[slurm-users] Resource limits

[Alex Bateman]

Hello all,

As I understand the resource limits is applied as follow:

1. Partition QOS limit
2. Job QOS limit
3. User association
4. Account association(s), ascending the hierarchy
5. Root/Cluster association
6. Partition limit
7. None

Here I set the limit on the cluster association-
sacctmgr modify cluster where name=dev set grptres=cpu=1

I set another limit on the user association-
sacctmgr modify user where name=testuser set grptres=cpu=2

The user limit should be the first point on the hierarchy tree and
therefore this limit should be used when job is submitted under that
account. Instead, I'm seeing the cpu=1 limit being applied to the
testuser. Am I misunderstanding something?

Regards,

A