For up to date information regarding the BlueBEAR cluster please subscribe to the bear-announce mailing list. See the IT Services status site to see the status of all IT Services services, including BlueBEAR Service.
Current Service Notice: Services are fully operational.
On December 12th
2017 we shall be taking a short maintenance window that will affect the BlueBEAR HPC service. This window is due to the annual, essential testing of the fire systems in the University’s data centre.
During the testing period, the air cooling systems have to be turned off in the data centre, and as one of the biggest generators of heat in the data centre, we are required to shut down the HPC compute systems to reduce the impact on other critical business systems in the shared space.
Users should continue to submit jobs as usual as we have placed a maintenance window in the HPC scheduler which will prevent it from starting or running jobs which will coincide with the period when the compute systems are powered off. We expect to continue to provide access to the login nodes etc as usual and it will only be HPC compute capacity that will not be available. We are aiming to keep the maintenance window as short as possible, but will also be taking the opportunity to re-instance the storage cache system we have which accelerates access to the Research Data Store (RDS).
We will confirm more details of the timing and duration of the maintenance window closer to the date. Users who submit jobs that are 10 days long or to the research-group specific 30 day queues should be aware that as we get closed to the maintenance window, their jobs will be held in the queueing status if they would overrun the maintenance window.
In other HPC related news, during the next year we will be moving to a new data centre which is currently being constructed, you can find our more information on the BEAR blog at:
| BlueBEAR HPC service All-day
This page will be updated with more information as it becomes available.
Last modified: 13 November 2017