Horton
Menu

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

David Steele <david@pgbackrest.org>

View thread
From: David Steele <david@pgbackrest.org>
To: Fujii Masao <masao.fujii@gmail.com>, Michael Paquier <michael@paquier.xyz>
Cc: Nikolay Samokhvalov <nik@postgres.ai>, pgsql-hackers <pgsql-hackers@lists.postgresql.org>
Date: 2026-02-13T15:45:35Z
Lists: pgsql-hackers 
On 2/13/26 20:27, Fujii Masao wrote:
> On Fri, Feb 13, 2026 at 3:18 PM Michael Paquier <michael@paquier.xyz> wrote:
>>
>> On Fri, Feb 13, 2026 at 03:05:45PM +0900, Fujii Masao wrote:
>>> Yeah, so I've added the test as suggested. The updated patch is attached.
>>
>>
>> What's the point in having the check for the files in data_dir?  The
>> second one for standby2 should be enough as this is to test only
>> readRecoverySignalFile().
> 
> I added that test to verify that both files are removed even in the normal
> standby case (i.e., when only standby.signal is present). However, if testing
> only the case where both signal files are present is sufficient, I'm fine with
> removing the data_dir check. Attached is an updated patch that checks only
> the latter case for standby2.
> 
> I will commit this patch.

I'm fine with the additional checks in v2. They are inexpensive and show 
that the changes (probably) don't have side effects.

But I don't feel strongly about it so either v2 or v3 is OK with me.

Regards,
-David

Commits

Same data as JSON: GET /api/v1/messages/:b64id/commits the thread's linked commits as JSON, with link sources. API reference →

  1. Remove recovery.signal at recovery end when both signal files are present.