Frequently asked questions
Collection of the frequently asked questions, shoutout to Sean Andrew Buchan (Digital Research Analyst at the University of Aberdeen) for answering most of the questions
How do I check my HPC usage?
$ sreport cluster UserUtilizationByAccount users=r04mr23 start=2000-01-01 format=Accounts%20,Cluster%16,Login,Used
--------------------------------------------------------------------------------
Cluster/User/Account Utilization 2000-01-01T00:00:00 - 2024-06-22T23:59:59 (772412400 secs)
Usage reported in CPU Minutes
--------------------------------------------------------------------------------
Account Cluster Login Used
-------------------- ---------------- --------- --------
mmsn maxwell_cluster r04mr23 281760
Change the r04mr23
with your Maxwell username, it's in minutes, so you have to divide it by 60.
I set my #SBATCH --time
correctly but my job is being stopped mid way?
The reason might be because you didn't set the partition in the first place and your job being assigned to spot-compute
which can only be used for running short job no longer than 36 hours.Add the #SBATCH --partition uoa-compute
to run job that is longer than 36 hours
How do I check how much storage have I used? du -sh
seems to take forever
Chek home (/uoa/home/r04mr23
) directory usage
$ beegfs-ctl --getquota --cfgFile=/etc/beegfs/beegfs-client-uoa-home.conf --uid r04mr23
Quota information for storage pool Default (ID: 1):
user/group || size || chunk files
name | id || used | hard || used | hard
--------------|------||------------|------------||---------|---------
r04mr23|193544|| 14.19 GiB| 56.00 GiB|| 75776| 200000
Check sharedscratch directory (/uoa/home/r04mr23/sharedscratch
) usage
$ beegfs-ctl --getquota --cfgFile=/etc/beegfs/beegfs-client-uoa-scratch.conf --uid r04mr23
user/group || size || chunk files
name | id || used | hard || used | hard
--------------|------||------------|------------||---------|---------
r04mr23|193544|| 2.52 TiB| 1024.00 GiB|| 1002345| 1000000
Again change r04mr23 with your Maxwell username, size
used
is your usage, size
hard
is the non-official storage limit, try to keep it below hard
as suggested by the Maxwell team. chunk files
is the file number limit.
How do I check how long did it take for my slurm job to finish?
sacct --format=JobID,JobName,State,Elapsed -j <JOB_ID>