[torqueusers] queue rejects jobs when over 700 are submitted

Garrick Staples garrick at usc.edu
Tue Dec 15 09:40:14 MST 2009


On Tue, Dec 15, 2009 at 08:41:03AM -0700, Sarah Mulholland alleged:
> Are there any built in limits for how many jobs can be queued?  One of my users is getting his jobs rejected when he submits over 700 jobs.  The job exits with a nonzero status and error message, "qsub: Invalid credential".
> 
> In the archives this error message was attributed to running as root (we don't) or running as a user who isn't in /etc/passwd (we use yellowpages, and why would the first 600+ jobs be okay?).
> 
> Any suggestions?
> 
> We are submitting our jobs via torque qsub using torque-2.1.6.   We are using the maui 3.2.6p16 scheduler with a priority queue.  We did not set any limits on queue size or on how many jobs a user can submit.  I am posting to the torque list instead of the maui list since it's a torque error that we receive.
> 

Submitting at the same time? You are running out of priv ports.

Slow down the submission a bit and you can queue up hundreds of thousands.

-- 
Garrick Staples, GNU/Linux HPCC SysAdmin
University of Southern California

Life is Good!
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
Url : http://www.supercluster.org/pipermail/torqueusers/attachments/20091215/b5fca79b/attachment.bin 


More information about the torqueusers mailing list