Re: recovery.signal not cleaned up when both signal files are present

From: David Steele <david(at)pgbackrest(dot)org>
To: Michael Paquier <michael(at)paquier(dot)xyz>, Fujii Masao <masao(dot)fujii(at)gmail(dot)com>
Cc: Nikolay Samokhvalov <nik(at)postgres(dot)ai>, pgsql-hackers <pgsql-hackers(at)lists(dot)postgresql(dot)org>
Subject: Re: recovery.signal not cleaned up when both signal files are present
Date: 2026-02-13 00:55:15
Message-ID: d914df67-626c-41ee-9ebe-8015404885a0@pgbackrest.org
Views: Whole Thread | Raw Message | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On 2/10/26 11:52, Michael Paquier wrote:
> On Tue, Feb 10, 2026 at 12:26:36PM +0900, Fujii Masao wrote:
>> +1 to apply this change only to the master branch. Patch attached.
>
> It looks like something we should have a test for, at least..

+1 for a test.

> + /*
> + *
> + * If both signal files are present, standby signal file takes precedence.
> + * If neither is present then we won't enter archive recovery.
> + */
>
> This comment's format is incorrect.

Other than the comment the patch looks good overall.

Reluctantly I have to agree to not back patch this. I'm not sure how
this change would break existing recovery processes but experience tells
me that it probably would.

Instead -- I wonder if we could add a warning in the back branches?

Regards,
-David

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Henson Choi 2026-02-13 01:14:26 Re: Row pattern recognition
Previous Message Andres Freund 2026-02-13 00:51:18 Re: Replace literal 0 values with the appropriate Invalid* constants