr_nsp.gentable;
CREATE PUBLICATION addr_pub_schema FOR TABLES IN SCHEMA addr_nsp;
+CREATE PUBLICATION addr_pub_schema2 FOR SEQUENCES IN SCHEMA addr_nsp;
RESET client_min_messages;
CREATE SUBSCRIPTION regress_addr_sub CONNECTION '' PUBLICATION bar WITH (connect = false, slot_name = NONE);
CREATE STATISTICS addr_nsp.gentable_stat ON a, b FROM addr_nsp.gentable;
@@ -206,7 +207,8 @@ WITH objects (type, name, args) AS (VALUES
('transform', '{int}', '{sql}'),
('access method', '{btree}', '{}'),
('publication', '{addr_pub}', '{}'),
- ('publication namespace', '{addr_nsp}', '{addr_pub_schema}'),
+ ('publication namespace', '{addr_nsp}', '{addr_pub_schema, t}'),
+ ('publication namespace', '{addr_nsp}', '{addr_pub_schema2, s}'),
('publication relation', '{addr_nsp, gentable}', '{addr_pub}'),
('subscription', '{regress_addr_sub}', '{}'),
('statistics object', '{addr_nsp, gentable_stat}', '{}')
@@ -227,6 +229,7 @@ ORDER BY addr1.classid, addr1.objid, addr1.objsubid;
DROP FOREIGN DATA WRAPPER addr_fdw CASCADE;
DROP PUBLICATION addr_pub;
DROP PUBLICATION addr_pub_schema;
+DROP PUBLICATION addr_pub_schema2;
DROP SUBSCRIPTION regress_addr_sub;
DROP SCHEMA addr_nsp CASCADE;
diff --git a/src/test/regress/sql/publication.sql b/src/test/regress/sql/publication.sql
index d5051a5e746..016b58c67ec 100644
--- a/src/test/regress/sql/publication.sql
+++ b/src/test/regress/sql/publication.sql
@@ -27,7 +27,7 @@ CREATE PUBLICATION testpub_xxx WITH (publish_via_partition_root = 'true', publis
\dRp
-ALTER PUBLICATION testpub_default SET (publish = 'insert, update, delete');
+ALTER PUBLICATION testpub_default SET (publish = 'insert, update, delete, sequence');
\dRp
@@ -46,6 +46,8 @@ ALTER PUBLICATION testpub_foralltables SET (publish = 'insert, update');
CREATE TABLE testpub_tbl2 (id serial primary key, data text);
-- fail - can't add to for all tables publication
ALTER PUBLICATION testpub_foralltables ADD TABLE testpub_tbl2;
+-- fail - can't add a table using ADD SEQUENCE command
+ALTER PUBLICATION testpub_foralltables ADD SEQUENCE testpub_tbl2;
-- fail - can't drop from all tables publication
ALTER PUBLICATION testpub_foralltables DROP TABLE testpub_tbl2;
-- fail - can't add to for all tables publication
@@ -117,6 +119,188 @@ RESET client_min_messages;
DROP TABLE testpub_tbl3, testpub_tbl3a;
DROP PUBLICATION testpub3, testpub4;
+--- adding sequences
+CREATE SEQUENCE testpub_seq0;
+CREATE SEQUENCE pub_test.testpub_seq1;
+
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_forallsequences FOR ALL SEQUENCES WITH (publish = 'sequence');
+RESET client_min_messages;
+ALTER PUBLICATION testpub_forallsequences SET (publish = 'insert, sequence');
+
+CREATE SEQUENCE testpub_seq2;
+-- fail - can't add to for all sequences publication
+ALTER PUBLICATION testpub_forallsequences ADD SEQUENCE testpub_seq2;
+-- fail - can't drop from all sequences publication
+ALTER PUBLICATION testpub_forallsequences DROP SEQUENCE testpub_seq2;
+-- fail - can't add to for all sequences publication
+ALTER PUBLICATION testpub_forallsequences SET SEQUENCE pub_test.testpub_seq1;
+
+-- fail - can't add schema to 'FOR ALL SEQUENCES' publication
+ALTER PUBLICATION testpub_forallsequences ADD SEQUENCES IN SCHEMA pub_test;
+-- fail - can't drop schema from 'FOR ALL SEQUENCES' publication
+ALTER PUBLICATION testpub_forallsequences DROP SEQUENCES IN SCHEMA pub_test;
+-- fail - can't set schema to 'FOR ALL SEQUENCES' publication
+ALTER PUBLICATION testpub_forallsequences SET SEQUENCES IN SCHEMA pub_test;
+
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_forsequence FOR SEQUENCE testpub_seq0;
+RESET client_min_messages;
+-- should be able to add schema to 'FOR SEQUENCE' publication
+ALTER PUBLICATION testpub_forsequence ADD SEQUENCES IN SCHEMA pub_test;
+\dRp+ testpub_forsequence
+-- add sequence from the schema we already added
+ALTER PUBLICATION testpub_forsequence ADD SEQUENCE pub_test.testpub_seq1;
+-- fail - can't add sequence using ADD TABLE command
+ALTER PUBLICATION testpub_forsequence ADD TABLE pub_test.testpub_seq1;
+-- should be able to drop schema from 'FOR SEQUENCE' publication
+ALTER PUBLICATION testpub_forsequence DROP SEQUENCES IN SCHEMA pub_test;
+\dRp+ testpub_forsequence
+-- should be able to set schema to 'FOR SEQUENCE' publication
+ALTER PUBLICATION testpub_forsequence SET SEQUENCES IN SCHEMA pub_test;
+\dRp+ testpub_forsequence
+
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_forschema FOR SEQUENCES IN SCHEMA pub_test;
+RESET client_min_messages;
+-- should be able to set sequence to schema publication
+ALTER PUBLICATION testpub_forschema SET SEQUENCE pub_test.testpub_seq1;
+\dRp+ testpub_forschema
+
+SELECT pubname, puballtables, puballsequences FROM pg_publication WHERE pubname = 'testpub_forallsequences';
+\d+ pub_test.testpub_seq1
+\dRp+ testpub_forallsequences
+DROP SEQUENCE testpub_seq0, pub_test.testpub_seq1, testpub_seq2;
+DROP PUBLICATION testpub_forallsequences, testpub_forsequence, testpub_forschema;
+
+
+-- publication testing multiple sequences at the same time
+CREATE SEQUENCE testpub_seq1;
+CREATE SEQUENCE testpub_seq2;
+
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_multi FOR SEQUENCE testpub_seq1, testpub_seq2;
+RESET client_min_messages;
+
+\dRp+ testpub_multi
+
+DROP PUBLICATION testpub_multi;
+DROP SEQUENCE testpub_seq1;
+DROP SEQUENCE testpub_seq2;
+
+
+-- Publication mixing tables and sequences
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_mix;
+RESET client_min_messages;
+
+CREATE SEQUENCE testpub_seq1;
+CREATE SEQUENCE pub_test.testpub_seq2;
+
+ALTER PUBLICATION testpub_mix ADD SEQUENCE testpub_seq1, TABLE testpub_tbl1;
+\dRp+ testpub_mix
+
+ALTER PUBLICATION testpub_mix ADD SEQUENCES IN SCHEMA pub_test, TABLES IN SCHEMA pub_test;
+\dRp+ testpub_mix
+
+ALTER PUBLICATION testpub_mix DROP SEQUENCES IN SCHEMA pub_test;
+\dRp+ testpub_mix
+
+ALTER PUBLICATION testpub_mix DROP TABLES IN SCHEMA pub_test;
+\dRp+ testpub_mix
+
+DROP PUBLICATION testpub_mix;
+DROP SEQUENCE testpub_seq1;
+DROP SEQUENCE pub_test.testpub_seq2;
+
+
+-- make sure we replicate only the correct relation type
+CREATE SCHEMA pub_test1;
+CREATE SEQUENCE pub_test1.test_seq1;
+CREATE TABLE pub_test1.test_tbl1 (a int primary key, b int);
+
+CREATE SCHEMA pub_test2;
+CREATE SEQUENCE pub_test2.test_seq2;
+CREATE TABLE pub_test2.test_tbl2 (a int primary key, b int);
+
+SET client_min_messages = 'ERROR';
+CREATE PUBLICATION testpub_schemas;
+RESET client_min_messages;
+
+-- add tables from one schema, sequences from the other
+ALTER PUBLICATION testpub_schemas ADD TABLES IN SCHEMA pub_test2;
+ALTER PUBLICATION testpub_schemas ADD SEQUENCES IN SCHEMA pub_test1;
+
+\dRp+ testpub_schemas
+
+\dn+ pub_test1
+\dn+ pub_test2
+
+\d+ pub_test1.test_seq1;
+\d+ pub_test1.test_tbl1;
+
+\d+ pub_test2.test_seq2;
+\d+ pub_test2.test_tbl2;
+
+-- add the other object type from each schema
+ALTER PUBLICATION testpub_schemas ADD TABLES IN SCHEMA pub_test1;
+ALTER PUBLICATION testpub_schemas ADD SEQUENCES IN SCHEMA pub_test2;
+
+\dRp+ testpub_schemas
+
+\dn+ pub_test1
+\dn+ pub_test2
+
+\d+ pub_test1.test_seq1;
+\d+ pub_test1.test_tbl1;
+
+\d+ pub_test2.test_seq2;
+\d+ pub_test2.test_tbl2;
+
+-- now drop the object type added first
+ALTER PUBLICATION testpub_schemas DROP TABLES IN SCHEMA pub_test2;
+ALTER PUBLICATION testpub_schemas DROP SEQUENCES IN SCHEMA pub_test1;
+
+\dRp+ testpub_schemas
+
+\dn+ pub_test1
+\dn+ pub_test2
+
+\d+ pub_test1.test_seq1;
+\d+ pub_test1.test_tbl1;
+
+\d+ pub_test2.test_seq2;
+\d+ pub_test2.test_tbl2;
+
+-- add a different schema (not including the already published sequences)
+ALTER PUBLICATION testpub_schemas ADD TABLE pub_test2.test_tbl2;
+ALTER PUBLICATION testpub_schemas ADD SEQUENCE pub_test1.test_seq1;
+
+\dRp+ testpub_schemas
+
+\d+ pub_test1.test_seq1;
+\d+ pub_test1.test_tbl1;
+
+\d+ pub_test2.test_seq2;
+\d+ pub_test2.test_tbl2;
+
+-- now drop the explicitly added objects again
+ALTER PUBLICATION testpub_schemas DROP TABLE pub_test2.test_tbl2;
+ALTER PUBLICATION testpub_schemas DROP SEQUENCE pub_test1.test_seq1;
+
+\dRp+ testpub_schemas
+
+\d+ pub_test1.test_seq1;
+\d+ pub_test1.test_tbl1;
+
+\d+ pub_test2.test_seq2;
+\d+ pub_test2.test_tbl2;
+
+DROP PUBLICATION testpub_schemas;
+DROP TABLE pub_test1.test_tbl1, pub_test2.test_tbl2;
+DROP SEQUENCE pub_test1.test_seq1, pub_test2.test_seq2;
+DROP SCHEMA pub_test1, pub_test2;
+
-- Tests for partitioned tables
SET client_min_messages = 'ERROR';
CREATE PUBLICATION testpub_forparted;
@@ -1052,32 +1236,51 @@ CREATE SCHEMA sch1;
CREATE SCHEMA sch2;
CREATE TABLE sch1.tbl1 (a int) PARTITION BY RANGE(a);
CREATE TABLE sch2.tbl1_part1 PARTITION OF sch1.tbl1 FOR VALUES FROM (1) to (10);
+CREATE SEQUENCE sch1.seq1;
+CREATE SEQUENCE sch2.seq2;
-- Schema publication that does not include the schema that has the parent table
CREATE PUBLICATION pub FOR TABLES IN SCHEMA sch2 WITH (PUBLISH_VIA_PARTITION_ROOT=1);
+ALTER PUBLICATION pub ADD SEQUENCES IN SCHEMA sch2;
SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
DROP PUBLICATION pub;
-- Table publication that does not include the parent table
CREATE PUBLICATION pub FOR TABLE sch2.tbl1_part1 WITH (PUBLISH_VIA_PARTITION_ROOT=1);
+ALTER PUBLICATION pub ADD SEQUENCE sch2.seq2;
SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
-- Table publication that includes both the parent table and the child table
ALTER PUBLICATION pub ADD TABLE sch1.tbl1;
+ALTER PUBLICATION pub ADD SEQUENCE sch1.seq1;
SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
DROP PUBLICATION pub;
-- Schema publication that does not include the schema that has the parent table
CREATE PUBLICATION pub FOR TABLES IN SCHEMA sch2 WITH (PUBLISH_VIA_PARTITION_ROOT=0);
+ALTER PUBLICATION pub ADD SEQUENCE sch1.seq1;
+SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
+
+DROP PUBLICATION pub;
+-- Sequence publication
+CREATE PUBLICATION pub FOR SEQUENCE sch2.seq2;
SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
DROP PUBLICATION pub;
-- Table publication that does not include the parent table
CREATE PUBLICATION pub FOR TABLE sch2.tbl1_part1 WITH (PUBLISH_VIA_PARTITION_ROOT=0);
SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
-- Table publication that includes both the parent table and the child table
ALTER PUBLICATION pub ADD TABLE sch1.tbl1;
+ALTER PUBLICATION pub ADD SEQUENCES IN SCHEMA sch2;
SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
DROP PUBLICATION pub;
DROP TABLE sch2.tbl1_part1;
@@ -1090,10 +1293,36 @@ CREATE TABLE sch1.tbl1_part3 (a int) PARTITION BY RANGE(a);
ALTER TABLE sch1.tbl1 ATTACH PARTITION sch1.tbl1_part3 FOR VALUES FROM (20) to (30);
CREATE PUBLICATION pub FOR TABLES IN SCHEMA sch1 WITH (PUBLISH_VIA_PARTITION_ROOT=1);
SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
+
+DROP PUBLICATION pub;
+-- Schema publication
+CREATE PUBLICATION pub FOR SEQUENCE sch2.seq2;
+SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
+
+DROP PUBLICATION pub;
+-- Sequence publication
+CREATE PUBLICATION pub FOR SEQUENCES IN SCHEMA sch2;
+SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
+
+ALTER PUBLICATION pub ADD SEQUENCE sch1.seq1;
+SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
+
+ALTER PUBLICATION pub DROP SEQUENCE sch1.seq1;
+SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
+
+ALTER PUBLICATION pub ADD SEQUENCES IN SCHEMA sch1;
+SELECT * FROM pg_publication_tables;
+SELECT * FROM pg_publication_sequences;
RESET client_min_messages;
DROP PUBLICATION pub;
DROP TABLE sch1.tbl1;
+DROP SEQUENCE sch1.seq1, sch2.seq2;
DROP SCHEMA sch1 cascade;
DROP SCHEMA sch2 cascade;
diff --git a/src/test/subscription/t/034_sequences.pl b/src/test/subscription/t/034_sequences.pl
new file mode 100644
index 00000000000..2e9317642fc
--- /dev/null
+++ b/src/test/subscription/t/034_sequences.pl
@@ -0,0 +1,344 @@
+
+# Copyright (c) 2021, PostgreSQL Global Development Group
+
+# This tests that sequences are replicated correctly by logical replication
+use strict;
+use warnings;
+use PostgreSQL::Test::Cluster;
+use PostgreSQL::Test::Utils;
+use Test::More;
+
+# Initialize publisher node
+my $node_publisher = PostgreSQL::Test::Cluster->new('publisher');
+$node_publisher->init(allows_streaming => 'logical');
+$node_publisher->start;
+
+# Create subscriber node
+my $node_subscriber = PostgreSQL::Test::Cluster->new('subscriber');
+$node_subscriber->init(allows_streaming => 'logical');
+$node_subscriber->start;
+
+# Create some preexisting content on publisher
+my $ddl = qq(
+ CREATE TABLE seq_test (v BIGINT);
+ CREATE SEQUENCE s;
+);
+
+# Setup structure on the publisher
+$node_publisher->safe_psql('postgres', $ddl);
+
+# Create some the same structure on subscriber, and an extra sequence that
+# we'll create on the publisher later
+$ddl = qq(
+ CREATE TABLE seq_test (v BIGINT);
+ CREATE SEQUENCE s;
+ CREATE SEQUENCE s2;
+);
+
+$node_subscriber->safe_psql('postgres', $ddl);
+
+# Setup logical replication
+my $publisher_connstr = $node_publisher->connstr . ' dbname=postgres';
+$node_publisher->safe_psql('postgres',
+ "CREATE PUBLICATION seq_pub");
+
+$node_publisher->safe_psql('postgres',
+ "ALTER PUBLICATION seq_pub ADD SEQUENCE s");
+
+# function that alters the sequence in a completed subtransaction
+$node_publisher->safe_psql('postgres',
+ "CREATE FUNCTION regress_alter_sequence(s regclass, v integer) RETURNS void as \$\$
+BEGIN
+EXECUTE format('alter sequence %s restart %s', s, v);
+EXCEPTION WHEN others THEN NULL;
+END; \$\$ LANGUAGE plpgsql");
+
+$node_subscriber->safe_psql('postgres',
+ "CREATE SUBSCRIPTION seq_sub CONNECTION '$publisher_connstr' PUBLICATION seq_pub"
+);
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Wait for initial sync to finish as well
+my $synced_query =
+ "SELECT count(1) = 0 FROM pg_subscription_rel WHERE srsubstate NOT IN ('s', 'r');";
+$node_subscriber->poll_query_until('postgres', $synced_query)
+ or die "Timed out while waiting for subscriber to synchronize data";
+
+# Insert initial test data
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ -- generate a number of values using the sequence
+ INSERT INTO seq_test SELECT nextval('s') FROM generate_series(1,100);
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+my $result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s;
+));
+
+is( $result, '132|0|t',
+ 'initial test data replicated');
+
+
+# advance the sequence in a rolled-back transaction - the rollback
+# does not wait for the replication, so we could see any intermediate state
+# so do something else after the test, to ensure we wait for everything
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ INSERT INTO seq_test SELECT nextval('s') FROM generate_series(1,100);
+ ROLLBACK;
+ INSERT INTO seq_test VALUES (-1);
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s;
+));
+
+is( $result, '231|0|t',
+ 'advance sequence in rolled-back transaction');
+
+
+# create a new sequence and roll it back - should not be replicated, due to
+# the transactional behavior
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ CREATE SEQUENCE s2;
+ ALTER PUBLICATION seq_pub ADD SEQUENCE s2;
+ INSERT INTO seq_test SELECT nextval('s2') FROM generate_series(1,100);
+ ROLLBACK;
+));
+
+# Refresh publication after sequence is added to publication
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ ALTER SUBSCRIPTION seq_sub REFRESH PUBLICATION
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s2;
+));
+
+is( $result, '1|0|f',
+ 'create new sequence and roll it back');
+
+
+# create a new sequence, advance it in a rolled-back transaction, but commit
+# the create - the advance should be replicated nevertheless
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ ALTER SEQUENCE s RESTART WITH 1000;
+ SAVEPOINT sp1;
+ INSERT INTO seq_test SELECT nextval('s') FROM generate_series(1,100);
+ ROLLBACK TO sp1;
+ COMMIT;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Wait for sync of the second sequence we just added to finish
+$synced_query =
+ "SELECT count(1) = 0 FROM pg_subscription_rel WHERE srsubstate NOT IN ('s', 'r');";
+$node_subscriber->poll_query_until('postgres', $synced_query)
+ or die "Timed out while waiting for subscriber to synchronize data";
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s;
+));
+
+is( $result, '1131|0|t',
+ 'create sequence, advance it in rolled-back transaction, but commit the create');
+
+
+# advance the new sequence in a transaction, and roll it back - the rollback
+# does not wait for the replication, so we could see any intermediate state
+# so do something else after the test, to ensure we wait for everything
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ INSERT INTO seq_test SELECT nextval('s') FROM generate_series(1,100);
+ ROLLBACK;
+ INSERT INTO seq_test VALUES (-1);
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s;
+));
+
+is( $result, '1230|0|t',
+ 'advance the new sequence in a transaction and roll it back');
+
+
+# advance the sequence in a subtransaction - the subtransaction gets rolled
+# back, but commit the main one - the changes should still be replicated
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ SAVEPOINT s1;
+ INSERT INTO seq_test SELECT nextval('s') FROM generate_series(1,100);
+ ROLLBACK TO s1;
+ COMMIT;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s;
+));
+
+is( $result, '1329|0|t',
+ 'advance sequence in a subtransaction');
+
+
+# alter the sequence twice in the same transaction
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ ALTER SEQUENCE s RESTART 1000;
+ SAVEPOINT s1;
+ ALTER SEQUENCE s RESTART 2000;
+ INSERT INTO seq_test SELECT nextval('s') FROM generate_series(1,40);
+ COMMIT;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s;
+));
+
+is( $result, '2065|0|t',
+ 'alter sequence twice in a subtransaction');
+
+
+# alter the sequence twice in the same transaction, rollback the second change
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ ALTER SEQUENCE s RESTART 1000;
+ SAVEPOINT s1;
+ ALTER SEQUENCE s RESTART 2000;
+ ROLLBACK TO s1;
+ INSERT INTO seq_test SELECT nextval('s') FROM generate_series(1,40);
+ COMMIT;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s;
+));
+
+is( $result, '1065|0|t',
+ 'alter sequence twice in a subtransaction');
+
+
+# alter the sequence in a subtransaction that commits, followed by advancing
+# the sequence in a subtransaction that gets rolled back, but then the main
+# one commits - only the second set of changes should be replicated, as it
+# belongs to the alter, which was in the committed part
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ SELECT regress_alter_sequence('s', 1000); -- reset to 1000
+ INSERT INTO seq_test SELECT nextval('s') FROM generate_series(1,100);
+ SAVEPOINT s1;
+ SELECT regress_alter_sequence('s', 2000); -- reset to 2000
+ INSERT INTO seq_test SELECT nextval('s') FROM generate_series(1,100);
+ ROLLBACK TO s1;
+ INSERT INTO seq_test SELECT nextval('s') FROM generate_series(1,100);
+ COMMIT;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s;
+));
+
+is( $result, '1230|0|t',
+ 'alter sequence in a subtransaction');
+
+
+# alter the sequence in a subtransaction that commits, followed by advancing
+# the sequence in another subtransactio. the changes should be replicated, as
+# it belongs to the second alter, and both alters commit
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ SELECT regress_alter_sequence('s', 1000); -- reset to 1000
+ SAVEPOINT s1;
+ SELECT regress_alter_sequence('s', 2000); -- reset to 2000
+ INSERT INTO seq_test SELECT nextval('s') FROM generate_series(1,100);
+ COMMIT;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s;
+));
+
+is( $result, '2131|0|t',
+ 'alter sequence in a subtransaction');
+
+
+# alter the sequence in a subtransaction that commits, followed by advancing
+# the sequence in a subtransaction that gets rolled back, but then the main
+# one commits - the changes should still be replicated, because they belong
+# to the alter, which was in the committed part
+$node_publisher->safe_psql(
+ 'postgres', qq(
+ BEGIN;
+ SELECT regress_alter_sequence('s', 1000); -- reset to 1000
+ INSERT INTO seq_test SELECT nextval('s') FROM generate_series(1,100);
+ SAVEPOINT s1;
+ INSERT INTO seq_test SELECT nextval('s') FROM generate_series(1,100);
+ ROLLBACK TO s1;
+ INSERT INTO seq_test SELECT nextval('s') FROM generate_series(1,100);
+ COMMIT;
+));
+
+$node_publisher->wait_for_catchup('seq_sub');
+
+# Check the data on subscriber
+$result = $node_subscriber->safe_psql(
+ 'postgres', qq(
+ SELECT * FROM s;
+));
+
+is( $result, '1329|0|t',
+ 'alter sequence in a subtransaction');
+
+
+done_testing();
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index b632fbd7cf3..4a37dec12a0 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -1507,6 +1507,7 @@ LogicalRepRelId
LogicalRepRelMapEntry
LogicalRepRelation
LogicalRepRollbackPreparedTxnData
+LogicalRepSequence
LogicalRepStreamAbortData
LogicalRepTupleData
LogicalRepTyp
--
2.43.0
[text/x-patch] v20240111-0004-global-hash-table-of-sequence-relfilenodes.patch (12.1K, ../../[email protected]/5-v20240111-0004-global-hash-table-of-sequence-relfilenodes.patch)
download | inline diff:
From 4ecda5e6debda82946669f500ad14c018b0eae57 Mon Sep 17 00:00:00 2001
From: Tomas Vondra <[email protected]>
Date: Sat, 2 Dec 2023 10:28:26 +0100
Subject: [PATCH v20240111 4/7] global hash table of sequence relfilenodes
---
.../replication/logical/reorderbuffer.c | 177 +++++++-----------
src/include/replication/reorderbuffer.h | 6 +
2 files changed, 74 insertions(+), 109 deletions(-)
diff --git a/src/backend/replication/logical/reorderbuffer.c b/src/backend/replication/logical/reorderbuffer.c
index 6ac91068799..acaa6dce8e4 100644
--- a/src/backend/replication/logical/reorderbuffer.c
+++ b/src/backend/replication/logical/reorderbuffer.c
@@ -415,6 +415,14 @@ ReorderBufferAllocate(void)
dlist_init(&buffer->txns_by_base_snapshot_lsn);
dclist_init(&buffer->catchange_txns);
+ /* Create hash of sequences, mapping relfilelocator to transaction. */
+ hash_ctl.keysize = sizeof(RelFileLocator);
+ hash_ctl.entrysize = sizeof(ReorderBufferSequenceEnt);
+ hash_ctl.hcxt = buffer->context;
+
+ buffer->sequences_hash = hash_create("ReorderBufferSequenceHash", 32, &hash_ctl,
+ HASH_ELEM | HASH_BLOBS | HASH_CONTEXT);
+
/*
* Ensure there's no stale data from prior uses of this slot, in case some
* prior exit avoided calling ReorderBufferFree. Failure to do this can
@@ -1015,45 +1023,27 @@ ReorderBufferSequenceIsTransactional(ReorderBuffer *rb,
RelFileLocator rlocator,
TransactionId *xid)
{
+ ReorderBufferSequenceEnt *entry;
bool found = false;
- dlist_iter iter;
AssertCheckSequences(rb);
+ entry = hash_search(rb->sequences_hash,
+ (void *) &rlocator,
+ HASH_FIND,
+ &found);
+
/*
- * Walk all top-level transactions (some of which may be subxacts, except
- * that we haven't processed the assignments yet), and check if any of
- * them created the relfilenode.
+ * If we found an entry with matching relfilenode, we're done - we
+ * have to treat the sequence change as transactional, and replay it
+ * in the same (sub)transaction just like any other change.
+ *
+ * Optionally set XID of the (sub)xact that created the relfilenode.
*/
- dlist_foreach(iter, &rb->toplevel_by_lsn)
+ if (found)
{
- ReorderBufferSequenceEnt *entry;
- ReorderBufferTXN *txn = dlist_container(ReorderBufferTXN, node,
- iter.cur);
-
- /* transaction has no relfilenodes at all */
- if (!txn->sequences_hash)
- continue;
-
- entry = hash_search(txn->sequences_hash,
- (void *) &rlocator,
- HASH_FIND,
- &found);
-
- /*
- * If we found an entry with matching relfilenode, we're done - we
- * have to treat the sequence change as transactional, and replay it
- * in the same (sub)transaction just like any other change.
