A reliable, step-by-step guide to bring a VirtualBox/VMware .ova into QEMU/KVM, with both GUI and CLI paths—and fixes for the most common pitfalls.Contents

  1. Prerequisites
  2. 1) Extract the OVA
  3. 2) Convert VMDK → QCOW2
  4. 3) Create the VM in virt-manager (GUI)
  5. 4) Alternative: virt-install (CLI)
  6. 5) Common gotchas & fixes
  7. Quick troubleshooting checklist

Prerequisites

  • QEMU/KVM, libvirt, and virt-manager installed on the host.
  • User is allowed to manage libvirt (typically being in the libvirt/libvirtd and kvm groups).
  • Enough disk space to hold the converted image (QCOW2).

1) Extract the OVA

An OVA is a tar archive containing an .ovf (metadata) and one or more .vmdk disks.

mkdir ~/ova-import && cd ~/ova-import
tar -xvf /path/to/your-image.ova
# Expect: appliance.ovf, disk0.vmdk, (maybe disk1.vmdk, .mf), etc.

2) Convert VMDK → QCOW2 (recommended)

QCOW2 supports snapshots and sparse allocation.

# Single-disk OVA
qemu-img convert -p -f vmdk disk0.vmdk -O qcow2 disk0.qcow2

# Inspect the result
qemu-img info disk0.qcow2

# (Optional) Grow before first boot
qemu-img resize disk0.qcow2 +20G

For multi-disk OVAs, convert each .vmdk to its own .qcow2.

3) Create the VM in virt-manager (GUI)

  1. Open Virtual Machine Manager → Create a new virtual machine.
  2. Select Import existing disk image.
  3. Browse to disk0.qcow2.
  4. Choose the closest Guest OS type (Linux/Windows and version).
  5. Firmware:
    • Try UEFI (OVMF) first if the source VM used UEFI.
    • If it fails to boot, recreate the VM using BIOS (SeaBIOS).
  6. Assign CPU/RAM to taste.
  7. Tick Customize configuration before install, then:
    • Chipset: q35 is a good default; try i440fx if needed.
    • Disk bus: start with VirtIO for performance. If the guest can’t boot/find the disk, switch to SATA. (Windows may need VirtIO drivers—see below.)
    • NIC model: virtio (or e1000 for maximum compatibility).
    • If using SCSI, add a VirtIO SCSI controller and attach the disk to it.
  8. Click Begin Installation.

4) Alternative: virt-install (CLI)

virt-install \
 --name my-imported-vm \
 --memory 4096 --vcpus 4 \
 --import \
 --disk path=/path/to/disk0.qcow2,format=qcow2,bus=virtio \
 --os-variant detect=on,require=off \
 --network network=default,model=virtio \
 --graphics spice \
 --boot uefi

5) Common gotchas & fixes

“network ‘default’ is not active”

If the default NAT network isn’t up:

virsh net-start default
virsh net-autostart default

If the network definition is missing, define it, then start and autostart:

# Path may vary by distro
virsh net-define /usr/share/libvirt/networks/default.xml
virsh net-start default
virsh net-autostart default

Black screen / “No bootable device”

  • Try switching Firmware (UEFI ↔ BIOS).
  • Change Disk bus from VirtIO → SATA to test.
  • Verify the disk has a bootloader/ESP. Attach it to a helper VM and inspect with lsblk/fdisk -l.

Windows guests (imported from OVA)

  • If using VirtIO disk/NIC, mount the virtio-win ISO in virt-manager and install drivers (during setup or via Device Manager).
  • If you see INACCESSIBLE_BOOT_DEVICE, switch the disk bus to SATA to boot, install VirtIO storage drivers, then switch back to VirtIO.

Linux guests boot but no network

  • Predictable interface names may differ. Check ip a inside the guest and update its network config.
  • For very old distros, try NIC model e1000.

CPU features / performance quirks

  • In VM CPU settings, use Copy host CPU configuration (aka host-passthrough) for best compatibility.

Multi-disk OVAs

  • Convert each VMDK and attach them in the same order as in the OVF (disk0 → vda/sda, disk1 → vdb/sdb, …).

Quick troubleshooting checklist

  • Used the right firmware? (Try UEFI first, then BIOS.)
  • Tried switching disk bus VirtIO ↔ SATA?
  • Does the libvirt default network exist and run?
  • For Windows: installed VirtIO drivers or temporarily used SATA?
  • For Linux: interface name changed? Adjust network config.
  • Set CPU mode to host-passthrough?

Tip: Keep the original OVA and the converted QCOW2 until the new VM boots cleanly.

Blogpost created with AI help