The workaround I have found at this point is to rebuild grub after migrating to XCP-NG. Adds a few minutes to each VM but I don't have that many.

Boot to live iso in UEFI mode (I used Debian12 given it was a Debian12 vm)

sudo su -
mount /dev/mapper/?????  /mnt 
mount /dev/xvda2 /mnt/boot
mount /dev/xvda1 /mnt/boot/efi
for i in /dev /dev/pts /proc /sys /run; do sudo mount -B $i /mnt$i; done
mount --bind /sys/firmware/efi/efivars /mnt/sys/firmware/efi/efivars
chroot /mnt
grub-install /dev/xvda
update-grub
exit
shutdown now

I'm using LVM so adjust the first mount to match your root partition.
I shutdown so that I can remove iso and boot when I'm ready.

Hope this saves somebody a bunch of frustration in the future.

@jebrown Thanks. I wondered about that. I just hadn't gotten around to trying that. Thanks for confirming!

OK, I built a fresh UEFI Debian 12 from ISO on esxi. I then cloned it 3 times. I removed open-vm-tools from #2 and I installed xcp tools on #3 and imported all 3. They all have the same error.

I poked around the internet more while waiting and see other instances with proxmox in particular for Debian and Ubuntu. Something isn't jiving between DEB based UEFI and either the conversion process or TianoCore or something.

Which is fair but I have moved 3 bios Debian 12 without issue. I have tried 2 different Debian 12 with UEFI and both have failed with the same error.

I use open-vm-tools on them. I'll try removing it but cannot start the vm for a while because something locks the vmdk on the esxi side after doing the xo migration and I have to either reboot the host or wait a while.

I'll try building a new uefi debian 12 vm and see if it happens from a clean build.

Unfortunately I don't believe that to be the case. I removed the pci device, rebooted the VM to ensure everything was fully baked in, shut it down and exported to OVA. I imported the OVA. It recognized it as BIOS which wedged at booting from hard disk with 100% CPU (as I expected it would but wanted to see what would happen). I shut it down and switched it to UEFI. That resulted in the previously seen missing device.

I then used the import from VMWare option. That did import it as UEFI but booting resulted in the exact same error.

I do have a usb hard drive attached to one of them but I believe I exported it with that removed from the VMware side. I'll try again.

The other admittedly has a pass-through GPU. I assumed (I know, I know haha) that it would just ignore that during the migration. I'll try removing that as well.

Boot firmware under advanced is set to UEFI, yes. I can escape and get into the uefi menu at boot as well.

VMware vm also has EFI selected under boot options/firmware.

I've been importing VMs from VCenter to XCP-ng with decent success but find that my UEFI Debian boxes fail to start saying the device isn't found.

BdsDxe: failed to load Boot0001 "UEFI Misc Device" from PciRoot90x0) /Pci (0x3,0x0)/VenHw (...) : Not Found.

I tried booting into live cd and rebuilding grub but that didn't help. I'm at a loss at this point and not finding much to help find a solution. Tried both import from VMware and exporting to ova and importing that.

I know this is quite old but wanted to update with additional findings.

I had the same error and it turns out you are using a non-standard port. In this case, tcp/8888 which I have blocked on my perimeter firewall.