[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Condor-users] Jobs remain idle long time!



On Aug 25, 2005, at 4:44 AM, Vinicius da Cunha Martins Borges wrote:

The jobs submitted by globus 4.0.0(GRAM - globusrun-ws) usually remain idle.
...
Would it be the universe vanilla? What is the universe I must use when I
submitted by globus? what should permissions exist in the directory /
usr/local/condor/local.labweb02/spool?

For the jobs submitted by Globus, there are a few things you can look at. Run condor_q -analyze on them to see if they're not matching.


It's possible that the jobs are being run repeatedly and failures keep happening (for example, if the executable don't exist). Run condor_q -l on the jobs and look for JobRunCount. That's the number of times Condor has tried to start the job. If it's present and larger than 3 or 4, then you have a problem.

Also, try submitting equivalent jobs by hand with condor_submit. See if those jobs behave differently.

+----------------------------------+---------------------------------+
|            Jaime Frey            |  Public Split on Whether        |
|        jfrey@xxxxxxxxxxx         |  Bush Is a Divider              |
|  http://www.cs.wisc.edu/~jfrey/  |         -- CNN Scrolling Banner |
+----------------------------------+---------------------------------+