public inbox for [email protected]
help / color / mirror / Atom feedFrom: Drouvot, Bertrand <[email protected]>
To: Andres Freund <[email protected]>
To: Robert Haas <[email protected]>
To: Thomas Munro <[email protected]>
To: Alvaro Herrera <[email protected]>
To: Ibrar Ahmed <[email protected]>
To: Amit Khandekar <[email protected]>
To: [email protected]
To: tushar <[email protected]>
To: Rahila Syed <[email protected]>
To: pgsql-hackers <[email protected]>
To: Melanie Plageman <[email protected]>
Subject: Re: Minimal logical decoding on standbys
Date: Mon, 30 Jan 2023 17:18:39 +0100
Message-ID: <[email protected]> (raw)
In-Reply-To: <[email protected]>
References: <CA+TgmoYTTsxP8y6uknZvCBNCRq+1FJ4zGbX8Px1TGW459fGsaQ@mail.gmail.com>
<[email protected]>
<[email protected]>
<[email protected]>
<CA+TgmoY8LAaTwAhS+YNQNrM+3K8eqxGvPT3gbJaHER86XShvFw@mail.gmail.com>
<[email protected]>
<[email protected]>
<[email protected]>
<[email protected]>
<[email protected]>
<20230123232121.5zs7dd45f442ozwu@liskov>
<[email protected]>
Hi,
Just realized that Melanie was missing in the up-thread reply to
her feedback (not sure what happened, sorry about that).... So, adding her here.
Please find attached V45 addressing Melanie's feedback.
On 1/24/23 3:59 PM, Drouvot, Bertrand wrote:
> Hi,
>
> On 1/24/23 12:21 AM, Melanie Plageman wrote:
>> I'm new to this thread and subject, but I had a few basic thoughts about
>> the first patch in the set.
>>
>
> Thanks for looking at it!
>
>> On Mon, Jan 23, 2023 at 12:03:35PM +0100, Drouvot, Bertrand wrote:
>>> 1. We want to enable logical decoding on standbys, but replay of WAL
>>> from the primary might remove data that is needed by logical decoding,
>>> causing replication conflicts much as hot standby does.
>>
>> It is a little confusing to mention replication conflicts in point 1. It
>> makes it sound like it already logs a recovery conflict. Without the
>> recovery conflict handling in this patchset, logical decoding of
>> statements using data that has been removed will fail with some error
>> like :
>> ERROR: could not map filenumber "xxx" to relation OID
>> Part of what this patchset does is introduce the concept of a new kind
>> of recovery conflict and a resolution process.
Changed the wording in V45's commit message.
>>> 3. To do this we need the latestRemovedXid for each change, just as we
>>> do for physical replication conflicts, but we also need to know
>>> whether any particular change was to data that logical replication
>>> might access.
>
>> It isn't clear from the above sentence why you would need both. I think
>> it has something to do with what is below (hot_standby_feedback being
>> off), but I'm not sure, so the order is confusing.
>
Trying to be more clear in the new commit message.
>
>>> Implementation:
>>>
>>> When a WAL replay on standby indicates that a catalog table tuple is
>>> to be deleted by an xid that is greater than a logical slot's
>>> catalog_xmin, then that means the slot's catalog_xmin conflicts with
>>> the xid, and we need to handle the conflict. While subsequent commits
>>> will do the actual conflict handling, this commit adds a new field
>>> isCatalogRel in such WAL records (and a new bit set in the
>>> xl_heap_visible flags field), that is true for catalog tables, so as to
>>> arrange for conflict handling.
>>
>> You do mention it a bit here, but I think it could be more clear and
>> specific.
>
Added the WAL record types impacted by the change in the new commit message.
>>
>> It is not very important, but I noticed you made "heaprel" the last
>> parameter to all of the btree-related functions but the first parameter
>> to the gist functions. I thought it might be nice to make the order
>> consistent.
>
> Agree, will do.
Done.
>
>> I also was wondering why you made it the last argument to
>> all the btree functions to begin with (i.e. instead of directly after
>> the first rel argument).
>>
>
> No real reasons, will put all of them after the first rel argument (that seems a better place).
Done.
>
>>> diff --git a/src/include/access/gist_private.h b/src/include/access/gist_private.h
>>> index 8af33d7b40..9bdac12baf 100644
>>> --- a/src/include/access/gist_private.h
>>> +++ b/src/include/access/gist_private.h
>>> @@ -440,7 +440,7 @@ extern XLogRecPtr gistXLogPageDelete(Buffer buffer,
>>> FullTransactionId xid, Buffer parentBuffer,
>>> OffsetNumber downlinkOffset);
>>> -extern void gistXLogPageReuse(Relation rel, BlockNumber blkno,
>>> +extern void gistXLogPageReuse(Relation heaprel, Relation rel, BlockNumber blkno,
>>> FullTransactionId deleteXid);
>>> extern XLogRecPtr gistXLogUpdate(Buffer buffer,
>>> @@ -485,7 +485,7 @@ extern bool gistproperty(Oid index_oid, int attno,
>>> extern bool gistfitpage(IndexTuple *itvec, int len);
>>> extern bool gistnospace(Page page, IndexTuple *itvec, int len, OffsetNumber todelete, Size freespace);
>>> extern void gistcheckpage(Relation rel, Buffer buf);
>>> -extern Buffer gistNewBuffer(Relation r);
>>> +extern Buffer gistNewBuffer(Relation heaprel, Relation r);
>>> extern bool gistPageRecyclable(Page page);
>>> extern void gistfillbuffer(Page page, IndexTuple *itup, int len,
>>> OffsetNumber off);
>>> diff --git a/src/include/access/gistxlog.h b/src/include/access/gistxlog.h
>>> index 09f9b0f8c6..191f0e5808 100644
>>> --- a/src/include/access/gistxlog.h
>>> +++ b/src/include/access/gistxlog.h
>>> @@ -51,13 +51,13 @@ typedef struct gistxlogDelete
>>> {
>>> TransactionId snapshotConflictHorizon;
>>> uint16 ntodelete; /* number of deleted offsets */
>>> + bool isCatalogRel;
>>
>> In some of these struct definitions, I think it would help comprehension
>> to have a comment explaining the purpose of this member.
>>
>
> Yeah, agree but it could be done in another patch (outside of this feature), agree?
Please forget about my previous reply (I misunderstood and thought you were mentioning the offset's Array).
Added comments about isCatalogRel in V45 attached.
Regards,
--
Bertrand Drouvot
PostgreSQL Contributors Team
RDS Open Source Databases
Amazon Web Services: https://aws.amazon.com
From f45d02edeeed24e604c1e48cf39c09224eac1644 Mon Sep 17 00:00:00 2001
From: bdrouvotAWS <[email protected]>
Date: Mon, 30 Jan 2023 15:44:00 +0000
Subject: [PATCH v45 6/6] Doc changes describing details about logical
decoding.
Author: Andres Freund (in an older version), Amit Khandekar, Bertrand Drouvot
Reviewed-By: Bertrand Drouvot, Andres Freund, Robert Haas, Fabrizio de Royes Mello
---
doc/src/sgml/logicaldecoding.sgml | 21 +++++++++++++++++++++
1 file changed, 21 insertions(+)
100.0% doc/src/sgml/
diff --git a/doc/src/sgml/logicaldecoding.sgml b/doc/src/sgml/logicaldecoding.sgml
index 4e912b4bd4..2e8bee033f 100644
--- a/doc/src/sgml/logicaldecoding.sgml
+++ b/doc/src/sgml/logicaldecoding.sgml
@@ -316,6 +316,27 @@ postgres=# select * from pg_logical_slot_get_changes('regression_slot', NULL, NU
may consume changes from a slot at any given time.
</para>
+ <para>
+ A logical replication slot can also be created on a hot standby. To prevent
+ <command>VACUUM</command> from removing required rows from the system
+ catalogs, <varname>hot_standby_feedback</varname> should be set on the
+ standby. In spite of that, if any required rows get removed, the slot gets
+ invalidated. It's highly recommended to use a physical slot between the primary
+ and the standby. Otherwise, hot_standby_feedback will work, but only while the
+ connection is alive (for example a node restart would break it). Existing
+ logical slots on standby also get invalidated if wal_level on primary is reduced to
+ less than 'logical'.
+ </para>
+
+ <para>
+ For a logical slot to be created, it builds a historic snapshot, for which
+ information of all the currently running transactions is essential. On
+ primary, this information is available, but on standby, this information
+ has to be obtained from primary. So, slot creation may wait for some
+ activity to happen on the primary. If the primary is idle, creating a
+ logical slot on standby may take a noticeable time.
+ </para>
+
<caution>
<para>
Replication slots persist across crashes and know nothing about the state
--
2.34.1
From 61471ca687aa95eb334fd23448836bc8100ed92f Mon Sep 17 00:00:00 2001
From: bdrouvotAWS <[email protected]>
Date: Mon, 30 Jan 2023 15:43:21 +0000
Subject: [PATCH v45 5/6] New TAP test for logical decoding on standby.
Author: Craig Ringer (in an older version), Amit Khandekar, Bertrand Drouvot
Reviewed-By: Bertrand Drouvot, Andres Freund, Robert Haas, Fabrizio de Royes Mello
---
src/test/perl/PostgreSQL/Test/Cluster.pm | 39 ++
src/test/recovery/meson.build | 1 +
.../t/034_standby_logical_decoding.pl | 658 ++++++++++++++++++
3 files changed, 698 insertions(+)
5.0% src/test/perl/PostgreSQL/Test/
94.8% src/test/recovery/t/
diff --git a/src/test/perl/PostgreSQL/Test/Cluster.pm b/src/test/perl/PostgreSQL/Test/Cluster.pm
index 04921ca3a3..fd81ddcf39 100644
--- a/src/test/perl/PostgreSQL/Test/Cluster.pm
+++ b/src/test/perl/PostgreSQL/Test/Cluster.pm
@@ -3037,6 +3037,45 @@ $SIG{TERM} = $SIG{INT} = sub {
=pod
+=item $node->create_logical_slot_on_standby(self, primary, slot_name, dbname)
+
+Create logical replication slot on given standby
+
+=cut
+
+sub create_logical_slot_on_standby
+{
+ my ($self, $primary, $slot_name, $dbname) = @_;
+ my ($stdout, $stderr);
+
+ my $handle;
+
+ $handle = IPC::Run::start(['pg_recvlogical', '-d', $self->connstr($dbname), '-P', 'test_decoding', '-S', $slot_name, '--create-slot'], '>', \$stdout, '2>', \$stderr);
+
+ # Once slot restart_lsn is created, the standby looks for xl_running_xacts
+ # WAL record from the restart_lsn onwards. So firstly, wait until the slot
+ # restart_lsn is evaluated.
+
+ $self->poll_query_until(
+ 'postgres', qq[
+ SELECT restart_lsn IS NOT NULL
+ FROM pg_catalog.pg_replication_slots WHERE slot_name = '$slot_name'
+ ]) or die "timed out waiting for logical slot to calculate its restart_lsn";
+
+ # Now arrange for the xl_running_xacts record for which pg_recvlogical
+ # is waiting.
+ # Note: Write a C helper function to call LogStandbySnapshot() instead
+ # of asking for a checkpoint.
+ $primary->safe_psql('postgres', 'CHECKPOINT');
+
+ $handle->finish();
+
+ is($self->slot($slot_name)->{'slot_type'}, 'logical', $slot_name . ' on standby created')
+ or die "could not create slot" . $slot_name;
+}
+
+=pod
+
=back
=cut
diff --git a/src/test/recovery/meson.build b/src/test/recovery/meson.build
index edaaa1a3ce..52b2816c7a 100644
--- a/src/test/recovery/meson.build
+++ b/src/test/recovery/meson.build
@@ -40,6 +40,7 @@ tests += {
't/031_recovery_conflict.pl',
't/032_relfilenode_reuse.pl',
't/033_replay_tsp_drops.pl',
+ 't/034_standby_logical_decoding.pl',
],
},
}
diff --git a/src/test/recovery/t/034_standby_logical_decoding.pl b/src/test/recovery/t/034_standby_logical_decoding.pl
new file mode 100644
index 0000000000..4370d595d8
--- /dev/null
+++ b/src/test/recovery/t/034_standby_logical_decoding.pl
@@ -0,0 +1,658 @@
+# logical decoding on standby : test logical decoding,
+# recovery conflict and standby promotion.
+
+use strict;
+use warnings;
+
+use PostgreSQL::Test::Cluster;
+use PostgreSQL::Test::Utils;
+use Test::More tests => 62;
+
+my ($stdin, $stdout, $stderr, $ret, $handle, $slot);
+
+my $node_primary = PostgreSQL::Test::Cluster->new('primary');
+my $node_standby = PostgreSQL::Test::Cluster->new('standby');
+my $default_timeout = $PostgreSQL::Test::Utils::timeout_default;
+my $res;
+
+# Name for the physical slot on primary
+my $primary_slotname = 'primary_physical';
+
+# find $pat in logfile of $node after $off-th byte
+sub find_in_log
+{
+ my ($node, $pat, $off) = @_;
+
+ $off = 0 unless defined $off;
+ my $log = PostgreSQL::Test::Utils::slurp_file($node->logfile);
+ return 0 if (length($log) <= $off);
+
+ $log = substr($log, $off);
+
+ return $log =~ m/$pat/;
+}
+
+# Fetch xmin columns from slot's pg_replication_slots row, after waiting for
+# given boolean condition to be true to ensure we've reached a quiescent state.
+sub wait_for_xmins
+{
+ my ($node, $slotname, $check_expr) = @_;
+
+ $node->poll_query_until(
+ 'postgres', qq[
+ SELECT $check_expr
+ FROM pg_catalog.pg_replication_slots
+ WHERE slot_name = '$slotname';
+ ]) or die "Timed out waiting for slot xmins to advance";
+}
+
+# Create the required logical slots on standby.
+sub create_logical_slots
+{
+ $node_standby->create_logical_slot_on_standby($node_primary, 'inactiveslot', 'testdb');
+ $node_standby->create_logical_slot_on_standby($node_primary, 'activeslot', 'testdb');
+}
+
+# Drop the logical slots on standby.
+sub drop_logical_slots
+{
+ $node_standby->psql('postgres', q[SELECT pg_drop_replication_slot('inactiveslot')]);
+ $node_standby->psql('postgres', q[SELECT pg_drop_replication_slot('activeslot')]);
+}
+
+# Acquire one of the standby logical slots created by create_logical_slots().
+# In case wait is true we are waiting for an active pid on the 'activeslot' slot.
+# If wait is not true it means we are testing a known failure scenario.
+sub make_slot_active
+{
+ my $wait = shift;
+ my $slot_user_handle;
+
+ $slot_user_handle = IPC::Run::start(['pg_recvlogical', '-d', $node_standby->connstr('testdb'), '-S', 'activeslot', '-o', 'include-xids=0', '-o', 'skip-empty-xacts=1', '--no-loop', '--start', '-f', '-'], '>', \$stdout, '2>', \$stderr);
+
+ if ($wait)
+ {
+ # make sure activeslot is in use
+ $node_standby->poll_query_until('testdb',
+ "SELECT EXISTS (SELECT 1 FROM pg_replication_slots WHERE slot_name = 'activeslot' AND active_pid IS NOT NULL)"
+ ) or die "slot never became active";
+ }
+ return $slot_user_handle;
+}
+
+# Check pg_recvlogical stderr
+sub check_pg_recvlogical_stderr
+{
+ my ($slot_user_handle, $check_stderr) = @_;
+ my $return;
+
+ # our client should've terminated in response to the walsender error
+ $slot_user_handle->finish;
+ $return = $?;
+ cmp_ok($return, "!=", 0, "pg_recvlogical exited non-zero");
+ if ($return) {
+ like($stderr, qr/$check_stderr/, 'slot has been invalidated');
+ }
+
+ return 0;
+}
+
+# Check if all the slots on standby are dropped. These include the 'activeslot'
+# that was acquired by make_slot_active(), and the non-active 'inactiveslot'.
+sub check_slots_dropped
+{
+ my ($slot_user_handle) = @_;
+
+ is($node_standby->slot('inactiveslot')->{'slot_type'}, '', 'inactiveslot on standby dropped');
+ is($node_standby->slot('activeslot')->{'slot_type'}, '', 'activeslot on standby dropped');
+
+ check_pg_recvlogical_stderr($slot_user_handle, "conflict with recovery");
+}
+
+# Check if all the slots on standby are dropped. These include the 'activeslot'
+# that was acquired by make_slot_active(), and the non-active 'inactiveslot'.
+sub change_hot_standby_feedback_and_wait_for_xmins
+{
+ my ($hsf, $invalidated) = @_;
+
+ $node_standby->append_conf('postgresql.conf',qq[
+ hot_standby_feedback = $hsf
+ ]);
+
+ $node_standby->reload;
+
+ if ($hsf && $invalidated)
+ {
+ # With hot_standby_feedback on, xmin should advance,
+ # but catalog_xmin should still remain NULL since there is no logical slot.
+ wait_for_xmins($node_primary, $primary_slotname,
+ "xmin IS NOT NULL AND catalog_xmin IS NULL");
+ }
+ elsif ($hsf)
+ {
+ # With hot_standby_feedback on, xmin and catalog_xmin should advance.
+ wait_for_xmins($node_primary, $primary_slotname,
+ "xmin IS NOT NULL AND catalog_xmin IS NOT NULL");
+ }
+ else
+ {
+ # Both should be NULL since hs_feedback is off
+ wait_for_xmins($node_primary, $primary_slotname,
+ "xmin IS NULL AND catalog_xmin IS NULL");
+
+ }
+}
+
+# Check conflicting status in pg_replication_slots.
+sub check_slots_conflicting_status
+{
+ my ($conflicting) = @_;
+
+ if ($conflicting)
+ {
+ $res = $node_standby->safe_psql(
+ 'postgres', qq(
+ select bool_and(conflicting) from pg_replication_slots;));
+
+ is($res, 't',
+ "Logical slots are reported as conflicting");
+ }
+ else
+ {
+ $res = $node_standby->safe_psql(
+ 'postgres', qq(
+ select bool_or(conflicting) from pg_replication_slots;));
+
+ is($res, 'f',
+ "Logical slots are reported as non conflicting");
+ }
+}
+
+########################
+# Initialize primary node
+########################
+
+$node_primary->init(allows_streaming => 1, has_archiving => 1);
+$node_primary->append_conf('postgresql.conf', q{
+wal_level = 'logical'
+max_replication_slots = 4
+max_wal_senders = 4
+log_min_messages = 'debug2'
+log_error_verbosity = verbose
+});
+$node_primary->dump_info;
+$node_primary->start;
+
+$node_primary->psql('postgres', q[CREATE DATABASE testdb]);
+
+$node_primary->safe_psql('testdb', qq[SELECT * FROM pg_create_physical_replication_slot('$primary_slotname');]);
+
+# Check conflicting is NULL for physical slot
+$res = $node_primary->safe_psql(
+ 'postgres', qq[
+ SELECT conflicting is null FROM pg_replication_slots where slot_name = '$primary_slotname';]);
+
+is($res, 't',
+ "Physical slot reports conflicting as NULL");
+
+my $backup_name = 'b1';
+$node_primary->backup($backup_name);
+
+#######################
+# Initialize standby node
+#######################
+
+$node_standby->init_from_backup(
+ $node_primary, $backup_name,
+ has_streaming => 1,
+ has_restoring => 1);
+$node_standby->append_conf('postgresql.conf',
+ qq[primary_slot_name = '$primary_slotname']);
+$node_standby->start;
+$node_primary->wait_for_catchup($node_standby, 'replay', $node_primary->lsn('flush'));
+
+
+##################################################
+# Test that logical decoding on the standby
+# behaves correctly.
+##################################################
+
+# create the logical slots
+create_logical_slots();
+
+$node_primary->safe_psql('testdb', qq[CREATE TABLE decoding_test(x integer, y text);]);
+$node_primary->safe_psql('testdb', qq[INSERT INTO decoding_test(x,y) SELECT s, s::text FROM generate_series(1,10) s;]);
+
+$node_primary->wait_for_catchup($node_standby, 'replay', $node_primary->lsn('flush'));
+
+my $result = $node_standby->safe_psql('testdb',
+ qq[SELECT pg_logical_slot_get_changes('activeslot', NULL, NULL);]);
+
+# test if basic decoding works
+is(scalar(my @foobar = split /^/m, $result),
+ 14, 'Decoding produced 14 rows (2 BEGIN/COMMIT and 10 rows)');
+
+# Insert some rows and verify that we get the same results from pg_recvlogical
+# and the SQL interface.
+$node_primary->safe_psql('testdb',
+ qq[INSERT INTO decoding_test(x,y) SELECT s, s::text FROM generate_series(1,4) s;]
+);
+
+my $expected = q{BEGIN
+table public.decoding_test: INSERT: x[integer]:1 y[text]:'1'
+table public.decoding_test: INSERT: x[integer]:2 y[text]:'2'
+table public.decoding_test: INSERT: x[integer]:3 y[text]:'3'
+table public.decoding_test: INSERT: x[integer]:4 y[text]:'4'
+COMMIT};
+
+$node_primary->wait_for_catchup($node_standby, 'replay', $node_primary->lsn('flush'));
+
+my $stdout_sql = $node_standby->safe_psql('testdb',
+ qq[SELECT data FROM pg_logical_slot_peek_changes('activeslot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');]
+);
+
+is($stdout_sql, $expected, 'got expected output from SQL decoding session');
+
+my $endpos = $node_standby->safe_psql('testdb',
+ "SELECT lsn FROM pg_logical_slot_peek_changes('activeslot', NULL, NULL) ORDER BY lsn DESC LIMIT 1;"
+);
+
+# Insert some rows after $endpos, which we won't read.
+$node_primary->safe_psql('testdb',
+ qq[INSERT INTO decoding_test(x,y) SELECT s, s::text FROM generate_series(5,50) s;]
+);
+
+$node_primary->wait_for_catchup($node_standby, 'replay', $node_primary->lsn('flush'));
+
+my $stdout_recv = $node_standby->pg_recvlogical_upto(
+ 'testdb', 'activeslot', $endpos, $default_timeout,
+ 'include-xids' => '0',
+ 'skip-empty-xacts' => '1');
+chomp($stdout_recv);
+is($stdout_recv, $expected,
+ 'got same expected output from pg_recvlogical decoding session');
+
+$node_standby->poll_query_until('testdb',
+ "SELECT EXISTS (SELECT 1 FROM pg_replication_slots WHERE slot_name = 'activeslot' AND active_pid IS NULL)"
+) or die "slot never became inactive";
+
+$stdout_recv = $node_standby->pg_recvlogical_upto(
+ 'testdb', 'activeslot', $endpos, $default_timeout,
+ 'include-xids' => '0',
+ 'skip-empty-xacts' => '1');
+chomp($stdout_recv);
+is($stdout_recv, '', 'pg_recvlogical acknowledged changes');
+
+$node_primary->safe_psql('postgres', 'CREATE DATABASE otherdb');
+
+is( $node_primary->psql(
+ 'otherdb',
+ "SELECT lsn FROM pg_logical_slot_peek_changes('activeslot', NULL, NULL) ORDER BY lsn DESC LIMIT 1;"
+ ),
+ 3,
+ 'replaying logical slot from another database fails');
+
+# drop the logical slots
+drop_logical_slots();
+
+##################################################
+# Recovery conflict: Invalidate conflicting slots, including in-use slots
+# Scenario 1: hot_standby_feedback off and vacuum FULL
+##################################################
+
+# create the logical slots
+create_logical_slots();
+
+# One way to produce recovery conflict is to create/drop a relation and
+# launch a vacuum full on pg_class with hot_standby_feedback turned off on
+# the standby.
+change_hot_standby_feedback_and_wait_for_xmins(0,1);
+
+$handle = make_slot_active(1);
+
+# This should trigger the conflict
+$node_primary->safe_psql('testdb', qq[CREATE TABLE conflict_test(x integer, y text);]);
+$node_primary->safe_psql('testdb', qq[DROP TABLE conflict_test;]);
+$node_primary->safe_psql('testdb', 'VACUUM full pg_class;');
+
+$node_primary->wait_for_catchup($node_standby, 'replay', $node_primary->lsn('flush'));
+
+# message should be issued
+ok( find_in_log(
+ $node_standby,
+ "invalidating slot \"inactiveslot\" because it conflicts with recovery"),
+ 'inactiveslot slot invalidation is logged with vacuum FULL on pg_class');
+
+ok( find_in_log(
+ $node_standby,
+ "invalidating slot \"activeslot\" because it conflicts with recovery"),
+ 'activeslot slot invalidation is logged with vacuum FULL on pg_class');
+
+# Verify that pg_stat_database_conflicts.confl_active_logicalslot has been updated
+ok( $node_standby->poll_query_until(
+ 'postgres',
+ "select (confl_active_logicalslot = 1) from pg_stat_database_conflicts where datname = 'testdb'", 't'),
+ 'confl_active_logicalslot updated') or die "Timed out waiting confl_active_logicalslot to be updated";
+
+# Verify slots are reported as conflicting in pg_replication_slots
+check_slots_conflicting_status(1);
+
+$handle = make_slot_active(0);
+# We are not able to read from the slot as it has been invalidated
+check_pg_recvlogical_stderr($handle, "cannot read from logical replication slot \"activeslot\"");
+
+# Turn hot_standby_feedback back on
+change_hot_standby_feedback_and_wait_for_xmins(1,1);
+
+##################################################
+# Recovery conflict: Invalidate conflicting slots, including in-use slots
+# Scenario 2: conflict due to row removal with hot_standby_feedback off.
+##################################################
+
+# get the position to search from in the standby logfile
+my $logstart = -s $node_standby->logfile;
+
+# drop the logical slots
+drop_logical_slots();
+
+# create the logical slots
+create_logical_slots();
+
+# One way to produce recovery conflict is to create/drop a relation and
+# launch a vacuum on pg_class with hot_standby_feedback turned off on the standby.
+change_hot_standby_feedback_and_wait_for_xmins(0,1);
+
+$handle = make_slot_active(1);
+
+# This should trigger the conflict
+$node_primary->safe_psql('testdb', qq[CREATE TABLE conflict_test(x integer, y text);]);
+$node_primary->safe_psql('testdb', qq[DROP TABLE conflict_test;]);
+$node_primary->safe_psql('testdb', 'VACUUM pg_class;');
+
+$node_primary->wait_for_catchup($node_standby, 'replay', $node_primary->lsn('flush'));
+
+# message should be issued
+ok( find_in_log(
+ $node_standby,
+ "invalidating slot \"inactiveslot\" because it conflicts with recovery", $logstart),
+ 'inactiveslot slot invalidation is logged with vacuum on pg_class');
+
+ok( find_in_log(
+ $node_standby,
+ "invalidating slot \"activeslot\" because it conflicts with recovery", $logstart),
+ 'activeslot slot invalidation is logged with vacuum on pg_class');
+
+# Verify that pg_stat_database_conflicts.confl_active_logicalslot has been updated
+# we now expect 2 conflicts reported as the counter persist across reloads
+ok( $node_standby->poll_query_until(
+ 'postgres',
+ "select (confl_active_logicalslot = 2) from pg_stat_database_conflicts where datname = 'testdb'", 't'),
+ 'confl_active_logicalslot updated') or die "Timed out waiting confl_active_logicalslot to be updated";
+
+# Verify slots are reported as conflicting in pg_replication_slots
+check_slots_conflicting_status(1);
+
+$handle = make_slot_active(0);
+
+# We are not able to read from the slot as it has been invalidated
+check_pg_recvlogical_stderr($handle, "cannot read from logical replication slot \"activeslot\"");
+
+##################################################
+# Recovery conflict: Same as Scenario 2 but on a non catalog table
+# Scenario 3: No conflict expected.
+##################################################
+
+# get the position to search from in the standby logfile
+$logstart = -s $node_standby->logfile;
+
+# drop the logical slots
+drop_logical_slots();
+
+# create the logical slots
+create_logical_slots();
+
+# put hot standby feedback to off
+change_hot_standby_feedback_and_wait_for_xmins(0,1);
+
+$handle = make_slot_active(1);
+
+# This should not trigger a conflict
+$node_primary->safe_psql('testdb', qq[CREATE TABLE conflict_test(x integer, y text);]);
+$node_primary->safe_psql('testdb', qq[INSERT INTO conflict_test(x,y) SELECT s, s::text FROM generate_series(1,4) s;]);
+$node_primary->safe_psql('testdb', qq[UPDATE conflict_test set x=1, y=1;]);
+$node_primary->safe_psql('testdb', 'VACUUM conflict_test;');
+
+$node_primary->wait_for_catchup($node_standby, 'replay', $node_primary->lsn('flush'));
+
+# message should not be issued
+ok( !find_in_log(
+ $node_standby,
+ "invalidating slot \"inactiveslot\" because it conflicts with recovery", $logstart),
+ 'inactiveslot slot invalidation is not logged with vacuum on conflict_test');
+
+ok( !find_in_log(
+ $node_standby,
+ "invalidating slot \"activeslot\" because it conflicts with recovery", $logstart),
+ 'activeslot slot invalidation is not logged with vacuum on conflict_test');
+
+# Verify that pg_stat_database_conflicts.confl_active_logicalslot has not been updated
+# we now still expect 2 conflicts reported as the counter persist across reloads
+ok( $node_standby->poll_query_until(
+ 'postgres',
+ "select (confl_active_logicalslot = 2) from pg_stat_database_conflicts where datname = 'testdb'", 't'),
+ 'confl_active_logicalslot not updated') or die "Timed out waiting confl_active_logicalslot to be updated";
+
+# Verify slots are reported as non conflicting in pg_replication_slots
+check_slots_conflicting_status(0);
+
+# Turn hot_standby_feedback back on
+change_hot_standby_feedback_and_wait_for_xmins(1, 0);
+
+##################################################
+# Recovery conflict: Invalidate conflicting slots, including in-use slots
+# Scenario 4: conflict due to on-access pruning.
+##################################################
+
+# get the position to search from in the standby logfile
+$logstart = -s $node_standby->logfile;
+
+# drop the logical slots
+drop_logical_slots();
+
+# create the logical slots
+create_logical_slots();
+
+# One way to produce recovery conflict is to trigger an on-access pruning
+# on a relation marked as user_catalog_table.
+change_hot_standby_feedback_and_wait_for_xmins(0,0);
+
+$handle = make_slot_active(1);
+
+# This should trigger the conflict
+$node_primary->safe_psql('testdb', qq[CREATE TABLE prun(id integer, s char(2000)) WITH (fillfactor = 75, user_catalog_table = true);]);
+$node_primary->safe_psql('testdb', qq[INSERT INTO prun VALUES (1, 'A');]);
+$node_primary->safe_psql('testdb', qq[UPDATE prun SET s = 'B';]);
+$node_primary->safe_psql('testdb', qq[UPDATE prun SET s = 'C';]);
+$node_primary->safe_psql('testdb', qq[UPDATE prun SET s = 'D';]);
+$node_primary->safe_psql('testdb', qq[UPDATE prun SET s = 'E';]);
+
+$node_primary->wait_for_catchup($node_standby, 'replay', $node_primary->lsn('flush'));
+
+# message should be issued
+ok( find_in_log(
+ $node_standby,
+ "invalidating slot \"inactiveslot\" because it conflicts with recovery", $logstart),
+ 'inactiveslot slot invalidation is logged with on-access pruning');
+
+ok( find_in_log(
+ $node_standby,
+ "invalidating slot \"activeslot\" because it conflicts with recovery", $logstart),
+ 'activeslot slot invalidation is logged with on-access pruning');
+
+# Verify that pg_stat_database_conflicts.confl_active_logicalslot has been updated
+# we now expect 3 conflicts reported as the counter persist across reloads
+ok( $node_standby->poll_query_until(
+ 'postgres',
+ "select (confl_active_logicalslot = 3) from pg_stat_database_conflicts where datname = 'testdb'", 't'),
+ 'confl_active_logicalslot updated') or die "Timed out waiting confl_active_logicalslot to be updated";
+
+# Verify slots are reported as conflicting in pg_replication_slots
+check_slots_conflicting_status(1);
+
+$handle = make_slot_active(0);
+
+# We are not able to read from the slot as it has been invalidated
+check_pg_recvlogical_stderr($handle, "cannot read from logical replication slot \"activeslot\"");
+
+# Turn hot_standby_feedback back on
+change_hot_standby_feedback_and_wait_for_xmins(1, 1);
+
+##################################################
+# Recovery conflict: Invalidate conflicting slots, including in-use slots
+# Scenario 5: incorrect wal_level on primary.
+##################################################
+
+# get the position to search from in the standby logfile
+$logstart = -s $node_standby->logfile;
+
+# drop the logical slots
+drop_logical_slots();
+
+# create the logical slots
+create_logical_slots();
+
+$handle = make_slot_active(1);
+
+# Make primary wal_level replica. This will trigger slot conflict.
+$node_primary->append_conf('postgresql.conf',q[
+wal_level = 'replica'
+]);
+$node_primary->restart;
+
+$node_primary->wait_for_catchup($node_standby, 'replay', $node_primary->lsn('flush'));
+
+# message should be issued
+ok( find_in_log(
+ $node_standby,
+ "invalidating slot \"inactiveslot\" because it conflicts with recovery", $logstart),
+ 'inactiveslot slot invalidation is logged due to wal_level');
+
+ok( find_in_log(
+ $node_standby,
+ "invalidating slot \"activeslot\" because it conflicts with recovery", $logstart),
+ 'activeslot slot invalidation is logged due to wal_level');
+
+# Verify that pg_stat_database_conflicts.confl_active_logicalslot has been updated
+# we now expect 3 conflicts reported as the counter persist across reloads
+ok( $node_standby->poll_query_until(
+ 'postgres',
+ "select (confl_active_logicalslot = 4) from pg_stat_database_conflicts where datname = 'testdb'", 't'),
+ 'confl_active_logicalslot updated') or die "Timed out waiting confl_active_logicalslot to be updated";
+
+# Verify slots are reported as conflicting in pg_replication_slots
+check_slots_conflicting_status(1);
+
+$handle = make_slot_active(0);
+# We are not able to read from the slot as it requires wal_level at least logical on the primary server
+check_pg_recvlogical_stderr($handle, "logical decoding on standby requires wal_level to be at least logical on the primary server");
+
+# Restore primary wal_level
+$node_primary->append_conf('postgresql.conf',q[
+wal_level = 'logical'
+]);
+$node_primary->restart;
+$node_primary->wait_for_catchup($node_standby, 'replay', $node_primary->lsn('flush'));
+
+$handle = make_slot_active(0);
+# as the slot has been invalidated we should not be able to read
+check_pg_recvlogical_stderr($handle, "cannot read from logical replication slot \"activeslot\"");
+
+##################################################
+# DROP DATABASE should drops it's slots, including active slots.
+##################################################
+
+# drop the logical slots
+drop_logical_slots();
+
+# create the logical slots
+create_logical_slots();
+
+$handle = make_slot_active(1);
+# Create a slot on a database that would not be dropped. This slot should not
+# get dropped.
+$node_standby->create_logical_slot_on_standby($node_primary, 'otherslot', 'postgres');
+
+# dropdb on the primary to verify slots are dropped on standby
+$node_primary->safe_psql('postgres', q[DROP DATABASE testdb]);
+
+$node_primary->wait_for_catchup($node_standby, 'replay', $node_primary->lsn('flush'));
+
+is($node_standby->safe_psql('postgres',
+ q[SELECT EXISTS(SELECT 1 FROM pg_database WHERE datname = 'testdb')]), 'f',
+ 'database dropped on standby');
+
+check_slots_dropped($handle);
+
+is($node_standby->slot('otherslot')->{'slot_type'}, 'logical',
+ 'otherslot on standby not dropped');
+
+# Cleanup : manually drop the slot that was not dropped.
+$node_standby->psql('postgres', q[SELECT pg_drop_replication_slot('otherslot')]);
+
+##################################################
+# Test standby promotion and logical decoding behavior
+# after the standby gets promoted.
+##################################################
+
+$node_primary->psql('postgres', q[CREATE DATABASE testdb]);
+$node_primary->safe_psql('testdb', qq[CREATE TABLE decoding_test(x integer, y text);]);
+
+# create the logical slots
+create_logical_slots();
+$handle = make_slot_active(1);
+
+# Insert some rows before the promotion
+$node_primary->safe_psql('testdb',
+ qq[INSERT INTO decoding_test(x,y) SELECT s, s::text FROM generate_series(1,4) s;]
+);
+
+$node_primary->wait_for_catchup($node_standby, 'replay', $node_primary->lsn('flush'));
+
+# promote
+$node_standby->promote;
+
+# insert some rows on promoted standby
+$node_standby->safe_psql('testdb',
+ qq[INSERT INTO decoding_test(x,y) SELECT s, s::text FROM generate_series(5,7) s;]
+);
+
+$expected = q{BEGIN
+table public.decoding_test: INSERT: x[integer]:1 y[text]:'1'
+table public.decoding_test: INSERT: x[integer]:2 y[text]:'2'
+table public.decoding_test: INSERT: x[integer]:3 y[text]:'3'
+table public.decoding_test: INSERT: x[integer]:4 y[text]:'4'
+COMMIT
+BEGIN
+table public.decoding_test: INSERT: x[integer]:5 y[text]:'5'
+table public.decoding_test: INSERT: x[integer]:6 y[text]:'6'
+table public.decoding_test: INSERT: x[integer]:7 y[text]:'7'
+COMMIT};
+
+# check that we are decoding pre and post promotion inserted rows
+$stdout_sql = $node_standby->safe_psql('testdb',
+ qq[SELECT data FROM pg_logical_slot_peek_changes('inactiveslot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');]
+);
+
+is($stdout_sql, $expected, 'got expected output from SQL decoding session on promoted standby');
+
+# check that we are decoding pre and post promotion inserted rows
+# with pg_recvlogical that has started before the promotion
+my $pump_timeout = IPC::Run::timer($PostgreSQL::Test::Utils::timeout_default);
+
+ok( pump_until(
+ $handle, $pump_timeout, \$stdout, qr/^.*COMMIT.*COMMIT$/s),
+ 'got 2 COMMIT from pg_recvlogical output');
+
+chomp($stdout);
+is($stdout, $expected,
+ 'got same expected output from pg_recvlogical decoding session');
--
2.34.1
From 31f671b55fc9eb9eba2799182ae50c20ba6ed6c9 Mon Sep 17 00:00:00 2001
From: bdrouvotAWS <[email protected]>
Date: Mon, 30 Jan 2023 15:42:26 +0000
Subject: [PATCH v45 4/6] Fixing Walsender corner case with logical decoding on
standby.
The problem is that WalSndWaitForWal() waits for the *replay* LSN to
increase, but gets woken up by walreceiver when new WAL has been
flushed. Which means that typically walsenders will get woken up at the
same time that the startup process will be - which means that by the
time the logical walsender checks GetXLogReplayRecPtr() it's unlikely
that the startup process already replayed the record and updated
XLogCtl->lastReplayedEndRecPtr.
Introducing a new condition variable to fix this corner case.
---
src/backend/access/transam/xlogrecovery.c | 28 ++++++++++++++++++++
src/backend/replication/walsender.c | 31 +++++++++++++++++------
src/backend/utils/activity/wait_event.c | 3 +++
src/include/access/xlogrecovery.h | 3 +++
src/include/replication/walsender.h | 1 +
src/include/utils/wait_event.h | 1 +
6 files changed, 59 insertions(+), 8 deletions(-)
41.2% src/backend/access/transam/
48.5% src/backend/replication/
3.6% src/backend/utils/activity/
3.4% src/include/access/
diff --git a/src/backend/access/transam/xlogrecovery.c b/src/backend/access/transam/xlogrecovery.c
index 2a5352f879..bb0de527ab 100644
--- a/src/backend/access/transam/xlogrecovery.c
+++ b/src/backend/access/transam/xlogrecovery.c
@@ -358,6 +358,9 @@ typedef struct XLogRecoveryCtlData
RecoveryPauseState recoveryPauseState;
ConditionVariable recoveryNotPausedCV;
+ /* Replay state (see getReplayedCV() for more explanation) */
+ ConditionVariable replayedCV;
+
slock_t info_lck; /* locks shared variables shown above */
} XLogRecoveryCtlData;
@@ -467,6 +470,7 @@ XLogRecoveryShmemInit(void)
SpinLockInit(&XLogRecoveryCtl->info_lck);
InitSharedLatch(&XLogRecoveryCtl->recoveryWakeupLatch);
ConditionVariableInit(&XLogRecoveryCtl->recoveryNotPausedCV);
+ ConditionVariableInit(&XLogRecoveryCtl->replayedCV);
}
/*
@@ -1916,6 +1920,11 @@ ApplyWalRecord(XLogReaderState *xlogreader, XLogRecord *record, TimeLineID *repl
XLogRecoveryCtl->lastReplayedTLI = *replayTLI;
SpinLockRelease(&XLogRecoveryCtl->info_lck);
+ /*
+ * wake up walsender(s) used by logical decoding on standby.
+ */
+ ConditionVariableBroadcast(&XLogRecoveryCtl->replayedCV);
+
/*
* If rm_redo called XLogRequestWalReceiverReply, then we wake up the
* receiver so that it notices the updated lastReplayedEndRecPtr and sends
@@ -4923,3 +4932,22 @@ assign_recovery_target_xid(const char *newval, void *extra)
else
recoveryTarget = RECOVERY_TARGET_UNSET;
}
+
+/*
+ * Return the ConditionVariable indicating that a replay has been done.
+ *
+ * This is needed for logical decoding on standby. Indeed the "problem" is that
+ * WalSndWaitForWal() waits for the *replay* LSN to increase, but gets woken up
+ * by walreceiver when new WAL has been flushed. Which means that typically
+ * walsenders will get woken up at the same time that the startup process
+ * will be - which means that by the time the logical walsender checks
+ * GetXLogReplayRecPtr() it's unlikely that the startup process already replayed
+ * the record and updated XLogCtl->lastReplayedEndRecPtr.
+ *
+ * The ConditionVariable XLogRecoveryCtl->replayedCV solves this corner case.
+ */
+ConditionVariable *
+getReplayedCV(void)
+{
+ return &XLogRecoveryCtl->replayedCV;
+}
diff --git a/src/backend/replication/walsender.c b/src/backend/replication/walsender.c
index 1e91cbc564..b3fe5dbeb2 100644
--- a/src/backend/replication/walsender.c
+++ b/src/backend/replication/walsender.c
@@ -1552,6 +1552,7 @@ WalSndWaitForWal(XLogRecPtr loc)
{
int wakeEvents;
static XLogRecPtr RecentFlushPtr = InvalidXLogRecPtr;
+ ConditionVariable *replayedCV = getReplayedCV();
/*
* Fast path to avoid acquiring the spinlock in case we already know we
@@ -1570,7 +1571,6 @@ WalSndWaitForWal(XLogRecPtr loc)
for (;;)
{
- long sleeptime;
/* Clear any already-pending wakeups */
ResetLatch(MyLatch);
@@ -1654,20 +1654,35 @@ WalSndWaitForWal(XLogRecPtr loc)
WalSndKeepaliveIfNecessary();
/*
- * Sleep until something happens or we time out. Also wait for the
- * socket becoming writable, if there's still pending output.
+ * When not in recovery, sleep until something happens or we time out.
+ * Also wait for the socket becoming writable, if there's still pending output.
* Otherwise we might sit on sendable output data while waiting for
* new WAL to be generated. (But if we have nothing to send, we don't
* want to wake on socket-writable.)
*/
- sleeptime = WalSndComputeSleeptime(GetCurrentTimestamp());
+ if (!RecoveryInProgress())
+ {
+ long sleeptime;
+ sleeptime = WalSndComputeSleeptime(GetCurrentTimestamp());
- wakeEvents = WL_SOCKET_READABLE;
+ wakeEvents = WL_SOCKET_READABLE;
- if (pq_is_send_pending())
- wakeEvents |= WL_SOCKET_WRITEABLE;
+ if (pq_is_send_pending())
+ wakeEvents |= WL_SOCKET_WRITEABLE;
- WalSndWait(wakeEvents, sleeptime, WAIT_EVENT_WAL_SENDER_WAIT_WAL);
+ WalSndWait(wakeEvents, sleeptime * 10, WAIT_EVENT_WAL_SENDER_WAIT_WAL);
+ }
+ else
+ /*
+ * We are in the logical decoding on standby case.
+ * We are waiting for the startup process to replay wal record(s) using
+ * a timeout in case we are requested to stop.
+ */
+ {
+ ConditionVariablePrepareToSleep(replayedCV);
+ ConditionVariableTimedSleep(replayedCV, 1000,
+ WAIT_EVENT_WAL_SENDER_WAIT_REPLAY);
+ }
}
/* reactivate latch so WalSndLoop knows to continue */
diff --git a/src/backend/utils/activity/wait_event.c b/src/backend/utils/activity/wait_event.c
index 6e4599278c..38c747b786 100644
--- a/src/backend/utils/activity/wait_event.c
+++ b/src/backend/utils/activity/wait_event.c
@@ -463,6 +463,9 @@ pgstat_get_wait_ipc(WaitEventIPC w)
case WAIT_EVENT_WAL_RECEIVER_WAIT_START:
event_name = "WalReceiverWaitStart";
break;
+ case WAIT_EVENT_WAL_SENDER_WAIT_REPLAY:
+ event_name = "WalReceiverWaitReplay";
+ break;
case WAIT_EVENT_XACT_GROUP_UPDATE:
event_name = "XactGroupUpdate";
break;
diff --git a/src/include/access/xlogrecovery.h b/src/include/access/xlogrecovery.h
index 47c29350f5..b65c2cf1f0 100644
--- a/src/include/access/xlogrecovery.h
+++ b/src/include/access/xlogrecovery.h
@@ -15,6 +15,7 @@
#include "catalog/pg_control.h"
#include "lib/stringinfo.h"
#include "utils/timestamp.h"
+#include "storage/condition_variable.h"
/*
* Recovery target type.
@@ -155,4 +156,6 @@ extern void RecoveryRequiresIntParameter(const char *param_name, int currValue,
extern void xlog_outdesc(StringInfo buf, XLogReaderState *record);
+extern ConditionVariable *getReplayedCV(void);
+
#endif /* XLOGRECOVERY_H */
diff --git a/src/include/replication/walsender.h b/src/include/replication/walsender.h
index 52bb3e2aae..2fd745fe72 100644
--- a/src/include/replication/walsender.h
+++ b/src/include/replication/walsender.h
@@ -13,6 +13,7 @@
#define _WALSENDER_H
#include <signal.h>
+#include "storage/condition_variable.h"
/*
* What to do with a snapshot in create replication slot command.
diff --git a/src/include/utils/wait_event.h b/src/include/utils/wait_event.h
index 6cacd6edaf..04a37feee4 100644
--- a/src/include/utils/wait_event.h
+++ b/src/include/utils/wait_event.h
@@ -130,6 +130,7 @@ typedef enum
WAIT_EVENT_SYNC_REP,
WAIT_EVENT_WAL_RECEIVER_EXIT,
WAIT_EVENT_WAL_RECEIVER_WAIT_START,
+ WAIT_EVENT_WAL_SENDER_WAIT_REPLAY,
WAIT_EVENT_XACT_GROUP_UPDATE
} WaitEventIPC;
--
2.34.1
From d184a98f9c9aeefdcfbeeebab3aba4bf0cee9815 Mon Sep 17 00:00:00 2001
From: bdrouvotAWS <[email protected]>
Date: Mon, 30 Jan 2023 15:41:38 +0000
Subject: [PATCH v45 3/6] Allow logical decoding on standby.
Allow a logical slot to be created on standby. Restrict its usage
or its creation if wal_level on primary is less than logical.
During slot creation, it's restart_lsn is set to the last replayed
LSN. Effectively, a logical slot creation on standby waits for an
xl_running_xact record to arrive from primary.
Author: Andres Freund (in an older version), Amit Khandekar, Bertrand Drouvot
Reviewed-By: Bertrand Drouvot, Andres Freund, Robert Haas, Fabrizio de Royes Mello
---
src/backend/access/transam/xlog.c | 11 +++++
src/backend/replication/logical/decode.c | 22 ++++++++-
src/backend/replication/logical/logical.c | 37 ++++++++-------
src/backend/replication/slot.c | 57 ++++++++++++-----------
src/backend/replication/walsender.c | 41 ++++++++++------
src/include/access/xlog.h | 1 +
6 files changed, 111 insertions(+), 58 deletions(-)
4.7% src/backend/access/transam/
38.7% src/backend/replication/logical/
55.6% src/backend/replication/
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 867675d5a1..1abe747cb5 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -4465,6 +4465,17 @@ LocalProcessControlFile(bool reset)
ReadControlFile();
}
+/*
+ * Get the wal_level from the control file. For a standby, this value should be
+ * considered as its active wal_level, because it may be different from what
+ * was originally configured on standby.
