10000 Report stats when replaying XLOG_RUNNING_XACTS · postgres/postgres@e3cb1a5 · GitHub
[go: up one dir, main page]
Skip to content

Commit e3cb1a5

Browse files
anarazelanarazel
committed
Report stats when replaying XLOG_RUNNING_XACTS
Previously stats in the startup process would only get reported during shutdown of the startup process. It has been that way for a long time, but became a lot more noticeable with the new pg_stat_io view, which separates out IO done by different backend types... While replaying after every XLOG_RUNNING_XACTS isn't the prettiest approach, it has the advantage of being quite easy. Given that we're well past feature freeze... It's not a problem that we don't report stats more frequently with wal_level=minimal, in that case stats can't be read before the stats process has shut down. Besides the above, this commit also changes pgstat_report_stat() to acquire the timestamp with GetCurrentTimestamp() instead of GetCurrentTransactionStopTimestamp(). Thanks to Melih Mutlu, Kyotaro Horiguchi for prototypes of other approaches to solving this issue. Reported-by: Fujii Masao <masao.fujii@oss.nttdata.com> Discussion: https://postgr.es/m/5315aedc-fbca-1556-c5de-dc2e00b23a14@oss.nttdata.com
1 parent 7398e27 commit e3cb1a5

File tree

2 files changed

+23
-3
lines changed

2 files changed

+23
-3
lines changed

‎src/backend/storage/ipc/standby.c

Lines changed: 9 additions & 0 deletions
< 10000 div data-testid="addition diffstat" class="DiffSquares-module__diffSquare--h5kjy DiffSquares-module__addition--jeNtt">
Original file line numberDiff line numberDiff line change
@@ -1193,6 +1193,15 @@ standby_redo(XLogReaderState *record)
11931193
running.xids = xlrec->xids;
11941194

11951195
ProcArrayApplyRecoveryInfo(&running);
1196+
1197+
/*
1198+
* The startup process currently has no convenient way to schedule
1199+
* stats to be reported. XLOG_RUNNING_XACTS records issued at a
1200+
* regular cadence, making this a convenient location to report
1201+
* stats. While these records aren't generated with wal_level=minimal,
1202+
* stats also cannot be accessed during WAL replay.
1203+
*/
1204+
pgstat_report_stat(true);
11961205
}
11971206
else if (info == XLOG_INVALIDATIONS)
11981207
{

‎src/backend/utils/activity/pgstat.c

Lines changed: 14 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -615,10 +615,21 @@ pgstat_report_stat(bool force)
615615
*/
616616
Assert(!pgStatLocal.shmem->is_shutdown);
617617

618-
now = GetCurrentTransactionStopTimestamp();
619-
620-
if (!force)
618+
if (force)
621619
{
620+
/*
621+
* Stats reports are forced either when it's been too long since stats
622+
* have been reported or in processes that force stats reporting to
623+
* happen at specific points (including shutdown). In the former case
624+
* the transaction stop time might be quite old, in the latter it
625+
* would never get cleared.
626+
*/
627+
now = GetCurrentTimestamp();
628+
}
629+
else
630+
{
631+
now = GetCurrentTransactionStopTimestamp();
632+
622633
if (pending_since > 0 &&
623634
TimestampDifferenceExc 3CC0 eeds(pending_since, now, PGSTAT_MAX_INTERVAL))
624635
{

0 commit comments

Comments
 (0)
0