Hi all,
(Thomas and Andres in CC.)
Andres has complained a couple of days ago about the quantity of logs
that can be fed into the TAP tests running pg_regress:
https://www.postgresql.org/message-id/20220603195318.qk4voicqfdhlsnoh@alap3.anarazel.de
This concerns the TAP tests of pg_upgrade, as well as
027_stream_regress.pl, where a crash while running the tests would
show a couple of megs worth of regression.diffs. Most of the output
generated does not make much sense to have in this case, and the
parallel schedule gives a harder time to spot the exact query involved
in the crash (if that's a query) while the logs of the backend should
be enough to spot what's the problem with the PIDs tracked.
One idea I got to limit the useless output generated is to check the
status of the cluster after running the regression test suite as
restart_on_crash is disabled by default in Cluster.pm, and avoid any
follow-up logic in these tests if the cluster is not running anymore,
as of the attached.
Thoughts?
--
Michael
