On Thursday afternoon, April 17, we performed scheduled system maintenance to apply a series of important upgrades. This required temporarily taking parts of the system offline—but it was worth it!

Here’s what’s new:

• Operating System Update: All nodes have been upgraded from AlmaLinux 9.4 to 9.5.
• cgroups v2 Enabled: The system now uses cgroups v2 for improved resource control and future compatibility.
• Slurm Upgrade: We’ve upgraded Slurm from version 22.05 to 24.11, fully configured to use cgroups v2. This brings better resource accounting and will allow us to include CPUs currently reserved for system processes into the regular Slurm partitions in the near future.
• NVIDIA Driver Update: GPU nodes are now running the latest NVIDIA drivers for improved stability and performance.
• CephFS Upgrade: Our storage backend has been upgraded from CephFS 19.2.1 to 19.2.2, including minor bug fixes and stability improvements.

Please let us know if you encounter any issues or have questions. Thanks again for your understanding during the downtime!