Description of problem:

I had an incident yesterday, I restarted vdsm during migration. And guest stayed in kind of paused "state" and was marked red (running vms are marked green).

When I tried to start or stop guest (I'm not sure if start option was even there) I got unexpected error. And I really wasn't able to get rid of that paused state.

vdsClient -s 0 .... destroy ... returned unhandled exception as well.


These things just happens in production and we're not properly handling this.

service vdsmd restart (on all nodes) did not help.

So I rebooted hosts and restarted RHEVM service.

Something that might have make somebody really angry.

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

rhevm ic104
vdsm-4.9-54.el6.x86_64

How reproducible:


Steps to Reproduce:
1. try to migrate host from node to node (there was kind of problem with migration, but this is not relevant now)
2. restart vdsmd on hosts
3. see guest in some kind of "unhandled" state
4. get rid of it without restarting any service or host
  
Actual results:

got to restart host and rhevm (not sure if rhevm was needed)

Expected results:

Stop button should stop guest in no matter state. All clusters have some kind of flush method that can help you check if deamon isn't already dead and so on ... you should really implement something like this.

Additional info: