Runnable jobs never start

[Franco Broi]

Hi

Seen this a few times now, we have jobs queued that should be able to
run but they wont start unless I restart the controller daemon. Other
jobs submitted more recently seem to working fine.

I can see from the slurmctld log file with debug=9 that they are not
being tested to see if they are runnable, does this mean that the daemon
has somehow forgotten about them?

I just restarted the daemon and they started immediately.

Any ideas how I can debug this if it happens again?

Cheers,