public inbox for [email protected]
help / color / mirror / Atom feed[PATCH v10 6/6] Move removal of old logical rewrite mapping files to custodian.
45+ messages / 8 participants
[nested] [flat]
* [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; 45+ 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] 45+ messages in thread
* Re: logical decoding and replication of sequences, take 2
@ 2023-11-29 14:41 Tomas Vondra <[email protected]>
0 siblings, 1 reply; 45+ 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] 45+ 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; 45+ 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] 45+ 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; 45+ 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] 45+ 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; 45+ 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] 45+ 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; 45+ 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] 45+ 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; 45+ 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] 45+ 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; 45+ 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] 45+ 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; 45+ 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] 45+ 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; 45+ 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] 45+ 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; 45+ 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] 45+ 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; 45+ 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] 45+ 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; 45+ 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] 45+ 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; 45+ 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] 45+ 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; 45+ 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] 45+ 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; 45+ 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] 45+ 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; 45+ 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] 45+ 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; 45+ 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] 45+ 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; 45+ 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] 45+ 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; 45+ 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] 45+ 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; 45+ 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] 45+ 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; 45+ 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] 45+ 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; 45+ 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] 45+ 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; 45+ 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] 45+ 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; 45+ 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] 45+ 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; 45+ 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] 45+ 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; 45+ 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] 45+ 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; 45+ 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] 45+ 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; 45+ 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] 45+ 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; 45+ 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] 45+ 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; 45+ 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] 45+ 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; 45+ 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] 45+ 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; 45+ 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] 45+ 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; 45+ 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] 45+ 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; 45+ 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] 45+ 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; 45+ 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] 45+ 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; 45+ 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] 45+ 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; 45+ 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] 45+ 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; 45+ 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] 45+ 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; 45+ 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] 45+ 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; 45+ 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] 45+ 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; 45+ 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] 45+ 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; 45+ 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] 45+ 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; 45+ 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] 45+ 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; 45+ 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] 45+ messages in thread
end of thread, other threads:[~2024-02-21 05:06 UTC | newest]
Thread overview: 45+ messages (download: mbox mbox.gz follow: Atom feed)
-- links below jump to the message on this page --
2021-12-13 06:07 [PATCH v10 6/6] Move removal of old logical rewrite mapping files to custodian. Nathan Bossart <[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