Hello Thomas and Michael,
14.03.2024 06:16, Thomas Munro wrote:
>
> Yeah, I was wondering if its checkpoint delaying logic might have
> got the checkpointer jammed or something like that, but I don't
> currently see how. Yeah, the replay of bulk newpages could be
> relevant, but it's not exactly new technology. One thing I wondered
> about is whether the Perl "wait for catchup" thing, which generates
> large volumes of useless log, could be somehow changed to actually
> show the progress after some time. Something like "I'm still waiting
> for this replica to reach LSN X, but it has so far only reported LSN
> Y, and here's a dump of the WAL around there"?
I have perhaps reproduced the issue here (at least I'm seeing something
similar), and going to investigate the issue in the coming days, but what
I'm confused with now is the duration of poll_query_until:
For the failure you referenced:
[15:55:54.740](418.725s) # poll_query_until timed out executing this query:
And a couple of others:
https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=skink&dt=2024-03-08%2000%3A34%3A06
[00:45:57.747](376.159s) # poll_query_until timed out executing this query:
https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=skink&dt=2024-03-04%2016%3A32%3A17
[16:45:24.870](407.970s) # poll_query_until timed out executing this query:
Could it be that the timeout (360 sec?) is just not enough for the test
under the current (changed due to switch to meson) conditions?
Best regards,
Alexander