Skip to content

CHPC DOWNTIME:  Downtime of general environemnt compute resources - Dec 6, 2022 starting at 7:30am

Date Published: November 11th, 2022

Update 12/06/2022, 9:32pm: 
Due to unexpected complications with the move to the new Infiniband Gateway switch, the notchpeak, kingspeak and ash clusters (both compute and interactive nodes) as well as the frisco nodes and the cryosparc nodes will remain offline overnight.  
We have been in contact with the vendor about the issues and expect to have feedback in the morning to allow us to determine the best way to proceed. Once we have more information and can determine the best way to proceed, we will provide an estimate on when the resources will be back in service.

There will be a general environment downtime on Tuesday December 6, 2022 starting at 7:30AM.    

The downtime is to replace the current infiniband (IB) gateway switch. This switch supports the three IB based clusters of notchpeak, kingspeak and ash. Therefore, the scope is limited to the interactive and compute nodes of these three clusters, along with the frisco nodes and the cryosparc stand-alone systems that utilize this gateway.. 

Reservations are in place on notchpeak, kingspeak and ash to drain these clusters of running jobs by 7:30AM on December 6. Any pending jobs will remain in the queue until after the downtime. 

The work will be done in stages. The first stage will be the move of the notchpeak cluster to the new gateway. Once that is complete and tested, we will release notchpeak and then proceed with making the same changes for kingspeak and ash. 

We anticipate that notchpeak will be returned to service midafternoon, with the other two clusters being back in service late afternoon. 

Last Updated: 1/11/23