Dear Users,
we have currently applied a workaround to solve the issue affecting the filesystems on Marconi nodes and the cluster is now back to its full production.
The workaround is to keep the cluster in production up to the next maintenance, scheduled for next week (we will send you a dedicated HPC news soon) when we plan to remove the source of the problem. Unfortunately the solution cannot be applied while the cluster is up and running.
The issue of this morning may have impacted some of the running jobs that may have crashed and need to be resubmitted.
Best regards
HPC User Support @ CINECA