Skip to content

Data Storage§

Our clusters have local parallel filesystems consisting of your home and Scratch directories where you can write data. These are "close to" the compute, connected to it with a fast network.

They may also have local storage on the compute node that can be used during your job.

On-cluster storage§

Home§

Every user has a home directory. This is the directory you are in when you first log in.

  • Location: /home/<username>
  • May also be referred to as: ~, $HOME.

Many programs will write hidden config files in here, with names beginning with . (eg .config, .cache). You can see these with ls -al.

Scratch§

Every user also has a Scratch directory. It is intended that this is a larger space to keep your working files as you do your research, but should not be relied on for secure long-term permanent storage.

Important data should be regularly backed up to another location.

  • Location: /scratch/scratch/<username>
  • Also at: /home/<username>/Scratch (a shortcut or symbolic link to the first location).

Temporary storage for jobs (TMPDIR)§

If the cluster you are on is not described in its about page as being diskless, the compute nodes will have local disks that can be written to during your job.

This is because a local disk is faster to write to than a parallel filesystem.

This location is created on the compute node when your job begins, and deleted again when it ends. You will need to copy the data back out of it before the end of your job into your Scratch. If your job fails or runs out of time, you will not be able to recover this data.

  • Location: $TMPDIR
  • Will only exist during your job and be deleted after.

ARC Cluster File System (ACFS)§

The ARC Cluster File System (ACFS) is ARC's centralised storage system that will be available from multiple ARC systems.

It is the backed-up location for data which you wish to keep.

The ACFS is available read-write on the login nodes but read-only on the compute nodes. This means that your jobs can read from it, but not write to it, and it is intended that you copy data onto it after deciding which outputs from your jobs are important to keep.

Initially rolled out on Kathleen, you will later be able to access it from Myriad too and see the same files from both clusters.

  • Location: /acfs/users/<username>
  • Also at: /home/<username>/ACFS (a shortcut or symbolic link to the first location).
  • Backed up daily.