database/postgres
1
0
Fork 0
mirror of https://github.com/postgres/postgres.git synced 2025-09-03 15:22:11 +03:00
Code Activity

Advance old-segment horizon properly after slot invalidation

Browse Source 
When some slots are invalidated due to the max_slot_wal_keep_size limit,
the old segment horizon should move forward to stay within the limit.
However, in commit c655077639 we forgot to call KeepLogSeg again to
recompute the horizon after invalidating replication slots.  In cases
where other slots remained, the limits would be recomputed eventually
for other reasons, but if all slots were invalidated, the limits would
not move at all afterwards.  Repair.

Backpatch to 13 where the feature was introduced.

Author: Kyotaro Horiguchi <horikyota.ntt@gmail.com>
Reported-by: Marcin Krupowicz <mk@071.ovh>
Discussion: https://postgr.es/m/17103-004130e8f27782c9@postgresql.org
This commit is contained in:
Alvaro Herrera
2021-07-16 12:07:30 -04:00
parent 525115a332
commit e5bcbb1070
4 changed files with 74 additions and 12 deletions
Download Patch File Download Diff File Expand all files Collapse all files

32
src/test/recovery/t/019_replslot_limit.pl
View File

@@ -11,7 +11,7 @@ use TestLib;
use PostgresNode;
use File::Path qw(rmtree);
use Test::More tests => $TestLib::windows_os ? 14 : 18;
use Test::More tests => $TestLib::windows_os ? 15 : 19;
use Time::HiRes qw(usleep);
$ENV{PGDATABASE} = 'postgres';
@@ -176,7 +176,12 @@ ok( !find_in_log(
# Advance WAL again, the slot loses the oldest segment.
my $logstart = get_log_size($node_primary);
advance_wal($node_primary, 7);
$node_primary->safe_psql('postgres', "CHECKPOINT;");
# This slot should be broken, wait for that to happen
$node_primary->poll_query_until(
'postgres',
qq[SELECT wal_status = 'lost' FROM pg_replication_slots
WHERE slot_name = 'rep1']);
# WARNING should be issued
ok( find_in_log(
@@ -185,13 +190,28 @@ ok( find_in_log(
$logstart),
'check that the warning is logged');
# This slot should be broken
$result = $node_primary->safe_psql('postgres',
"SELECT slot_name, active, restart_lsn IS NULL, wal_status, safe_wal_size FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
$result = $node_primary->safe_psql(
'postgres',
qq[
SELECT slot_name, active, restart_lsn IS NULL, wal_status, safe_wal_size
FROM pg_replication_slots WHERE slot_name = 'rep1']);
is($result, "rep1|f|t|lost|",
'check that the slot became inactive and the state "lost" persists');
# The invalidated slot shouldn't keep the old-segment horizon back;
# see bug #17103: https://postgr.es/m/17103-004130e8f27782c9@postgresql.org
# Test for this by creating a new slot and comparing its restart LSN
# to the oldest existing file.
my $redoseg = $node_primary->safe_psql('postgres',
"SELECT pg_walfile_name(lsn) FROM pg_create_physical_replication_slot('s2', true)"
);
my $oldestseg = $node_primary->safe_psql('postgres',
"SELECT pg_ls_dir AS f FROM pg_ls_dir('pg_wal') WHERE pg_ls_dir ~ '^[0-9A-F]{24}\$' ORDER BY 1 LIMIT 1"
);
$node_primary->safe_psql('postgres',
qq[SELECT pg_drop_replication_slot('s2')]);
is($oldestseg, $redoseg, "check that segments have been removed");
# The standby no longer can connect to the primary
$logstart = get_log_size($node_standby);
$node_standby->start;