Message-ID: 201604220930
Time:  since 21.04.2016, 16:00 Uhr
Affected: HPC cluster, users on gwdu05
Impact: partial access failures on homedirectories of HPC users

The troubleshooting of the filesystem UNI05 is progressing, but will most probably take until Monday. About 190 of the 585 HPC users and functional accounts are affected by this issue.

Due to a software error, increasing the capacity of the filesystem failed. By analyzing the issue so far, partial inconsistencies within the filesystem were detected. Yesterday at around 4:00 pm the system began to disable any access to protect the data.

For fully resolving the failure, a full filesystem check is advised by the manufacturer which are expected to be finished on the course of the weekend. After that, final checks will be performed and the storage increased. If everything runs as expected, the filesystem will be back at full service on Monday.

We are very sorry for this inconvenience and apologize for being vague on providing a specific time for the repair – due to the very complex directory structure of the filesystem, we can only estimate the required time for the checks.