+ */
+WalLevel
+GetActiveWalLevelOnStandby(void)
+{
+ return ControlFile->wal_level;
+}
+
/*
* Initialization of shared memory for XLOG
*/
diff --git a/src/backend/replication/logical/decode.c b/src/backend/replication/logical/decode.c
index a53e23c679..6b66a971ba 100644
--- a/src/backend/replication/logical/decode.c
+++ b/src/backend/replication/logical/decode.c
@@ -152,11 +152,31 @@ xlog_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
* can restart from there.
*/
break;
+ case XLOG_PARAMETER_CHANGE:
+ {
+ xl_parameter_change *xlrec =
+ (xl_parameter_change *) XLogRecGetData(buf->record);
+
+ /*
+ * If wal_level on primary is reduced to less than logical, then we
+ * want to prevent existing logical slots from being used.
+ * Existing logical slots on standby get invalidated when this WAL
+ * record is replayed; and further, slot creation fails when the
+ * wal level is not sufficient; but all these operations are not
+ * synchronized, so a logical slot may creep in while the wal_level
+ * is being reduced. Hence this extra check.
+ */
+ if (xlrec->wal_level < WAL_LEVEL_LOGICAL)
+ ereport(ERROR,
+ (errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
+ errmsg("logical decoding on standby requires wal_level "
+ "to be at least logical on the primary server")));
+ break;
+ }
case XLOG_NOOP:
case XLOG_NEXTOID:
case XLOG_SWITCH:
case XLOG_BACKUP_END:
- case XLOG_PARAMETER_CHANGE:
case XLOG_RESTORE_POINT:
case XLOG_FPW_CHANGE:
case XLOG_FPI_FOR_HINT:
diff --git a/src/backend/replication/logical/logical.c b/src/backend/replication/logical/logical.c
index 1a58dd7649..91acc0c155 100644
--- a/src/backend/replication/logical/logical.c
+++ b/src/backend/replication/logical/logical.c
@@ -119,23 +119,22 @@ CheckLogicalDecodingRequirements(void)
(errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
errmsg("logical decoding requires a database connection")));
- /* ----
- * TODO: We got to change that someday soon...
- *
- * There's basically three things missing to allow this:
- * 1) We need to be able to correctly and quickly identify the timeline a
- * LSN belongs to
- * 2) We need to force hot_standby_feedback to be enabled at all times so
- * the primary cannot remove rows we need.
- * 3) support dropping replication slots referring to a database, in
- * dbase_redo. There can't be any active ones due to HS recovery
- * conflicts, so that should be relatively easy.
- * ----
- */
if (RecoveryInProgress())
- ereport(ERROR,
- (errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
- errmsg("logical decoding cannot be used while in recovery")));
+ {
+ /*
+ * This check may have race conditions, but whenever
+ * XLOG_PARAMETER_CHANGE indicates that wal_level has changed, we
+ * verify that there are no existing logical replication slots. And to
+ * avoid races around creating a new slot,
+ * CheckLogicalDecodingRequirements() is called once before creating
+ * the slot, and once when logical decoding is initially starting up.
+ */
+ if (GetActiveWalLevelOnStandby() < WAL_LEVEL_LOGICAL)
+ ereport(ERROR,
+ (errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
+ errmsg("logical decoding on standby requires wal_level "
+ "to be at least logical on the primary server")));
+ }
}
/*
@@ -331,6 +330,12 @@ CreateInitDecodingContext(const char *plugin,
LogicalDecodingContext *ctx;
MemoryContext old_context;
+ /*
+ * On standby, this check is also required while creating the slot. Check
+ * the comments in this function.
+ */
+ CheckLogicalDecodingRequirements();
+
/* shorter lines... */
slot = MyReplicationSlot;
diff --git a/src/backend/replication/slot.c b/src/backend/replication/slot.c
index 38c6f18886..290d4b45f4 100644
--- a/src/backend/replication/slot.c
+++ b/src/backend/replication/slot.c
@@ -51,6 +51,7 @@
#include "storage/proc.h"
#include "storage/procarray.h"
#include "utils/builtins.h"
+#include "access/xlogrecovery.h"
/*
* Replication slot on-disk data structure.
@@ -1177,37 +1178,28 @@ ReplicationSlotReserveWal(void)
/*
* For logical slots log a standby snapshot and start logical decoding
* at exactly that position. That allows the slot to start up more
- * quickly.
+ * quickly. But on a standby we cannot do WAL writes, so just use the
+ * replay pointer; effectively, an attempt to create a logical slot on
+ * standby will cause it to wait for an xl_running_xact record to be
+ * logged independently on the primary, so that a snapshot can be built
+ * using the record.
*
- * That's not needed (or indeed helpful) for physical slots as they'll
- * start replay at the last logged checkpoint anyway. Instead return
- * the location of the last redo LSN. While that slightly increases
- * the chance that we have to retry, it's where a base backup has to
- * start replay at.
+ * None of this is needed (or indeed helpful) for physical slots as
+ * they'll start replay at the last logged checkpoint anyway. Instead
+ * return the location of the last redo LSN. While that slightly
+ * increases the chance that we have to retry, it's where a base backup
+ * has to start replay at.
*/
- if (!RecoveryInProgress() && SlotIsLogical(slot))
- {
- XLogRecPtr flushptr;
-
- /* start at current insert position */
+ if (SlotIsPhysical(slot))
+ restart_lsn = GetRedoRecPtr();
+ else if (RecoveryInProgress())
+ restart_lsn = GetXLogReplayRecPtr(NULL);
+ else
restart_lsn = GetXLogInsertRecPtr();
- SpinLockAcquire(&slot->mutex);
- slot->data.restart_lsn = restart_lsn;
- SpinLockRelease(&slot->mutex);
-
- /* make sure we have enough information to start */
- flushptr = LogStandbySnapshot();
- /* and make sure it's fsynced to disk */
- XLogFlush(flushptr);
- }
- else
- {
- restart_lsn = GetRedoRecPtr();
- SpinLockAcquire(&slot->mutex);
- slot->data.restart_lsn = restart_lsn;
- SpinLockRelease(&slot->mutex);
- }
+ SpinLockAcquire(&slot->mutex);
+ slot->data.restart_lsn = restart_lsn;
+ SpinLockRelease(&slot->mutex);
/* prevent WAL removal as fast as possible */
ReplicationSlotsComputeRequiredLSN();
@@ -1223,6 +1215,17 @@ ReplicationSlotReserveWal(void)
if (XLogGetLastRemovedSegno() < segno)
break;
}
+
+ if (!RecoveryInProgress() && SlotIsLogical(slot))
+ {
+ XLogRecPtr flushptr;
+
+ /* make sure we have enough information to start */
+ flushptr = LogStandbySnapshot();
+
+ /* and make sure it's fsynced to disk */
+ XLogFlush(flushptr);
+ }
}
/*
diff --git a/src/backend/replication/walsender.c b/src/backend/replication/walsender.c
index 8885cdeebc..1e91cbc564 100644
--- a/src/backend/replication/walsender.c
+++ b/src/backend/replication/walsender.c
@@ -906,23 +906,31 @@ logical_read_xlog_page(XLogReaderState *state, XLogRecPtr targetPagePtr, int req
int count;
WALReadError errinfo;
XLogSegNo segno;
- TimeLineID currTLI = GetWALInsertionTimeLine();
+ TimeLineID currTLI;
/*
- * Since logical decoding is only permitted on a primary server, we know
- * that the current timeline ID can't be changing any more. If we did this
- * on a standby, we'd have to worry about the values we compute here
- * becoming invalid due to a promotion or timeline change.
+ * Since logical decoding is also permitted on a standby server, we need
+ * to check if the server is in recovery to decide how to get the current
+ * timeline ID (so that it also cover the promotion or timeline change cases).
*/
+
+ /* make sure we have enough WAL available */
+ flushptr = WalSndWaitForWal(targetPagePtr + reqLen);
+
+ /* the standby could have been promoted, so check if still in recovery */
+ am_cascading_walsender = RecoveryInProgress();
+
+ if (am_cascading_walsender)
+ GetXLogReplayRecPtr(&currTLI);
+ else
+ currTLI = GetWALInsertionTimeLine();
+
XLogReadDetermineTimeline(state, targetPagePtr, reqLen, currTLI);
sendTimeLineIsHistoric = (state->currTLI != currTLI);
sendTimeLine = state->currTLI;
sendTimeLineValidUpto = state->currTLIValidUntil;
sendTimeLineNextTLI = state->nextTLI;
- /* make sure we have enough WAL available */
- flushptr = WalSndWaitForWal(targetPagePtr + reqLen);
-
/* fail if not (implies we are going to shut down) */
if (flushptr < targetPagePtr + reqLen)
return -1;
@@ -937,7 +945,7 @@ logical_read_xlog_page(XLogReaderState *state, XLogRecPtr targetPagePtr, int req
cur_page,
targetPagePtr,
XLOG_BLCKSZ,
- state->seg.ws_tli, /* Pass the current TLI because only
+ currTLI, /* Pass the current TLI because only
* WalSndSegmentOpen controls whether new
* TLI is needed. */
&errinfo))
@@ -3074,10 +3082,14 @@ XLogSendLogical(void)
* If first time through in this session, initialize flushPtr. Otherwise,
* we only need to update flushPtr if EndRecPtr is past it.
*/
- if (flushPtr == InvalidXLogRecPtr)
- flushPtr = GetFlushRecPtr(NULL);
- else if (logical_decoding_ctx->reader->EndRecPtr >= flushPtr)
- flushPtr = GetFlushRecPtr(NULL);
+ if (flushPtr == InvalidXLogRecPtr ||
+ logical_decoding_ctx->reader->EndRecPtr >= flushPtr)
+ {
+ if (am_cascading_walsender)
+ flushPtr = GetStandbyFlushRecPtr(NULL);
+ else
+ flushPtr = GetFlushRecPtr(NULL);
+ }
/* If EndRecPtr is still past our flushPtr, it means we caught up. */
if (logical_decoding_ctx->reader->EndRecPtr >= flushPtr)
@@ -3168,7 +3180,8 @@ GetStandbyFlushRecPtr(TimeLineID *tli)
receivePtr = GetWalRcvFlushRecPtr(NULL, &receiveTLI);
replayPtr = GetXLogReplayRecPtr(&replayTLI);
- *tli = replayTLI;
+ if (tli)
+ *tli = replayTLI;
result = replayPtr;
if (receiveTLI == replayTLI && receivePtr > replayPtr)
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index cfe5409738..48ca852381 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -230,6 +230,7 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(void);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
extern void CreateCheckPoint(int flags);
--
2.34.1
From 700c26ef8ee1c5caa40c6e0ce14796f7662c6f97 Mon Sep 17 00:00:00 2001
From: bdrouvotAWS <[email protected]>
Date: Mon, 30 Jan 2023 15:40:56 +0000
Subject: [PATCH v45 2/6] Handle logical slot conflicts on standby.
During WAL replay on standby, when slot conflict is identified,
invalidate such slots. Also do the same thing if wal_level on the primary server
is reduced to below logical and there are existing logical slots
on standby. Introduce a new ProcSignalReason value for slot
conflict recovery. Arrange for a new pg_stat_database_conflicts field:
confl_active_logicalslot.
Add a new field "conflicting" in pg_replication_slots.
Author: Andres Freund (in an older version), Amit Khandekar, Bertrand Drouvot
Reviewed-By: Bertrand Drouvot, Andres Freund, Robert Haas, Fabrizio de Royes Mello,
Bharath Rupireddy
---
doc/src/sgml/monitoring.sgml | 11 +
doc/src/sgml/system-views.sgml | 10 +
src/backend/access/gist/gistxlog.c | 2 +
src/backend/access/hash/hash_xlog.c | 1 +
src/backend/access/heap/heapam.c | 3 +
src/backend/access/nbtree/nbtxlog.c | 2 +
src/backend/access/spgist/spgxlog.c | 1 +
src/backend/access/transam/xlog.c | 24 ++-
src/backend/catalog/system_views.sql | 6 +-
.../replication/logical/logicalfuncs.c | 13 +-
src/backend/replication/slot.c | 198 +++++++++++++-----
src/backend/replication/slotfuncs.c | 13 +-
src/backend/replication/walsender.c | 8 +
src/backend/storage/ipc/procsignal.c | 3 +
src/backend/storage/ipc/standby.c | 13 +-
src/backend/tcop/postgres.c | 24 +++
src/backend/utils/activity/pgstat_database.c | 4 +
src/backend/utils/adt/pgstatfuncs.c | 3 +
src/include/catalog/pg_proc.dat | 11 +-
src/include/pgstat.h | 1 +
src/include/replication/slot.h | 5 +-
src/include/storage/procsignal.h | 1 +
src/include/storage/standby.h | 2 +
src/test/regress/expected/rules.out | 8 +-
24 files changed, 304 insertions(+), 63 deletions(-)
5.4% doc/src/sgml/
7.2% src/backend/access/transam/
4.7% src/backend/replication/logical/
56.8% src/backend/replication/
4.5% src/backend/storage/ipc/
6.5% src/backend/tcop/
5.4% src/backend/
3.9% src/include/catalog/
3.0% src/include/replication/
diff --git a/doc/src/sgml/monitoring.sgml b/doc/src/sgml/monitoring.sgml
index 1756f1a4b6..e25f71a776 100644
--- a/doc/src/sgml/monitoring.sgml
+++ b/doc/src/sgml/monitoring.sgml
@@ -4365,6 +4365,17 @@ SELECT pid, wait_event_type, wait_event FROM pg_stat_activity WHERE wait_event i
deadlocks
</para></entry>
</row>
+
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ <structfield>confl_active_logicalslot</structfield> <type>bigint</type>
+ </para>
+ <para>
+ Number of active logical slots in this database that have been
+ invalidated because they conflict with recovery (note that inactive ones
+ are also invalidated but do not increment this counter)
+ </para></entry>
+ </row>
</tbody>
</tgroup>
</table>
diff --git a/doc/src/sgml/system-views.sgml b/doc/src/sgml/system-views.sgml
index 7c8fc3f654..239f713295 100644
--- a/doc/src/sgml/system-views.sgml
+++ b/doc/src/sgml/system-views.sgml
@@ -2516,6 +2516,16 @@ SELECT * FROM pg_locks pl LEFT JOIN pg_prepared_xacts ppx
false for physical slots.
</para></entry>
</row>
+
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ <structfield>conflicting</structfield> <type>bool</type>
+ </para>
+ <para>
+ True if this logical slot conflicted with recovery (and so is now
+ invalidated). Always NULL for physical slots.
+ </para></entry>
+ </row>
</tbody>
</tgroup>
</table>
diff --git a/src/backend/access/gist/gistxlog.c b/src/backend/access/gist/gistxlog.c
index b7678f3c14..9a86fb3fef 100644
--- a/src/backend/access/gist/gistxlog.c
+++ b/src/backend/access/gist/gistxlog.c
@@ -197,6 +197,7 @@ gistRedoDeleteRecord(XLogReaderState *record)
XLogRecGetBlockTag(record, 0, &rlocator, NULL, NULL);
ResolveRecoveryConflictWithSnapshot(xldata->snapshotConflictHorizon,
+ xldata->isCatalogRel,
rlocator);
}
@@ -390,6 +391,7 @@ gistRedoPageReuse(XLogReaderState *record)
*/
if (InHotStandby)
ResolveRecoveryConflictWithSnapshotFullXid(xlrec->snapshotConflictHorizon,
+ xlrec->isCatalogRel,
xlrec->locator);
}
diff --git a/src/backend/access/hash/hash_xlog.c b/src/backend/access/hash/hash_xlog.c
index 08ceb91288..b856304746 100644
--- a/src/backend/access/hash/hash_xlog.c
+++ b/src/backend/access/hash/hash_xlog.c
@@ -1003,6 +1003,7 @@ hash_xlog_vacuum_one_page(XLogReaderState *record)
XLogRecGetBlockTag(record, 0, &rlocator, NULL, NULL);
ResolveRecoveryConflictWithSnapshot(xldata->snapshotConflictHorizon,
+ xldata->isCatalogRel,
rlocator);
}
diff --git a/src/backend/access/heap/heapam.c b/src/backend/access/heap/heapam.c
index d478724b9d..d64fb4cc84 100644
--- a/src/backend/access/heap/heapam.c
+++ b/src/backend/access/heap/heapam.c
@@ -8891,6 +8891,7 @@ heap_xlog_prune(XLogReaderState *record)
*/
if (InHotStandby)
ResolveRecoveryConflictWithSnapshot(xlrec->snapshotConflictHorizon,
+ xlrec->isCatalogRel,
rlocator);
/*
@@ -9060,6 +9061,7 @@ heap_xlog_visible(XLogReaderState *record)
*/
if (InHotStandby)
ResolveRecoveryConflictWithSnapshot(xlrec->snapshotConflictHorizon,
+ xlrec->flags & VISIBILITYMAP_IS_CATALOG_REL,
rlocator);
/*
@@ -9177,6 +9179,7 @@ heap_xlog_freeze_page(XLogReaderState *record)
XLogRecGetBlockTag(record, 0, &rlocator, NULL, NULL);
ResolveRecoveryConflictWithSnapshot(xlrec->snapshotConflictHorizon,
+ xlrec->isCatalogRel,
rlocator);
}
diff --git a/src/backend/access/nbtree/nbtxlog.c b/src/backend/access/nbtree/nbtxlog.c
index 414ca4f6de..c87e46ed66 100644
--- a/src/backend/access/nbtree/nbtxlog.c
+++ b/src/backend/access/nbtree/nbtxlog.c
@@ -669,6 +669,7 @@ btree_xlog_delete(XLogReaderState *record)
XLogRecGetBlockTag(record, 0, &rlocator, NULL, NULL);
ResolveRecoveryConflictWithSnapshot(xlrec->snapshotConflictHorizon,
+ xlrec->isCatalogRel,
rlocator);
}
@@ -1007,6 +1008,7 @@ btree_xlog_reuse_page(XLogReaderState *record)
if (InHotStandby)
ResolveRecoveryConflictWithSnapshotFullXid(xlrec->snapshotConflictHorizon,
+ xlrec->isCatalogRel,
xlrec->locator);
}
diff --git a/src/backend/access/spgist/spgxlog.c b/src/backend/access/spgist/spgxlog.c
index b071b59c8a..459ac929ba 100644
--- a/src/backend/access/spgist/spgxlog.c
+++ b/src/backend/access/spgist/spgxlog.c
@@ -879,6 +879,7 @@ spgRedoVacuumRedirect(XLogReaderState *record)
XLogRecGetBlockTag(record, 0, &locator, NULL, NULL);
ResolveRecoveryConflictWithSnapshot(xldata->snapshotConflictHorizon,
+ xldata->isCatalogRel,
locator);
}
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index fb4c860bde..867675d5a1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -6447,6 +6447,7 @@ CreateCheckPoint(int flags)
VirtualTransactionId *vxids;
int nvxids;
int oldXLogAllowed = 0;
+ bool invalidated = false;
/*
* An end-of-recovery checkpoint is really a shutdown checkpoint, just
@@ -6807,7 +6808,8 @@ CreateCheckPoint(int flags)
*/
XLByteToSeg(RedoRecPtr, _logSegNo, wal_segment_size);
KeepLogSeg(recptr, &_logSegNo);
- if (InvalidateObsoleteReplicationSlots(_logSegNo))
+ InvalidateObsoleteReplicationSlots(_logSegNo, &invalidated, InvalidOid, NULL);
+ if (invalidated)
{
/*
* Some slots have been invalidated; recalculate the old-segment
@@ -7086,6 +7088,7 @@ CreateRestartPoint(int flags)
XLogRecPtr endptr;
XLogSegNo _logSegNo;
TimestampTz xtime;
+ bool invalidated = false;
/* Concurrent checkpoint/restartpoint cannot happen */
Assert(!IsUnderPostmaster || MyBackendType == B_CHECKPOINTER);
@@ -7251,7 +7254,8 @@ CreateRestartPoint(int flags)
replayPtr = GetXLogReplayRecPtr(&replayTLI);
endptr = (receivePtr < replayPtr) ? replayPtr : receivePtr;
KeepLogSeg(endptr, &_logSegNo);
- if (InvalidateObsoleteReplicationSlots(_logSegNo))
+ InvalidateObsoleteReplicationSlots(_logSegNo, &invalidated, InvalidOid, NULL);
+ if (invalidated)
{
/*
* Some slots have been invalidated; recalculate the old-segment
@@ -7966,6 +7970,22 @@ xlog_redo(XLogReaderState *record)
/* Update our copy of the parameters in pg_control */
memcpy(&xlrec, XLogRecGetData(record), sizeof(xl_parameter_change));
+ /*
+ * Invalidate logical slots if we are in hot standby and the primary does not
+ * have a WAL level sufficient for logical decoding. No need to search
+ * for potentially conflicting logically slots if standby is running
+ * with wal_level lower than logical, because in that case, we would
+ * have either disallowed creation of logical slots or invalidated existing
+ * ones.
+ */
+ if (InRecovery && InHotStandby &&
+ xlrec.wal_level < WAL_LEVEL_LOGICAL &&
+ wal_level >= WAL_LEVEL_LOGICAL)
+ {
+ TransactionId ConflictHorizon = InvalidTransactionId;
+ InvalidateObsoleteReplicationSlots(InvalidXLogRecPtr, NULL, InvalidOid, &ConflictHorizon);
+ }
+
LWLockAcquire(ControlFileLock, LW_EXCLUSIVE);
ControlFile->MaxConnections = xlrec.MaxConnections;
ControlFile->max_worker_processes = xlrec.max_worker_processes;
diff --git a/src/backend/catalog/system_views.sql b/src/backend/catalog/system_views.sql
index 8608e3fa5b..a272bd4a88 100644
--- a/src/backend/catalog/system_views.sql
+++ b/src/backend/catalog/system_views.sql
@@ -997,7 +997,8 @@ CREATE VIEW pg_replication_slots AS
L.confirmed_flush_lsn,
L.wal_status,
L.safe_wal_size,
- L.two_phase
+ L.two_phase,
+ L.conflicting
FROM pg_get_replication_slots() AS L
LEFT JOIN pg_database D ON (L.datoid = D.oid);
@@ -1065,7 +1066,8 @@ CREATE VIEW pg_stat_database_conflicts AS
pg_stat_get_db_conflict_lock(D.oid) AS confl_lock,
pg_stat_get_db_conflict_snapshot(D.oid) AS confl_snapshot,
pg_stat_get_db_conflict_bufferpin(D.oid) AS confl_bufferpin,
- pg_stat_get_db_conflict_startup_deadlock(D.oid) AS confl_deadlock
+ pg_stat_get_db_conflict_startup_deadlock(D.oid) AS confl_deadlock,
+ pg_stat_get_db_conflict_logicalslot(D.oid) AS confl_active_logicalslot
FROM pg_database D;
CREATE VIEW pg_stat_user_functions AS
diff --git a/src/backend/replication/logical/logicalfuncs.c b/src/backend/replication/logical/logicalfuncs.c
index fa1b641a2b..070fd378e8 100644
--- a/src/backend/replication/logical/logicalfuncs.c
+++ b/src/backend/replication/logical/logicalfuncs.c
@@ -216,9 +216,9 @@ pg_logical_slot_get_changes_guts(FunctionCallInfo fcinfo, bool confirm, bool bin
/*
* After the sanity checks in CreateDecodingContext, make sure the
- * restart_lsn is valid. Avoid "cannot get changes" wording in this
- * errmsg because that'd be confusingly ambiguous about no changes
- * being available.
+ * restart_lsn is valid or both xmin and catalog_xmin are valid. Avoid
+ * "cannot get changes" wording in this errmsg because that'd be
+ * confusingly ambiguous about no changes being available.
*/
if (XLogRecPtrIsInvalid(MyReplicationSlot->data.restart_lsn))
ereport(ERROR,
@@ -227,6 +227,13 @@ pg_logical_slot_get_changes_guts(FunctionCallInfo fcinfo, bool confirm, bool bin
NameStr(*name)),
errdetail("This slot has never previously reserved WAL, or it has been invalidated.")));
+ if (LogicalReplicationSlotIsInvalid(MyReplicationSlot))
+ ereport(ERROR,
+ (errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
+ errmsg("cannot read from logical replication slot \"%s\"",
+ NameStr(*name)),
+ errdetail("This slot has been invalidated because it was conflicting with recovery.")));
+
MemoryContextSwitchTo(oldcontext);
/*
diff --git a/src/backend/replication/slot.c b/src/backend/replication/slot.c
index f286918f69..38c6f18886 100644
--- a/src/backend/replication/slot.c
+++ b/src/backend/replication/slot.c
@@ -855,8 +855,10 @@ ReplicationSlotsComputeRequiredXmin(bool already_locked)
SpinLockAcquire(&s->mutex);
effective_xmin = s->effective_xmin;
effective_catalog_xmin = s->effective_catalog_xmin;
- invalidated = (!XLogRecPtrIsInvalid(s->data.invalidated_at) &&
- XLogRecPtrIsInvalid(s->data.restart_lsn));
+ invalidated = ((!XLogRecPtrIsInvalid(s->data.invalidated_at) &&
+ XLogRecPtrIsInvalid(s->data.restart_lsn))
+ || (!TransactionIdIsValid(s->data.xmin) &&
+ !TransactionIdIsValid(s->data.catalog_xmin)));
SpinLockRelease(&s->mutex);
/* invalidated slots need not apply */
@@ -1224,20 +1226,21 @@ ReplicationSlotReserveWal(void)
}
/*
- * Helper for InvalidateObsoleteReplicationSlots -- acquires the given slot
- * and mark it invalid, if necessary and possible.
+ * Helper for InvalidateObsoleteReplicationSlots
+ *
+ * Acquires the given slot and mark it invalid, if necessary and possible.
*
* Returns whether ReplicationSlotControlLock was released in the interim (and
* in that case we're not holding the lock at return, otherwise we are).
*
- * Sets *invalidated true if the slot was invalidated. (Untouched otherwise.)
+ * Sets *invalidated true if an obsolete slot was invalidated. (Untouched otherwise.)
*
* This is inherently racy, because we release the LWLock
* for syscalls, so caller must restart if we return true.
*/
static bool
-InvalidatePossiblyObsoleteSlot(ReplicationSlot *s, XLogRecPtr oldestLSN,
- bool *invalidated)
+InvalidatePossiblyObsoleteOrConflictingLogicalSlot(ReplicationSlot *s, XLogRecPtr oldestLSN,
+ bool *invalidated, TransactionId *xid)
{
int last_signaled_pid = 0;
bool released_lock = false;
@@ -1245,6 +1248,9 @@ InvalidatePossiblyObsoleteSlot(ReplicationSlot *s, XLogRecPtr oldestLSN,
for (;;)
{
XLogRecPtr restart_lsn;
+ TransactionId slot_xmin;
+ TransactionId slot_catalog_xmin;
+
NameData slotname;
int active_pid = 0;
@@ -1261,18 +1267,33 @@ InvalidatePossiblyObsoleteSlot(ReplicationSlot *s, XLogRecPtr oldestLSN,
* Check if the slot needs to be invalidated. If it needs to be
* invalidated, and is not currently acquired, acquire it and mark it
* as having been invalidated. We do this with the spinlock held to
- * avoid race conditions -- for example the restart_lsn could move
- * forward, or the slot could be dropped.
+ * avoid race conditions -- for example the restart_lsn (or the
+ * xmin(s) could) move forward or the slot could be dropped.
*/
SpinLockAcquire(&s->mutex);
restart_lsn = s->data.restart_lsn;
+ slot_xmin = s->data.xmin;
+ slot_catalog_xmin = s->data.catalog_xmin;
+
+ /* slot has been invalidated (logical decoding conflict case) */
+ if ((xid &&
+ ((LogicalReplicationSlotIsInvalid(s))
+ ||
/*
- * If the slot is already invalid or is fresh enough, we don't need to
- * do anything.
+ * We are not forcing for invalidation because the xid is valid and
+ * this is a non conflicting slot.
*/
- if (XLogRecPtrIsInvalid(restart_lsn) || restart_lsn >= oldestLSN)
+ (TransactionIdIsValid(*xid) && !(
+ (TransactionIdIsValid(slot_xmin) && TransactionIdPrecedesOrEquals(slot_xmin, *xid))
+ ||
+ (TransactionIdIsValid(slot_catalog_xmin) && TransactionIdPrecedesOrEquals(slot_catalog_xmin, *xid))
+ ))
+ ))
+ ||
+ /* slot has been invalidated (obsolete LSN case) */
+ (!xid && (XLogRecPtrIsInvalid(restart_lsn) || restart_lsn >= oldestLSN)))
{
SpinLockRelease(&s->mutex);
if (released_lock)
@@ -1292,9 +1313,16 @@ InvalidatePossiblyObsoleteSlot(ReplicationSlot *s, XLogRecPtr oldestLSN,
{
MyReplicationSlot = s;
s->active_pid = MyProcPid;
- s->data.invalidated_at = restart_lsn;
- s->data.restart_lsn = InvalidXLogRecPtr;
-
+ if (xid)
+ {
+ s->data.xmin = InvalidTransactionId;
+ s->data.catalog_xmin = InvalidTransactionId;
+ }
+ else
+ {
+ s->data.invalidated_at = restart_lsn;
+ s->data.restart_lsn = InvalidXLogRecPtr;
+ }
/* Let caller know */
*invalidated = true;
}
@@ -1327,15 +1355,39 @@ InvalidatePossiblyObsoleteSlot(ReplicationSlot *s, XLogRecPtr oldestLSN,
*/
if (last_signaled_pid != active_pid)
{
- ereport(LOG,
- errmsg("terminating process %d to release replication slot \"%s\"",
- active_pid, NameStr(slotname)),
- errdetail("The slot's restart_lsn %X/%X exceeds the limit by %llu bytes.",
- LSN_FORMAT_ARGS(restart_lsn),
- (unsigned long long) (oldestLSN - restart_lsn)),
- errhint("You might need to increase max_slot_wal_keep_size."));
+ if (xid)
+ {
+ if (TransactionIdIsValid(*xid))
+ {
+ ereport(LOG,
+ errmsg("terminating process %d because replication slot \"%s\" conflicts with recovery",
+ active_pid, NameStr(slotname)),
+ errdetail("The slot conflicted with xid horizon %u.",
+ *xid));
+ }
+ else
+ {
+ ereport(LOG,
+ errmsg("terminating process %d because replication slot \"%s\" conflicts with recovery",
+ active_pid, NameStr(slotname)),
+ errdetail("Logical decoding on standby requires wal_level to be at least logical on the primary server"));
+ }
+
+ (void) SendProcSignal(active_pid, PROCSIG_RECOVERY_CONFLICT_LOGICALSLOT, InvalidBackendId);
+ }
+ else
+ {
+ ereport(LOG,
+ errmsg("terminating process %d to release replication slot \"%s\"",
+ active_pid, NameStr(slotname)),
+ errdetail("The slot's restart_lsn %X/%X exceeds the limit by %llu bytes.",
+ LSN_FORMAT_ARGS(restart_lsn),
+ (unsigned long long) (oldestLSN - restart_lsn)),
+ errhint("You might need to increase max_slot_wal_keep_size."));
+
+ (void) kill(active_pid, SIGTERM);
+ }
- (void) kill(active_pid, SIGTERM);
last_signaled_pid = active_pid;
}
@@ -1369,13 +1421,33 @@ InvalidatePossiblyObsoleteSlot(ReplicationSlot *s, XLogRecPtr oldestLSN,
ReplicationSlotSave();
ReplicationSlotRelease();
- ereport(LOG,
- errmsg("invalidating obsolete replication slot \"%s\"",
- NameStr(slotname)),
- errdetail("The slot's restart_lsn %X/%X exceeds the limit by %llu bytes.",
- LSN_FORMAT_ARGS(restart_lsn),
- (unsigned long long) (oldestLSN - restart_lsn)),
- errhint("You might need to increase max_slot_wal_keep_size."));
+ if (xid)
+ {
+ pgstat_drop_replslot(s);
+
+ if (TransactionIdIsValid(*xid))
+ {
+ ereport(LOG,
+ errmsg("invalidating slot \"%s\" because it conflicts with recovery", NameStr(slotname)),
+ errdetail("The slot conflicted with xid horizon %u.", *xid));
+ }
+ else
+ {
+ ereport(LOG,
+ errmsg("invalidating slot \"%s\" because it conflicts with recovery", NameStr(slotname)),
+ errdetail("Logical decoding on standby requires wal_level to be at least logical on the primary server"));
+ }
+ }
+ else
+ {
+ ereport(LOG,
+ errmsg("invalidating obsolete replication slot \"%s\"",
+ NameStr(slotname)),
+ errdetail("The slot's restart_lsn %X/%X exceeds the limit by %llu bytes.",
+ LSN_FORMAT_ARGS(restart_lsn),
+ (unsigned long long) (oldestLSN - restart_lsn)),
+ errhint("You might need to increase max_slot_wal_keep_size."));
+ }
/* done with this slot for now */
break;
@@ -1388,20 +1460,40 @@ InvalidatePossiblyObsoleteSlot(ReplicationSlot *s, XLogRecPtr oldestLSN,
}
/*
- * Mark any slot that points to an LSN older than the given segment
- * as invalid; it requires WAL that's about to be removed.
+ * Invalidate Obsolete slots or resolve recovery conflicts with logical slots.
*
- * Returns true when any slot have got invalidated.
+ * Obsolete case (aka xid is NULL):
*
- * NB - this runs as part of checkpoint, so avoid raising errors if possible.
+ * Mark any slot that points to an LSN older than the given segment
+ * as invalid; it requires WAL that's about to be removed.
+ * invalidated is set to true when any slot have got invalidated.
+ *
+ * Logical replication slot case:
+ *
+ * When xid is valid, it means that we are about to remove rows older than xid.
+ * Therefore we need to invalidate slots that depend on seeing those rows.
+ * When xid is invalid, invalidate all logical slots. This is required when the
+ * master wal_level is set back to replica, so existing logical slots need to
+ * be invalidated.
*/
-bool
-InvalidateObsoleteReplicationSlots(XLogSegNo oldestSegno)
+void
+InvalidateObsoleteReplicationSlots(XLogSegNo oldestSegno, bool *invalidated, Oid dboid, TransactionId *xid)
{
- XLogRecPtr oldestLSN;
- bool invalidated = false;
- XLogSegNoOffsetToRecPtr(oldestSegno, 0, wal_segment_size, oldestLSN);
+ XLogRecPtr oldestLSN = InvalidXLogRecPtr;
+ bool logical_slot_invalidated = false;
+
+ Assert(max_replication_slots >= 0);
+
+ if (max_replication_slots == 0)
+ return;
+
+ if (!xid)
+ {
+ Assert(invalidated);
+ *invalidated = false;
+ XLogSegNoOffsetToRecPtr(oldestSegno, 0, wal_segment_size, oldestLSN);
+ }
restart:
LWLockAcquire(ReplicationSlotControlLock, LW_SHARED);
@@ -1412,24 +1504,36 @@ restart:
if (!s->in_use)
continue;
- if (InvalidatePossiblyObsoleteSlot(s, oldestLSN, &invalidated))
+ if (xid)
{
- /* if the lock was released, start from scratch */
- goto restart;
+ /* we are only dealing with *logical* slot conflicts */
+ if (!SlotIsLogical(s))
+ continue;
+
+ /*
+ * not the database of interest and we don't want all the
+ * database, skip
+ */
+ if (s->data.database != dboid && TransactionIdIsValid(*xid))
+ continue;
}
+
+ if (InvalidatePossiblyObsoleteOrConflictingLogicalSlot(s, oldestLSN, invalidated ? invalidated : &logical_slot_invalidated, xid))
+ goto restart;
}
+
LWLockRelease(ReplicationSlotControlLock);
/*
- * If any slots have been invalidated, recalculate the resource limits.
+ * If any slots have been invalidated, recalculate the required xmin
+ * and the required lsn (if appropriate).
*/
- if (invalidated)
+ if ((!xid && *invalidated) || (xid && logical_slot_invalidated))
{
ReplicationSlotsComputeRequiredXmin(false);
- ReplicationSlotsComputeRequiredLSN();
+ if (!xid && *invalidated)
+ ReplicationSlotsComputeRequiredLSN();
}
-
- return invalidated;
}
/*
diff --git a/src/backend/replication/slotfuncs.c b/src/backend/replication/slotfuncs.c
index 2f3c964824..44192bc32d 100644
--- a/src/backend/replication/slotfuncs.c
+++ b/src/backend/replication/slotfuncs.c
@@ -232,7 +232,7 @@ pg_drop_replication_slot(PG_FUNCTION_ARGS)
Datum
pg_get_replication_slots(PG_FUNCTION_ARGS)
{
-#define PG_GET_REPLICATION_SLOTS_COLS 14
+#define PG_GET_REPLICATION_SLOTS_COLS 15
ReturnSetInfo *rsinfo = (ReturnSetInfo *) fcinfo->resultinfo;
XLogRecPtr currlsn;
int slotno;
@@ -404,6 +404,17 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
values[i++] = BoolGetDatum(slot_contents.data.two_phase);
+ if (slot_contents.data.database == InvalidOid)
+ nulls[i++] = true;
+ else
+ {
+ if (slot_contents.data.xmin == InvalidTransactionId &&
+ slot_contents.data.catalog_xmin == InvalidTransactionId)
+ values[i++] = BoolGetDatum(true);
+ else
+ values[i++] = BoolGetDatum(false);
+ }
+
Assert(i == PG_GET_REPLICATION_SLOTS_COLS);
tuplestore_putvalues(rsinfo->setResult, rsinfo->setDesc,
diff --git a/src/backend/replication/walsender.c b/src/backend/replication/walsender.c
index 4ed3747e3f..8885cdeebc 100644
--- a/src/backend/replication/walsender.c
+++ b/src/backend/replication/walsender.c
@@ -1253,6 +1253,14 @@ StartLogicalReplication(StartReplicationCmd *cmd)
ReplicationSlotAcquire(cmd->slotname, true);
+ if (!TransactionIdIsValid(MyReplicationSlot->data.xmin)
+ && !TransactionIdIsValid(MyReplicationSlot->data.catalog_xmin))
+ ereport(ERROR,
+ (errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
+ errmsg("cannot read from logical replication slot \"%s\"",
+ cmd->slotname),
+ errdetail("This slot has been invalidated because it was conflicting with recovery.")));
+
if (XLogRecPtrIsInvalid(MyReplicationSlot->data.restart_lsn))
ereport(ERROR,
(errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
diff --git a/src/backend/storage/ipc/procsignal.c b/src/backend/storage/ipc/procsignal.c
index 395b2cf690..c85cb5cc18 100644
--- a/src/backend/storage/ipc/procsignal.c
+++ b/src/backend/storage/ipc/procsignal.c
@@ -673,6 +673,9 @@ procsignal_sigusr1_handler(SIGNAL_ARGS)
if (CheckProcSignal(PROCSIG_RECOVERY_CONFLICT_SNAPSHOT))
RecoveryConflictInterrupt(PROCSIG_RECOVERY_CONFLICT_SNAPSHOT);
+ if (CheckProcSignal(PROCSIG_RECOVERY_CONFLICT_LOGICALSLOT))
+ RecoveryConflictInterrupt(PROCSIG_RECOVERY_CONFLICT_LOGICALSLOT);
+
if (CheckProcSignal(PROCSIG_RECOVERY_CONFLICT_STARTUP_DEADLOCK))
RecoveryConflictInterrupt(PROCSIG_RECOVERY_CONFLICT_STARTUP_DEADLOCK);
diff --git a/src/backend/storage/ipc/standby.c b/src/backend/storage/ipc/standby.c
index 94cc860f5f..ec817381a1 100644
--- a/src/backend/storage/ipc/standby.c
+++ b/src/backend/storage/ipc/standby.c
@@ -35,6 +35,7 @@
#include "utils/ps_status.h"
#include "utils/timeout.h"
#include "utils/timestamp.h"
+#include "replication/slot.h"
/* User-settable GUC parameters */
int vacuum_defer_cleanup_age;
@@ -475,6 +476,7 @@ ResolveRecoveryConflictWithVirtualXIDs(VirtualTransactionId *waitlist,
*/
void
ResolveRecoveryConflictWithSnapshot(TransactionId snapshotConflictHorizon,
+ bool isCatalogRel,
RelFileLocator locator)
{
VirtualTransactionId *backends;
@@ -500,6 +502,9 @@ ResolveRecoveryConflictWithSnapshot(TransactionId snapshotConflictHorizon,
PROCSIG_RECOVERY_CONFLICT_SNAPSHOT,
WAIT_EVENT_RECOVERY_CONFLICT_SNAPSHOT,
true);
+
+ if (wal_level >= WAL_LEVEL_LOGICAL && isCatalogRel)
+ InvalidateObsoleteReplicationSlots(InvalidXLogRecPtr, NULL, locator.dbOid, &snapshotConflictHorizon);
}
/*
@@ -508,6 +513,7 @@ ResolveRecoveryConflictWithSnapshot(TransactionId snapshotConflictHorizon,
*/
void
ResolveRecoveryConflictWithSnapshotFullXid(FullTransactionId snapshotConflictHorizon,
+ bool isCatalogRel,
RelFileLocator locator)
{
/*
@@ -526,7 +532,9 @@ ResolveRecoveryConflictWithSnapshotFullXid(FullTransactionId snapshotConflictHor
TransactionId truncated;
truncated = XidFromFullTransactionId(snapshotConflictHorizon);
- ResolveRecoveryConflictWithSnapshot(truncated, locator);
+ ResolveRecoveryConflictWithSnapshot(truncated,
+ isCatalogRel,
+ locator);
}
}
@@ -1487,6 +1495,9 @@ get_recovery_conflict_desc(ProcSignalReason reason)
case PROCSIG_RECOVERY_CONFLICT_SNAPSHOT:
reasonDesc = _("recovery conflict on snapshot");
break;
+ case PROCSIG_RECOVERY_CONFLICT_LOGICALSLOT:
+ reasonDesc = _("recovery conflict on replication slot");
+ break;
case PROCSIG_RECOVERY_CONFLICT_STARTUP_DEADLOCK:
reasonDesc = _("recovery conflict on buffer deadlock");
break;
diff --git a/src/backend/tcop/postgres.c b/src/backend/tcop/postgres.c
index 470b734e9e..0041896620 100644
--- a/src/backend/tcop/postgres.c
+++ b/src/backend/tcop/postgres.c
@@ -2481,6 +2481,9 @@ errdetail_recovery_conflict(void)
case PROCSIG_RECOVERY_CONFLICT_SNAPSHOT:
errdetail("User query might have needed to see row versions that must be removed.");
break;
+ case PROCSIG_RECOVERY_CONFLICT_LOGICALSLOT:
+ errdetail("User was using the logical slot that must be dropped.");
+ break;
case PROCSIG_RECOVERY_CONFLICT_STARTUP_DEADLOCK:
errdetail("User transaction caused buffer deadlock with recovery.");
break;
@@ -3050,6 +3053,27 @@ RecoveryConflictInterrupt(ProcSignalReason reason)
case PROCSIG_RECOVERY_CONFLICT_LOCK:
case PROCSIG_RECOVERY_CONFLICT_TABLESPACE:
case PROCSIG_RECOVERY_CONFLICT_SNAPSHOT:
+ case PROCSIG_RECOVERY_CONFLICT_LOGICALSLOT:
+
+ /*
+ * For conflicts that require a logical slot to be
+ * invalidated, the requirement is for the signal receiver to
+ * release the slot, so that it could be invalidated by the
+ * signal sender. So for normal backends, the transaction
+ * should be aborted, just like for other recovery conflicts.
+ * But if it's walsender on standby, we don't want to go
+ * through the following IsTransactionOrTransactionBlock()
+ * check, so break here.
+ */
+ if (am_cascading_walsender &&
+ reason == PROCSIG_RECOVERY_CONFLICT_LOGICALSLOT &&
+ MyReplicationSlot && SlotIsLogical(MyReplicationSlot))
+ {
+ RecoveryConflictPending = true;
+ QueryCancelPending = true;
+ InterruptPending = true;
+ break;
+ }
/*
* If we aren't in a transaction any longer then ignore.
diff --git a/src/backend/utils/activity/pgstat_database.c b/src/backend/utils/activity/pgstat_database.c
index 6e650ceaad..7149f22f72 100644
--- a/src/backend/utils/activity/pgstat_database.c
+++ b/src/backend/utils/activity/pgstat_database.c
@@ -109,6 +109,9 @@ pgstat_report_recovery_conflict(int reason)
case PROCSIG_RECOVERY_CONFLICT_BUFFERPIN:
dbentry->conflict_bufferpin++;
break;
+ case PROCSIG_RECOVERY_CONFLICT_LOGICALSLOT:
+ dbentry->conflict_logicalslot++;
+ break;
case PROCSIG_RECOVERY_CONFLICT_STARTUP_DEADLOCK:
dbentry->conflict_startup_deadlock++;
break;
@@ -387,6 +390,7 @@ pgstat_database_flush_cb(PgStat_EntryRef *entry_ref, bool nowait)
PGSTAT_ACCUM_DBCOUNT(conflict_tablespace);
PGSTAT_ACCUM_DBCOUNT(conflict_lock);
PGSTAT_ACCUM_DBCOUNT(conflict_snapshot);
+ PGSTAT_ACCUM_DBCOUNT(conflict_logicalslot);
PGSTAT_ACCUM_DBCOUNT(conflict_bufferpin);
PGSTAT_ACCUM_DBCOUNT(conflict_startup_deadlock);
diff --git a/src/backend/utils/adt/pgstatfuncs.c b/src/backend/utils/adt/pgstatfuncs.c
index 6737493402..afd62d3cc0 100644
--- a/src/backend/utils/adt/pgstatfuncs.c
+++ b/src/backend/utils/adt/pgstatfuncs.c
@@ -1066,6 +1066,8 @@ PG_STAT_GET_DBENTRY_INT64(xact_commit)
/* pg_stat_get_db_xact_rollback */
PG_STAT_GET_DBENTRY_INT64(xact_rollback)
+/* pg_stat_get_db_conflict_logicalslot */
+PG_STAT_GET_DBENTRY_INT64(conflict_logicalslot)
Datum
pg_stat_get_db_stat_reset_time(PG_FUNCTION_ARGS)
@@ -1099,6 +1101,7 @@ pg_stat_get_db_conflict_all(PG_FUNCTION_ARGS)
result = (int64) (dbentry->conflict_tablespace +
dbentry->conflict_lock +
dbentry->conflict_snapshot +
+ dbentry->conflict_logicalslot +
dbentry->conflict_bufferpin +
dbentry->conflict_startup_deadlock);
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index c0f2a8a77c..c8e11ab710 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -5577,6 +5577,11 @@
proname => 'pg_stat_get_db_conflict_snapshot', provolatile => 's',
proparallel => 'r', prorettype => 'int8', proargtypes => 'oid',
prosrc => 'pg_stat_get_db_conflict_snapshot' },
+{ oid => '9901',
+ descr => 'statistics: recovery conflicts in database caused by logical replication slot',
+ proname => 'pg_stat_get_db_conflict_logicalslot', provolatile => 's',
+ proparallel => 'r', prorettype => 'int8', proargtypes => 'oid',
+ prosrc => 'pg_stat_get_db_conflict_logicalslot' },
{ oid => '3068',
descr => 'statistics: recovery conflicts in database caused by shared buffer pin',
proname => 'pg_stat_get_db_conflict_bufferpin', provolatile => 's',
@@ -10946,9 +10951,9 @@
proname => 'pg_get_replication_slots', prorows => '10', proisstrict => 'f',
proretset => 't', provolatile => 's', prorettype => 'record',
proargtypes => '',
- proallargtypes => '{name,name,text,oid,bool,bool,int4,xid,xid,pg_lsn,pg_lsn,text,int8,bool}',
- proargmodes => '{o,o,o,o,o,o,o,o,o,o,o,o,o,o}',
- proargnames => '{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn,wal_status,safe_wal_size,two_phase}',
+ proallargtypes => '{name,name,text,oid,bool,bool,int4,xid,xid,pg_lsn,pg_lsn,text,int8,bool,bool}',
+ proargmodes => '{o,o,o,o,o,o,o,o,o,o,o,o,o,o,o}',
+ proargnames => '{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn,wal_status,safe_wal_size,two_phase,conflicting}',
prosrc => 'pg_get_replication_slots' },
{ oid => '3786', descr => 'set up a logical replication slot',
proname => 'pg_create_logical_replication_slot', provolatile => 'v',
diff --git a/src/include/pgstat.h b/src/include/pgstat.h
index 5e3326a3b9..872eb35757 100644
--- a/src/include/pgstat.h
+++ b/src/include/pgstat.h
@@ -291,6 +291,7 @@ typedef struct PgStat_StatDBEntry
PgStat_Counter conflict_tablespace;
PgStat_Counter conflict_lock;
PgStat_Counter conflict_snapshot;
+ PgStat_Counter conflict_logicalslot;
PgStat_Counter conflict_bufferpin;
PgStat_Counter conflict_startup_deadlock;
PgStat_Counter temp_files;
diff --git a/src/include/replication/slot.h b/src/include/replication/slot.h
index 8872c80cdf..236ebcdbdb 100644
--- a/src/include/replication/slot.h
+++ b/src/include/replication/slot.h
@@ -17,6 +17,8 @@
#include "storage/spin.h"
#include "replication/walreceiver.h"
+#define LogicalReplicationSlotIsInvalid(s) (!TransactionIdIsValid(s->data.xmin) && \
+ !TransactionIdIsValid(s->data.catalog_xmin))
/*
* Behaviour of replication slots, upon release or crash.
