Horton

Re: Add two missing tests in 035_standby_logical_decoding.pl

Bertrand Drouvot <bertranddrouvot.pg@gmail.com>

View thread View on postgr.es
From: "Drouvot, Bertrand" <bertranddrouvot.pg@gmail.com>
To: "Yu Shi (Fujitsu)" <shiy.fnst@fujitsu.com>, Amit Kapila <amit.kapila16@gmail.com>
Cc: vignesh C <vignesh21@gmail.com>, PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Date: 2023-04-26T14:23:06Z
Lists: pgsql-hackers

Attachments

Hi,

On 4/26/23 11:58 AM, Yu Shi (Fujitsu) wrote:
> On Mon, Apr 24, 2023 8:07 PM Drouvot, Bertrand <bertranddrouvot.pg@gmail.com> wrote:

> I think that's because when replaying a checkpoint record, the startup process
> of standby only saves the information of the checkpoint, and we need to wait for
> the checkpointer to perform a restartpoint (see RecoveryRestartPoint), right? If
> so, could we force a checkpoint on standby? After this, the standby should have
> completed the restartpoint and we don't need to wait.
> 

Thanks for looking at it!

Oh right, that looks like good a good way to ensure the WAL file is removed on the standby
so that we don't need to wait.

Implemented that way in V6 attached and that works fine.

> Besides, would it be better to wait for the cascading standby? If the wal log
> file needed for cascading standby is removed on the standby, the subsequent test
> will fail. 

Good catch! I agree that we have to wait on the cascading standby before removing
the WAL files. It's done in V6 (and the test is not failing anymore if we set a
recovery_min_apply_delay to 5s on the cascading standby).

Regards,

-- 
Bertrand Drouvot
PostgreSQL Contributors Team
RDS Open Source Databases
Amazon Web Services: https://aws.amazon.com

Commits

  1. Fix the race condition in the test case added by commit a6e04b1d20.

  2. Test that invalidated logical slots doesn't retain WAL.

  3. Add a test to verify that subscription to the standby works.

  4. Reduce the log level in 035_standby_logical_decoding.pl.