Slurm Nodes Automatically Going to DRAIN State Due to TRIX-Drainer

[shivshankar7977]

Hi All,

I am facing an issue with Slurm after installing TrinityX 15.2.

I am able to manually bring the Slurm nodes into the IDLE state, but after a few minutes, the nodes automatically go back into the DRAIN state.
The drain reason shown is:

Reason = TRIX-DRAINER: Multiple reasons for drained node

Both slurmctld on the controller node and slurmd on the compute nodes are running fine.

I have attached a screenshot for your reference.

Could you please help me understand why TrinityX drainer is forcing the nodes into DRAIN state and how we can fix this issue so that the nodes remain in the IDLE state?

Thanks!

[aphmschonewille]

Hi,

The nodes are drained when there is a believed reason. E.g. hardware problems or others. Can you verify with alertx overview?

Also have a look at alertx docu.

-Antoine