[slurm-users] diagnosing why interactive/non-interactive job waits are so long with State=MIXED

[Robert Kudyba via slurm-users]

At the moment we have 2 nodes that are having long wait times. Generally this is when the nodes are fully allocated. What would be the other reasons if there is still enough available memory and CPU available, that a job would take so long? Slurm version is  23.02.4 via Bright Computing. Note the compute nodes have hyperthreading enabled but that should be irrelevant. Is there a way to determine what else could be holding jobs up?

srun --pty  -t 0-01:00:00 --nodelist=node001 --gres=gpu:1 -A ourts -p short /bin/bash
srun: job 672204 queued and waiting for resources

 scontrol show node node001

NodeName=m001 Arch=x86_64 CoresPerSocket=48
   CPUAlloc=24 CPUEfctv=192 CPUTot=192 CPULoad=20.37
   AvailableFeatures=location=local
   ActiveFeatures=location=local
   Gres=gpu:A6000:8
   NodeAddr=node001 NodeHostName=node001 Version=23.02.4
   OS=Linux 5.14.0-70.13.1.el9_0.x86_64 #1 SMP PREEMPT Thu Apr 14 12:42:38 EDT 2022
   RealMemory=1031883 AllocMem=1028096 FreeMem=222528 Sockets=2 Boards=1
   State=MIXED ThreadsPerCore=2 TmpDisk=0 Weight=1 Owner=N/A MCS_label=N/A
   Partitions=ours,short
   BootTime=2024-04-29T16:18:30 SlurmdStartTime=2024-05-18T16:48:11
   LastBusyTime=2024-06-03T10:49:49 ResumeAfterTime=None
   CfgTRES=cpu=192,mem=1031883M,billing=192,gres/gpu=8
   AllocTRES=cpu=24,mem=1004G,gres/gpu=2,gres/gpu:a6000=2
   CapWatts=n/a
   CurrentWatts=0 AveWatts=0
   ExtSensorsJoules=n/s ExtSensorsWatts=0 ExtSensorsTemp=n/s

grep 672204 /var/log/slurmctld
[2024-06-04T15:50:35.627] sched: _slurm_rpc_allocate_resources JobId=672204 NodeList=(null) usec=852

[Ryan Novosielski via slurm-users]

This is relatively true of my system as well, and I believe it’s that the backfill schedule is slower than the main scheduler.

--
#BlackLivesMatter

____
|| \\UTGERS,     |---------------------------*O*---------------------------
||_// the State  |         Ryan Novosielski - novo...@rutgers.edu
|| \\ University | Sr. Technologist - 973/972.0922 (2x0922) ~*~ RBHS Campus
||  \\    of NJ  | Office of Advanced Research Computing - MSB A555B, Newark
     `'

--
slurm-users mailing list -- slurm...@lists.schedmd.com
To unsubscribe send an email to slurm-us...@lists.schedmd.com

[Robert Kudyba via slurm-users]

Thanks for the quick response Ryan!

Are there any recommendations for bf_ options from https://slurm.schedmd.com/sched_config.html that could help with this? bf_continue? Decreasing bf_interval= to a value lower than 30?

[Ryan Novosielski via slurm-users]

We do have bf_continue set. And also bf_max_job_user=50, because we discovered that one user can submit so many jobs that it will hit the limit of the number it’s going to consider and not run some jobs that it could otherwise run.

[Loris Bennett via slurm-users]

Ryan Novosielski via slurm-users <slurm...@lists.schedmd.com> writes:

> We do have bf_continue set. And also bf_max_job_user=50, because we discovered that one user can submit so many jobs that it will hit the limit of the number
> it’s going to consider and not run some jobs that it could otherwise run.
>
> On Jun 4, 2024, at 16:20, Robert Kudyba <rku...@fordham.edu> wrote:
>
> Thanks for the quick response Ryan!
>
> Are there any recommendations for bf_ options from https://slurm.schedmd.com/sched_config.html that could help with this? bf_continue? Decreasing
> bf_interval= to a value lower than 30?

Your bf_window may be too small. From 'man slurm.conf':

bf_window=#

The number of minutes into the future to look when considering
jobs to schedule. Higher values result in more overhead and
less responsiveness. A value at least as long as the highest
allowed time limit is generally advisable to prevent job
starvation. In order to limit the amount of data managed by
the backfill scheduler, if the value of bf_window is increased,
then it is generally advisable to also increase bf_resolution.
This option applies only to SchedulerType=sched/backfill.
Default: 1440 (1 day), Min: 1, Max: 43200 (30 days).

--
Dr. Loris Bennett (Herr/Mr)
FUB-IT (ex-ZEDAT), Freie Universität Berlin

[Robert Kudyba via slurm-users]

Your bf_window may be too small.  From 'man slurm.conf':

  bf_window=#

         The number of minutes into the future to look when considering
         jobs to schedule.  Higher values result in more overhead and
         less responsiveness.  A value at least as long as the highest
         allowed time limit is generally advisable to prevent job
         starvation.  In order to limit the amount of data managed by
         the backfill scheduler, if the value of bf_window is increased,
         then it is generally advisable to also increase bf_resolution.
         This option applies only to SchedulerType=sched/backfill.
         Default: 1440 (1 day), Min: 1, Max: 43200 (30 days).

So since we have a 5 day option should bf_window=7200? What should bf_resolution be set to then?

But how does this affect/improve wait times?