All nodes drain state on bamboo

Raphael.Rubino · April 24, 2025, 11:34am

Primary informations

Username: rubinor
Cluster: bamboo

Description

All nodes are in drain state on bamboo.

Steps to Reproduce

sinfo on bamboo

Yann.Sagon · April 24, 2025, 12:20pm

Dear @Raphael.Rubino

this is fixed. [2025] Current issues on HPC Cluster - #10 by Yann.Sagon

Raphael.Rubino · April 25, 2025, 9:03am

Thanks @Yann.Sagon

Still on bamboo, gpu003 is currently unusable.

scontrol shows
Reason=gres/gpu GRES core specification 12-15 doesn't match socket boundaries. (Socket 0 is cores 0-16) [slurm@2025-04-25T10:52:59]

Yann.Sagon · April 25, 2025, 9:08am

Yes, fixed. This is an annoying bug that will be fixed in the next Slurm release.

Best

Raphael.Rubino · April 25, 2025, 2:49pm

Thanks @Yann.Sagon
It seems that bamboo login node is unreachable outside unige network at the moment.
Best regards

Yann.Sagon · April 25, 2025, 3:12pm

This is fixed. Time for the week-end!