Horton

Re: wake up logical workers after ALTER SUBSCRIPTION

Thomas Munro <thomas.munro@gmail.com>

View thread View on postgr.es
From: Thomas Munro <thomas.munro@gmail.com>
To: Nathan Bossart <nathandbossart@gmail.com>
Cc: pgsql-hackers@postgresql.org
Date: 2022-11-22T02:16:05Z
Lists: pgsql-hackers 
On Tue, Nov 22, 2022 at 1:41 PM Nathan Bossart <nathandbossart@gmail.com> wrote:
> On my machine, the attached patch
> improved 'check-world -j8' run time by ~12 seconds (from 3min 8sec to 2min
> 56 sec) and src/test/subscription test time by ~17 seconds (from 139
> seconds to 122 seconds).

Nice!

Maybe a comment to explain why a single variable is enough?  And an
assertion that it wasn't already set?  And a note to future self: this
would be a candidate user of the nearby SetLatches() patch (which is
about moving SetLatch() syscalls out from under LWLocks, though this
one may not be very hot).

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.