Linux-Cluster: limited Operation

Update (Oct 16, 11:00)

CoolMUC-2 has been returned to limited operation. This means that on the compute nodes, only HOME and SCRATCH are available. Users with data on DSS will need to manually copy data to SCRATCH on the login nodes. Some nodes remain reserved for further investigation of the observed file system / node crashes, which is expected to take some time still.

Investigation on CoolMUC-3 continues. The KCS segment has been returned to limited user operation, with only the KCS-DSS and HOME file systems available.

Dear users of the Linux Cluster,

We regret the need to inform you that operation of the Cluster systems is currently disrupted. The symptoms for this are user jobs that fail or show very bad performance.
For now, it will likely be necessary to stop batch processing on most cluster segments.

Investigation of the root cause for the disruption has started. We will keep you updated about the progress via this document.


Verfasser: Bader
veröffentlicht: 2019-10-04