Users not logged off

[lawrencej]

Hi all,

We have setup a floating pool with the policy of refreshing the desktop immediately after disconnect.  However users are not always being logged off properly and are still showing as an active session after disconnect. 

This causes issues when the user tries to log on to a new session, we have to force log off from the admin console. 

Also a setting to forcibly disconnect users after 12 hours is set, but we can have sessions connected for up to 2 days, this should not happen?

Thanks,

James.

[PhilNyeTheITGuy]

lawrencej‌

I had the same issue, even though I had forcibly disconnect users set to 720 minutes i still had sessions time longer then 12 hours.  Ends up the Automatically logoff after disconnect in Remote Settings in my Floating link Clone pools were set to 75 minutes. Vmware advised to set it to Immediate.

I would check there first.

Unfortunately, I could only go down to 15 minutes and I'm still having issues one or two sessions.  Ends up after logging into the console of the problem machines I noticed its stuck logging off because of Taskhost.exe being hung.   You may want to check that as well.  I brought this finding up to vmware and of course they where completely hands off because its a "guest os issue". Even though their system should be forcing it past this point.

[cH1LL1]

This can help -

AutoEndTasks

https://technet.microsoft.com/en-gb/library/cc978604.aspx

usually we find its Outlook holding open the Calender reminders window.

So the desktop waits for the apps to shutdown.

[PhilNyeTheITGuy]

cH1LL1‌

That's a good technet, I found that also but was hoping to avoid making changes to the registry.  My only question with that is if the entries are changed in  HKCU on the parent will this take into effect on the child clones?  Or should the entries/keys be made on  HKU?

[cH1LL1]

I have this setup as a Registry add via GPO that applies to the OU where the desktops are sitting  - This works becuase I have the Loopback policy enabled.

User settings> Preferences> Windows Settings>Registry

Image attached of setting.

[PhilNyeTheITGuy]

I'm defiantly going to try this, thanks cH1LL1‌!

[cH1LL1]

Of course - but also remember that if those installed applications are still running "or the process is running" on the master image than that may be your problem.

It might be an idea to power the Master image up and make sure no processes are running that shouldnt be-  like an idle IExplore.exe or Outlook.exe or etc.....

Also when you shut your Master template down remember to release ip / flush dns and shutdown. (some think its over kill but it should be best practice.)

Goodluck man

[Redbridge]

We've had that issue since early this year and haven't been able to find a link until now.  We noticed that the SQL Database Integrity Check was taking up to 11 hours every night as part of the SQL Maintenance Plan on the vCenter Database, since we took that out of the nightly plan and moved it to once a week we have had no problems with sessions overrunning the timeout settings.

[PhilNyeTheITGuy]

@Redbridge I wish that was the case for us, our DB Integrity check duration is normally 5-10min.  Finally vmware replied in the ticket I have regarding this and sent me a GPO ADM template that they created to hopefully to fix the issue.  Sadly it also didn't work for us but I figured I attach it encase anyone else once to give it a shot.  Its got the AutoEndTasks Ch1ll1 recommended in his post along with a couple other registry changes to kill tasks/apps.

[EricNichols]

Put one of these "logging off" VM's into maintenance mode, RDP into it and have a look in the Application event log to see if this event exists: "The winlogon notification subscriber <Sens> is taking long time to handle the notification event (Logoff). "  Our solution to this problem is in this post: View 6.2 Windows 7 Ent 64bit fails to log off some users