Re: [Condor-users] nfs and condor

[Steven Timm <timm@xxxxxxxx>]

On Sun, 12 Jun 2011, Mag Gam wrote:

> How are you adding the writes after each job?

I was wondering if there are any tricks I can do such as writing log
files to local filesystems instead of NFS. Can I place, "log",
"output", and "error" to local directories?


Yes you can and should.  Some very busy sites even get a SSD
device to put them on.
There are also tricks you can do with condor_transfer_files
to transfer your executable to the worker nodes aand
to bring back the stdout/stderr files.

Steve Timm



On Sun, Jun 12, 2011 at 6:27 AM, Erik Aronesty <erik@xxxxxxx> wrote:
> I've noticed one issue with NFS that may or may not be related.  Some NFS
> systems are "clusters" of storage nodes that are striped across the set of
> machines.   In these NFS systems, a client connects to one node, and the
> information may propagate to other nodes after some time lag.
>
> If you're using condor, this means that your NFS mounts on the submitter and
> execution nodes may be out of sync...since each mount may be to a different
> physical storage cluster node.
>
> Adding small delays after writing files that both the execution and
> submitter node need to see fixed the problem for us.
>
> On Sun, Jun 12, 2011 at 6:15 AM, Mag Gam <magawake@xxxxxxxxx> wrote:
> > At our university we are a heavy NFS user. When we run run long jobs
> > with condor and there is a performance problem with our home
> > directories (which on are NFS). It seems the job gets requeued.
> >
> > I was wondering if anyone else out there have a similar problem and
> > what they did to fix it :-)
> > _______________________________________________
> > 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:
> > https://lists.cs.wisc.edu/archive/condor-users/
>
>
> _______________________________________________
> 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:
> https://lists.cs.wisc.edu/archive/condor-users/
_______________________________________________
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:
https://lists.cs.wisc.edu/archive/condor-users/

--
------------------------------------------------------------------
Steven C. Timm, Ph.D  (630) 840-8525
timm@xxxxxxxx  http://home.fnal.gov/~timm/
Fermilab Computing Division, Scientific Computing Facilities,
Grid Facilities Department, FermiGrid Services Group, Group Leader.
Lead of FermiCloud project.