All HPC Resources will undergo maintenance November 4-5, 2025. Derecho, Casper, JupyterHub, GLADE file systems, data access nodes, and Globus transfers will all be unavailable for various durations over these two days.
The primary maintenance activity will be upgrading the operating system version deployed on Casper and Derecho. This critical activity is required in order to satisfy security and support requirements, as well as enable new versions of user-facing software. We intend to deploy a new ‘ncarenv’ user software version during this outage, while also ensuring current versions continue to function properly.
As usual, scheduler reservations will be implemented across all systems to ensure that user jobs complete before downtime begins on November 4.
However, a notable difference from usual outages is that queued jobs will be purged and must be resubmitted after the maintenance period. This disruption is necessary to implement changes to the queue structure for selecting GPGPU resources, which will simplify the process of choosing resource types. For details on the upcoming changes please refer to the October NCAR HPC Users Group
discussion and
preliminary documentation.
Progress updates will be communicated through the Notifier system throughout the outage.
