[slurm-users] How to limit cpu hours with multiple clusters

Fri Jan 25 16:32:05 UTC 2019

Hi,

We have a setup with two clusters connected with Slurm, both using the multifactor plugin. I made a QOS with the NoDecay flag to limit cpu time for paying users, but when I was testing it, even if it reached the limit in one cluster it allowed to compute in the other one.

Is there any way to syncronize the Raw Usage of the QOS? Or at least another way to limit cpu hours for both clusters using Slurm?

We are currently using Slurm version 18.08.1

Thanks!
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.schedmd.com/pipermail/slurm-users/attachments/20190125/364aa220/attachment.html>