Horton

Re: wake up logical workers after ALTER SUBSCRIPTION

Nathan Bossart <nathandbossart@gmail.com>

View thread View on postgr.es
From: Nathan Bossart <nathandbossart@gmail.com>
To: Tom Lane <tgl@sss.pgh.pa.us>
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: 2022-12-31T23:50:19Z
Lists: pgsql-hackers

Attachments

On Sun, Dec 18, 2022 at 03:36:07PM -0800, Nathan Bossart wrote:
> This seems to have somehow broken the archiving tests on Windows, so
> obviously I owe some better analysis here.  I didn't see anything obvious
> in the logs, but I will continue to dig.

On Windows, WaitForWALToBecomeAvailable() seems to depend on the call to
WaitLatch() for wal_retrieve_retry_interval to ensure that signals are
dispatched (i.e., pgwin32_dispatch_queued_signals()).  My first instinct is
to just always call WaitLatch() in this code path, even if
wal_retrieve_rety_interval milliseconds have already elapsed.  The attached
0003 does this.

-- 
Nathan Bossart
Amazon Web Services: https://aws.amazon.com

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.