[slurm-users] Slurm fair share priority not being applied

[Bruno Santos]

Loris, I think you hit the nail on the head.

Running sshare -l only shows the root user:
             Account       User  RawShares  NormShares    RawUsage
 NormUsage  EffectvUsage  FairShare    LevelFS
GrpTRESMins                    TRESRunMins
-------------------- ---------- ---------- ----------- -----------
----------- ------------- ---------- ----------
------------------------------ ------------------------------
root                                          0.000000           0
        1.000000
 cpu=0,mem=0,energy=0,node=0

So I am guessing the user information is indeed not being linked to the
slurmdbd. What do I need to do to set up this?
To be honest the documentation is very sparse in details of how to set up
the slurmdb.

Best,
Bruno

On 1 December 2017 at 06:54, Loris Bennett <loris.bennett at fu-berlin.de>
wrote:

> Hi Bruno,
>
> Bruno Santos <bacmsantos at gmail.com> writes:
>
> > Hi everyone,
> >
> > I have recently set-up slurm to use priority/multifactor giving fair
> share the major weight.
> >
> > Since the queue is up I have submitted about 1500 small jobs and just
> today 2 other users jumped on the queue with their first job. However, it
> seems that slurm is not
> > taking into account the fair share factor as they still went to the
> bottom of the queue.
> >
> > Looking at sprio it looks like the FAIRSHARE is still 0 for all of us.
> Shouldn't they be getting a much higher for that?
> > JOBID USER PRIORITY AGE FAIRSHARE JOBSIZE PARTITION QOS NICE TRES
> > 7435 bsantos 1770 40 0 731 1000 0 0
> > 7436 bsantos 1770 40 0 731 1000 0 0
> > 7445 pruperao 1735 15 0 721 1000 0 0
> > 7446 fkhan 1706 14 0 692 1000 0 0
>
> [snip (21 lines)]
>
> You have used 'sacctmgr' to assign shares to the associations
> corresponding to the users, right?  Check the output of 'sshare' to see
> information on how the shares have been set up and what the actual
> fairshare values are.  Unfortunately, the web page for the multifactor
> plugin
>
>   https://slurm.schedmd.com/priority_multifactor.html
>
> doesn't seem to mention setting things up with 'sacct'.
>
> Cheers,
>
> Loris
> --
> Dr. Loris Bennett (Mr.)
> ZEDAT, Freie Universität Berlin         Email loris.bennett at fu-berlin.de
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.schedmd.com/pipermail/slurm-users/attachments/20171201/8f90c787/attachment.html>