Thanks, Changing the NEGOTIATOR_PRE_JOB_RANK default value solves my problem.
Werner:
In the negotiator, the NEGOTIATOR_PRE_JOB_RANK has more power
than the job's RANK _expression_. The default
NEGOTIATATOR_PRE_JOB_RANK _expression_ tries to minimize
fragmentation of a pool. I suspect this is why your job RANK
isn't having the effect you want. Try changing this knob in your
negotiator, and I think you'll see the policy you need.
-greg
On 12/14/20 9:02 AM, Thomas Hartmann
wrote:
Hi Werner,
are resources available with the desired kflops value?
If only slots on weaker (kflops) nodes are available, condor will
broker your jobs to these to get them running.
If you want to really run only on nodes with a minimum number of
kflops, you can add this as a minimum in the job requirements 9and
combine them with the rank _expression_ to aim for the highest
performing slots within your selection).
Cheers,
Thomas
On 10/12/2020 21.24, Werner KoppelstÃtter wrote:
Hi all,
I tried to use the rank _expression_ to define the best matching
slot for a job.
I added "Rank = kflops" to my submit-file as described in the
HTCondor Manual, but nothing changed.
My jobs were assigned to a slow machine and filled all the
slots. Then all the slots of a next machine were filled.
How can I adjust the order of the slots?
Thanks in advance,
Werner
_______________________________________________
HTCondor-users mailing list
subject: Unsubscribe
You can also unsubscribe by visiting
The archives can be found at:
_______________________________________________
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/
![[Computer Systems Lab]](http://www.cs.wisc.edu/pics/csl_logo.gif){kind=logo}