Horton

Re: Make use of assign_checkpoint_completion_target() to calculate CheckPointSegments correctly

Michael Paquier <michael@paquier.xyz>

View thread View on postgr.es
From: Michael Paquier <michael@paquier.xyz>
To: Bharath Rupireddy <bharath.rupireddyforpostgres@gmail.com>
Cc: PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Date: 2023-01-19T04:16:50Z
Lists: pgsql-hackers 
On Tue, Jan 17, 2023 at 07:55:53PM +0530, Bharath Rupireddy wrote:
> On Tue, Jan 17, 2023 at 12:31 PM Michael Paquier <michael@paquier.xyz> wrote:
>> Oops.  It looks like you are right here.  This would impact the
>> calculation of CheckPointSegments on reload when
>> checkpoint_completion_target is updated.  This is wrong since we've
>> switched to max_wal_size as of 88e9823, so this had better be
>> backpatched all the way down.
>>
>> Thoughts?
> 
> +1 to backpatch as setting checkpoint_completion_target will not take
> effect immediately.

Okay, applied down to 11.  I have double-checked the surroundings to
see if there was a similar mistake or something hiding around
CheckpointSegments but noticed nothing (also did some primary/standby
pgbench-ing with periodic reloads of checkpoint_completion_target,
while on it).
--
Michael

Commits

  1. Add missing assign hook for GUC checkpoint_completion_target

  2. Replace checkpoint_segments with min_wal_size and max_wal_size.