Re: [Condor-users] HELP: condor_schedd memory leak???

["Robert E. Parrott" <parrott@xxxxxxxxxxxxxxxx>]

I have stopped and restarted the condor_schedd multiple times, and  
each time I get the same behavior.
Essentially, it goes through its cycle, then the memory use sky- 
rockets and once it hits swap brings the host to its knees. There is  
no error reported in the logs.

It looks like a beahvior that would result from a corrupt data file,  
or similar, since it happens in a consistent fashion.

BTW, the version is 7.0.1 on this node, but the compute nodes are  
6.8.5 (I think).

rob


On Apr 7, 2008, at 12:41 PM, Dan Bradley wrote:

> If you stop and restart condor, the existing contents of your job  
> queue
> are preserved.  For vanilla universe jobs with sufficient lease times
> (20 minutes by default), the schedd should be able to reconnect to
> running jobs after it restarts without the jobs being interrupted by  
> the
> restart.
>
> Is your schedd logging anything strange?  Is it responding to  
> condor_q?
>
> --Dan
>
> Robert E. Parrott wrote:
>
> > Hi Folks,
> >
> > I have somewhat of an emergency situation.
> >
> > After a DOS attempt on ssh on our login node, and subsequent system
> > unresponsiveness an then reboot, the condor_schedd process now grows
> > to exhaust all of physical memory on the host (6 GB + at present).
> > This causes swapping issues etc.  ... it ain't pretty.
> >
> > Is there some appraoch I can take to try to resolve this issue,
> > without losing the previous queue of submitted jobs?
> >
> > thanks in advance for the quick response,
> > rob
> >
> >
> > ==========================
> > Robert E. Parrott, Ph.D. (Phys. '06)
> > Associate Director, Grid and
> >       Supercomputing Platforms
> > Project Manager, CrimsonGrid Initiative
> > Harvard University Sch. of Eng. and App. Sci.
> > Maxwell-Dworkin  211,
> > 33 Oxford St.
> > Cambridge, MA 02138
> > (617)-495-5045
> >
> > _______________________________________________
> > 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/


==========================
Robert E. Parrott, Ph.D. (Phys. '06)
Associate Director, Grid and
       Supercomputing Platforms
Project Manager, CrimsonGrid Initiative
Harvard University Sch. of Eng. and App. Sci.
Maxwell-Dworkin  211,
33 Oxford St.
Cambridge, MA 02138
(617)-495-5045