System update in Hydra

Warning We are carrying out a routine update of the operating system in Hydra. The nodes will be progressively updated without impacting running jobs, but queue times during the roll out might be longer.

We perform such routine updates in Hydra regularly. They include new versions of several core system libraries, including the kernel, the network drivers and the shared filesystem. The most important change in this update is the upgrade of the Nvidia drivers for all GPUs in Hydra to version 470.82.01, which will allow the installation of software depending on the recent CUDA 11.4.

The update of the compute nodes will be rolled out gradually, meaning that part of the cluster will not execute new jobs until the node is empty and it can be updated. Users will see a slight decrease in the number of nodes available, but all partitions in hydra will continue to operate without interruption. Depending on the load of the cluster, queue times in the following 7-10 days might be slightly longer.

Our apologies for the inconvenience.