The main separate data storage locations in our HPC machines are:
- The /home/$USER volume can be accessed from all machines. It is a permanent storage space, with regular back-up and a personal quota of 4GB. A group quota can also apply. This is the storage unit that should be used for permanent inputs, model files, scripts, and other such files.
- The /scratch/$USER volume can be accessed from all machines. It provides a convenient location for produce large data files (such as long simulation outputs), with a personal quota of 1TB. Files in /scratch/ can be automatically deleted after 30 days, and the unit is not backed-up.
- The /tmp/$USER/ is the local disk space in each node, allowing for fast I/O rate, but can only be accessed from the node itself. Files in /tmp/ are cleared after 7 days and space usage is limited by job. Once job is finished, temporal files are copied automatically to /scratch/$USER/tmp/$JOBID for a better availability of data.
So, as a summary:
| Name | Variable | Availability | Quota | Time limit | Backup |
|---|---|---|---|---|---|
| /home/$USER | $HOME | Global | 4GB* | Unlimited | Yes |
| /scratch/$USER | - | Global | 1TB | 30 days | No |
| /scratch/$USER/tmp/$JOBID | $SHAREDSCRATCH | 7 days | |||
| /tmp/$USER/$JOBID | $LOCALSCRATCH | Local to node | Job file limit | 7 days | No |
* There is a group quota, depending on the number of users. The limits are 4, 8 and 16 GB for groups up to 5, 10 and 20 users, respectively.
The users’ home directories are offered by a NetApp FAS8020 NFS server and the shared scratch by the BeeGFS cluster.
Related articles