Hello,
I recently set up a small cluster at work using Warewulf/Slurm. Currently,
I am not able to get the scheduler to
work well with GPU's (Gres).
While slurm is able to filter by GPU type, it allocates all the GPU's on
the node. See below:
[abhiram@whale ~]$ srun --gres=gpu:p100:2 -n 1 --part
Hi,
We want to modify a job's priority right after it is submitted automatically in
slurm. Job submit plugin was considered, but it seems it haven't even had the
jobid of a new job yet (probably because the job have't actually been submitted
yet), while we need to use scontrol with jobid to get
Thanks for the replies,
I can dump the file and it has all the data I need. I really appreciate the
help.
Chad
Chad Julius
Cyberinfrastructure Engineer Specialist
South Dakota State University
Old Horticulture 205
Brookings, SD 57007
Phone: 605.688.5767
Email: chad.jul...@sdstate.edu
You can also run both clusters from the same database (with or without
federating them).
Might not be what you want - you can then e.g. in sinfo see both
clusters, and/or submit to all clusters from everywhere and such.
Tina
On 13/01/2021 18:19, Mando Rodriguez wrote:
Hi chad,
You can ‘dum
Hi chad,
You can ‘dump’ the info in the slurm database to a file and reload the file
(here named cluster.cfg).
Dump the info with:
sacctmgr dump slurm_cluster file=cluster.cfg
You can load the info with:
sacctmgr load file=cluster.cfg
It saves all accounts, associations, and partition permis
All,
Curious if it is possible to copy my sacctmgr information (accounts, users,
etc...) from one cluster to another? For example slurm_cluster1 to
slurm_cluster2.
Thanks,
Chad
Chad Julius
Cyberinfrastructure Engineer Specialist
South Dakota State University
Old Horticulture 205
Brookings, S