Scheduled maintenance of Leonardo – UPDATE – Caveat on SCRATCH FS

Dear Users,
this is to give you an update about the Leonardo maintenance. It is almost completed and we expect to bring the cluster back to production within the afternoon. However, there are important information to be taken into account.
As already announced, during the maintenance a technical issue to the cooling system occurred that affected a few controller components of the scratch filesystem, preventing the access to a small number of disks and the mount of the filesystem on the compute nodes. So, at the moment, when opening production, the entire SCRATCH area will not be mounted on the compute nodes. It will be accessible only from the logins. As a consequence all the jobs making use of the SCRATCH area will fail until the issue is solved.We suggest, thus, to cancel all of your jobs using SCRATCH area and resubmit them making use of the WORK area that will be fully operative. You may need to copy data from SCRATCH to WORK before resubmitting the job.
While copying, it is possible that some files may not be accessible because present in the physical disks of the storage with the controllers affected by the issue described above. We are working to restore them.
We will promptly inform you as soon as we have important updates on this.
We apologies for the inconvenience.
Best regards
HPC User Support @ CINECA