[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Condor-users] node priority



Hi Pasquale,

   This works for me
      NEGOTIATOR_POST_JOB_RANK = ( -1 * Memory )

tsan
On Sun, 20 Aug 2006 13:35:14 -0700, Pasquale Tricarico wrote
> Thank you, Erik.
> 
> OK, so I changed NEGOTIATOR_PRE_JOB_RANK from its default value of
> $(UWCS_NEGOTIATOR_PRE_JOB_RANK) to $(MEMORY), but this didn't change
> the standard behaviour. So I changed it to
> 
> NEGOTIATOR_PRE_JOB_RANK = -$(MEMORY)
> 
> but now the negotiator complains:  ERROR "Error parsing
> NEGOTIATOR_PRE_JOB_RANK expression: -" at line 304 in file
> matchmaker.C
> 
> I'm sorry, I'm still learning Condor, can you help me figure out how
> to change NEGOTIATOR_PRE_JOB_RANK so that the jobs are started on the
> nodes with low memory first?
> 
> Thanks,
> Pasquale
> 
> On 8/20/06, Erik Paulson <epaulson@xxxxxxxxxxx> wrote:
> > On Sun, Aug 20, 2006 at 06:01:02AM -0700, Pasquale Tricarico wrote:
> > > Hi,
> > >
> > > I'm using Condor 6.8 to manage a cluster composed by a total of 10
> > > nodes: 5 with 4 GB of ram (node01 to node05), and 5 with 2 GB of ram
> > > (node06 to node10). This is an X86-64 cluster, so no checkpointing or
> > > migration is available (yet). Since the nodes appear in this same
> > > order in the queue, when the first job is submitted it usually starts
> > > on node01, then node02 and so on. The "problem" is that usually these
> > > jobs start on the high memory nodes even when this is not a
> > > requirement, i.e. even when the job would run perfectly with only say
> > > 20 MB of ram. So what I'd like to do is configure condor so that when
> > > there is not special memory requirement, a job is started on one of
> > > the five nodes with low memory (node06 to node10). This because I'd
> > > like to keep the high-memory nodes available for parallel jobs and for
> > > jobs that really require a lot of memory. Can you please help me to
> > > figure out how to achieve this configuration? Of course, when
> > > migration will be available this should not be a problem anymore, but
> > > this is not the case yet. And yes, renaming the nodes is not really an
> > > option for me, sorry.
> > >
> >
> > http://www.cs.wisc.edu/condor/manual/v6.8.0/3_3Configuration.html#13009
> >
> > -Erik
> >
> > > Thank you.
> > >
> > > Pasquale Tricarico
> > > Associate Research Scientist
> > > Planetary Science Institute
> > > _______________________________________________
> > > Condor-users mailing list
> > > To unsubscribe, send a message to condor-users-request@xxxxxxxxxxx with a
> > > subject: Unsubscribe
> > > You can also unsubscribe by visiting
> > > https://lists.cs.wisc.edu/mailman/listinfo/condor-users
> > >
> > > The archives can be found at either
> > > https://lists.cs.wisc.edu/archive/condor-users/
> > > http://www.opencondor.org/spaces/viewmailarchive.action?key=CONDOR
> > _______________________________________________
> > Condor-users mailing list
> > To unsubscribe, send a message to condor-users-request@xxxxxxxxxxx with a
> > subject: Unsubscribe
> > You can also unsubscribe by visiting
> > https://lists.cs.wisc.edu/mailman/listinfo/condor-users
> >
> > The archives can be found at either
> > https://lists.cs.wisc.edu/archive/condor-users/
> > http://www.opencondor.org/spaces/viewmailarchive.action?key=CONDOR
> >
> _______________________________________________
> Condor-users mailing list
> To unsubscribe, send a message to condor-users-request@xxxxxxxxxxx with a
> subject: Unsubscribe
> You can also unsubscribe by visiting
> https://lists.cs.wisc.edu/mailman/listinfo/condor-users
> 
> The archives can be found at either
> https://lists.cs.wisc.edu/archive/condor-users/
> http://www.opencondor.org/spaces/viewmailarchive.action?key=CONDOR


---
Room 411, Institute of Physics, Academia Sinica, Nankang, Taipei, Taiwan, 11529, R.O.C.

[TEL: 886-2-2788-0058#4045] [Mobil: 0955092291] [FAX: 886-2-2788-9828] [MSN: yuwanyi@xxxxxxx] [Skype: tsanlung]