*
@@ -215,7 +217,7 @@ extern void ReplicationSlotsComputeRequiredLSN(void);
extern XLogRecPtr ReplicationSlotsComputeLogicalRestartLSN(void);
extern bool ReplicationSlotsCountDBSlots(Oid dboid, int *nslots, int *nactive);
extern void ReplicationSlotsDropDBSlots(Oid dboid);
-extern bool InvalidateObsoleteReplicationSlots(XLogSegNo oldestSegno);
+extern void InvalidateObsoleteReplicationSlots(XLogSegNo oldestSegno, bool *invalidated, Oid dboid, TransactionId *xid);
extern ReplicationSlot *SearchNamedReplicationSlot(const char *name, bool need_lock);
extern int ReplicationSlotIndex(ReplicationSlot *slot);
extern bool ReplicationSlotName(int index, Name name);
@@ -227,5 +229,6 @@ extern void CheckPointReplicationSlots(void);
extern void CheckSlotRequirements(void);
extern void CheckSlotPermissions(void);
+extern void ResolveRecoveryConflictWithLogicalSlots(Oid dboid, TransactionId xid, char *reason);
#endif /* SLOT_H */
diff --git a/src/include/storage/procsignal.h b/src/include/storage/procsignal.h
index 905af2231b..2f52100b00 100644
--- a/src/include/storage/procsignal.h
+++ b/src/include/storage/procsignal.h
@@ -42,6 +42,7 @@ typedef enum
PROCSIG_RECOVERY_CONFLICT_TABLESPACE,
PROCSIG_RECOVERY_CONFLICT_LOCK,
PROCSIG_RECOVERY_CONFLICT_SNAPSHOT,
+ PROCSIG_RECOVERY_CONFLICT_LOGICALSLOT,
PROCSIG_RECOVERY_CONFLICT_BUFFERPIN,
PROCSIG_RECOVERY_CONFLICT_STARTUP_DEADLOCK,
diff --git a/src/include/storage/standby.h b/src/include/storage/standby.h
index 2effdea126..41f4dc372e 100644
--- a/src/include/storage/standby.h
+++ b/src/include/storage/standby.h
@@ -30,8 +30,10 @@ extern void InitRecoveryTransactionEnvironment(void);
extern void ShutdownRecoveryTransactionEnvironment(void);
extern void ResolveRecoveryConflictWithSnapshot(TransactionId snapshotConflictHorizon,
+ bool isCatalogRel,
RelFileLocator locator);
extern void ResolveRecoveryConflictWithSnapshotFullXid(FullTransactionId snapshotConflictHorizon,
+ bool isCatalogRel,
RelFileLocator locator);
extern void ResolveRecoveryConflictWithTablespace(Oid tsid);
extern void ResolveRecoveryConflictWithDatabase(Oid dbid);
diff --git a/src/test/regress/expected/rules.out b/src/test/regress/expected/rules.out
index e7a2f5856a..11ea206337 100644
--- a/src/test/regress/expected/rules.out
+++ b/src/test/regress/expected/rules.out
@@ -1472,8 +1472,9 @@ pg_replication_slots| SELECT l.slot_name,
l.confirmed_flush_lsn,
l.wal_status,
l.safe_wal_size,
- l.two_phase
- FROM (pg_get_replication_slots() l(slot_name, plugin, slot_type, datoid, temporary, active, active_pid, xmin, catalog_xmin, restart_lsn, confirmed_flush_lsn, wal_status, safe_wal_size, two_phase)
+ l.two_phase,
+ l.conflicting
+ FROM (pg_get_replication_slots() l(slot_name, plugin, slot_type, datoid, temporary, active, active_pid, xmin, catalog_xmin, restart_lsn, confirmed_flush_lsn, wal_status, safe_wal_size, two_phase, conflicting)
LEFT JOIN pg_database d ON ((l.datoid = d.oid)));
pg_roles| SELECT pg_authid.rolname,
pg_authid.rolsuper,
@@ -1868,7 +1869,8 @@ pg_stat_database_conflicts| SELECT oid AS datid,
pg_stat_get_db_conflict_lock(oid) AS confl_lock,
pg_stat_get_db_conflict_snapshot(oid) AS confl_snapshot,
pg_stat_get_db_conflict_bufferpin(oid) AS confl_bufferpin,
- pg_stat_get_db_conflict_startup_deadlock(oid) AS confl_deadlock
+ pg_stat_get_db_conflict_startup_deadlock(oid) AS confl_deadlock,
+ pg_stat_get_db_conflict_logicalslot(oid) AS confl_active_logicalslot
FROM pg_database d;
pg_stat_gssapi| SELECT pid,
gss_auth AS gss_authenticated,
--
2.34.1
From 7ab0a6ec046f0c181c8f3b96cb18ae938f573ba9 Mon Sep 17 00:00:00 2001
From: bdrouvotAWS <[email protected]>
Date: Mon, 30 Jan 2023 15:30:15 +0000
Subject: [PATCH v45 1/6] Add info in WAL records in preparation for logical
slot conflict handling.
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Overall design:
1. We want to enable logical decoding on standbys, but replay of WAL
from the primary might remove data that is needed by logical decoding,
causing error(s) on the standby. To prevent those errors, a new replication
conflict scenario needs to be addressed (as much as hot standby does).
2. Our chosen strategy for dealing with this type of replication slot
is to invalidate logical slots for which needed data has been removed.
3. To do this we need the latestRemovedXid for each change, just as we
do for physical replication conflicts, but we also need to know
whether any particular change was to data that logical replication
might access. That way, during WAL replay, we know when there is a risk of
conflict and, if so, if there is a conflict.
4. We can't rely on the standby's relcache entries for this purpose in
any way, because the startup process can't access catalog contents.
5. Therefore every WAL record that potentially removes data from the
index or heap must carry a flag indicating whether or not it is one
that might be accessed during logical decoding.
Why do we need this for logical decoding on standby?
First, let's forget about logical decoding on standby and recall that
on a primary database, any catalog rows that may be needed by a logical
decoding replication slot are not removed.
This is done thanks to the catalog_xmin associated with the logical
replication slot.
But, with logical decoding on standby, in the following cases:
- hot_standby_feedback is off
- hot_standby_feedback is on but there is no a physical slot between
the primary and the standby. Then, hot_standby_feedback will work,
but only while the connection is alive (for example a node restart
would break it)
Then, the primary may delete system catalog rows that could be needed
by the logical decoding on the standby (as it does not know about the
catalog_xmin on the standby).
So, it’s mandatory to identify those rows and invalidate the slots
that may need them if any. Identifying those rows is the purpose of
this commit.
Implementation:
When a WAL replay on standby indicates that a catalog table tuple is
to be deleted by an xid that is greater than a logical slot's
catalog_xmin, then that means the slot's catalog_xmin conflicts with
the xid, and we need to handle the conflict. While subsequent commits
will do the actual conflict handling, this commit adds a new field
isCatalogRel in such WAL records (and a new bit set in the
xl_heap_visible flags field), that is true for catalog tables, so as to
arrange for conflict handling.
The affected WAL records are the ones that already contain the
snapshotConflictHorizon field, namely:
- gistxlogDelete
- gistxlogPageReuse
- xl_hash_vacuum_one_page
- xl_heap_prune
- xl_heap_freeze_page
- xl_heap_visible
- xl_btree_reuse_page
- xl_btree_delete
- spgxlogVacuumRedirect
Due to this new field being added, xl_hash_vacuum_one_page and
gistxlogDelete do now contain the offsets to be deleted as a
FLEXIBLE_ARRAY_MEMBER. This is needed to ensure correct alignement.
It's not needed on the others struct where isCatalogRel has
been added.
Author: Andres Freund (in an older version), Amit Khandekar, Bertrand
Drouvot
Reviewed-By: Bertrand Drouvot, Andres Freund, Robert Haas, Fabrizio de
Royes Mello, Melanie Plageman
---
contrib/amcheck/verify_nbtree.c | 15 +--
src/backend/access/gist/gist.c | 5 +-
src/backend/access/gist/gistbuild.c | 2 +-
src/backend/access/gist/gistutil.c | 4 +-
src/backend/access/gist/gistxlog.c | 17 ++--
src/backend/access/hash/hash_xlog.c | 12 +--
src/backend/access/hash/hashinsert.c | 1 +
src/backend/access/heap/heapam.c | 5 +-
src/backend/access/heap/heapam_handler.c | 9 +-
src/backend/access/heap/pruneheap.c | 1 +
src/backend/access/heap/vacuumlazy.c | 2 +
src/backend/access/heap/visibilitymap.c | 3 +-
src/backend/access/nbtree/nbtinsert.c | 91 +++++++++--------
src/backend/access/nbtree/nbtpage.c | 111 +++++++++++----------
src/backend/access/nbtree/nbtree.c | 4 +-
src/backend/access/nbtree/nbtsearch.c | 50 ++++++----
src/backend/access/nbtree/nbtsort.c | 2 +-
src/backend/access/nbtree/nbtutils.c | 7 +-
src/backend/access/spgist/spgvacuum.c | 9 +-
src/backend/catalog/index.c | 1 +
src/backend/commands/analyze.c | 1 +
src/backend/commands/vacuumparallel.c | 6 ++
src/backend/optimizer/util/plancat.c | 2 +-
src/backend/utils/sort/tuplesortvariants.c | 5 +-
src/include/access/genam.h | 1 +
src/include/access/gist_private.h | 7 +-
src/include/access/gistxlog.h | 13 ++-
src/include/access/hash_xlog.h | 8 +-
src/include/access/heapam_xlog.h | 10 +-
src/include/access/nbtree.h | 37 ++++---
src/include/access/nbtxlog.h | 8 +-
src/include/access/spgxlog.h | 2 +
src/include/access/visibilitymapdefs.h | 10 +-
src/include/utils/rel.h | 1 +
src/include/utils/tuplesort.h | 4 +-
35 files changed, 263 insertions(+), 203 deletions(-)
3.3% contrib/amcheck/
4.7% src/backend/access/gist/
4.1% src/backend/access/heap/
59.0% src/backend/access/nbtree/
3.7% src/backend/access/
22.0% src/include/access/
diff --git a/contrib/amcheck/verify_nbtree.c b/contrib/amcheck/verify_nbtree.c
index 257cff671b..eb280d4893 100644
--- a/contrib/amcheck/verify_nbtree.c
+++ b/contrib/amcheck/verify_nbtree.c
@@ -183,6 +183,7 @@ static inline bool invariant_l_nontarget_offset(BtreeCheckState *state,
OffsetNumber upperbound);
static Page palloc_btree_page(BtreeCheckState *state, BlockNumber blocknum);
static inline BTScanInsert bt_mkscankey_pivotsearch(Relation rel,
+ Relation heaprel,
IndexTuple itup);
static ItemId PageGetItemIdCareful(BtreeCheckState *state, BlockNumber block,
Page page, OffsetNumber offset);
@@ -331,7 +332,7 @@ bt_index_check_internal(Oid indrelid, bool parentcheck, bool heapallindexed,
RelationGetRelationName(indrel))));
/* Extract metadata from metapage, and sanitize it in passing */
- _bt_metaversion(indrel, &heapkeyspace, &allequalimage);
+ _bt_metaversion(indrel, heaprel, &heapkeyspace, &allequalimage);
if (allequalimage && !heapkeyspace)
ereport(ERROR,
(errcode(ERRCODE_INDEX_CORRUPTED),
@@ -1258,7 +1259,7 @@ bt_target_page_check(BtreeCheckState *state)
}
/* Build insertion scankey for current page offset */
- skey = bt_mkscankey_pivotsearch(state->rel, itup);
+ skey = bt_mkscankey_pivotsearch(state->rel, state->heaprel, itup);
/*
* Make sure tuple size does not exceed the relevant BTREE_VERSION
@@ -1768,7 +1769,7 @@ bt_right_page_check_scankey(BtreeCheckState *state)
* memory remaining allocated.
*/
firstitup = (IndexTuple) PageGetItem(rightpage, rightitem);
- return bt_mkscankey_pivotsearch(state->rel, firstitup);
+ return bt_mkscankey_pivotsearch(state->rel, state->heaprel, firstitup);
}
/*
@@ -2681,7 +2682,7 @@ bt_rootdescend(BtreeCheckState *state, IndexTuple itup)
Buffer lbuf;
bool exists;
- key = _bt_mkscankey(state->rel, itup);
+ key = _bt_mkscankey(state->rel, state->heaprel, itup);
Assert(key->heapkeyspace && key->scantid != NULL);
/*
@@ -2694,7 +2695,7 @@ bt_rootdescend(BtreeCheckState *state, IndexTuple itup)
*/
Assert(state->readonly && state->rootdescend);
exists = false;
- stack = _bt_search(state->rel, key, &lbuf, BT_READ, NULL);
+ stack = _bt_search(state->rel, state->heaprel, key, &lbuf, BT_READ, NULL);
if (BufferIsValid(lbuf))
{
@@ -3133,11 +3134,11 @@ palloc_btree_page(BtreeCheckState *state, BlockNumber blocknum)
* the scankey is greater.
*/
static inline BTScanInsert
-bt_mkscankey_pivotsearch(Relation rel, IndexTuple itup)
+bt_mkscankey_pivotsearch(Relation rel, Relation heaprel, IndexTuple itup)
{
BTScanInsert skey;
- skey = _bt_mkscankey(rel, itup);
+ skey = _bt_mkscankey(rel, heaprel, itup);
skey->pivotsearch = true;
return skey;
diff --git a/src/backend/access/gist/gist.c b/src/backend/access/gist/gist.c
index ba394f08f6..3ac68ec3b4 100644
--- a/src/backend/access/gist/gist.c
+++ b/src/backend/access/gist/gist.c
@@ -348,7 +348,7 @@ gistplacetopage(Relation rel, Size freespace, GISTSTATE *giststate,
for (; ptr; ptr = ptr->next)
{
/* Allocate new page */
- ptr->buffer = gistNewBuffer(rel);
+ ptr->buffer = gistNewBuffer(rel, heapRel);
GISTInitBuffer(ptr->buffer, (is_leaf) ? F_LEAF : 0);
ptr->page = BufferGetPage(ptr->buffer);
ptr->block.blkno = BufferGetBlockNumber(ptr->buffer);
@@ -1694,7 +1694,8 @@ gistprunepage(Relation rel, Page page, Buffer buffer, Relation heapRel)
recptr = gistXLogDelete(buffer,
deletable, ndeletable,
- snapshotConflictHorizon);
+ snapshotConflictHorizon,
+ heapRel);
PageSetLSN(page, recptr);
}
diff --git a/src/backend/access/gist/gistbuild.c b/src/backend/access/gist/gistbuild.c
index d21a308d41..4462022904 100644
--- a/src/backend/access/gist/gistbuild.c
+++ b/src/backend/access/gist/gistbuild.c
@@ -298,7 +298,7 @@ gistbuild(Relation heap, Relation index, IndexInfo *indexInfo)
Page page;
/* initialize the root page */
- buffer = gistNewBuffer(index);
+ buffer = gistNewBuffer(index, heap);
Assert(BufferGetBlockNumber(buffer) == GIST_ROOT_BLKNO);
page = BufferGetPage(buffer);
diff --git a/src/backend/access/gist/gistutil.c b/src/backend/access/gist/gistutil.c
index 56451fede1..aad14a401d 100644
--- a/src/backend/access/gist/gistutil.c
+++ b/src/backend/access/gist/gistutil.c
@@ -821,7 +821,7 @@ gistcheckpage(Relation rel, Buffer buf)
* Caller is responsible for initializing the page by calling GISTInitBuffer
*/
Buffer
-gistNewBuffer(Relation r)
+gistNewBuffer(Relation r, Relation heaprel)
{
Buffer buffer;
bool needLock;
@@ -865,7 +865,7 @@ gistNewBuffer(Relation r)
* page's deleteXid.
*/
if (XLogStandbyInfoActive() && RelationNeedsWAL(r))
- gistXLogPageReuse(r, blkno, GistPageGetDeleteXid(page));
+ gistXLogPageReuse(r, heaprel, blkno, GistPageGetDeleteXid(page));
return buffer;
}
diff --git a/src/backend/access/gist/gistxlog.c b/src/backend/access/gist/gistxlog.c
index f65864254a..b7678f3c14 100644
--- a/src/backend/access/gist/gistxlog.c
+++ b/src/backend/access/gist/gistxlog.c
@@ -177,6 +177,7 @@ gistRedoDeleteRecord(XLogReaderState *record)
gistxlogDelete *xldata = (gistxlogDelete *) XLogRecGetData(record);
Buffer buffer;
Page page;
+ OffsetNumber *toDelete = xldata->offsets;
/*
* If we have any conflict processing to do, it must happen before we
@@ -203,14 +204,7 @@ gistRedoDeleteRecord(XLogReaderState *record)
{
page = (Page) BufferGetPage(buffer);
- if (XLogRecGetDataLen(record) > SizeOfGistxlogDelete)
- {
- OffsetNumber *todelete;
-
- todelete = (OffsetNumber *) ((char *) xldata + SizeOfGistxlogDelete);
-
- PageIndexMultiDelete(page, todelete, xldata->ntodelete);
- }
+ PageIndexMultiDelete(page, toDelete, xldata->ntodelete);
GistClearPageHasGarbage(page);
GistMarkTuplesDeleted(page);
@@ -597,7 +591,8 @@ gistXLogAssignLSN(void)
* Write XLOG record about reuse of a deleted page.
*/
void
-gistXLogPageReuse(Relation rel, BlockNumber blkno, FullTransactionId deleteXid)
+gistXLogPageReuse(Relation rel, Relation heaprel,
+ BlockNumber blkno, FullTransactionId deleteXid)
{
gistxlogPageReuse xlrec_reuse;
@@ -608,6 +603,7 @@ gistXLogPageReuse(Relation rel, BlockNumber blkno, FullTransactionId deleteXid)
*/
/* XLOG stuff */
+ xlrec_reuse.isCatalogRel = RelationIsAccessibleInLogicalDecoding(heaprel);
xlrec_reuse.locator = rel->rd_locator;
xlrec_reuse.block = blkno;
xlrec_reuse.snapshotConflictHorizon = deleteXid;
@@ -672,11 +668,12 @@ gistXLogUpdate(Buffer buffer,
*/
XLogRecPtr
gistXLogDelete(Buffer buffer, OffsetNumber *todelete, int ntodelete,
- TransactionId snapshotConflictHorizon)
+ TransactionId snapshotConflictHorizon, Relation heaprel)
{
gistxlogDelete xlrec;
XLogRecPtr recptr;
+ xlrec.isCatalogRel = RelationIsAccessibleInLogicalDecoding(heaprel);
xlrec.snapshotConflictHorizon = snapshotConflictHorizon;
xlrec.ntodelete = ntodelete;
diff --git a/src/backend/access/hash/hash_xlog.c b/src/backend/access/hash/hash_xlog.c
index f38b42efb9..08ceb91288 100644
--- a/src/backend/access/hash/hash_xlog.c
+++ b/src/backend/access/hash/hash_xlog.c
@@ -980,8 +980,10 @@ hash_xlog_vacuum_one_page(XLogReaderState *record)
Page page;
XLogRedoAction action;
HashPageOpaque pageopaque;
+ OffsetNumber *toDelete;
xldata = (xl_hash_vacuum_one_page *) XLogRecGetData(record);
+ toDelete = xldata->offsets;
/*
* If we have any conflict processing to do, it must happen before we
@@ -1010,15 +1012,7 @@ hash_xlog_vacuum_one_page(XLogReaderState *record)
{
page = (Page) BufferGetPage(buffer);
- if (XLogRecGetDataLen(record) > SizeOfHashVacuumOnePage)
- {
- OffsetNumber *unused;
-
- unused = (OffsetNumber *) ((char *) xldata + SizeOfHashVacuumOnePage);
-
- PageIndexMultiDelete(page, unused, xldata->ntuples);
- }
-
+ PageIndexMultiDelete(page, toDelete, xldata->ntuples);
/*
* Mark the page as not containing any LP_DEAD items. See comments in
* _hash_vacuum_one_page() for details.
diff --git a/src/backend/access/hash/hashinsert.c b/src/backend/access/hash/hashinsert.c
index a604e31891..22656b24e2 100644
--- a/src/backend/access/hash/hashinsert.c
+++ b/src/backend/access/hash/hashinsert.c
@@ -432,6 +432,7 @@ _hash_vacuum_one_page(Relation rel, Relation hrel, Buffer metabuf, Buffer buf)
xl_hash_vacuum_one_page xlrec;
XLogRecPtr recptr;
+ xlrec.isCatalogRel = RelationIsAccessibleInLogicalDecoding(hrel);
xlrec.snapshotConflictHorizon = snapshotConflictHorizon;
xlrec.ntuples = ndeletable;
diff --git a/src/backend/access/heap/heapam.c b/src/backend/access/heap/heapam.c
index e6024a980b..d478724b9d 100644
--- a/src/backend/access/heap/heapam.c
+++ b/src/backend/access/heap/heapam.c
@@ -6872,6 +6872,7 @@ heap_freeze_execute_prepared(Relation rel, Buffer buffer,
nplans = heap_log_freeze_plan(tuples, ntuples, plans, offsets);
xlrec.snapshotConflictHorizon = snapshotConflictHorizon;
+ xlrec.isCatalogRel = RelationIsAccessibleInLogicalDecoding(rel);
xlrec.nplans = nplans;
XLogBeginInsert();
@@ -8442,7 +8443,7 @@ bottomup_sort_and_shrink(TM_IndexDeleteOp *delstate)
* update the heap page's LSN.
*/
XLogRecPtr
-log_heap_visible(RelFileLocator rlocator, Buffer heap_buffer, Buffer vm_buffer,
+log_heap_visible(Relation rel, Buffer heap_buffer, Buffer vm_buffer,
TransactionId snapshotConflictHorizon, uint8 vmflags)
{
xl_heap_visible xlrec;
@@ -8454,6 +8455,8 @@ log_heap_visible(RelFileLocator rlocator, Buffer heap_buffer, Buffer vm_buffer,
xlrec.snapshotConflictHorizon = snapshotConflictHorizon;
xlrec.flags = vmflags;
+ if (RelationIsAccessibleInLogicalDecoding(rel))
+ xlrec.flags |= VISIBILITYMAP_IS_CATALOG_REL;
XLogBeginInsert();
XLogRegisterData((char *) &xlrec, SizeOfHeapVisible);
diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index c4b1916d36..392c6e659c 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -720,9 +720,14 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
*multi_cutoff);
- /* Set up sorting if wanted */
+ /*
+ * Set up sorting if wanted. NewHeap is being passed to
+ * tuplesort_begin_cluster(), it could have been OldHeap too. It does not
+ * really matter, as the goal is to have a heap relation being passed to
+ * _bt_log_reuse_page() (which should not be called from this code path).
+ */
if (use_sort)
- tuplesort = tuplesort_begin_cluster(oldTupDesc, OldIndex,
+ tuplesort = tuplesort_begin_cluster(oldTupDesc, OldIndex, NewHeap,
maintenance_work_mem,
NULL, TUPLESORT_NONE);
else
diff --git a/src/backend/access/heap/pruneheap.c b/src/backend/access/heap/pruneheap.c
index 4e65cbcadf..3f0342351f 100644
--- a/src/backend/access/heap/pruneheap.c
+++ b/src/backend/access/heap/pruneheap.c
@@ -418,6 +418,7 @@ heap_page_prune(Relation relation, Buffer buffer,
xl_heap_prune xlrec;
XLogRecPtr recptr;
+ xlrec.isCatalogRel = RelationIsAccessibleInLogicalDecoding(relation);
xlrec.snapshotConflictHorizon = prstate.snapshotConflictHorizon;
xlrec.nredirected = prstate.nredirected;
xlrec.ndead = prstate.ndead;
diff --git a/src/backend/access/heap/vacuumlazy.c b/src/backend/access/heap/vacuumlazy.c
index 8f14cf85f3..ae628d747d 100644
--- a/src/backend/access/heap/vacuumlazy.c
+++ b/src/backend/access/heap/vacuumlazy.c
@@ -2710,6 +2710,7 @@ lazy_vacuum_one_index(Relation indrel, IndexBulkDeleteResult *istat,
ivinfo.message_level = DEBUG2;
ivinfo.num_heap_tuples = reltuples;
ivinfo.strategy = vacrel->bstrategy;
+ ivinfo.heaprel = vacrel->rel;
/*
* Update error traceback information.
@@ -2759,6 +2760,7 @@ lazy_cleanup_one_index(Relation indrel, IndexBulkDeleteResult *istat,
ivinfo.num_heap_tuples = reltuples;
ivinfo.strategy = vacrel->bstrategy;
+ ivinfo.heaprel = vacrel->rel;
/*
* Update error traceback information.
diff --git a/src/backend/access/heap/visibilitymap.c b/src/backend/access/heap/visibilitymap.c
index 74ff01bb17..d1ba859851 100644
--- a/src/backend/access/heap/visibilitymap.c
+++ b/src/backend/access/heap/visibilitymap.c
@@ -288,8 +288,7 @@ visibilitymap_set(Relation rel, BlockNumber heapBlk, Buffer heapBuf,
if (XLogRecPtrIsInvalid(recptr))
{
Assert(!InRecovery);
- recptr = log_heap_visible(rel->rd_locator, heapBuf, vmBuf,
- cutoff_xid, flags);
+ recptr = log_heap_visible(rel, heapBuf, vmBuf, cutoff_xid, flags);
/*
* If data checksums are enabled (or wal_log_hints=on), we
diff --git a/src/backend/access/nbtree/nbtinsert.c b/src/backend/access/nbtree/nbtinsert.c
index f4c1a974ef..8c6e867c61 100644
--- a/src/backend/access/nbtree/nbtinsert.c
+++ b/src/backend/access/nbtree/nbtinsert.c
@@ -30,7 +30,8 @@
#define BTREE_FASTPATH_MIN_LEVEL 2
-static BTStack _bt_search_insert(Relation rel, BTInsertState insertstate);
+static BTStack _bt_search_insert(Relation rel, Relation heaprel,
+ BTInsertState insertstate);
static TransactionId _bt_check_unique(Relation rel, BTInsertState insertstate,
Relation heapRel,
IndexUniqueCheck checkUnique, bool *is_unique,
@@ -41,8 +42,9 @@ static OffsetNumber _bt_findinsertloc(Relation rel,
bool indexUnchanged,
BTStack stack,
Relation heapRel);
-static void _bt_stepright(Relation rel, BTInsertState insertstate, BTStack stack);
-static void _bt_insertonpg(Relation rel, BTScanInsert itup_key,
+static void _bt_stepright(Relation rel, Relation heaprel,
+ BTInsertState insertstate, BTStack stack);
+static void _bt_insertonpg(Relation rel, Relation heaprel, BTScanInsert itup_key,
Buffer buf,
Buffer cbuf,
BTStack stack,
@@ -51,13 +53,13 @@ static void _bt_insertonpg(Relation rel, BTScanInsert itup_key,
OffsetNumber newitemoff,
int postingoff,
bool split_only_page);
-static Buffer _bt_split(Relation rel, BTScanInsert itup_key, Buffer buf,
- Buffer cbuf, OffsetNumber newitemoff, Size newitemsz,
- IndexTuple newitem, IndexTuple orignewitem,
+static Buffer _bt_split(Relation rel, Relation heaprel, BTScanInsert itup_key,
+ Buffer buf, Buffer cbuf, OffsetNumber newitemoff,
+ Size newitemsz, IndexTuple newitem, IndexTuple orignewitem,
IndexTuple nposting, uint16 postingoff);
-static void _bt_insert_parent(Relation rel, Buffer buf, Buffer rbuf,
- BTStack stack, bool isroot, bool isonly);
-static Buffer _bt_newroot(Relation rel, Buffer lbuf, Buffer rbuf);
+static void _bt_insert_parent(Relation rel, Relation heaprel, Buffer buf,
+ Buffer rbuf, BTStack stack, bool isroot, bool isonly);
+static Buffer _bt_newroot(Relation rel, Relation heaprel, Buffer lbuf, Buffer rbuf);
static inline bool _bt_pgaddtup(Page page, Size itemsize, IndexTuple itup,
OffsetNumber itup_off, bool newfirstdataitem);
static void _bt_delete_or_dedup_one_page(Relation rel, Relation heapRel,
@@ -108,7 +110,7 @@ _bt_doinsert(Relation rel, IndexTuple itup,
bool checkingunique = (checkUnique != UNIQUE_CHECK_NO);
/* we need an insertion scan key to do our search, so build one */
- itup_key = _bt_mkscankey(rel, itup);
+ itup_key = _bt_mkscankey(rel, heapRel, itup);
if (checkingunique)
{
@@ -162,7 +164,7 @@ search:
* searching from the root page. insertstate.buf will hold a buffer that
* is locked in exclusive mode afterwards.
*/
- stack = _bt_search_insert(rel, &insertstate);
+ stack = _bt_search_insert(rel, heapRel, &insertstate);
/*
* checkingunique inserts are not allowed to go ahead when two tuples with
@@ -255,8 +257,8 @@ search:
*/
newitemoff = _bt_findinsertloc(rel, &insertstate, checkingunique,
indexUnchanged, stack, heapRel);
- _bt_insertonpg(rel, itup_key, insertstate.buf, InvalidBuffer, stack,
- itup, insertstate.itemsz, newitemoff,
+ _bt_insertonpg(rel, heapRel, itup_key, insertstate.buf, InvalidBuffer,
+ stack, itup, insertstate.itemsz, newitemoff,
insertstate.postingoff, false);
}
else
@@ -312,7 +314,7 @@ search:
* since each per-backend cache won't stay valid for long.
*/
static BTStack
-_bt_search_insert(Relation rel, BTInsertState insertstate)
+_bt_search_insert(Relation rel, Relation heaprel, BTInsertState insertstate)
{
Assert(insertstate->buf == InvalidBuffer);
Assert(!insertstate->bounds_valid);
@@ -375,8 +377,8 @@ _bt_search_insert(Relation rel, BTInsertState insertstate)
}
/* Cannot use optimization -- descend tree, return proper descent stack */
- return _bt_search(rel, insertstate->itup_key, &insertstate->buf, BT_WRITE,
- NULL);
+ return _bt_search(rel, heaprel, insertstate->itup_key, &insertstate->buf,
+ BT_WRITE, NULL);
}
/*
@@ -885,7 +887,7 @@ _bt_findinsertloc(Relation rel,
_bt_compare(rel, itup_key, page, P_HIKEY) <= 0)
break;
- _bt_stepright(rel, insertstate, stack);
+ _bt_stepright(rel, heapRel, insertstate, stack);
/* Update local state after stepping right */
page = BufferGetPage(insertstate->buf);
opaque = BTPageGetOpaque(page);
@@ -969,7 +971,7 @@ _bt_findinsertloc(Relation rel,
pg_prng_uint32(&pg_global_prng_state) <= (PG_UINT32_MAX / 100))
break;
- _bt_stepright(rel, insertstate, stack);
+ _bt_stepright(rel, heapRel, insertstate, stack);
/* Update local state after stepping right */
page = BufferGetPage(insertstate->buf);
opaque = BTPageGetOpaque(page);
@@ -1022,7 +1024,7 @@ _bt_findinsertloc(Relation rel,
* indexes.
*/
static void
-_bt_stepright(Relation rel, BTInsertState insertstate, BTStack stack)
+_bt_stepright(Relation rel, Relation heaprel, BTInsertState insertstate, BTStack stack)
{
Page page;
BTPageOpaque opaque;
@@ -1048,7 +1050,7 @@ _bt_stepright(Relation rel, BTInsertState insertstate, BTStack stack)
*/
if (P_INCOMPLETE_SPLIT(opaque))
{
- _bt_finish_split(rel, rbuf, stack);
+ _bt_finish_split(rel, heaprel, rbuf, stack);
rbuf = InvalidBuffer;
continue;
}
@@ -1099,6 +1101,7 @@ _bt_stepright(Relation rel, BTInsertState insertstate, BTStack stack)
*/
static void
_bt_insertonpg(Relation rel,
+ Relation heaprel,
BTScanInsert itup_key,
Buffer buf,
Buffer cbuf,
@@ -1209,8 +1212,8 @@ _bt_insertonpg(Relation rel,
Assert(!split_only_page);
/* split the buffer into left and right halves */
- rbuf = _bt_split(rel, itup_key, buf, cbuf, newitemoff, itemsz, itup,
- origitup, nposting, postingoff);
+ rbuf = _bt_split(rel, heaprel, itup_key, buf, cbuf, newitemoff, itemsz,
+ itup, origitup, nposting, postingoff);
PredicateLockPageSplit(rel,
BufferGetBlockNumber(buf),
BufferGetBlockNumber(rbuf));
@@ -1233,7 +1236,7 @@ _bt_insertonpg(Relation rel,
* page.
*----------
*/
- _bt_insert_parent(rel, buf, rbuf, stack, isroot, isonly);
+ _bt_insert_parent(rel, heaprel, buf, rbuf, stack, isroot, isonly);
}
else
{
@@ -1254,7 +1257,7 @@ _bt_insertonpg(Relation rel,
Assert(!isleaf);
Assert(BufferIsValid(cbuf));
- metabuf = _bt_getbuf(rel, BTREE_METAPAGE, BT_WRITE);
+ metabuf = _bt_getbuf(rel, heaprel, BTREE_METAPAGE, BT_WRITE);
metapg = BufferGetPage(metabuf);
metad = BTPageGetMeta(metapg);
@@ -1418,7 +1421,7 @@ _bt_insertonpg(Relation rel,
* call _bt_getrootheight while holding a buffer lock.
*/
if (BlockNumberIsValid(blockcache) &&
- _bt_getrootheight(rel) >= BTREE_FASTPATH_MIN_LEVEL)
+ _bt_getrootheight(rel, heaprel) >= BTREE_FASTPATH_MIN_LEVEL)
RelationSetTargetBlock(rel, blockcache);
}
@@ -1459,8 +1462,8 @@ _bt_insertonpg(Relation rel,
* The pin and lock on buf are maintained.
*/
static Buffer
-_bt_split(Relation rel, BTScanInsert itup_key, Buffer buf, Buffer cbuf,
- OffsetNumber newitemoff, Size newitemsz, IndexTuple newitem,
+_bt_split(Relation rel, Relation heaprel, BTScanInsert itup_key, Buffer buf,
+ Buffer cbuf, OffsetNumber newitemoff, Size newitemsz, IndexTuple newitem,
IndexTuple orignewitem, IndexTuple nposting, uint16 postingoff)
{
Buffer rbuf;
@@ -1712,7 +1715,7 @@ _bt_split(Relation rel, BTScanInsert itup_key, Buffer buf, Buffer cbuf,
* way because it avoids an unnecessary PANIC when either origpage or its
* existing sibling page are corrupt.
*/
- rbuf = _bt_getbuf(rel, P_NEW, BT_WRITE);
+ rbuf = _bt_getbuf(rel, heaprel, P_NEW, BT_WRITE);
rightpage = BufferGetPage(rbuf);
rightpagenumber = BufferGetBlockNumber(rbuf);
/* rightpage was initialized by _bt_getbuf */
@@ -1885,7 +1888,7 @@ _bt_split(Relation rel, BTScanInsert itup_key, Buffer buf, Buffer cbuf,
*/
if (!isrightmost)
{
- sbuf = _bt_getbuf(rel, oopaque->btpo_next, BT_WRITE);
+ sbuf = _bt_getbuf(rel, heaprel, oopaque->btpo_next, BT_WRITE);
spage = BufferGetPage(sbuf);
sopaque = BTPageGetOpaque(spage);
if (sopaque->btpo_prev != origpagenumber)
@@ -2092,6 +2095,7 @@ _bt_split(Relation rel, BTScanInsert itup_key, Buffer buf, Buffer cbuf,
*/
static void
_bt_insert_parent(Relation rel,
+ Relation heaprel,
Buffer buf,
Buffer rbuf,
BTStack stack,
@@ -2118,7 +2122,7 @@ _bt_insert_parent(Relation rel,
Assert(stack == NULL);
Assert(isonly);
/* create a new root node and update the metapage */
- rootbuf = _bt_newroot(rel, buf, rbuf);
+ rootbuf = _bt_newroot(rel, heaprel, buf, rbuf);
/* release the split buffers */
_bt_relbuf(rel, rootbuf);
_bt_relbuf(rel, rbuf);
@@ -2157,7 +2161,8 @@ _bt_insert_parent(Relation rel,
BlockNumberIsValid(RelationGetTargetBlock(rel))));
/* Find the leftmost page at the next level up */
- pbuf = _bt_get_endpoint(rel, opaque->btpo_level + 1, false, NULL);
+ pbuf = _bt_get_endpoint(rel, heaprel, opaque->btpo_level + 1, false,
+ NULL);
/* Set up a phony stack entry pointing there */
stack = &fakestack;
stack->bts_blkno = BufferGetBlockNumber(pbuf);
@@ -2183,7 +2188,7 @@ _bt_insert_parent(Relation rel,
* new downlink will be inserted at the correct offset. Even buf's
* parent may have changed.
*/
- pbuf = _bt_getstackbuf(rel, stack, bknum);
+ pbuf = _bt_getstackbuf(rel, heaprel, stack, bknum);
/*
* Unlock the right child. The left child will be unlocked in
@@ -2207,7 +2212,7 @@ _bt_insert_parent(Relation rel,
RelationGetRelationName(rel), bknum, rbknum)));
/* Recursively insert into the parent */
- _bt_insertonpg(rel, NULL, pbuf, buf, stack->bts_parent,
+ _bt_insertonpg(rel, heaprel, NULL, pbuf, buf, stack->bts_parent,
new_item, MAXALIGN(IndexTupleSize(new_item)),
stack->bts_offset + 1, 0, isonly);
@@ -2227,7 +2232,7 @@ _bt_insert_parent(Relation rel,
* and unpinned.
*/
void
-_bt_finish_split(Relation rel, Buffer lbuf, BTStack stack)
+_bt_finish_split(Relation rel, Relation heaprel, Buffer lbuf, BTStack stack)
{
Page lpage = BufferGetPage(lbuf);
BTPageOpaque lpageop = BTPageGetOpaque(lpage);
@@ -2240,7 +2245,7 @@ _bt_finish_split(Relation rel, Buffer lbuf, BTStack stack)
Assert(P_INCOMPLETE_SPLIT(lpageop));
/* Lock right sibling, the one missing the downlink */
- rbuf = _bt_getbuf(rel, lpageop->btpo_next, BT_WRITE);
+ rbuf = _bt_getbuf(rel, heaprel, lpageop->btpo_next, BT_WRITE);
rpage = BufferGetPage(rbuf);
rpageop = BTPageGetOpaque(rpage);
@@ -2252,7 +2257,7 @@ _bt_finish_split(Relation rel, Buffer lbuf, BTStack stack)
BTMetaPageData *metad;
/* acquire lock on the metapage */
- metabuf = _bt_getbuf(rel, BTREE_METAPAGE, BT_WRITE);
+ metabuf = _bt_getbuf(rel, heaprel, BTREE_METAPAGE, BT_WRITE);
metapg = BufferGetPage(metabuf);
metad = BTPageGetMeta(metapg);
@@ -2269,7 +2274,7 @@ _bt_finish_split(Relation rel, Buffer lbuf, BTStack stack)
elog(DEBUG1, "finishing incomplete split of %u/%u",
BufferGetBlockNumber(lbuf), BufferGetBlockNumber(rbuf));
- _bt_insert_parent(rel, lbuf, rbuf, stack, wasroot, wasonly);
+ _bt_insert_parent(rel, heaprel, lbuf, rbuf, stack, wasroot, wasonly);
}
/*
@@ -2304,7 +2309,7 @@ _bt_finish_split(Relation rel, Buffer lbuf, BTStack stack)
* offset number bts_offset + 1.
*/
Buffer
-_bt_getstackbuf(Relation rel, BTStack stack, BlockNumber child)
+_bt_getstackbuf(Relation rel, Relation heaprel, BTStack stack, BlockNumber child)
{
BlockNumber blkno;
OffsetNumber start;
@@ -2318,13 +2323,13 @@ _bt_getstackbuf(Relation rel, BTStack stack, BlockNumber child)
Page page;
BTPageOpaque opaque;
- buf = _bt_getbuf(rel, blkno, BT_WRITE);
+ buf = _bt_getbuf(rel, heaprel, blkno, BT_WRITE);
page = BufferGetPage(buf);
opaque = BTPageGetOpaque(page);
if (P_INCOMPLETE_SPLIT(opaque))
{
- _bt_finish_split(rel, buf, stack->bts_parent);
+ _bt_finish_split(rel, heaprel, buf, stack->bts_parent);
continue;
}
@@ -2428,7 +2433,7 @@ _bt_getstackbuf(Relation rel, BTStack stack, BlockNumber child)
* lbuf, rbuf & rootbuf.
*/
static Buffer
-_bt_newroot(Relation rel, Buffer lbuf, Buffer rbuf)
+_bt_newroot(Relation rel, Relation heaprel, Buffer lbuf, Buffer rbuf)
{
Buffer rootbuf;
Page lpage,
@@ -2454,12 +2459,12 @@ _bt_newroot(Relation rel, Buffer lbuf, Buffer rbuf)
lopaque = BTPageGetOpaque(lpage);
/* get a new root page */
- rootbuf = _bt_getbuf(rel, P_NEW, BT_WRITE);
+ rootbuf = _bt_getbuf(rel, heaprel, P_NEW, BT_WRITE);
rootpage = BufferGetPage(rootbuf);
rootblknum = BufferGetBlockNumber(rootbuf);
/* acquire lock on the metapage */
- metabuf = _bt_getbuf(rel, BTREE_METAPAGE, BT_WRITE);
+ metabuf = _bt_getbuf(rel, heaprel, BTREE_METAPAGE, BT_WRITE);
metapg = BufferGetPage(metabuf);
metad = BTPageGetMeta(metapg);
diff --git a/src/backend/access/nbtree/nbtpage.c b/src/backend/access/nbtree/nbtpage.c
index 3feee28d19..151ad37a54 100644
--- a/src/backend/access/nbtree/nbtpage.c
+++ b/src/backend/access/nbtree/nbtpage.c
@@ -38,25 +38,24 @@
#include "utils/snapmgr.h"
static BTMetaPageData *_bt_getmeta(Relation rel, Buffer metabuf);
-static void _bt_log_reuse_page(Relation rel, BlockNumber blkno,
+static void _bt_log_reuse_page(Relation rel, Relation heaprel, BlockNumber blkno,
FullTransactionId safexid);
-static void _bt_delitems_delete(Relation rel, Buffer buf,
+static void _bt_delitems_delete(Relation rel, Relation heaprel, Buffer buf,
TransactionId snapshotConflictHorizon,
OffsetNumber *deletable, int ndeletable,
BTVacuumPosting *updatable, int nupdatable);
static char *_bt_delitems_update(BTVacuumPosting *updatable, int nupdatable,
OffsetNumber *updatedoffsets,
Size *updatedbuflen, bool needswal);
-static bool _bt_mark_page_halfdead(Relation rel, Buffer leafbuf,
- BTStack stack);
+static bool _bt_mark_page_halfdead(Relation rel, Relation heaprel,
+ Buffer leafbuf, BTStack stack);
static bool _bt_unlink_halfdead_page(Relation rel, Buffer leafbuf,
BlockNumber scanblkno,
bool *rightsib_empty,
BTVacState *vstate);
-static bool _bt_lock_subtree_parent(Relation rel, BlockNumber child,
- BTStack stack,
- Buffer *subtreeparent,
- OffsetNumber *poffset,
+static bool _bt_lock_subtree_parent(Relation rel, Relation heaprel,
+ BlockNumber child, BTStack stack,
+ Buffer *subtreeparent, OffsetNumber *poffset,
BlockNumber *topparent,
BlockNumber *topparentrightsib);
static void _bt_pendingfsm_add(BTVacState *vstate, BlockNumber target,
@@ -178,7 +177,7 @@ _bt_getmeta(Relation rel, Buffer metabuf)
* index tuples needed to be deleted.
*/
bool
-_bt_vacuum_needs_cleanup(Relation rel)
+_bt_vacuum_needs_cleanup(Relation rel, Relation heaprel)
{
Buffer metabuf;
Page metapg;
@@ -191,7 +190,7 @@ _bt_vacuum_needs_cleanup(Relation rel)
*
* Note that we deliberately avoid using cached version of metapage here.
*/
- metabuf = _bt_getbuf(rel, BTREE_METAPAGE, BT_READ);
+ metabuf = _bt_getbuf(rel, heaprel, BTREE_METAPAGE, BT_READ);
metapg = BufferGetPage(metabuf);
metad = BTPageGetMeta(metapg);
btm_version = metad->btm_version;
@@ -231,7 +230,7 @@ _bt_vacuum_needs_cleanup(Relation rel)
* finalized.
*/
void
-_bt_set_cleanup_info(Relation rel, BlockNumber num_delpages)
+_bt_set_cleanup_info(Relation rel, Relation heaprel, BlockNumber num_delpages)
{
Buffer metabuf;
Page metapg;
@@ -255,7 +254,7 @@ _bt_set_cleanup_info(Relation rel, BlockNumber num_delpages)
* no longer used as of PostgreSQL 14. We set it to -1.0 on rewrite, just
* to be consistent.
*/
- metabuf = _bt_getbuf(rel, BTREE_METAPAGE, BT_READ);
+ metabuf = _bt_getbuf(rel, heaprel, BTREE_METAPAGE, BT_READ);
metapg = BufferGetPage(metabuf);
metad = BTPageGetMeta(metapg);
@@ -340,7 +339,7 @@ _bt_set_cleanup_info(Relation rel, BlockNumber num_delpages)
* The metadata page is not locked or pinned on exit.
*/
Buffer
-_bt_getroot(Relation rel, int access)
+_bt_getroot(Relation rel, Relation heaprel, int access)
{
Buffer metabuf;
Buffer rootbuf;
@@ -370,7 +369,7 @@ _bt_getroot(Relation rel, int access)
Assert(rootblkno != P_NONE);
rootlevel = metad->btm_fastlevel;
- rootbuf = _bt_getbuf(rel, rootblkno, BT_READ);
+ rootbuf = _bt_getbuf(rel, heaprel, rootblkno, BT_READ);
rootpage = BufferGetPage(rootbuf);
rootopaque = BTPageGetOpaque(rootpage);
@@ -396,7 +395,7 @@ _bt_getroot(Relation rel, int access)
rel->rd_amcache = NULL;
}
- metabuf = _bt_getbuf(rel, BTREE_METAPAGE, BT_READ);
+ metabuf = _bt_getbuf(rel, heaprel, BTREE_METAPAGE, BT_READ);
metad = _bt_getmeta(rel, metabuf);
/* if no root page initialized yet, do it */
@@ -429,7 +428,7 @@ _bt_getroot(Relation rel, int access)
* to optimize this case.)
*/
_bt_relbuf(rel, metabuf);
- return _bt_getroot(rel, access);
+ return _bt_getroot(rel, heaprel, access);
}
/*
@@ -437,7 +436,7 @@ _bt_getroot(Relation rel, int access)
* the new root page. Since this is the first page in the tree, it's
* a leaf as well as the root.
*/
- rootbuf = _bt_getbuf(rel, P_NEW, BT_WRITE);
+ rootbuf = _bt_getbuf(rel, heaprel, P_NEW, BT_WRITE);
rootblkno = BufferGetBlockNumber(rootbuf);
rootpage = BufferGetPage(rootbuf);
rootopaque = BTPageGetOpaque(rootpage);
@@ -574,7 +573,7 @@ _bt_getroot(Relation rel, int access)
* moving to the root --- that'd deadlock against any concurrent root split.)
