Re: Address the bug in 041_checkpoint_at_promote.pl

From: Michael Paquier <michael(at)paquier(dot)xyz>
To: Nitin Jadhav <nitinjadhavpostgres(at)gmail(dot)com>
Cc: Pg Hackers <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: Address the bug in 041_checkpoint_at_promote.pl
Date: 2025-02-12 08:24:21
Message-ID: Z6xatZsXN_J2JYNO@paquier.xyz
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On Wed, Feb 12, 2025 at 01:28:55PM +0530, Nitin Jadhav wrote:
> The code is intended to wait for the restart point to complete before
> proceeding. However, it doesn't actually wait. Regardless of whether
> the restart point completes, the loop exits after the first iteration
> because the if condition always evaluates to true. This happens
> because $logstart is not passed as an argument to log_contains() by
> mistake. If the restart point operation is quick, this issue might not
> be noticeable, which is often the case.
>
> I've attached a patch to fix this issue. Please review and share your feedback.

Oops, you're right. I am measuring 2ms or so in average between the wakeup
and the restartpoint complete. Removing the wakeup makes the test
complete, but it should wait in its lookup loop.

Will fix down to v17 where this error has been introduced.
--
Michael

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Julien Rouhaud 2025-02-12 08:25:38 Unneeded volatile qualifier in fmgr.c
Previous Message Nitin Jadhav 2025-02-12 07:58:55 Address the bug in 041_checkpoint_at_promote.pl