VM accepts invalid instance type not satisfying preference requirements, ends up in patching failures. This is reproduced when creating centos VM with u1.small and then hotplug it to u1.nano(Done with UI) Even after restart the vmi/pods we are not able to recover VM.

Behavior Observed: 
1. Unable to stop VM
2. Unable to delete VM
3. Unable to patch VM
4. Unable to move VM to a higher IT . example : nano to medium/small 

virtctl stop centos-small-test -n default
error stopping VirtualMachine Internal error occurred: admission webhook "virtualmachine-validator.kubevirt.io" denied the request: failure checking preference requirements: insufficient Memory resources of 512Mi provided by instance type, preference requires 1536Mi


$ oc patch vm centos-small-test --type=merge -p '{"spec":{"running":false}}'
The request is invalid: spec.instancetype: failure checking preference requirements: insufficient Memory resources of 512Mi provided by instance type, preference requires 1536Mi

4.19

always

1. start VM and hotplug to nano for testing this use case
2.
3.

VM cannot be recovered

Should allow such a stage where it can't be recovered . Deleting VMI, pods nothing helped. 

No w/a .