Re: Use of backup_label not noted in log

David Steele <david@pgmasters.net>

From: David Steele <david@pgmasters.net>
To: Michael Paquier <michael@paquier.xyz>
Cc: Andres Freund <andres@anarazel.de>, Pg Hackers <pgsql-hackers@postgresql.org>, Laurenz Albe <laurenz.albe@cybertec.at>
Date: 2024-01-26T12:20:05Z
Lists: pgsql-hackers

On 1/25/24 20:52, Michael Paquier wrote:
> On Thu, Jan 25, 2024 at 08:56:52AM -0400, David Steele wrote:
>> I would still advocate for a back patch here. It is frustrating to get logs
>> from users that just say:
>>
>> LOG:  invalid checkpoint record
>> PANIC:  could not locate a valid checkpoint record
>>
>> It would be very helpful to know what the checkpoint record LSN was in this
>> case.
> 
> Yes, I've pested over this one in the past when debugging corruption
> issues.  To me, this would just mean to appens to the PANIC an "at
> %X/%X", but perhaps you have more in mind for these code paths?

I think adding the LSN to the panic message would be a good change for HEAD.

However, that still would not take the place of the additional messages 
in 1d35f705e showing that the LSN came from a backup_label.

Regards,
-David



Commits

  1. Add more LOG messages when starting and ending recovery from a backup