Leaving VMware for Proxmox without downtime

After VMware's exit from the Russian market and Broadcom's move to expensive subscriptions, many IT departments face the same question: what should run their virtual machines next. Extending support legally is impossible, running without security updates is risky, and a fleet of dozens of VMs cannot simply be wished away.

Proxmox VE is the most natural replacement. It is a mature open platform built on KVM with the same key capabilities: multi-node clusters, high availability, live migration of machines between servers, snapshots and built-in backups via Proxmox Backup Server. The web interface covers the whole daily administration cycle, and a full API is there for automation.

The move itself is simpler than it looks. First we deploy a Proxmox cluster next to the existing infrastructure — two or three freed-up or new servers are usually enough. Then machines move in waves: test and secondary systems first, production last. ESXi disks (VMDK) are converted with standard tools, and for critical systems we use re-synchronisation: the bulk of the data is copied ahead of time, so the cutover downtime is measured in minutes.

The usual pitfalls are known in advance. Drivers: Windows machines need VirtIO drivers installed before the move, or they won't see their disks. Networking: VLANs and addressing should be carried over exactly to avoid "floating" issues afterwards. Backups: old Veeam jobs have to be replaced — Proxmox ships a very capable mechanism of its own. And most importantly, application licences tied to hardware or machine UUIDs: that list must be made before, not after.

In our experience, migrating a fleet of a few dozen virtual machines takes two to four weeks including the audit, a test wave and monitoring setup — with services running the whole time. If you are planning to leave VMware, write to us via the contact page: we will look at your configuration and draw up a migration plan.