Before jumping into the more technical aspects of conducting research on Raj there are a few important things to note.

Running Jobs

The cluster is a shared resource. As such, users cannot run jobs like they would on a personal desktop or lab owned cluster. You cannot log onto the system and start running jobs. All jobs must be run using the cluster workload manager Slurm. For more on adapting your research to run using Slurm see the section on Running Jobs on Raj with Slurm.

Protecting Your Data

It is very important for all Raj users and group leaders to maintain a copy of all critical data, including code, safely on private storage. In the event of catastrophic data loss due to unforeseen hardware issues, software failure, or a user accidentally deleting their own data, that data may not be recoverable. All users are responsible for maintaining the integrity of their own data. As a consequence, ITS highly recommends an off-cluster backup of all critical data and code. 

In addition to the recommended off-cluster backup, we provide an NFS share backup location for critical data and code. This backup is located at /backup on the cluster. Each user has a 50 GB directory at this location that can be used as an additional backup. This is not a performance-based storage location. Please do not run any jobs or major computations from code or data stored there. This is not it's intended purpose.

---