I have two slurm partitions (lhpc and lgpu) with a shared node (n16-90). I have configured one partition with higher priority. I want that if one job uses the shared node through the lgpu partition and there is already one job in the lhpc partition, the latter is suspended and the former allocates the shared node.

However, it doesn’t happen, the job waits until the lower priority ends.

Here I leave some config:

SchedulerTimeSlice=10
SchedulerType=sched/backfill
SelectType=select/cons_tres

PriorityType=priority/multifactor
PriorityWeightPartition=100

PreemptType=preempt/partition_prio
PreemptMode=suspend,gang

PartitionName=lhpc Nodes=n16-[80-83,90-93] Default=YES MaxTime=24:00:00 State=UP PriorityTier=100 PriorityJobFactor=100 OverSubscribe=FORCE
PartitionName=lgpu Nodes=n16-90 Default=YES MaxTime=24:00:00 State=UP PriorityTier=200 PriorityJobFactor=200 OverSubscribe=FORCE

I have tried a lot of configurations. I have also checked that the job in lgpu has higher priority than the lhpc job.

PS: it’s my first publication, forgive me if there is something wrong with it.