On Tue, Dec 9, 2008 at 10:41 AM, Bogdan Costescu <
[EMAIL PROTECTED]> wrote:

>
> Most likely these are jobs that should have finished but didn't; there are
> several situations in which they can appear, the 2 most common in my setup
> are:
> - one or more of the nodes from that job has died and the queueing system
> cannot terminate the job
> - a job not started via the TM interface (f.e. one using rsh to connect to
> remote nodes) for which the queueing system is not able to clean up
>


Thanks Bogdan! I think the first diagnosis is for me. I think we did have a
couple of dead nodes recently. I can test the hypothesis when I get a
negative time job on my queue again. Shouldn't be too long!

-- 
Rahul
_______________________________________________
mauiusers mailing list
[email protected]
http://www.supercluster.org/mailman/listinfo/mauiusers

Reply via email to