
On Fri, Jun 24, 2022 at 4:33 PM Michal Skrivanek <mskrivan@redhat.com> wrote:
On 24. 6. 2022, at 11:58, Gianluca Cecchi <gianluca.cecchi@gmail.com> wrote:
Hello, I need to transfer a VM from oVirt 4.4 to vSphere. I see that the "Export as OVA" in the GUI exports in a format compatible with oVirt but not vSphere. Any hint? Any way to easily convert the oVirt generated OVA to a vSPhere compatible one?
not that i know of. it depends on how resilient is vmware's ova import ... but if it is a single or a few vms then you can give up on settings and just import raw disks?
Ok, in fact I went sort of that way. Simple VM with RH EL 8.5 and only one 10Gb boot disk that worked as the quorum device host of a rhel 8 cluster. More in detail, if it can be of any help for others: . Creation on vSphere of a rhel7 VM (no choice of rhel8 because it is currently on 6.5) without disks and with paravirtual controller . On the source rhel8 oVirt VM, update of initramfs, adding the vmw_pvscsi kernel driver cd /boot cp -p initramfs-$(uname -r).img initramfs-$(uname -r).img.orig dracut --force --kver $(uname -r) --add-drivers vmw_pvscsi /boot/initramfs-$(uname -r).img . export VM as OVA on oVirt . extract virtual disk from OVA (tar -xf qdevice.ova) . convert virtual disk into vmdk format (done on a Fedora36 workstation with qemu-img-6.2.0-12.fc36.x86_64) qemu-img convert -O vmdk 3750043c-100c-42e6-8bd5-2cff2f81ee79 qdevice1.vmdk . upload of qdevice1.vmdk to the ESXi server from vSphere Web Client, inside the target VM directory of the related datastore . connect via ssh to ESXi server and further convert the file to a more "acceptable" vmdk format cd /vmfs/volumes/vmfslocal-myesxi-2TB/qdevice vmkfstools -i qdevice1.vmdk qdevice.vmdk see here: https://kb.vmware.com/s/article/1028943 I got "Unsupported and/or invalid disk type 2" when booting, without this further conversion step . removal of qdevice1.vmdk from vSphere Web Client interface . From vSphere Web Client Add existing disk to the VM, using the newly converted disk (qdevice.vmdk) . boot VM (with cluster related services disabled) . customizations such as network reconfiguration (due to network hw change) and VMware Tools guest managed install dnf install open-vm-tools systemctl enable vmtoolsd --now In my case I didn't have qemu-guest-agent service enabled in the VM, otherwise I should have disabled it too Enablement of cluster related services systemctl enable corosync-qnetd systemctl enable pcsd reboot . At the end the 2-nodes rhel8 cluster automatically reconnects without problems to the quorum device host Gianluca