Problem: randomly users login and fslogix does not load their previous profile it creates a new one so any files are missing.
When I look at the profile logs I see the same events in the logs as a good login except I get this towards the end of the login process.
[13:08:19.509][tid:00000e34.00021af8][INFO] ===== Begin Session: Profile Directory Changed: S-1-5-21-982881632-1095161353-1905309208-1406
[13:08:19.509][tid:00000e34.00021af8][INFO] New Profile Directory: C:\Users\Mobleyxx.XXXTEAM
[13:08:19.519][tid:00000e34.00021af8][INFO] ===== End Session: Profile Directory Changed: S-1-5-21-982881632-1095161353-1905309208-1406
Also if the user logs off and back on often fslogix creates a temp profile which we block and they can't login.
We are running version 2.9.7979.62170. We updated to this version from a version that was about 3-4 months older hoping this would fix the problem.
Users login to a 2016 RDS server VM. The profiles exist on another VM in the same ESXi cluster.
So far we can't determine what is causing any of the behaviors because as verbose as the logs are for fslogix there is no indication as to why the profile is changed.
I've compared good logins to bad ones in Compareit and its line for line the same til you get to the event logged above.
In order to fix the problem we delete the domain account and create it again from scratch and usually that fixes the problem. Luckily users are not suppose to save files on the profiles so we are not concerned with loss of files. We know its missing files really because shortcuts that are made through a script are missing.
While testing this while there were no users on the server and fixing some accounts that would not get fixed by the method just mentioned, I recreated the accounts again and the accounts worked fine. I suspect there is some performance issue that is causing this but again I don't see a performance problem when looking at the terminal server or the file server that is storing the VHDX
storage response time is between 1-10ms average is sub 1ms.
network graphs only show about 150mb/s of traffic on a 10gb port and network.
There are about 150 users logging in and only about 30-40 users with fslogix profiles atm. We have set this up to use local before moving to an fslogix so most users have not moved to a VHDX.
We are using GPO to manage the profiles.
profile type is set to Try for read-write profile and fallback to read-only
size of profiles is 20GB
we are not using cloud cache.