Operating Status

Percent Utilization over the past 24 hours (Red Cloud)

Status and Downtime Notices

• Downtime for regular Web server maintenance will occur the Wednesday following the second Tuesday of every month.
• Only one e-mail will be sent per downtime, to all project members.
• Information concerning completion of the work or extension of the downtime will be posted on this page.

OUTAGE REASON(S): To apply November 2026 security patches (Windows and Linux)

WHO WILL BE AFFECTED? Users of CAC managed resources housed in Rhodes Hall

WHAT WILL BE UNAVAILABLE WHILE THEY ARE PATCHED AND REBOOTED?

  • Servers/Clusters: Linuxlogin, Mysql2, RT, seneca
  • The following HPC cluster head node and login nodes will be rebooted to patch against the CVE-2026-98432 security vulnerability: aida, altas, capecrystal, cayuga, pool2, tardis, mbot, praxis1, hopper, Yuegroup
  • Red Cloud instance users should run 'dnf update' or 'apt update; apt upgrade' and reboot their instances ASAP to patch this vulnerability.

WHAT WILL NOT BE AFFECTED THIS MONTH?

  • Archival Storage

STATUS: Downtime status will be posted on this page.

QUESTIONS: https://portal.cac.cornell.edu/help.

NEXT SCHEDULED MAINTENANCE: Wednesday, December 9, 2026

WHY: Install new PDUs to support additional compute nodes

WHO WILL BE AFFECTED? All MBot cluster users (nd357_0001 and ayh24_0001 projects)

DETAILS:

  • At 9 pm on November 2, the entire mbot cluster will be powered off.
  • New PDUs will be installed in the early morning of November 3 to provide more power to accommodate additional compute nodes.
  • After the cluster is powered back on and passes verification tests, user logins will be re-enabled. Users will be notified by e-mail.

NOTE:

  • A "2026_11_02_Install_PDU" reservation starting at 9 pm on November 2 has been created to prevent slurm from scheduling jobs running past that time.

STATUS: Downtime status will be posted on this page.

QUESTIONS: https://portal.cac.cornell.edu/help.

 Seneca cluster configuration updates have been completed.

WHAT CHANGED? 

  • All cluster nodes are now connected by 400 Gbit NDR infiniband.
  • All nodes in the "cac-cpu" partition now has 1 TB of RAM.
  • A few nodes remain offline until some hardware issues can be resolved.

STATUS: Downtime status will be posted on this page.

QUESTIONS: https://portal.cac.cornell.edu/help.

WHY: We need to add infiniband interconnect and relocate some servers to a different rack

WHO WILL BE AFFECTED? Users of the "cac-cpu" and "cac-gpu" partitions

WHAT WILL BE AFFECTED?

  • All nodes in the "cac-cpu" partition now have 1 TB of RAM, infiniband interconnect, and direct access to the BeeGFS parallel file system at /mnt/beegfs. However, fewer nodes are available until the transition is completed.
  • The "h100pcie" and "a100" GPU resources are not available in the "cac-gpu" partition until the transition is completed. Users should request the more capable "h100" GPU resource meanwhile.
  • We are looking to complete the transition in the next few days and will update this page when it is completed.

STATUS: Downtime status will be posted on this page.

QUESTIONS: https://portal.cac.cornell.edu/help.

OUTAGE REASON(S): To apply August 2026 security patches (Windows and Linux)

WHO WILL BE AFFECTED? Users of CAC managed resources housed in Rhodes Hall

WHAT WILL BE UNAVAILABLE WHILE THEY ARE PATCHED AND REBOOTED?

  • Servers: Linuxlogin, Mysql2, RT
  • Clusters: seneca
  • Web: https://www.cac.cornell.edu/

WHAT WILL NOT BE AFFECTED THIS MONTH?

  • Linux Clusters: Aclab, AIDA, Altas, CapeCrystal, Cayuga, CommCloud, Cqmetrics, Hopper, Marvin, Marx1, Mbot, Pool, Tardis3, Yuegroup
  • Red Cloud
  • Archival Storage

STATUS: Downtime status will be posted on this page.

QUESTIONS: https://portal.cac.cornell.edu/help.

NEXT SCHEDULED MAINTENANCE: Wednesday, September 9th, 2026

 Jetstream2 Cornell Region is back online running Openstack 2025.1 "Epoxy" release. Please email help@cac.cornell.edu if you encounter problems. Thank you for your patience.

 

STATUS: Downtime status will be posted on this page.

QUESTIONS: https://portal.cac.cornell.edu/help.

WHY: Upgrade Jetstream2 Cornell Region to run Openstack "Epoxy" release 2025.1

WHO WILL BE AFFECTED? Users of Jetstream2 Cornell Region

WHAT WILL BE AFFECTED:

  • Users should save their work and shelve all running instances in Jetstream2 Cornell region. We have been live migrating instances to minimize disruptions in the last few upgrade cycles. This is no longer possible for this upgrade cycle and all running instances must be shelved.
  • At the start of the maintenance window, all running instances will be shelved.
  • Users will be notified at the end of maintenance window so they can unshelve their instances.

STATUS: Downtime status will be posted on this page.

QUESTIONS: https://portal.cac.cornell.edu/help.

WHY: Upgrade OS and apply software patches on cloud controller for Jetstream2 Cornell Region

WHO WILL BE AFFECTED: Users of Jetstream2 Cornell Region

WHAT WILL BE AFFECTED:

  • During the maintenance window, Horizon and Exosphere web consoles and Openstack API will not be available.
  • External network access for instances on private networks will not be available. Instances should continue to run without disruption.

STATUS: Downtime status will be posted on this page.

QUESTIONS: https://portal.cac.cornell.edu/help.

 Hopper cluster is back in service. If you continue to experience problems, please email help@cac.cornell.edu for assistance.

 

STATUS: Downtime status will be posted on this page.

QUESTIONS: https://portal.cac.cornell.edu/help.

 Hopper cluster is unavailable due to a file system issue. CAC system staff is investigating.

 

STATUS: Downtime status will be posted on this page.

QUESTIONS: https://portal.cac.cornell.edu/help.