Hey Everyone,
Rather strange issue we've run into in the past month or two. I just submitted a case to MS about the issue, but I'm curious if anyone here has seen this.
Environment: 5 2012R2 RDS Servers, up to date with all updates. Clients - Win 7 all up to date. Users have roaming TS profiles.
In late december, we started to see an issue where RDS servers seem "stuck", and nobody can log into them. A user will sit and wait at the RDP login screen 'Please Wait for User profile Service' indefinitely. It's not that logins are slow, they don't work at all. Even administrators can't login (RDP, RDP console, local console). The only way to 'unstick' a server is to hard-reset it. Users can log in again, but typically only for a day or two.
We noticed in January that certain users seemed to be causing the issue. When user John Doe would log into a server (often unsuccessfully), that would 'lock' a server, and then nobody else could login (as per my description above). As such, we deleted all users roaming RDS profiles. This seemed to 'fix' the issue for 3-4 weeks, but has returned today.
We've researched this issue to death! Lots of forum threads about 08R2 having this issue, but also lots of hotfixes. Apparently in 08R2 this is caused by filesystem deadlocks. Lots of people seem to have this issue with 2012R2 as well, but very little information and very few hotfixes. Most of the threads go nowhere and have no apparent fixes. We keep peoples RDS profiles very small (most under 5-20MB), so it's not like these profiles are very large. When the farm is working properly, most people can login completely within about 10 seconds).
We've seen a couple hotfixes that seem related to this issue, but none of them have worked (3047296, 3053667). We've installed all Windows updates (Except February 2016), but nothing has resolved this issue. The farm was working fine for about 8 months in up to december when we started to see this issue.
We built entirely new RDS servers from scratch mid-january. That did not fix the issue.
Event logs are pretty clean. The only major errors are group policy taking too long to apply. Again, logins aren't slow when this happens, they dont workat all. You can sit and wait all day the the 'please wait for user profile service' screen.
We've also tried the usual group policy settings (detect slow network connections, wait for network at startup, synchronous/asynchonous login, etc), nothing has helped.
There's a few other threads with seemingly similar issues, but no real fix.
Thoughts? The only hunch I have so far is it's somehow user profile or group policy based. But again, we keep profiles clean and small.
