[torquedev] torque/PBS <-> job attribute mem

Jens Kuehlberg jens.kuehlberg at gmx.de
Thu Aug 31 11:04:45 MDT 2006


Hi,

that's right I use pbs_sched. When I submit the job, then I can see the correct attribute in the queue (qstat -f ). How can I discover if PBS is working with this attributes, or not?

Bye
Jens Kuehlberg


    Job_Name = test.sub
    Job_Owner = griduser at uni-vo2-cs
    job_state = R
    queue = workq
    server = uni-vo2-cs
    Checkpoint = u
    ctime = Thu Aug 31 18:52:25 2006
    Error_Path = uni-vo2-cs:/usr/local/FILESPACE_PBS/test.sub.e10
    exec_host = uni-vo2-cn1/0
    Hold_Types = n
    Join_Path = n
    Keep_Files = n
    Mail_Points = a
    mtime = Thu Aug 31 18:52:25 2006
    Output_Path = uni-vo2-cs:/usr/local/FILESPACE_PBS/test.sub.o10
    Priority = 0
    qtime = Thu Aug 31 18:52:25 2006
    Rerunable = True
    Resource_List.mem = 512mb
    Resource_List.neednodes = 1
    Resource_List.nodect = 1
    Resource_List.nodes = 1
    session_id = 1035
    substate = 42
    Variable_List = PBS_O_HOME=/home/griduser,PBS_O_LOGNAME=griduser,
        PBS_O_PATH=/usr/local/torque/bin/:/usr/local/j2sdk/bin:/usr/local/bin
        :/usr/bin:/bin:/usr/bin/X11:/usr/games,PBS_O_MAIL=/var/mail/griduser,
        PBS_O_SHELL=/bin/bash,PBS_O_HOST=uni-vo2-cs,
        PBS_O_WORKDIR=/usr/local/FILESPACE_PBS,PBS_O_QUEUE=workq
    euser = griduser
    egroup = griduser
    hashname = 10.uni-vo2-
    queue_rank = 1
    queue_type = E
    comment = Job started on Thu Aug 31 at 18:52
    etime = Thu Aug 31 18:52:25 2006
    submit_args = -l mem=512mb



-------- Original-Nachricht --------
Datum: Wed, 30 Aug 2006 20:28:49 -0600
Von: Garrick Staples <garrick at clusterresources.com>
An: torquedev at supercluster.org
Betreff: Re: [torquedev] torque/PBS <-> job attribute mem

> On Wed, Aug 30, 2006 at 08:11:22PM +0200, Jens Kuehlberg alleged:
> > Hi,
> > 
> > I have got some compute nodes and a manage node in a testbed. The
> compute nodes are heterogeneous (MEM,CPU,..)
> > 
> > The Submit-Job' are described by job attributes. Torque don't processed
> mem attributes correctly. Some special jobs need many memories, so I'm
> setting mem attributes. After some attemps, torque schedules the job on every
> node. Torque don't select a node with enough memory. But node name attribute
> schedule works fine.
> > 
> > Could anybody help me?
> > 
> > Best regards
> > Jens Kuehlberg
> 
> That is up to the scheduler, it is not internal to TORQUE.  I'm not even
> sure if pbs_sched supports it.  Which scheduler are you using?
> 
> _______________________________________________
> torquedev mailing list
> torquedev at supercluster.org
> http://www.supercluster.org/mailman/listinfo/torquedev

-- 


"Feel free" – 10 GB Mailbox, 100 FreeSMS/Monat ...
Jetzt GMX TopMail testen: http://www.gmx.net/de/go/topmail


More information about the torquedev mailing list