[Mauiusers] insufficient idle procs available ?

Jan Ploski Jan.Ploski at offis.de
Tue Jan 22 06:02:39 MST 2008


"Itay M" <itaym.tau at gmail.com> schrieb am 01/22/2008 01:52:38 PM:

> Here is the qstat -f post. Note that node28 has physically 4 procs but
> the moment it only runs these two jobs (each one with 1 proc)  :
> 
> /==============================/
> $ qstat -f 191768 191769
> Job Id: 191768.cluster
>     Job_Name = vpu.GEP_1
>     Job_Owner = ad_user at cluster
>     resources_used.cput = 16:34:45
>     resources_used.mem = 572892kb
>     resources_used.vmem = 689168kb
>     resources_used.walltime = 19:37:00
>     job_state = R
>     queue = heavy
>     server = cluster
>     Checkpoint = u
>     ctime = Mon Jan 21 15:54:06 2008
>     Error_Path = ...(deleted)
>     exec_host = node28/1
>     Hold_Types = n
>     Join_Path = n
>     Keep_Files = n
>     Mail_Points = a
>     mtime = Tue Jan 22 09:17:58 2008
>     Output_Path = ...(deleted)
>     Priority = 0
>     qtime = Mon Jan 21 15:54:06 2008
>     Rerunable = True
>     Resource_List.cput = 240:00:00
>     Resource_List.mem = 512mb
>     Resource_List.ncpus = 1
>     Resource_List.neednodes = 1
>     Resource_List.nice = 13
>     Resource_List.nodect = 1
>     Resource_List.nodes = 1
>     Resource_List.walltime = 100:00:00

Get rid of the 'ncpus' piece (by changing the job scripts and/or 
server/queue parameters with qmgr). I can't tell you why exactly, but I am 
quite sure that ncpus caused problems for me in the past. I'm also 
completely sure that this attribute is not needed (based on our 
TORQUE+Maui installation, which successfully runs a mix of single and 
multi-processor jobs).

Regards,
Jan Ploski


More information about the mauiusers mailing list