A four-node, 41-machine Proxmox cluster taken over and documented

Context

A cluster grown organically over years: four physical nodes, forty-one virtual machines, several administrators along the way and no current documentation.

Problem

The audit surfaced duplicate MAC addresses across machines, missing backup jobs for production systems, a high availability configuration still pinned to a decommissioned node, and storage allocated with no record of it.

What I did

1. 01Full inventory: node by node, machine by machine, with owner and purpose identified
2. 02Fixed the duplicate identifiers that were causing intermittent network conflicts
3. 03Brought every production system under backup, on a dedicated backup server
4. 04Renamed a node live, without stopping the machines hosted on it
5. 05Standardised administrative access with dedicated accounts and controlled privileges