Incident description

Created by Martin Siegert on

Title

Scheduler problem - Problème d'ordonnanceur

Summary

A switch failed and because of that several management nodes lost connectivity to the cluster. This includes the scheduler. Until the switch can be replaced the scheduler cannot be contacted, i.e., no jobs can be submitted, and all scheduler related commands fail, e.g., sbatch, squeue, salloc, sinfo, etc.

Updated by Ali Kerrache on