How to set nodes to drain so they are not reused

I’m using AWS parallel cluster to launch nodes. Ideally a new job would use a new node, we are seeing SLURM schedule on top of existing nodes that no longer have a job but haven’t been terminated yet.

AWS recommends the following be set - scontrol update NodeName=$SLURM_NODELIST State=drain Reason=“Job completed”

Question is - Is it possible to set that somewhere in OoD or would I have to use the SLURM epilog for example (what AWS recommended actually) ?

Yea there’s no way for an unprivileged user to drain the nodes (or at least I hope not!). Sounds like an epilogue is your best bet.

1 Like