Skip to main content
Have a request for an upcoming news/science story? Submit a Request

MATH data center cooling outage

Link to update at August 3, 2022 9:28pm EDT UPDATE:

As of 9:28pm EDT, the Anvil, Bell, Geddes, Gilbreth, and Halstead clusters have been brought up and returned to normal service. Job queues have been enabled and job scheduling has resumed.

Please note that the vast majority of compute nodes were powered off to decrease thermal load, and jobs that were running when the outage occurred most likely have been interrupted. Please check your outputs and resubmit if necessary.

We apologize for the disruption of service. Please report any issues to rcac-help@purdue.edu (for Purdue clusters) and help@xsede.org (for the XSEDE Anvil system).

Link to update at August 3, 2022 8:34pm EDT UPDATE:

The Anvil cluster has been returned to service.

Link to update at August 3, 2022 8:06pm EDT UPDATE:

The Bell cluster has been returned to service.

Link to update at August 3, 2022 7:29pm EDT UPDATE:

The Halstead cluster has been returned to service.

Link to update at August 3, 2022 6:14pm EDT UPDATE:

Cooling is restored in the Math data center, but most affected systems remain powered off. Staff began procedures to bring affected administrative and research systems online. Gilbreth cluster has been returned into operational status.

We will provide further update by 9pm.

Link to update at August 3, 2022 2:06pm EDT UPDATE:

Temperatures in the data center were rising rapidly and all nodes are being powered down.

Anvil, Bell, Geddes, Gilbreth, and Halstead are down until the cooling issues can be resolved.

Link to original posting ORIGINAL:

The Math building data center began experience issues with its cooling system around 1:40pm EDT. To minimize thermal load on the cooling infrastructure, job scheduling has been paused and all idle compute nodes on Anvil, Bell, Geddes, Gilbreth, and Halstead clusters are being powered down while data center engineers and physical facilities crew are diagnosing the issue and are working to identify a fix.

We will provide an update by 6pm.

Originally posted: