Reputation: 1097
Our cluster used SLURM to manage the workload. I would like to extract a report that reports the CPU, RAM and disk usage per account, for all jobs completed after a given date. Is it possible?
I have been looking at sreport and sacct, but can't seem to be able to customise it as I wish, best I can get is
sreport job SizesByAccount start=2015-01-01 -t percent Format=account Grouping=5,9
for the CPU usage. Any suggestions?
Thanks.
Upvotes: 3
Views: 3603
Reputation: 2518
If that can help someone in the future, I use something close to
$ sacct -S2020-01-01 -E2020-10-07 -X -oJobID,Elapsed,ReqMem,ReqCPUS,User,Node --parsable
69836|00:00:22|3200Mc|1|user1|slurm-node20|
69837|00:03:06|1200Mc|1|user2|slurm-node15|
69838|00:03:08|3200Mc|1|user1|slurm-node12|
69839|00:00:11|1200Mc|1|user1|slurm-node12|
69840|00:03:01|3200Mc|1|user1|slurm-node10|
[...]
And then parse the results with Python.
Upvotes: 3