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

Re: [Condor-users] MaxJobRetirementTime




Yes, higher priority jobs may have to wait around for lower priority jobs to finish when you grant long retirement times. When this happens, the schedd will periodically go back to the negotiator to get a new match. Therefore, if you configure your negotiator to steer jobs to idle or quickly preemptable machines, you can achieve better scheduling. The configuration expressions that you may wish to tune are the following:

REQUEST_CLAIM_TIMEOUT

NEGOTIATOR_PRE_JOB_RANK

NEGOTIATOR_POST_JOB_RANK

--Dan

On Jan 30, 2006, at 1:27 AM, Andrew Stubbings wrote:

MaxJobRetirementTime is described in the manual as the 'maximum time in
seconds to let this job run uninterrupted before kicking it off when it
is being preempted.' If MaxJobRetirementTime is set to some large value,
does this mean that a job which has a higher priority than the one
currently running on the node has to wait until the MaxJobRetirementTime
is reached or will the negotiator choose another machine that has a
lower MaxJobRetirementTime hence allowing a waiting job to be scheduled
on a resource that is going to allow it to be allocated faster?

Cheers,
Andrew

--
Andrew Stubbings
Grid Research, Inc.
Tsukuba, Japan
Tel: +81 29 863 3462
E-mail: ajs@xxxxxxxxxxxxxxxxx
www.grid-research.com www.bestsystems.co.jp


_______________________________________________
Condor-users mailing list
Condor-users@xxxxxxxxxxx
https://lists.cs.wisc.edu/mailman/listinfo/condor-users