- *
- * Optionally set XID of the (sub)xact that created the relfilenode.
- */
- if (found)
- {
- if (xid)
- *xid = entry->txn->xid;
-
- break;
- }
+ if (xid)
+ *xid = entry->txn->xid;
}
return found;
@@ -1068,27 +1058,22 @@ ReorderBufferSequenceIsTransactional(ReorderBuffer *rb,
* the committed/aborted one.
*/
static void
-ReorderBufferSequenceCleanup(ReorderBufferTXN *txn)
+ReorderBufferSequenceCleanup(ReorderBuffer *rb, ReorderBufferTXN *txn)
{
HASH_SEQ_STATUS scan_status;
ReorderBufferSequenceEnt *ent;
- /* Bail out if not a subxact, or if there are no entries. */
- if (!rbtxn_is_known_subxact(txn))
- return;
-
+ /* Bail out if there are no sequence entries for this xact. */
if (!txn->sequences_hash)
return;
/*
- * Scan the top-level transaction hash and remove the entries from it. If
- * we have entries for subxact, the top-level hash must have been
- * initialized.
+ * Scan the global transaction hash and remove the entries from it.
*/
hash_seq_init(&scan_status, txn->sequences_hash);
while ((ent = (ReorderBufferSequenceEnt *) hash_seq_search(&scan_status)) != NULL)
{
- if (hash_search(txn->toptxn->sequences_hash,
+ if (hash_search(rb->sequences_hash,
(void *) &ent->rlocator,
HASH_REMOVE, NULL) == NULL)
elog(ERROR, "hash table corrupted");
@@ -1286,28 +1271,23 @@ ReorderBufferAddRelFileLocator(ReorderBuffer *rb, TransactionId xid,
txn = ReorderBufferTXNByXid(rb, xid, true, NULL, lsn, true);
/*
- * First add it to the top-level transaction, but make sure it links to
+ * First add it to the global hash table, and make sure it links to
* the correct subtransaction (so that we add later changes to it).
*/
- if (txn->toptxn)
- {
- /* make sure the hash table is initialized */
- ReorderBufferTXNSequencesInit(rb, txn->toptxn);
+ entry = hash_search(rb->sequences_hash,
+ (void *) &rlocator,
+ HASH_ENTER,
+ &found);
- /* search the lookup table */
- entry = hash_search(txn->toptxn->sequences_hash,
- (void *) &rlocator,
- HASH_ENTER,
- &found);
+ /*
+ * We've just decoded creation of the relfilenode, so if we found it
+ * in the hash table, something is wrong.
+ */
+ Assert(!found);
- /*
- * We've just decoded creation of the relfilenode, so if we found it
- * in the hash table, something is wrong.
- */
- Assert(!found);
+ entry->txn = txn;
- entry->txn = txn;
- }
+ /* now add it to the hash table in the transaction itself */
/* make sure the hash table is initialized */
ReorderBufferTXNSequencesInit(rb, txn);
@@ -1417,7 +1397,6 @@ AssertCheckSequences(ReorderBuffer *rb)
{
#ifdef USE_ASSERT_CHECKING
LogicalDecodingContext *ctx = rb->private_data;
- dlist_iter iter;
/*
* Skip the verification if we don't reach the LSN at which we start
@@ -1433,57 +1412,53 @@ AssertCheckSequences(ReorderBuffer *rb)
return;
/*
- * Make sure the relfilenodes from subxacts are properly recorded in the
- * top-level transaction hash table.
+ * Make sure the relfilenodes from all xacts are properly recorded in the
+ * global hash table of sequences.
*/
- dlist_foreach(iter, &rb->toplevel_by_lsn)
{
+ HASH_SEQ_STATUS hash_seq;
+ ReorderBufferTXNByIdEnt *ent;
int nentries = 0,
nsubentries = 0;
- dlist_iter subiter;
- ReorderBufferTXN *txn = dlist_container(ReorderBufferTXN, node,
- iter.cur);
/*
- * We don't skip top-level transactions without relfilenodes, because
- * there might be a subtransaction with some, and we want to detect
+ * We do this check even if the global hash table is empty, because
+ * there might be a transaction with some, and we want to detect
* such cases too.
*/
- if (txn->sequences_hash)
- nentries = hash_get_num_entries(txn->sequences_hash);
+ nentries = hash_get_num_entries(rb->sequences_hash);
- /* walk all subxacts, and check the hash table in each one */
- dlist_foreach(subiter, &txn->subtxns)
+ /* walk all xacts, and check the hash table in each one */
+ hash_seq_init(&hash_seq, rb->by_txn);
+ while ((ent = hash_seq_search(&hash_seq)) != NULL)
{
- HASH_SEQ_STATUS scan_status;
+ ReorderBufferTXN *txn = ent->txn;
ReorderBufferSequenceEnt *entry;
-
- ReorderBufferTXN *subtxn = dlist_container(ReorderBufferTXN, node,
- subiter.cur);
+ HASH_SEQ_STATUS scan_status;
/*
* If this subxact has no relfilenodes, skip it. We'll do the
- * check in the opposite direction (that all top-level
- * relfilenodes are in the correct subxact) later.
+ * check in the opposite direction (that all relfilenodes are
+ * also in hash for the correct xact) later.
*/
- if (!subtxn->sequences_hash)
+ if (!txn->sequences_hash)
continue;
- /* add number of relfilenodes in this subxact */
- nsubentries += hash_get_num_entries(subtxn->sequences_hash);
+ /* account for relfilenodes in this xact */
+ nsubentries += hash_get_num_entries(txn->sequences_hash);
/*
* Check that all subxact relfilenodes are in the top-level txn
* too, and are pointing to this subtransaction.
*/
- hash_seq_init(&scan_status, subtxn->sequences_hash);
+ hash_seq_init(&scan_status, txn->sequences_hash);
while ((entry = (ReorderBufferSequenceEnt *) hash_seq_search(&scan_status)) != NULL)
{
bool found = false;
ReorderBufferSequenceEnt *subentry;
- /* search for the same relfilenode in the top-level txn */
- subentry = hash_search(txn->sequences_hash,
+ /* search for the same relfilenode in the global hash table */
+ subentry = hash_search(rb->sequences_hash,
(void *) &entry->rlocator,
HASH_FIND,
&found);
@@ -1498,7 +1473,7 @@ AssertCheckSequences(ReorderBuffer *rb)
* The entry has to point to the subxact - there's no subxact
* "below" this one to which the relfilenode could belong.
*/
- Assert(subentry->txn == subtxn);
+ Assert(subentry->txn == txn);
}
}
@@ -1508,32 +1483,25 @@ AssertCheckSequences(ReorderBuffer *rb)
* directly, so this needs to be inequality.
*/
Assert(nentries >= nsubentries);
+ }
+ {
/*
* Now do the check in the opposite direction - check that every entry
- * in the top-level txn (except those pointing to the top-level txn
- * itself) point to one of the subxacts, and there's an entry in the
- * subxact hash.
+ * in the global hash table point to one of the running xacts, and
+ * there's an entry in the transaction hash table.
*/
- if (txn->sequences_hash)
+ if (rb->sequences_hash)
{
HASH_SEQ_STATUS scan_status;
ReorderBufferSequenceEnt *entry;
- hash_seq_init(&scan_status, txn->sequences_hash);
+ hash_seq_init(&scan_status, rb->sequences_hash);
while ((entry = (ReorderBufferSequenceEnt *) hash_seq_search(&scan_status)) != NULL)
{
bool found = false;
ReorderBufferSequenceEnt *subentry;
- /* Skip entries for the top-level txn itself. */
- if (entry->txn == txn)
- continue;
-
- /* Is it a subxact of this txn? */
- Assert(rbtxn_is_known_subxact(entry->txn));
- Assert(entry->txn->toptxn == txn);
-
/*
* Search for the same relfilenode in the subxact (it should
* be initialized, as we expect it to contain the
@@ -2181,6 +2149,9 @@ ReorderBufferCleanupTXN(ReorderBuffer *rb, ReorderBufferTXN *txn)
if (rbtxn_is_serialized(txn))
ReorderBufferRestoreCleanup(rb, txn);
+ /* remove the sequence relfilenode from the global hash table */
+ ReorderBufferSequenceCleanup(rb, txn);
+
/* deallocate */
ReorderBufferReturnTXN(rb, txn);
}
@@ -3519,9 +3490,6 @@ ReorderBufferFinishPrepared(ReorderBuffer *rb, TransactionId xid,
ReorderBufferExecuteInvalidations(txn->ninvalidations,
txn->invalidations);
- /* cleanup: remove sequence relfilenodes from the top-level txn */
- ReorderBufferSequenceCleanup(txn);
-
ReorderBufferCleanupTXN(rb, txn);
}
@@ -3571,9 +3539,6 @@ ReorderBufferAbort(ReorderBuffer *rb, TransactionId xid, XLogRecPtr lsn,
/* cosmetic... */
txn->final_lsn = lsn;
- /* remove sequence relfilenodes from the top-level txn */
- ReorderBufferSequenceCleanup(txn);
-
/* remove potential on-disk data, and deallocate */
ReorderBufferCleanupTXN(rb, txn);
@@ -3613,9 +3578,6 @@ ReorderBufferAbortOld(ReorderBuffer *rb, TransactionId oldestRunningXid)
if (rbtxn_is_streamed(txn))
rb->stream_abort(rb, txn, InvalidXLogRecPtr);
- /* remove sequence relfilenodes from the top-level txn */
- ReorderBufferSequenceCleanup(txn);
-
/* remove potential on-disk data, and deallocate this tx */
ReorderBufferCleanupTXN(rb, txn);
@@ -3668,9 +3630,6 @@ ReorderBufferForget(ReorderBuffer *rb, TransactionId xid, XLogRecPtr lsn)
else
Assert(txn->ninvalidations == 0);
- /* remove sequence relfilenodes from a top-level txn */
- ReorderBufferSequenceCleanup(txn);
-
/* remove potential on-disk data, and deallocate */
ReorderBufferCleanupTXN(rb, txn);
}
diff --git a/src/include/replication/reorderbuffer.h b/src/include/replication/reorderbuffer.h
index 0134133517a..ad206b3f7d3 100644
--- a/src/include/replication/reorderbuffer.h
+++ b/src/include/replication/reorderbuffer.h
@@ -704,6 +704,12 @@ struct ReorderBuffer
*/
int64 totalTxns; /* total number of transactions sent */
int64 totalBytes; /* total amount of data decoded */
+
+ /*
+ * Sequence relfilenodes created in any transaction (also includes
+ * altered sequences, which assigns new relfilenode).
+ */
+ HTAB *sequences_hash;
};
--
2.43.0
[text/x-patch] v20240111-0005-CREATE-SUBSCRIPTION-flag-to-enable-sequenc.patch (19.0K, ../../[email protected]/6-v20240111-0005-CREATE-SUBSCRIPTION-flag-to-enable-sequenc.patch)
download | inline diff:
From 58fd0bd09edfa2814e163e75e05449946ad1795b Mon Sep 17 00:00:00 2001
From: Tomas Vondra <[email protected]>
Date: Wed, 10 Jan 2024 17:22:07 +0100
Subject: [PATCH v20240111 5/7] CREATE SUBSCRIPTION flag to enable sequences
---
contrib/test_decoding/test_decoding.c | 13 ++----
doc/src/sgml/logicaldecoding.sgml | 8 ++++
doc/src/sgml/ref/alter_subscription.sgml | 5 ++-
doc/src/sgml/ref/create_subscription.sgml | 12 ++++++
src/backend/catalog/pg_subscription.c | 1 +
src/backend/commands/subscriptioncmds.c | 43 ++++++++++++++++---
.../libpqwalreceiver/libpqwalreceiver.c | 4 ++
src/backend/replication/logical/logical.c | 20 +++++++++
src/backend/replication/logical/worker.c | 10 +++++
src/backend/replication/pgoutput/pgoutput.c | 14 ++++++
src/include/catalog/pg_subscription.h | 4 ++
src/include/replication/output_plugin.h | 1 +
src/include/replication/pgoutput.h | 1 +
src/include/replication/walreceiver.h | 1 +
src/test/subscription/t/034_sequences.pl | 2 +-
15 files changed, 121 insertions(+), 18 deletions(-)
diff --git a/contrib/test_decoding/test_decoding.c b/contrib/test_decoding/test_decoding.c
index 200caad6863..072bbba575b 100644
--- a/contrib/test_decoding/test_decoding.c
+++ b/contrib/test_decoding/test_decoding.c
@@ -29,7 +29,6 @@ typedef struct
MemoryContext context;
bool include_xids;
bool include_timestamp;
- bool include_sequences;
bool skip_empty_xacts;
bool only_local;
} TestDecodingData;
@@ -179,12 +178,12 @@ pg_decode_startup(LogicalDecodingContext *ctx, OutputPluginOptions *opt,
data->include_timestamp = false;
data->skip_empty_xacts = false;
data->only_local = false;
- data->include_sequences = false;
ctx->output_plugin_private = data;
opt->output_type = OUTPUT_PLUGIN_TEXTUAL_OUTPUT;
opt->receive_rewrites = false;
+ opt->receive_sequences = false;
foreach(option, ctx->output_plugin_options)
{
@@ -275,8 +274,8 @@ pg_decode_startup(LogicalDecodingContext *ctx, OutputPluginOptions *opt,
{
if (elem->arg == NULL)
- data->include_sequences = true;
- else if (!parse_bool(strVal(elem->arg), &data->include_sequences))
+ opt->receive_sequences = true;
+ else if (!parse_bool(strVal(elem->arg), &opt->receive_sequences))
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("could not parse value \"%s\" for parameter \"%s\"",
@@ -794,9 +793,6 @@ pg_decode_sequence(LogicalDecodingContext *ctx, ReorderBufferTXN *txn,
TestDecodingData *data = ctx->output_plugin_private;
TestDecodingTxnData *txndata = txn->output_plugin_private;
- if (!data->include_sequences)
- return;
-
/* output BEGIN if we haven't yet, but only for the transactional case */
if (transactional)
{
@@ -1038,9 +1034,6 @@ pg_decode_stream_sequence(LogicalDecodingContext *ctx, ReorderBufferTXN *txn,
TestDecodingData *data = ctx->output_plugin_private;
TestDecodingTxnData *txndata = txn->output_plugin_private;
- if (!data->include_sequences)
- return;
-
/* output BEGIN if we haven't yet, but only for the transactional case */
if (transactional)
{
diff --git a/doc/src/sgml/logicaldecoding.sgml b/doc/src/sgml/logicaldecoding.sgml
index 90e99a95128..31f8aa7f932 100644
--- a/doc/src/sgml/logicaldecoding.sgml
+++ b/doc/src/sgml/logicaldecoding.sgml
@@ -646,6 +646,7 @@ typedef struct OutputPluginOptions
{
OutputPluginOutputType output_type;
bool receive_rewrites;
+ bool receive_sequences;
} OutputPluginOptions;
</programlisting>
<literal>output_type</literal> has to either be set to
@@ -658,6 +659,13 @@ typedef struct OutputPluginOptions
replication, but they require special handling.
