Re: test: avoid redundant standby catchup in 049_wait_for_lsn

From: Alexander Korotkov <aekorotkov(at)gmail(dot)com>
To: Xuneng Zhou <xunengzhou(at)gmail(dot)com>
Cc: Jobin Augustine <jobin(dot)augustine(at)percona(dot)com>, Michael Paquier <michael(at)paquier(dot)xyz>, pgsql-hackers <pgsql-hackers(at)lists(dot)postgresql(dot)org>
Subject: Re: test: avoid redundant standby catchup in 049_wait_for_lsn
Date: 2026-11-18 14:27:04
Message-ID: CAPpHfdu+R3UvVAiTvDyYSL=Z92N9uJH2qZWfMhhz=c6eGSM_nQ@mail.gmail.com
Views: Whole Thread | Raw Message | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On Sun, Nov 15, 2026 at 5:03 PM Alexander Korotkov <aekorotkov(at)gmail(dot)com> wrote:
>
> On Sun, Nov 15, 2026 at 8:09 AM Xuneng Zhou <xunengzhou(at)gmail(dot)com> wrote:
> >
> > On Sat, Nov 14, 2026 at 7:00 PM Alexander Korotkov <aekorotkov(at)gmail(dot)com> wrote:
> > >
> > > On Sat, Nov 14, 2026 at 7:53 AM Xuneng Zhou <xunengzhou(at)gmail(dot)com> wrote:
> > > > Thanks for looking into this!
> > > >
> > > > > Thanks to Xuneng for investigation, and thanks to Jobin for the
> > > > > additional report.
> > > > >
> > > > > I actually don't like idea of new pg_current_wal_insert_end_lsn()
> > > > > function for the following reasons:
> > > > > 1) We should fix this for PG 19. But it's RC already, we should avoid
> > > > > catversion bump if possible.
> > > > > 2) There is probability that users would still use
> > > > > pg_current_wal_insert_lsn() and hit to this problem.
> > > > > 3) Additional compexity.
> > > >
> > > > Makes sense to me. I did not consider the potential issue of early-adapters..
> > > >
> > > > > I suggest that WAIT FOR LSN can normalize the LSN to be waited on its
> > > > > side. There is no point one should wait for the LSN inside the page
> > > > > header. The attached patch implements that. Optinions?
> > > >
> > > > Normalizing the LSN in ExecWaitStmt() before invoking WaitForLSN()
> > > > seems good to me. It adds bits of extra cost which seems small and the
> > > > WAIT does not sit in a hot path.
> > > >
> > > > The new test looks like it will fail on 32-bit systems, since it
> > > > assumes 64-bit record and page header sizes. I'm wondering if we could
> > > > use pg_switch_wal() instead? After a WAL switch, the insert LSN is
> > > > just past the first segment page header, so size arithmetic is not
> > > > needed. This only covers the long-header case, not ordinary pages, but
> > > > that seems fine, since the mapping logic stays the same.
> > >
> > > Good point, thank you!
> > > Please, check v2 where pg_switch_wal() is used.
> > >
> >
> > Thanks for updating the patch. In my testing, it exercised the bug
> > well. I made a minor adjustment to the patch to stop the standby at
> > the boundary before checking, so an unrelated background WAL would not
> > succeed the test accidently. If this seems unnecessary, please feel
> > free to ignore it. Otherwise, v2 looks good to me.
>
> I'm good with your update, thank you. I'm going to push this once
> release freeze is lifted.

Pushed and marked Open Item as resolved.

------
Regards,
Alexander Korotkov
Supabase

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Nikhil Sontakke 2026-11-18 14:27:51 Re: Dropping a composite attribute causes data integrity violations
Previous Message Heikki Linnakangas 2026-11-18 14:25:17 Re: Compile warnings under -DMEMORY_CONTEXT_CHECKING and no assertions