Fix version_update_single_node to send shutdown request multiple times.

[anazobec]

Some VMs didn't want to receive a shutdown request, so version_update_single_node was updated so that the shutdown request is sent multiple times - in a loop.

[justinc1]

For some reason Windows VMs sometimes ignore ACPI shutdown. Current PR would:

• rapidly send multiple shutdown request
• wait 300 sec on VMs to shutdown

It would be better if send shutdown is moved into wait on VM shutdown loop. So if some VM does not shutdown, then we retry shutdown. That is inside shutdown_vms.yml, move taks name: Shutdown running VMs into:

# Wait up to 300 sec (30*10)
- name: Wait until VMs shutdown
  include_tasks: wait_vm_shutdown.yml
  loop: "{{ range(0, 30) | list }}"
  when: version_update_all_vms_stopped | default(true)

Would that work?