[Condor-users] priority weirdness

[Grant Goodyear <grant@xxxxxxxxxxxxxxxxx>]

We're running Condor 6.8.5 on a small windows cluster, and we've been using
accounting groups to schedule our various users' jobs.  It's been working
beautifully until now, and we haven't made any changes to the Condor
configuration, so I'm at a loss to understand what's happened.

What we're seeing is that we currently have two users with jobs in the queue. 
One, who has an effective priority of 50, has some 500 jobs either running or in
the queue.  The other, who has an effective priority of 2, has a half dozen jobs
that aren't getting started when currently-running jobs finish.  Instead,
priority seems to be going to the priority=50 jobs instead.  Worse, we weren't
seeing this behavior before now, and we've been running Condor non-stop since
May or so.

Any suggestions on how we can isolate the cause of this problem?

Thanks,
Grant Goodyear