Re: pgsql: Test replay of regression tests, attempt II. - Mailing list pgsql-committers

On Tue, Jan 18, 2022 at 9:37 AM Andres Freund <andres@anarazel.de> wrote:
> diff -w -U3 c:/cirrus/src/test/recovery/../regress/expected/reloptions.out
c:/cirrus/src/test/recovery/results/reloptions.out
> --- c:/cirrus/src/test/recovery/../regress/expected/reloptions.out      2022-01-17 07:08:52.779337800 +0000
> +++ c:/cirrus/src/test/recovery/results/reloptions.out  2022-01-17 07:24:46.785394900 +0000
> @@ -133,7 +133,7 @@
>  SELECT pg_relation_size('reloptions_test') = 0;
>   ?column?
>  ----------
> - t
> + f
>  (1 row)
>
> and the server log:
>
https://api.cirrus-ci.com/v1/artifact/task/6460573058596864/log/src/test/recovery/tmp_check/log/027_stream_regress_primary.log
>
> I'm not entirely clear on what the regression test is trying to test here.
>
>
> It's plausible the standby was a bit behind at that point and prevented
> VACUUM from truncating via hot_standby_feeback?

I don't think hot_standby_feedback is enabled.  There is some history
of flakiness here, recently adjusted by the following commit.  I'm not
entirely sure what replication has to do with it... perhaps it's just
messing with CPU scheduling.

commit fe246d1c111d43fd60a1b0afff25ed09b7ae11eb
Author: Michael Paquier <michael@paquier.xyz>
Date:   Fri Apr 2 09:44:42 2021 +0900

    Improve stability of test with vacuum_truncate in reloptions.sql

Hmm... looking at that commit and the referenced discussion, I'm not
sure why that commit added only FREEZE and not also
DISABLE_PAGE_SKIPPING, following the example of:

commit c2dc1a79767a0f947e1145f82eb65dfe4360d25f
Author: Tom Lane <tgl@sss.pgh.pa.us>
Date:   Wed Jan 20 11:49:29 2021 -0500

    Disable vacuum page skipping in selected test cases.