A vm that receives a shutdown requests moves to 'Powering down' state. However, the request may be ignored if there is no guest agent, or no acpi support, or user cancels the request. In that case, VM state changes back to 'Up' after a couple of minutes.
RHEV-M should notice this transition (Powering down --> Up) and notify admin.
+++ This bug was initially created as a clone of Bug #529884 +++
--- Additional comment from acathrow on 2009-11-18 10:08:50 EDT ---
(In reply to comment #6)
> Andy,
> #1: Having a RHEL agent is a nice goal. do we have a formal feature request for
> that?
> #2: my patch, already in vdsm-4.5-5, adds a 5 minutes timeout. VM will appear
> as 'Up' in RHEV-M 5 minutes after shutdown request is ignored by guest.
#1 Yes, requested for 2.2 but still up in the air
#2 - will we get an error logged in RHEV-M to indicate that the action failed?
closing old bugs as part of cleanup.
you can reopen the bug if you find it relevant and important still.
please consider opening a new bug on upstream ovirt component if the issue is still relevant.
Every month or so, people come to me ranting "I've asked to shut down a VM, but it did not not shut down, why do I see it Up again?". An engine event would warn the admin that the VM does not want to be shut down cleanly, and should be either killed or the guest addressed.