Thanks Christoph for chiming in and a second voice that confirms what I said.
So to summarize: - we have found the root cause to be sure what is going on (stored procedures triggered late lib loading of new on-disk .so files by an old instance of the server) - Packaging takes care of default-installations restarts - Any more complex server installations not restarted by upgrades need to be restarted by the admin/automation or hold off the upgrade until the time when restarting is possible. - This also matches upstreams intended supported scenarios Further more - and I'm really NOT recommending this - if anyone has a setup where it is known in advance that he can/want in no way ever restart the server (TBH what is the update for then, but anyway) there a configuration option to preload shared libraries - that could (not tested) allow to load the lib on server start avoiding the late-load which here was the issue after upgrades, see: https://postgresqlco.nf/doc/en/param/shared_preload_libraries/ https://www.postgresql.org/docs/current/runtime-config-client.html#RUNTIME-CONFIG-CLIENT-PRELOAD Thereby I think we can close the postgresql task on this bug as Won't Fix. ** Changed in: postgresql-12 (Ubuntu) Status: Confirmed => Won't Fix -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/1939898 Title: Unnatended postgresql-12 upgrade caused MAAS internal error To manage notifications about this bug go to: https://bugs.launchpad.net/maas/+bug/1939898/+subscriptions -- ubuntu-bugs mailing list [email protected] https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
