Thank Nobuto,
I agree if not restarted there could be issues. 

In addition to the general habit of restarting services if possible there now 
also is the needrestart service in newer releases to further nudge users/admins 
to realize that.
For databases all of that isn't so easy - as Nobuto pointed out you'd usually 
use some tool to start your DB - and if set up for HA then that becomes more 
complex.

But even if the server isn't restarted I think there must be something more 
complex.
The error message that is reported indicates that it tires to load new code 
which mismatches the old code and therefore fails to load.
Like
 1. load server
 2. update without restart
 3. make the server load a plugin (new build, fails to load as it needs new 
builds symbols)

I'm still not sure the package could do anything about it, as I said
even if the symbols match there could be issues in the API not being the
same. So I appreciate the discussion and wonder if you do happen to know
if the tests that you run (or snap.maas.supervisor restart as that is
what Nobuto did to then trigger it) will make it late-load any plugins
or .so files?


/me wants to setup such a maas later, but now calls are inbound ...

-- 
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

Reply via email to