Horton

Re: Intermittent buildfarm failures on wrasse

Noah Misch <noah@leadboat.com>

View thread View on postgr.es
From: Noah Misch <noah@leadboat.com>
To: Andres Freund <andres@anarazel.de>
Cc: Tom Lane <tgl@sss.pgh.pa.us>, Peter Geoghegan <pg@bowt.ie>, David Rowley <dgrowleyml@gmail.com>, PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Date: 2022-04-15T05:21:25Z
Lists: pgsql-hackers 
On Thu, Apr 14, 2022 at 10:12:05PM -0700, Andres Freund wrote:
> On 2022-04-14 19:45:15 -0700, Noah Misch wrote:
> > I suspect the failure is somehow impossible in "check".  Yesterday, I cranked
> > up the number of locales, so there are now a lot more installcheck.  Before
> > that, each farm run had one "check" and two "installcheck".  Those days saw
> > ten installcheck failures, zero check failures.
> 
> I notice that the buildfarm appears to run initdb with syncing enabled
> ("syncing data to disk ... ok" in the initdb steps).  Whereas pg_regress
> uses --no-sync.

Yep.

> I wonder if that's what makes the difference? Now that you reproduced
> it, does it still reproduce with --no-sync added?

It does; the last version of my script used "initdb -N ...".

> Also worth noting that pg_regress doesn't go through pg_ctl...

Hmmm.

Commits

  1. Repurpose PROC_COPYABLE_FLAGS as PROC_XMIN_FLAGS

  2. Tighten ComputeXidHorizons' handling of walsenders.

  3. Adjust VACUUM's removable cutoff log message.

  4. Temporarily add some probes of tenk1's relallvisible in create_index.sql.

  5. Set synchronous_commit=on in test_setup.sql.

  6. Rearrange core regression tests to reduce cross-script dependencies.

  7. snapshot scalability: Don't compute global horizons while building snapshots.