Horton

Re: t/035_standby_logical_decoding.pl might fail on attempt to read wrong timeline

Xuneng Zhou <xunengzhou@gmail.com>

View thread View on postgr.es
From: Xuneng Zhou <xunengzhou@gmail.com>
To: Bertrand Drouvot <bertranddrouvot.pg@gmail.com>, "Hayato Kuroda (Fujitsu)" <kuroda.hayato@fujitsu.com>
Cc: Alexander Lakhin <exclusion@gmail.com>, pgsql-hackers <pgsql-hackers@postgresql.org>
Date: 2026-06-09T06:13:54Z
Lists: pgsql-hackers 
On Tue, Jun 9, 2026 at 1:57 PM Bertrand Drouvot
<bertranddrouvot.pg@gmail.com> wrote:
>
> Hi Hayato-san,
>
> On Tue, Jun 09, 2026 at 04:04:20AM +0000, Hayato Kuroda (Fujitsu) wrote:
> > Dear Bertrand, Xuneng,
> >
> > Thanks for sharing the reproducer and opinion. I could now understand.
> >
> > While seeing, I felt the same issue can happen even for the streaming case.
> > Please see attached reproducer and possible fix.

Good catch!

> I just had a quick look and I feel that's probably an issue too (good find!).
> I think that your fix should use GetWALInsertionTimeLineIfSet() though (but I
> only did a quick scan on your patch).
>
> > Can you also include that point? Or this thread discusses only for the logical case?
> > If separated, I can take initiative for the streaming part, after the logical decoding
> > case.
>
> I would says it's probably better to keep this thread focusing on Alexander's issue
> (logical decoding). Once this thread is fixed, then I'd suggest to open a dedicated
> thread for the streaming one (mentioning this thread and probably trying to mimic
> its fix as close as possible).
>

+1 for starting a dedicated thread. Please CC me in that thread in
case I miss it.

-- 
Regards,
Xuneng Zhou
HighGo Software Co., Ltd.

Commits

  1. Fix second race with timeline selection during promotion

  2. Remove TAP test for timeline lookup race with logical decoding on standbys

  3. Fix race with timeline selection in logical decoding during promotion

  4. Backport GetWALInsertionTimeLineIfSet()

  5. TAP test for logical decoding on standby