*/
Buffer
-_bt_gettrueroot(Relation rel)
+_bt_gettrueroot(Relation rel, Relation heaprel)
{
Buffer metabuf;
Page metapg;
@@ -596,7 +595,7 @@ _bt_gettrueroot(Relation rel)
pfree(rel->rd_amcache);
rel->rd_amcache = NULL;
- metabuf = _bt_getbuf(rel, BTREE_METAPAGE, BT_READ);
+ metabuf = _bt_getbuf(rel, heaprel, BTREE_METAPAGE, BT_READ);
metapg = BufferGetPage(metabuf);
metaopaque = BTPageGetOpaque(metapg);
metad = BTPageGetMeta(metapg);
@@ -669,7 +668,7 @@ _bt_gettrueroot(Relation rel)
* about updating previously cached data.
*/
int
-_bt_getrootheight(Relation rel)
+_bt_getrootheight(Relation rel, Relation heaprel)
{
BTMetaPageData *metad;
@@ -677,7 +676,7 @@ _bt_getrootheight(Relation rel)
{
Buffer metabuf;
- metabuf = _bt_getbuf(rel, BTREE_METAPAGE, BT_READ);
+ metabuf = _bt_getbuf(rel, heaprel, BTREE_METAPAGE, BT_READ);
metad = _bt_getmeta(rel, metabuf);
/*
@@ -733,7 +732,7 @@ _bt_getrootheight(Relation rel)
* pg_upgrade'd from Postgres 12.
*/
void
-_bt_metaversion(Relation rel, bool *heapkeyspace, bool *allequalimage)
+_bt_metaversion(Relation rel, Relation heaprel, bool *heapkeyspace, bool *allequalimage)
{
BTMetaPageData *metad;
@@ -741,7 +740,7 @@ _bt_metaversion(Relation rel, bool *heapkeyspace, bool *allequalimage)
{
Buffer metabuf;
- metabuf = _bt_getbuf(rel, BTREE_METAPAGE, BT_READ);
+ metabuf = _bt_getbuf(rel, heaprel, BTREE_METAPAGE, BT_READ);
metad = _bt_getmeta(rel, metabuf);
/*
@@ -825,7 +824,8 @@ _bt_checkpage(Relation rel, Buffer buf)
* Log the reuse of a page from the FSM.
*/
static void
-_bt_log_reuse_page(Relation rel, BlockNumber blkno, FullTransactionId safexid)
+_bt_log_reuse_page(Relation rel, Relation heaprel, BlockNumber blkno,
+ FullTransactionId safexid)
{
xl_btree_reuse_page xlrec_reuse;
@@ -836,6 +836,7 @@ _bt_log_reuse_page(Relation rel, BlockNumber blkno, FullTransactionId safexid)
*/
/* XLOG stuff */
+ xlrec_reuse.isCatalogRel = RelationIsAccessibleInLogicalDecoding(heaprel);
xlrec_reuse.locator = rel->rd_locator;
xlrec_reuse.block = blkno;
xlrec_reuse.snapshotConflictHorizon = safexid;
@@ -868,7 +869,7 @@ _bt_log_reuse_page(Relation rel, BlockNumber blkno, FullTransactionId safexid)
* as _bt_lockbuf().
*/
Buffer
-_bt_getbuf(Relation rel, BlockNumber blkno, int access)
+_bt_getbuf(Relation rel, Relation heaprel, BlockNumber blkno, int access)
{
Buffer buf;
@@ -943,7 +944,7 @@ _bt_getbuf(Relation rel, BlockNumber blkno, int access)
* than safexid value
*/
if (XLogStandbyInfoActive() && RelationNeedsWAL(rel))
- _bt_log_reuse_page(rel, blkno,
+ _bt_log_reuse_page(rel, heaprel, blkno,
BTPageGetDeleteXid(page));
/* Okay to use page. Re-initialize and return it. */
@@ -1293,7 +1294,7 @@ _bt_delitems_vacuum(Relation rel, Buffer buf,
* clear page's VACUUM cycle ID.
*/
static void
-_bt_delitems_delete(Relation rel, Buffer buf,
+_bt_delitems_delete(Relation rel, Relation heaprel, Buffer buf,
TransactionId snapshotConflictHorizon,
OffsetNumber *deletable, int ndeletable,
BTVacuumPosting *updatable, int nupdatable)
@@ -1358,6 +1359,7 @@ _bt_delitems_delete(Relation rel, Buffer buf,
XLogRecPtr recptr;
xl_btree_delete xlrec_delete;
+ xlrec_delete.isCatalogRel = RelationIsAccessibleInLogicalDecoding(heaprel);
xlrec_delete.snapshotConflictHorizon = snapshotConflictHorizon;
xlrec_delete.ndeleted = ndeletable;
xlrec_delete.nupdated = nupdatable;
@@ -1684,8 +1686,8 @@ _bt_delitems_delete_check(Relation rel, Buffer buf, Relation heapRel,
}
/* Physically delete tuples (or TIDs) using deletable (or updatable) */
- _bt_delitems_delete(rel, buf, snapshotConflictHorizon,
- deletable, ndeletable, updatable, nupdatable);
+ _bt_delitems_delete(rel, heapRel, buf, snapshotConflictHorizon, deletable,
+ ndeletable, updatable, nupdatable);
/* be tidy */
for (int i = 0; i < nupdatable; i++)
@@ -1706,7 +1708,8 @@ _bt_delitems_delete_check(Relation rel, Buffer buf, Relation heapRel,
* same level must always be locked left to right to avoid deadlocks.
*/
static bool
-_bt_leftsib_splitflag(Relation rel, BlockNumber leftsib, BlockNumber target)
+_bt_leftsib_splitflag(Relation rel, Relation heaprel, BlockNumber leftsib,
+ BlockNumber target)
{
Buffer buf;
Page page;
@@ -1717,7 +1720,7 @@ _bt_leftsib_splitflag(Relation rel, BlockNumber leftsib, BlockNumber target)
if (leftsib == P_NONE)
return false;
- buf = _bt_getbuf(rel, leftsib, BT_READ);
+ buf = _bt_getbuf(rel, heaprel, leftsib, BT_READ);
page = BufferGetPage(buf);
opaque = BTPageGetOpaque(page);
@@ -1763,7 +1766,7 @@ _bt_leftsib_splitflag(Relation rel, BlockNumber leftsib, BlockNumber target)
* to-be-deleted subtree.)
*/
static bool
-_bt_rightsib_halfdeadflag(Relation rel, BlockNumber leafrightsib)
+_bt_rightsib_halfdeadflag(Relation rel, Relation heaprel, BlockNumber leafrightsib)
{
Buffer buf;
Page page;
@@ -1772,7 +1775,7 @@ _bt_rightsib_halfdeadflag(Relation rel, BlockNumber leafrightsib)
Assert(leafrightsib != P_NONE);
- buf = _bt_getbuf(rel, leafrightsib, BT_READ);
+ buf = _bt_getbuf(rel, heaprel, leafrightsib, BT_READ);
page = BufferGetPage(buf);
opaque = BTPageGetOpaque(page);
@@ -1961,17 +1964,18 @@ _bt_pagedel(Relation rel, Buffer leafbuf, BTVacState *vstate)
* marked with INCOMPLETE_SPLIT flag before proceeding
*/
Assert(leafblkno == scanblkno);
- if (_bt_leftsib_splitflag(rel, leftsib, leafblkno))
+ if (_bt_leftsib_splitflag(rel, vstate->info->heaprel, leftsib, leafblkno))
{
ReleaseBuffer(leafbuf);
return;
}
/* we need an insertion scan key for the search, so build one */
- itup_key = _bt_mkscankey(rel, targetkey);
+ itup_key = _bt_mkscankey(rel, vstate->info->heaprel, targetkey);
/* find the leftmost leaf page with matching pivot/high key */
itup_key->pivotsearch = true;
- stack = _bt_search(rel, itup_key, &sleafbuf, BT_READ, NULL);
+ stack = _bt_search(rel, vstate->info->heaprel, itup_key,
+ &sleafbuf, BT_READ, NULL);
/* won't need a second lock or pin on leafbuf */
_bt_relbuf(rel, sleafbuf);
@@ -2002,7 +2006,7 @@ _bt_pagedel(Relation rel, Buffer leafbuf, BTVacState *vstate)
* leafbuf page half-dead.
*/
Assert(P_ISLEAF(opaque) && !P_IGNORE(opaque));
- if (!_bt_mark_page_halfdead(rel, leafbuf, stack))
+ if (!_bt_mark_page_halfdead(rel, vstate->info->heaprel, leafbuf, stack))
{
_bt_relbuf(rel, leafbuf);
return;
@@ -2065,7 +2069,7 @@ _bt_pagedel(Relation rel, Buffer leafbuf, BTVacState *vstate)
if (!rightsib_empty)
break;
- leafbuf = _bt_getbuf(rel, rightsib, BT_WRITE);
+ leafbuf = _bt_getbuf(rel, vstate->info->heaprel, rightsib, BT_WRITE);
}
}
@@ -2084,7 +2088,8 @@ _bt_pagedel(Relation rel, Buffer leafbuf, BTVacState *vstate)
* successfully.
*/
static bool
-_bt_mark_page_halfdead(Relation rel, Buffer leafbuf, BTStack stack)
+_bt_mark_page_halfdead(Relation rel, Relation heaprel, Buffer leafbuf,
+ BTStack stack)
{
BlockNumber leafblkno;
BlockNumber leafrightsib;
@@ -2119,7 +2124,7 @@ _bt_mark_page_halfdead(Relation rel, Buffer leafbuf, BTStack stack)
* delete the downlink. It would fail the "right sibling of target page
* is also the next child in parent page" cross-check below.
*/
- if (_bt_rightsib_halfdeadflag(rel, leafrightsib))
+ if (_bt_rightsib_halfdeadflag(rel, heaprel, leafrightsib))
{
elog(DEBUG1, "could not delete page %u because its right sibling %u is half-dead",
leafblkno, leafrightsib);
@@ -2143,7 +2148,7 @@ _bt_mark_page_halfdead(Relation rel, Buffer leafbuf, BTStack stack)
*/
topparent = leafblkno;
topparentrightsib = leafrightsib;
- if (!_bt_lock_subtree_parent(rel, leafblkno, stack,
+ if (!_bt_lock_subtree_parent(rel, heaprel, leafblkno, stack,
&subtreeparent, &poffset,
&topparent, &topparentrightsib))
return false;
@@ -2363,7 +2368,7 @@ _bt_unlink_halfdead_page(Relation rel, Buffer leafbuf, BlockNumber scanblkno,
Assert(target != leafblkno);
/* Fetch the block number of the target's left sibling */
- buf = _bt_getbuf(rel, target, BT_READ);
+ buf = _bt_getbuf(rel, vstate->info->heaprel, target, BT_READ);
page = BufferGetPage(buf);
opaque = BTPageGetOpaque(page);
leftsib = opaque->btpo_prev;
@@ -2390,7 +2395,7 @@ _bt_unlink_halfdead_page(Relation rel, Buffer leafbuf, BlockNumber scanblkno,
_bt_lockbuf(rel, leafbuf, BT_WRITE);
if (leftsib != P_NONE)
{
- lbuf = _bt_getbuf(rel, leftsib, BT_WRITE);
+ lbuf = _bt_getbuf(rel, vstate->info->heaprel, leftsib, BT_WRITE);
page = BufferGetPage(lbuf);
opaque = BTPageGetOpaque(page);
while (P_ISDELETED(opaque) || opaque->btpo_next != target)
@@ -2440,7 +2445,7 @@ _bt_unlink_halfdead_page(Relation rel, Buffer leafbuf, BlockNumber scanblkno,
CHECK_FOR_INTERRUPTS();
/* step right one page */
- lbuf = _bt_getbuf(rel, leftsib, BT_WRITE);
+ lbuf = _bt_getbuf(rel, vstate->info->heaprel, leftsib, BT_WRITE);
page = BufferGetPage(lbuf);
opaque = BTPageGetOpaque(page);
}
@@ -2504,7 +2509,7 @@ _bt_unlink_halfdead_page(Relation rel, Buffer leafbuf, BlockNumber scanblkno,
* And next write-lock the (current) right sibling.
*/
rightsib = opaque->btpo_next;
- rbuf = _bt_getbuf(rel, rightsib, BT_WRITE);
+ rbuf = _bt_getbuf(rel, vstate->info->heaprel, rightsib, BT_WRITE);
page = BufferGetPage(rbuf);
opaque = BTPageGetOpaque(page);
if (opaque->btpo_prev != target)
@@ -2533,7 +2538,8 @@ _bt_unlink_halfdead_page(Relation rel, Buffer leafbuf, BlockNumber scanblkno,
if (P_RIGHTMOST(opaque))
{
/* rightsib will be the only one left on the level */
- metabuf = _bt_getbuf(rel, BTREE_METAPAGE, BT_WRITE);
+ metabuf = _bt_getbuf(rel, vstate->info->heaprel, BTREE_METAPAGE,
+ BT_WRITE);
metapg = BufferGetPage(metabuf);
metad = BTPageGetMeta(metapg);
@@ -2773,9 +2779,10 @@ _bt_unlink_halfdead_page(Relation rel, Buffer leafbuf, BlockNumber scanblkno,
* parent block in the leafbuf page using BTreeTupleSetTopParent()).
*/
static bool
-_bt_lock_subtree_parent(Relation rel, BlockNumber child, BTStack stack,
- Buffer *subtreeparent, OffsetNumber *poffset,
- BlockNumber *topparent, BlockNumber *topparentrightsib)
+_bt_lock_subtree_parent(Relation rel, Relation heaprel, BlockNumber child,
+ BTStack stack, Buffer *subtreeparent,
+ OffsetNumber *poffset, BlockNumber *topparent,
+ BlockNumber *topparentrightsib)
{
BlockNumber parent,
leftsibparent;
@@ -2789,7 +2796,7 @@ _bt_lock_subtree_parent(Relation rel, BlockNumber child, BTStack stack,
* Locate the pivot tuple whose downlink points to "child". Write lock
* the parent page itself.
*/
- pbuf = _bt_getstackbuf(rel, stack, child);
+ pbuf = _bt_getstackbuf(rel, heaprel, stack, child);
if (pbuf == InvalidBuffer)
{
/*
@@ -2889,11 +2896,11 @@ _bt_lock_subtree_parent(Relation rel, BlockNumber child, BTStack stack,
*
* Note: We deliberately avoid completing incomplete splits here.
*/
- if (_bt_leftsib_splitflag(rel, leftsibparent, parent))
+ if (_bt_leftsib_splitflag(rel, heaprel, leftsibparent, parent))
return false;
/* Recurse to examine child page's grandparent page */
- return _bt_lock_subtree_parent(rel, parent, stack->bts_parent,
+ return _bt_lock_subtree_parent(rel, heaprel, parent, stack->bts_parent,
subtreeparent, poffset,
topparent, topparentrightsib);
}
diff --git a/src/backend/access/nbtree/nbtree.c b/src/backend/access/nbtree/nbtree.c
index 1cc88da032..4e8a85fb5d 100644
--- a/src/backend/access/nbtree/nbtree.c
+++ b/src/backend/access/nbtree/nbtree.c
@@ -834,7 +834,7 @@ btvacuumcleanup(IndexVacuumInfo *info, IndexBulkDeleteResult *stats)
if (stats == NULL)
{
/* Check if VACUUM operation can entirely avoid btvacuumscan() call */
- if (!_bt_vacuum_needs_cleanup(info->index))
+ if (!_bt_vacuum_needs_cleanup(info->index, info->heaprel))
return NULL;
/*
@@ -870,7 +870,7 @@ btvacuumcleanup(IndexVacuumInfo *info, IndexBulkDeleteResult *stats)
*/
Assert(stats->pages_deleted >= stats->pages_free);
num_delpages = stats->pages_deleted - stats->pages_free;
- _bt_set_cleanup_info(info->index, num_delpages);
+ _bt_set_cleanup_info(info->index, info->heaprel, num_delpages);
/*
* It's quite possible for us to be fooled by concurrent page splits into
diff --git a/src/backend/access/nbtree/nbtsearch.c b/src/backend/access/nbtree/nbtsearch.c
index c43c1a2830..5c728e353d 100644
--- a/src/backend/access/nbtree/nbtsearch.c
+++ b/src/backend/access/nbtree/nbtsearch.c
@@ -42,7 +42,8 @@ static bool _bt_steppage(IndexScanDesc scan, ScanDirection dir);
static bool _bt_readnextpage(IndexScanDesc scan, BlockNumber blkno, ScanDirection dir);
static bool _bt_parallel_readpage(IndexScanDesc scan, BlockNumber blkno,
ScanDirection dir);
-static Buffer _bt_walk_left(Relation rel, Buffer buf, Snapshot snapshot);
+static Buffer _bt_walk_left(Relation rel, Relation heaprel, Buffer buf,
+ Snapshot snapshot);
static bool _bt_endpoint(IndexScanDesc scan, ScanDirection dir);
static inline void _bt_initialize_more_data(BTScanOpaque so, ScanDirection dir);
@@ -93,14 +94,14 @@ _bt_drop_lock_and_maybe_pin(IndexScanDesc scan, BTScanPos sp)
* during the search will be finished.
*/
BTStack
-_bt_search(Relation rel, BTScanInsert key, Buffer *bufP, int access,
- Snapshot snapshot)
+_bt_search(Relation rel, Relation heaprel, BTScanInsert key, Buffer *bufP,
+ int access, Snapshot snapshot)
{
BTStack stack_in = NULL;
int page_access = BT_READ;
/* Get the root page to start with */
- *bufP = _bt_getroot(rel, access);
+ *bufP = _bt_getroot(rel, heaprel, access);
/* If index is empty and access = BT_READ, no root page is created. */
if (!BufferIsValid(*bufP))
@@ -129,8 +130,8 @@ _bt_search(Relation rel, BTScanInsert key, Buffer *bufP, int access,
* also taken care of in _bt_getstackbuf). But this is a good
* opportunity to finish splits of internal pages too.
*/
- *bufP = _bt_moveright(rel, key, *bufP, (access == BT_WRITE), stack_in,
- page_access, snapshot);
+ *bufP = _bt_moveright(rel, heaprel, key, *bufP, (access == BT_WRITE),
+ stack_in, page_access, snapshot);
/* if this is a leaf page, we're done */
page = BufferGetPage(*bufP);
@@ -190,7 +191,7 @@ _bt_search(Relation rel, BTScanInsert key, Buffer *bufP, int access,
* but before we acquired a write lock. If it has, we may need to
* move right to its new sibling. Do that.
*/
- *bufP = _bt_moveright(rel, key, *bufP, true, stack_in, BT_WRITE,
+ *bufP = _bt_moveright(rel, heaprel, key, *bufP, true, stack_in, BT_WRITE,
snapshot);
}
@@ -234,6 +235,7 @@ _bt_search(Relation rel, BTScanInsert key, Buffer *bufP, int access,
*/
Buffer
_bt_moveright(Relation rel,
+ Relation heaprel,
BTScanInsert key,
Buffer buf,
bool forupdate,
@@ -288,12 +290,12 @@ _bt_moveright(Relation rel,
}
if (P_INCOMPLETE_SPLIT(opaque))
- _bt_finish_split(rel, buf, stack);
+ _bt_finish_split(rel, heaprel, buf, stack);
else
_bt_relbuf(rel, buf);
/* re-acquire the lock in the right mode, and re-check */
- buf = _bt_getbuf(rel, blkno, access);
+ buf = _bt_getbuf(rel, heaprel, blkno, access);
continue;
}
@@ -860,6 +862,7 @@ bool
_bt_first(IndexScanDesc scan, ScanDirection dir)
{
Relation rel = scan->indexRelation;
+ Relation heaprel = scan->heapRelation;
BTScanOpaque so = (BTScanOpaque) scan->opaque;
Buffer buf;
BTStack stack;
@@ -1352,7 +1355,7 @@ _bt_first(IndexScanDesc scan, ScanDirection dir)
}
/* Initialize remaining insertion scan key fields */
- _bt_metaversion(rel, &inskey.heapkeyspace, &inskey.allequalimage);
+ _bt_metaversion(rel, heaprel, &inskey.heapkeyspace, &inskey.allequalimage);
inskey.anynullkeys = false; /* unused */
inskey.nextkey = nextkey;
inskey.pivotsearch = false;
@@ -1363,7 +1366,7 @@ _bt_first(IndexScanDesc scan, ScanDirection dir)
* Use the manufactured insertion scan key to descend the tree and
* position ourselves on the target leaf page.
*/
- stack = _bt_search(rel, &inskey, &buf, BT_READ, scan->xs_snapshot);
+ stack = _bt_search(rel, heaprel, &inskey, &buf, BT_READ, scan->xs_snapshot);
/* don't need to keep the stack around... */
_bt_freestack(stack);
@@ -2004,7 +2007,7 @@ _bt_readnextpage(IndexScanDesc scan, BlockNumber blkno, ScanDirection dir)
/* check for interrupts while we're not holding any buffer lock */
CHECK_FOR_INTERRUPTS();
/* step right one page */
- so->currPos.buf = _bt_getbuf(rel, blkno, BT_READ);
+ so->currPos.buf = _bt_getbuf(rel, scan->heapRelation, blkno, BT_READ);
page = BufferGetPage(so->currPos.buf);
TestForOldSnapshot(scan->xs_snapshot, rel, page);
opaque = BTPageGetOpaque(page);
@@ -2078,7 +2081,8 @@ _bt_readnextpage(IndexScanDesc scan, BlockNumber blkno, ScanDirection dir)
if (BTScanPosIsPinned(so->currPos))
_bt_lockbuf(rel, so->currPos.buf, BT_READ);
else
- so->currPos.buf = _bt_getbuf(rel, so->currPos.currPage, BT_READ);
+ so->currPos.buf = _bt_getbuf(rel, scan->heapRelation,
+ so->currPos.currPage, BT_READ);
for (;;)
{
@@ -2092,8 +2096,8 @@ _bt_readnextpage(IndexScanDesc scan, BlockNumber blkno, ScanDirection dir)
}
/* Step to next physical page */
- so->currPos.buf = _bt_walk_left(rel, so->currPos.buf,
- scan->xs_snapshot);
+ so->currPos.buf = _bt_walk_left(rel, scan->heapRelation,
+ so->currPos.buf, scan->xs_snapshot);
/* if we're physically at end of index, return failure */
if (so->currPos.buf == InvalidBuffer)
@@ -2140,7 +2144,8 @@ _bt_readnextpage(IndexScanDesc scan, BlockNumber blkno, ScanDirection dir)
BTScanPosInvalidate(so->currPos);
return false;
}
- so->currPos.buf = _bt_getbuf(rel, blkno, BT_READ);
+ so->currPos.buf = _bt_getbuf(rel, scan->heapRelation, blkno,
+ BT_READ);
}
}
}
@@ -2185,7 +2190,7 @@ _bt_parallel_readpage(IndexScanDesc scan, BlockNumber blkno, ScanDirection dir)
* again if it's important.
*/
static Buffer
-_bt_walk_left(Relation rel, Buffer buf, Snapshot snapshot)
+_bt_walk_left(Relation rel, Relation heaprel, Buffer buf, Snapshot snapshot)
{
Page page;
BTPageOpaque opaque;
@@ -2213,7 +2218,7 @@ _bt_walk_left(Relation rel, Buffer buf, Snapshot snapshot)
_bt_relbuf(rel, buf);
/* check for interrupts while we're not holding any buffer lock */
CHECK_FOR_INTERRUPTS();
- buf = _bt_getbuf(rel, blkno, BT_READ);
+ buf = _bt_getbuf(rel, heaprel, blkno, BT_READ);
page = BufferGetPage(buf);
TestForOldSnapshot(snapshot, rel, page);
opaque = BTPageGetOpaque(page);
@@ -2304,7 +2309,7 @@ _bt_walk_left(Relation rel, Buffer buf, Snapshot snapshot)
* The returned buffer is pinned and read-locked.
*/
Buffer
-_bt_get_endpoint(Relation rel, uint32 level, bool rightmost,
+_bt_get_endpoint(Relation rel, Relation heaprel, uint32 level, bool rightmost,
Snapshot snapshot)
{
Buffer buf;
@@ -2320,9 +2325,9 @@ _bt_get_endpoint(Relation rel, uint32 level, bool rightmost,
* smarter about intermediate levels.)
*/
if (level == 0)
- buf = _bt_getroot(rel, BT_READ);
+ buf = _bt_getroot(rel, heaprel, BT_READ);
else
- buf = _bt_gettrueroot(rel);
+ buf = _bt_gettrueroot(rel, heaprel);
if (!BufferIsValid(buf))
return InvalidBuffer;
@@ -2403,7 +2408,8 @@ _bt_endpoint(IndexScanDesc scan, ScanDirection dir)
* version of _bt_search(). We don't maintain a stack since we know we
* won't need it.
*/
- buf = _bt_get_endpoint(rel, 0, ScanDirectionIsBackward(dir), scan->xs_snapshot);
+ buf = _bt_get_endpoint(rel, scan->heapRelation, 0,
+ ScanDirectionIsBackward(dir), scan->xs_snapshot);
if (!BufferIsValid(buf))
{
diff --git a/src/backend/access/nbtree/nbtsort.c b/src/backend/access/nbtree/nbtsort.c
index 67b7b1710c..8c58fdb8d1 100644
--- a/src/backend/access/nbtree/nbtsort.c
+++ b/src/backend/access/nbtree/nbtsort.c
@@ -566,7 +566,7 @@ _bt_leafbuild(BTSpool *btspool, BTSpool *btspool2)
wstate.heap = btspool->heap;
wstate.index = btspool->index;
- wstate.inskey = _bt_mkscankey(wstate.index, NULL);
+ wstate.inskey = _bt_mkscankey(wstate.index, btspool->heap, NULL);
/* _bt_mkscankey() won't set allequalimage without metapage */
wstate.inskey->allequalimage = _bt_allequalimage(wstate.index, true);
wstate.btws_use_wal = RelationNeedsWAL(wstate.index);
diff --git a/src/backend/access/nbtree/nbtutils.c b/src/backend/access/nbtree/nbtutils.c
index 8003583c0a..70a0c2418a 100644
--- a/src/backend/access/nbtree/nbtutils.c
+++ b/src/backend/access/nbtree/nbtutils.c
@@ -87,7 +87,7 @@ static int _bt_keep_natts(Relation rel, IndexTuple lastleft,
* field themselves.
*/
BTScanInsert
-_bt_mkscankey(Relation rel, IndexTuple itup)
+_bt_mkscankey(Relation rel, Relation heaprel, IndexTuple itup)
{
BTScanInsert key;
ScanKey skey;
@@ -112,7 +112,7 @@ _bt_mkscankey(Relation rel, IndexTuple itup)
key = palloc(offsetof(BTScanInsertData, scankeys) +
sizeof(ScanKeyData) * indnkeyatts);
if (itup)
- _bt_metaversion(rel, &key->heapkeyspace, &key->allequalimage);
+ _bt_metaversion(rel, heaprel, &key->heapkeyspace, &key->allequalimage);
else
{
/* Utility statement callers can set these fields themselves */
@@ -1761,7 +1761,8 @@ _bt_killitems(IndexScanDesc scan)
droppedpin = true;
/* Attempt to re-read the buffer, getting pin and lock. */
- buf = _bt_getbuf(scan->indexRelation, so->currPos.currPage, BT_READ);
+ buf = _bt_getbuf(scan->indexRelation, scan->heapRelation,
+ so->currPos.currPage, BT_READ);
page = BufferGetPage(buf);
if (BufferGetLSNAtomic(buf) == so->currPos.lsn)
diff --git a/src/backend/access/spgist/spgvacuum.c b/src/backend/access/spgist/spgvacuum.c
index 3adb18f2d8..2f4a4aad24 100644
--- a/src/backend/access/spgist/spgvacuum.c
+++ b/src/backend/access/spgist/spgvacuum.c
@@ -489,7 +489,7 @@ vacuumLeafRoot(spgBulkDeleteState *bds, Relation index, Buffer buffer)
* Unlike the routines above, this works on both leaf and inner pages.
*/
static void
-vacuumRedirectAndPlaceholder(Relation index, Buffer buffer)
+vacuumRedirectAndPlaceholder(Relation index, Relation heaprel, Buffer buffer)
{
Page page = BufferGetPage(buffer);
SpGistPageOpaque opaque = SpGistPageGetOpaque(page);
@@ -503,6 +503,7 @@ vacuumRedirectAndPlaceholder(Relation index, Buffer buffer)
spgxlogVacuumRedirect xlrec;
GlobalVisState *vistest;
+ xlrec.isCatalogRel = RelationIsAccessibleInLogicalDecoding(heaprel);
xlrec.nToPlaceholder = 0;
xlrec.snapshotConflictHorizon = InvalidTransactionId;
@@ -643,13 +644,13 @@ spgvacuumpage(spgBulkDeleteState *bds, BlockNumber blkno)
else
{
vacuumLeafPage(bds, index, buffer, false);
- vacuumRedirectAndPlaceholder(index, buffer);
+ vacuumRedirectAndPlaceholder(index, bds->info->heaprel, buffer);
}
}
else
{
/* inner page */
- vacuumRedirectAndPlaceholder(index, buffer);
+ vacuumRedirectAndPlaceholder(index, bds->info->heaprel, buffer);
}
/*
@@ -719,7 +720,7 @@ spgprocesspending(spgBulkDeleteState *bds)
/* deal with any deletable tuples */
vacuumLeafPage(bds, index, buffer, true);
/* might as well do this while we are here */
- vacuumRedirectAndPlaceholder(index, buffer);
+ vacuumRedirectAndPlaceholder(index, bds->info->heaprel, buffer);
SpGistSetLastUsedPage(index, buffer);
diff --git a/src/backend/catalog/index.c b/src/backend/catalog/index.c
index 41b16cb89b..48d1d6b506 100644
--- a/src/backend/catalog/index.c
+++ b/src/backend/catalog/index.c
@@ -3352,6 +3352,7 @@ validate_index(Oid heapId, Oid indexId, Snapshot snapshot)
ivinfo.message_level = DEBUG2;
ivinfo.num_heap_tuples = heapRelation->rd_rel->reltuples;
ivinfo.strategy = NULL;
+ ivinfo.heaprel = heapRelation;
/*
* Encode TIDs as int8 values for the sort, rather than directly sorting
diff --git a/src/backend/commands/analyze.c b/src/backend/commands/analyze.c
index c86e690980..321fc0d31b 100644
--- a/src/backend/commands/analyze.c
+++ b/src/backend/commands/analyze.c
@@ -712,6 +712,7 @@ do_analyze_rel(Relation onerel, VacuumParams *params,
ivinfo.message_level = elevel;
ivinfo.num_heap_tuples = onerel->rd_rel->reltuples;
ivinfo.strategy = vac_strategy;
+ ivinfo.heaprel = onerel;
stats = index_vacuum_cleanup(&ivinfo, NULL);
diff --git a/src/backend/commands/vacuumparallel.c b/src/backend/commands/vacuumparallel.c
index bcd40c80a1..2cdbd182b6 100644
--- a/src/backend/commands/vacuumparallel.c
+++ b/src/backend/commands/vacuumparallel.c
@@ -148,6 +148,9 @@ struct ParallelVacuumState
/* NULL for worker processes */
ParallelContext *pcxt;
+ /* Parent Heap Relation */
+ Relation heaprel;
+
/* Target indexes */
Relation *indrels;
int nindexes;
@@ -266,6 +269,7 @@ parallel_vacuum_init(Relation rel, Relation *indrels, int nindexes,
pvs->nindexes = nindexes;
pvs->will_parallel_vacuum = will_parallel_vacuum;
pvs->bstrategy = bstrategy;
+ pvs->heaprel = rel;
EnterParallelMode();
pcxt = CreateParallelContext("postgres", "parallel_vacuum_main",
@@ -838,6 +842,7 @@ parallel_vacuum_process_one_index(ParallelVacuumState *pvs, Relation indrel,
ivinfo.estimated_count = pvs->shared->estimated_count;
ivinfo.num_heap_tuples = pvs->shared->reltuples;
ivinfo.strategy = pvs->bstrategy;
+ ivinfo.heaprel = pvs->heaprel;
/* Update error traceback information */
pvs->indname = pstrdup(RelationGetRelationName(indrel));
@@ -1007,6 +1012,7 @@ parallel_vacuum_main(dsm_segment *seg, shm_toc *toc)
pvs.dead_items = dead_items;
pvs.relnamespace = get_namespace_name(RelationGetNamespace(rel));
pvs.relname = pstrdup(RelationGetRelationName(rel));
+ pvs.heaprel = rel;
/* These fields will be filled during index vacuum or cleanup */
pvs.indname = NULL;
diff --git a/src/backend/optimizer/util/plancat.c b/src/backend/optimizer/util/plancat.c
index d58c4a1078..e3824efe9b 100644
--- a/src/backend/optimizer/util/plancat.c
+++ b/src/backend/optimizer/util/plancat.c
@@ -462,7 +462,7 @@ get_relation_info(PlannerInfo *root, Oid relationObjectId, bool inhparent,
* For btrees, get tree height while we have the index
* open
*/
- info->tree_height = _bt_getrootheight(indexRelation);
+ info->tree_height = _bt_getrootheight(indexRelation, relation);
}
else
{
diff --git a/src/backend/utils/sort/tuplesortvariants.c b/src/backend/utils/sort/tuplesortvariants.c
index eb6cfcfd00..0188106925 100644
--- a/src/backend/utils/sort/tuplesortvariants.c
+++ b/src/backend/utils/sort/tuplesortvariants.c
@@ -207,6 +207,7 @@ tuplesort_begin_heap(TupleDesc tupDesc,
Tuplesortstate *
tuplesort_begin_cluster(TupleDesc tupDesc,
Relation indexRel,
+ Relation heaprel,
int workMem,
SortCoordinate coordinate, int sortopt)
{
@@ -260,7 +261,7 @@ tuplesort_begin_cluster(TupleDesc tupDesc,
arg->tupDesc = tupDesc; /* assume we need not copy tupDesc */
- indexScanKey = _bt_mkscankey(indexRel, NULL);
+ indexScanKey = _bt_mkscankey(indexRel, heaprel, NULL);
if (arg->indexInfo->ii_Expressions != NULL)
{
@@ -361,7 +362,7 @@ tuplesort_begin_index_btree(Relation heapRel,
arg->enforceUnique = enforceUnique;
arg->uniqueNullsNotDistinct = uniqueNullsNotDistinct;
- indexScanKey = _bt_mkscankey(indexRel, NULL);
+ indexScanKey = _bt_mkscankey(indexRel, heapRel, NULL);
/* Prepare SortSupport data for each column */
base->sortKeys = (SortSupport) palloc0(base->nKeys *
diff --git a/src/include/access/genam.h b/src/include/access/genam.h
index 83dbee0fe6..7708b82d7d 100644
--- a/src/include/access/genam.h
+++ b/src/include/access/genam.h
@@ -50,6 +50,7 @@ typedef struct IndexVacuumInfo
int message_level; /* ereport level for progress messages */
double num_heap_tuples; /* tuples remaining in heap */
BufferAccessStrategy strategy; /* access strategy for reads */
+ Relation heaprel; /* the heap relation the index belongs to */
} IndexVacuumInfo;
/*
diff --git a/src/include/access/gist_private.h b/src/include/access/gist_private.h
index 8af33d7b40..ee275650bd 100644
--- a/src/include/access/gist_private.h
+++ b/src/include/access/gist_private.h
@@ -440,7 +440,7 @@ extern XLogRecPtr gistXLogPageDelete(Buffer buffer,
FullTransactionId xid, Buffer parentBuffer,
OffsetNumber downlinkOffset);
-extern void gistXLogPageReuse(Relation rel, BlockNumber blkno,
+extern void gistXLogPageReuse(Relation rel, Relation heaprel, BlockNumber blkno,
FullTransactionId deleteXid);
extern XLogRecPtr gistXLogUpdate(Buffer buffer,
@@ -449,7 +449,8 @@ extern XLogRecPtr gistXLogUpdate(Buffer buffer,
Buffer leftchildbuf);
extern XLogRecPtr gistXLogDelete(Buffer buffer, OffsetNumber *todelete,
- int ntodelete, TransactionId snapshotConflictHorizon);
+ int ntodelete, TransactionId snapshotConflictHorizon,
+ Relation heaprel);
extern XLogRecPtr gistXLogSplit(bool page_is_leaf,
SplitedPageLayout *dist,
@@ -485,7 +486,7 @@ extern bool gistproperty(Oid index_oid, int attno,
extern bool gistfitpage(IndexTuple *itvec, int len);
extern bool gistnospace(Page page, IndexTuple *itvec, int len, OffsetNumber todelete, Size freespace);
extern void gistcheckpage(Relation rel, Buffer buf);
-extern Buffer gistNewBuffer(Relation r);
+extern Buffer gistNewBuffer(Relation r, Relation heaprel);
extern bool gistPageRecyclable(Page page);
extern void gistfillbuffer(Page page, IndexTuple *itup, int len,
OffsetNumber off);
diff --git a/src/include/access/gistxlog.h b/src/include/access/gistxlog.h
index 09f9b0f8c6..2eea866f06 100644
--- a/src/include/access/gistxlog.h
+++ b/src/include/access/gistxlog.h
@@ -51,13 +51,14 @@ typedef struct gistxlogDelete
{
TransactionId snapshotConflictHorizon;
uint16 ntodelete; /* number of deleted offsets */
+ bool isCatalogRel; /* to handle recovery conflict during logical
+ * decoding on standby */
- /*
- * In payload of blk 0 : todelete OffsetNumbers
- */
+ /* TODELETE OFFSET NUMBERS */
+ OffsetNumber offsets[FLEXIBLE_ARRAY_MEMBER];
} gistxlogDelete;
-#define SizeOfGistxlogDelete (offsetof(gistxlogDelete, ntodelete) + sizeof(uint16))
+#define SizeOfGistxlogDelete offsetof(gistxlogDelete, offsets)
/*
* Backup Blk 0: If this operation completes a page split, by inserting a
@@ -100,9 +101,11 @@ typedef struct gistxlogPageReuse
RelFileLocator locator;
BlockNumber block;
FullTransactionId snapshotConflictHorizon;
+ bool isCatalogRel; /* to handle recovery conflict during logical
+ * decoding on standby */
} gistxlogPageReuse;
-#define SizeOfGistxlogPageReuse (offsetof(gistxlogPageReuse, snapshotConflictHorizon) + sizeof(FullTransactionId))
+#define SizeOfGistxlogPageReuse (offsetof(gistxlogPageReuse, isCatalogRel) + sizeof(bool))
extern void gist_redo(XLogReaderState *record);
extern void gist_desc(StringInfo buf, XLogReaderState *record);
diff --git a/src/include/access/hash_xlog.h b/src/include/access/hash_xlog.h
index a2f0f39213..7e9e47ce67 100644
--- a/src/include/access/hash_xlog.h
+++ b/src/include/access/hash_xlog.h
@@ -252,12 +252,14 @@ typedef struct xl_hash_vacuum_one_page
{
TransactionId snapshotConflictHorizon;
int ntuples;
+ bool isCatalogRel; /* to handle recovery conflict during logical
+ * decoding on standby */
- /* TARGET OFFSET NUMBERS FOLLOW AT THE END */
+ /* TARGET OFFSET NUMBERS */
+ OffsetNumber offsets[FLEXIBLE_ARRAY_MEMBER];
} xl_hash_vacuum_one_page;
-#define SizeOfHashVacuumOnePage \
- (offsetof(xl_hash_vacuum_one_page, ntuples) + sizeof(int))
+#define SizeOfHashVacuumOnePage offsetof(xl_hash_vacuum_one_page, offsets)
extern void hash_redo(XLogReaderState *record);
extern void hash_desc(StringInfo buf, XLogReaderState *record);
diff --git a/src/include/access/heapam_xlog.h b/src/include/access/heapam_xlog.h
index 8cb0d8da19..223db4b199 100644
--- a/src/include/access/heapam_xlog.h
+++ b/src/include/access/heapam_xlog.h
@@ -245,10 +245,12 @@ typedef struct xl_heap_prune
TransactionId snapshotConflictHorizon;
uint16 nredirected;
uint16 ndead;
+ bool isCatalogRel; /* to handle recovery conflict during logical
+ * decoding on standby */
/* OFFSET NUMBERS are in the block reference 0 */
} xl_heap_prune;
-#define SizeOfHeapPrune (offsetof(xl_heap_prune, ndead) + sizeof(uint16))
+#define SizeOfHeapPrune (offsetof(xl_heap_prune, isCatalogRel) + sizeof(bool))
/*
* The vacuum page record is similar to the prune record, but can only mark
@@ -344,12 +346,14 @@ typedef struct xl_heap_freeze_page
{
TransactionId snapshotConflictHorizon;
uint16 nplans;
+ bool isCatalogRel; /* to handle recovery conflict during logical
+ * decoding on standby */
/* FREEZE PLANS FOLLOW */
/* OFFSET NUMBER ARRAY FOLLOWS */
} xl_heap_freeze_page;
-#define SizeOfHeapFreezePage (offsetof(xl_heap_freeze_page, nplans) + sizeof(uint16))
+#define SizeOfHeapFreezePage (offsetof(xl_heap_freeze_page, isCatalogRel) + sizeof(bool))
/*
* This is what we need to know about setting a visibility map bit
@@ -408,7 +412,7 @@ extern void heap2_desc(StringInfo buf, XLogReaderState *record);
extern const char *heap2_identify(uint8 info);
extern void heap_xlog_logical_rewrite(XLogReaderState *r);
-extern XLogRecPtr log_heap_visible(RelFileLocator rlocator, Buffer heap_buffer,
+extern XLogRecPtr log_heap_visible(Relation rel, Buffer heap_buffer,
Buffer vm_buffer,
TransactionId snapshotConflictHorizon,
uint8 vmflags);
diff --git a/src/include/access/nbtree.h b/src/include/access/nbtree.h
index 8f48960f9d..6dee307042 100644
--- a/src/include/access/nbtree.h
+++ b/src/include/access/nbtree.h
@@ -1182,8 +1182,10 @@ extern IndexTuple _bt_swap_posting(IndexTuple newitem, IndexTuple oposting,
extern bool _bt_doinsert(Relation rel, IndexTuple itup,
IndexUniqueCheck checkUnique, bool indexUnchanged,
Relation heapRel);
-extern void _bt_finish_split(Relation rel, Buffer lbuf, BTStack stack);
-extern Buffer _bt_getstackbuf(Relation rel, BTStack stack, BlockNumber child);
+extern void _bt_finish_split(Relation rel, Relation heaprel, Buffer lbuf,
+ BTStack stack);
+extern Buffer _bt_getstackbuf(Relation rel, Relation heaprel, BTStack stack,
+ BlockNumber child);
/*
* prototypes for functions in nbtsplitloc.c
@@ -1197,16 +1199,18 @@ extern OffsetNumber _bt_findsplitloc(Relation rel, Page origpage,
*/
extern void _bt_initmetapage(Page page, BlockNumber rootbknum, uint32 level,
bool allequalimage);
-extern bool _bt_vacuum_needs_cleanup(Relation rel);
-extern void _bt_set_cleanup_info(Relation rel, BlockNumber num_delpages);
+extern bool _bt_vacuum_needs_cleanup(Relation rel, Relation heaprel);
+extern void _bt_set_cleanup_info(Relation rel, Relation heaprel,
+ BlockNumber num_delpages);
extern void _bt_upgrademetapage(Page page);
-extern Buffer _bt_getroot(Relation rel, int access);
-extern Buffer _bt_gettrueroot(Relation rel);
-extern int _bt_getrootheight(Relation rel);
-extern void _bt_metaversion(Relation rel, bool *heapkeyspace,
+extern Buffer _bt_getroot(Relation rel, Relation heaprel, int access);
+extern Buffer _bt_gettrueroot(Relation rel, Relation heaprel);
+extern int _bt_getrootheight(Relation rel, Relation heaprel);
+extern void _bt_metaversion(Relation rel, Relation heaprel, bool *heapkeyspace,
bool *allequalimage);
extern void _bt_checkpage(Relation rel, Buffer buf);
-extern Buffer _bt_getbuf(Relation rel, BlockNumber blkno, int access);
+extern Buffer _bt_getbuf(Relation rel, Relation heaprel, BlockNumber blkno,
+ int access);
extern Buffer _bt_relandgetbuf(Relation rel, Buffer obuf,
BlockNumber blkno, int access);
extern void _bt_relbuf(Relation rel, Buffer buf);
@@ -1229,21 +1233,22 @@ extern void _bt_pendingfsm_finalize(Relation rel, BTVacState *vstate);
/*
* prototypes for functions in nbtsearch.c
*/
-extern BTStack _bt_search(Relation rel, BTScanInsert key, Buffer *bufP,
- int access, Snapshot snapshot);
-extern Buffer _bt_moveright(Relation rel, BTScanInsert key, Buffer buf,
- bool forupdate, BTStack stack, int access, Snapshot snapshot);
+extern BTStack _bt_search(Relation rel, Relation heaprel, BTScanInsert key,
+ Buffer *bufP, int access, Snapshot snapshot);
+extern Buffer _bt_moveright(Relation rel, Relation heaprel, BTScanInsert key,
+ Buffer buf, bool forupdate, BTStack stack,
+ int access, Snapshot snapshot);
extern OffsetNumber _bt_binsrch_insert(Relation rel, BTInsertState insertstate);
extern int32 _bt_compare(Relation rel, BTScanInsert key, Page page, OffsetNumber offnum);
extern bool _bt_first(IndexScanDesc scan, ScanDirection dir);
extern bool _bt_next(IndexScanDesc scan, ScanDirection dir);
-extern Buffer _bt_get_endpoint(Relation rel, uint32 level, bool rightmost,
- Snapshot snapshot);
+extern Buffer _bt_get_endpoint(Relation rel, Relation heaprel, uint32 level,
+ bool rightmost, Snapshot snapshot);
/*
* prototypes for functions in nbtutils.c
*/
-extern BTScanInsert _bt_mkscankey(Relation rel, IndexTuple itup);
+extern BTScanInsert _bt_mkscankey(Relation rel, Relation heaprel, IndexTuple itup);
extern void _bt_freestack(BTStack stack);
extern void _bt_preprocess_array_keys(IndexScanDesc scan);
extern void _bt_start_array_keys(IndexScanDesc scan, ScanDirection dir);
diff --git a/src/include/access/nbtxlog.h b/src/include/access/nbtxlog.h
index edd1333d9b..1e45d58845 100644
--- a/src/include/access/nbtxlog.h
+++ b/src/include/access/nbtxlog.h
@@ -188,9 +188,11 @@ typedef struct xl_btree_reuse_page
RelFileLocator locator;
BlockNumber block;
FullTransactionId snapshotConflictHorizon;
+ bool isCatalogRel; /* to handle recovery conflict during logical
+ * decoding on standby */
} xl_btree_reuse_page;
-#define SizeOfBtreeReusePage (sizeof(xl_btree_reuse_page))
+#define SizeOfBtreeReusePage (offsetof(xl_btree_reuse_page, isCatalogRel) + sizeof(bool))
/*
* xl_btree_vacuum and xl_btree_delete records describe deletion of index
@@ -235,13 +237,15 @@ typedef struct xl_btree_delete
TransactionId snapshotConflictHorizon;
uint16 ndeleted;
uint16 nupdated;
+ bool isCatalogRel; /* to handle recovery conflict during logical
+ * decoding on standby */
/* DELETED TARGET OFFSET NUMBERS FOLLOW */
/* UPDATED TARGET OFFSET NUMBERS FOLLOW */
/* UPDATED TUPLES METADATA (xl_btree_update) ARRAY FOLLOWS */
} xl_btree_delete;
-#define SizeOfBtreeDelete (offsetof(xl_btree_delete, nupdated) + sizeof(uint16))
+#define SizeOfBtreeDelete (offsetof(xl_btree_delete, isCatalogRel) + sizeof(bool))
/*
* The offsets that appear in xl_btree_update metadata are offsets into the
diff --git a/src/include/access/spgxlog.h b/src/include/access/spgxlog.h
index b9d6753533..75267a4914 100644
--- a/src/include/access/spgxlog.h
+++ b/src/include/access/spgxlog.h
@@ -240,6 +240,8 @@ typedef struct spgxlogVacuumRedirect
uint16 nToPlaceholder; /* number of redirects to make placeholders */
OffsetNumber firstPlaceholder; /* first placeholder tuple to remove */
TransactionId snapshotConflictHorizon; /* newest XID of removed redirects */
+ bool isCatalogRel; /* to handle recovery conflict during logical
+ * decoding on standby */
/* offsets of redirect tuples to make placeholders follow */
OffsetNumber offsets[FLEXIBLE_ARRAY_MEMBER];
diff --git a/src/include/access/visibilitymapdefs.h b/src/include/access/visibilitymapdefs.h
index 9165b9456b..7306a1c3ee 100644
--- a/src/include/access/visibilitymapdefs.h
+++ b/src/include/access/visibilitymapdefs.h
@@ -17,9 +17,11 @@
#define BITS_PER_HEAPBLOCK 2
/* Flags for bit map */
-#define VISIBILITYMAP_ALL_VISIBLE 0x01
-#define VISIBILITYMAP_ALL_FROZEN 0x02
-#define VISIBILITYMAP_VALID_BITS 0x03 /* OR of all valid visibilitymap
- * flags bits */
+#define VISIBILITYMAP_ALL_VISIBLE 0x01
+#define VISIBILITYMAP_ALL_FROZEN 0x02
+#define VISIBILITYMAP_VALID_BITS 0x03 /* OR of all valid visibilitymap
+ * flags bits */
+#define VISIBILITYMAP_IS_CATALOG_REL 0x04 /* to handle recovery conflict during logical
+ * decoding on standby */
#endif /* VISIBILITYMAPDEFS_H */
diff --git a/src/include/utils/rel.h b/src/include/utils/rel.h
index af9785038d..0cfe02aa4a 100644
--- a/src/include/utils/rel.h
+++ b/src/include/utils/rel.h
@@ -27,6 +27,7 @@
#include "storage/smgr.h"
#include "utils/relcache.h"
#include "utils/reltrigger.h"
+#include "catalog/catalog.h"
/*
diff --git a/src/include/utils/tuplesort.h b/src/include/utils/tuplesort.h
index 12578e42bc..395abfe596 100644
--- a/src/include/utils/tuplesort.h
+++ b/src/include/utils/tuplesort.h
@@ -399,7 +399,9 @@ extern Tuplesortstate *tuplesort_begin_heap(TupleDesc tupDesc,
int workMem, SortCoordinate coordinate,
int sortopt);
extern Tuplesortstate *tuplesort_begin_cluster(TupleDesc tupDesc,
- Relation indexRel, int workMem,
+ Relation indexRel,
+ Relation heaprel,
+ int workMem,
SortCoordinate coordinate,
int sortopt);
extern Tuplesortstate *tuplesort_begin_index_btree(Relation heapRel,
--
2.34.1
Attachments:
[text/plain] v45-0006-Doc-changes-describing-details-about-logical-dec.patch (2.1K, ../[email protected]/2-v45-0006-Doc-changes-describing-details-about-logical-dec.patch)
download | inline diff:
From f45d02edeeed24e604c1e48cf39c09224eac1644 Mon Sep 17 00:00:00 2001
From: bdrouvotAWS <[email protected]>
Date: Mon, 30 Jan 2023 15:44:00 +0000
Subject: [PATCH v45 6/6] Doc changes describing details about logical
decoding.
