Movatterモバイル変換

[0]ホーム
URL:

Skip to content

Navigation Menu

Sign in
Appearance settings

Search code, repositories, users, issues, pull requests...

Provide feedback

We read every piece of feedback, and take your input very seriously.

Saved searches

Use saved searches to filter your results more quickly

 Sign up
Appearance settings

Commit91c0570

Browse files
committed
Don't fail for > 1 walsenders in 019_replslot_limit, add debug messages.
So far the first of the retries introduced inf28bf66 resolves theissue. But I (Andres) am still suspicious that the start of the failures mightindicate a problem.To reduce noise, stop reporting a failure if a retry resolves the problem. Toallow figuring out what causes the slow slot drop, add a few more debugmessages to ReplicationSlotDropPtr.See also commitafdeff1,fe0972e andf28bf66.Discussion:https://postgr.es/m/20220327213219.smdvfkq2fl74flow@alap3.anarazel.de
1 parentda4b566 commit91c0570

File tree

2 files changed

+16
-3
lines changed

2 files changed

+16
-3
lines changed

‎src/backend/replication/slot.c

Lines changed: 9 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -702,15 +702,22 @@ ReplicationSlotDropPtr(ReplicationSlot *slot)
702702
slot->active_pid=0;
703703
slot->in_use= false;
704704
LWLockRelease(ReplicationSlotControlLock);
705+
706+
elog(DEBUG3,"replication slot drop: %s: marked as not in use",NameStr(slot->data.name));
707+
705708
ConditionVariableBroadcast(&slot->active_cv);
706709

710+
elog(DEBUG3,"replication slot drop: %s: notified others",NameStr(slot->data.name));
711+
707712
/*
708713
* Slot is dead and doesn't prevent resource removal anymore, recompute
709714
* limits.
710715
*/
711716
ReplicationSlotsComputeRequiredXmin(false);
712717
ReplicationSlotsComputeRequiredLSN();
713718

719+
elog(DEBUG3,"replication slot drop: %s: computed required",NameStr(slot->data.name));
720+
714721
/*
715722
* If removing the directory fails, the worst thing that will happen is
716723
* that the user won't be able to create a new slot with the same name
@@ -720,6 +727,8 @@ ReplicationSlotDropPtr(ReplicationSlot *slot)
720727
ereport(WARNING,
721728
(errmsg("could not remove directory \"%s\"",tmppath)));
722729

730+
elog(DEBUG3,"replication slot drop: %s: removed directory",NameStr(slot->data.name));
731+
723732
/*
724733
* Send a message to drop the replication slot to the stats collector.
725734
* Since there is no guarantee of the order of message transfer on a UDP

‎src/test/recovery/t/019_replslot_limit.pl

Lines changed: 7 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -339,8 +339,8 @@
339339
# We've seen occasional cases where multiple walsender pids are active. It
340340
# could be that we're just observing process shutdown being slow. To collect
341341
# more information, retry a couple times, print a bit of debugging information
342-
# each iteration.For now report atestfailure even if later iterations
343-
#succeed.
342+
# each iteration.Don't fail thetestif retries find just one pid, the
343+
#buildfarm failures are too noisy.
344344
my$i = 0;
345345
while (1)
346346
{
@@ -349,7 +349,9 @@
349349
$senderpid =$node_primary3->safe_psql('postgres',
350350
"SELECT pid FROM pg_stat_activity WHERE backend_type = 'walsender'");
351351

352-
lastif like($senderpid,qr/^[0-9]+$/,"have walsender pid$senderpid");
352+
lastif$senderpid =~qr/^[0-9]+$/;
353+
354+
diag"multiple walsenders active in iteration$i";
353355

354356
# show information about all active connections
355357
$node_primary3->psql('postgres',
@@ -370,6 +372,8 @@
370372
usleep(100_000);
371373
}
372374

375+
like($senderpid,qr/^[0-9]+$/,"have walsender pid$senderpid");
376+
373377
my$receiverpid =$node_standby3->safe_psql('postgres',
374378
"SELECT pid FROM pg_stat_activity WHERE backend_type = 'walreceiver'");
375379
like($receiverpid,qr/^[0-9]+$/,"have walreceiver pid$receiverpid");

0 commit comments

Comments
 (0)
[8]ページ先頭
©2009-2025 Movatter.jp