I have a weird issues that I'm hoping someone can help with. Upgrading the vmware view agent from 7.0.3 to any version above on any of our Grid pools (Nvidia Grid - Win 10) causes any user that logs in to be automatically logged out. For some reason this does not apply to admins, but only non admin users (99% of the people in the company. The same issue appears when the agent above 7.0.3 is used on a non floating pool as well.
Here is another piece of random behavior: When installing the view agent by just double clicking, the base image reboots right after the install begins. Just goes black and restarts the image. In order to install it, I have to right click and run as admin to get the install to complete.
Rolling back fixes the issue, but I'm worried that this issue follows us with each new agent version (so far 7.1 and 7.2). I have not been able to load Is there something I should check? I could not find anything that stated that the whole view environment needed to match the agent (the rest of the environment would remain on 7.0.3 for the time being), but I'm thinking I may need to upgrade everything.
This could be some compatibility issues with mismatched versions between the View agent and View Connection Server.
This also could be caused by permission issues or GPO's applied for these users.
Okay. Long time between posts. We have upgraded our environment to 7.6 and our agents to 7.5. The issue persists. VMware support has no idea what is causing this and so far we have been wasting support resources creating new pools to test. Any agent beyond 7.0.3 (we tested, 7.1, 7.4, and now 7.5) causes a situation where a user signs in and is immediately logged off. This issue does not happen when domain admins or users that are given domain admin rights login. There is also a new wrinkle. We have a third pool on our grid servers that is not having the issue. The pool is bound by the same policies as the other 2, but does not have the nvidia grid vgpu drivers installed. I sent the agent logs from one of the pools to the support agent, but they may be headed in the direction of it being a windows issue.
Anything else I can look at? I couldn't imagine any policy that would disagree with the agent.
Can you tell us more about the environment? I'm really curious what display protocol you are using and if tunneling is enabled on the connection servers or if UAG are used. It almost sounds like a firewall issue.
I think we figured it out. Our support agent recommended that we follow this KB and sure enough our userinit string looked like swiss cheese. The example had a normal one as
C:\Windows\System32\userinit.exe,"C:\Program Files\VMware\VMware View\Agent\bin\wssm.exe", and ours was "Userinit"=""C:\Program Files\ProfileUnity\userinit.exe",,",,"C:\Program Files (x86)\VMware\ScannerRedirection\diagutil.exe" /c /q /ds","C:\Program Files\VMware\VMware View\Agent\bin\vmlm_helper.exe","C:\Program Files\VMware\VMware View\Agent\bin\wssm.exe"," . As you can see there are number of extra commas and quotes that don't seem to go to anything.
When I cleaned it up and recomposed, we went from being instantly logged out to being able to login on the second attempt correctly. I then used our test pool without GPOs and removed the profile unity software and pasted the example userinit in our registry (what did I have to lose?) and recomposed again that worked correctly right away. I then did the same thing on a production pool with GPOs and again, it worked. So our issue is either with profile unity or a screwed up userinit string. Either way all of the issues we had with logouts, displays, etc. went away. We will probably need to test profile unity more to see if that is an issue, but at least we are not pulling our hair out over the issue. Hopefully this helps someone else.