Server move and brief outage

The Bowman lab at U Penn recently underwent a massive migration of our physical hardware.

The move was motivated by Penn’s investment in a new, shared compute cluster called Betty. This cluster boasts 31 nodes that each have 8x B200 NVIDIA GPUs. Along with this hardware investment, the university decided to consolidate all its compute systems at a data center that could better support the new shared cluster.

In preparation for the move, we shifted infrastructure that is normally hosted through Penn elsewhere. Shared resources like the core server were migrated to AWS and new projects were started on the cloud or servers from other labs.

The move largely went without a hitch. We got all our work servers back up and running, as well as the core server and other shared services hosted at Penn.

Unfortunately, our data storage system went down this past week. It had been running smoothly, so it’s unclear if the move or some other factor was causative. Regardless, we apologize for a temporary lack of CPU work units. All functionality is now restored and we’re keeping a close eye on it.

Moving forward, we’re excited to see how we can leverage Penn’s new shared resources to accelerate science on Folding@home.

Thanks for your contributions, and for your patience if you were impacted by our outages over the past week.