Hello, thank you, this cleared up my confusion.

Do you have experience with migrating VMs?

I am trying to migrate a very old qemu installation/VM, for which its machine architecture is pc-q35-5.2. This Architecture is currently deprecatred and I am getting the error: "error: unsupported configuration: Target domain virt type qemu does not match source kvm".

I am a little afraid my qemu installation or the VM could break with an update.
Is there a way to still migrate?

On 4/7/26 14:54, Peter Krempa via Users wrote:
On Tue, Apr 07, 2026 at 12:55:06 +0200, [email protected] wrote:
Hello,

I want to migrate multiple VMs to a new host. In the process, several
questions came up, where I could not find an answer.  I am using
virt-manager version 1:5.0.0-5+deb13u1.

I reordered and renamed my storage pools: When migrating a VM, how can I
move the VM to a specific storage pool? Will the VM be moved to a storage
pool with the same name by default?
While virt-manager (likely) doesn't support this, with 'virsh migrate'
you can provide a new XML with updated definition where you change the
storage pool names. Look for 'virsh migrate --xml'. To get a XML
suitable to be modified you can use 'virsh dumpxml --migratable'


When migration finishes, the VM on the source host will still be accessible
but will shut down. Is that true?
If the VM was running originally (thus it's not offline migration) the
VM will be running on the source host up until the "switchover" point.
After that the source VM process is terminated and continues running on
the destination in the exact state it had right before switchover.

Can I prevent the VM from shutting down
during the migration?
Do you mean 'pausing' of the VM during migration (the virtual CPUs being
stopped while its being migrated?)? -> yes you can do that use '--live'.

Note that it will increase the amount of data copied over as any memory
page which was already migrated to destination but was written to by the
guest OS while it was running on the source will need to be re-copied.

Reply via email to