3: HPC Compute Cluster

Administering a compute cluster in a production environment is a niche area of system administration. Challenges include cluster management, customer usage, and specialized software. The topics will be based on the attendees’ interest but may include OS deployment, software deployment, disk-less compute nodes, management tools, schedulers and resource managers, and customer issues. Attendees should be admins of a compute cluster or interested in adminning this type of cluster. They should come prepared to discuss openly their admin experiences, problems and successes with this class of machine, and the pros and cons of their existing cluster management tools.