Working File System

Due to multiple issues on our filesystem that hosts project directories /rcf-proj2 and /rcf-proj3, we have currently made the project filesystem read-only to prevent further file access issues. The filesystem will remain in this state until we can deploy a new, previously planned filesystem and migrate all project data to it. The new project filesystem is expected to be up and running by mid-July 2020.

All users should now use the /scratch storage area as their working space for jobs.

/scratch is a high-performing, parallel filesystem that is hosted on dedicated storage machines. This is a shared resource for use by all HPC researchers. Data stored in /scratch is retained and not deleted between jobs, but it is not backed up. Data stored in /scratch should be periodically copied to a permanent project filesystem to decrease the risk of data loss.

A directory is automatically created for each HPC user under /scratch so that data can be stored there temporarily. Your scratch directory is accessible only to you, and each user account is limited to a 10TB quota.

For more information on /scratch and instructions for data transfer and software installation, please see our Project Filesystem Status GitHub page.

Managing Files in /scratch

Although files on /scratch are not deleted between jobs, they are never backed up and it is very important for users to periodically copy results to a permanent project filesystem. While HPC will try to protect user files on these systems, any software or hardware failures can result in data loss.

Please remember that /scratch is a resource shared among all users, so it will be monitored for overuse. Please only copy necessary to files to /scratch.