</para>
+ <para>
+ If <literal>receive_sequences</literal> is true, the output plugin will
+ also be called for changes made by operations on sequences. These are
+ of interest to plugins that need to maintain sequences consistent with
+ the rest of the data, for example for purposes of upgrade or failover.
+ </para>
+
<para>
The startup callback should validate the options present in
<literal>ctx->output_plugin_options</literal>. If the output plugin
diff --git a/doc/src/sgml/ref/alter_subscription.sgml b/doc/src/sgml/ref/alter_subscription.sgml
index 479ec495896..2d24ff5768e 100644
--- a/doc/src/sgml/ref/alter_subscription.sgml
+++ b/doc/src/sgml/ref/alter_subscription.sgml
@@ -226,8 +226,9 @@ ALTER SUBSCRIPTION <replaceable class="parameter">name</replaceable> RENAME TO <
<link linkend="sql-createsubscription-params-with-streaming"><literal>streaming</literal></link>,
<link linkend="sql-createsubscription-params-with-disable-on-error"><literal>disable_on_error</literal></link>,
<link linkend="sql-createsubscription-params-with-password-required"><literal>password_required</literal></link>,
- <link linkend="sql-createsubscription-params-with-run-as-owner"><literal>run_as_owner</literal></link>, and
- <link linkend="sql-createsubscription-params-with-origin"><literal>origin</literal></link>.
+ <link linkend="sql-createsubscription-params-with-run-as-owner"><literal>run_as_owner</literal></link>,
+ <link linkend="sql-createsubscription-params-with-origin"><literal>origin</literal></link> and
+ <link linkend="sql-createsubscription-params-with-sequences"><literal>sequences</literal></link>.
Only a superuser can set <literal>password_required = false</literal>.
</para>
</listitem>
diff --git a/doc/src/sgml/ref/create_subscription.sgml b/doc/src/sgml/ref/create_subscription.sgml
index f1c20b3a465..4591e37dac7 100644
--- a/doc/src/sgml/ref/create_subscription.sgml
+++ b/doc/src/sgml/ref/create_subscription.sgml
@@ -399,6 +399,18 @@ CREATE SUBSCRIPTION <replaceable class="parameter">subscription_name</replaceabl
</para>
</listitem>
</varlistentry>
+
+ <varlistentry id="sql-createsubscription-params-with-sequences">
+ <term><literal>sequences</literal> (<type>boolean</type>)</term>
+ <listitem>
+ <para>
+ Specifies whether the subscription will request the publisher to
+ decode and send sequence changes. Note that for sequences to be
+ decoded, the sequence also has to be added to the publication.
+ The default is <literal>false</literal>.
+ </para>
+ </listitem>
+ </varlistentry>
</variablelist></para>
</listitem>
diff --git a/src/backend/catalog/pg_subscription.c b/src/backend/catalog/pg_subscription.c
index c516c25ac7b..e52abf61791 100644
--- a/src/backend/catalog/pg_subscription.c
+++ b/src/backend/catalog/pg_subscription.c
@@ -73,6 +73,7 @@ GetSubscription(Oid subid, bool missing_ok)
sub->disableonerr = subform->subdisableonerr;
sub->passwordrequired = subform->subpasswordrequired;
sub->runasowner = subform->subrunasowner;
+ sub->sequences = subform->subsequences;
/* Get conninfo */
datum = SysCacheGetAttrNotNull(SUBSCRIPTIONOID,
diff --git a/src/backend/commands/subscriptioncmds.c b/src/backend/commands/subscriptioncmds.c
index 144609c1b9f..76b1bbd9b2f 100644
--- a/src/backend/commands/subscriptioncmds.c
+++ b/src/backend/commands/subscriptioncmds.c
@@ -71,6 +71,7 @@
#define SUBOPT_RUN_AS_OWNER 0x00001000
#define SUBOPT_LSN 0x00002000
#define SUBOPT_ORIGIN 0x00004000
+#define SUBOPT_SEQUENCES 0x00008000
/* check if the 'val' has 'bits' set */
#define IsSet(val, bits) (((val) & (bits)) == (bits))
@@ -97,6 +98,7 @@ typedef struct SubOpts
bool runasowner;
char *origin;
XLogRecPtr lsn;
+ bool sequences;
} SubOpts;
static List *fetch_table_list(WalReceiverConn *wrconn, List *publications);
@@ -158,6 +160,8 @@ parse_subscription_options(ParseState *pstate, List *stmt_options,
opts->runasowner = false;
if (IsSet(supported_opts, SUBOPT_ORIGIN))
opts->origin = pstrdup(LOGICALREP_ORIGIN_ANY);
+ if (IsSet(supported_opts, SUBOPT_SEQUENCES))
+ opts->sequences = false;
/* Parse options */
foreach(lc, stmt_options)
@@ -354,6 +358,15 @@ parse_subscription_options(ParseState *pstate, List *stmt_options,
opts->specified_opts |= SUBOPT_LSN;
opts->lsn = lsn;
}
+ else if (IsSet(supported_opts, SUBOPT_SEQUENCES) &&
+ strcmp(defel->defname, "sequences") == 0)
+ {
+ if (IsSet(opts->specified_opts, SUBOPT_SEQUENCES))
+ errorConflictingDefElem(defel, pstate);
+
+ opts->specified_opts |= SUBOPT_SEQUENCES;
+ opts->sequences = defGetBoolean(defel);
+ }
else
ereport(ERROR,
(errcode(ERRCODE_SYNTAX_ERROR),
@@ -592,7 +605,7 @@ CreateSubscription(ParseState *pstate, CreateSubscriptionStmt *stmt,
SUBOPT_SYNCHRONOUS_COMMIT | SUBOPT_BINARY |
SUBOPT_STREAMING | SUBOPT_TWOPHASE_COMMIT |
SUBOPT_DISABLE_ON_ERR | SUBOPT_PASSWORD_REQUIRED |
- SUBOPT_RUN_AS_OWNER | SUBOPT_ORIGIN);
+ SUBOPT_RUN_AS_OWNER | SUBOPT_ORIGIN | SUBOPT_SEQUENCES);
parse_subscription_options(pstate, stmt->options, supported_opts, &opts);
/*
@@ -698,6 +711,7 @@ CreateSubscription(ParseState *pstate, CreateSubscriptionStmt *stmt,
values[Anum_pg_subscription_subdisableonerr - 1] = BoolGetDatum(opts.disableonerr);
values[Anum_pg_subscription_subpasswordrequired - 1] = BoolGetDatum(opts.passwordrequired);
values[Anum_pg_subscription_subrunasowner - 1] = BoolGetDatum(opts.runasowner);
+ values[Anum_pg_subscription_subsequences - 1] = BoolGetDatum(opts.sequences);
values[Anum_pg_subscription_subconninfo - 1] =
CStringGetTextDatum(conninfo);
if (opts.slot_name)
@@ -762,8 +776,14 @@ CreateSubscription(ParseState *pstate, CreateSubscriptionStmt *stmt,
* relation sync status info.
*/
relations = fetch_table_list(wrconn, publications);
- relations = list_concat(relations,
- fetch_sequence_list(wrconn, publications));
+
+ /*
+ * Add sequences, but only if the subscription explicitly enabled
+ * them to be replicated.
+ */
+ if (opts.sequences)
+ relations = list_concat(relations,
+ fetch_sequence_list(wrconn, publications));
foreach(lc, relations)
{
@@ -888,8 +908,14 @@ AlterSubscription_refresh(Subscription *sub, bool copy_data,
/* Get the list of relations from publisher. */
pubrel_names = fetch_table_list(wrconn, sub->publications);
- pubrel_names = list_concat(pubrel_names,
- fetch_sequence_list(wrconn, sub->publications));
+
+ /*
+ * Add sequences, but only if the subscription explicitly enabled
+ * them to be replicated.
+ */
+ if (sub->sequences)
+ pubrel_names = list_concat(pubrel_names,
+ fetch_sequence_list(wrconn, sub->publications));
/* Get local table list. */
subrel_states = GetSubscriptionRelations(sub->oid, false);
@@ -1224,6 +1250,13 @@ AlterSubscription(ParseState *pstate, AlterSubscriptionStmt *stmt,
replaces[Anum_pg_subscription_suborigin - 1] = true;
}
+ if (IsSet(opts.specified_opts, SUBOPT_SEQUENCES))
+ {
+ values[Anum_pg_subscription_subsequences - 1] =
+ BoolGetDatum(opts.sequences);
+ replaces[Anum_pg_subscription_subsequences - 1] = true;
+ }
+
update_tuple = true;
break;
}
diff --git a/src/backend/replication/libpqwalreceiver/libpqwalreceiver.c b/src/backend/replication/libpqwalreceiver/libpqwalreceiver.c
index 78344a03615..07a0e55c21d 100644
--- a/src/backend/replication/libpqwalreceiver/libpqwalreceiver.c
+++ b/src/backend/replication/libpqwalreceiver/libpqwalreceiver.c
@@ -475,6 +475,10 @@ libpqrcv_startstreaming(WalReceiverConn *conn,
appendStringInfo(&cmd, ", origin '%s'",
options->proto.logical.origin);
+ if (options->proto.logical.sequences &&
+ PQserverVersion(conn->streamConn) >= 170000)
+ appendStringInfo(&cmd, ", sequences 'on'");
+
pubnames = options->proto.logical.publication_names;
pubnames_str = stringlist_to_identifierstr(conn->streamConn, pubnames);
if (!pubnames_str)
diff --git a/src/backend/replication/logical/logical.c b/src/backend/replication/logical/logical.c
index 8cadc0e9e6f..86557745e18 100644
--- a/src/backend/replication/logical/logical.c
+++ b/src/backend/replication/logical/logical.c
@@ -475,6 +475,16 @@ CreateInitDecodingContext(const char *plugin,
ctx->reorder->output_rewrites = ctx->options.receive_rewrites;
+ /*
+ * StartupDecodingContext only checks which callbacks are defined, and
+ * startup_cb_wrapper sets receiver_sequences flag. We need to combine
+ * those two pieces and disable sequences if receiver_sequences=false.
+ *
+ * XXX We could also leave this up to the plugin startup, but it seems
+ * cleaner to just do it here.
+ */
+ ctx->sequences &= ctx->options.receive_sequences;
+
return ctx;
}
@@ -622,6 +632,16 @@ CreateDecodingContext(XLogRecPtr start_lsn,
ctx->reorder->output_rewrites = ctx->options.receive_rewrites;
+ /*
+ * StartupDecodingContext only checks which callbacks are defined, and
+ * startup_cb_wrapper sets receiver_sequences flag. We need to combine
+ * those two pieces and disable sequences if receiver_sequences=false.
+ *
+ * XXX We could also leave this up to the plugin startup, but it seems
+ * cleaner to just do it here.
+ */
+ ctx->sequences &= ctx->options.receive_sequences;
+
ereport(LOG,
(errmsg("starting logical decoding for slot \"%s\"",
NameStr(slot->data.name)),
diff --git a/src/backend/replication/logical/worker.c b/src/backend/replication/logical/worker.c
index 01fe9527b6b..6a71db354a3 100644
--- a/src/backend/replication/logical/worker.c
+++ b/src/backend/replication/logical/worker.c
@@ -4040,6 +4040,7 @@ maybe_reread_subscription(void)
strcmp(newsub->slotname, MySubscription->slotname) != 0 ||
newsub->binary != MySubscription->binary ||
newsub->stream != MySubscription->stream ||
+ newsub->sequences != MySubscription->sequences ||
newsub->passwordrequired != MySubscription->passwordrequired ||
strcmp(newsub->origin, MySubscription->origin) != 0 ||
newsub->owner != MySubscription->owner ||
@@ -4470,6 +4471,15 @@ set_stream_options(WalRcvStreamOptions *options,
options->proto.logical.publication_names = MySubscription->publications;
options->proto.logical.binary = MySubscription->binary;
+ /*
+ * FIXME maybe this should depend on server_version too? We don't want
+ * to request sequences from old releases, but maybe we should not just
+ * ignore that? Might easily lead to surprises.
+ *
+ * XXX Now there's a server_version check in libpqrcv_startstreaming.
