Description of problem:
After starting a live-migration and if the process has not been started due to new node availability (virt-launcher pod in 'pending' state), then trying to cancel the live-migration (delete the resource VirtualMachineInstanceMigration kind) hangs forever.

Version-Release number of selected component (if applicable):

How reproducible:

Steps to Reproduce:
1. oc adm cordon <all-nodes-except-the-vmi-running-node>
2. start a live-migration (oc create -f vmim.yaml...)
3. watch a new pod virt-launcher is created and in 'pending' state
4. try delete the vmim cr (oc delete vmim...)

Actual results:
step 4 above hangs forever

Expected results:
live-migration to be cancelled

Additional info:
The way to mitigate this is to first delete the pod then delete the vmim cr.