[torqueusers] a newcomer in trouble, again

Claudio Destri Claudio.Destri at mib.infn.it
Thu Jul 3 14:24:32 MDT 2008


I have torque-2.1.6 installed and running on a cluster.
Suppose the following:

Qmgr: list queue small
Queue small
         queue_type = Execution
         total_jobs = 0
         state_count = Transit:0 Queued:0 Held:0 Waiting:0 Running:0  Exiting:0
         acl_host_enable = False
         resources_max.ncpus = 12
         resources_default.nodes = 1
         acl_group_enable = True
         acl_groups = roma2
         mtime = Wed Jun 25 15:46:25 2008
         resources_available.nodes = c0-10
         resources_assigned.ncpus = 0
         resources_assigned.nodect = 0
         enabled = True
         started = True


Is it correct to expect that job submitted to queue small should be
restricted to run on node c0-10 (which has 16 cores) and use at most 12
cores?

I ask because no such limitation is really taking place: jobs can be 
subitted on any node and with any usage of cores, for instance with
the lines:

 	#PBS -q small
 	#PBS -l nodes=c0-8:ppn=4+c0-9:ppn=4+c0-10:ppn=16

in the submission script.

So I take this to be a bug in my torque installation.
Can anyone help me out on this, suggesting new steps?


 			CLAUDIO DESTRI   Claudio.Destri at mib.infn.it
                         DIPARTIMENTO DI FISICA G. OCCHIALINI,
                         UNIVERSITA` MILANO-BICOCCA
                         PIAZZA DELLA SCIENZA 3, 20126 MILANO
                         tel: (39)(02)64482539  fax: (39)(02)64482585


More information about the torqueusers mailing list