+ */
+ options->proto.logical.sequences = MySubscription->sequences;
+
/*
* Assign the appropriate option value for streaming option according to
* the 'streaming' mode and the publisher's ability to support that mode.
diff --git a/src/backend/replication/pgoutput/pgoutput.c b/src/backend/replication/pgoutput/pgoutput.c
index 8b8f81c5f8e..5a19306d26c 100644
--- a/src/backend/replication/pgoutput/pgoutput.c
+++ b/src/backend/replication/pgoutput/pgoutput.c
@@ -291,6 +291,7 @@ parse_output_parameters(List *options, PGOutputData *data)
bool streaming_given = false;
bool two_phase_option_given = false;
bool origin_option_given = false;
+ bool sequences_option_given = false;
data->binary = false;
data->streaming = LOGICALREP_STREAM_OFF;
@@ -404,6 +405,16 @@ parse_output_parameters(List *options, PGOutputData *data)
errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("unrecognized origin value: \"%s\"", origin));
}
+ else if (strcmp(defel->defname, "sequences") == 0)
+ {
+ if (sequences_option_given)
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("conflicting or redundant options")));
+ sequences_option_given = true;
+
+ data->sequences = defGetBoolean(defel);
+ }
else
elog(ERROR, "unrecognized pgoutput option: %s", defel->defname);
}
@@ -442,6 +453,7 @@ pgoutput_startup(LogicalDecodingContext *ctx, OutputPluginOptions *opt,
/* This plugin uses binary protocol. */
opt->output_type = OUTPUT_PLUGIN_BINARY_OUTPUT;
+ opt->receive_sequences = false;
/*
* This is replication start and not slot initialization.
@@ -540,6 +552,8 @@ pgoutput_startup(LogicalDecodingContext *ctx, OutputPluginOptions *opt,
ctx->streaming = false;
ctx->twophase = false;
}
+
+ opt->receive_sequences = data->sequences;
}
/*
diff --git a/src/include/catalog/pg_subscription.h b/src/include/catalog/pg_subscription.h
index ca326255852..7205f9ed13b 100644
--- a/src/include/catalog/pg_subscription.h
+++ b/src/include/catalog/pg_subscription.h
@@ -93,6 +93,9 @@ CATALOG(pg_subscription,6100,SubscriptionRelationId) BKI_SHARED_RELATION BKI_ROW
bool subrunasowner; /* True if replication should execute as the
* subscription owner */
+ bool subsequences; /* True if replication stream should contain
+ * decoded sequence changes */
+
#ifdef CATALOG_VARLEN /* variable-length fields start here */
/* Connection string to the publisher */
text subconninfo BKI_FORCE_NOT_NULL;
@@ -139,6 +142,7 @@ typedef struct Subscription
* occurs */
bool passwordrequired; /* Must connection use a password? */
bool runasowner; /* Run replication as subscription owner */
+ bool sequences; /* Request sequences from the publisher */
char *conninfo; /* Connection string to the publisher */
char *slotname; /* Name of the replication slot */
char *synccommit; /* Synchronous commit setting for worker */
diff --git a/src/include/replication/output_plugin.h b/src/include/replication/output_plugin.h
index 2ae25c6f4fe..f9ca2994fa7 100644
--- a/src/include/replication/output_plugin.h
+++ b/src/include/replication/output_plugin.h
@@ -27,6 +27,7 @@ typedef struct OutputPluginOptions
{
OutputPluginOutputType output_type;
bool receive_rewrites;
+ bool receive_sequences;
} OutputPluginOptions;
/*
diff --git a/src/include/replication/pgoutput.h b/src/include/replication/pgoutput.h
index 89f94e11472..a7461277a00 100644
--- a/src/include/replication/pgoutput.h
+++ b/src/include/replication/pgoutput.h
@@ -33,6 +33,7 @@ typedef struct PGOutputData
bool messages;
bool two_phase;
bool publish_no_origin;
+ bool sequences;
} PGOutputData;
#endif /* PGOUTPUT_H */
diff --git a/src/include/replication/walreceiver.h b/src/include/replication/walreceiver.h
index 0899891cdb8..c0526fd2cd3 100644
--- a/src/include/replication/walreceiver.h
+++ b/src/include/replication/walreceiver.h
@@ -186,6 +186,7 @@ typedef struct
* prepare time */
char *origin; /* Only publish data originating from the
* specified origin */
+ bool sequences; /* Request sequences from publisher */
} logical;
} proto;
} WalRcvStreamOptions;
diff --git a/src/test/subscription/t/034_sequences.pl b/src/test/subscription/t/034_sequences.pl
index 2e9317642fc..ca706631203 100644
--- a/src/test/subscription/t/034_sequences.pl
+++ b/src/test/subscription/t/034_sequences.pl
@@ -54,7 +54,7 @@ EXCEPTION WHEN others THEN NULL;
END; \$\$ LANGUAGE plpgsql");
$node_subscriber->safe_psql('postgres',
- "CREATE SUBSCRIPTION seq_sub CONNECTION '$publisher_connstr' PUBLICATION seq_pub"
+ "CREATE SUBSCRIPTION seq_sub CONNECTION '$publisher_connstr' PUBLICATION seq_pub WITH (sequences = true)"
);
$node_publisher->wait_for_catchup('seq_sub');
--
2.43.0
[text/x-patch] v20240111-0006-add-comment-about-SNAPBUILD_FULL.patch (1.9K, ../../[email protected]/7-v20240111-0006-add-comment-about-SNAPBUILD_FULL.patch)
download | inline diff:
From 2d96af7b89008d906ffc2a76ca906e0f7eac0a68 Mon Sep 17 00:00:00 2001
From: Tomas Vondra <[email protected]>
Date: Thu, 11 Jan 2024 14:35:29 +0100
Subject: [PATCH v20240111 6/7] add comment about SNAPBUILD_FULL
---
src/backend/replication/logical/decode.c | 17 ++++++++++++++++-
1 file changed, 16 insertions(+), 1 deletion(-)
diff --git a/src/backend/replication/logical/decode.c b/src/backend/replication/logical/decode.c
index 3462c545bb7..95f224df4a7 100644
--- a/src/backend/replication/logical/decode.c
+++ b/src/backend/replication/logical/decode.c
@@ -1432,12 +1432,27 @@ seq_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
* change as transactional and queue it. Otherwise it needs to be treated
* as non-transactional, in which case we just send it to the plugin right
* away.
+ *
+ * XXX This may incorrectly identify some changes as non-transactional,
+ * if the relfilenode was created before the snapshot switched to FULL
+ * (so that smgr_decode skipped it, and it's not in the hash table).
+ * But we still handle this correctly, because the snapshot can't be
+ * in CONSISTENT state yet, so the following check will skip the change
+ * too, and we won't try to apply it (which would fail due to unknown
+ * relfilenode).
*/
transactional = ReorderBufferSequenceIsTransactional(ctx->reorder,
target_locator,
NULL);
- /* Skip the change if already processed (per the snapshot). */
+ /*
+ * Skip the change if already processed (per the snapshot).
+ *
+ * XXX It's important to skip all non-transactional changes with snapshot
+ * that is not consistent yet. The sequence change may be incorrectly
+ * identified as non-transactional (see above), and we must not attempt
+ * to process it.
+ */
if (transactional &&
!SnapBuildProcessChange(builder, xid, buf->origptr))
return;
--
2.43.0
[text/x-patch] v20240111-0007-decode-all-sequence-relfilenodes.patch (4.7K, ../../[email protected]/8-v20240111-0007-decode-all-sequence-relfilenodes.patch)
download | inline diff:
From 8f1598ac7795d70a9baecc3492b2930dcd828f35 Mon Sep 17 00:00:00 2001
From: Tomas Vondra <[email protected]>
Date: Thu, 11 Jan 2024 15:48:44 +0100
Subject: [PATCH v20240111 7/7] decode all sequence relfilenodes
---
src/backend/replication/logical/decode.c | 31 ++++---------------
.../replication/logical/reorderbuffer.c | 26 ++++++++++++++--
2 files changed, 29 insertions(+), 28 deletions(-)
diff --git a/src/backend/replication/logical/decode.c b/src/backend/replication/logical/decode.c
index 95f224df4a7..a141f444772 100644
--- a/src/backend/replication/logical/decode.c
+++ b/src/backend/replication/logical/decode.c
@@ -1432,27 +1432,12 @@ seq_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
* change as transactional and queue it. Otherwise it needs to be treated
* as non-transactional, in which case we just send it to the plugin right
* away.
- *
- * XXX This may incorrectly identify some changes as non-transactional,
- * if the relfilenode was created before the snapshot switched to FULL
- * (so that smgr_decode skipped it, and it's not in the hash table).
- * But we still handle this correctly, because the snapshot can't be
- * in CONSISTENT state yet, so the following check will skip the change
- * too, and we won't try to apply it (which would fail due to unknown
- * relfilenode).
*/
transactional = ReorderBufferSequenceIsTransactional(ctx->reorder,
target_locator,
NULL);
- /*
- * Skip the change if already processed (per the snapshot).
- *
- * XXX It's important to skip all non-transactional changes with snapshot
- * that is not consistent yet. The sequence change may be incorrectly
- * identified as non-transactional (see above), and we must not attempt
- * to process it.
- */
+ /* Skip the change if already processed (per the snapshot). */
if (transactional &&
!SnapBuildProcessChange(builder, xid, buf->origptr))
return;
@@ -1516,11 +1501,15 @@ seq_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
* We only care about XLOG_SMGR_CREATE records for "our" database (logical
* decoding is restricted to a single database), and we do the filtering
* and skipping, as appropriate.
+ *
+ * Note that the relfilenode is decoded and added to the reorder buffer even
+ * before the snapshot switches to SNAPBUILD_FULL_SNAPSHOT. This is necessary
+ * so that seq_decode can correctly identify transactional changes when the
+ * state changes to FULL after the relfilenode is created.
*/
void
smgr_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
{
- SnapBuild *builder = ctx->snapshot_builder;
XLogReaderState *r = buf->record;
TransactionId xid = XLogRecGetXid(r);
uint8 info = XLogRecGetInfo(buf->record) & ~XLR_INFO_MASK;
@@ -1542,14 +1531,6 @@ smgr_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
ReorderBufferProcessXid(ctx->reorder, XLogRecGetXid(r), buf->origptr);
- /*
- * If we don't have snapshot or we are just fast-forwarding, there is no
- * point in decoding relfilenode information.
- */
- if (SnapBuildCurrentState(builder) < SNAPBUILD_FULL_SNAPSHOT ||
- ctx->fast_forward)
- return;
-
/* only interested in our database */
xlrec = (xl_smgr_create *) XLogRecGetData(r);
if (xlrec->rlocator.dbOid != ctx->slot->data.database)
diff --git a/src/backend/replication/logical/reorderbuffer.c b/src/backend/replication/logical/reorderbuffer.c
index acaa6dce8e4..017fd9c09d7 100644
--- a/src/backend/replication/logical/reorderbuffer.c
+++ b/src/backend/replication/logical/reorderbuffer.c
@@ -1280,10 +1280,30 @@ ReorderBufferAddRelFileLocator(ReorderBuffer *rb, TransactionId xid,
&found);
/*
- * We've just decoded creation of the relfilenode, so if we found it
- * in the hash table, something is wrong.
+ * We've just decoded creation of the relfilenode, but it's possible a
+ * the transaction aborted without triggering a reorderbuffer cleanup,
+ * and the relfilenode was allocated again. While quite unlikely (the
+ * hash table entry should be cleaned when processing RUNNING_XACTS,
+ * and we should not reuse the relfilenode this fast).
+ *
+ * So if we find an entry for the relfilenode, we simply point it to
+ * the new transaction, but we need to remove it from the hash table
+ * in the original transaction (otherwise it'd confuse cleanup).
*/
- Assert(!found);
+ if (found)
+ {
+ /*
+ * The entry must be pointing to a transaction, and there must be
+ * must be a hash table of relfilenodes.
+ */
+ Assert(entry->txn);
+ Assert(entry->txn->sequences_hash);
+
+ if (hash_search(entry->txn->sequences_hash,
+ (void *) &entry->rlocator,
+ HASH_REMOVE, NULL) == NULL)
+ elog(ERROR, "hash table of sequence relfilenode corrupted");
+ }
entry->txn = txn;
--
2.43.0
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2024-01-23 20:47 Robert Haas <[email protected]>
parent: Tomas Vondra <[email protected]>
0 siblings, 1 reply; 127+ messages in thread
From: Robert Haas @ 2024-01-23 20:47 UTC (permalink / raw)
To: Tomas Vondra <[email protected]>; +Cc: Amit Kapila <[email protected]>; Ashutosh Bapat <[email protected]>; Dilip Kumar <[email protected]>; Hayato Kuroda (Fujitsu) <[email protected]>; Zhijie Hou (Fujitsu) <[email protected]>; PostgreSQL Hackers <[email protected]>; Masahiko Sawada <[email protected]>; Peter Eisentraut <[email protected]>
On Thu, Jan 11, 2024 at 11:27 AM Tomas Vondra
<[email protected]> wrote:
> 1) desirability: We want a built-in way to handle sequences in logical
> replication. I think everyone agrees this is not a way to do distributed
> sequences in an active-active setups, but that there are other use cases
> that need this feature - typically upgrades / logical failover.
Yeah. I find it extremely hard to take seriously the idea that this
isn't a valuable feature. How else are you supposed to do a logical
failover without having your entire application break?
> 2) performance: There was concern about the performance impact, and that
> it affects everyone, including those who don't replicate sequences (as
> the overhead is mostly incurred before calls to output plugin etc.).
>
> The agreement was that the best way is to have a CREATE SUBSCRIPTION
> option that would instruct the upstream to decode sequences. By default
> this option is 'off' (because that's the no-overhead case), but it can
> be enabled for each subscription.
Seems reasonable, at least unless and until we come up with something better.
> 3) correctness: The last point is about making "transactional" flag
> correct when the snapshot state changes mid-transaction, originally
> pointed out by Dilip [4]. Per [5] this however happens to work
> correctly, because while we identify the change as 'non-transactional'
> (which is incorrect), we immediately throw it again (so we don't try to
> apply it, which would error-out).
