Shared Storage Usage Policy
To use NREL’s high-performance computing (HPC) systems, you must abide by the Shared Storage Usage Policy.
NREL HPC allocations include storage space in the /projects filesystems. This is a shared resource, and project leads should periodically monitor usage by running grep <allocation_id> /projects/space_usage, which provides project-total usage of the /projects filesystem in kB units for the associated system (i.e., Peregrine or Eagle). This data is updated daily.
HPC Operations will notify allocation leads if utilization exceeds the allocated amount. For allocations which also include node hours, job queue priority will be reduced until utilization falls within the allocated space. If utilization remains out of compliance after seven (7) days, the HPC job queue will stop accepting or running jobs associated with the project allocation until utilization decreases to the allocated capacity.
There is no fixed usage policy for the scratch filesystems aside from the Appropriate Use and Inappropriate Use policies. As described in the Data Retention Policy, data older than 28 days will be automatically purged from /scratch.