[PVE] Vmware VM recovery leaves it unbootable due to virtual HW being replaced

[dejan.ilic]

During tests of Vmware ESXi VM recovery on Proxmox 8.4 I noticed that the Veeam recovered Vmware VMs are not bootable out of the box.
It was fairly quickly narrowed down to Veeam replacing Vmware virtual HW with alternatives.
As we haven't installed VirtIO drivers in advance, the Windows VM (UEFI boot) simply would boot as it was missing drivers.

By exchanging the assigned virtual HW in PVE with the same as Vmware VM had in the first place, the VM booted without issues and any additional software.
* SCSI with Vmware Paravirtual
* network with VMXNET3
* added back vTPM
* display set to Default instead of "VMware compatible" (did work thou without change, still why guess what "default" setting implies)

If there are any plans to support SureBackup on PVE in the future, this Virtual HW replacement behavior will make it harder also.

As all these virtual HW emulations exists in Proxmox, I see no reason for replacing them with anything other.
If it is due to compatibility reason with older Proxmox, then there should be a selection button to switch virtual HW to proxmox corresponding.

Also, removing the vTPM without a hint/warning is also a bad behaviour. If it cannot be restored, Veeam should warn that the vTPM content might not be correct but functional.

The other virtual hardware being removed is less problematic as it is seldom used, but should also stay if possible (serial port, PS2 controller etc..)

[rovshan.pashayev]

Hello Dejan,

Thank you for your report!

I wanted to let you know that our QA team has recently identified the SCSI issue, and we are actively working on a fix.
Regarding the other concerns you mentioned, we are looking into them as well.

[FrancWest]

Hi,

I noticed this to. Veeam creates the VM with the virtIO SCSI controller instead of the SATA controller. I'm not sure if VMWare paravirtual will work, since we uninstall vmware tools before the migration. Anyhow, it's a bit of work to get the VM to boot after migration. You have to detach the virtual disk, re-attach it as SATA. Add a smaal 1GB virtual disk as VirtIO SCSI, so that after installing the virtio tools the controller drivers are installed and Windows is able to boot from it. Boot the VM, install virtio tools, shutdow the vm. Detach the SATA disks and re-attach them using the virtio SCSI controller. I also changed the CPU to x86-64-v2-AES, while Veeam sets it to a processor with less capabilities.

I opened a support call for this issue, case #07690336 and the support engineer created a feature request for this, but it seems to be a known issue.