I've said this before, but I still find this really scary. It's
unclear to me that we can simply classify updates as transactional or
non-transactional and expect things to work. If it's possible, I hope
we have a really good explanation somewhere of how and why it's
possible. If we do, can somebody point me to it so I can read it?
To be possibly slightly more clear about my concern, I think the scary
case is where we have transactional and non-transactional things
happening to the same sequence in close temporal proximity, either
within the same session or across two or more sessions. If a
non-transactional change can get reordered ahead of some transactional
change upon which it logically depends, or behind some transactional
change that logically depends on it, then we have trouble. I also
wonder if there are any cases where the same operation is partly
transactional and partly non-transactional.
--
Robert Haas
EDB: http://www.enterprisedb.com
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2024-01-24 17:46 Tomas Vondra <[email protected]>
parent: Robert Haas <[email protected]>
0 siblings, 1 reply; 127+ messages in thread
From: Tomas Vondra @ 2024-01-24 17:46 UTC (permalink / raw)
To: Robert Haas <[email protected]>; +Cc: Amit Kapila <[email protected]>; Ashutosh Bapat <[email protected]>; Dilip Kumar <[email protected]>; Hayato Kuroda (Fujitsu) <[email protected]>; Zhijie Hou (Fujitsu) <[email protected]>; PostgreSQL Hackers <[email protected]>; Masahiko Sawada <[email protected]>; Peter Eisentraut <[email protected]>
On 1/23/24 21:47, Robert Haas wrote:
> On Thu, Jan 11, 2024 at 11:27 AM Tomas Vondra
> <[email protected]> wrote:
>> 1) desirability: We want a built-in way to handle sequences in logical
>> replication. I think everyone agrees this is not a way to do distributed
>> sequences in an active-active setups, but that there are other use cases
>> that need this feature - typically upgrades / logical failover.
>
> Yeah. I find it extremely hard to take seriously the idea that this
> isn't a valuable feature. How else are you supposed to do a logical
> failover without having your entire application break?
>
>> 2) performance: There was concern about the performance impact, and that
>> it affects everyone, including those who don't replicate sequences (as
>> the overhead is mostly incurred before calls to output plugin etc.).
>>
>> The agreement was that the best way is to have a CREATE SUBSCRIPTION
>> option that would instruct the upstream to decode sequences. By default
>> this option is 'off' (because that's the no-overhead case), but it can
>> be enabled for each subscription.
>
> Seems reasonable, at least unless and until we come up with something better.
>
>> 3) correctness: The last point is about making "transactional" flag
>> correct when the snapshot state changes mid-transaction, originally
>> pointed out by Dilip [4]. Per [5] this however happens to work
>> correctly, because while we identify the change as 'non-transactional'
>> (which is incorrect), we immediately throw it again (so we don't try to
>> apply it, which would error-out).
>
> I've said this before, but I still find this really scary. It's
> unclear to me that we can simply classify updates as transactional or
> non-transactional and expect things to work. If it's possible, I hope
> we have a really good explanation somewhere of how and why it's
> possible. If we do, can somebody point me to it so I can read it?
>
I did try to explain how this works (and why) in a couple places:
1) the commit message
2) reorderbuffer header comment
3) ReorderBufferSequenceIsTransactional comment (and nearby)
It's possible this does not meet your expectations, ofc. Maybe there
should be a separate README for this - I haven't found anything like
that for logical decoding in general, which is why I did (1)-(3).
> To be possibly slightly more clear about my concern, I think the scary
> case is where we have transactional and non-transactional things
> happening to the same sequence in close temporal proximity, either
> within the same session or across two or more sessions. If a
> non-transactional change can get reordered ahead of some transactional
> change upon which it logically depends, or behind some transactional
> change that logically depends on it, then we have trouble. I also
> wonder if there are any cases where the same operation is partly
> transactional and partly non-transactional.
>
I certainly understand this concern, and to some extent I even share it.
Having to differentiate between transactional and non-transactional
changes certainly confused me more than once. It's especially confusing,
because the decoding implicitly changes the perceived ordering/atomicity
of the events.
That being said, I don't think it get reordered the way you're concerned
about. The "transactionality" is determined by relfilenode change, so
how could the reordering happen? We'd have to misidentify change in
either direction - and for nontransactional->transactional change that's
clearly not possible. There has to be a new relfilenode in that xact.
In the other direction (transactional->nontransactional), it can happen
if we fail to decode the relfilenode record. Which is what we discussed
earlier, but came to the conclusion that it actually works OK.
Of course, there might be bugs. I spent quite a bit of effort reviewing
and testing this, but there still might be something wrong. But I think
that applies to any feature.
What would be worse is some sort of thinko in the approach in general. I
don't have a good answer to that, unfortunately - I think it works, but
how would I know for sure? We explored multiple alternative approaches
and all of them crashed and burned ...
regards
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2024-01-26 14:39 Robert Haas <[email protected]>
parent: Tomas Vondra <[email protected]>
0 siblings, 1 reply; 127+ messages in thread
From: Robert Haas @ 2024-01-26 14:39 UTC (permalink / raw)
To: Tomas Vondra <[email protected]>; +Cc: Amit Kapila <[email protected]>; Ashutosh Bapat <[email protected]>; Dilip Kumar <[email protected]>; Hayato Kuroda (Fujitsu) <[email protected]>; Zhijie Hou (Fujitsu) <[email protected]>; PostgreSQL Hackers <[email protected]>; Masahiko Sawada <[email protected]>; Peter Eisentraut <[email protected]>
On Wed, Jan 24, 2024 at 12:46 PM Tomas Vondra
<[email protected]> wrote:
> I did try to explain how this works (and why) in a couple places:
>
> 1) the commit message
> 2) reorderbuffer header comment
> 3) ReorderBufferSequenceIsTransactional comment (and nearby)
>
> It's possible this does not meet your expectations, ofc. Maybe there
> should be a separate README for this - I haven't found anything like
> that for logical decoding in general, which is why I did (1)-(3).
I read over these and I do think they answer a bunch of questions, but
I don't think they answer all of the questions.
Suppose T1 creates a sequence and commits. Then T2 calls nextval().
Then T3 drops the sequence. According to the commit message, T2's
change will be "replayed immediately after decoding". But it's
essential to replay T2's change after we replay T1 and before we
replay T3, and the comments don't explain why that's guaranteed.
The answer might be "locks". If we always replay a transaction
immediately when we see it's commit record then in the example above
we're fine, because the commit record for the transaction that creates
the sequence must precede the nextval() call, since the sequence won't
be visible until the transaction commits, and also because T1 holds a
lock on it at that point sufficient to hedge out nextval. And the
nextval record must precede the point where T3 takes an exclusive lock
on the sequence.
Note, however, that this change of reasoning critically depends on us
never delaying application of a transaction. If we might reach T1's
commit record and say "hey, let's hold on to this for a bit and replay
it after we've decoded some more," everything immediately breaks,
unless we also delay application of T2's non-transactional update in
such a way that it's still guaranteed to happen after T1. I wonder if
this kind of situation would be a problem for a future parallel-apply
feature. It wouldn't work, for example, to hand T1 and T3 off (in that
order) to a separate apply process but handle T2's "non-transactional"
message directly, because it might handle that message before the
application of T1 got completed.
This also seems to depend on every transactional operation that might
affect a future non-transactional operation holding a lock that would
conflict with that non-transactional operation. For example, if ALTER
SEQUENCE .. RESTART WITH didn't take a strong lock on the sequence,
then you could have: T1 does nextval, T2 does ALTER SEQUENCE RESTART
WITH, T1 does nextval again, T1 commits, T2 commits. It's unclear what
the semantics of that would be -- would T1's second nextval() see the
sequence restart, or what? But if the effect of T1's second nextval
does depend in some way on the ALTER SEQUENCE operation which precedes
it in the WAL stream, then we might have some trouble here, because
both nextvals precede the commit of T2. Fortunately, this sequence of
events is foreclosed by locking.
But I did find one somewhat-similar case in which that's not so.
S1: create table withseq (a bigint generated always as identity);
S1: begin;
S2: select nextval('withseq_a_seq');
S1: alter table withseq set unlogged;
S2: select nextval('withseq_a_seq');
I think this is a bug in the code that supports owned sequences rather
than a problem that this patch should have to do something about. When
a sequence is flipped between logged and unlogged directly, we take a
stronger lock than we do here when it's done in this indirect way.
Also, I'm not quite sure if it would pose a problem for sequence
decoding anyway: it changes the relfilenode, but not the value. But
this is the *kind* of problem that could make the approach unsafe:
supposedly transactional changes being interleaved with supposedly
non-transctional changes, in such a way that the non-transactional
changes might get applied at the wrong time relative to the
transactional changes.
--
Robert Haas
EDB: http://www.enterprisedb.com
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2024-01-27 19:37 Tomas Vondra <[email protected]>
parent: Robert Haas <[email protected]>
0 siblings, 0 replies; 127+ messages in thread
From: Tomas Vondra @ 2024-01-27 19:37 UTC (permalink / raw)
To: Robert Haas <[email protected]>; +Cc: Amit Kapila <[email protected]>; Ashutosh Bapat <[email protected]>; Dilip Kumar <[email protected]>; Hayato Kuroda (Fujitsu) <[email protected]>; Zhijie Hou (Fujitsu) <[email protected]>; PostgreSQL Hackers <[email protected]>; Masahiko Sawada <[email protected]>; Peter Eisentraut <[email protected]>
On 1/26/24 15:39, Robert Haas wrote:
> On Wed, Jan 24, 2024 at 12:46 PM Tomas Vondra
> <[email protected]> wrote:
>> I did try to explain how this works (and why) in a couple places:
>>
>> 1) the commit message
>> 2) reorderbuffer header comment
>> 3) ReorderBufferSequenceIsTransactional comment (and nearby)
>>
>> It's possible this does not meet your expectations, ofc. Maybe there
>> should be a separate README for this - I haven't found anything like
>> that for logical decoding in general, which is why I did (1)-(3).
>
> I read over these and I do think they answer a bunch of questions, but
> I don't think they answer all of the questions.
>
> Suppose T1 creates a sequence and commits. Then T2 calls nextval().
> Then T3 drops the sequence. According to the commit message, T2's
> change will be "replayed immediately after decoding". But it's
> essential to replay T2's change after we replay T1 and before we
> replay T3, and the comments don't explain why that's guaranteed.
>
> The answer might be "locks". If we always replay a transaction
> immediately when we see it's commit record then in the example above
> we're fine, because the commit record for the transaction that creates
> the sequence must precede the nextval() call, since the sequence won't
> be visible until the transaction commits, and also because T1 holds a
> lock on it at that point sufficient to hedge out nextval. And the
> nextval record must precede the point where T3 takes an exclusive lock
> on the sequence.
>
Right, locks + apply in commit order gives us this guarantee (I can't
think of a case where it wouldn't be the case).
> Note, however, that this change of reasoning critically depends on us
> never delaying application of a transaction. If we might reach T1's
> commit record and say "hey, let's hold on to this for a bit and replay
> it after we've decoded some more," everything immediately breaks,
> unless we also delay application of T2's non-transactional update in
> such a way that it's still guaranteed to happen after T1. I wonder if
> this kind of situation would be a problem for a future parallel-apply
> feature. It wouldn't work, for example, to hand T1 and T3 off (in that
> order) to a separate apply process but handle T2's "non-transactional"
> message directly, because it might handle that message before the
> application of T1 got completed.
>
Doesn't the whole logical replication critically depend on the commit
order? If you decide to arbitrarily reorder/delay the transactions, all
kinds of really bad things can happen. That's a generic problem, it
applies to all kinds of objects, not just sequences - a parallel apply
would need to detect this sort of dependencies (e.g. INSERT + DELETE of
the same key), and do something about it.
Similar for sequences, where the important event is allocation of a new
relfilenode.
If anything, it's easier for sequences, because the relfilenode tracking
gives us an explicit (and easy) way to detect these dependencies between
transactions.
> This also seems to depend on every transactional operation that might
> affect a future non-transactional operation holding a lock that would
> conflict with that non-transactional operation. For example, if ALTER
> SEQUENCE .. RESTART WITH didn't take a strong lock on the sequence,
> then you could have: T1 does nextval, T2 does ALTER SEQUENCE RESTART
> WITH, T1 does nextval again, T1 commits, T2 commits. It's unclear what
> the semantics of that would be -- would T1's second nextval() see the
> sequence restart, or what? But if the effect of T1's second nextval
> does depend in some way on the ALTER SEQUENCE operation which precedes
> it in the WAL stream, then we might have some trouble here, because
> both nextvals precede the commit of T2. Fortunately, this sequence of
> events is foreclosed by locking.
>
I don't quite follow :-(
AFAIK this theory hinges on not having the right lock, but I believe
ALTER SEQUENCE does obtain the lock (at least in cases that assign a new
relfilenode). Which means such reordering should not be possible,
because nextval() in other transactions will then wait until commit. And
all nextval() calls in the same transaction will be treated as
transactional.
So I think this works OK. If something does not lock the sequence in a
way that would prevent other xacts to do nextval() on it, it's not a
change that would change the relfilenode - and so it does not switch the
sequence into a transactional mode.
> But I did find one somewhat-similar case in which that's not so.
>
> S1: create table withseq (a bigint generated always as identity);
> S1: begin;
> S2: select nextval('withseq_a_seq');
> S1: alter table withseq set unlogged;
> S2: select nextval('withseq_a_seq');
>
> I think this is a bug in the code that supports owned sequences rather
> than a problem that this patch should have to do something about. When
> a sequence is flipped between logged and unlogged directly, we take a
> stronger lock than we do here when it's done in this indirect way.
Yes, I think this is a bug in handling of owned sequences - from the
moment the "ALTER TABLE ... SET UNLOGGED" is executed, the two sessions
generate duplicate values (until the S1 is committed, at which point the
values generated in S2 get "forgotten").
