On 2025-07-05 Sa 5:00 AM, Alexander Lakhin wrote:
> Hello Andrew,
>
> I've noticed that prion fails with timeout rather frequently this week.
> Probably this is caused just by increased total number and duration of
> tests, as we can see timeouts at different stages:
> [1]
> # +++ tap install-check in src/test/modules/test_misc +++
> t/001_constraint_validation.pl .. ok
> t/002_tablespace.pl ............. ok
> t/003_check_guc.pl .............. ok
> t/004_io_direct.pl .............. ok
> --- testing was interrupted ---
>
> [2]
> # +++ regress install-check in contrib/intarray +++
> # using postmaster on /home/ec2-user/bf/root/tmp/buildfarm-D20LSW,
> port 5678
> --- testing was interrupted ---
>
> [3]
> # +++ regress install-check in contrib/postgres_fdw +++
> # using postmaster on /home/ec2-user/bf/root/tmp/buildfarm-eRpoH_,
> port 5678
> --- testing was interrupted ---
>
> Most of successful runs on master and REL_18_STABLE take around 2 hours,
> but there is also [4], with:
> snapshot taken 2025-07-01 21:36:23
> and the last line of postmasterl.log at the last (ecpg) stage is:
> 2025-07-02 00:39:58.519 UTC [3557281:14] LOCATION: UnlinkLockFiles,
> miscinit.c:1196
> (More than 3 hours.)
>
> and [5], with:
> snapshot taken 2025-07-04 16:28:07
> and the last line of postmasterl.log at the last (ecpg) stage is:
> 2025-07-04 20:13:51.355 UTC [2804973:14] LOCATION: UnlinkLockFiles,
> miscinit.c:1196
> (Almost 4 hours.)
>
> So maybe it's time to increase timeout for that animal, say, to 6
> hours...
> What do you think?
>
> [1]
> https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=prion&dt=2025-07-01%2017%3A28%3A07
> [2]
> https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=prion&dt=2025-07-03%2018%3A49%3A07
> [3]
> https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=prion&dt=2025-07-04%2020%3A39%3A06
> [4]
> https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=prion&dt=2025-07-01%2021%3A36%3A23
> [5]
> https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=prion&dt=2025-07-04%2016%3A28%3A07
Yes, done. It's kind of odd how much the performance of this EC2
instance varies, but I guess this should avoid the timeouts.
cheers
andrew
--
Andrew Dunstan
EDB: https://www.enterprisedb.com
