Upcoming NERC system maintenance and upgrade - Tuesday, August 12, 2025 8:00 AM - 5:00 PM.

GENERAL MAINTENANCE

- We will be upgrading our Red Hat OpenShift container platform (OCP) version on NERC to 4.18.

- As part of the upcoming maintenance, we’ll be rolling out several key enhancements to Red Hat OpenShift AI (RHOAI). These include a redesigned dashboard layout, support for both single and multi-model serving, and distributed training capabilities using CodeFlare and KubeRay. The platform will also support high-performance inference runtimes beyond the currently offered OpenVINO, such as vLLM. Additional improvements include a GPU pruner for more efficient GPU utilization, enhanced resource scheduling with Kueue, and an upgraded version of Kubeflow Pipelines. This update will further strengthen MLOps capabilities with an integrated model registry and catalog to manage the full model lifecycle.

- We will be upgrading our ColdFront deployment.

NOTICES

- The OpenShift and ColdFront services will be unavailable during this maintenance.

- Please keep a backup of your critical data or application running on your project on NERC’s OpenShift cluster.

- Project allocation requests and change requests for both OpenStack and OpenShift Resources will not be accepted or approved during this time.

- The estimated time to complete this maintenance is 9 hours. It can take more or less time, so we urge you to keep an eye on https://nerc.instatus.com/ for real-time updates on progress.

- Please do subscribe to https://nerc.instatus.com/subscribe/email in order to get the progress updates during this time.

Our priority is to help make science happen, so If you or your research team have any questions or need to escalate an issue, please contact us via email (help@nerc.mghpcc.org) or, by submitting a new ticket at the NERC's Support Ticketing System.