Running Proxmox in production requires more than a single node. This guide walks through building a 3-node HA cluster with shared Ceph storage and proper corosync quorum configuration.
We cover network design (separate cluster, storage and migration networks), fencing configuration, and the sysctl/kernel tunings that prevent split-brain scenarios.
Includes a pre-flight checklist, rolling upgrade procedure, and lessons learned from a live migration of 120 VMs without downtime.