[torqueusers] several questions

[Garrick Staples]

On Fri, Sep 16, 2005 at 01:33:58PM -0600, Doug Smith alleged:
> 
> 
> Hello I am new to the Torque world and have a few questions:
> 
> 1. Is there a definitive Torque reference that outlines all the resource 
> attributes for queue configuration? The main one I have been using is 
> clusterresource but the documentation on configuring queues seems to be 
> a bit thin.

The 'pbs_queue_attributes' manpage should be definitive.

 
> 2. I have noticed a problem on our cluster where qdel removes the job 
> from the queue but leaves processes running on the individual nodes. Are 
> others having this issue? shouldn't qdel be killing cancelled jobs off 
> of the nodes?

Only the "top level" process in the job is signalled to be killed.
You'll probably need an epilogue script to clean up the jobs.

 
> 3. Jobs submitted to a queue with defined with a resource_max.cput limit 
> aren't getting killed by the schedular after their time expires why? Is 
> there something else that needs to be set inconjunction with this setting?

Which scheduler?  maui/moab should definitely handle this.  I'm not sure
about pbs_sched.

-- 
Garrick Staples, Linux/HPCC Administrator
University of Southern California
-------------- 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/20050920/07ceeffd/attachment.bin