Bug 1373567

Description of problem:

I had a host die. I couldn't get it back up. I couldn't put it in maintenance mode b/c it was down.  I didn't select  Confirm host has been rebooted b/c I know it wasn't. Suggestion to put a "Force Maintenance Mode" Checkbox in.

See IRC transcript from #rhev-dev.

<mpusater> Host still has running VMs on it and is in Non Responsive state."
<mpusater> I should add this is the only host in environment
<mpusater> and I don't care about the running vm's
<mpusater> I just don't want to rebuild the manager as well if I can help it.
<awels> mpusater: Okay, you can right click on the host and say the host has been rebooted, that will clear the running VMs in the database, at which point you should be able to put the host in maintenance.
<mpusater> aweis, that worked
<awels> mpusater: Once in maintaince you can remove it.
<mpusater> awels, I mean :(
<awels> mpusater: The confirm host has been rebooted is basically a manual fence.
<mpusater> Yes, I was able to remove it. Thank you.  I'm not sure how intuitive that is, as the host wasn't rebooted.
<mpusater> Ok, good to know.
<awels> mpusater: Its to stop the deadlock between not being able to talk to the host, and the engine thinking stuff is still running on it.
<mpusater> Ok, I understand now - Thank you.
<mpusater> Though I don't think any admin would ever guess that.
<awels> mpusater: If you have a better wording for the 'confirm host has been rebooted' feel free to open a bugzilla with some better wording.
<mpusater> Have we ever considered a "force to maint mode" check box, that does all that behind the scenes.
* omachace has quit (Client Quit)
<mpusater> As an admin I know the host hasn't been rebooted. so I wouldn't think to confirm it's been rebooted.
<awels> mpusater: That is a good idea.
<awels> mpusater: I figured it out when having some issues with my hosts and trying to re-install and was unable to get it in maintaince mode. I figured what the hell, what can it hurt.
<awels> mpusater: Maybe the option should say confirm host has no VMs running on it.
<mpusater> well that too is slightly mis-leading.
<awels> mpusater: Since its dead already.
<mpusater> yeah, but would that be the case all the time?
<awels> mpusater: The problem is, the engine has no way to know, so it needs the human to tell it
<mpusater> Sure I get that.



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

3.5, but I suspect it's on all future versions. 

How reproducible:


Steps to Reproduce:
1. configure 1 host in rhev-m.
2. Turn host off and act like it's unrecoverable(aka, hardware failure)
3. How to remove host from manager when it's completely down.

Actual results:


Expected results:


Additional info: