Information notices
The NIRD Data Peak filesystem is currently available under both /nird/projects and /nird/datapeak. This will be permanently changed to /nird/datapeakonly on all HPC clusters and on NIRD.
The legacy path (/nird/projects) was maintained for convenience during a transitional period and will be removed on 1 December 2025 from both NIRD login nodes and the HPC machines.
Please ensure that any scripts or workflows you are using are updated to use /nird/datapeak before this date.
This change...
A functionality for automatically copying files from NIRD to Olivia before a job starts, and copying files back again afterwards, has been implemented on Olivia now. See Staging In / Out Files from / to NIRD using Slurm for description and an example of how to use it.
Current incidents
This maintenance is now starting as planned.
We have some noticed issues with the filesystem (dusage not working, users getting quota exceeded errors, etc.) from user reports. We are currently troubleshooting this, and will give an update as soon as possible.
Apologies for the inconvenience this has caused!
/NRIS team
This maintenance is now starting as planned.
A couple of the storage targets for the /cluster/work filesystem on Olivia have become full. This can lead to write errors.
We have started to clean up. This means that some of the oldest files on /cluser/work will have to be deleted.
After this, we will start the automatic file deletion on /cluster/work, and implement filesystem policy changes to try and avoid this in the future.
So from now on, old files on /cluster/work will be deleted, just like on Fram, Saga and Betzy.