Target audience

This documentation’s intended audience is system administrators, network administrators, ops and devops. Skills needed to understand and use this documentation to install a new HPC machine are at least (in no specific order):

• Networking: TCP/IP, vlans, routing

• System: Linux (preferably Redhat or Centos), yum or dnf package managers, systemd

• Filesystems: xfs, GlusterFS

• Automation software: Puppet, stackstorm

• DNS: bind, dnsmasq

• DHCP: ISC dhcp

• Load balancing and web: HAProxy, apache

• Monitoring: prometheus, grafana

• VM instanciation: qemu, KVM, cloud-init

• File formats: Yaml and json

• Versioning: git

• Key-value store: etcd

Having some knowledge on how HPC machines are configured and deployed is a plus.

There are a number of inhouse software that are part of the ocean distribution that you’ll need to learn and get used to:

• ClusterShell

• Confiture and BMGR

• fleet

• pcocc