It seems we end up updating both relfilenodes, which is clearly wrong.
Seems like a bug independent of the decoding, IMO.
> Also, I'm not quite sure if it would pose a problem for sequence
> decoding anyway: it changes the relfilenode, but not the value. But
> this is the *kind* of problem that could make the approach unsafe:
> supposedly transactional changes being interleaved with supposedly
> non-transctional changes, in such a way that the non-transactional
> changes might get applied at the wrong time relative to the
> transactional changes.
>
I'm not sure what you mean by "changes relfilenode, not value" but I
suspect it might break the sequence decoding - or at least confuse it. I
haven't thecked what exactly happens when we change logged/unlogged for
a sequence, but I assume it does change the relfilenode, which already
is a change of a value - we WAL-log the new sequence state, at least.
But it should be treated as "transactional" in the transaction that did
the ALTER TABLE, because it created the relfilenode.
However, I'm not sure this is a valid argument against the sequence
decoding patch. If something does not acquire the correct lock, it's not
surprising something else breaks, if it relies on the lock.
Of course, I understand you're trying to make a broader point - that if
something like this could happen in "correct" case, it'd be a problem.
But I don't think that's possible. The whole "transactional" thing is
determined by having a new relfilenode for the sequence, and I can't
imagine a case where we could assign a new relfilenode without a lock.
regards
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2024-02-20 05:54 Amit Kapila <[email protected]>
parent: Tomas Vondra <[email protected]>
0 siblings, 1 reply; 127+ messages in thread
From: Amit Kapila @ 2024-02-20 05:54 UTC (permalink / raw)
To: Tomas Vondra <[email protected]>; +Cc: Christophe Pettus <[email protected]>; Dilip Kumar <[email protected]>; Hayato Kuroda (Fujitsu) <[email protected]>; Zhijie Hou (Fujitsu) <[email protected]>; Ashutosh Bapat <[email protected]>; PostgreSQL Hackers <[email protected]>; Masahiko Sawada <[email protected]>; Peter Eisentraut <[email protected]>
On Thu, Dec 21, 2023 at 6:47 PM Tomas Vondra
<[email protected]> wrote:
>
> On 12/19/23 13:54, Christophe Pettus wrote:
> > Hi,
> >
> > I wanted to hop in here on one particular issue:
> >
> >> On Dec 12, 2023, at 02:01, Tomas Vondra <[email protected]> wrote:
> >> - desirability of the feature: Random IDs (UUIDs etc.) are likely a much
> >> better solution for distributed (esp. active-active) systems. But there
> >> are important use cases that are likely to keep using regular sequences
> >> (online upgrades of single-node instances, existing systems, ...).
> >
> > +1.
> >
> > Right now, the lack of sequence replication is a rather large
> > foot-gun on logical replication upgrades. Copying the sequences
> > over during the cutover period is doable, of course, but:
> >
> > (a) There's no out-of-the-box tooling that does it, so everyone has
> > to write some scripts just for that one function.
> >
> > (b) It's one more thing that extends the cutover window.
> >
>
> I agree it's an annoying gap for this use case. But if this is the only
> use cases, maybe a better solution would be to provide such tooling
> instead of adding it to the logical decoding?
>
> It might seem a bit strange if most data is copied by replication
> directly, while sequences need special handling, ofc.
>
One difference between the logical replication of tables and sequences
is that we can guarantee with synchronous_commit (and
synchronous_standby_names) that after failover transactions data is
replicated or not whereas for sequences we can't guarantee that
because of their non-transactional nature. Say, there are two
transactions T1 and T2, it is possible that T1's entire table data and
sequence data are committed and replicated but T2's sequence data is
replicated. So, after failover to logical subscriber in such a case if
one routes T2 again to the new node as it was not successful
previously then it would needlessly perform the sequence changes
again. I don't how much that matters but that would probably be the
difference between the replication of tables and sequences.
I agree with your point above that for upgrades some tool like
pg_copysequence where we can provide a way to copy sequence data to
subscribers from the publisher would suffice the need.
--
With Regards,
Amit Kapila.
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2024-02-20 12:09 Tomas Vondra <[email protected]>
parent: Amit Kapila <[email protected]>
0 siblings, 1 reply; 127+ messages in thread
From: Tomas Vondra @ 2024-02-20 12:09 UTC (permalink / raw)
To: Amit Kapila <[email protected]>; +Cc: Christophe Pettus <[email protected]>; Dilip Kumar <[email protected]>; Hayato Kuroda (Fujitsu) <[email protected]>; Zhijie Hou (Fujitsu) <[email protected]>; Ashutosh Bapat <[email protected]>; PostgreSQL Hackers <[email protected]>; Masahiko Sawada <[email protected]>; Peter Eisentraut <[email protected]>
On 2/20/24 06:54, Amit Kapila wrote:
> On Thu, Dec 21, 2023 at 6:47 PM Tomas Vondra
> <[email protected]> wrote:
>>
>> On 12/19/23 13:54, Christophe Pettus wrote:
>>> Hi,
>>>
>>> I wanted to hop in here on one particular issue:
>>>
>>>> On Dec 12, 2023, at 02:01, Tomas Vondra <[email protected]> wrote:
>>>> - desirability of the feature: Random IDs (UUIDs etc.) are likely a much
>>>> better solution for distributed (esp. active-active) systems. But there
>>>> are important use cases that are likely to keep using regular sequences
>>>> (online upgrades of single-node instances, existing systems, ...).
>>>
>>> +1.
>>>
>>> Right now, the lack of sequence replication is a rather large
>>> foot-gun on logical replication upgrades. Copying the sequences
>>> over during the cutover period is doable, of course, but:
>>>
>>> (a) There's no out-of-the-box tooling that does it, so everyone has
>>> to write some scripts just for that one function.
>>>
>>> (b) It's one more thing that extends the cutover window.
>>>
>>
>> I agree it's an annoying gap for this use case. But if this is the only
>> use cases, maybe a better solution would be to provide such tooling
>> instead of adding it to the logical decoding?
>>
>> It might seem a bit strange if most data is copied by replication
>> directly, while sequences need special handling, ofc.
>>
>
> One difference between the logical replication of tables and sequences
> is that we can guarantee with synchronous_commit (and
> synchronous_standby_names) that after failover transactions data is
> replicated or not whereas for sequences we can't guarantee that
> because of their non-transactional nature. Say, there are two
> transactions T1 and T2, it is possible that T1's entire table data and
> sequence data are committed and replicated but T2's sequence data is
> replicated. So, after failover to logical subscriber in such a case if
> one routes T2 again to the new node as it was not successful
> previously then it would needlessly perform the sequence changes
> again. I don't how much that matters but that would probably be the
> difference between the replication of tables and sequences.
>
I don't quite follow what the problem with synchronous_commit is :-(
For sequences, we log the changes ahead, i.e. even if nextval() did not
write anything into WAL, it's still safe because these changes are
covered by the WAL generated some time ago (up to ~32 values back). And
that's certainly subject to synchronous_commit, right?
There certainly are issues with sequences and syncrep:
https://www.postgresql.org/message-id/[email protected]
but that's unrelated to logical replication.
FWIW I don't think we'd re-apply sequence changes needlessly, because
the worker does update the origin after applying non-transactional
changes. So after the replication gets restarted, we'd skip what we
already applied, no?
But maybe there is an issue and I'm just not getting it. Could you maybe
share an example of T1/T2, with a replication restart and what you think
would happen?
> I agree with your point above that for upgrades some tool like
> pg_copysequence where we can provide a way to copy sequence data to
> subscribers from the publisher would suffice the need.
>
Perhaps. Unfortunately it doesn't quite work for failovers, and it's yet
another tool users would need to use.
regards
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
^ permalink raw reply [nested|flat] 127+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2024-02-21 05:06 Amit Kapila <[email protected]>
parent: Tomas Vondra <[email protected]>
0 siblings, 0 replies; 127+ messages in thread
From: Amit Kapila @ 2024-02-21 05:06 UTC (permalink / raw)
To: Tomas Vondra <[email protected]>; +Cc: Christophe Pettus <[email protected]>; Dilip Kumar <[email protected]>; Hayato Kuroda (Fujitsu) <[email protected]>; Zhijie Hou (Fujitsu) <[email protected]>; Ashutosh Bapat <[email protected]>; PostgreSQL Hackers <[email protected]>; Masahiko Sawada <[email protected]>; Peter Eisentraut <[email protected]>
On Tue, Feb 20, 2024 at 5:39 PM Tomas Vondra
<[email protected]> wrote:
>
> On 2/20/24 06:54, Amit Kapila wrote:
> > On Thu, Dec 21, 2023 at 6:47 PM Tomas Vondra
> > <[email protected]> wrote:
> >>
> >> On 12/19/23 13:54, Christophe Pettus wrote:
> >>> Hi,
> >>>
> >>> I wanted to hop in here on one particular issue:
> >>>
> >>>> On Dec 12, 2023, at 02:01, Tomas Vondra <[email protected]> wrote:
> >>>> - desirability of the feature: Random IDs (UUIDs etc.) are likely a much
> >>>> better solution for distributed (esp. active-active) systems. But there
> >>>> are important use cases that are likely to keep using regular sequences
> >>>> (online upgrades of single-node instances, existing systems, ...).
> >>>
> >>> +1.
> >>>
> >>> Right now, the lack of sequence replication is a rather large
> >>> foot-gun on logical replication upgrades. Copying the sequences
> >>> over during the cutover period is doable, of course, but:
> >>>
> >>> (a) There's no out-of-the-box tooling that does it, so everyone has
> >>> to write some scripts just for that one function.
> >>>
> >>> (b) It's one more thing that extends the cutover window.
> >>>
> >>
> >> I agree it's an annoying gap for this use case. But if this is the only
> >> use cases, maybe a better solution would be to provide such tooling
> >> instead of adding it to the logical decoding?
> >>
> >> It might seem a bit strange if most data is copied by replication
> >> directly, while sequences need special handling, ofc.
> >>
> >
> > One difference between the logical replication of tables and sequences
> > is that we can guarantee with synchronous_commit (and
> > synchronous_standby_names) that after failover transactions data is
> > replicated or not whereas for sequences we can't guarantee that
> > because of their non-transactional nature. Say, there are two
> > transactions T1 and T2, it is possible that T1's entire table data and
> > sequence data are committed and replicated but T2's sequence data is
> > replicated. So, after failover to logical subscriber in such a case if
> > one routes T2 again to the new node as it was not successful
> > previously then it would needlessly perform the sequence changes
> > again. I don't how much that matters but that would probably be the
> > difference between the replication of tables and sequences.
> >
>
> I don't quite follow what the problem with synchronous_commit is :-(
>
> For sequences, we log the changes ahead, i.e. even if nextval() did not
> write anything into WAL, it's still safe because these changes are
> covered by the WAL generated some time ago (up to ~32 values back). And
> that's certainly subject to synchronous_commit, right?
>
> There certainly are issues with sequences and syncrep:
>
> https://www.postgresql.org/message-id/[email protected]
>
> but that's unrelated to logical replication.
>
> FWIW I don't think we'd re-apply sequence changes needlessly, because
> the worker does update the origin after applying non-transactional
> changes. So after the replication gets restarted, we'd skip what we
> already applied, no?
>
It will work for restarts but I was trying to discuss what happens in
the scenario after the publisher node goes down and we failover to the
subscriber node and make it a primary node (or a failover case). After
that, all unfinished transactions will be re-routed to the new
primary. Consider a theoretical case where we send sequence changes of
the yet uncommitted transactions directly from wal buffers (something
like 91f2cae7a4 does for physical replication) and then immediately
the primary or publisher node crashes. After failover to the
subscriber node, the application will re-route unfinished transactions
to the new primary. In such a situation, I think there is a chance
that we will update the sequence value when it would have already
received/applied that update via replication. This is what I was
saying that there is probably a difference between tables and
sequences, for tables such a replicated change would be rolled back.
Having said that, this is probably no different from what would happen
in the case of physical replication.
> But maybe there is an issue and I'm just not getting it. Could you maybe
> share an example of T1/T2, with a replication restart and what you think
> would happen?
>
> > I agree with your point above that for upgrades some tool like
> > pg_copysequence where we can provide a way to copy sequence data to
> > subscribers from the publisher would suffice the need.
> >
>
> Perhaps. Unfortunately it doesn't quite work for failovers, and it's yet
> another tool users would need to use.
>
But can logical replica be used for failover? We don't have any way to
replicate/sync the slots on subscribers and neither we have a
mechanism to replicate existing publications. I think if we want to
achieve failover to a logical subscriber we need to replicate/sync the
required logical and physical slots to the subscribers. I haven't
thought through it completely so there would probably be more things
to consider for allowing logical subscribers to be used as failover
candidates.
--
With Regards,
Amit Kapila.
