1
0
mirror of https://github.com/postgres/postgres.git synced 2025-05-02 11:44:50 +03:00

Adjust pg_wal_replay_wait() procedure behavior on promoted standby

pg_wal_replay_wait() is intended to be called on standby.  However, standby
can be promoted to primary at any moment, even concurrently with the
pg_wal_replay_wait() call.  If recovery is not currently in progress
that doesn't mean the wait was unsuccessful.  Thus, we always need to recheck
if the target LSN is replayed.

Reported-by: Kevin Hale Boyes
Discussion: https://postgr.es/m/CAPpHfdu5QN%2BZGACS%2B7foxmr8_nekgA2PA%2B-G3BuOUrdBLBFb6Q%40mail.gmail.com
Author: Alexander Korotkov
This commit is contained in:
Alexander Korotkov 2024-08-10 21:43:02 +03:00
parent bbf668d66f
commit 867d396ccd
3 changed files with 55 additions and 11 deletions

View File

@ -28969,6 +28969,15 @@ postgres=# SELECT '0/0'::pg_lsn + pd.segment_number * ps.setting::int + :offset
connection pooler side.
</para>
<para>
<function>pg_wal_replay_wait</function> should be called on standby.
If a user calls <function>pg_wal_replay_wait</function> on primary, it
will error out. However, if <function>pg_wal_replay_wait</function> is
called on primary promoted from standby and <literal>target_lsn</literal>
was already replayed, then <function>pg_wal_replay_wait</function> just
exits immediately.
</para>
<para>
You can use <function>pg_wal_replay_wait</function> to wait for
the <type>pg_lsn</type> value. For example, an application could update

View File

@ -230,14 +230,27 @@ WaitForLSNReplay(XLogRecPtr targetLSN, int64 timeout)
Assert(MyProcNumber >= 0 && MyProcNumber < MaxBackends);
if (!RecoveryInProgress())
{
/*
* Recovery is not in progress. Given that we detected this in the
* very first check, this procedure was mistakenly called on primary.
* However, it's possible that standby was promoted concurrently to
* the procedure call, while target LSN is replayed. So, we still
* check the last replay LSN before reporting an error.
*/
if (targetLSN <= GetXLogReplayRecPtr(NULL))
return;
ereport(ERROR,
(errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
errmsg("recovery is not in progress"),
errhint("Waiting for LSN can only be executed during recovery.")));
/* If target LSN is already replayed, exit immediately */
if (targetLSN <= GetXLogReplayRecPtr(NULL))
return;
}
else
{
/* If target LSN is already replayed, exit immediately */
if (targetLSN <= GetXLogReplayRecPtr(NULL))
return;
}
if (timeout > 0)
{
@ -257,19 +270,30 @@ WaitForLSNReplay(XLogRecPtr targetLSN, int64 timeout)
int rc;
long delay_ms = 0;
/* Check if the waited LSN has been replayed */
currentLSN = GetXLogReplayRecPtr(NULL);
if (targetLSN <= currentLSN)
break;
/* Recheck that recovery is still in-progress */
if (!RecoveryInProgress())
{
/*
* Recovery was ended, but recheck if target LSN was already
* replayed.
*/
currentLSN = GetXLogReplayRecPtr(NULL);
if (targetLSN <= currentLSN)
return;
ereport(ERROR,
(errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
errmsg("recovery is not in progress"),
errdetail("Recovery ended before replaying target LSN %X/%X; last replay LSN %X/%X.",
LSN_FORMAT_ARGS(targetLSN),
LSN_FORMAT_ARGS(currentLSN))));
}
else
{
/* Check if the waited LSN has been replayed */
currentLSN = GetXLogReplayRecPtr(NULL);
if (targetLSN <= currentLSN)
break;
}
/*
* If the timeout value is specified, calculate the number of

View File

@ -126,12 +126,18 @@ ok(1, 'multiple LSN waiters reported consistent data');
# 5. Check that the standby promotion terminates the wait on LSN. Start
# waiting for an unreachable LSN then promote. Check the log for the relevant
# error message.
# error message. Also, check that waiting for already replayed LSN doesn't
# cause an error even after promotion.
my $lsn4 =
$node_primary->safe_psql('postgres',
"SELECT pg_current_wal_insert_lsn() + 10000000000");
my $lsn5 =
$node_primary->safe_psql('postgres', "SELECT pg_current_wal_insert_lsn()");
my $psql_session = $node_standby1->background_psql('postgres');
$psql_session->query_until(
qr/start/, qq[
\\echo start
CALL pg_wal_replay_wait('${lsn3}');
CALL pg_wal_replay_wait('${lsn4}');
]);
$log_offset = -s $node_standby1->logfile;
@ -140,6 +146,11 @@ $node_standby1->wait_for_log('recovery is not in progress', $log_offset);
ok(1, 'got error after standby promote');
$node_standby1->safe_psql('postgres', "CALL pg_wal_replay_wait('${lsn5}');");
ok(1,
'wait for already replayed LSN exists immediately even after promotion');
$node_standby1->stop;
$node_primary->stop;