Author: Andres Freund (in an older version), Amit Khandekar, Bertrand Drouvot
Reviewed-By: Bertrand Drouvot, Andres Freund, Robert Haas, Fabrizio de Royes Mello
---
doc/src/sgml/logicaldecoding.sgml | 21 +++++++++++++++++++++
1 file changed, 21 insertions(+)
100.0% doc/src/sgml/
diff --git a/doc/src/sgml/logicaldecoding.sgml b/doc/src/sgml/logicaldecoding.sgml
index 4e912b4bd4..2e8bee033f 100644
--- a/doc/src/sgml/logicaldecoding.sgml
+++ b/doc/src/sgml/logicaldecoding.sgml
@@ -316,6 +316,27 @@ postgres=# select * from pg_logical_slot_get_changes('regression_slot', NULL, NU
may consume changes from a slot at any given time.
</para>
+ <para>
+ A logical replication slot can also be created on a hot standby. To prevent
+ <command>VACUUM</command> from removing required rows from the system
+ catalogs, <varname>hot_standby_feedback</varname> should be set on the
+ standby. In spite of that, if any required rows get removed, the slot gets
+ invalidated. It's highly recommended to use a physical slot between the primary
+ and the standby. Otherwise, hot_standby_feedback will work, but only while the
+ connection is alive (for example a node restart would break it). Existing
+ logical slots on standby also get invalidated if wal_level on primary is reduced to
+ less than 'logical'.
+ </para>
+
+ <para>
+ For a logical slot to be created, it builds a historic snapshot, for which
+ information of all the currently running transactions is essential. On
+ primary, this information is available, but on standby, this information
+ has to be obtained from primary. So, slot creation may wait for some
+ activity to happen on the primary. If the primary is idle, creating a
+ logical slot on standby may take a noticeable time.
+ </para>
+
<caution>
<para>
Replication slots persist across crashes and know nothing about the state
--
2.34.1
[text/plain] v45-0005-New-TAP-test-for-logical-decoding-on-standby.patch (26.5K, ../[email protected]/3-v45-0005-New-TAP-test-for-logical-decoding-on-standby.patch)
download | inline diff:
From 61471ca687aa95eb334fd23448836bc8100ed92f Mon Sep 17 00:00:00 2001
From: bdrouvotAWS <[email protected]>
Date: Mon, 30 Jan 2023 15:43:21 +0000
Subject: [PATCH v45 5/6] New TAP test for logical decoding on standby.
Author: Craig Ringer (in an older version), Amit Khandekar, Bertrand Drouvot
Reviewed-By: Bertrand Drouvot, Andres Freund, Robert Haas, Fabrizio de Royes Mello
---
src/test/perl/PostgreSQL/Test/Cluster.pm | 39 ++
src/test/recovery/meson.build | 1 +
.../t/034_standby_logical_decoding.pl | 658 ++++++++++++++++++
3 files changed, 698 insertions(+)
5.0% src/test/perl/PostgreSQL/Test/
94.8% src/test/recovery/t/
diff --git a/src/test/perl/PostgreSQL/Test/Cluster.pm b/src/test/perl/PostgreSQL/Test/Cluster.pm
index 04921ca3a3..fd81ddcf39 100644
--- a/src/test/perl/PostgreSQL/Test/Cluster.pm
+++ b/src/test/perl/PostgreSQL/Test/Cluster.pm
@@ -3037,6 +3037,45 @@ $SIG{TERM} = $SIG{INT} = sub {
=pod
+=item $node->create_logical_slot_on_standby(self, primary, slot_name, dbname)
+
+Create logical replication slot on given standby
+
+=cut
+
+sub create_logical_slot_on_standby
+{
+ my ($self, $primary, $slot_name, $dbname) = @_;
+ my ($stdout, $stderr);
+
+ my $handle;
+
+ $handle = IPC::Run::start(['pg_recvlogical', '-d', $self->connstr($dbname), '-P', 'test_decoding', '-S', $slot_name, '--create-slot'], '>', \$stdout, '2>', \$stderr);
+
+ # Once slot restart_lsn is created, the standby looks for xl_running_xacts
+ # WAL record from the restart_lsn onwards. So firstly, wait until the slot
+ # restart_lsn is evaluated.
+
+ $self->poll_query_until(
+ 'postgres', qq[
+ SELECT restart_lsn IS NOT NULL
+ FROM pg_catalog.pg_replication_slots WHERE slot_name = '$slot_name'
+ ]) or die "timed out waiting for logical slot to calculate its restart_lsn";
+
+ # Now arrange for the xl_running_xacts record for which pg_recvlogical
+ # is waiting.
+ # Note: Write a C helper function to call LogStandbySnapshot() instead
+ # of asking for a checkpoint.
+ $primary->safe_psql('postgres', 'CHECKPOINT');
+
+ $handle->finish();
+
+ is($self->slot($slot_name)->{'slot_type'}, 'logical', $slot_name . ' on standby created')
+ or die "could not create slot" . $slot_name;
+}
+
+=pod
+
=back
=cut
diff --git a/src/test/recovery/meson.build b/src/test/recovery/meson.build
index edaaa1a3ce..52b2816c7a 100644
--- a/src/test/recovery/meson.build
+++ b/src/test/recovery/meson.build
@@ -40,6 +40,7 @@ tests += {
't/031_recovery_conflict.pl',
't/032_relfilenode_reuse.pl',
't/033_replay_tsp_drops.pl',
+ 't/034_standby_logical_decoding.pl',
],
},
}
diff --git a/src/test/recovery/t/034_standby_logical_decoding.pl b/src/test/recovery/t/034_standby_logical_decoding.pl
new file mode 100644
index 0000000000..4370d595d8
--- /dev/null
+++ b/src/test/recovery/t/034_standby_logical_decoding.pl
@@ -0,0 +1,658 @@
+# logical decoding on standby : test logical decoding,
+# recovery conflict and standby promotion.
+
+use strict;
+use warnings;
+
+use PostgreSQL::Test::Cluster;
+use PostgreSQL::Test::Utils;
+use Test::More tests => 62;
+
+my ($stdin, $stdout, $stderr, $ret, $handle, $slot);
+
+my $node_primary = PostgreSQL::Test::Cluster->new('primary');
+my $node_standby = PostgreSQL::Test::Cluster->new('standby');
+my $default_timeout = $PostgreSQL::Test::Utils::timeout_default;
+my $res;
+
+# Name for the physical slot on primary
+my $primary_slotname = 'primary_physical';
+
+# find $pat in logfile of $node after $off-th byte
+sub find_in_log
+{
+ my ($node, $pat, $off) = @_;
+
+ $off = 0 unless defined $off;
+ my $log = PostgreSQL::Test::Utils::slurp_file($node->logfile);
+ return 0 if (length($log) <= $off);
+
+ $log = substr($log, $off);
+
+ return $log =~ m/$pat/;
+}
+
+# Fetch xmin columns from slot's pg_replication_slots row, after waiting for
+# given boolean condition to be true to ensure we've reached a quiescent state.
+sub wait_for_xmins
+{
+ my ($node, $slotname, $check_expr) = @_;
+
+ $node->poll_query_until(
+ 'postgres', qq[
+ SELECT $check_expr
+ FROM pg_catalog.pg_replication_slots
+ WHERE slot_name = '$slotname';
+ ]) or die "Timed out waiting for slot xmins to advance";
+}
+
+# Create the required logical slots on standby.
+sub create_logical_slots
+{
+ $node_standby->create_logical_slot_on_standby($node_primary, 'inactiveslot', 'testdb');
+ $node_standby->create_logical_slot_on_standby($node_primary, 'activeslot', 'testdb');
+}
+
+# Drop the logical slots on standby.
+sub drop_logical_slots
+{
+ $node_standby->psql('postgres', q[SELECT pg_drop_replication_slot('inactiveslot')]);
+ $node_standby->psql('postgres', q[SELECT pg_drop_replication_slot('activeslot')]);
+}
+
+# Acquire one of the standby logical slots created by create_logical_slots().
+# In case wait is true we are waiting for an active pid on the 'activeslot' slot.
+# If wait is not true it means we are testing a known failure scenario.
+sub make_slot_active
+{
+ my $wait = shift;
+ my $slot_user_handle;
+
+ $slot_user_handle = IPC::Run::start(['pg_recvlogical', '-d', $node_standby->connstr('testdb'), '-S', 'activeslot', '-o', 'include-xids=0', '-o', 'skip-empty-xacts=1', '--no-loop', '--start', '-f', '-'], '>', \$stdout, '2>', \$stderr);
+
+ if ($wait)
+ {
+ # make sure activeslot is in use
+ $node_standby->poll_query_until('testdb',
+ "SELECT EXISTS (SELECT 1 FROM pg_replication_slots WHERE slot_name = 'activeslot' AND active_pid IS NOT NULL)"
+ ) or die "slot never became active";
+ }
+ return $slot_user_handle;
+}
+
+# Check pg_recvlogical stderr
+sub check_pg_recvlogical_stderr
+{
+ my ($slot_user_handle, $check_stderr) = @_;
+ my $return;
+
+ # our client should've terminated in response to the walsender error
+ $slot_user_handle->finish;
+ $return = $?;
+ cmp_ok($return, "!=", 0, "pg_recvlogical exited non-zero");
+ if ($return) {
+ like($stderr, qr/$check_stderr/, 'slot has been invalidated');
+ }
+
+ return 0;
+}
+
+# Check if all the slots on standby are dropped. These include the 'activeslot'
+# that was acquired by make_slot_active(), and the non-active 'inactiveslot'.
+sub check_slots_dropped
+{
+ my ($slot_user_handle) = @_;
+
+ is($node_standby->slot('inactiveslot')->{'slot_type'}, '', 'inactiveslot on standby dropped');
+ is($node_standby->slot('activeslot')->{'slot_type'}, '', 'activeslot on standby dropped');
+
+ check_pg_recvlogical_stderr($slot_user_handle, "conflict with recovery");
+}
+
+# Check if all the slots on standby are dropped. These include the 'activeslot'
+# that was acquired by make_slot_active(), and the non-active 'inactiveslot'.
+sub change_hot_standby_feedback_and_wait_for_xmins
+{
+ my ($hsf, $invalidated) = @_;
+
+ $node_standby->append_conf('postgresql.conf',qq[
+ hot_standby_feedback = $hsf
+ ]);
+
+ $node_standby->reload;
+
+ if ($hsf && $invalidated)
+ {
+ # With hot_standby_feedback on, xmin should advance,
+ # but catalog_xmin should still remain NULL since there is no logical slot.
+ wait_for_xmins($node_primary, $primary_slotname,
+ "xmin IS NOT NULL AND catalog_xmin IS NULL");
+ }
+ elsif ($hsf)
+ {
+ # With hot_standby_feedback on, xmin and catalog_xmin should advance.
+ wait_for_xmins($node_primary, $primary_slotname,
+ "xmin IS NOT NULL AND catalog_xmin IS NOT NULL");
+ }
+ else
+ {
+ # Both should be NULL since hs_feedback is off
+ wait_for_xmins($node_primary, $primary_slotname,
+ "xmin IS NULL AND catalog_xmin IS NULL");
+
+ }
+}
+
+# Check conflicting status in pg_replication_slots.
+sub check_slots_conflicting_status
+{
+ my ($conflicting) = @_;
+
+ if ($conflicting)
+ {
+ $res = $node_standby->safe_psql(
+ 'postgres', qq(
+ select bool_and(conflicting) from pg_replication_slots;));
+
+ is($res, 't',
+ "Logical slots are reported as conflicting");
+ }
+ else
+ {
+ $res = $node_standby->safe_psql(
+ 'postgres', qq(
+ select bool_or(conflicting) from pg_replication_slots;));
+
+ is($res, 'f',
+ "Logical slots are reported as non conflicting");
+ }
+}
+
+########################
+# Initialize primary node
+########################
+
+$node_primary->init(allows_streaming => 1, has_archiving => 1);
+$node_primary->append_conf('postgresql.conf', q{
+wal_level = 'logical'
+max_replication_slots = 4
+max_wal_senders = 4
+log_min_messages = 'debug2'
+log_error_verbosity = verbose
+});
+$node_primary->dump_info;
+$node_primary->start;
+
+$node_primary->psql('postgres', q[CREATE DATABASE testdb]);
+
+$node_primary->safe_psql('testdb', qq[SELECT * FROM pg_create_physical_replication_slot('$primary_slotname');]);
+
+# Check conflicting is NULL for physical slot
+$res = $node_primary->safe_psql(
+ 'postgres', qq[
+ SELECT conflicting is null FROM pg_replication_slots where slot_name = '$primary_slotname';]);
+
+is($res, 't',
+ "Physical slot reports conflicting as NULL");
+
+my $backup_name = 'b1';
+$node_primary->backup($backup_name);
+
+#######################
+# Initialize standby node
+#######################
+
+$node_standby->init_from_backup(
+ $node_primary, $backup_name,
+ has_streaming => 1,
+ has_restoring => 1);
+$node_standby->append_conf('postgresql.conf',
+ qq[primary_slot_name = '$primary_slotname']);
+$node_standby->start;
+$node_primary->wait_for_catchup($node_standby, 'replay', $node_primary->lsn('flush'));
+
+
+##################################################
+# Test that logical decoding on the standby
+# behaves correctly.
+##################################################
+
+# create the logical slots
+create_logical_slots();
+
+$node_primary->safe_psql('testdb', qq[CREATE TABLE decoding_test(x integer, y text);]);
+$node_primary->safe_psql('testdb', qq[INSERT INTO decoding_test(x,y) SELECT s, s::text FROM generate_series(1,10) s;]);
+
+$node_primary->wait_for_catchup($node_standby, 'replay', $node_primary->lsn('flush'));
+
+my $result = $node_standby->safe_psql('testdb',
+ qq[SELECT pg_logical_slot_get_changes('activeslot', NULL, NULL);]);
+
+# test if basic decoding works
+is(scalar(my @foobar = split /^/m, $result),
+ 14, 'Decoding produced 14 rows (2 BEGIN/COMMIT and 10 rows)');
+
+# Insert some rows and verify that we get the same results from pg_recvlogical
+# and the SQL interface.
+$node_primary->safe_psql('testdb',
+ qq[INSERT INTO decoding_test(x,y) SELECT s, s::text FROM generate_series(1,4) s;]
+);
+
+my $expected = q{BEGIN
+table public.decoding_test: INSERT: x[integer]:1 y[text]:'1'
+table public.decoding_test: INSERT: x[integer]:2 y[text]:'2'
+table public.decoding_test: INSERT: x[integer]:3 y[text]:'3'
+table public.decoding_test: INSERT: x[integer]:4 y[text]:'4'
+COMMIT};
+
+$node_primary->wait_for_catchup($node_standby, 'replay', $node_primary->lsn('flush'));
+
+my $stdout_sql = $node_standby->safe_psql('testdb',
+ qq[SELECT data FROM pg_logical_slot_peek_changes('activeslot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');]
+);
+
+is($stdout_sql, $expected, 'got expected output from SQL decoding session');
+
+my $endpos = $node_standby->safe_psql('testdb',
+ "SELECT lsn FROM pg_logical_slot_peek_changes('activeslot', NULL, NULL) ORDER BY lsn DESC LIMIT 1;"
+);
+
+# Insert some rows after $endpos, which we won't read.
+$node_primary->safe_psql('testdb',
+ qq[INSERT INTO decoding_test(x,y) SELECT s, s::text FROM generate_series(5,50) s;]
+);
+
+$node_primary->wait_for_catchup($node_standby, 'replay', $node_primary->lsn('flush'));
+
+my $stdout_recv = $node_standby->pg_recvlogical_upto(
+ 'testdb', 'activeslot', $endpos, $default_timeout,
+ 'include-xids' => '0',
+ 'skip-empty-xacts' => '1');
+chomp($stdout_recv);
+is($stdout_recv, $expected,
+ 'got same expected output from pg_recvlogical decoding session');
+
+$node_standby->poll_query_until('testdb',
+ "SELECT EXISTS (SELECT 1 FROM pg_replication_slots WHERE slot_name = 'activeslot' AND active_pid IS NULL)"
+) or die "slot never became inactive";
+
+$stdout_recv = $node_standby->pg_recvlogical_upto(
+ 'testdb', 'activeslot', $endpos, $default_timeout,
+ 'include-xids' => '0',
+ 'skip-empty-xacts' => '1');
+chomp($stdout_recv);
+is($stdout_recv, '', 'pg_recvlogical acknowledged changes');
+
+$node_primary->safe_psql('postgres', 'CREATE DATABASE otherdb');
+
+is( $node_primary->psql(
+ 'otherdb',
+ "SELECT lsn FROM pg_logical_slot_peek_changes('activeslot', NULL, NULL) ORDER BY lsn DESC LIMIT 1;"
+ ),
+ 3,
+ 'replaying logical slot from another database fails');
+
+# drop the logical slots
+drop_logical_slots();
+
+##################################################
+# Recovery conflict: Invalidate conflicting slots, including in-use slots
+# Scenario 1: hot_standby_feedback off and vacuum FULL
+##################################################
+
+# create the logical slots
+create_logical_slots();
+
+# One way to produce recovery conflict is to create/drop a relation and
+# launch a vacuum full on pg_class with hot_standby_feedback turned off on
+# the standby.
+change_hot_standby_feedback_and_wait_for_xmins(0,1);
+
+$handle = make_slot_active(1);
+
+# This should trigger the conflict
+$node_primary->safe_psql('testdb', qq[CREATE TABLE conflict_test(x integer, y text);]);
+$node_primary->safe_psql('testdb', qq[DROP TABLE conflict_test;]);
+$node_primary->safe_psql('testdb', 'VACUUM full pg_class;');
+
+$node_primary->wait_for_catchup($node_standby, 'replay', $node_primary->lsn('flush'));
+
+# message should be issued
+ok( find_in_log(
+ $node_standby,
+ "invalidating slot \"inactiveslot\" because it conflicts with recovery"),
+ 'inactiveslot slot invalidation is logged with vacuum FULL on pg_class');
+
+ok( find_in_log(
+ $node_standby,
+ "invalidating slot \"activeslot\" because it conflicts with recovery"),
+ 'activeslot slot invalidation is logged with vacuum FULL on pg_class');
+
+# Verify that pg_stat_database_conflicts.confl_active_logicalslot has been updated
+ok( $node_standby->poll_query_until(
+ 'postgres',
+ "select (confl_active_logicalslot = 1) from pg_stat_database_conflicts where datname = 'testdb'", 't'),
+ 'confl_active_logicalslot updated') or die "Timed out waiting confl_active_logicalslot to be updated";
+
+# Verify slots are reported as conflicting in pg_replication_slots
+check_slots_conflicting_status(1);
+
+$handle = make_slot_active(0);
+# We are not able to read from the slot as it has been invalidated
+check_pg_recvlogical_stderr($handle, "cannot read from logical replication slot \"activeslot\"");
+
+# Turn hot_standby_feedback back on
+change_hot_standby_feedback_and_wait_for_xmins(1,1);
+
+##################################################
+# Recovery conflict: Invalidate conflicting slots, including in-use slots
+# Scenario 2: conflict due to row removal with hot_standby_feedback off.
+##################################################
+
+# get the position to search from in the standby logfile
+my $logstart = -s $node_standby->logfile;
+
+# drop the logical slots
+drop_logical_slots();
+
+# create the logical slots
+create_logical_slots();
+
+# One way to produce recovery conflict is to create/drop a relation and
+# launch a vacuum on pg_class with hot_standby_feedback turned off on the standby.
+change_hot_standby_feedback_and_wait_for_xmins(0,1);
+
+$handle = make_slot_active(1);
+
+# This should trigger the conflict
+$node_primary->safe_psql('testdb', qq[CREATE TABLE conflict_test(x integer, y text);]);
+$node_primary->safe_psql('testdb', qq[DROP TABLE conflict_test;]);
+$node_primary->safe_psql('testdb', 'VACUUM pg_class;');
+
+$node_primary->wait_for_catchup($node_standby, 'replay', $node_primary->lsn('flush'));
+
+# message should be issued
+ok( find_in_log(
+ $node_standby,
+ "invalidating slot \"inactiveslot\" because it conflicts with recovery", $logstart),
+ 'inactiveslot slot invalidation is logged with vacuum on pg_class');
+
+ok( find_in_log(
+ $node_standby,
+ "invalidating slot \"activeslot\" because it conflicts with recovery", $logstart),
+ 'activeslot slot invalidation is logged with vacuum on pg_class');
+
+# Verify that pg_stat_database_conflicts.confl_active_logicalslot has been updated
+# we now expect 2 conflicts reported as the counter persist across reloads
+ok( $node_standby->poll_query_until(
+ 'postgres',
+ "select (confl_active_logicalslot = 2) from pg_stat_database_conflicts where datname = 'testdb'", 't'),
+ 'confl_active_logicalslot updated') or die "Timed out waiting confl_active_logicalslot to be updated";
+
+# Verify slots are reported as conflicting in pg_replication_slots
+check_slots_conflicting_status(1);
+
+$handle = make_slot_active(0);
+
+# We are not able to read from the slot as it has been invalidated
+check_pg_recvlogical_stderr($handle, "cannot read from logical replication slot \"activeslot\"");
+
+##################################################
+# Recovery conflict: Same as Scenario 2 but on a non catalog table
+# Scenario 3: No conflict expected.
+##################################################
+
+# get the position to search from in the standby logfile
+$logstart = -s $node_standby->logfile;
+
+# drop the logical slots
+drop_logical_slots();
+
+# create the logical slots
+create_logical_slots();
+
+# put hot standby feedback to off
+change_hot_standby_feedback_and_wait_for_xmins(0,1);
+
+$handle = make_slot_active(1);
+
+# This should not trigger a conflict
+$node_primary->safe_psql('testdb', qq[CREATE TABLE conflict_test(x integer, y text);]);
+$node_primary->safe_psql('testdb', qq[INSERT INTO conflict_test(x,y) SELECT s, s::text FROM generate_series(1,4) s;]);
+$node_primary->safe_psql('testdb', qq[UPDATE conflict_test set x=1, y=1;]);
+$node_primary->safe_psql('testdb', 'VACUUM conflict_test;');
+
+$node_primary->wait_for_catchup($node_standby, 'replay', $node_primary->lsn('flush'));
+
+# message should not be issued
+ok( !find_in_log(
+ $node_standby,
+ "invalidating slot \"inactiveslot\" because it conflicts with recovery", $logstart),
+ 'inactiveslot slot invalidation is not logged with vacuum on conflict_test');
+
+ok( !find_in_log(
+ $node_standby,
+ "invalidating slot \"activeslot\" because it conflicts with recovery", $logstart),
+ 'activeslot slot invalidation is not logged with vacuum on conflict_test');
+
+# Verify that pg_stat_database_conflicts.confl_active_logicalslot has not been updated
+# we now still expect 2 conflicts reported as the counter persist across reloads
+ok( $node_standby->poll_query_until(
+ 'postgres',
+ "select (confl_active_logicalslot = 2) from pg_stat_database_conflicts where datname = 'testdb'", 't'),
+ 'confl_active_logicalslot not updated') or die "Timed out waiting confl_active_logicalslot to be updated";
+
+# Verify slots are reported as non conflicting in pg_replication_slots
+check_slots_conflicting_status(0);
+
+# Turn hot_standby_feedback back on
+change_hot_standby_feedback_and_wait_for_xmins(1, 0);
+
+##################################################
+# Recovery conflict: Invalidate conflicting slots, including in-use slots
+# Scenario 4: conflict due to on-access pruning.
+##################################################
+
+# get the position to search from in the standby logfile
+$logstart = -s $node_standby->logfile;
+
+# drop the logical slots
+drop_logical_slots();
+
+# create the logical slots
+create_logical_slots();
+
+# One way to produce recovery conflict is to trigger an on-access pruning
+# on a relation marked as user_catalog_table.
+change_hot_standby_feedback_and_wait_for_xmins(0,0);
+
+$handle = make_slot_active(1);
+
+# This should trigger the conflict
+$node_primary->safe_psql('testdb', qq[CREATE TABLE prun(id integer, s char(2000)) WITH (fillfactor = 75, user_catalog_table = true);]);
+$node_primary->safe_psql('testdb', qq[INSERT INTO prun VALUES (1, 'A');]);
+$node_primary->safe_psql('testdb', qq[UPDATE prun SET s = 'B';]);
+$node_primary->safe_psql('testdb', qq[UPDATE prun SET s = 'C';]);
+$node_primary->safe_psql('testdb', qq[UPDATE prun SET s = 'D';]);
+$node_primary->safe_psql('testdb', qq[UPDATE prun SET s = 'E';]);
+
+$node_primary->wait_for_catchup($node_standby, 'replay', $node_primary->lsn('flush'));
+
+# message should be issued
+ok( find_in_log(
+ $node_standby,
+ "invalidating slot \"inactiveslot\" because it conflicts with recovery", $logstart),
+ 'inactiveslot slot invalidation is logged with on-access pruning');
+
+ok( find_in_log(
+ $node_standby,
+ "invalidating slot \"activeslot\" because it conflicts with recovery", $logstart),
+ 'activeslot slot invalidation is logged with on-access pruning');
+
+# Verify that pg_stat_database_conflicts.confl_active_logicalslot has been updated
+# we now expect 3 conflicts reported as the counter persist across reloads
+ok( $node_standby->poll_query_until(
+ 'postgres',
+ "select (confl_active_logicalslot = 3) from pg_stat_database_conflicts where datname = 'testdb'", 't'),
+ 'confl_active_logicalslot updated') or die "Timed out waiting confl_active_logicalslot to be updated";
+
+# Verify slots are reported as conflicting in pg_replication_slots
+check_slots_conflicting_status(1);
+
+$handle = make_slot_active(0);
+
+# We are not able to read from the slot as it has been invalidated
+check_pg_recvlogical_stderr($handle, "cannot read from logical replication slot \"activeslot\"");
+
+# Turn hot_standby_feedback back on
+change_hot_standby_feedback_and_wait_for_xmins(1, 1);
+
+##################################################
+# Recovery conflict: Invalidate conflicting slots, including in-use slots
+# Scenario 5: incorrect wal_level on primary.
+##################################################
+
+# get the position to search from in the standby logfile
+$logstart = -s $node_standby->logfile;
+
+# drop the logical slots
+drop_logical_slots();
+
+# create the logical slots
+create_logical_slots();
+
+$handle = make_slot_active(1);
+
+# Make primary wal_level replica. This will trigger slot conflict.
+$node_primary->append_conf('postgresql.conf',q[
+wal_level = 'replica'
+]);
+$node_primary->restart;
+
+$node_primary->wait_for_catchup($node_standby, 'replay', $node_primary->lsn('flush'));
+
+# message should be issued
+ok( find_in_log(
+ $node_standby,
+ "invalidating slot \"inactiveslot\" because it conflicts with recovery", $logstart),
+ 'inactiveslot slot invalidation is logged due to wal_level');
+
+ok( find_in_log(
+ $node_standby,
+ "invalidating slot \"activeslot\" because it conflicts with recovery", $logstart),
+ 'activeslot slot invalidation is logged due to wal_level');
+
+# Verify that pg_stat_database_conflicts.confl_active_logicalslot has been updated
+# we now expect 3 conflicts reported as the counter persist across reloads
+ok( $node_standby->poll_query_until(
+ 'postgres',
+ "select (confl_active_logicalslot = 4) from pg_stat_database_conflicts where datname = 'testdb'", 't'),
+ 'confl_active_logicalslot updated') or die "Timed out waiting confl_active_logicalslot to be updated";
+
+# Verify slots are reported as conflicting in pg_replication_slots
+check_slots_conflicting_status(1);
+
+$handle = make_slot_active(0);
+# We are not able to read from the slot as it requires wal_level at least logical on the primary server
+check_pg_recvlogical_stderr($handle, "logical decoding on standby requires wal_level to be at least logical on the primary server");
+
+# Restore primary wal_level
+$node_primary->append_conf('postgresql.conf',q[
+wal_level = 'logical'
+]);
+$node_primary->restart;
+$node_primary->wait_for_catchup($node_standby, 'replay', $node_primary->lsn('flush'));
+
+$handle = make_slot_active(0);
+# as the slot has been invalidated we should not be able to read
+check_pg_recvlogical_stderr($handle, "cannot read from logical replication slot \"activeslot\"");
+
+##################################################
+# DROP DATABASE should drops it's slots, including active slots.
+##################################################
+
+# drop the logical slots
+drop_logical_slots();
+
+# create the logical slots
+create_logical_slots();
+
+$handle = make_slot_active(1);
+# Create a slot on a database that would not be dropped. This slot should not
+# get dropped.
+$node_standby->create_logical_slot_on_standby($node_primary, 'otherslot', 'postgres');
+
+# dropdb on the primary to verify slots are dropped on standby
+$node_primary->safe_psql('postgres', q[DROP DATABASE testdb]);
+
+$node_primary->wait_for_catchup($node_standby, 'replay', $node_primary->lsn('flush'));
+
+is($node_standby->safe_psql('postgres',
+ q[SELECT EXISTS(SELECT 1 FROM pg_database WHERE datname = 'testdb')]), 'f',
+ 'database dropped on standby');
+
+check_slots_dropped($handle);
+
+is($node_standby->slot('otherslot')->{'slot_type'}, 'logical',
+ 'otherslot on standby not dropped');
+
+# Cleanup : manually drop the slot that was not dropped.
+$node_standby->psql('postgres', q[SELECT pg_drop_replication_slot('otherslot')]);
+
+##################################################
+# Test standby promotion and logical decoding behavior
+# after the standby gets promoted.
+##################################################
+
+$node_primary->psql('postgres', q[CREATE DATABASE testdb]);
+$node_primary->safe_psql('testdb', qq[CREATE TABLE decoding_test(x integer, y text);]);
+
+# create the logical slots
+create_logical_slots();
+$handle = make_slot_active(1);
+
+# Insert some rows before the promotion
+$node_primary->safe_psql('testdb',
+ qq[INSERT INTO decoding_test(x,y) SELECT s, s::text FROM generate_series(1,4) s;]
+);
+
+$node_primary->wait_for_catchup($node_standby, 'replay', $node_primary->lsn('flush'));
+
+# promote
+$node_standby->promote;
+
+# insert some rows on promoted standby
+$node_standby->safe_psql('testdb',
+ qq[INSERT INTO decoding_test(x,y) SELECT s, s::text FROM generate_series(5,7) s;]
+);
+
+$expected = q{BEGIN
+table public.decoding_test: INSERT: x[integer]:1 y[text]:'1'
+table public.decoding_test: INSERT: x[integer]:2 y[text]:'2'
+table public.decoding_test: INSERT: x[integer]:3 y[text]:'3'
+table public.decoding_test: INSERT: x[integer]:4 y[text]:'4'
+COMMIT
+BEGIN
+table public.decoding_test: INSERT: x[integer]:5 y[text]:'5'
+table public.decoding_test: INSERT: x[integer]:6 y[text]:'6'
+table public.decoding_test: INSERT: x[integer]:7 y[text]:'7'
+COMMIT};
+
+# check that we are decoding pre and post promotion inserted rows
+$stdout_sql = $node_standby->safe_psql('testdb',
+ qq[SELECT data FROM pg_logical_slot_peek_changes('inactiveslot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');]
+);
+
+is($stdout_sql, $expected, 'got expected output from SQL decoding session on promoted standby');
+
+# check that we are decoding pre and post promotion inserted rows
+# with pg_recvlogical that has started before the promotion
+my $pump_timeout = IPC::Run::timer($PostgreSQL::Test::Utils::timeout_default);
+
+ok( pump_until(
+ $handle, $pump_timeout, \$stdout, qr/^.*COMMIT.*COMMIT$/s),
+ 'got 2 COMMIT from pg_recvlogical output');
+
+chomp($stdout);
+is($stdout, $expected,
+ 'got same expected output from pg_recvlogical decoding session');
--
2.34.1
[text/plain] v45-0004-Fixing-Walsender-corner-case-with-logical-decodi.patch (7.5K, ../[email protected]/4-v45-0004-Fixing-Walsender-corner-case-with-logical-decodi.patch)
download | inline diff:
From 31f671b55fc9eb9eba2799182ae50c20ba6ed6c9 Mon Sep 17 00:00:00 2001
From: bdrouvotAWS <[email protected]>
Date: Mon, 30 Jan 2023 15:42:26 +0000
Subject: [PATCH v45 4/6] Fixing Walsender corner case with logical decoding on
standby.
The problem is that WalSndWaitForWal() waits for the *replay* LSN to
increase, but gets woken up by walreceiver when new WAL has been
flushed. Which means that typically walsenders will get woken up at the
same time that the startup process will be - which means that by the
time the logical walsender checks GetXLogReplayRecPtr() it's unlikely
that the startup process already replayed the record and updated
XLogCtl->lastReplayedEndRecPtr.
Introducing a new condition variable to fix this corner case.
---
src/backend/access/transam/xlogrecovery.c | 28 ++++++++++++++++++++
src/backend/replication/walsender.c | 31 +++++++++++++++++------
src/backend/utils/activity/wait_event.c | 3 +++
src/include/access/xlogrecovery.h | 3 +++
src/include/replication/walsender.h | 1 +
src/include/utils/wait_event.h | 1 +
6 files changed, 59 insertions(+), 8 deletions(-)
41.2% src/backend/access/transam/
48.5% src/backend/replication/
3.6% src/backend/utils/activity/
3.4% src/include/access/
diff --git a/src/backend/access/transam/xlogrecovery.c b/src/backend/access/transam/xlogrecovery.c
index 2a5352f879..bb0de527ab 100644
--- a/src/backend/access/transam/xlogrecovery.c
+++ b/src/backend/access/transam/xlogrecovery.c
@@ -358,6 +358,9 @@ typedef struct XLogRecoveryCtlData
RecoveryPauseState recoveryPauseState;
ConditionVariable recoveryNotPausedCV;
+ /* Replay state (see getReplayedCV() for more explanation) */
+ ConditionVariable replayedCV;
+
slock_t info_lck; /* locks shared variables shown above */
} XLogRecoveryCtlData;
@@ -467,6 +470,7 @@ XLogRecoveryShmemInit(void)
SpinLockInit(&XLogRecoveryCtl->info_lck);
InitSharedLatch(&XLogRecoveryCtl->recoveryWakeupLatch);
ConditionVariableInit(&XLogRecoveryCtl->recoveryNotPausedCV);
+ ConditionVariableInit(&XLogRecoveryCtl->replayedCV);
}
/*
@@ -1916,6 +1920,11 @@ ApplyWalRecord(XLogReaderState *xlogreader, XLogRecord *record, TimeLineID *repl
XLogRecoveryCtl->lastReplayedTLI = *replayTLI;
SpinLockRelease(&XLogRecoveryCtl->info_lck);
+ /*
+ * wake up walsender(s) used by logical decoding on standby.
+ */
+ ConditionVariableBroadcast(&XLogRecoveryCtl->replayedCV);
+
/*
* If rm_redo called XLogRequestWalReceiverReply, then we wake up the
* receiver so that it notices the updated lastReplayedEndRecPtr and sends
@@ -4923,3 +4932,22 @@ assign_recovery_target_xid(const char *newval, void *extra)
else
recoveryTarget = RECOVERY_TARGET_UNSET;
}
+
+/*
+ * Return the ConditionVariable indicating that a replay has been done.
+ *
+ * This is needed for logical decoding on standby. Indeed the "problem" is that
+ * WalSndWaitForWal() waits for the *replay* LSN to increase, but gets woken up
+ * by walreceiver when new WAL has been flushed. Which means that typically
+ * walsenders will get woken up at the same time that the startup process
+ * will be - which means that by the time the logical walsender checks
+ * GetXLogReplayRecPtr() it's unlikely that the startup process already replayed
+ * the record and updated XLogCtl->lastReplayedEndRecPtr.
+ *
+ * The ConditionVariable XLogRecoveryCtl->replayedCV solves this corner case.
+ */
+ConditionVariable *
+getReplayedCV(void)
+{
+ return &XLogRecoveryCtl->replayedCV;
+}
diff --git a/src/backend/replication/walsender.c b/src/backend/replication/walsender.c
index 1e91cbc564..b3fe5dbeb2 100644
--- a/src/backend/replication/walsender.c
+++ b/src/backend/replication/walsender.c
@@ -1552,6 +1552,7 @@ WalSndWaitForWal(XLogRecPtr loc)
{
int wakeEvents;
static XLogRecPtr RecentFlushPtr = InvalidXLogRecPtr;
+ ConditionVariable *replayedCV = getReplayedCV();
/*
* Fast path to avoid acquiring the spinlock in case we already know we
@@ -1570,7 +1571,6 @@ WalSndWaitForWal(XLogRecPtr loc)
for (;;)
{
- long sleeptime;
/* Clear any already-pending wakeups */
ResetLatch(MyLatch);
@@ -1654,20 +1654,35 @@ WalSndWaitForWal(XLogRecPtr loc)
WalSndKeepaliveIfNecessary();
/*
- * Sleep until something happens or we time out. Also wait for the
- * socket becoming writable, if there's still pending output.
+ * When not in recovery, sleep until something happens or we time out.
+ * Also wait for the socket becoming writable, if there's still pending output.
* Otherwise we might sit on sendable output data while waiting for
* new WAL to be generated. (But if we have nothing to send, we don't
* want to wake on socket-writable.)
*/
- sleeptime = WalSndComputeSleeptime(GetCurrentTimestamp());
+ if (!RecoveryInProgress())
+ {
+ long sleeptime;
+ sleeptime = WalSndComputeSleeptime(GetCurrentTimestamp());
- wakeEvents = WL_SOCKET_READABLE;
+ wakeEvents = WL_SOCKET_READABLE;
- if (pq_is_send_pending())
- wakeEvents |= WL_SOCKET_WRITEABLE;
+ if (pq_is_send_pending())
+ wakeEvents |= WL_SOCKET_WRITEABLE;
- WalSndWait(wakeEvents, sleeptime, WAIT_EVENT_WAL_SENDER_WAIT_WAL);
+ WalSndWait(wakeEvents, sleeptime * 10, WAIT_EVENT_WAL_SENDER_WAIT_WAL);
+ }
+ else
+ /*
+ * We are in the logical decoding on standby case.
+ * We are waiting for the startup process to replay wal record(s) using
+ * a timeout in case we are requested to stop.
+ */
+ {
+ ConditionVariablePrepareToSleep(replayedCV);
+ ConditionVariableTimedSleep(replayedCV, 1000,
+ WAIT_EVENT_WAL_SENDER_WAIT_REPLAY);
+ }
}
/* reactivate latch so WalSndLoop knows to continue */
diff --git a/src/backend/utils/activity/wait_event.c b/src/backend/utils/activity/wait_event.c
index 6e4599278c..38c747b786 100644
--- a/src/backend/utils/activity/wait_event.c
+++ b/src/backend/utils/activity/wait_event.c
@@ -463,6 +463,9 @@ pgstat_get_wait_ipc(WaitEventIPC w)
case WAIT_EVENT_WAL_RECEIVER_WAIT_START:
event_name = "WalReceiverWaitStart";
break;
+ case WAIT_EVENT_WAL_SENDER_WAIT_REPLAY:
+ event_name = "WalReceiverWaitReplay";
+ break;
case WAIT_EVENT_XACT_GROUP_UPDATE:
event_name = "XactGroupUpdate";
break;
diff --git a/src/include/access/xlogrecovery.h b/src/include/access/xlogrecovery.h
index 47c29350f5..b65c2cf1f0 100644
--- a/src/include/access/xlogrecovery.h
+++ b/src/include/access/xlogrecovery.h
@@ -15,6 +15,7 @@
#include "catalog/pg_control.h"
#include "lib/stringinfo.h"
#include "utils/timestamp.h"
+#include "storage/condition_variable.h"
/*
* Recovery target type.
@@ -155,4 +156,6 @@ extern void RecoveryRequiresIntParameter(const char *param_name, int currValue,
extern void xlog_outdesc(StringInfo buf, XLogReaderState *record);
+extern ConditionVariable *getReplayedCV(void);
+
#endif /* XLOGRECOVERY_H */
diff --git a/src/include/replication/walsender.h b/src/include/replication/walsender.h
index 52bb3e2aae..2fd745fe72 100644
--- a/src/include/replication/walsender.h
+++ b/src/include/replication/walsender.h
@@ -13,6 +13,7 @@
#define _WALSENDER_H
#include <signal.h>
+#include "storage/condition_variable.h"
/*
* What to do with a snapshot in create replication slot command.
diff --git a/src/include/utils/wait_event.h b/src/include/utils/wait_event.h
index 6cacd6edaf..04a37feee4 100644
--- a/src/include/utils/wait_event.h
+++ b/src/include/utils/wait_event.h
@@ -130,6 +130,7 @@ typedef enum
WAIT_EVENT_SYNC_REP,
WAIT_EVENT_WAL_RECEIVER_EXIT,
WAIT_EVENT_WAL_RECEIVER_WAIT_START,
+ WAIT_EVENT_WAL_SENDER_WAIT_REPLAY,
WAIT_EVENT_XACT_GROUP_UPDATE
} WaitEventIPC;
--
2.34.1
[text/plain] v45-0003-Allow-logical-decoding-on-standby.patch (11.8K, ../[email protected]/5-v45-0003-Allow-logical-decoding-on-standby.patch)
download | inline diff:
From d184a98f9c9aeefdcfbeeebab3aba4bf0cee9815 Mon Sep 17 00:00:00 2001
From: bdrouvotAWS <[email protected]>
Date: Mon, 30 Jan 2023 15:41:38 +0000
Subject: [PATCH v45 3/6] Allow logical decoding on standby.
Allow a logical slot to be created on standby. Restrict its usage
or its creation if wal_level on primary is less than logical.
During slot creation, it's restart_lsn is set to the last replayed
LSN. Effectively, a logical slot creation on standby waits for an
xl_running_xact record to arrive from primary.
Author: Andres Freund (in an older version), Amit Khandekar, Bertrand Drouvot
Reviewed-By: Bertrand Drouvot, Andres Freund, Robert Haas, Fabrizio de Royes Mello
---
src/backend/access/transam/xlog.c | 11 +++++
src/backend/replication/logical/decode.c | 22 ++++++++-
src/backend/replication/logical/logical.c | 37 ++++++++-------
src/backend/replication/slot.c | 57 ++++++++++++-----------
src/backend/replication/walsender.c | 41 ++++++++++------
src/include/access/xlog.h | 1 +
6 files changed, 111 insertions(+), 58 deletions(-)
4.7% src/backend/access/transam/
38.7% src/backend/replication/logical/
55.6% src/backend/replication/
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 867675d5a1..1abe747cb5 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -4465,6 +4465,17 @@ LocalProcessControlFile(bool reset)
ReadControlFile();
}
+/*
+ * Get the wal_level from the control file. For a standby, this value should be
+ * considered as its active wal_level, because it may be different from what
+ * was originally configured on standby.
+ */
+WalLevel
+GetActiveWalLevelOnStandby(void)
+{
+ return ControlFile->wal_level;
+}
+
/*
* Initialization of shared memory for XLOG
*/
diff --git a/src/backend/replication/logical/decode.c b/src/backend/replication/logical/decode.c
index a53e23c679..6b66a971ba 100644
--- a/src/backend/replication/logical/decode.c
+++ b/src/backend/replication/logical/decode.c
@@ -152,11 +152,31 @@ xlog_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
* can restart from there.
*/
break;
+ case XLOG_PARAMETER_CHANGE:
+ {
+ xl_parameter_change *xlrec =
+ (xl_parameter_change *) XLogRecGetData(buf->record);
+
+ /*
+ * If wal_level on primary is reduced to less than logical, then we
+ * want to prevent existing logical slots from being used.
+ * Existing logical slots on standby get invalidated when this WAL
+ * record is replayed; and further, slot creation fails when the
+ * wal level is not sufficient; but all these operations are not
+ * synchronized, so a logical slot may creep in while the wal_level
+ * is being reduced. Hence this extra check.
+ */
+ if (xlrec->wal_level < WAL_LEVEL_LOGICAL)
+ ereport(ERROR,
+ (errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
+ errmsg("logical decoding on standby requires wal_level "
+ "to be at least logical on the primary server")));
+ break;
+ }
case XLOG_NOOP:
case XLOG_NEXTOID:
case XLOG_SWITCH:
case XLOG_BACKUP_END:
- case XLOG_PARAMETER_CHANGE:
case XLOG_RESTORE_POINT:
case XLOG_FPW_CHANGE:
case XLOG_FPI_FOR_HINT:
diff --git a/src/backend/replication/logical/logical.c b/src/backend/replication/logical/logical.c
index 1a58dd7649..91acc0c155 100644
--- a/src/backend/replication/logical/logical.c
+++ b/src/backend/replication/logical/logical.c
@@ -119,23 +119,22 @@ CheckLogicalDecodingRequirements(void)
(errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
errmsg("logical decoding requires a database connection")));
- /* ----
- * TODO: We got to change that someday soon...
- *
- * There's basically three things missing to allow this:
- * 1) We need to be able to correctly and quickly identify the timeline a
- * LSN belongs to
- * 2) We need to force hot_standby_feedback to be enabled at all times so
- * the primary cannot remove rows we need.
- * 3) support dropping replication slots referring to a database, in
- * dbase_redo. There can't be any active ones due to HS recovery
- * conflicts, so that should be relatively easy.
- * ----
- */
if (RecoveryInProgress())
- ereport(ERROR,
- (errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
- errmsg("logical decoding cannot be used while in recovery")));
+ {
+ /*
+ * This check may have race conditions, but whenever
+ * XLOG_PARAMETER_CHANGE indicates that wal_level has changed, we
+ * verify that there are no existing logical replication slots. And to
+ * avoid races around creating a new slot,
+ * CheckLogicalDecodingRequirements() is called once before creating
+ * the slot, and once when logical decoding is initially starting up.
