Re: [HTCondor-users] help needed to troubleshoot why suddenly an user is running less jobs than it used to

[jcaballero.hep@xxxxxxxxx]

Hi,

I am still in the middle of the video. Quite helpful so far, indeed.
Thanks a lot for that.

Before I finish it, I already have a question.
Could GROUP_SORT_EXPR and GROUP_ACCEPT_SURPLUS=True "mess" with the
group quotas?
For example, let's say you have 2 groups, G1 and G2, and each one of
them has the same quota: 50% of the pool.
GROUP_QUOTA_DYNAMIC_group_G1 = 0.5
GROUP_QUOTA_DYNAMIC_group_G2 = 0.5

Is it possible that, if both have IDLE jobs, G1 takes over more than
50% of the pool and kicks G2 out if G1 is considered first by the
Negotiator and surplus is enabled?

Cheers,
Jose


El lun, 1 mar 2021 a las 16:14, Greg Thain (<gthain@xxxxxxxxxxx>) escribiÃ:
>
>
> On 3/1/21 9:59 AM, jcaballero.hep@xxxxxxxxx wrote:
> > Hi,
> >
> > condor_userprio indeed seems to be telling me something. Even though
> > all users have the same Priority Factor, the Effective Priorities are
> > very different. I need to learn how the Effective Priorities values
> > are calculated, and why are so different.
> > Thanks a lot for that tip !!
>
>
> Jose:
>
> If you'd like to know exactly how Effective Priorities are calculated,
> we put up a YouTube video of a Condor Week talk we did on the subject at:
>
> https://www.youtube.com/watch?v=NNnrCjFV0tM
>
> -greg
>
> >
> > Cheers,
> > Jose
> >
> > El vie, 26 feb 2021 a las 10:32, Jose Caballero
> > (<jcaballero.hep@xxxxxxxxx>) escribiÃ:
> >> Hi,
> >>
> >> I am going to have a look. Thanks for the hint.
> >>
> >> Cheers,
> >> Jose
> >>
> >>
> >> El vie, 26 feb 2021 a las 9:21, <thomas.hartmann@xxxxxxx> escribiÃ:
> >>> Hi Jose,
> >>>
> >>> can you check, what
> >>>     condor_userprio
> >>> says for your users to see Condor's current opinion about the current
> >>> usage/priority?
> >>>
> >>> Cheers,
> >>>     Thomas
> >>>
> >>> On 26/02/2021 09.24, jcaballero.hep@xxxxxxxxx wrote:
> >>>> Hello,
> >>>>
> >>>>
> >>>> Running condor 8.6.13 on Scientific Linux 7.9.
> >>>>
> >>>> Without changing the Central Managers configuration, as far as I can
> >>>> tell, suddenly one of the users (A) is running less and less jobs in
> >>>> our farm. Order of magnitude of 40% less jobs than it used to.
> >>>>
> >>>> It has been always setup to have the same fairshare values than a
> >>>> couple other users (B and C). However, B and C are both now running
> >>>> more than twice than A. This has been happening since a couple of
> >>>> weeks ago.
> >>>>
> >>>> The Requirement expression for A's jobs hasn't changed.
> >>>>
> >>>> The user always has about 1000 IDLE jobs waiting, so it is not a
> >>>> shortage of waiting jobs.
> >>>>
> >>>>
> >>>> Any tip or suggestion on how to troubleshoot this is more than
> >>>> welcome. I am running out of ideas where to look.
> >>>> Thanks a lot in advance.
> >>>> Cheers,
> >>>> Jose
> >>>> _______________________________________________
> >>>> 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/
> >>>>
> >>> _______________________________________________
> >>> 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/
> > _______________________________________________
> > 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/
> _______________________________________________
> 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/