"Already Used" after every log off since upgrade to 5.1.3

[GCIT2013]

Last week we performed the update from 5.1.0 to 5.1.3 to resolve a minor issue but now we are having a larger one. Every linked clone machine goes into an Already Used state after the user logs off of it. Or even with our pools that have provisioned machines sitting, when someone logs on one, and the provisioned machine then attempts to become available it goes to Already Used. I used to find machines in this state every once in a while and I would just remove them but now it is nearly every machine every time it attempts to refresh.

Through the update I patched the View Connection Server, Security Server, and the View Agent all to 5.1.3. View Composer is at 3.0 and vSphere at 5.0.

I have found the KBs about changing the pae-DirtyVmPolicy from 0 to 1 or 2. If I do that, yes it will make the desktops available and not held up until I can manually remove them but I would like to fix the real reason this is happening since it is almost every single desktop every single time someone logs off. If I change the policy it'll be deleting machines constantly rather than refreshing properly like they should be,

We cannot upgrade higher than 5.1.3 because our storage won't support it at this time. Anyone have any ideas?

[nzorn]

I ran into this issue as well, but it's no longer an issue with version 5.3.  We removed the options to shutdown or restart the desktops, that seemed to cause the issue, instead we made the user use the log off button.  See this:  Solution for "Already Used" state

[GCIT2013]

Thank you! I will test with having the users just log off and see if it helps. Is there somewhere to remove the options for Shutdown and Restart? I think I remember seeing it somewhere but can't think of where.

[nzorn]

We use group policy: User Config > Policies > Admin Templates > Start Menu and Taskbar > Change Start Menu Power button and Remove and prevent access to the Shut Down, Restart, Sleep, and Hibernate commands

[GCIT2013]

Great! Thank you!

[mpryor]

Sounds timing related - 5.1.3 subscribes to an additional notification from Windows to be sure that logoff has fully completed before passing the ENDED notification back to the broker, as in 5.1.0 it was found that a long profile sync could sometimes be interrupted if we start the shutdown process too early. As nzorn says, if that's the case then not allowing users to shutdown but instead only logoff may help alleviate the issue.

I'm interested in the below statement in your original post, could you clarify what the exact reason is? There shouldn't really be any reason to not move to the latest 5.x View release even if you keep the same vSphere version for hosting the VMs.

"We cannot upgrade higher than 5.1.3 because our storage won't support it at this time. Anyone have any ideas?"

[GCIT2013]

I set the group policy for one of my pooled users to test the log off suggestion. And so far (only a day) the machines haven't gone into the already used state except for one. In this particular pool, I have a max number of 8 desktops, 3 always powered on, and 5 provisioned. So when one user signs on a machine another attempts to become available. That machine that was supposed to come out of the provisioned state to the available state is what went directly into an already used state.

So the log off only option may help for the users signing off and the machine refreshing better but this creates another situation where the machine gets stuck. You can see the event information of the machine in the attached image. I can make all of the machines available instead of powered off I suppose.

But to answer your question I was told that we had to stay on the same level of View as vSphere. Thank you for questioning it because now I see we should be fine going 5.3

[astamand2]

Hello,

I have encountered this in the past and was able to resolve it by adjusting a few key configuration settings in the ADAM database. VMware now has a KB Article (1000590) outlining almost these changes:

VMware KB: The View virtual machine is not accessible and the View Administration console shows the ...

Hopefully this will be useful for your situation.

[PCTechStream]

My only solution was "Remove" the VM from the pool, nothing else "Reset, Recompose, Log off/Disconnect  session" worked for me!