^ permalink raw reply [nested|flat] 127+ messages in thread
end of thread, other threads:[~2024-02-21 05:06 UTC | newest]
Thread overview: 127+ messages (download: mbox mbox.gz follow: Atom feed)
-- links below jump to the message on this page --
2019-07-23 15:28 [PATCH] Support pg_walfile_name on standby Jehan-Guillaume de Rorthais <[email protected]>
2019-07-23 15:28 [PATCH] Support pg_walfile_name on standby Jehan-Guillaume de Rorthais <[email protected]>
2021-06-07 22:28 logical decoding and replication of sequences Tomas Vondra <[email protected]>
2021-06-13 21:15 ` Re: logical decoding and replication of sequences Tomas Vondra <[email protected]>
2021-06-23 14:14 ` Re: logical decoding and replication of sequences Tomas Vondra <[email protected]>
2021-06-23 14:25 ` Re: logical decoding and replication of sequences Tomas Vondra <[email protected]>
2021-07-15 12:17 ` Re: logical decoding and replication of sequences vignesh C <[email protected]>
2021-07-20 21:41 ` Re: logical decoding and replication of sequences Tomas Vondra <[email protected]>
2021-07-30 18:26 ` Re: logical decoding and replication of sequences Tomas Vondra <[email protected]>
2021-09-23 10:27 ` Re: logical decoding and replication of sequences Peter Eisentraut <[email protected]>
2021-09-24 19:16 ` Re: logical decoding and replication of sequences Tomas Vondra <[email protected]>
2021-09-25 20:05 ` Re: logical decoding and replication of sequences Hannu Krosing <[email protected]>
2021-10-31 20:44 ` Re: logical decoding and replication of sequences Tomas Vondra <[email protected]>
2021-11-19 19:54 ` Re: logical decoding and replication of sequences Tomas Vondra <[email protected]>
2021-11-22 00:47 ` Re: logical decoding and replication of sequences Tomas Vondra <[email protected]>
2021-11-22 15:44 ` Re: logical decoding and replication of sequences Peter Eisentraut <[email protected]>
2021-11-22 16:01 ` Re: logical decoding and replication of sequences Petr Jelinek <[email protected]>
2021-12-07 14:38 ` Re: logical decoding and replication of sequences Peter Eisentraut <[email protected]>
2021-12-08 00:23 ` Re: logical decoding and replication of sequences Tomas Vondra <[email protected]>
2021-12-08 09:04 ` Re: logical decoding and replication of sequences Peter Eisentraut <[email protected]>
2021-12-14 01:31 ` Re: logical decoding and replication of sequences Tomas Vondra <[email protected]>
2021-12-15 13:20 ` Re: logical decoding and replication of sequences Amit Kapila <[email protected]>
2021-12-15 13:51 ` Re: logical decoding and replication of sequences Tomas Vondra <[email protected]>
2021-12-15 13:58 ` Re: logical decoding and replication of sequences Tomas Vondra <[email protected]>
2021-12-16 11:59 ` Re: logical decoding and replication of sequences Amit Kapila <[email protected]>
2021-12-16 14:54 ` Re: logical decoding and replication of sequences Tomas Vondra <[email protected]>
2021-12-18 01:56 ` Re: logical decoding and replication of sequences Tomas Vondra <[email protected]>
2021-12-22 15:40 ` Re: logical decoding and replication of sequences Tomas Vondra <[email protected]>
2021-12-15 13:44 ` Re: logical decoding and replication of sequences Peter Eisentraut <[email protected]>
2021-12-15 16:42 ` Re: logical decoding and replication of sequences Alvaro Herrera <[email protected]>
2021-12-15 16:56 ` Re: logical decoding and replication of sequences Tomas Vondra <[email protected]>
2021-11-23 01:01 ` Re: logical decoding and replication of sequences Andres Freund <[email protected]>
2021-11-23 13:36 ` Re: logical decoding and replication of sequences Tomas Vondra <[email protected]>
2021-07-30 18:58 ` Re: logical decoding and replication of sequences Robert Haas <[email protected]>
2021-07-20 15:30 ` Re: logical decoding and replication of sequences Peter Eisentraut <[email protected]>
2021-07-20 20:27 ` Re: logical decoding and replication of sequences Tomas Vondra <[email protected]>
2022-03-23 11:50 ` Re: logical decoding and replication of sequences Amit Kapila <[email protected]>
2022-03-23 12:46 ` Re: logical decoding and replication of sequences Petr Jelinek <[email protected]>
2022-03-23 22:30 ` Re: logical decoding and replication of sequences Tomas Vondra <[email protected]>
2022-03-24 21:52 ` Re: logical decoding and replication of sequences Greg Stark <[email protected]>
2022-03-24 22:01 ` Re: logical decoding and replication of sequences Tomas Vondra <[email protected]>
2022-03-24 21:59 ` Re: logical decoding and replication of sequences Tomas Vondra <[email protected]>
2022-03-25 04:01 ` Re: logical decoding and replication of sequences Amit Kapila <[email protected]>
2022-03-25 10:26 ` Re: logical decoding and replication of sequences Tomas Vondra <[email protected]>
2022-03-25 11:21 ` Re: logical decoding and replication of sequences Amit Kapila <[email protected]>
2022-03-25 11:59 ` Re: logical decoding and replication of sequences Tomas Vondra <[email protected]>
2022-03-25 16:50 ` Re: logical decoding and replication of sequences Tomas Vondra <[email protected]>
2022-03-26 07:28 ` Re: logical decoding and replication of sequences Amit Kapila <[email protected]>
2022-03-26 09:56 ` Re: logical decoding and replication of sequences Tomas Vondra <[email protected]>
2022-03-28 05:29 ` Re: logical decoding and replication of sequences Amit Kapila <[email protected]>
2022-05-25 20:42 ` Re: logical decoding and replication of sequences Zheng Li <[email protected]>
2022-08-07 00:36 ` Re: logical decoding and replication of sequences Noah Misch <[email protected]>
2022-08-07 13:18 ` Re: logical decoding and replication of sequences Tomas Vondra <[email protected]>
2022-08-07 19:12 ` Re: logical decoding and replication of sequences Noah Misch <[email protected]>
2022-08-07 21:09 ` Re: logical decoding and replication of sequences Thomas Munro <[email protected]>
2022-08-08 06:15 ` Re: logical decoding and replication of sequences Thomas Munro <[email protected]>
2022-08-08 08:33 ` Re: logical decoding and replication of sequences Kyotaro Horiguchi <[email protected]>
2022-08-08 08:56 ` Re: logical decoding and replication of sequences Kyotaro Horiguchi <[email protected]>
2022-08-29 10:21 ` Re: logical decoding and replication of sequences Thomas Munro <[email protected]>
2022-08-29 18:04 ` Re: logical decoding and replication of sequences Tomas Vondra <[email protected]>
2022-08-29 21:42 ` Re: logical decoding and replication of sequences Thomas Munro <[email protected]>
2022-08-23 04:21 ` Re: logical decoding and replication of sequences Michael Paquier <[email protected]>
2022-08-23 04:32 ` Re: logical decoding and replication of sequences Thomas Munro <[email protected]>
2022-08-23 15:04 ` Re: logical decoding and replication of sequences Robert Haas <[email protected]>
2022-08-24 00:51 ` Re: logical decoding and replication of sequences Thomas Munro <[email protected]>
2022-03-25 20:10 ` Re: logical decoding and replication of sequences Tomas Vondra <[email protected]>
2022-03-25 07:00 ` Re: logical decoding and replication of sequences Masahiko Sawada <[email protected]>
2022-03-25 10:34 ` Re: logical decoding and replication of sequences Tomas Vondra <[email protected]>
2022-03-25 14:34 ` Re: logical decoding and replication of sequences vignesh C <[email protected]>
2022-03-25 19:58 ` Re: logical decoding and replication of sequences Tomas Vondra <[email protected]>
2021-12-13 06:07 [PATCH v10 6/6] Move removal of old logical rewrite mapping files to custodian. Nathan Bossart <[email protected]>
2022-08-18 21:10 logical decoding and replication of sequences, take 2 Tomas Vondra <[email protected]>
2022-08-19 11:11 ` Re: logical decoding and replication of sequences, take 2 Tomas Vondra <[email protected]>
2022-11-11 22:49 ` Re: logical decoding and replication of sequences, take 2 Tomas Vondra <[email protected]>
2022-11-16 04:43 ` Re: logical decoding and replication of sequences, take 2 Ian Lawrence Barwick <[email protected]>
2022-11-16 21:05 ` Re: logical decoding and replication of sequences, take 2 Robert Haas <[email protected]>
2022-11-17 01:41 ` Re: logical decoding and replication of sequences, take 2 Tomas Vondra <[email protected]>
2022-11-17 02:43 ` Re: logical decoding and replication of sequences, take 2 Andres Freund <[email protected]>
2022-11-17 11:39 ` Re: logical decoding and replication of sequences, take 2 Tomas Vondra <[email protected]>
2022-11-17 17:07 ` Re: logical decoding and replication of sequences, take 2 Andres Freund <[email protected]>
2022-11-17 21:13 ` Re: logical decoding and replication of sequences, take 2 Tomas Vondra <[email protected]>
2022-11-18 03:03 ` Re: logical decoding and replication of sequences, take 2 Andres Freund <[email protected]>
2022-11-17 16:04 ` Re: logical decoding and replication of sequences, take 2 Robert Haas <[email protected]>
2023-11-29 14:41 ` Re: logical decoding and replication of sequences, take 2 Tomas Vondra <[email protected]>
2023-11-29 23:58 ` Re: logical decoding and replication of sequences, take 2 Tomas Vondra <[email protected]>
2023-11-30 11:56 ` Re: logical decoding and replication of sequences, take 2 Amit Kapila <[email protected]>
2023-12-02 00:10 ` Re: logical decoding and replication of sequences, take 2 Tomas Vondra <[email protected]>
2023-12-01 11:08 ` RE: logical decoding and replication of sequences, take 2 Hayato Kuroda (Fujitsu) <[email protected]>
2023-12-02 00:23 ` Re: logical decoding and replication of sequences, take 2 Tomas Vondra <[email protected]>
2023-12-03 12:55 ` RE: logical decoding and replication of sequences, take 2 Hayato Kuroda (Fujitsu) <[email protected]>
2023-12-03 17:52 ` Re: logical decoding and replication of sequences, take 2 Tomas Vondra <[email protected]>
2023-12-03 18:26 ` Re: logical decoding and replication of sequences, take 2 Tomas Vondra <[email protected]>
2023-12-06 10:19 ` Re: logical decoding and replication of sequences, take 2 Amit Kapila <[email protected]>
2023-12-06 13:50 ` Re: logical decoding and replication of sequences, take 2 Tomas Vondra <[email protected]>
2023-12-07 04:03 ` Re: logical decoding and replication of sequences, take 2 Amit Kapila <[email protected]>
2023-12-05 12:17 ` Re: logical decoding and replication of sequences, take 2 Amit Kapila <[email protected]>
2023-12-06 05:42 ` Re: logical decoding and replication of sequences, take 2 Dilip Kumar <[email protected]>
2023-12-06 09:05 ` Re: logical decoding and replication of sequences, take 2 Dilip Kumar <[email protected]>
2023-12-06 13:39 ` Re: logical decoding and replication of sequences, take 2 Tomas Vondra <[email protected]>
2023-12-07 05:11 ` Re: logical decoding and replication of sequences, take 2 Dilip Kumar <[email protected]>
2023-12-13 12:56 ` Re: logical decoding and replication of sequences, take 2 Amit Kapila <[email protected]>
2023-12-14 03:44 ` RE: logical decoding and replication of sequences, take 2 Hayato Kuroda (Fujitsu) <[email protected]>
2023-12-14 05:22 ` Re: logical decoding and replication of sequences, take 2 Dilip Kumar <[email protected]>
2023-12-14 07:01 ` Re: logical decoding and replication of sequences, take 2 Ashutosh Bapat <[email protected]>
2023-12-14 07:07 ` Re: logical decoding and replication of sequences, take 2 Dilip Kumar <[email protected]>
2023-12-14 09:06 ` Re: logical decoding and replication of sequences, take 2 Amit Kapila <[email protected]>
2023-12-14 09:14 ` Re: logical decoding and replication of sequences, take 2 Ashutosh Bapat <[email protected]>
2023-12-14 09:20 ` Re: logical decoding and replication of sequences, take 2 Amit Kapila <[email protected]>
2023-12-14 15:44 ` Re: logical decoding and replication of sequences, take 2 Ashutosh Bapat <[email protected]>
2023-12-15 02:33 ` Re: logical decoding and replication of sequences, take 2 Amit Kapila <[email protected]>
2023-12-21 14:04 ` Re: logical decoding and replication of sequences, take 2 Tomas Vondra <[email protected]>
2024-01-11 16:26 ` Re: logical decoding and replication of sequences, take 2 Tomas Vondra <[email protected]>
2024-01-23 20:47 ` Re: logical decoding and replication of sequences, take 2 Robert Haas <[email protected]>
2024-01-24 17:46 ` Re: logical decoding and replication of sequences, take 2 Tomas Vondra <[email protected]>
2024-01-26 14:39 ` Re: logical decoding and replication of sequences, take 2 Robert Haas <[email protected]>
2024-01-27 19:37 ` Re: logical decoding and replication of sequences, take 2 Tomas Vondra <[email protected]>
2023-12-06 10:05 ` Re: logical decoding and replication of sequences, take 2 Amit Kapila <[email protected]>
2023-12-06 11:05 ` Re: logical decoding and replication of sequences, take 2 Dilip Kumar <[email protected]>
2023-12-06 13:47 ` Re: logical decoding and replication of sequences, take 2 Tomas Vondra <[email protected]>
2023-12-07 04:35 ` Re: logical decoding and replication of sequences, take 2 Dilip Kumar <[email protected]>
2023-12-07 04:56 ` Re: logical decoding and replication of sequences, take 2 Amit Kapila <[email protected]>
2023-12-12 10:01 ` Re: logical decoding and replication of sequences, take 2 Tomas Vondra <[email protected]>
2023-12-19 12:54 ` Re: logical decoding and replication of sequences, take 2 Christophe Pettus <[email protected]>
2023-12-21 13:17 ` Re: logical decoding and replication of sequences, take 2 Tomas Vondra <[email protected]>
2024-02-20 05:54 ` Re: logical decoding and replication of sequences, take 2 Amit Kapila <[email protected]>
2024-02-20 12:09 ` Re: logical decoding and replication of sequences, take 2 Tomas Vondra <[email protected]>
2024-02-21 05:06 ` Re: logical decoding and replication of sequences, take 2 Amit Kapila <[email protected]>
This inbox is served by agora; see mirroring instructions
for how to clone and mirror all data and code used for this inbox