HPC Grid PoliciesShare & Print
Please read the following information before attempting to use the Grid:
1. The WSU Grid uses Slurm to schedule jobs. You must run all processes on compute nodes by requesting a node through Slurm. This may be done with either an interactive job or a job script. If you are caught running processes on Warrior, the Grid's head node, they will be deleted immediately and your account will be temporarily suspended. If you continuously run processes on Warrior then you will be banned from using the Grid.
2. You may run no more than 20 processes on Warrior. Some things like interactive jobs and screen sessions may use more than a single process, so please pay attention to what you leave running on the Grid.
3. All home directories are shared across all clusters by a global file system that spans all Grid nodes and the master node. Your home directory is where you should place your files and work. Your home directory is protected from other users being able to see your files. If you need to share files then request a group directory by sending an email to the High Performance Computing team including the group name, the AccessIDs of the members, and the amount of space needed.
4. If you are a member of an existing group then the head of the group must send a request to the High Performance Computing team to have you added.
5. When running jobs you should keep all temporary files in $TMPDIR. $TMPDIR is erased once your job is finished. Both /tmp and /wsu/tmp are cleaned out on a rotating monthly basis, so if your files are not stored in $TMPDIR then they are in danger of being erased. We do not backup /tmp or /wsu/tmp, so be warned not to store your files there.
6. SSH-2 has been enabled on all nodes and SSH keys are used for you to be able to move between nodes without re-authenticating each time. The first time you log in, keys are automatically generated. We suggest you let the system auto-generate your keys and that you use a blank password for the SSH-key pair. This will allow you to move around the Grid easily.
7. The Grid may be brought down for maintenance at the following times:
- Third week of May
- First week of December
- First week of March
8. The Grid may be brought down at other times for major hardware upgrades as well. The Grid Computing Team will always provide as much notification as possible so that you may prepare your workload to account for the downtime.
9. If you have any questions, please send an email to the High Performance Computing team at firstname.lastname@example.org.