+ */
+ if (GetActiveWalLevelOnStandby() < WAL_LEVEL_LOGICAL)
+ ereport(ERROR,
+ (errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
+ errmsg("logical decoding on standby requires wal_level "
+ "to be at least logical on the primary server")));
+ }
}
/*
@@ -331,6 +330,12 @@ CreateInitDecodingContext(const char *plugin,
LogicalDecodingContext *ctx;
MemoryContext old_context;
+ /*
+ * On standby, this check is also required while creating the slot. Check
+ * the comments in this function.
+ */
+ CheckLogicalDecodingRequirements();
+
/* shorter lines... */
slot = MyReplicationSlot;
diff --git a/src/backend/replication/slot.c b/src/backend/replication/slot.c
index 38c6f18886..290d4b45f4 100644
--- a/src/backend/replication/slot.c
+++ b/src/backend/replication/slot.c
@@ -51,6 +51,7 @@
#include "storage/proc.h"
#include "storage/procarray.h"
#include "utils/builtins.h"
+#include "access/xlogrecovery.h"
/*
* Replication slot on-disk data structure.
@@ -1177,37 +1178,28 @@ ReplicationSlotReserveWal(void)
/*
* For logical slots log a standby snapshot and start logical decoding
* at exactly that position. That allows the slot to start up more
- * quickly.
+ * quickly. But on a standby we cannot do WAL writes, so just use the
+ * replay pointer; effectively, an attempt to create a logical slot on
+ * standby will cause it to wait for an xl_running_xact record to be
+ * logged independently on the primary, so that a snapshot can be built
+ * using the record.
*
- * That's not needed (or indeed helpful) for physical slots as they'll
- * start replay at the last logged checkpoint anyway. Instead return
- * the location of the last redo LSN. While that slightly increases
- * the chance that we have to retry, it's where a base backup has to
- * start replay at.
+ * None of this is needed (or indeed helpful) for physical slots as
+ * they'll start replay at the last logged checkpoint anyway. Instead
+ * return the location of the last redo LSN. While that slightly
+ * increases the chance that we have to retry, it's where a base backup
+ * has to start replay at.
*/
- if (!RecoveryInProgress() && SlotIsLogical(slot))
- {
- XLogRecPtr flushptr;
-
- /* start at current insert position */
+ if (SlotIsPhysical(slot))
+ restart_lsn = GetRedoRecPtr();
+ else if (RecoveryInProgress())
+ restart_lsn = GetXLogReplayRecPtr(NULL);
+ else
restart_lsn = GetXLogInsertRecPtr();
- SpinLockAcquire(&slot->mutex);
- slot->data.restart_lsn = restart_lsn;
- SpinLockRelease(&slot->mutex);
-
- /* make sure we have enough information to start */
- flushptr = LogStandbySnapshot();
- /* and make sure it's fsynced to disk */
- XLogFlush(flushptr);
- }
- else
- {
- restart_lsn = GetRedoRecPtr();
- SpinLockAcquire(&slot->mutex);
- slot->data.restart_lsn = restart_lsn;
- SpinLockRelease(&slot->mutex);
- }
+ SpinLockAcquire(&slot->mutex);
+ slot->data.restart_lsn = restart_lsn;
+ SpinLockRelease(&slot->mutex);
/* prevent WAL removal as fast as possible */
ReplicationSlotsComputeRequiredLSN();
@@ -1223,6 +1215,17 @@ ReplicationSlotReserveWal(void)
if (XLogGetLastRemovedSegno() < segno)
break;
}
+
+ if (!RecoveryInProgress() && SlotIsLogical(slot))
+ {
+ XLogRecPtr flushptr;
+
+ /* make sure we have enough information to start */
+ flushptr = LogStandbySnapshot();
+
+ /* and make sure it's fsynced to disk */
+ XLogFlush(flushptr);
+ }
}
/*
diff --git a/src/backend/replication/walsender.c b/src/backend/replication/walsender.c
index 8885cdeebc..1e91cbc564 100644
--- a/src/backend/replication/walsender.c
+++ b/src/backend/replication/walsender.c
@@ -906,23 +906,31 @@ logical_read_xlog_page(XLogReaderState *state, XLogRecPtr targetPagePtr, int req
int count;
WALReadError errinfo;
XLogSegNo segno;
- TimeLineID currTLI = GetWALInsertionTimeLine();
+ TimeLineID currTLI;
/*
- * Since logical decoding is only permitted on a primary server, we know
- * that the current timeline ID can't be changing any more. If we did this
- * on a standby, we'd have to worry about the values we compute here
- * becoming invalid due to a promotion or timeline change.
+ * Since logical decoding is also permitted on a standby server, we need
+ * to check if the server is in recovery to decide how to get the current
+ * timeline ID (so that it also cover the promotion or timeline change cases).
*/
+
+ /* make sure we have enough WAL available */
+ flushptr = WalSndWaitForWal(targetPagePtr + reqLen);
+
+ /* the standby could have been promoted, so check if still in recovery */
+ am_cascading_walsender = RecoveryInProgress();
+
+ if (am_cascading_walsender)
+ GetXLogReplayRecPtr(&currTLI);
+ else
+ currTLI = GetWALInsertionTimeLine();
+
XLogReadDetermineTimeline(state, targetPagePtr, reqLen, currTLI);
sendTimeLineIsHistoric = (state->currTLI != currTLI);
sendTimeLine = state->currTLI;
sendTimeLineValidUpto = state->currTLIValidUntil;
sendTimeLineNextTLI = state->nextTLI;
- /* make sure we have enough WAL available */
- flushptr = WalSndWaitForWal(targetPagePtr + reqLen);
-
/* fail if not (implies we are going to shut down) */
if (flushptr < targetPagePtr + reqLen)
return -1;
@@ -937,7 +945,7 @@ logical_read_xlog_page(XLogReaderState *state, XLogRecPtr targetPagePtr, int req
cur_page,
targetPagePtr,
XLOG_BLCKSZ,
- state->seg.ws_tli, /* Pass the current TLI because only
+ currTLI, /* Pass the current TLI because only
* WalSndSegmentOpen controls whether new
* TLI is needed. */
&errinfo))
@@ -3074,10 +3082,14 @@ XLogSendLogical(void)
* If first time through in this session, initialize flushPtr. Otherwise,
* we only need to update flushPtr if EndRecPtr is past it.
*/
- if (flushPtr == InvalidXLogRecPtr)
- flushPtr = GetFlushRecPtr(NULL);
- else if (logical_decoding_ctx->reader->EndRecPtr >= flushPtr)
- flushPtr = GetFlushRecPtr(NULL);
+ if (flushPtr == InvalidXLogRecPtr ||
+ logical_decoding_ctx->reader->EndRecPtr >= flushPtr)
+ {
+ if (am_cascading_walsender)
+ flushPtr = GetStandbyFlushRecPtr(NULL);
+ else
+ flushPtr = GetFlushRecPtr(NULL);
+ }
/* If EndRecPtr is still past our flushPtr, it means we caught up. */
if (logical_decoding_ctx->reader->EndRecPtr >= flushPtr)
@@ -3168,7 +3180,8 @@ GetStandbyFlushRecPtr(TimeLineID *tli)
receivePtr = GetWalRcvFlushRecPtr(NULL, &receiveTLI);
replayPtr = GetXLogReplayRecPtr(&replayTLI);
- *tli = replayTLI;
+ if (tli)
+ *tli = replayTLI;
result = replayPtr;
if (receiveTLI == replayTLI && receivePtr > replayPtr)
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index cfe5409738..48ca852381 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -230,6 +230,7 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(void);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
extern void CreateCheckPoint(int flags);
--
2.34.1
[text/plain] v45-0002-Handle-logical-slot-conflicts-on-standby.patch (37.0K, ../[email protected]/6-v45-0002-Handle-logical-slot-conflicts-on-standby.patch)
download | inline diff:
From 700c26ef8ee1c5caa40c6e0ce14796f7662c6f97 Mon Sep 17 00:00:00 2001
From: bdrouvotAWS <[email protected]>
Date: Mon, 30 Jan 2023 15:40:56 +0000
Subject: [PATCH v45 2/6] Handle logical slot conflicts on standby.
During WAL replay on standby, when slot conflict is identified,
invalidate such slots. Also do the same thing if wal_level on the primary server
is reduced to below logical and there are existing logical slots
on standby. Introduce a new ProcSignalReason value for slot
conflict recovery. Arrange for a new pg_stat_database_conflicts field:
confl_active_logicalslot.
Add a new field "conflicting" in pg_replication_slots.
Author: Andres Freund (in an older version), Amit Khandekar, Bertrand Drouvot
Reviewed-By: Bertrand Drouvot, Andres Freund, Robert Haas, Fabrizio de Royes Mello,
Bharath Rupireddy
---
doc/src/sgml/monitoring.sgml | 11 +
doc/src/sgml/system-views.sgml | 10 +
src/backend/access/gist/gistxlog.c | 2 +
src/backend/access/hash/hash_xlog.c | 1 +
src/backend/access/heap/heapam.c | 3 +
src/backend/access/nbtree/nbtxlog.c | 2 +
src/backend/access/spgist/spgxlog.c | 1 +
src/backend/access/transam/xlog.c | 24 ++-
src/backend/catalog/system_views.sql | 6 +-
.../replication/logical/logicalfuncs.c | 13 +-
src/backend/replication/slot.c | 198 +++++++++++++-----
src/backend/replication/slotfuncs.c | 13 +-
src/backend/replication/walsender.c | 8 +
src/backend/storage/ipc/procsignal.c | 3 +
src/backend/storage/ipc/standby.c | 13 +-
src/backend/tcop/postgres.c | 24 +++
src/backend/utils/activity/pgstat_database.c | 4 +
src/backend/utils/adt/pgstatfuncs.c | 3 +
src/include/catalog/pg_proc.dat | 11 +-
src/include/pgstat.h | 1 +
src/include/replication/slot.h | 5 +-
src/include/storage/procsignal.h | 1 +
src/include/storage/standby.h | 2 +
src/test/regress/expected/rules.out | 8 +-
24 files changed, 304 insertions(+), 63 deletions(-)
5.4% doc/src/sgml/
7.2% src/backend/access/transam/
4.7% src/backend/replication/logical/
56.8% src/backend/replication/
4.5% src/backend/storage/ipc/
6.5% src/backend/tcop/
5.4% src/backend/
3.9% src/include/catalog/
3.0% src/include/replication/
diff --git a/doc/src/sgml/monitoring.sgml b/doc/src/sgml/monitoring.sgml
index 1756f1a4b6..e25f71a776 100644
--- a/doc/src/sgml/monitoring.sgml
+++ b/doc/src/sgml/monitoring.sgml
@@ -4365,6 +4365,17 @@ SELECT pid, wait_event_type, wait_event FROM pg_stat_activity WHERE wait_event i
deadlocks
</para></entry>
</row>
+
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ <structfield>confl_active_logicalslot</structfield> <type>bigint</type>
+ </para>
+ <para>
+ Number of active logical slots in this database that have been
+ invalidated because they conflict with recovery (note that inactive ones
+ are also invalidated but do not increment this counter)
+ </para></entry>
+ </row>
</tbody>
</tgroup>
</table>
diff --git a/doc/src/sgml/system-views.sgml b/doc/src/sgml/system-views.sgml
index 7c8fc3f654..239f713295 100644
--- a/doc/src/sgml/system-views.sgml
+++ b/doc/src/sgml/system-views.sgml
@@ -2516,6 +2516,16 @@ SELECT * FROM pg_locks pl LEFT JOIN pg_prepared_xacts ppx
false for physical slots.
</para></entry>
</row>
+
+ <row>
+ <entry role="catalog_table_entry"><para role="column_definition">
+ <structfield>conflicting</structfield> <type>bool</type>
+ </para>
+ <para>
+ True if this logical slot conflicted with recovery (and so is now
+ invalidated). Always NULL for physical slots.
+ </para></entry>
+ </row>
</tbody>
</tgroup>
</table>
diff --git a/src/backend/access/gist/gistxlog.c b/src/backend/access/gist/gistxlog.c
index b7678f3c14..9a86fb3fef 100644
--- a/src/backend/access/gist/gistxlog.c
+++ b/src/backend/access/gist/gistxlog.c
@@ -197,6 +197,7 @@ gistRedoDeleteRecord(XLogReaderState *record)
XLogRecGetBlockTag(record, 0, &rlocator, NULL, NULL);
ResolveRecoveryConflictWithSnapshot(xldata->snapshotConflictHorizon,
+ xldata->isCatalogRel,
rlocator);
}
@@ -390,6 +391,7 @@ gistRedoPageReuse(XLogReaderState *record)
*/
if (InHotStandby)
ResolveRecoveryConflictWithSnapshotFullXid(xlrec->snapshotConflictHorizon,
+ xlrec->isCatalogRel,
xlrec->locator);
}
diff --git a/src/backend/access/hash/hash_xlog.c b/src/backend/access/hash/hash_xlog.c
index 08ceb91288..b856304746 100644
--- a/src/backend/access/hash/hash_xlog.c
+++ b/src/backend/access/hash/hash_xlog.c
@@ -1003,6 +1003,7 @@ hash_xlog_vacuum_one_page(XLogReaderState *record)
XLogRecGetBlockTag(record, 0, &rlocator, NULL, NULL);
ResolveRecoveryConflictWithSnapshot(xldata->snapshotConflictHorizon,
+ xldata->isCatalogRel,
rlocator);
}
diff --git a/src/backend/access/heap/heapam.c b/src/backend/access/heap/heapam.c
index d478724b9d..d64fb4cc84 100644
--- a/src/backend/access/heap/heapam.c
+++ b/src/backend/access/heap/heapam.c
@@ -8891,6 +8891,7 @@ heap_xlog_prune(XLogReaderState *record)
*/
if (InHotStandby)
ResolveRecoveryConflictWithSnapshot(xlrec->snapshotConflictHorizon,
+ xlrec->isCatalogRel,
rlocator);
/*
@@ -9060,6 +9061,7 @@ heap_xlog_visible(XLogReaderState *record)
*/
if (InHotStandby)
ResolveRecoveryConflictWithSnapshot(xlrec->snapshotConflictHorizon,
+ xlrec->flags & VISIBILITYMAP_IS_CATALOG_REL,
rlocator);
/*
@@ -9177,6 +9179,7 @@ heap_xlog_freeze_page(XLogReaderState *record)
XLogRecGetBlockTag(record, 0, &rlocator, NULL, NULL);
ResolveRecoveryConflictWithSnapshot(xlrec->snapshotConflictHorizon,
+ xlrec->isCatalogRel,
rlocator);
}
diff --git a/src/backend/access/nbtree/nbtxlog.c b/src/backend/access/nbtree/nbtxlog.c
index 414ca4f6de..c87e46ed66 100644
--- a/src/backend/access/nbtree/nbtxlog.c
+++ b/src/backend/access/nbtree/nbtxlog.c
@@ -669,6 +669,7 @@ btree_xlog_delete(XLogReaderState *record)
XLogRecGetBlockTag(record, 0, &rlocator, NULL, NULL);
ResolveRecoveryConflictWithSnapshot(xlrec->snapshotConflictHorizon,
+ xlrec->isCatalogRel,
rlocator);
}
@@ -1007,6 +1008,7 @@ btree_xlog_reuse_page(XLogReaderState *record)
if (InHotStandby)
ResolveRecoveryConflictWithSnapshotFullXid(xlrec->snapshotConflictHorizon,
+ xlrec->isCatalogRel,
xlrec->locator);
}
diff --git a/src/backend/access/spgist/spgxlog.c b/src/backend/access/spgist/spgxlog.c
index b071b59c8a..459ac929ba 100644
--- a/src/backend/access/spgist/spgxlog.c
+++ b/src/backend/access/spgist/spgxlog.c
@@ -879,6 +879,7 @@ spgRedoVacuumRedirect(XLogReaderState *record)
XLogRecGetBlockTag(record, 0, &locator, NULL, NULL);
ResolveRecoveryConflictWithSnapshot(xldata->snapshotConflictHorizon,
+ xldata->isCatalogRel,
locator);
}
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index fb4c860bde..867675d5a1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -6447,6 +6447,7 @@ CreateCheckPoint(int flags)
VirtualTransactionId *vxids;
int nvxids;
int oldXLogAllowed = 0;
+ bool invalidated = false;
/*
* An end-of-recovery checkpoint is really a shutdown checkpoint, just
@@ -6807,7 +6808,8 @@ CreateCheckPoint(int flags)
*/
XLByteToSeg(RedoRecPtr, _logSegNo, wal_segment_size);
KeepLogSeg(recptr, &_logSegNo);
- if (InvalidateObsoleteReplicationSlots(_logSegNo))
+ InvalidateObsoleteReplicationSlots(_logSegNo, &invalidated, InvalidOid, NULL);
+ if (invalidated)
{
/*
* Some slots have been invalidated; recalculate the old-segment
@@ -7086,6 +7088,7 @@ CreateRestartPoint(int flags)
XLogRecPtr endptr;
XLogSegNo _logSegNo;
TimestampTz xtime;
+ bool invalidated = false;
/* Concurrent checkpoint/restartpoint cannot happen */
Assert(!IsUnderPostmaster || MyBackendType == B_CHECKPOINTER);
@@ -7251,7 +7254,8 @@ CreateRestartPoint(int flags)
replayPtr = GetXLogReplayRecPtr(&replayTLI);
endptr = (receivePtr < replayPtr) ? replayPtr : receivePtr;
KeepLogSeg(endptr, &_logSegNo);
- if (InvalidateObsoleteReplicationSlots(_logSegNo))
+ InvalidateObsoleteReplicationSlots(_logSegNo, &invalidated, InvalidOid, NULL);
+ if (invalidated)
{
/*
* Some slots have been invalidated; recalculate the old-segment
@@ -7966,6 +7970,22 @@ xlog_redo(XLogReaderState *record)
/* Update our copy of the parameters in pg_control */
memcpy(&xlrec, XLogRecGetData(record), sizeof(xl_parameter_change));
+ /*
+ * Invalidate logical slots if we are in hot standby and the primary does not
+ * have a WAL level sufficient for logical decoding. No need to search
+ * for potentially conflicting logically slots if standby is running
+ * with wal_level lower than logical, because in that case, we would
+ * have either disallowed creation of logical slots or invalidated existing
+ * ones.
+ */
+ if (InRecovery && InHotStandby &&
+ xlrec.wal_level < WAL_LEVEL_LOGICAL &&
+ wal_level >= WAL_LEVEL_LOGICAL)
+ {
+ TransactionId ConflictHorizon = InvalidTransactionId;
+ InvalidateObsoleteReplicationSlots(InvalidXLogRecPtr, NULL, InvalidOid, &ConflictHorizon);
+ }
+
LWLockAcquire(ControlFileLock, LW_EXCLUSIVE);
ControlFile->MaxConnections = xlrec.MaxConnections;
ControlFile->max_worker_processes = xlrec.max_worker_processes;
diff --git a/src/backend/catalog/system_views.sql b/src/backend/catalog/system_views.sql
index 8608e3fa5b..a272bd4a88 100644
--- a/src/backend/catalog/system_views.sql
+++ b/src/backend/catalog/system_views.sql
@@ -997,7 +997,8 @@ CREATE VIEW pg_replication_slots AS
L.confirmed_flush_lsn,
L.wal_status,
L.safe_wal_size,
- L.two_phase
+ L.two_phase,
+ L.conflicting
FROM pg_get_replication_slots() AS L
LEFT JOIN pg_database D ON (L.datoid = D.oid);
@@ -1065,7 +1066,8 @@ CREATE VIEW pg_stat_database_conflicts AS
pg_stat_get_db_conflict_lock(D.oid) AS confl_lock,
pg_stat_get_db_conflict_snapshot(D.oid) AS confl_snapshot,
pg_stat_get_db_conflict_bufferpin(D.oid) AS confl_bufferpin,
- pg_stat_get_db_conflict_startup_deadlock(D.oid) AS confl_deadlock
+ pg_stat_get_db_conflict_startup_deadlock(D.oid) AS confl_deadlock,
+ pg_stat_get_db_conflict_logicalslot(D.oid) AS confl_active_logicalslot
FROM pg_database D;
CREATE VIEW pg_stat_user_functions AS
diff --git a/src/backend/replication/logical/logicalfuncs.c b/src/backend/replication/logical/logicalfuncs.c
index fa1b641a2b..070fd378e8 100644
--- a/src/backend/replication/logical/logicalfuncs.c
+++ b/src/backend/replication/logical/logicalfuncs.c
@@ -216,9 +216,9 @@ pg_logical_slot_get_changes_guts(FunctionCallInfo fcinfo, bool confirm, bool bin
/*
* After the sanity checks in CreateDecodingContext, make sure the
- * restart_lsn is valid. Avoid "cannot get changes" wording in this
- * errmsg because that'd be confusingly ambiguous about no changes
- * being available.
+ * restart_lsn is valid or both xmin and catalog_xmin are valid. Avoid
+ * "cannot get changes" wording in this errmsg because that'd be
+ * confusingly ambiguous about no changes being available.
*/
if (XLogRecPtrIsInvalid(MyReplicationSlot->data.restart_lsn))
ereport(ERROR,
@@ -227,6 +227,13 @@ pg_logical_slot_get_changes_guts(FunctionCallInfo fcinfo, bool confirm, bool bin
NameStr(*name)),
errdetail("This slot has never previously reserved WAL, or it has been invalidated.")));
+ if (LogicalReplicationSlotIsInvalid(MyReplicationSlot))
+ ereport(ERROR,
+ (errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
+ errmsg("cannot read from logical replication slot \"%s\"",
+ NameStr(*name)),
+ errdetail("This slot has been invalidated because it was conflicting with recovery.")));
+
MemoryContextSwitchTo(oldcontext);
/*
diff --git a/src/backend/replication/slot.c b/src/backend/replication/slot.c
index f286918f69..38c6f18886 100644
--- a/src/backend/replication/slot.c
+++ b/src/backend/replication/slot.c
@@ -855,8 +855,10 @@ ReplicationSlotsComputeRequiredXmin(bool already_locked)
SpinLockAcquire(&s->mutex);
effective_xmin = s->effective_xmin;
effective_catalog_xmin = s->effective_catalog_xmin;
- invalidated = (!XLogRecPtrIsInvalid(s->data.invalidated_at) &&
- XLogRecPtrIsInvalid(s->data.restart_lsn));
+ invalidated = ((!XLogRecPtrIsInvalid(s->data.invalidated_at) &&
+ XLogRecPtrIsInvalid(s->data.restart_lsn))
+ || (!TransactionIdIsValid(s->data.xmin) &&
+ !TransactionIdIsValid(s->data.catalog_xmin)));
SpinLockRelease(&s->mutex);
/* invalidated slots need not apply */
@@ -1224,20 +1226,21 @@ ReplicationSlotReserveWal(void)
}
/*
- * Helper for InvalidateObsoleteReplicationSlots -- acquires the given slot
- * and mark it invalid, if necessary and possible.
+ * Helper for InvalidateObsoleteReplicationSlots
+ *
+ * Acquires the given slot and mark it invalid, if necessary and possible.
*
* Returns whether ReplicationSlotControlLock was released in the interim (and
* in that case we're not holding the lock at return, otherwise we are).
*
- * Sets *invalidated true if the slot was invalidated. (Untouched otherwise.)
+ * Sets *invalidated true if an obsolete slot was invalidated. (Untouched otherwise.)
*
* This is inherently racy, because we release the LWLock
* for syscalls, so caller must restart if we return true.
*/
static bool
-InvalidatePossiblyObsoleteSlot(ReplicationSlot *s, XLogRecPtr oldestLSN,
- bool *invalidated)
+InvalidatePossiblyObsoleteOrConflictingLogicalSlot(ReplicationSlot *s, XLogRecPtr oldestLSN,
+ bool *invalidated, TransactionId *xid)
{
int last_signaled_pid = 0;
bool released_lock = false;
@@ -1245,6 +1248,9 @@ InvalidatePossiblyObsoleteSlot(ReplicationSlot *s, XLogRecPtr oldestLSN,
for (;;)
{
XLogRecPtr restart_lsn;
+ TransactionId slot_xmin;
+ TransactionId slot_catalog_xmin;
+
NameData slotname;
int active_pid = 0;
@@ -1261,18 +1267,33 @@ InvalidatePossiblyObsoleteSlot(ReplicationSlot *s, XLogRecPtr oldestLSN,
* Check if the slot needs to be invalidated. If it needs to be
* invalidated, and is not currently acquired, acquire it and mark it
* as having been invalidated. We do this with the spinlock held to
- * avoid race conditions -- for example the restart_lsn could move
- * forward, or the slot could be dropped.
+ * avoid race conditions -- for example the restart_lsn (or the
+ * xmin(s) could) move forward or the slot could be dropped.
*/
SpinLockAcquire(&s->mutex);
restart_lsn = s->data.restart_lsn;
+ slot_xmin = s->data.xmin;
+ slot_catalog_xmin = s->data.catalog_xmin;
+
+ /* slot has been invalidated (logical decoding conflict case) */
+ if ((xid &&
+ ((LogicalReplicationSlotIsInvalid(s))
+ ||
/*
- * If the slot is already invalid or is fresh enough, we don't need to
- * do anything.
+ * We are not forcing for invalidation because the xid is valid and
+ * this is a non conflicting slot.
*/
- if (XLogRecPtrIsInvalid(restart_lsn) || restart_lsn >= oldestLSN)
+ (TransactionIdIsValid(*xid) && !(
+ (TransactionIdIsValid(slot_xmin) && TransactionIdPrecedesOrEquals(slot_xmin, *xid))
+ ||
+ (TransactionIdIsValid(slot_catalog_xmin) && TransactionIdPrecedesOrEquals(slot_catalog_xmin, *xid))
+ ))
+ ))
+ ||
+ /* slot has been invalidated (obsolete LSN case) */
+ (!xid && (XLogRecPtrIsInvalid(restart_lsn) || restart_lsn >= oldestLSN)))
{
SpinLockRelease(&s->mutex);
if (released_lock)
@@ -1292,9 +1313,16 @@ InvalidatePossiblyObsoleteSlot(ReplicationSlot *s, XLogRecPtr oldestLSN,
{
MyReplicationSlot = s;
s->active_pid = MyProcPid;
- s->data.invalidated_at = restart_lsn;
- s->data.restart_lsn = InvalidXLogRecPtr;
-
+ if (xid)
+ {
+ s->data.xmin = InvalidTransactionId;
+ s->data.catalog_xmin = InvalidTransactionId;
+ }
+ else
+ {
+ s->data.invalidated_at = restart_lsn;
+ s->data.restart_lsn = InvalidXLogRecPtr;
+ }
/* Let caller know */
*invalidated = true;
}
@@ -1327,15 +1355,39 @@ InvalidatePossiblyObsoleteSlot(ReplicationSlot *s, XLogRecPtr oldestLSN,
*/
if (last_signaled_pid != active_pid)
{
- ereport(LOG,
- errmsg("terminating process %d to release replication slot \"%s\"",
- active_pid, NameStr(slotname)),
- errdetail("The slot's restart_lsn %X/%X exceeds the limit by %llu bytes.",
- LSN_FORMAT_ARGS(restart_lsn),
- (unsigned long long) (oldestLSN - restart_lsn)),
- errhint("You might need to increase max_slot_wal_keep_size."));
+ if (xid)
+ {
+ if (TransactionIdIsValid(*xid))
+ {
+ ereport(LOG,
+ errmsg("terminating process %d because replication slot \"%s\" conflicts with recovery",
+ active_pid, NameStr(slotname)),
+ errdetail("The slot conflicted with xid horizon %u.",
+ *xid));
+ }
+ else
+ {
+ ereport(LOG,
+ errmsg("terminating process %d because replication slot \"%s\" conflicts with recovery",
+ active_pid, NameStr(slotname)),
+ errdetail("Logical decoding on standby requires wal_level to be at least logical on the primary server"));
+ }
+
+ (void) SendProcSignal(active_pid, PROCSIG_RECOVERY_CONFLICT_LOGICALSLOT, InvalidBackendId);
+ }
+ else
+ {
+ ereport(LOG,
+ errmsg("terminating process %d to release replication slot \"%s\"",
+ active_pid, NameStr(slotname)),
+ errdetail("The slot's restart_lsn %X/%X exceeds the limit by %llu bytes.",
+ LSN_FORMAT_ARGS(restart_lsn),
+ (unsigned long long) (oldestLSN - restart_lsn)),
+ errhint("You might need to increase max_slot_wal_keep_size."));
+
+ (void) kill(active_pid, SIGTERM);
+ }
- (void) kill(active_pid, SIGTERM);
last_signaled_pid = active_pid;
}
@@ -1369,13 +1421,33 @@ InvalidatePossiblyObsoleteSlot(ReplicationSlot *s, XLogRecPtr oldestLSN,
ReplicationSlotSave();
ReplicationSlotRelease();
- ereport(LOG,
- errmsg("invalidating obsolete replication slot \"%s\"",
- NameStr(slotname)),
- errdetail("The slot's restart_lsn %X/%X exceeds the limit by %llu bytes.",
- LSN_FORMAT_ARGS(restart_lsn),
- (unsigned long long) (oldestLSN - restart_lsn)),
- errhint("You might need to increase max_slot_wal_keep_size."));
+ if (xid)
+ {
+ pgstat_drop_replslot(s);
+
+ if (TransactionIdIsValid(*xid))
+ {
+ ereport(LOG,
+ errmsg("invalidating slot \"%s\" because it conflicts with recovery", NameStr(slotname)),
+ errdetail("The slot conflicted with xid horizon %u.", *xid));
+ }
+ else
+ {
+ ereport(LOG,
+ errmsg("invalidating slot \"%s\" because it conflicts with recovery", NameStr(slotname)),
+ errdetail("Logical decoding on standby requires wal_level to be at least logical on the primary server"));
+ }
+ }
+ else
+ {
+ ereport(LOG,
+ errmsg("invalidating obsolete replication slot \"%s\"",
+ NameStr(slotname)),
+ errdetail("The slot's restart_lsn %X/%X exceeds the limit by %llu bytes.",
+ LSN_FORMAT_ARGS(restart_lsn),
+ (unsigned long long) (oldestLSN - restart_lsn)),
+ errhint("You might need to increase max_slot_wal_keep_size."));
+ }
/* done with this slot for now */
break;
@@ -1388,20 +1460,40 @@ InvalidatePossiblyObsoleteSlot(ReplicationSlot *s, XLogRecPtr oldestLSN,
}
/*
- * Mark any slot that points to an LSN older than the given segment
- * as invalid; it requires WAL that's about to be removed.
+ * Invalidate Obsolete slots or resolve recovery conflicts with logical slots.
*
- * Returns true when any slot have got invalidated.
+ * Obsolete case (aka xid is NULL):
*
- * NB - this runs as part of checkpoint, so avoid raising errors if possible.
+ * Mark any slot that points to an LSN older than the given segment
+ * as invalid; it requires WAL that's about to be removed.
+ * invalidated is set to true when any slot have got invalidated.
+ *
+ * Logical replication slot case:
+ *
+ * When xid is valid, it means that we are about to remove rows older than xid.
+ * Therefore we need to invalidate slots that depend on seeing those rows.
+ * When xid is invalid, invalidate all logical slots. This is required when the
+ * master wal_level is set back to replica, so existing logical slots need to
+ * be invalidated.
*/
-bool
-InvalidateObsoleteReplicationSlots(XLogSegNo oldestSegno)
+void
+InvalidateObsoleteReplicationSlots(XLogSegNo oldestSegno, bool *invalidated, Oid dboid, TransactionId *xid)
{
- XLogRecPtr oldestLSN;
- bool invalidated = false;
- XLogSegNoOffsetToRecPtr(oldestSegno, 0, wal_segment_size, oldestLSN);
+ XLogRecPtr oldestLSN = InvalidXLogRecPtr;
+ bool logical_slot_invalidated = false;
+
+ Assert(max_replication_slots >= 0);
+
+ if (max_replication_slots == 0)
+ return;
+
+ if (!xid)
+ {
+ Assert(invalidated);
+ *invalidated = false;
+ XLogSegNoOffsetToRecPtr(oldestSegno, 0, wal_segment_size, oldestLSN);
+ }
restart:
LWLockAcquire(ReplicationSlotControlLock, LW_SHARED);
@@ -1412,24 +1504,36 @@ restart:
if (!s->in_use)
continue;
- if (InvalidatePossiblyObsoleteSlot(s, oldestLSN, &invalidated))
+ if (xid)
{
- /* if the lock was released, start from scratch */
- goto restart;
+ /* we are only dealing with *logical* slot conflicts */
+ if (!SlotIsLogical(s))
+ continue;
+
+ /*
+ * not the database of interest and we don't want all the
+ * database, skip
+ */
+ if (s->data.database != dboid && TransactionIdIsValid(*xid))
+ continue;
}
+
+ if (InvalidatePossiblyObsoleteOrConflictingLogicalSlot(s, oldestLSN, invalidated ? invalidated : &logical_slot_invalidated, xid))
+ goto restart;
}
+
LWLockRelease(ReplicationSlotControlLock);
/*
- * If any slots have been invalidated, recalculate the resource limits.
+ * If any slots have been invalidated, recalculate the required xmin
+ * and the required lsn (if appropriate).
*/
- if (invalidated)
+ if ((!xid && *invalidated) || (xid && logical_slot_invalidated))
{
ReplicationSlotsComputeRequiredXmin(false);
- ReplicationSlotsComputeRequiredLSN();
+ if (!xid && *invalidated)
+ ReplicationSlotsComputeRequiredLSN();
}
-
- return invalidated;
}
/*
diff --git a/src/backend/replication/slotfuncs.c b/src/backend/replication/slotfuncs.c
index 2f3c964824..44192bc32d 100644
--- a/src/backend/replication/slotfuncs.c
+++ b/src/backend/replication/slotfuncs.c
@@ -232,7 +232,7 @@ pg_drop_replication_slot(PG_FUNCTION_ARGS)
Datum
pg_get_replication_slots(PG_FUNCTION_ARGS)
{
-#define PG_GET_REPLICATION_SLOTS_COLS 14
+#define PG_GET_REPLICATION_SLOTS_COLS 15
ReturnSetInfo *rsinfo = (ReturnSetInfo *) fcinfo->resultinfo;
XLogRecPtr currlsn;
int slotno;
@@ -404,6 +404,17 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
values[i++] = BoolGetDatum(slot_contents.data.two_phase);
+ if (slot_contents.data.database == InvalidOid)
+ nulls[i++] = true;
+ else
+ {
+ if (slot_contents.data.xmin == InvalidTransactionId &&
+ slot_contents.data.catalog_xmin == InvalidTransactionId)
+ values[i++] = BoolGetDatum(true);
+ else
+ values[i++] = BoolGetDatum(false);
+ }
+
Assert(i == PG_GET_REPLICATION_SLOTS_COLS);
tuplestore_putvalues(rsinfo->setResult, rsinfo->setDesc,
diff --git a/src/backend/replication/walsender.c b/src/backend/replication/walsender.c
index 4ed3747e3f..8885cdeebc 100644
--- a/src/backend/replication/walsender.c
+++ b/src/backend/replication/walsender.c
@@ -1253,6 +1253,14 @@ StartLogicalReplication(StartReplicationCmd *cmd)
ReplicationSlotAcquire(cmd->slotname, true);
+ if (!TransactionIdIsValid(MyReplicationSlot->data.xmin)
+ && !TransactionIdIsValid(MyReplicationSlot->data.catalog_xmin))
+ ereport(ERROR,
+ (errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
+ errmsg("cannot read from logical replication slot \"%s\"",
+ cmd->slotname),
+ errdetail("This slot has been invalidated because it was conflicting with recovery.")));
+
if (XLogRecPtrIsInvalid(MyReplicationSlot->data.restart_lsn))
ereport(ERROR,
(errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
diff --git a/src/backend/storage/ipc/procsignal.c b/src/backend/storage/ipc/procsignal.c
index 395b2cf690..c85cb5cc18 100644
--- a/src/backend/storage/ipc/procsignal.c
+++ b/src/backend/storage/ipc/procsignal.c
@@ -673,6 +673,9 @@ procsignal_sigusr1_handler(SIGNAL_ARGS)
if (CheckProcSignal(PROCSIG_RECOVERY_CONFLICT_SNAPSHOT))
RecoveryConflictInterrupt(PROCSIG_RECOVERY_CONFLICT_SNAPSHOT);
+ if (CheckProcSignal(PROCSIG_RECOVERY_CONFLICT_LOGICALSLOT))
+ RecoveryConflictInterrupt(PROCSIG_RECOVERY_CONFLICT_LOGICALSLOT);
+
if (CheckProcSignal(PROCSIG_RECOVERY_CONFLICT_STARTUP_DEADLOCK))
RecoveryConflictInterrupt(PROCSIG_RECOVERY_CONFLICT_STARTUP_DEADLOCK);
diff --git a/src/backend/storage/ipc/standby.c b/src/backend/storage/ipc/standby.c
index 94cc860f5f..ec817381a1 100644
--- a/src/backend/storage/ipc/standby.c
+++ b/src/backend/storage/ipc/standby.c
@@ -35,6 +35,7 @@
#include "utils/ps_status.h"
#include "utils/timeout.h"
#include "utils/timestamp.h"
+#include "replication/slot.h"
/* User-settable GUC parameters */
int vacuum_defer_cleanup_age;
@@ -475,6 +476,7 @@ ResolveRecoveryConflictWithVirtualXIDs(VirtualTransactionId *waitlist,
*/
void
ResolveRecoveryConflictWithSnapshot(TransactionId snapshotConflictHorizon,
+ bool isCatalogRel,
RelFileLocator locator)
{
VirtualTransactionId *backends;
@@ -500,6 +502,9 @@ ResolveRecoveryConflictWithSnapshot(TransactionId snapshotConflictHorizon,
PROCSIG_RECOVERY_CONFLICT_SNAPSHOT,
WAIT_EVENT_RECOVERY_CONFLICT_SNAPSHOT,
true);
+
+ if (wal_level >= WAL_LEVEL_LOGICAL && isCatalogRel)
+ InvalidateObsoleteReplicationSlots(InvalidXLogRecPtr, NULL, locator.dbOid, &snapshotConflictHorizon);
}
/*
@@ -508,6 +513,7 @@ ResolveRecoveryConflictWithSnapshot(TransactionId snapshotConflictHorizon,
*/
void
ResolveRecoveryConflictWithSnapshotFullXid(FullTransactionId snapshotConflictHorizon,
+ bool isCatalogRel,
RelFileLocator locator)
{
/*
@@ -526,7 +532,9 @@ ResolveRecoveryConflictWithSnapshotFullXid(FullTransactionId snapshotConflictHor
TransactionId truncated;
truncated = XidFromFullTransactionId(snapshotConflictHorizon);
- ResolveRecoveryConflictWithSnapshot(truncated, locator);
+ ResolveRecoveryConflictWithSnapshot(truncated,
+ isCatalogRel,
+ locator);
}
}
@@ -1487,6 +1495,9 @@ get_recovery_conflict_desc(ProcSignalReason reason)
case PROCSIG_RECOVERY_CONFLICT_SNAPSHOT:
reasonDesc = _("recovery conflict on snapshot");
break;
+ case PROCSIG_RECOVERY_CONFLICT_LOGICALSLOT:
+ reasonDesc = _("recovery conflict on replication slot");
+ break;
case PROCSIG_RECOVERY_CONFLICT_STARTUP_DEADLOCK:
reasonDesc = _("recovery conflict on buffer deadlock");
break;
diff --git a/src/backend/tcop/postgres.c b/src/backend/tcop/postgres.c
index 470b734e9e..0041896620 100644
--- a/src/backend/tcop/postgres.c
+++ b/src/backend/tcop/postgres.c
@@ -2481,6 +2481,9 @@ errdetail_recovery_conflict(void)
case PROCSIG_RECOVERY_CONFLICT_SNAPSHOT:
errdetail("User query might have needed to see row versions that must be removed.");
break;
+ case PROCSIG_RECOVERY_CONFLICT_LOGICALSLOT:
+ errdetail("User was using the logical slot that must be dropped.");
+ break;
case PROCSIG_RECOVERY_CONFLICT_STARTUP_DEADLOCK:
errdetail("User transaction caused buffer deadlock with recovery.");
break;
@@ -3050,6 +3053,27 @@ RecoveryConflictInterrupt(ProcSignalReason reason)
case PROCSIG_RECOVERY_CONFLICT_LOCK:
case PROCSIG_RECOVERY_CONFLICT_TABLESPACE:
case PROCSIG_RECOVERY_CONFLICT_SNAPSHOT:
+ case PROCSIG_RECOVERY_CONFLICT_LOGICALSLOT:
+
+ /*
+ * For conflicts that require a logical slot to be
+ * invalidated, the requirement is for the signal receiver to
+ * release the slot, so that it could be invalidated by the
+ * signal sender. So for normal backends, the transaction
+ * should be aborted, just like for other recovery conflicts.
+ * But if it's walsender on standby, we don't want to go
+ * through the following IsTransactionOrTransactionBlock()
+ * check, so break here.
+ */
+ if (am_cascading_walsender &&
+ reason == PROCSIG_RECOVERY_CONFLICT_LOGICALSLOT &&
+ MyReplicationSlot && SlotIsLogical(MyReplicationSlot))
+ {
+ RecoveryConflictPending = true;
+ QueryCancelPending = true;
+ InterruptPending = true;
+ break;
+ }
/*
* If we aren't in a transaction any longer then ignore.
diff --git a/src/backend/utils/activity/pgstat_database.c b/src/backend/utils/activity/pgstat_database.c
index 6e650ceaad..7149f22f72 100644
--- a/src/backend/utils/activity/pgstat_database.c
+++ b/src/backend/utils/activity/pgstat_database.c
@@ -109,6 +109,9 @@ pgstat_report_recovery_conflict(int reason)
case PROCSIG_RECOVERY_CONFLICT_BUFFERPIN:
dbentry->conflict_bufferpin++;
break;
+ case PROCSIG_RECOVERY_CONFLICT_LOGICALSLOT:
+ dbentry->conflict_logicalslot++;
+ break;
case PROCSIG_RECOVERY_CONFLICT_STARTUP_DEADLOCK:
dbentry->conflict_startup_deadlock++;
break;
@@ -387,6 +390,7 @@ pgstat_database_flush_cb(PgStat_EntryRef *entry_ref, bool nowait)
PGSTAT_ACCUM_DBCOUNT(conflict_tablespace);
PGSTAT_ACCUM_DBCOUNT(conflict_lock);
PGSTAT_ACCUM_DBCOUNT(conflict_snapshot);
+ PGSTAT_ACCUM_DBCOUNT(conflict_logicalslot);
PGSTAT_ACCUM_DBCOUNT(conflict_bufferpin);
PGSTAT_ACCUM_DBCOUNT(conflict_startup_deadlock);
diff --git a/src/backend/utils/adt/pgstatfuncs.c b/src/backend/utils/adt/pgstatfuncs.c
index 6737493402..afd62d3cc0 100644
--- a/src/backend/utils/adt/pgstatfuncs.c
+++ b/src/backend/utils/adt/pgstatfuncs.c
@@ -1066,6 +1066,8 @@ PG_STAT_GET_DBENTRY_INT64(xact_commit)
/* pg_stat_get_db_xact_rollback */
PG_STAT_GET_DBENTRY_INT64(xact_rollback)
+/* pg_stat_get_db_conflict_logicalslot */
+PG_STAT_GET_DBENTRY_INT64(conflict_logicalslot)
Datum
pg_stat_get_db_stat_reset_time(PG_FUNCTION_ARGS)
@@ -1099,6 +1101,7 @@ pg_stat_get_db_conflict_all(PG_FUNCTION_ARGS)
result = (int64) (dbentry->conflict_tablespace +
dbentry->conflict_lock +
dbentry->conflict_snapshot +
+ dbentry->conflict_logicalslot +
dbentry->conflict_bufferpin +
dbentry->conflict_startup_deadlock);
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index c0f2a8a77c..c8e11ab710 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -5577,6 +5577,11 @@
proname => 'pg_stat_get_db_conflict_snapshot', provolatile => 's',
proparallel => 'r', prorettype => 'int8', proargtypes => 'oid',
prosrc => 'pg_stat_get_db_conflict_snapshot' },
+{ oid => '9901',
+ descr => 'statistics: recovery conflicts in database caused by logical replication slot',
+ proname => 'pg_stat_get_db_conflict_logicalslot', provolatile => 's',
+ proparallel => 'r', prorettype => 'int8', proargtypes => 'oid',
+ prosrc => 'pg_stat_get_db_conflict_logicalslot' },
{ oid => '3068',
descr => 'statistics: recovery conflicts in database caused by shared buffer pin',
proname => 'pg_stat_get_db_conflict_bufferpin', provolatile => 's',
@@ -10946,9 +10951,9 @@
proname => 'pg_get_replication_slots', prorows => '10', proisstrict => 'f',
proretset => 't', provolatile => 's', prorettype => 'record',
proargtypes => '',
- proallargtypes => '{name,name,text,oid,bool,bool,int4,xid,xid,pg_lsn,pg_lsn,text,int8,bool}',
- proargmodes => '{o,o,o,o,o,o,o,o,o,o,o,o,o,o}',
- proargnames => '{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn,wal_status,safe_wal_size,two_phase}',
+ proallargtypes => '{name,name,text,oid,bool,bool,int4,xid,xid,pg_lsn,pg_lsn,text,int8,bool,bool}',
+ proargmodes => '{o,o,o,o,o,o,o,o,o,o,o,o,o,o,o}',
+ proargnames => '{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn,wal_status,safe_wal_size,two_phase,conflicting}',
prosrc => 'pg_get_replication_slots' },
{ oid => '3786', descr => 'set up a logical replication slot',
proname => 'pg_create_logical_replication_slot', provolatile => 'v',
diff --git a/src/include/pgstat.h b/src/include/pgstat.h
index 5e3326a3b9..872eb35757 100644
--- a/src/include/pgstat.h
+++ b/src/include/pgstat.h
@@ -291,6 +291,7 @@ typedef struct PgStat_StatDBEntry
PgStat_Counter conflict_tablespace;
PgStat_Counter conflict_lock;
PgStat_Counter conflict_snapshot;
+ PgStat_Counter conflict_logicalslot;
PgStat_Counter conflict_bufferpin;
PgStat_Counter conflict_startup_deadlock;
PgStat_Counter temp_files;
diff --git a/src/include/replication/slot.h b/src/include/replication/slot.h
index 8872c80cdf..236ebcdbdb 100644
--- a/src/include/replication/slot.h
+++ b/src/include/replication/slot.h
@@ -17,6 +17,8 @@
#include "storage/spin.h"
#include "replication/walreceiver.h"
+#define LogicalReplicationSlotIsInvalid(s) (!TransactionIdIsValid(s->data.xmin) && \
+ !TransactionIdIsValid(s->data.catalog_xmin))
/*
* Behaviour of replication slots, upon release or crash.
