[torqueusers] Torque only assigning one task per node

[Matt Marziale]

I'm trying to get torque up and running on a cluster I built. I'm using 
Maui for the scheduler, but every time I run a job, it appears, at least 
to the Torque-aware version of mpiexec from the Ohio Supercomputer 
Center and qstat, that there is only one task per node, no matter how 
many processors I specify in qsub. The following is the output from list 
queue batch:

         queue_type = Execution
         total_jobs = 1
         state_count = Transit:0 Queued:0 Held:0 Waiting:0 Running:1 
Exiting:0
         resources_max.ncpus = 36
         resources_default.nodect = 1
         resources_default.walltime = 720:00:00
         mtime = Wed Nov 28 23:21:53 2012
         resources_available.ncpus = 36
         resources_available.nodect = 2
         resources_assigned.nodect = 1
         enabled = True
         started = True


Despite mpiexec and qstat only showing one task per node allocated, 
showq marks all processors requested as active. Does anyone know what I 
might be doing wrong?

-- 
Matt Marziale
mmarzial at purdue.edu