Today (from morning until early afternoon), the SCRATCH file system was not available. Within that timeframe, jobs may have crashed. SCRATCH file system was made available again.
Short lasting outages of SCRATCH have been observed today on CoolMUC-2 and hosted cluster systems as well.
SCRATCH unavailable on CoolMUC-3
To avoid file system crashes, we have decided to unmount SCRATCH from all systems associated with CoolMUC-3 for now. We expect to revert this measure again at the next scheduled maintenance .
End of service for NAS systems
NAS paths will be taken offline at the beginning of January, 2020. Please contact the Service Desk for outstanding data migration issues.
You will find all HOME data in the new DSS HOME area, data migration will be performed by LRZ (unless you are specifically notified that you need to perform HOME data migration yourself). For emergency recoveries, the legacy NAS area (pointed to by the HOME_LEGACY variable) will remain available in read-only mode until the end of the year.
Following action is required on your part: Make necessary adjustments to (job) scripts and configuration data to account for the changes in path names. LRZ strongly recommends using relative path names because this minimizes the required work.
Migrate data from the legacy NAS area (pointed to by the PROJECT_LEGACY variable) to the new DSS area. LRZ strongly advises to get rid of unneeded data sets, and/or archive data sets to tape.
Step-by-step procedure for migration:
On any cluster login node, issue the command dssusrinfo all This will list paths to accessible containers, as well as quota information etc.
Edit your shell profile and set the PROJECT and/or WORK variable to a suitable path value based on the above output, typically one of the DSS paths with your account name appended to it.
Use the cp or rsync or tar command to migrate your data from PROJECT_LEGACY to the new storage area.
If your scripts use absolute path names instead of the PROJECT or WORK variable, they need appropriate updates
Confirm the e-Mail invitation to validate your access,
After the Linux Cluster maintenance (see below), store path information in an environment variable on the Cluster (e.g. by setting the PROJECT variable in ~/.bashrc).
One this is done, migrating data from NAS PROJECT to DSS PROJECT can start
Please read the change description on how to handle the significant changes to the Linux cluster configuration performed end of September, 2019
Messages for Cloud and other HPC systems
The OpenNebula Compute Cloud was decommissioned on