*
@@ -215,7 +217,7 @@ extern void ReplicationSlotsComputeRequiredLSN(void);
extern XLogRecPtr ReplicationSlotsComputeLogicalRestartLSN(void);
extern bool ReplicationSlotsCountDBSlots(Oid dboid, int *nslots, int *nactive);
extern void ReplicationSlotsDropDBSlots(Oid dboid);
-extern bool InvalidateObsoleteReplicationSlots(XLogSegNo oldestSegno);
+extern void InvalidateObsoleteReplicationSlots(XLogSegNo oldestSegno, bool *invalidated, Oid dboid, TransactionId *xid);
extern ReplicationSlot *SearchNamedReplicationSlot(const char *name, bool need_lock);
extern int ReplicationSlotIndex(ReplicationSlot *slot);
extern bool ReplicationSlotName(int index, Name name);
@@ -227,5 +229,6 @@ extern void CheckPointReplicationSlots(void);
extern void CheckSlotRequirements(void);
extern void CheckSlotPermissions(void);
+extern void ResolveRecoveryConflictWithLogicalSlots(Oid dboid, TransactionId xid, char *reason);
#endif /* SLOT_H */
diff --git a/src/include/storage/procsignal.h b/src/include/storage/procsignal.h
index 905af2231b..2f52100b00 100644
--- a/src/include/storage/procsignal.h
+++ b/src/include/storage/procsignal.h
@@ -42,6 +42,7 @@ typedef enum
PROCSIG_RECOVERY_CONFLICT_TABLESPACE,
PROCSIG_RECOVERY_CONFLICT_LOCK,
PROCSIG_RECOVERY_CONFLICT_SNAPSHOT,
+ PROCSIG_RECOVERY_CONFLICT_LOGICALSLOT,
PROCSIG_RECOVERY_CONFLICT_BUFFERPIN,
PROCSIG_RECOVERY_CONFLICT_STARTUP_DEADLOCK,
diff --git a/src/include/storage/standby.h b/src/include/storage/standby.h
index 2effdea126..41f4dc372e 100644
--- a/src/include/storage/standby.h
+++ b/src/include/storage/standby.h
@@ -30,8 +30,10 @@ extern void InitRecoveryTransactionEnvironment(void);
extern void ShutdownRecoveryTransactionEnvironment(void);
extern void ResolveRecoveryConflictWithSnapshot(TransactionId snapshotConflictHorizon,
+ bool isCatalogRel,
RelFileLocator locator);
extern void ResolveRecoveryConflictWithSnapshotFullXid(FullTransactionId snapshotConflictHorizon,
+ bool isCatalogRel,
RelFileLocator locator);
extern void ResolveRecoveryConflictWithTablespace(Oid tsid);
extern void ResolveRecoveryConflictWithDatabase(Oid dbid);
diff --git a/src/test/regress/expected/rules.out b/src/test/regress/expected/rules.out
index e7a2f5856a..11ea206337 100644
--- a/src/test/regress/expected/rules.out
+++ b/src/test/regress/expected/rules.out
@@ -1472,8 +1472,9 @@ pg_replication_slots| SELECT l.slot_name,
l.confirmed_flush_lsn,
l.wal_status,
l.safe_wal_size,
- l.two_phase
- FROM (pg_get_replication_slots() l(slot_name, plugin, slot_type, datoid, temporary, active, active_pid, xmin, catalog_xmin, restart_lsn, confirmed_flush_lsn, wal_status, safe_wal_size, two_phase)
+ l.two_phase,
+ l.conflicting
+ FROM (pg_get_replication_slots() l(slot_name, plugin, slot_type, datoid, temporary, active, active_pid, xmin, catalog_xmin, restart_lsn, confirmed_flush_lsn, wal_status, safe_wal_size, two_phase, conflicting)
LEFT JOIN pg_database d ON ((l.datoid = d.oid)));
pg_roles| SELECT pg_authid.rolname,
pg_authid.rolsuper,
@@ -1868,7 +1869,8 @@ pg_stat_database_conflicts| SELECT oid AS datid,
pg_stat_get_db_conflict_lock(oid) AS confl_lock,
pg_stat_get_db_conflict_snapshot(oid) AS confl_snapshot,
pg_stat_get_db_conflict_bufferpin(oid) AS confl_bufferpin,
- pg_stat_get_db_conflict_startup_deadlock(oid) AS confl_deadlock
+ pg_stat_get_db_conflict_startup_deadlock(oid) AS confl_deadlock,
+ pg_stat_get_db_conflict_logicalslot(oid) AS confl_active_logicalslot
FROM pg_database d;
pg_stat_gssapi| SELECT pid,
gss_auth AS gss_authenticated,
--
2.34.1
[text/plain] v45-0001-Add-info-in-WAL-records-in-preparation-for-logic.patch (76.2K, ../[email protected]/7-v45-0001-Add-info-in-WAL-records-in-preparation-for-logic.patch)
download | inline diff:
From 7ab0a6ec046f0c181c8f3b96cb18ae938f573ba9 Mon Sep 17 00:00:00 2001
From: bdrouvotAWS <[email protected]>
Date: Mon, 30 Jan 2023 15:30:15 +0000
Subject: [PATCH v45 1/6] Add info in WAL records in preparation for logical
slot conflict handling.
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Overall design:
1. We want to enable logical decoding on standbys, but replay of WAL
from the primary might remove data that is needed by logical decoding,
causing error(s) on the standby. To prevent those errors, a new replication
conflict scenario needs to be addressed (as much as hot standby does).
2. Our chosen strategy for dealing with this type of replication slot
is to invalidate logical slots for which needed data has been removed.
3. To do this we need the latestRemovedXid for each change, just as we
do for physical replication conflicts, but we also need to know
whether any particular change was to data that logical replication
might access. That way, during WAL replay, we know when there is a risk of
conflict and, if so, if there is a conflict.
4. We can't rely on the standby's relcache entries for this purpose in
any way, because the startup process can't access catalog contents.
5. Therefore every WAL record that potentially removes data from the
index or heap must carry a flag indicating whether or not it is one
that might be accessed during logical decoding.
Why do we need this for logical decoding on standby?
First, let's forget about logical decoding on standby and recall that
on a primary database, any catalog rows that may be needed by a logical
decoding replication slot are not removed.
This is done thanks to the catalog_xmin associated with the logical
replication slot.
But, with logical decoding on standby, in the following cases:
- hot_standby_feedback is off
- hot_standby_feedback is on but there is no a physical slot between
the primary and the standby. Then, hot_standby_feedback will work,
but only while the connection is alive (for example a node restart
would break it)
Then, the primary may delete system catalog rows that could be needed
by the logical decoding on the standby (as it does not know about the
catalog_xmin on the standby).
So, it’s mandatory to identify those rows and invalidate the slots
that may need them if any. Identifying those rows is the purpose of
this commit.
Implementation:
When a WAL replay on standby indicates that a catalog table tuple is
to be deleted by an xid that is greater than a logical slot's
catalog_xmin, then that means the slot's catalog_xmin conflicts with
the xid, and we need to handle the conflict. While subsequent commits
will do the actual conflict handling, this commit adds a new field
isCatalogRel in such WAL records (and a new bit set in the
xl_heap_visible flags field), that is true for catalog tables, so as to
arrange for conflict handling.
The affected WAL records are the ones that already contain the
snapshotConflictHorizon field, namely:
- gistxlogDelete
- gistxlogPageReuse
- xl_hash_vacuum_one_page
- xl_heap_prune
- xl_heap_freeze_page
- xl_heap_visible
- xl_btree_reuse_page
- xl_btree_delete
- spgxlogVacuumRedirect
Due to this new field being added, xl_hash_vacuum_one_page and
gistxlogDelete do now contain the offsets to be deleted as a
FLEXIBLE_ARRAY_MEMBER. This is needed to ensure correct alignement.
It's not needed on the others struct where isCatalogRel has
been added.
Author: Andres Freund (in an older version), Amit Khandekar, Bertrand
Drouvot
Reviewed-By: Bertrand Drouvot, Andres Freund, Robert Haas, Fabrizio de
Royes Mello, Melanie Plageman
---
contrib/amcheck/verify_nbtree.c | 15 +--
src/backend/access/gist/gist.c | 5 +-
src/backend/access/gist/gistbuild.c | 2 +-
src/backend/access/gist/gistutil.c | 4 +-
src/backend/access/gist/gistxlog.c | 17 ++--
src/backend/access/hash/hash_xlog.c | 12 +--
src/backend/access/hash/hashinsert.c | 1 +
src/backend/access/heap/heapam.c | 5 +-
src/backend/access/heap/heapam_handler.c | 9 +-
src/backend/access/heap/pruneheap.c | 1 +
src/backend/access/heap/vacuumlazy.c | 2 +
src/backend/access/heap/visibilitymap.c | 3 +-
src/backend/access/nbtree/nbtinsert.c | 91 +++++++++--------
src/backend/access/nbtree/nbtpage.c | 111 +++++++++++----------
src/backend/access/nbtree/nbtree.c | 4 +-
src/backend/access/nbtree/nbtsearch.c | 50 ++++++----
src/backend/access/nbtree/nbtsort.c | 2 +-
src/backend/access/nbtree/nbtutils.c | 7 +-
src/backend/access/spgist/spgvacuum.c | 9 +-
src/backend/catalog/index.c | 1 +
src/backend/commands/analyze.c | 1 +
src/backend/commands/vacuumparallel.c | 6 ++
src/backend/optimizer/util/plancat.c | 2 +-
src/backend/utils/sort/tuplesortvariants.c | 5 +-
src/include/access/genam.h | 1 +
src/include/access/gist_private.h | 7 +-
src/include/access/gistxlog.h | 13 ++-
src/include/access/hash_xlog.h | 8 +-
src/include/access/heapam_xlog.h | 10 +-
src/include/access/nbtree.h | 37 ++++---
src/include/access/nbtxlog.h | 8 +-
src/include/access/spgxlog.h | 2 +
src/include/access/visibilitymapdefs.h | 10 +-
src/include/utils/rel.h | 1 +
src/include/utils/tuplesort.h | 4 +-
35 files changed, 263 insertions(+), 203 deletions(-)
3.3% contrib/amcheck/
4.7% src/backend/access/gist/
4.1% src/backend/access/heap/
59.0% src/backend/access/nbtree/
3.7% src/backend/access/
22.0% src/include/access/
diff --git a/contrib/amcheck/verify_nbtree.c b/contrib/amcheck/verify_nbtree.c
index 257cff671b..eb280d4893 100644
--- a/contrib/amcheck/verify_nbtree.c
+++ b/contrib/amcheck/verify_nbtree.c
@@ -183,6 +183,7 @@ static inline bool invariant_l_nontarget_offset(BtreeCheckState *state,
OffsetNumber upperbound);
static Page palloc_btree_page(BtreeCheckState *state, BlockNumber blocknum);
static inline BTScanInsert bt_mkscankey_pivotsearch(Relation rel,
+ Relation heaprel,
IndexTuple itup);
static ItemId PageGetItemIdCareful(BtreeCheckState *state, BlockNumber block,
Page page, OffsetNumber offset);
@@ -331,7 +332,7 @@ bt_index_check_internal(Oid indrelid, bool parentcheck, bool heapallindexed,
RelationGetRelationName(indrel))));
/* Extract metadata from metapage, and sanitize it in passing */
- _bt_metaversion(indrel, &heapkeyspace, &allequalimage);
+ _bt_metaversion(indrel, heaprel, &heapkeyspace, &allequalimage);
if (allequalimage && !heapkeyspace)
ereport(ERROR,
(errcode(ERRCODE_INDEX_CORRUPTED),
@@ -1258,7 +1259,7 @@ bt_target_page_check(BtreeCheckState *state)
}
/* Build insertion scankey for current page offset */
- skey = bt_mkscankey_pivotsearch(state->rel, itup);
+ skey = bt_mkscankey_pivotsearch(state->rel, state->heaprel, itup);
/*
* Make sure tuple size does not exceed the relevant BTREE_VERSION
@@ -1768,7 +1769,7 @@ bt_right_page_check_scankey(BtreeCheckState *state)
* memory remaining allocated.
*/
firstitup = (IndexTuple) PageGetItem(rightpage, rightitem);
- return bt_mkscankey_pivotsearch(state->rel, firstitup);
+ return bt_mkscankey_pivotsearch(state->rel, state->heaprel, firstitup);
}
/*
@@ -2681,7 +2682,7 @@ bt_rootdescend(BtreeCheckState *state, IndexTuple itup)
Buffer lbuf;
bool exists;
- key = _bt_mkscankey(state->rel, itup);
+ key = _bt_mkscankey(state->rel, state->heaprel, itup);
Assert(key->heapkeyspace && key->scantid != NULL);
/*
@@ -2694,7 +2695,7 @@ bt_rootdescend(BtreeCheckState *state, IndexTuple itup)
*/
Assert(state->readonly && state->rootdescend);
exists = false;
- stack = _bt_search(state->rel, key, &lbuf, BT_READ, NULL);
+ stack = _bt_search(state->rel, state->heaprel, key, &lbuf, BT_READ, NULL);
if (BufferIsValid(lbuf))
{
@@ -3133,11 +3134,11 @@ palloc_btree_page(BtreeCheckState *state, BlockNumber blocknum)
* the scankey is greater.
*/
static inline BTScanInsert
-bt_mkscankey_pivotsearch(Relation rel, IndexTuple itup)
+bt_mkscankey_pivotsearch(Relation rel, Relation heaprel, IndexTuple itup)
{
BTScanInsert skey;
- skey = _bt_mkscankey(rel, itup);
+ skey = _bt_mkscankey(rel, heaprel, itup);
skey->pivotsearch = true;
return skey;
diff --git a/src/backend/access/gist/gist.c b/src/backend/access/gist/gist.c
index ba394f08f6..3ac68ec3b4 100644
--- a/src/backend/access/gist/gist.c
+++ b/src/backend/access/gist/gist.c
@@ -348,7 +348,7 @@ gistplacetopage(Relation rel, Size freespace, GISTSTATE *giststate,
for (; ptr; ptr = ptr->next)
{
/* Allocate new page */
- ptr->buffer = gistNewBuffer(rel);
+ ptr->buffer = gistNewBuffer(rel, heapRel);
GISTInitBuffer(ptr->buffer, (is_leaf) ? F_LEAF : 0);
ptr->page = BufferGetPage(ptr->buffer);
ptr->block.blkno = BufferGetBlockNumber(ptr->buffer);
@@ -1694,7 +1694,8 @@ gistprunepage(Relation rel, Page page, Buffer buffer, Relation heapRel)
recptr = gistXLogDelete(buffer,
deletable, ndeletable,
- snapshotConflictHorizon);
+ snapshotConflictHorizon,
+ heapRel);
PageSetLSN(page, recptr);
}
diff --git a/src/backend/access/gist/gistbuild.c b/src/backend/access/gist/gistbuild.c
index d21a308d41..4462022904 100644
--- a/src/backend/access/gist/gistbuild.c
+++ b/src/backend/access/gist/gistbuild.c
@@ -298,7 +298,7 @@ gistbuild(Relation heap, Relation index, IndexInfo *indexInfo)
Page page;
/* initialize the root page */
- buffer = gistNewBuffer(index);
+ buffer = gistNewBuffer(index, heap);
Assert(BufferGetBlockNumber(buffer) == GIST_ROOT_BLKNO);
page = BufferGetPage(buffer);
diff --git a/src/backend/access/gist/gistutil.c b/src/backend/access/gist/gistutil.c
index 56451fede1..aad14a401d 100644
--- a/src/backend/access/gist/gistutil.c
+++ b/src/backend/access/gist/gistutil.c
@@ -821,7 +821,7 @@ gistcheckpage(Relation rel, Buffer buf)
* Caller is responsible for initializing the page by calling GISTInitBuffer
*/
Buffer
-gistNewBuffer(Relation r)
+gistNewBuffer(Relation r, Relation heaprel)
{
Buffer buffer;
bool needLock;
@@ -865,7 +865,7 @@ gistNewBuffer(Relation r)
* page's deleteXid.
*/
if (XLogStandbyInfoActive() && RelationNeedsWAL(r))
- gistXLogPageReuse(r, blkno, GistPageGetDeleteXid(page));
+ gistXLogPageReuse(r, heaprel, blkno, GistPageGetDeleteXid(page));
return buffer;
}
diff --git a/src/backend/access/gist/gistxlog.c b/src/backend/access/gist/gistxlog.c
index f65864254a..b7678f3c14 100644
--- a/src/backend/access/gist/gistxlog.c
+++ b/src/backend/access/gist/gistxlog.c
@@ -177,6 +177,7 @@ gistRedoDeleteRecord(XLogReaderState *record)
gistxlogDelete *xldata = (gistxlogDelete *) XLogRecGetData(record);
Buffer buffer;
Page page;
+ OffsetNumber *toDelete = xldata->offsets;
/*
* If we have any conflict processing to do, it must happen before we
@@ -203,14 +204,7 @@ gistRedoDeleteRecord(XLogReaderState *record)
{
page = (Page) BufferGetPage(buffer);
- if (XLogRecGetDataLen(record) > SizeOfGistxlogDelete)
- {
- OffsetNumber *todelete;
-
- todelete = (OffsetNumber *) ((char *) xldata + SizeOfGistxlogDelete);
-
- PageIndexMultiDelete(page, todelete, xldata->ntodelete);
- }
+ PageIndexMultiDelete(page, toDelete, xldata->ntodelete);
GistClearPageHasGarbage(page);
GistMarkTuplesDeleted(page);
@@ -597,7 +591,8 @@ gistXLogAssignLSN(void)
* Write XLOG record about reuse of a deleted page.
*/
void
-gistXLogPageReuse(Relation rel, BlockNumber blkno, FullTransactionId deleteXid)
+gistXLogPageReuse(Relation rel, Relation heaprel,
+ BlockNumber blkno, FullTransactionId deleteXid)
{
gistxlogPageReuse xlrec_reuse;
@@ -608,6 +603,7 @@ gistXLogPageReuse(Relation rel, BlockNumber blkno, FullTransactionId deleteXid)
*/
/* XLOG stuff */
+ xlrec_reuse.isCatalogRel = RelationIsAccessibleInLogicalDecoding(heaprel);
xlrec_reuse.locator = rel->rd_locator;
xlrec_reuse.block = blkno;
xlrec_reuse.snapshotConflictHorizon = deleteXid;
@@ -672,11 +668,12 @@ gistXLogUpdate(Buffer buffer,
*/
XLogRecPtr
gistXLogDelete(Buffer buffer, OffsetNumber *todelete, int ntodelete,
- TransactionId snapshotConflictHorizon)
+ TransactionId snapshotConflictHorizon, Relation heaprel)
{
gistxlogDelete xlrec;
XLogRecPtr recptr;
+ xlrec.isCatalogRel = RelationIsAccessibleInLogicalDecoding(heaprel);
xlrec.snapshotConflictHorizon = snapshotConflictHorizon;
xlrec.ntodelete = ntodelete;
diff --git a/src/backend/access/hash/hash_xlog.c b/src/backend/access/hash/hash_xlog.c
index f38b42efb9..08ceb91288 100644
--- a/src/backend/access/hash/hash_xlog.c
+++ b/src/backend/access/hash/hash_xlog.c
@@ -980,8 +980,10 @@ hash_xlog_vacuum_one_page(XLogReaderState *record)
Page page;
XLogRedoAction action;
HashPageOpaque pageopaque;
+ OffsetNumber *toDelete;
xldata = (xl_hash_vacuum_one_page *) XLogRecGetData(record);
+ toDelete = xldata->offsets;
/*
* If we have any conflict processing to do, it must happen before we
@@ -1010,15 +1012,7 @@ hash_xlog_vacuum_one_page(XLogReaderState *record)
{
page = (Page) BufferGetPage(buffer);
- if (XLogRecGetDataLen(record) > SizeOfHashVacuumOnePage)
- {
- OffsetNumber *unused;
-
- unused = (OffsetNumber *) ((char *) xldata + SizeOfHashVacuumOnePage);
-
- PageIndexMultiDelete(page, unused, xldata->ntuples);
- }
-
+ PageIndexMultiDelete(page, toDelete, xldata->ntuples);
/*
* Mark the page as not containing any LP_DEAD items. See comments in
* _hash_vacuum_one_page() for details.
diff --git a/src/backend/access/hash/hashinsert.c b/src/backend/access/hash/hashinsert.c
index a604e31891..22656b24e2 100644
--- a/src/backend/access/hash/hashinsert.c
+++ b/src/backend/access/hash/hashinsert.c
@@ -432,6 +432,7 @@ _hash_vacuum_one_page(Relation rel, Relation hrel, Buffer metabuf, Buffer buf)
xl_hash_vacuum_one_page xlrec;
XLogRecPtr recptr;
+ xlrec.isCatalogRel = RelationIsAccessibleInLogicalDecoding(hrel);
xlrec.snapshotConflictHorizon = snapshotConflictHorizon;
xlrec.ntuples = ndeletable;
diff --git a/src/backend/access/heap/heapam.c b/src/backend/access/heap/heapam.c
index e6024a980b..d478724b9d 100644
--- a/src/backend/access/heap/heapam.c
+++ b/src/backend/access/heap/heapam.c
@@ -6872,6 +6872,7 @@ heap_freeze_execute_prepared(Relation rel, Buffer buffer,
nplans = heap_log_freeze_plan(tuples, ntuples, plans, offsets);
xlrec.snapshotConflictHorizon = snapshotConflictHorizon;
+ xlrec.isCatalogRel = RelationIsAccessibleInLogicalDecoding(rel);
xlrec.nplans = nplans;
XLogBeginInsert();
@@ -8442,7 +8443,7 @@ bottomup_sort_and_shrink(TM_IndexDeleteOp *delstate)
* update the heap page's LSN.
*/
XLogRecPtr
-log_heap_visible(RelFileLocator rlocator, Buffer heap_buffer, Buffer vm_buffer,
+log_heap_visible(Relation rel, Buffer heap_buffer, Buffer vm_buffer,
TransactionId snapshotConflictHorizon, uint8 vmflags)
{
xl_heap_visible xlrec;
@@ -8454,6 +8455,8 @@ log_heap_visible(RelFileLocator rlocator, Buffer heap_buffer, Buffer vm_buffer,
xlrec.snapshotConflictHorizon = snapshotConflictHorizon;
xlrec.flags = vmflags;
+ if (RelationIsAccessibleInLogicalDecoding(rel))
+ xlrec.flags |= VISIBILITYMAP_IS_CATALOG_REL;
XLogBeginInsert();
XLogRegisterData((char *) &xlrec, SizeOfHeapVisible);
diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index c4b1916d36..392c6e659c 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -720,9 +720,14 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
*multi_cutoff);
- /* Set up sorting if wanted */
+ /*
+ * Set up sorting if wanted. NewHeap is being passed to
+ * tuplesort_begin_cluster(), it could have been OldHeap too. It does not
+ * really matter, as the goal is to have a heap relation being passed to
+ * _bt_log_reuse_page() (which should not be called from this code path).
+ */
if (use_sort)
- tuplesort = tuplesort_begin_cluster(oldTupDesc, OldIndex,
+ tuplesort = tuplesort_begin_cluster(oldTupDesc, OldIndex, NewHeap,
maintenance_work_mem,
NULL, TUPLESORT_NONE);
else
diff --git a/src/backend/access/heap/pruneheap.c b/src/backend/access/heap/pruneheap.c
index 4e65cbcadf..3f0342351f 100644
--- a/src/backend/access/heap/pruneheap.c
+++ b/src/backend/access/heap/pruneheap.c
@@ -418,6 +418,7 @@ heap_page_prune(Relation relation, Buffer buffer,
xl_heap_prune xlrec;
XLogRecPtr recptr;
+ xlrec.isCatalogRel = RelationIsAccessibleInLogicalDecoding(relation);
xlrec.snapshotConflictHorizon = prstate.snapshotConflictHorizon;
xlrec.nredirected = prstate.nredirected;
xlrec.ndead = prstate.ndead;
diff --git a/src/backend/access/heap/vacuumlazy.c b/src/backend/access/heap/vacuumlazy.c
index 8f14cf85f3..ae628d747d 100644
--- a/src/backend/access/heap/vacuumlazy.c
+++ b/src/backend/access/heap/vacuumlazy.c
@@ -2710,6 +2710,7 @@ lazy_vacuum_one_index(Relation indrel, IndexBulkDeleteResult *istat,
ivinfo.message_level = DEBUG2;
ivinfo.num_heap_tuples = reltuples;
ivinfo.strategy = vacrel->bstrategy;
+ ivinfo.heaprel = vacrel->rel;
/*
* Update error traceback information.
@@ -2759,6 +2760,7 @@ lazy_cleanup_one_index(Relation indrel, IndexBulkDeleteResult *istat,
ivinfo.num_heap_tuples = reltuples;
ivinfo.strategy = vacrel->bstrategy;
+ ivinfo.heaprel = vacrel->rel;
/*
* Update error traceback information.
diff --git a/src/backend/access/heap/visibilitymap.c b/src/backend/access/heap/visibilitymap.c
index 74ff01bb17..d1ba859851 100644
--- a/src/backend/access/heap/visibilitymap.c
+++ b/src/backend/access/heap/visibilitymap.c
@@ -288,8 +288,7 @@ visibilitymap_set(Relation rel, BlockNumber heapBlk, Buffer heapBuf,
if (XLogRecPtrIsInvalid(recptr))
{
Assert(!InRecovery);
- recptr = log_heap_visible(rel->rd_locator, heapBuf, vmBuf,
- cutoff_xid, flags);
+ recptr = log_heap_visible(rel, heapBuf, vmBuf, cutoff_xid, flags);
/*
* If data checksums are enabled (or wal_log_hints=on), we
diff --git a/src/backend/access/nbtree/nbtinsert.c b/src/backend/access/nbtree/nbtinsert.c
index f4c1a974ef..8c6e867c61 100644
--- a/src/backend/access/nbtree/nbtinsert.c
+++ b/src/backend/access/nbtree/nbtinsert.c
@@ -30,7 +30,8 @@
#define BTREE_FASTPATH_MIN_LEVEL 2
-static BTStack _bt_search_insert(Relation rel, BTInsertState insertstate);
+static BTStack _bt_search_insert(Relation rel, Relation heaprel,
+ BTInsertState insertstate);
static TransactionId _bt_check_unique(Relation rel, BTInsertState insertstate,
Relation heapRel,
IndexUniqueCheck checkUnique, bool *is_unique,
@@ -41,8 +42,9 @@ static OffsetNumber _bt_findinsertloc(Relation rel,
bool indexUnchanged,
BTStack stack,
Relation heapRel);
-static void _bt_stepright(Relation rel, BTInsertState insertstate, BTStack stack);
-static void _bt_insertonpg(Relation rel, BTScanInsert itup_key,
+static void _bt_stepright(Relation rel, Relation heaprel,
+ BTInsertState insertstate, BTStack stack);
+static void _bt_insertonpg(Relation rel, Relation heaprel, BTScanInsert itup_key,
Buffer buf,
Buffer cbuf,
BTStack stack,
@@ -51,13 +53,13 @@ static void _bt_insertonpg(Relation rel, BTScanInsert itup_key,
OffsetNumber newitemoff,
int postingoff,
bool split_only_page);
-static Buffer _bt_split(Relation rel, BTScanInsert itup_key, Buffer buf,
- Buffer cbuf, OffsetNumber newitemoff, Size newitemsz,
- IndexTuple newitem, IndexTuple orignewitem,
+static Buffer _bt_split(Relation rel, Relation heaprel, BTScanInsert itup_key,
+ Buffer buf, Buffer cbuf, OffsetNumber newitemoff,
+ Size newitemsz, IndexTuple newitem, IndexTuple orignewitem,
IndexTuple nposting, uint16 postingoff);
-static void _bt_insert_parent(Relation rel, Buffer buf, Buffer rbuf,
- BTStack stack, bool isroot, bool isonly);
-static Buffer _bt_newroot(Relation rel, Buffer lbuf, Buffer rbuf);
+static void _bt_insert_parent(Relation rel, Relation heaprel, Buffer buf,
+ Buffer rbuf, BTStack stack, bool isroot, bool isonly);
+static Buffer _bt_newroot(Relation rel, Relation heaprel, Buffer lbuf, Buffer rbuf);
static inline bool _bt_pgaddtup(Page page, Size itemsize, IndexTuple itup,
OffsetNumber itup_off, bool newfirstdataitem);
static void _bt_delete_or_dedup_one_page(Relation rel, Relation heapRel,
@@ -108,7 +110,7 @@ _bt_doinsert(Relation rel, IndexTuple itup,
bool checkingunique = (checkUnique != UNIQUE_CHECK_NO);
/* we need an insertion scan key to do our search, so build one */
- itup_key = _bt_mkscankey(rel, itup);
+ itup_key = _bt_mkscankey(rel, heapRel, itup);
if (checkingunique)
{
@@ -162,7 +164,7 @@ search:
* searching from the root page. insertstate.buf will hold a buffer that
* is locked in exclusive mode afterwards.
*/
- stack = _bt_search_insert(rel, &insertstate);
+ stack = _bt_search_insert(rel, heapRel, &insertstate);
/*
* checkingunique inserts are not allowed to go ahead when two tuples with
@@ -255,8 +257,8 @@ search:
*/
newitemoff = _bt_findinsertloc(rel, &insertstate, checkingunique,
indexUnchanged, stack, heapRel);
- _bt_insertonpg(rel, itup_key, insertstate.buf, InvalidBuffer, stack,
- itup, insertstate.itemsz, newitemoff,
+ _bt_insertonpg(rel, heapRel, itup_key, insertstate.buf, InvalidBuffer,
+ stack, itup, insertstate.itemsz, newitemoff,
insertstate.postingoff, false);
}
else
@@ -312,7 +314,7 @@ search:
* since each per-backend cache won't stay valid for long.
*/
static BTStack
-_bt_search_insert(Relation rel, BTInsertState insertstate)
+_bt_search_insert(Relation rel, Relation heaprel, BTInsertState insertstate)
{
Assert(insertstate->buf == InvalidBuffer);
Assert(!insertstate->bounds_valid);
@@ -375,8 +377,8 @@ _bt_search_insert(Relation rel, BTInsertState insertstate)
}
/* Cannot use optimization -- descend tree, return proper descent stack */
- return _bt_search(rel, insertstate->itup_key, &insertstate->buf, BT_WRITE,
- NULL);
+ return _bt_search(rel, heaprel, insertstate->itup_key, &insertstate->buf,
+ BT_WRITE, NULL);
}
/*
@@ -885,7 +887,7 @@ _bt_findinsertloc(Relation rel,
_bt_compare(rel, itup_key, page, P_HIKEY) <= 0)
break;
- _bt_stepright(rel, insertstate, stack);
+ _bt_stepright(rel, heapRel, insertstate, stack);
/* Update local state after stepping right */
page = BufferGetPage(insertstate->buf);
opaque = BTPageGetOpaque(page);
@@ -969,7 +971,7 @@ _bt_findinsertloc(Relation rel,
pg_prng_uint32(&pg_global_prng_state) <= (PG_UINT32_MAX / 100))
break;
- _bt_stepright(rel, insertstate, stack);
+ _bt_stepright(rel, heapRel, insertstate, stack);
/* Update local state after stepping right */
page = BufferGetPage(insertstate->buf);
opaque = BTPageGetOpaque(page);
@@ -1022,7 +1024,7 @@ _bt_findinsertloc(Relation rel,
* indexes.
*/
static void
-_bt_stepright(Relation rel, BTInsertState insertstate, BTStack stack)
+_bt_stepright(Relation rel, Relation heaprel, BTInsertState insertstate, BTStack stack)
{
Page page;
BTPageOpaque opaque;
@@ -1048,7 +1050,7 @@ _bt_stepright(Relation rel, BTInsertState insertstate, BTStack stack)
*/
if (P_INCOMPLETE_SPLIT(opaque))
{
- _bt_finish_split(rel, rbuf, stack);
+ _bt_finish_split(rel, heaprel, rbuf, stack);
rbuf = InvalidBuffer;
continue;
}
@@ -1099,6 +1101,7 @@ _bt_stepright(Relation rel, BTInsertState insertstate, BTStack stack)
*/
static void
_bt_insertonpg(Relation rel,
+ Relation heaprel,
BTScanInsert itup_key,
Buffer buf,
Buffer cbuf,
@@ -1209,8 +1212,8 @@ _bt_insertonpg(Relation rel,
Assert(!split_only_page);
/* split the buffer into left and right halves */
- rbuf = _bt_split(rel, itup_key, buf, cbuf, newitemoff, itemsz, itup,
- origitup, nposting, postingoff);
+ rbuf = _bt_split(rel, heaprel, itup_key, buf, cbuf, newitemoff, itemsz,
+ itup, origitup, nposting, postingoff);
PredicateLockPageSplit(rel,
BufferGetBlockNumber(buf),
BufferGetBlockNumber(rbuf));
@@ -1233,7 +1236,7 @@ _bt_insertonpg(Relation rel,
* page.
*----------
*/
- _bt_insert_parent(rel, buf, rbuf, stack, isroot, isonly);
+ _bt_insert_parent(rel, heaprel, buf, rbuf, stack, isroot, isonly);
}
else
{
@@ -1254,7 +1257,7 @@ _bt_insertonpg(Relation rel,
Assert(!isleaf);
Assert(BufferIsValid(cbuf));
- metabuf = _bt_getbuf(rel, BTREE_METAPAGE, BT_WRITE);
+ metabuf = _bt_getbuf(rel, heaprel, BTREE_METAPAGE, BT_WRITE);
metapg = BufferGetPage(metabuf);
metad = BTPageGetMeta(metapg);
@@ -1418,7 +1421,7 @@ _bt_insertonpg(Relation rel,
* call _bt_getrootheight while holding a buffer lock.
*/
if (BlockNumberIsValid(blockcache) &&
- _bt_getrootheight(rel) >= BTREE_FASTPATH_MIN_LEVEL)
+ _bt_getrootheight(rel, heaprel) >= BTREE_FASTPATH_MIN_LEVEL)
RelationSetTargetBlock(rel, blockcache);
}
@@ -1459,8 +1462,8 @@ _bt_insertonpg(Relation rel,
* The pin and lock on buf are maintained.
*/
static Buffer
-_bt_split(Relation rel, BTScanInsert itup_key, Buffer buf, Buffer cbuf,
- OffsetNumber newitemoff, Size newitemsz, IndexTuple newitem,
+_bt_split(Relation rel, Relation heaprel, BTScanInsert itup_key, Buffer buf,
+ Buffer cbuf, OffsetNumber newitemoff, Size newitemsz, IndexTuple newitem,
IndexTuple orignewitem, IndexTuple nposting, uint16 postingoff)
{
Buffer rbuf;
@@ -1712,7 +1715,7 @@ _bt_split(Relation rel, BTScanInsert itup_key, Buffer buf, Buffer cbuf,
* way because it avoids an unnecessary PANIC when either origpage or its
* existing sibling page are corrupt.
*/
- rbuf = _bt_getbuf(rel, P_NEW, BT_WRITE);
+ rbuf = _bt_getbuf(rel, heaprel, P_NEW, BT_WRITE);
rightpage = BufferGetPage(rbuf);
rightpagenumber = BufferGetBlockNumber(rbuf);
/* rightpage was initialized by _bt_getbuf */
@@ -1885,7 +1888,7 @@ _bt_split(Relation rel, BTScanInsert itup_key, Buffer buf, Buffer cbuf,
*/
if (!isrightmost)
{
- sbuf = _bt_getbuf(rel, oopaque->btpo_next, BT_WRITE);
+ sbuf = _bt_getbuf(rel, heaprel, oopaque->btpo_next, BT_WRITE);
spage = BufferGetPage(sbuf);
sopaque = BTPageGetOpaque(spage);
if (sopaque->btpo_prev != origpagenumber)
@@ -2092,6 +2095,7 @@ _bt_split(Relation rel, BTScanInsert itup_key, Buffer buf, Buffer cbuf,
*/
static void
_bt_insert_parent(Relation rel,
+ Relation heaprel,
Buffer buf,
Buffer rbuf,
BTStack stack,
@@ -2118,7 +2122,7 @@ _bt_insert_parent(Relation rel,
Assert(stack == NULL);
Assert(isonly);
/* create a new root node and update the metapage */
- rootbuf = _bt_newroot(rel, buf, rbuf);
+ rootbuf = _bt_newroot(rel, heaprel, buf, rbuf);
/* release the split buffers */
_bt_relbuf(rel, rootbuf);
_bt_relbuf(rel, rbuf);
@@ -2157,7 +2161,8 @@ _bt_insert_parent(Relation rel,
BlockNumberIsValid(RelationGetTargetBlock(rel))));
/* Find the leftmost page at the next level up */
- pbuf = _bt_get_endpoint(rel, opaque->btpo_level + 1, false, NULL);
+ pbuf = _bt_get_endpoint(rel, heaprel, opaque->btpo_level + 1, false,
+ NULL);
/* Set up a phony stack entry pointing there */
stack = &fakestack;
stack->bts_blkno = BufferGetBlockNumber(pbuf);
@@ -2183,7 +2188,7 @@ _bt_insert_parent(Relation rel,
* new downlink will be inserted at the correct offset. Even buf's
* parent may have changed.
*/
- pbuf = _bt_getstackbuf(rel, stack, bknum);
+ pbuf = _bt_getstackbuf(rel, heaprel, stack, bknum);
/*
* Unlock the right child. The left child will be unlocked in
@@ -2207,7 +2212,7 @@ _bt_insert_parent(Relation rel,
RelationGetRelationName(rel), bknum, rbknum)));
/* Recursively insert into the parent */
- _bt_insertonpg(rel, NULL, pbuf, buf, stack->bts_parent,
+ _bt_insertonpg(rel, heaprel, NULL, pbuf, buf, stack->bts_parent,
new_item, MAXALIGN(IndexTupleSize(new_item)),
stack->bts_offset + 1, 0, isonly);
@@ -2227,7 +2232,7 @@ _bt_insert_parent(Relation rel,
* and unpinned.
*/
void
-_bt_finish_split(Relation rel, Buffer lbuf, BTStack stack)
+_bt_finish_split(Relation rel, Relation heaprel, Buffer lbuf, BTStack stack)
{
Page lpage = BufferGetPage(lbuf);
BTPageOpaque lpageop = BTPageGetOpaque(lpage);
@@ -2240,7 +2245,7 @@ _bt_finish_split(Relation rel, Buffer lbuf, BTStack stack)
Assert(P_INCOMPLETE_SPLIT(lpageop));
/* Lock right sibling, the one missing the downlink */
- rbuf = _bt_getbuf(rel, lpageop->btpo_next, BT_WRITE);
+ rbuf = _bt_getbuf(rel, heaprel, lpageop->btpo_next, BT_WRITE);
rpage = BufferGetPage(rbuf);
rpageop = BTPageGetOpaque(rpage);
@@ -2252,7 +2257,7 @@ _bt_finish_split(Relation rel, Buffer lbuf, BTStack stack)
BTMetaPageData *metad;
/* acquire lock on the metapage */
- metabuf = _bt_getbuf(rel, BTREE_METAPAGE, BT_WRITE);
+ metabuf = _bt_getbuf(rel, heaprel, BTREE_METAPAGE, BT_WRITE);
metapg = BufferGetPage(metabuf);
metad = BTPageGetMeta(metapg);
@@ -2269,7 +2274,7 @@ _bt_finish_split(Relation rel, Buffer lbuf, BTStack stack)
elog(DEBUG1, "finishing incomplete split of %u/%u",
BufferGetBlockNumber(lbuf), BufferGetBlockNumber(rbuf));
- _bt_insert_parent(rel, lbuf, rbuf, stack, wasroot, wasonly);
+ _bt_insert_parent(rel, heaprel, lbuf, rbuf, stack, wasroot, wasonly);
}
/*
@@ -2304,7 +2309,7 @@ _bt_finish_split(Relation rel, Buffer lbuf, BTStack stack)
* offset number bts_offset + 1.
*/
Buffer
-_bt_getstackbuf(Relation rel, BTStack stack, BlockNumber child)
+_bt_getstackbuf(Relation rel, Relation heaprel, BTStack stack, BlockNumber child)
{
BlockNumber blkno;
OffsetNumber start;
@@ -2318,13 +2323,13 @@ _bt_getstackbuf(Relation rel, BTStack stack, BlockNumber child)
Page page;
BTPageOpaque opaque;
- buf = _bt_getbuf(rel, blkno, BT_WRITE);
+ buf = _bt_getbuf(rel, heaprel, blkno, BT_WRITE);
page = BufferGetPage(buf);
opaque = BTPageGetOpaque(page);
if (P_INCOMPLETE_SPLIT(opaque))
{
- _bt_finish_split(rel, buf, stack->bts_parent);
+ _bt_finish_split(rel, heaprel, buf, stack->bts_parent);
continue;
}
@@ -2428,7 +2433,7 @@ _bt_getstackbuf(Relation rel, BTStack stack, BlockNumber child)
* lbuf, rbuf & rootbuf.
*/
static Buffer
-_bt_newroot(Relation rel, Buffer lbuf, Buffer rbuf)
+_bt_newroot(Relation rel, Relation heaprel, Buffer lbuf, Buffer rbuf)
{
Buffer rootbuf;
Page lpage,
@@ -2454,12 +2459,12 @@ _bt_newroot(Relation rel, Buffer lbuf, Buffer rbuf)
lopaque = BTPageGetOpaque(lpage);
/* get a new root page */
- rootbuf = _bt_getbuf(rel, P_NEW, BT_WRITE);
+ rootbuf = _bt_getbuf(rel, heaprel, P_NEW, BT_WRITE);
rootpage = BufferGetPage(rootbuf);
rootblknum = BufferGetBlockNumber(rootbuf);
/* acquire lock on the metapage */
- metabuf = _bt_getbuf(rel, BTREE_METAPAGE, BT_WRITE);
+ metabuf = _bt_getbuf(rel, heaprel, BTREE_METAPAGE, BT_WRITE);
metapg = BufferGetPage(metabuf);
metad = BTPageGetMeta(metapg);
diff --git a/src/backend/access/nbtree/nbtpage.c b/src/backend/access/nbtree/nbtpage.c
index 3feee28d19..151ad37a54 100644
--- a/src/backend/access/nbtree/nbtpage.c
+++ b/src/backend/access/nbtree/nbtpage.c
@@ -38,25 +38,24 @@
#include "utils/snapmgr.h"
static BTMetaPageData *_bt_getmeta(Relation rel, Buffer metabuf);
-static void _bt_log_reuse_page(Relation rel, BlockNumber blkno,
+static void _bt_log_reuse_page(Relation rel, Relation heaprel, BlockNumber blkno,
FullTransactionId safexid);
-static void _bt_delitems_delete(Relation rel, Buffer buf,
+static void _bt_delitems_delete(Relation rel, Relation heaprel, Buffer buf,
TransactionId snapshotConflictHorizon,
OffsetNumber *deletable, int ndeletable,
BTVacuumPosting *updatable, int nupdatable);
static char *_bt_delitems_update(BTVacuumPosting *updatable, int nupdatable,
OffsetNumber *updatedoffsets,
Size *updatedbuflen, bool needswal);
-static bool _bt_mark_page_halfdead(Relation rel, Buffer leafbuf,
- BTStack stack);
+static bool _bt_mark_page_halfdead(Relation rel, Relation heaprel,
+ Buffer leafbuf, BTStack stack);
static bool _bt_unlink_halfdead_page(Relation rel, Buffer leafbuf,
BlockNumber scanblkno,
bool *rightsib_empty,
BTVacState *vstate);
-static bool _bt_lock_subtree_parent(Relation rel, BlockNumber child,
- BTStack stack,
- Buffer *subtreeparent,
- OffsetNumber *poffset,
+static bool _bt_lock_subtree_parent(Relation rel, Relation heaprel,
+ BlockNumber child, BTStack stack,
+ Buffer *subtreeparent, OffsetNumber *poffset,
BlockNumber *topparent,
BlockNumber *topparentrightsib);
static void _bt_pendingfsm_add(BTVacState *vstate, BlockNumber target,
@@ -178,7 +177,7 @@ _bt_getmeta(Relation rel, Buffer metabuf)
* index tuples needed to be deleted.
*/
bool
-_bt_vacuum_needs_cleanup(Relation rel)
+_bt_vacuum_needs_cleanup(Relation rel, Relation heaprel)
{
Buffer metabuf;
Page metapg;
@@ -191,7 +190,7 @@ _bt_vacuum_needs_cleanup(Relation rel)
*
* Note that we deliberately avoid using cached version of metapage here.
*/
- metabuf = _bt_getbuf(rel, BTREE_METAPAGE, BT_READ);
+ metabuf = _bt_getbuf(rel, heaprel, BTREE_METAPAGE, BT_READ);
metapg = BufferGetPage(metabuf);
metad = BTPageGetMeta(metapg);
btm_version = metad->btm_version;
@@ -231,7 +230,7 @@ _bt_vacuum_needs_cleanup(Relation rel)
* finalized.
*/
void
-_bt_set_cleanup_info(Relation rel, BlockNumber num_delpages)
+_bt_set_cleanup_info(Relation rel, Relation heaprel, BlockNumber num_delpages)
{
Buffer metabuf;
Page metapg;
@@ -255,7 +254,7 @@ _bt_set_cleanup_info(Relation rel, BlockNumber num_delpages)
* no longer used as of PostgreSQL 14. We set it to -1.0 on rewrite, just
* to be consistent.
*/
- metabuf = _bt_getbuf(rel, BTREE_METAPAGE, BT_READ);
+ metabuf = _bt_getbuf(rel, heaprel, BTREE_METAPAGE, BT_READ);
metapg = BufferGetPage(metabuf);
metad = BTPageGetMeta(metapg);
@@ -340,7 +339,7 @@ _bt_set_cleanup_info(Relation rel, BlockNumber num_delpages)
* The metadata page is not locked or pinned on exit.
*/
Buffer
-_bt_getroot(Relation rel, int access)
+_bt_getroot(Relation rel, Relation heaprel, int access)
{
Buffer metabuf;
Buffer rootbuf;
@@ -370,7 +369,7 @@ _bt_getroot(Relation rel, int access)
Assert(rootblkno != P_NONE);
rootlevel = metad->btm_fastlevel;
- rootbuf = _bt_getbuf(rel, rootblkno, BT_READ);
+ rootbuf = _bt_getbuf(rel, heaprel, rootblkno, BT_READ);
rootpage = BufferGetPage(rootbuf);
rootopaque = BTPageGetOpaque(rootpage);
@@ -396,7 +395,7 @@ _bt_getroot(Relation rel, int access)
rel->rd_amcache = NULL;
}
- metabuf = _bt_getbuf(rel, BTREE_METAPAGE, BT_READ);
+ metabuf = _bt_getbuf(rel, heaprel, BTREE_METAPAGE, BT_READ);
metad = _bt_getmeta(rel, metabuf);
/* if no root page initialized yet, do it */
@@ -429,7 +428,7 @@ _bt_getroot(Relation rel, int access)
* to optimize this case.)
*/
_bt_relbuf(rel, metabuf);
- return _bt_getroot(rel, access);
+ return _bt_getroot(rel, heaprel, access);
}
/*
@@ -437,7 +436,7 @@ _bt_getroot(Relation rel, int access)
* the new root page. Since this is the first page in the tree, it's
* a leaf as well as the root.
*/
- rootbuf = _bt_getbuf(rel, P_NEW, BT_WRITE);
+ rootbuf = _bt_getbuf(rel, heaprel, P_NEW, BT_WRITE);
rootblkno = BufferGetBlockNumber(rootbuf);
rootpage = BufferGetPage(rootbuf);
rootopaque = BTPageGetOpaque(rootpage);
@@ -574,7 +573,7 @@ _bt_getroot(Relation rel, int access)
* moving to the root --- that'd deadlock against any concurrent root split.)
*/
Buffer
-_bt_gettrueroot(Relation rel)
+_bt_gettrueroot(Relation rel, Relation heaprel)
{
Buffer metabuf;
Page metapg;
@@ -596,7 +595,7 @@ _bt_gettrueroot(Relation rel)
pfree(rel->rd_amcache);
rel->rd_amcache = NULL;
- metabuf = _bt_getbuf(rel, BTREE_METAPAGE, BT_READ);
+ metabuf = _bt_getbuf(rel, heaprel, BTREE_METAPAGE, BT_READ);
metapg = BufferGetPage(metabuf);
metaopaque = BTPageGetOpaque(metapg);
metad = BTPageGetMeta(metapg);
@@ -669,7 +668,7 @@ _bt_gettrueroot(Relation rel)
* about updating previously cached data.
*/
int
-_bt_getrootheight(Relation rel)
+_bt_getrootheight(Relation rel, Relation heaprel)
{
BTMetaPageData *metad;
@@ -677,7 +676,7 @@ _bt_getrootheight(Relation rel)
{
Buffer metabuf;
- metabuf = _bt_getbuf(rel, BTREE_METAPAGE, BT_READ);
+ metabuf = _bt_getbuf(rel, heaprel, BTREE_METAPAGE, BT_READ);
metad = _bt_getmeta(rel, metabuf);
/*
@@ -733,7 +732,7 @@ _bt_getrootheight(Relation rel)
* pg_upgrade'd from Postgres 12.
*/
void
-_bt_metaversion(Relation rel, bool *heapkeyspace, bool *allequalimage)
+_bt_metaversion(Relation rel, Relation heaprel, bool *heapkeyspace, bool *allequalimage)
{
BTMetaPageData *metad;
@@ -741,7 +740,7 @@ _bt_metaversion(Relation rel, bool *heapkeyspace, bool *allequalimage)
{
Buffer metabuf;
- metabuf = _bt_getbuf(rel, BTREE_METAPAGE, BT_READ);
+ metabuf = _bt_getbuf(rel, heaprel, BTREE_METAPAGE, BT_READ);
metad = _bt_getmeta(rel, metabuf);
/*
@@ -825,7 +824,8 @@ _bt_checkpage(Relation rel, Buffer buf)
* Log the reuse of a page from the FSM.
*/
static void
-_bt_log_reuse_page(Relation rel, BlockNumber blkno, FullTransactionId safexid)
+_bt_log_reuse_page(Relation rel, Relation heaprel, BlockNumber blkno,
+ FullTransactionId safexid)
{
xl_btree_reuse_page xlrec_reuse;
@@ -836,6 +836,7 @@ _bt_log_reuse_page(Relation rel, BlockNumber blkno, FullTransactionId safexid)
*/
/* XLOG stuff */
+ xlrec_reuse.isCatalogRel = RelationIsAccessibleInLogicalDecoding(heaprel);
xlrec_reuse.locator = rel->rd_locator;
xlrec_reuse.block = blkno;
xlrec_reuse.snapshotConflictHorizon = safexid;
@@ -868,7 +869,7 @@ _bt_log_reuse_page(Relation rel, BlockNumber blkno, FullTransactionId safexid)
* as _bt_lockbuf().
