Horton

Re: wake up logical workers after ALTER SUBSCRIPTION

Amit Kapila <amit.kapila16@gmail.com>

View thread View on postgr.es
From: Amit Kapila <amit.kapila16@gmail.com>
To: Nathan Bossart <nathandbossart@gmail.com>
Cc: Melih Mutlu <m.melihmutlu@gmail.com>, Thomas Munro <thomas.munro@gmail.com>, "Hayato Kuroda (Fujitsu)" <kuroda.hayato@fujitsu.com>, "pgsql-hackers@postgresql.org" <pgsql-hackers@postgresql.org>
Date: 2023-01-06T05:00:18Z
Lists: pgsql-hackers 
On Thu, Jan 5, 2023 at 10:49 PM Nathan Bossart <nathandbossart@gmail.com> wrote:
>
> On Thu, Jan 05, 2023 at 11:34:37AM +0530, Amit Kapila wrote:
> > On Thu, Jan 5, 2023 at 10:16 AM Nathan Bossart <nathandbossart@gmail.com> wrote:
> >> In v12, I moved the restart for two_phase mode to the end of
> >> process_syncing_tables_for_apply() so that we don't need to rely on another
> >> iteration of the loop.
> >
> > This should work but it is better to add a comment before calling
> > CommandCounterIncrement() to indicate that this is for making changes
> > to the relation state visible.
>
> Will do.
>

Isn't it better to move this part into a separate patch as this is
useful even without the main patch to improve wakeups?

> > Thinking along similar lines, won't apply worker need to be notified
> > of SUBREL_STATE_SYNCWAIT state change by the tablesync worker?
>
> wait_for_worker_state_change() should notify the apply worker in this case.
>

I think this is yet to be included in the patch, right?

-- 
With Regards,
Amit Kapila.

Commits

  1. Avoid type cheats for invalid dsa_handles and dshash_table_handles.

  2. Track logrep apply workers' last start times to avoid useless waits.

  3. Check for two_phase change at end of process_syncing_tables_for_apply.

  4. Wake up a subscription's replication worker processes after DDL.