ESXi to Proxmox VE migration: where to start

You have decided to leave VMware and your hosts run ESXi. The good news: Proxmox VE can connect directly to an ESXi host and import its VMs. So the real question is not how to convert a disk, but where to start so you do not cut production. Here is the order that works.

Start with the audit, not the conversion

The classic mistake is to import a VM right away. A migration is won earlier, in the inventory. Before touching ESXi, list:

• Every VM: vCPU, RAM, disk, OS, how critical it is.
• Dependencies: which VMs talk to which, and what cannot tolerate downtime.
• Networking: VLANs, port groups, firewall rules.
• Storage: datastores, capacity, IOPS profile.

This inventory dictates the migration order and the target cluster sizing. You always start with the lowest-risk VMs.

Use the Proxmox VE import wizard

Proxmox VE ships an import wizard that connects to an ESXi host, reads its VMs and imports them, disks included. It converts VMDK to qcow2 or raw automatically and removes most of the manual conversion work. It is the fastest and most reliable route for the majority of estates. Whether your ESXi hosts are driven by vCenter or standalone does not change the method: the migration runs host by host.

Two guest-side points not to miss

The technical import is the easy part. What makes a boot fail is almost always one of these two points:

• VirtIO drivers: install them for paravirtualised disk and network performance. On Windows guests, do it before cutover while VMware Tools is still present.
• Boot mode: match BIOS or UEFI to what the VM expects, or it will not boot.

Migrate a pilot before anything else

Before migrating anything critical, run a small group of non-sensitive VMs end to end: import, boot, network, application, performance measured against the baseline you recorded during the audit. This pilot proves the procedure and surfaces surprises while they are harmless. Keep the ESXi copy until validation: that is your rollback.

Cut over in waves

Once the pilot passes, migrate in waves grouped by dependency. Each wave moves over in a planned window while the rest of production keeps running. Critical services go last, once the procedure is proven. That is how you migrate without a global interruption window.

Where support changes things

The mechanics are documented. What stays hard is sequencing dependencies, sizing the target cluster correctly, and getting high availability and storage right on day one. That is what our VMware to Proxmox migration service takes on, from the audit to wave-by-wave validation. To set the budget up front, see how much a VMware to Proxmox VE migration costs.