Horton

Re: Statement timeout in pg_rewind

Michael Paquier <michael@paquier.xyz>

View thread View on postgr.es
From: Michael Paquier <michael@paquier.xyz>
To: Alexander Kukushkin <cyberdemn@gmail.com>
Cc: David Fetter <david@fetter.org>, Tom Lane <tgl@sss.pgh.pa.us>, PostgreSQL-development <pgsql-hackers@postgresql.org>
Date: 2019-08-28T02:51:33Z
Lists: pgsql-hackers 
On Tue, Aug 27, 2019 at 10:45:27AM +0200, Alexander Kukushkin wrote:
> Done, please see the next version attached.

I have made the new error message consistent with run_simple_query to
avoid more work to translators and because it is possible to know
immediately the code path involved thanks to the SQL query, then
applied the fix down to 9.5 where pg_rewind has been added.  Please
note that your patch had a warning as "result" is not needed in
run_simple_command().

idle_in_transaction_session_timeout only applies to 9.6 and newer
versions.  lock_timeout (imagine a concurrent lock on pg_class for
example) and statement_timeout can cause issues, but the full set gets
disabled as your patch did and as mentioned upthread.
--
Michael

Commits

  1. Disable timeouts when running pg_rewind with online source cluster