Re: [HTCondor-users] jobs to start on busy machines first

[Marcus Ebert <mebert@xxxxxxx>]

Hi Rolf,

You could use NEGOTIATOR_POST_JOB_RANK.
I didn't find a way to get the actual number of running condor jobs to be 
used, but depending what your jobs do you could prefer to send jobs 
to machine with the the highest load (~=most running condor jobs):

NEGOTIATOR_POST_JOB_RANK = (RemoteOwner =?= UNDEFINED) * 
(CondorLoadAvg - 1.0e10*(Offline=?=True))

Cheers,
 Marcus

On Tue, 7 Aug 2018, Rolf Seuster wrote:

> Hi all,
>
> we have a small cluster of worker nodes. For testing, I start
> short jobs (<1min) every minute. From the manual, I would assume
> that the jobs populate first all slots on one machine, then move to
> the next machine. That's not what we observe. All machines
> get more or less the same number of jobs.
>
> We use
>
> #condor_version
> $CondorVersion: 8.6.9 Jan 03 2018 BuildID: 428149 $
> $CondorPlatform: x86_64_RedHat7 $
>
> and according to the manual, this is the default:
>
> # condor_config_val NEGOTIATOR_PRE_JOB_RANK
> (10000000 * My.Rank) + (1000000 * (TARGET.RemoteOwner =?= UNDEFINED)) -
> (100000 * TARGET.Cpus) - TARGET.Memory
>
> Question: How do I configure condor to start jobs on the least number of
> workers ?
>
> Thanks, Rolf
>
>
>
> _______________________________________________
> HTCondor-users mailing list
> To unsubscribe, send a message to htcondor-users-request@xxxxxxxxxxx with a
> subject: Unsubscribe
> You can also unsubscribe by visiting
> https://lists.cs.wisc.edu/mailman/listinfo/htcondor-users
>
> The archives can be found at:
> https://lists.cs.wisc.edu/archive/htcondor-users/