The move to subscription licensing has pushed many small and mid-sized environments to look for an exit from vSphere, and Proxmox VE 9.2 (May 2026) is the most common landing point: KVM and QEMU underneath, ZFS or Ceph for storage, and an integrated ESXi import feature that pulls VMs directly from a running host. This guide walks through the import from ESXi 7.0, 8.0 U3 or 9.x into Proxmox VE 9, with the guest-side preparation that decides whether the VM boots first time.
Applies to ESXi 7.0, 8.0 U3 and 9.x to Proxmox VE 9.2
Table of Contents
Short answer: In Proxmox add the ESXi host as storage under Datacenter » Storage » Add » ESXi with root credentials, browse the VM list, and click Import on each VM after choosing the target storage, network bridge and (for Windows) the “Prepare for VirtIO-SCSI” option; the VM is copied disk by disk with its configuration mapped to a Proxmox VM. Before importing Windows guests, install the VirtIO drivers and QEMU guest agent inside the guest while it still runs on ESXi, remove VMware Tools, and note the IP configuration. Power the source off before the final import so the copy is consistent.
Prepare the Proxmox side
The ESXi import plugin ships with Proxmox VE 9. Create the target storage (ZFS pool or LVM-thin), a bridge that matches the VLAN layout of the VMware port groups, and confirm the Proxmox host can reach the ESXi management address on TCP 443. Add the ESXi source:
Datacenter » Storage » Add » ESXi » server address, root username and password, and tick “Skip certificate verification” if the host uses a self-signed certificate. The new storage entry then lists every VM registered on that host. Proxmox reads disks through the ESXi HTTPS datastore interface; expect 50–150 MB/s per stream.
Prepare each guest before migration
Windows guests need VirtIO drivers present before the first boot on KVM or they blue-screen with INACCESSIBLE_BOOT_DEVICE. While the VM still runs on ESXi:
- Mount the latest virtio-win ISO and run the
virtio-win-guest-tools.exeinstaller, which installs the SCSI, network, balloon and guest agent components. - Uninstall VMware Tools.
- Record the IP configuration with
ipconfig /all, because the VMXNET3 adapter will disappear and a new VirtIO adapter will need the addresses. - Note whether the VM boots with EFI (
bcdedit | findstr pathshowing winload.efi) or BIOS.
Linux guests generally have VirtIO modules in the initramfs already; swap open-vm-tools for qemu-guest-agent after the move.
Import from the wizard
Power off the source VM in vCenter. In Proxmox, select the ESXi storage, pick the VM and click Import. The dialog offers a VM ID, target storage for each disk, the network bridge and model, and an Advanced tab where you set the SCSI controller to VirtIO SCSI single, the disk bus to SCSI, and the CPU type to x86-64-v2-AES or host. For Windows, tick “Prepare for VirtIO-SCSI”, which attaches the boot disk as SATA for the first boot so Windows can activate the VirtIO driver, after which you switch it to SCSI.
Set the BIOS to OVMF (UEFI) if the source was EFI and add an EFI disk; for BIOS sources keep SeaBIOS. Start the import; progress shows per disk in the task log.
Import manually with qm
For hosts where the plugin is unavailable, or for very large disks, copy the VMDK to the Proxmox host and import it directly:
qm create 201 --name app01 --memory 8192 --cores 4 --net0 virtio,bridge=vmbr0 --scsihw virtio-scsi-single --bios ovmf --efidisk0 local-zfs:1
qm importdisk 201 /mnt/vmware/app01/app01.vmdk local-zfs --format raw
qm set 201 --scsi0 local-zfs:vm-201-disk-1
qm set 201 --boot order=scsi0
qm set 201 --agent enabled=1Copy the .vmdk and -flat.vmdk together, or export an OVF from vCenter and use qm importovf to create the VM from the manifest.
First boot and fix-up
Start the VM and open the console. On Windows, confirm the VirtIO SCSI controller and network adapter appear in Device Manager, re-enter the IP settings, and reactivate Windows if the licence check trips on the hardware change. Then shut down, move the boot disk from SATA to SCSI in the Hardware tab, update the boot order and start again. On Linux, verify /etc/network/interfaces or NetworkManager profiles reference the new interface name, and check systemctl status qemu-guest-agent.
Then verify from the host:
qm status 201
qm agent 201 ping
qm config 201A successful agent ping confirms the guest tools are working; the config should show the intended controller, bridge and firmware. Test the application, then delete the ESXi storage entry from Proxmox once the migration wave is complete so the credentials are not left behind.
Common pitfall
Importing a Windows VM without the VirtIO drivers installed beforehand is the single most common failure; the “Prepare for VirtIO-SCSI” option only helps if the driver files are already in the guest. The second is a running source: importing a powered-on VM gives a crash-consistent image at best, so power it off before the import starts. For staying on vSphere instead, the current upgrade path is in Upgrade ESXi 8.0 U3 to ESXi 9.1.
Migrate VMware to Proxmox at a glance
Official documentation: Broadcom TechDocs (VMware), Proxmox VE documentation, Linux man pages.
Related guides: Upgrade ESXi 8.0 U3 to ESXi 9.1 with esxcli and Lifecycle Manager · Proxmox VE 9 vs ESXi 9 for hosting providers: licensing, storage, backup and migration realities · Upgrade vCenter Server Appliance 8.0 to 9.x with the Migration Assistant.
Frequently asked questions
Does the Proxmox ESXi import work with vCenter or only with individual hosts?
It connects to an ESXi host directly, not to vCenter, so add each host as its own storage entry. Live imports from vCenter are not supported; VMs must be reachable through the host that runs them.
How long does a VMware to Proxmox migration take per VM?
Budget roughly 15–30 minutes per 100 GB of allocated disk through the import plugin, plus 15 minutes of guest preparation and fix-up. Thin-provisioned disks copy only their used blocks, which shortens the transfer.
Can I undo the migration and move back to VMware?
The source VM remains untouched on ESXi as long as you do not delete it, so reverting is simply powering the original back on and discarding the Proxmox copy. Data written on the Proxmox side after cutover has to be copied back manually.
Maintenance record
This guide changes servers, data or security settings, so we re-check it against current versions on a fixed schedule. Take a backup or snapshot before you start.
- Maintained by
- srvScripts editorial team
- Supported versions
- ESXi 7.0, 8.0 U3 and 9.x to Proxmox VE 9.2
- Last full review
- Next review