Abusive User Policy
For NREL's high-performance computing (HPC) systems, learn what constitutes an abusive user and how incidents of abuse are handled.
An abusive user is a user who is making the system difficult to use for other users by exhausting a shared resource (such as bandwidth to /home) or working around established controls (like scheduler policy or /scratch file aging) to their benefit. Inadvertent abuse (most common situation) will be handled as a first incident described below.
- The user’s ability to run new jobs or store new data will be suspended temporarily. For situations severely impacting a common resource, a user’s jobs may be terminated if required to return the system to normal operation.
- A member of the operations team will get in touch with the user, either via email or phone, to explain the situation, educate, and help fix the situation.
- Once user has acknowledged and participated in a remedy, ability to run new jobs or store new data will be restored.
- Suspend running new jobs or storing new data. Terminate jobs if necessary.
- The system and operations group manager will talk to the user and the user’s line manager (NREL staff) or NREL sponsor (non-staff).
- Once user and his or her manager / sponsor has acknowledged and participated in a remedy, ability to run new jobs or store new data will be restored.
- Suspend new jobs or storing new data for 1 month.
- Suspend new jobs or storing new data for 1 year.