*/
Buffer
-_bt_getbuf(Relation rel, BlockNumber blkno, int access)
+_bt_getbuf(Relation rel, Relation heaprel, BlockNumber blkno, int access)
{
Buffer buf;
@@ -943,7 +944,7 @@ _bt_getbuf(Relation rel, BlockNumber blkno, int access)
* than safexid value
*/
if (XLogStandbyInfoActive() && RelationNeedsWAL(rel))
- _bt_log_reuse_page(rel, blkno,
+ _bt_log_reuse_page(rel, heaprel, blkno,
BTPageGetDeleteXid(page));
/* Okay to use page. Re-initialize and return it. */
@@ -1293,7 +1294,7 @@ _bt_delitems_vacuum(Relation rel, Buffer buf,
* clear page's VACUUM cycle ID.
*/
static void
-_bt_delitems_delete(Relation rel, Buffer buf,
+_bt_delitems_delete(Relation rel, Relation heaprel, Buffer buf,
TransactionId snapshotConflictHorizon,
OffsetNumber *deletable, int ndeletable,
BTVacuumPosting *updatable, int nupdatable)
@@ -1358,6 +1359,7 @@ _bt_delitems_delete(Relation rel, Buffer buf,
XLogRecPtr recptr;
xl_btree_delete xlrec_delete;
+ xlrec_delete.isCatalogRel = RelationIsAccessibleInLogicalDecoding(heaprel);
xlrec_delete.snapshotConflictHorizon = snapshotConflictHorizon;
xlrec_delete.ndeleted = ndeletable;
xlrec_delete.nupdated = nupdatable;
@@ -1684,8 +1686,8 @@ _bt_delitems_delete_check(Relation rel, Buffer buf, Relation heapRel,
}
/* Physically delete tuples (or TIDs) using deletable (or updatable) */
- _bt_delitems_delete(rel, buf, snapshotConflictHorizon,
- deletable, ndeletable, updatable, nupdatable);
+ _bt_delitems_delete(rel, heapRel, buf, snapshotConflictHorizon, deletable,
+ ndeletable, updatable, nupdatable);
/* be tidy */
for (int i = 0; i < nupdatable; i++)
@@ -1706,7 +1708,8 @@ _bt_delitems_delete_check(Relation rel, Buffer buf, Relation heapRel,
* same level must always be locked left to right to avoid deadlocks.
*/
static bool
-_bt_leftsib_splitflag(Relation rel, BlockNumber leftsib, BlockNumber target)
+_bt_leftsib_splitflag(Relation rel, Relation heaprel, BlockNumber leftsib,
+ BlockNumber target)
{
Buffer buf;
Page page;
@@ -1717,7 +1720,7 @@ _bt_leftsib_splitflag(Relation rel, BlockNumber leftsib, BlockNumber target)
if (leftsib == P_NONE)
return false;
- buf = _bt_getbuf(rel, leftsib, BT_READ);
+ buf = _bt_getbuf(rel, heaprel, leftsib, BT_READ);
page = BufferGetPage(buf);
opaque = BTPageGetOpaque(page);
@@ -1763,7 +1766,7 @@ _bt_leftsib_splitflag(Relation rel, BlockNumber leftsib, BlockNumber target)
* to-be-deleted subtree.)
*/
static bool
-_bt_rightsib_halfdeadflag(Relation rel, BlockNumber leafrightsib)
+_bt_rightsib_halfdeadflag(Relation rel, Relation heaprel, BlockNumber leafrightsib)
{
Buffer buf;
Page page;
@@ -1772,7 +1775,7 @@ _bt_rightsib_halfdeadflag(Relation rel, BlockNumber leafrightsib)
Assert(leafrightsib != P_NONE);
- buf = _bt_getbuf(rel, leafrightsib, BT_READ);
+ buf = _bt_getbuf(rel, heaprel, leafrightsib, BT_READ);
page = BufferGetPage(buf);
opaque = BTPageGetOpaque(page);
@@ -1961,17 +1964,18 @@ _bt_pagedel(Relation rel, Buffer leafbuf, BTVacState *vstate)
* marked with INCOMPLETE_SPLIT flag before proceeding
*/
Assert(leafblkno == scanblkno);
- if (_bt_leftsib_splitflag(rel, leftsib, leafblkno))
+ if (_bt_leftsib_splitflag(rel, vstate->info->heaprel, leftsib, leafblkno))
{
ReleaseBuffer(leafbuf);
return;
}
/* we need an insertion scan key for the search, so build one */
- itup_key = _bt_mkscankey(rel, targetkey);
+ itup_key = _bt_mkscankey(rel, vstate->info->heaprel, targetkey);
/* find the leftmost leaf page with matching pivot/high key */
itup_key->pivotsearch = true;
- stack = _bt_search(rel, itup_key, &sleafbuf, BT_READ, NULL);
+ stack = _bt_search(rel, vstate->info->heaprel, itup_key,
+ &sleafbuf, BT_READ, NULL);
/* won't need a second lock or pin on leafbuf */
_bt_relbuf(rel, sleafbuf);
@@ -2002,7 +2006,7 @@ _bt_pagedel(Relation rel, Buffer leafbuf, BTVacState *vstate)
* leafbuf page half-dead.
*/
Assert(P_ISLEAF(opaque) && !P_IGNORE(opaque));
- if (!_bt_mark_page_halfdead(rel, leafbuf, stack))
+ if (!_bt_mark_page_halfdead(rel, vstate->info->heaprel, leafbuf, stack))
{
_bt_relbuf(rel, leafbuf);
return;
@@ -2065,7 +2069,7 @@ _bt_pagedel(Relation rel, Buffer leafbuf, BTVacState *vstate)
if (!rightsib_empty)
break;
- leafbuf = _bt_getbuf(rel, rightsib, BT_WRITE);
+ leafbuf = _bt_getbuf(rel, vstate->info->heaprel, rightsib, BT_WRITE);
}
}
@@ -2084,7 +2088,8 @@ _bt_pagedel(Relation rel, Buffer leafbuf, BTVacState *vstate)
* successfully.
*/
static bool
-_bt_mark_page_halfdead(Relation rel, Buffer leafbuf, BTStack stack)
+_bt_mark_page_halfdead(Relation rel, Relation heaprel, Buffer leafbuf,
+ BTStack stack)
{
BlockNumber leafblkno;
BlockNumber leafrightsib;
@@ -2119,7 +2124,7 @@ _bt_mark_page_halfdead(Relation rel, Buffer leafbuf, BTStack stack)
* delete the downlink. It would fail the "right sibling of target page
* is also the next child in parent page" cross-check below.
*/
- if (_bt_rightsib_halfdeadflag(rel, leafrightsib))
+ if (_bt_rightsib_halfdeadflag(rel, heaprel, leafrightsib))
{
elog(DEBUG1, "could not delete page %u because its right sibling %u is half-dead",
leafblkno, leafrightsib);
@@ -2143,7 +2148,7 @@ _bt_mark_page_halfdead(Relation rel, Buffer leafbuf, BTStack stack)
*/
topparent = leafblkno;
topparentrightsib = leafrightsib;
- if (!_bt_lock_subtree_parent(rel, leafblkno, stack,
+ if (!_bt_lock_subtree_parent(rel, heaprel, leafblkno, stack,
&subtreeparent, &poffset,
&topparent, &topparentrightsib))
return false;
@@ -2363,7 +2368,7 @@ _bt_unlink_halfdead_page(Relation rel, Buffer leafbuf, BlockNumber scanblkno,
Assert(target != leafblkno);
/* Fetch the block number of the target's left sibling */
- buf = _bt_getbuf(rel, target, BT_READ);
+ buf = _bt_getbuf(rel, vstate->info->heaprel, target, BT_READ);
page = BufferGetPage(buf);
opaque = BTPageGetOpaque(page);
leftsib = opaque->btpo_prev;
@@ -2390,7 +2395,7 @@ _bt_unlink_halfdead_page(Relation rel, Buffer leafbuf, BlockNumber scanblkno,
_bt_lockbuf(rel, leafbuf, BT_WRITE);
if (leftsib != P_NONE)
{
- lbuf = _bt_getbuf(rel, leftsib, BT_WRITE);
+ lbuf = _bt_getbuf(rel, vstate->info->heaprel, leftsib, BT_WRITE);
page = BufferGetPage(lbuf);
opaque = BTPageGetOpaque(page);
while (P_ISDELETED(opaque) || opaque->btpo_next != target)
@@ -2440,7 +2445,7 @@ _bt_unlink_halfdead_page(Relation rel, Buffer leafbuf, BlockNumber scanblkno,
CHECK_FOR_INTERRUPTS();
/* step right one page */
- lbuf = _bt_getbuf(rel, leftsib, BT_WRITE);
+ lbuf = _bt_getbuf(rel, vstate->info->heaprel, leftsib, BT_WRITE);
page = BufferGetPage(lbuf);
opaque = BTPageGetOpaque(page);
}
@@ -2504,7 +2509,7 @@ _bt_unlink_halfdead_page(Relation rel, Buffer leafbuf, BlockNumber scanblkno,
* And next write-lock the (current) right sibling.
*/
rightsib = opaque->btpo_next;
- rbuf = _bt_getbuf(rel, rightsib, BT_WRITE);
+ rbuf = _bt_getbuf(rel, vstate->info->heaprel, rightsib, BT_WRITE);
page = BufferGetPage(rbuf);
opaque = BTPageGetOpaque(page);
if (opaque->btpo_prev != target)
@@ -2533,7 +2538,8 @@ _bt_unlink_halfdead_page(Relation rel, Buffer leafbuf, BlockNumber scanblkno,
if (P_RIGHTMOST(opaque))
{
/* rightsib will be the only one left on the level */
- metabuf = _bt_getbuf(rel, BTREE_METAPAGE, BT_WRITE);
+ metabuf = _bt_getbuf(rel, vstate->info->heaprel, BTREE_METAPAGE,
+ BT_WRITE);
metapg = BufferGetPage(metabuf);
metad = BTPageGetMeta(metapg);
@@ -2773,9 +2779,10 @@ _bt_unlink_halfdead_page(Relation rel, Buffer leafbuf, BlockNumber scanblkno,
* parent block in the leafbuf page using BTreeTupleSetTopParent()).
*/
static bool
-_bt_lock_subtree_parent(Relation rel, BlockNumber child, BTStack stack,
- Buffer *subtreeparent, OffsetNumber *poffset,
- BlockNumber *topparent, BlockNumber *topparentrightsib)
+_bt_lock_subtree_parent(Relation rel, Relation heaprel, BlockNumber child,
+ BTStack stack, Buffer *subtreeparent,
+ OffsetNumber *poffset, BlockNumber *topparent,
+ BlockNumber *topparentrightsib)
{
BlockNumber parent,
leftsibparent;
@@ -2789,7 +2796,7 @@ _bt_lock_subtree_parent(Relation rel, BlockNumber child, BTStack stack,
* Locate the pivot tuple whose downlink points to "child". Write lock
* the parent page itself.
*/
- pbuf = _bt_getstackbuf(rel, stack, child);
+ pbuf = _bt_getstackbuf(rel, heaprel, stack, child);
if (pbuf == InvalidBuffer)
{
/*
@@ -2889,11 +2896,11 @@ _bt_lock_subtree_parent(Relation rel, BlockNumber child, BTStack stack,
*
* Note: We deliberately avoid completing incomplete splits here.
*/
- if (_bt_leftsib_splitflag(rel, leftsibparent, parent))
+ if (_bt_leftsib_splitflag(rel, heaprel, leftsibparent, parent))
return false;
/* Recurse to examine child page's grandparent page */
- return _bt_lock_subtree_parent(rel, parent, stack->bts_parent,
+ return _bt_lock_subtree_parent(rel, heaprel, parent, stack->bts_parent,
subtreeparent, poffset,
topparent, topparentrightsib);
}
diff --git a/src/backend/access/nbtree/nbtree.c b/src/backend/access/nbtree/nbtree.c
index 1cc88da032..4e8a85fb5d 100644
--- a/src/backend/access/nbtree/nbtree.c
+++ b/src/backend/access/nbtree/nbtree.c
@@ -834,7 +834,7 @@ btvacuumcleanup(IndexVacuumInfo *info, IndexBulkDeleteResult *stats)
if (stats == NULL)
{
/* Check if VACUUM operation can entirely avoid btvacuumscan() call */
- if (!_bt_vacuum_needs_cleanup(info->index))
+ if (!_bt_vacuum_needs_cleanup(info->index, info->heaprel))
return NULL;
/*
@@ -870,7 +870,7 @@ btvacuumcleanup(IndexVacuumInfo *info, IndexBulkDeleteResult *stats)
*/
Assert(stats->pages_deleted >= stats->pages_free);
num_delpages = stats->pages_deleted - stats->pages_free;
- _bt_set_cleanup_info(info->index, num_delpages);
+ _bt_set_cleanup_info(info->index, info->heaprel, num_delpages);
/*
* It's quite possible for us to be fooled by concurrent page splits into
diff --git a/src/backend/access/nbtree/nbtsearch.c b/src/backend/access/nbtree/nbtsearch.c
index c43c1a2830..5c728e353d 100644
--- a/src/backend/access/nbtree/nbtsearch.c
+++ b/src/backend/access/nbtree/nbtsearch.c
@@ -42,7 +42,8 @@ static bool _bt_steppage(IndexScanDesc scan, ScanDirection dir);
static bool _bt_readnextpage(IndexScanDesc scan, BlockNumber blkno, ScanDirection dir);
static bool _bt_parallel_readpage(IndexScanDesc scan, BlockNumber blkno,
ScanDirection dir);
-static Buffer _bt_walk_left(Relation rel, Buffer buf, Snapshot snapshot);
+static Buffer _bt_walk_left(Relation rel, Relation heaprel, Buffer buf,
+ Snapshot snapshot);
static bool _bt_endpoint(IndexScanDesc scan, ScanDirection dir);
static inline void _bt_initialize_more_data(BTScanOpaque so, ScanDirection dir);
@@ -93,14 +94,14 @@ _bt_drop_lock_and_maybe_pin(IndexScanDesc scan, BTScanPos sp)
* during the search will be finished.
*/
BTStack
-_bt_search(Relation rel, BTScanInsert key, Buffer *bufP, int access,
- Snapshot snapshot)
+_bt_search(Relation rel, Relation heaprel, BTScanInsert key, Buffer *bufP,
+ int access, Snapshot snapshot)
{
BTStack stack_in = NULL;
int page_access = BT_READ;
/* Get the root page to start with */
- *bufP = _bt_getroot(rel, access);
+ *bufP = _bt_getroot(rel, heaprel, access);
/* If index is empty and access = BT_READ, no root page is created. */
if (!BufferIsValid(*bufP))
@@ -129,8 +130,8 @@ _bt_search(Relation rel, BTScanInsert key, Buffer *bufP, int access,
* also taken care of in _bt_getstackbuf). But this is a good
* opportunity to finish splits of internal pages too.
*/
- *bufP = _bt_moveright(rel, key, *bufP, (access == BT_WRITE), stack_in,
- page_access, snapshot);
+ *bufP = _bt_moveright(rel, heaprel, key, *bufP, (access == BT_WRITE),
+ stack_in, page_access, snapshot);
/* if this is a leaf page, we're done */
page = BufferGetPage(*bufP);
@@ -190,7 +191,7 @@ _bt_search(Relation rel, BTScanInsert key, Buffer *bufP, int access,
* but before we acquired a write lock. If it has, we may need to
* move right to its new sibling. Do that.
*/
- *bufP = _bt_moveright(rel, key, *bufP, true, stack_in, BT_WRITE,
+ *bufP = _bt_moveright(rel, heaprel, key, *bufP, true, stack_in, BT_WRITE,
snapshot);
}
@@ -234,6 +235,7 @@ _bt_search(Relation rel, BTScanInsert key, Buffer *bufP, int access,
*/
Buffer
_bt_moveright(Relation rel,
+ Relation heaprel,
BTScanInsert key,
Buffer buf,
bool forupdate,
@@ -288,12 +290,12 @@ _bt_moveright(Relation rel,
}
if (P_INCOMPLETE_SPLIT(opaque))
- _bt_finish_split(rel, buf, stack);
+ _bt_finish_split(rel, heaprel, buf, stack);
else
_bt_relbuf(rel, buf);
/* re-acquire the lock in the right mode, and re-check */
- buf = _bt_getbuf(rel, blkno, access);
+ buf = _bt_getbuf(rel, heaprel, blkno, access);
continue;
}
@@ -860,6 +862,7 @@ bool
_bt_first(IndexScanDesc scan, ScanDirection dir)
{
Relation rel = scan->indexRelation;
+ Relation heaprel = scan->heapRelation;
BTScanOpaque so = (BTScanOpaque) scan->opaque;
Buffer buf;
BTStack stack;
@@ -1352,7 +1355,7 @@ _bt_first(IndexScanDesc scan, ScanDirection dir)
}
/* Initialize remaining insertion scan key fields */
- _bt_metaversion(rel, &inskey.heapkeyspace, &inskey.allequalimage);
+ _bt_metaversion(rel, heaprel, &inskey.heapkeyspace, &inskey.allequalimage);
inskey.anynullkeys = false; /* unused */
inskey.nextkey = nextkey;
inskey.pivotsearch = false;
@@ -1363,7 +1366,7 @@ _bt_first(IndexScanDesc scan, ScanDirection dir)
* Use the manufactured insertion scan key to descend the tree and
* position ourselves on the target leaf page.
*/
- stack = _bt_search(rel, &inskey, &buf, BT_READ, scan->xs_snapshot);
+ stack = _bt_search(rel, heaprel, &inskey, &buf, BT_READ, scan->xs_snapshot);
/* don't need to keep the stack around... */
_bt_freestack(stack);
@@ -2004,7 +2007,7 @@ _bt_readnextpage(IndexScanDesc scan, BlockNumber blkno, ScanDirection dir)
/* check for interrupts while we're not holding any buffer lock */
CHECK_FOR_INTERRUPTS();
/* step right one page */
- so->currPos.buf = _bt_getbuf(rel, blkno, BT_READ);
+ so->currPos.buf = _bt_getbuf(rel, scan->heapRelation, blkno, BT_READ);
page = BufferGetPage(so->currPos.buf);
TestForOldSnapshot(scan->xs_snapshot, rel, page);
opaque = BTPageGetOpaque(page);
@@ -2078,7 +2081,8 @@ _bt_readnextpage(IndexScanDesc scan, BlockNumber blkno, ScanDirection dir)
if (BTScanPosIsPinned(so->currPos))
_bt_lockbuf(rel, so->currPos.buf, BT_READ);
else
- so->currPos.buf = _bt_getbuf(rel, so->currPos.currPage, BT_READ);
+ so->currPos.buf = _bt_getbuf(rel, scan->heapRelation,
+ so->currPos.currPage, BT_READ);
for (;;)
{
@@ -2092,8 +2096,8 @@ _bt_readnextpage(IndexScanDesc scan, BlockNumber blkno, ScanDirection dir)
}
/* Step to next physical page */
- so->currPos.buf = _bt_walk_left(rel, so->currPos.buf,
- scan->xs_snapshot);
+ so->currPos.buf = _bt_walk_left(rel, scan->heapRelation,
+ so->currPos.buf, scan->xs_snapshot);
/* if we're physically at end of index, return failure */
if (so->currPos.buf == InvalidBuffer)
@@ -2140,7 +2144,8 @@ _bt_readnextpage(IndexScanDesc scan, BlockNumber blkno, ScanDirection dir)
BTScanPosInvalidate(so->currPos);
return false;
}
- so->currPos.buf = _bt_getbuf(rel, blkno, BT_READ);
+ so->currPos.buf = _bt_getbuf(rel, scan->heapRelation, blkno,
+ BT_READ);
}
}
}
@@ -2185,7 +2190,7 @@ _bt_parallel_readpage(IndexScanDesc scan, BlockNumber blkno, ScanDirection dir)
* again if it's important.
*/
static Buffer
-_bt_walk_left(Relation rel, Buffer buf, Snapshot snapshot)
+_bt_walk_left(Relation rel, Relation heaprel, Buffer buf, Snapshot snapshot)
{
Page page;
BTPageOpaque opaque;
@@ -2213,7 +2218,7 @@ _bt_walk_left(Relation rel, Buffer buf, Snapshot snapshot)
_bt_relbuf(rel, buf);
/* check for interrupts while we're not holding any buffer lock */
CHECK_FOR_INTERRUPTS();
- buf = _bt_getbuf(rel, blkno, BT_READ);
+ buf = _bt_getbuf(rel, heaprel, blkno, BT_READ);
page = BufferGetPage(buf);
TestForOldSnapshot(snapshot, rel, page);
opaque = BTPageGetOpaque(page);
@@ -2304,7 +2309,7 @@ _bt_walk_left(Relation rel, Buffer buf, Snapshot snapshot)
* The returned buffer is pinned and read-locked.
*/
Buffer
-_bt_get_endpoint(Relation rel, uint32 level, bool rightmost,
+_bt_get_endpoint(Relation rel, Relation heaprel, uint32 level, bool rightmost,
Snapshot snapshot)
{
Buffer buf;
@@ -2320,9 +2325,9 @@ _bt_get_endpoint(Relation rel, uint32 level, bool rightmost,
* smarter about intermediate levels.)
*/
if (level == 0)
- buf = _bt_getroot(rel, BT_READ);
+ buf = _bt_getroot(rel, heaprel, BT_READ);
else
- buf = _bt_gettrueroot(rel);
+ buf = _bt_gettrueroot(rel, heaprel);
if (!BufferIsValid(buf))
return InvalidBuffer;
@@ -2403,7 +2408,8 @@ _bt_endpoint(IndexScanDesc scan, ScanDirection dir)
* version of _bt_search(). We don't maintain a stack since we know we
* won't need it.
*/
- buf = _bt_get_endpoint(rel, 0, ScanDirectionIsBackward(dir), scan->xs_snapshot);
+ buf = _bt_get_endpoint(rel, scan->heapRelation, 0,
+ ScanDirectionIsBackward(dir), scan->xs_snapshot);
if (!BufferIsValid(buf))
{
diff --git a/src/backend/access/nbtree/nbtsort.c b/src/backend/access/nbtree/nbtsort.c
index 67b7b1710c..8c58fdb8d1 100644
--- a/src/backend/access/nbtree/nbtsort.c
+++ b/src/backend/access/nbtree/nbtsort.c
@@ -566,7 +566,7 @@ _bt_leafbuild(BTSpool *btspool, BTSpool *btspool2)
wstate.heap = btspool->heap;
wstate.index = btspool->index;
- wstate.inskey = _bt_mkscankey(wstate.index, NULL);
+ wstate.inskey = _bt_mkscankey(wstate.index, btspool->heap, NULL);
/* _bt_mkscankey() won't set allequalimage without metapage */
wstate.inskey->allequalimage = _bt_allequalimage(wstate.index, true);
wstate.btws_use_wal = RelationNeedsWAL(wstate.index);
diff --git a/src/backend/access/nbtree/nbtutils.c b/src/backend/access/nbtree/nbtutils.c
index 8003583c0a..70a0c2418a 100644
--- a/src/backend/access/nbtree/nbtutils.c
+++ b/src/backend/access/nbtree/nbtutils.c
@@ -87,7 +87,7 @@ static int _bt_keep_natts(Relation rel, IndexTuple lastleft,
* field themselves.
*/
BTScanInsert
-_bt_mkscankey(Relation rel, IndexTuple itup)
+_bt_mkscankey(Relation rel, Relation heaprel, IndexTuple itup)
{
BTScanInsert key;
ScanKey skey;
@@ -112,7 +112,7 @@ _bt_mkscankey(Relation rel, IndexTuple itup)
key = palloc(offsetof(BTScanInsertData, scankeys) +
sizeof(ScanKeyData) * indnkeyatts);
if (itup)
- _bt_metaversion(rel, &key->heapkeyspace, &key->allequalimage);
+ _bt_metaversion(rel, heaprel, &key->heapkeyspace, &key->allequalimage);
else
{
/* Utility statement callers can set these fields themselves */
@@ -1761,7 +1761,8 @@ _bt_killitems(IndexScanDesc scan)
droppedpin = true;
/* Attempt to re-read the buffer, getting pin and lock. */
- buf = _bt_getbuf(scan->indexRelation, so->currPos.currPage, BT_READ);
+ buf = _bt_getbuf(scan->indexRelation, scan->heapRelation,
+ so->currPos.currPage, BT_READ);
page = BufferGetPage(buf);
if (BufferGetLSNAtomic(buf) == so->currPos.lsn)
diff --git a/src/backend/access/spgist/spgvacuum.c b/src/backend/access/spgist/spgvacuum.c
index 3adb18f2d8..2f4a4aad24 100644
--- a/src/backend/access/spgist/spgvacuum.c
+++ b/src/backend/access/spgist/spgvacuum.c
@@ -489,7 +489,7 @@ vacuumLeafRoot(spgBulkDeleteState *bds, Relation index, Buffer buffer)
* Unlike the routines above, this works on both leaf and inner pages.
*/
static void
-vacuumRedirectAndPlaceholder(Relation index, Buffer buffer)
+vacuumRedirectAndPlaceholder(Relation index, Relation heaprel, Buffer buffer)
{
Page page = BufferGetPage(buffer);
SpGistPageOpaque opaque = SpGistPageGetOpaque(page);
@@ -503,6 +503,7 @@ vacuumRedirectAndPlaceholder(Relation index, Buffer buffer)
spgxlogVacuumRedirect xlrec;
GlobalVisState *vistest;
+ xlrec.isCatalogRel = RelationIsAccessibleInLogicalDecoding(heaprel);
xlrec.nToPlaceholder = 0;
xlrec.snapshotConflictHorizon = InvalidTransactionId;
@@ -643,13 +644,13 @@ spgvacuumpage(spgBulkDeleteState *bds, BlockNumber blkno)
else
{
vacuumLeafPage(bds, index, buffer, false);
- vacuumRedirectAndPlaceholder(index, buffer);
+ vacuumRedirectAndPlaceholder(index, bds->info->heaprel, buffer);
}
}
else
{
/* inner page */
- vacuumRedirectAndPlaceholder(index, buffer);
+ vacuumRedirectAndPlaceholder(index, bds->info->heaprel, buffer);
}
/*
@@ -719,7 +720,7 @@ spgprocesspending(spgBulkDeleteState *bds)
/* deal with any deletable tuples */
vacuumLeafPage(bds, index, buffer, true);
/* might as well do this while we are here */
- vacuumRedirectAndPlaceholder(index, buffer);
+ vacuumRedirectAndPlaceholder(index, bds->info->heaprel, buffer);
SpGistSetLastUsedPage(index, buffer);
diff --git a/src/backend/catalog/index.c b/src/backend/catalog/index.c
index 41b16cb89b..48d1d6b506 100644
--- a/src/backend/catalog/index.c
+++ b/src/backend/catalog/index.c
@@ -3352,6 +3352,7 @@ validate_index(Oid heapId, Oid indexId, Snapshot snapshot)
ivinfo.message_level = DEBUG2;
ivinfo.num_heap_tuples = heapRelation->rd_rel->reltuples;
ivinfo.strategy = NULL;
+ ivinfo.heaprel = heapRelation;
/*
* Encode TIDs as int8 values for the sort, rather than directly sorting
diff --git a/src/backend/commands/analyze.c b/src/backend/commands/analyze.c
index c86e690980..321fc0d31b 100644
--- a/src/backend/commands/analyze.c
+++ b/src/backend/commands/analyze.c
@@ -712,6 +712,7 @@ do_analyze_rel(Relation onerel, VacuumParams *params,
ivinfo.message_level = elevel;
ivinfo.num_heap_tuples = onerel->rd_rel->reltuples;
ivinfo.strategy = vac_strategy;
+ ivinfo.heaprel = onerel;
stats = index_vacuum_cleanup(&ivinfo, NULL);
diff --git a/src/backend/commands/vacuumparallel.c b/src/backend/commands/vacuumparallel.c
index bcd40c80a1..2cdbd182b6 100644
--- a/src/backend/commands/vacuumparallel.c
+++ b/src/backend/commands/vacuumparallel.c
@@ -148,6 +148,9 @@ struct ParallelVacuumState
/* NULL for worker processes */
ParallelContext *pcxt;
+ /* Parent Heap Relation */
+ Relation heaprel;
+
/* Target indexes */
Relation *indrels;
int nindexes;
@@ -266,6 +269,7 @@ parallel_vacuum_init(Relation rel, Relation *indrels, int nindexes,
pvs->nindexes = nindexes;
pvs->will_parallel_vacuum = will_parallel_vacuum;
pvs->bstrategy = bstrategy;
+ pvs->heaprel = rel;
EnterParallelMode();
pcxt = CreateParallelContext("postgres", "parallel_vacuum_main",
@@ -838,6 +842,7 @@ parallel_vacuum_process_one_index(ParallelVacuumState *pvs, Relation indrel,
ivinfo.estimated_count = pvs->shared->estimated_count;
ivinfo.num_heap_tuples = pvs->shared->reltuples;
ivinfo.strategy = pvs->bstrategy;
+ ivinfo.heaprel = pvs->heaprel;
/* Update error traceback information */
pvs->indname = pstrdup(RelationGetRelationName(indrel));
@@ -1007,6 +1012,7 @@ parallel_vacuum_main(dsm_segment *seg, shm_toc *toc)
pvs.dead_items = dead_items;
pvs.relnamespace = get_namespace_name(RelationGetNamespace(rel));
pvs.relname = pstrdup(RelationGetRelationName(rel));
+ pvs.heaprel = rel;
/* These fields will be filled during index vacuum or cleanup */
pvs.indname = NULL;
diff --git a/src/backend/optimizer/util/plancat.c b/src/backend/optimizer/util/plancat.c
index d58c4a1078..e3824efe9b 100644
--- a/src/backend/optimizer/util/plancat.c
+++ b/src/backend/optimizer/util/plancat.c
@@ -462,7 +462,7 @@ get_relation_info(PlannerInfo *root, Oid relationObjectId, bool inhparent,
* For btrees, get tree height while we have the index
* open
*/
- info->tree_height = _bt_getrootheight(indexRelation);
+ info->tree_height = _bt_getrootheight(indexRelation, relation);
}
else
{
diff --git a/src/backend/utils/sort/tuplesortvariants.c b/src/backend/utils/sort/tuplesortvariants.c
index eb6cfcfd00..0188106925 100644
--- a/src/backend/utils/sort/tuplesortvariants.c
+++ b/src/backend/utils/sort/tuplesortvariants.c
@@ -207,6 +207,7 @@ tuplesort_begin_heap(TupleDesc tupDesc,
Tuplesortstate *
tuplesort_begin_cluster(TupleDesc tupDesc,
Relation indexRel,
+ Relation heaprel,
int workMem,
SortCoordinate coordinate, int sortopt)
{
@@ -260,7 +261,7 @@ tuplesort_begin_cluster(TupleDesc tupDesc,
arg->tupDesc = tupDesc; /* assume we need not copy tupDesc */
- indexScanKey = _bt_mkscankey(indexRel, NULL);
+ indexScanKey = _bt_mkscankey(indexRel, heaprel, NULL);
if (arg->indexInfo->ii_Expressions != NULL)
{
@@ -361,7 +362,7 @@ tuplesort_begin_index_btree(Relation heapRel,
arg->enforceUnique = enforceUnique;
arg->uniqueNullsNotDistinct = uniqueNullsNotDistinct;
- indexScanKey = _bt_mkscankey(indexRel, NULL);
+ indexScanKey = _bt_mkscankey(indexRel, heapRel, NULL);
/* Prepare SortSupport data for each column */
base->sortKeys = (SortSupport) palloc0(base->nKeys *
diff --git a/src/include/access/genam.h b/src/include/access/genam.h
index 83dbee0fe6..7708b82d7d 100644
--- a/src/include/access/genam.h
+++ b/src/include/access/genam.h
@@ -50,6 +50,7 @@ typedef struct IndexVacuumInfo
int message_level; /* ereport level for progress messages */
double num_heap_tuples; /* tuples remaining in heap */
BufferAccessStrategy strategy; /* access strategy for reads */
+ Relation heaprel; /* the heap relation the index belongs to */
} IndexVacuumInfo;
/*
diff --git a/src/include/access/gist_private.h b/src/include/access/gist_private.h
index 8af33d7b40..ee275650bd 100644
--- a/src/include/access/gist_private.h
+++ b/src/include/access/gist_private.h
@@ -440,7 +440,7 @@ extern XLogRecPtr gistXLogPageDelete(Buffer buffer,
FullTransactionId xid, Buffer parentBuffer,
OffsetNumber downlinkOffset);
-extern void gistXLogPageReuse(Relation rel, BlockNumber blkno,
+extern void gistXLogPageReuse(Relation rel, Relation heaprel, BlockNumber blkno,
FullTransactionId deleteXid);
extern XLogRecPtr gistXLogUpdate(Buffer buffer,
@@ -449,7 +449,8 @@ extern XLogRecPtr gistXLogUpdate(Buffer buffer,
Buffer leftchildbuf);
extern XLogRecPtr gistXLogDelete(Buffer buffer, OffsetNumber *todelete,
- int ntodelete, TransactionId snapshotConflictHorizon);
+ int ntodelete, TransactionId snapshotConflictHorizon,
+ Relation heaprel);
extern XLogRecPtr gistXLogSplit(bool page_is_leaf,
SplitedPageLayout *dist,
@@ -485,7 +486,7 @@ extern bool gistproperty(Oid index_oid, int attno,
extern bool gistfitpage(IndexTuple *itvec, int len);
extern bool gistnospace(Page page, IndexTuple *itvec, int len, OffsetNumber todelete, Size freespace);
extern void gistcheckpage(Relation rel, Buffer buf);
-extern Buffer gistNewBuffer(Relation r);
+extern Buffer gistNewBuffer(Relation r, Relation heaprel);
extern bool gistPageRecyclable(Page page);
extern void gistfillbuffer(Page page, IndexTuple *itup, int len,
OffsetNumber off);
diff --git a/src/include/access/gistxlog.h b/src/include/access/gistxlog.h
index 09f9b0f8c6..2eea866f06 100644
--- a/src/include/access/gistxlog.h
+++ b/src/include/access/gistxlog.h
@@ -51,13 +51,14 @@ typedef struct gistxlogDelete
{
TransactionId snapshotConflictHorizon;
uint16 ntodelete; /* number of deleted offsets */
+ bool isCatalogRel; /* to handle recovery conflict during logical
+ * decoding on standby */
- /*
- * In payload of blk 0 : todelete OffsetNumbers
- */
+ /* TODELETE OFFSET NUMBERS */
+ OffsetNumber offsets[FLEXIBLE_ARRAY_MEMBER];
} gistxlogDelete;
-#define SizeOfGistxlogDelete (offsetof(gistxlogDelete, ntodelete) + sizeof(uint16))
+#define SizeOfGistxlogDelete offsetof(gistxlogDelete, offsets)
/*
* Backup Blk 0: If this operation completes a page split, by inserting a
@@ -100,9 +101,11 @@ typedef struct gistxlogPageReuse
RelFileLocator locator;
BlockNumber block;
FullTransactionId snapshotConflictHorizon;
+ bool isCatalogRel; /* to handle recovery conflict during logical
+ * decoding on standby */
} gistxlogPageReuse;
-#define SizeOfGistxlogPageReuse (offsetof(gistxlogPageReuse, snapshotConflictHorizon) + sizeof(FullTransactionId))
+#define SizeOfGistxlogPageReuse (offsetof(gistxlogPageReuse, isCatalogRel) + sizeof(bool))
extern void gist_redo(XLogReaderState *record);
extern void gist_desc(StringInfo buf, XLogReaderState *record);
diff --git a/src/include/access/hash_xlog.h b/src/include/access/hash_xlog.h
index a2f0f39213..7e9e47ce67 100644
--- a/src/include/access/hash_xlog.h
+++ b/src/include/access/hash_xlog.h
@@ -252,12 +252,14 @@ typedef struct xl_hash_vacuum_one_page
{
TransactionId snapshotConflictHorizon;
int ntuples;
+ bool isCatalogRel; /* to handle recovery conflict during logical
+ * decoding on standby */
- /* TARGET OFFSET NUMBERS FOLLOW AT THE END */
+ /* TARGET OFFSET NUMBERS */
+ OffsetNumber offsets[FLEXIBLE_ARRAY_MEMBER];
} xl_hash_vacuum_one_page;
-#define SizeOfHashVacuumOnePage \
- (offsetof(xl_hash_vacuum_one_page, ntuples) + sizeof(int))
+#define SizeOfHashVacuumOnePage offsetof(xl_hash_vacuum_one_page, offsets)
extern void hash_redo(XLogReaderState *record);
extern void hash_desc(StringInfo buf, XLogReaderState *record);
diff --git a/src/include/access/heapam_xlog.h b/src/include/access/heapam_xlog.h
index 8cb0d8da19..223db4b199 100644
--- a/src/include/access/heapam_xlog.h
+++ b/src/include/access/heapam_xlog.h
@@ -245,10 +245,12 @@ typedef struct xl_heap_prune
TransactionId snapshotConflictHorizon;
uint16 nredirected;
uint16 ndead;
+ bool isCatalogRel; /* to handle recovery conflict during logical
+ * decoding on standby */
/* OFFSET NUMBERS are in the block reference 0 */
} xl_heap_prune;
-#define SizeOfHeapPrune (offsetof(xl_heap_prune, ndead) + sizeof(uint16))
+#define SizeOfHeapPrune (offsetof(xl_heap_prune, isCatalogRel) + sizeof(bool))
/*
* The vacuum page record is similar to the prune record, but can only mark
@@ -344,12 +346,14 @@ typedef struct xl_heap_freeze_page
{
TransactionId snapshotConflictHorizon;
uint16 nplans;
+ bool isCatalogRel; /* to handle recovery conflict during logical
+ * decoding on standby */
/* FREEZE PLANS FOLLOW */
/* OFFSET NUMBER ARRAY FOLLOWS */
} xl_heap_freeze_page;
-#define SizeOfHeapFreezePage (offsetof(xl_heap_freeze_page, nplans) + sizeof(uint16))
+#define SizeOfHeapFreezePage (offsetof(xl_heap_freeze_page, isCatalogRel) + sizeof(bool))
/*
* This is what we need to know about setting a visibility map bit
@@ -408,7 +412,7 @@ extern void heap2_desc(StringInfo buf, XLogReaderState *record);
extern const char *heap2_identify(uint8 info);
extern void heap_xlog_logical_rewrite(XLogReaderState *r);
-extern XLogRecPtr log_heap_visible(RelFileLocator rlocator, Buffer heap_buffer,
+extern XLogRecPtr log_heap_visible(Relation rel, Buffer heap_buffer,
Buffer vm_buffer,
TransactionId snapshotConflictHorizon,
uint8 vmflags);
diff --git a/src/include/access/nbtree.h b/src/include/access/nbtree.h
index 8f48960f9d..6dee307042 100644
--- a/src/include/access/nbtree.h
+++ b/src/include/access/nbtree.h
@@ -1182,8 +1182,10 @@ extern IndexTuple _bt_swap_posting(IndexTuple newitem, IndexTuple oposting,
extern bool _bt_doinsert(Relation rel, IndexTuple itup,
IndexUniqueCheck checkUnique, bool indexUnchanged,
Relation heapRel);
-extern void _bt_finish_split(Relation rel, Buffer lbuf, BTStack stack);
-extern Buffer _bt_getstackbuf(Relation rel, BTStack stack, BlockNumber child);
+extern void _bt_finish_split(Relation rel, Relation heaprel, Buffer lbuf,
+ BTStack stack);
+extern Buffer _bt_getstackbuf(Relation rel, Relation heaprel, BTStack stack,
+ BlockNumber child);
/*
* prototypes for functions in nbtsplitloc.c
@@ -1197,16 +1199,18 @@ extern OffsetNumber _bt_findsplitloc(Relation rel, Page origpage,
*/
extern void _bt_initmetapage(Page page, BlockNumber rootbknum, uint32 level,
bool allequalimage);
-extern bool _bt_vacuum_needs_cleanup(Relation rel);
-extern void _bt_set_cleanup_info(Relation rel, BlockNumber num_delpages);
+extern bool _bt_vacuum_needs_cleanup(Relation rel, Relation heaprel);
+extern void _bt_set_cleanup_info(Relation rel, Relation heaprel,
+ BlockNumber num_delpages);
extern void _bt_upgrademetapage(Page page);
-extern Buffer _bt_getroot(Relation rel, int access);
-extern Buffer _bt_gettrueroot(Relation rel);
-extern int _bt_getrootheight(Relation rel);
-extern void _bt_metaversion(Relation rel, bool *heapkeyspace,
+extern Buffer _bt_getroot(Relation rel, Relation heaprel, int access);
+extern Buffer _bt_gettrueroot(Relation rel, Relation heaprel);
+extern int _bt_getrootheight(Relation rel, Relation heaprel);
+extern void _bt_metaversion(Relation rel, Relation heaprel, bool *heapkeyspace,
bool *allequalimage);
extern void _bt_checkpage(Relation rel, Buffer buf);
-extern Buffer _bt_getbuf(Relation rel, BlockNumber blkno, int access);
+extern Buffer _bt_getbuf(Relation rel, Relation heaprel, BlockNumber blkno,
+ int access);
extern Buffer _bt_relandgetbuf(Relation rel, Buffer obuf,
BlockNumber blkno, int access);
extern void _bt_relbuf(Relation rel, Buffer buf);
@@ -1229,21 +1233,22 @@ extern void _bt_pendingfsm_finalize(Relation rel, BTVacState *vstate);
/*
* prototypes for functions in nbtsearch.c
*/
-extern BTStack _bt_search(Relation rel, BTScanInsert key, Buffer *bufP,
- int access, Snapshot snapshot);
-extern Buffer _bt_moveright(Relation rel, BTScanInsert key, Buffer buf,
- bool forupdate, BTStack stack, int access, Snapshot snapshot);
+extern BTStack _bt_search(Relation rel, Relation heaprel, BTScanInsert key,
+ Buffer *bufP, int access, Snapshot snapshot);
+extern Buffer _bt_moveright(Relation rel, Relation heaprel, BTScanInsert key,
+ Buffer buf, bool forupdate, BTStack stack,
+ int access, Snapshot snapshot);
extern OffsetNumber _bt_binsrch_insert(Relation rel, BTInsertState insertstate);
extern int32 _bt_compare(Relation rel, BTScanInsert key, Page page, OffsetNumber offnum);
extern bool _bt_first(IndexScanDesc scan, ScanDirection dir);
extern bool _bt_next(IndexScanDesc scan, ScanDirection dir);
-extern Buffer _bt_get_endpoint(Relation rel, uint32 level, bool rightmost,
- Snapshot snapshot);
+extern Buffer _bt_get_endpoint(Relation rel, Relation heaprel, uint32 level,
+ bool rightmost, Snapshot snapshot);
/*
* prototypes for functions in nbtutils.c
*/
-extern BTScanInsert _bt_mkscankey(Relation rel, IndexTuple itup);
+extern BTScanInsert _bt_mkscankey(Relation rel, Relation heaprel, IndexTuple itup);
extern void _bt_freestack(BTStack stack);
extern void _bt_preprocess_array_keys(IndexScanDesc scan);
extern void _bt_start_array_keys(IndexScanDesc scan, ScanDirection dir);
diff --git a/src/include/access/nbtxlog.h b/src/include/access/nbtxlog.h
index edd1333d9b..1e45d58845 100644
--- a/src/include/access/nbtxlog.h
+++ b/src/include/access/nbtxlog.h
@@ -188,9 +188,11 @@ typedef struct xl_btree_reuse_page
RelFileLocator locator;
BlockNumber block;
FullTransactionId snapshotConflictHorizon;
+ bool isCatalogRel; /* to handle recovery conflict during logical
+ * decoding on standby */
} xl_btree_reuse_page;
-#define SizeOfBtreeReusePage (sizeof(xl_btree_reuse_page))
+#define SizeOfBtreeReusePage (offsetof(xl_btree_reuse_page, isCatalogRel) + sizeof(bool))
/*
* xl_btree_vacuum and xl_btree_delete records describe deletion of index
@@ -235,13 +237,15 @@ typedef struct xl_btree_delete
TransactionId snapshotConflictHorizon;
uint16 ndeleted;
uint16 nupdated;
+ bool isCatalogRel; /* to handle recovery conflict during logical
+ * decoding on standby */
/* DELETED TARGET OFFSET NUMBERS FOLLOW */
/* UPDATED TARGET OFFSET NUMBERS FOLLOW */
/* UPDATED TUPLES METADATA (xl_btree_update) ARRAY FOLLOWS */
} xl_btree_delete;
-#define SizeOfBtreeDelete (offsetof(xl_btree_delete, nupdated) + sizeof(uint16))
+#define SizeOfBtreeDelete (offsetof(xl_btree_delete, isCatalogRel) + sizeof(bool))
/*
* The offsets that appear in xl_btree_update metadata are offsets into the
diff --git a/src/include/access/spgxlog.h b/src/include/access/spgxlog.h
index b9d6753533..75267a4914 100644
--- a/src/include/access/spgxlog.h
+++ b/src/include/access/spgxlog.h
@@ -240,6 +240,8 @@ typedef struct spgxlogVacuumRedirect
uint16 nToPlaceholder; /* number of redirects to make placeholders */
OffsetNumber firstPlaceholder; /* first placeholder tuple to remove */
TransactionId snapshotConflictHorizon; /* newest XID of removed redirects */
+ bool isCatalogRel; /* to handle recovery conflict during logical
+ * decoding on standby */
/* offsets of redirect tuples to make placeholders follow */
OffsetNumber offsets[FLEXIBLE_ARRAY_MEMBER];
diff --git a/src/include/access/visibilitymapdefs.h b/src/include/access/visibilitymapdefs.h
index 9165b9456b..7306a1c3ee 100644
--- a/src/include/access/visibilitymapdefs.h
+++ b/src/include/access/visibilitymapdefs.h
@@ -17,9 +17,11 @@
#define BITS_PER_HEAPBLOCK 2
/* Flags for bit map */
-#define VISIBILITYMAP_ALL_VISIBLE 0x01
-#define VISIBILITYMAP_ALL_FROZEN 0x02
-#define VISIBILITYMAP_VALID_BITS 0x03 /* OR of all valid visibilitymap
- * flags bits */
+#define VISIBILITYMAP_ALL_VISIBLE 0x01
+#define VISIBILITYMAP_ALL_FROZEN 0x02
+#define VISIBILITYMAP_VALID_BITS 0x03 /* OR of all valid visibilitymap
+ * flags bits */
+#define VISIBILITYMAP_IS_CATALOG_REL 0x04 /* to handle recovery conflict during logical
+ * decoding on standby */
#endif /* VISIBILITYMAPDEFS_H */
diff --git a/src/include/utils/rel.h b/src/include/utils/rel.h
index af9785038d..0cfe02aa4a 100644
--- a/src/include/utils/rel.h
+++ b/src/include/utils/rel.h
@@ -27,6 +27,7 @@
#include "storage/smgr.h"
#include "utils/relcache.h"
#include "utils/reltrigger.h"
+#include "catalog/catalog.h"
/*
diff --git a/src/include/utils/tuplesort.h b/src/include/utils/tuplesort.h
index 12578e42bc..395abfe596 100644
--- a/src/include/utils/tuplesort.h
+++ b/src/include/utils/tuplesort.h
@@ -399,7 +399,9 @@ extern Tuplesortstate *tuplesort_begin_heap(TupleDesc tupDesc,
int workMem, SortCoordinate coordinate,
int sortopt);
extern Tuplesortstate *tuplesort_begin_cluster(TupleDesc tupDesc,
- Relation indexRel, int workMem,
+ Relation indexRel,
+ Relation heaprel,
+ int workMem,
SortCoordinate coordinate,
int sortopt);
extern Tuplesortstate *tuplesort_begin_index_btree(Relation heapRel,
--
2.34.1
view thread (41+ messages) latest in thread
reply
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Reply to all the recipients using the --to and --cc options:
reply via email
To: [email protected]
Cc: [email protected], [email protected], [email protected], [email protected], [email protected], [email protected], [email protected], [email protected], [email protected], [email protected], [email protected]
Subject: Re: Minimal logical decoding on standbys
In-Reply-To: <[email protected]>
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
This inbox is served by agora; see mirroring instructions
for how to clone and mirror all data and code used for this inbox