Unable to logoff, disconnect, or reset terminal server user in production environment

What worked for me to resolve this same issue was to kill off all the processes running under the locked account from under Task Manager and then I was able to simply log off that account (from an Administrator account).

The user was then able to log back on under the account.

No reboot was necessary and no third party software needed to be downloaded.

I want to share how I reset of the account without the need to reboot the server. First of all you need to have administrator access to the server. I use the following logon option: mstsc /v:servername /console /admin in order to access the server. Then in "Windows Taks Manager", go to the Users tab and proceed to do a right click over the account that you want to "Log Off", select log off. This should free the locked session used by that account.

The simple answer is to run an elevated command prompt and type "Taskmgr" and then it will allow you to logoff the sessions under the USERS tab. It will not work without being in the elevated session.

I suppose the same happened today on my Win2008R2 Terminal Server. Sympthoms were: 1. He phoned me with "'connecting' message just hangs forever". He's just a simple user so I can't expect detailed problem description. 2. Tried logging off/resetting session (which usually helps in these cases) - did not work. The session still hangs in the list with 'disconnected' status. 3. Tried killing all processes for that user - did not help. Session persists and refuses to get killed.

Solution was - connect as user (login with his credentials if you can reset his password or use some kind of remote assistance to see what happens on his computer) and see what happens in logon window. When connecting I clicked on RDP Client's 'details' button - and here it was, a error message that winlogon did something wrong, it was waiting for user to click on 'retry/ignore/etc' buttons and since it's the omnipotent winlogon it caused all that weird behavior.

p.s. I could not found any way to really force kill a session :(

We just had a similar issue with our Windows Server 2008 R2 Remote Desktop server. The user session showed "Active" when looking at RDS Manager, but did not have the associated session ID# or connected device showing (both were blank).

All of the tricks above did not resolve the issue. When connecting as the user in question, an error message came back stating that the Terminal Server was busy and to try again later or contact the administrator.

We wound up rebooting the server as well.

I had the same issue in Windows Server 2016. The user was not able to login.

So I tried the following steps to disconnect the orphaned session:

1. on the CLI qwinsta lists all available sessions, inactive and active ones, there is one disconnected session (called "getr." in the screenshot) without a user name, but an session id.

show active sessions and kill them

2. with the session id (7) from 1. I tried to kill this session with either reset session 7 (fyi: rwinsta is an alias for reset session)

3. it worked for one session, but the next time it just had no effect, so I opened the task manager and the user tab. There you find one expandable list assigned to every remote desktop user - one list had no user name in it and only showed 4 tasks running.

4. I tried the obvious one: Logging off the user. Without any effect.

trying to logoff the user

5. So I tried to end those 4 tasks assigned to this user. Be careful, because some tasks, most importantly csrss.exe, when killed, also would result in a reboot of the system. I skipped them and just killed some obvious RDP-tasks.

After Step / Try 4 even the last crashed session was killed and the user was able to login again

6. If this is still not working, try this solution from another question: When you connect with the RDP client, click it's "details"-button. There you should see an error and you can even click retry or ignore.

You can start a cmd, do a query session, check the id of the session to be killed and then do a reset session. For instance, if with query session you get that the session name rdp-tcp#1 is the one you want to kill, then you can execute reset session rdp-tcp#1 and get it killed.

I'd the same situation: a Windows Server 2008 R2 with Remote Desktop Services, RDP connection set up to log off users after the session is inactive or disconnected 3 hours, and still some sessions remained locked. I tried to log them off with both Remote Desktop Manager and qwinsta/quser, with no success.

Here's how I resolved it:

1. I've located the session ID with qwinsta.
2. I've located the PID of winlogon.exe for the hung session with query process /ID:yourid.
3. I've killed the process with taskkill /f /PID yourPID.

Way to go. I'd like otherwise to find a solution for this to not happen.

For those who prefer a UI way of doing this.

1. Download SysInternal suite. (https://docs.microsoft.com/en-us/sysinternals/downloads/sysinternals-suite)

2. Launch procexp64.exe in Admin mode

3. Click on "Users" and select the user in question. Click on "Properties"

4. Note the Session ID

5. Go back to the main screen. Right click on the header of a column. Click on "Select Column"

6. Go to "Process Image" tab. Click on "Session" check box. Click OK

7. Click on the Session column header and find the Session from the box above.

8. Start killing each process. Some will say you cannot stop it, just continue to the next one. Eventually the ones that could not be stopped will stop by themselves.

1. Locate the session ID with qwinsta.
2. Kill all processes under the session taskkill /FI "SESSION eq 1" /F, assuming the session ID you wanted to end returned from qwinsta was 1.

This worked on Server 2012 Version 6.2 Build 9200, I would expect it to work on all version of windows.