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

Re: [Condor-users] Sharing HPC nodes with Condor



Set a couple things in your condor_config files (s/PBS/SGE/g):

After 6.7.18, you'll have to do the following to let this work:
ENABLE_RUNTIME_CONFIG = TRUE

Then
STARTD_SETTABLE_ATTRS_OWNER = PBSRunning

PBSRunning = False
PURDUE_RCAC_START_NOPBS = ( $(PBSRunning) == False )

START = $(PURDUE_RCAC_START) && $(PURDUE_RCAC_START_NOPBS)


Then in your prologue, when PBS starts a job, mark it unavailable to Condor, and run condor_vacate if necessary. Or, Dan Bradley suggested to me at Condor Week that it might be a better option to just manually suspend any running jobs when PBS butts in, if your environment will allow it. And then in the epilogue, after PBS is done, mark it as available to Condor again.

Prologue:

        if [ -x /opt/condor/bin/condor_config_val ]; then
                /opt/condor/bin/condor_config_val -rset \
                        -startd PBSRunning=True > /dev/null
                /opt/condor/sbin/condor_reconfig -startd > /dev/null
                if ( /opt/condor/bin/condor_status -claimed \
                        -direct $(hostname) 2>/dev/null \
                        | grep -q Machines )
                then
                        /opt/condor/sbin/condor_vacate > /dev/null
                fi
        fi



Epilogue:

                if [ -x /opt/condor/bin/condor_config_val ]; then
                        /opt/condor/bin/condor_config_val -rset \
                                -startd PBSRunning=False > /dev/null
/opt/condor/sbin/condor_reconfig -startd > / dev/null
                fi


On May 9, 2007, at 2:29 PM, Natarajan, Senthil wrote:

Hi Preston,
Thanks for offering your help.
Yes, we are trying to do something like that.

One of the department has a cluster managed by Sun Grid Engine, and they plan to share around 20 nodes to condor, so how can we configure in such a way so that, if the nodes are not used by Sun Grid Engine then condor can use it. Suppose if the jobs submitted to Sun Grid Engine needs nodes have being used by condor, then condor should terminate the job and send
it back to central manager like if somebody login the machine.

Thanks,
Senthil

-----Original Message-----
From: condor-users-bounces@xxxxxxxxxxx
[mailto:condor-users-bounces@xxxxxxxxxxx] On Behalf Of Preston Smith
Sent: Tuesday, May 08, 2007 9:38 PM
To: Condor-Users Mail List
Subject: Re: [Condor-users] Sharing HPC nodes with Condor

Sounds like you're wanting to backfill behind your primary scheduler
with Condor, correct? We do this on all of our PBS clusters at
Purdue, I can help with that if that's what you're looking for.

-Preston

On May 8, 2007, at 5:37 PM, Natarajan, Senthil wrote:

Hi,

We are planning to use 20 nodes from a High performance Computing
cluster to run condor jobs if there is no jobs running on the HPC
nodes.

Suppose if HPC users submits jobs and it requires the nodes where
condor jobs are running, condor jobs should be terminated and send
back to the central manager, like if some real user login the
machine. I was wondering how to configure condor to do this. Could
you please let me know.

Thanks,

Senthil

_______________________________________________
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



--
Preston Smith    <psmith@xxxxxxxxxx>
Systems Research Engineer
Rosen Center for Advanced Computing, Purdue University



_______________________________________________
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

--
Preston Smith  <psmith@xxxxxxxxxx>
Systems Research Engineer
Rosen Center for Advanced Computing, Purdue University