Restarting cloud-worker left in-flight requests dangling

[Iain Lane]

I'm not sure how this happened, as any unclean exit is supposed to cause the jobs to be put back in the amqp queue.

ubuntu@juju-prod-ues-proposed-migration-machine-11:~$ uptime
 13:26:25 up 1 day, 43 min, 1 user, load average: 3.52, 2.79, 2.57

The autopkgtest-cloud-worker machine was restarted for maintenance, and jobs/workers which were running at the time appear to have been SIGTERMed and they weren't requeued. They were left dangling (RUNNING) in proposed-migration.

Also the status page was frozen in time - logtail and duration - for these jobs at the time the event happened.