[slurm-users] [EXT] Weird issues with slurm's Priority

Sean Crosby scrosby at unimelb.edu.au
Wed Jul 8 09:36:19 UTC 2020


To disable the backfill scheduler, add/change

SchedulerType=sched/builtin

in your slurm.conf, and restart your slurmctld. The default SchedulerType
is sched/backfill

In terms of it is bad that you don't specify the Timelimit for jobs
accurately, disabling backfill scheduler (and not specifying timelimit
accurately) means that cores will go idle when there are jobs that could
use them. If you're happy with that, then all is fine.

Sean

--
Sean Crosby | Senior DevOpsHPC Engineer and HPC Team Lead
Research Computing Services | Business Services
The University of Melbourne, Victoria 3010 Australia



On Wed, 8 Jul 2020 at 18:47, zaxs84 <sciuscianebbia at gmail.com> wrote:

> *UoM notice: External email. Be cautious of links, attachments, or
>> impersonation attempts.*
>> ------------------------------
>> Can you see if it is set? Using (e.g. scontrol show job 337475 or sacct
>> -j 337475 -o Timelimit)
>>
>
> At the moment my users don't set the TimeLimit, because they don't know
> how long their jobs will take. If I check with sacct -o TimeLimit, I get :
>
>  Timelimit
> ----------
> 365-00:00+
>
> Is not setting a TimeLimit a bad thing with this configuration?
> Do you know how I can instruct the backfill scheduler NOT to run smaller
> jobs before my high priority job? Because what's happening right now is
> that all my high priority jobs are held in "PD" state, and low-priority
> jobs are still run.
>
> Many thanks!
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.schedmd.com/pipermail/slurm-users/attachments/20200708/0289f746/attachment.htm>


More information about the slurm-users mailing list