The Extreme High Performance Computing Resource (“Cluster”) is managed by administrative personnel (“Administrators”) at the Academic Computing & Communication Center (ACCC).
Administrators reserve the right to take the Cluster offline monthly for routine system maintenance. These maintenance periods are defined as the first Tuesday of each month from 1:00 PM and can last up to 48 hours without additional notification. During this time, user logins will be disabled and the job scheduler will be suspended.
System maintenance tasks include (but are not limited to): operating system upgrades, software installation, hardware replacements, recabling, network reconfiguration, filesystem maintenance, security policy implementation, and changes mandated by facilities & infrastructure.
When advantageous, a maintenance period may be rescheduled for a different time based on various criteria; this may be driven by external factors (e.g., vendor-assisted maintenance) or general convenience (e.g., academic schedule). Users will get advance notice of any changes in the standard maintenance schedule.
By default, users should assume the maintenance period will be in effect on a monthly basis. Administrators will notify the users via the Cluster mailing list regardless of whether maintenance will be taking place for the month.