Twice in the last week, we've experienced issues logging in to our VDIs. Horizon 6.1. In both cases, rebooting the Connection and Security servers has fixed the problems.
I could log in via horizon client and blast 100% to a users' VDI, yet when the user did it on his zero client, it would log in, he got a blank screen and about 20 seconds later it jumped back out asking him for username and password again. I could log in fine to my own VDI on the zero client - it just didn't like his. I restarted the troublesome VDI from within view. Removed and re-added from inventory but nothing would make it work from the zero client. Horizon said the machine was available and agent reachable.
Events in monitoring log are pretty useless:
User xxx has logged in to a new session on machine xxxxx
30 secs later, User xxx has disconnected from machine xxxx
Yet I could still log in reliably 100% of the time from my computer.
Rebooting the servers and the user can now login in 100% reliably using the thin client.
Machine that horizon says is reachable, available, assigned to me, not flagged as a problem. Yet when I go to connect it said no desktop sources available. Restart connection/sec servers and it worked fine.
Could log into the machine from a browser internal, but not a browser outside the network. Restart servers and it works fine.
Event log files don't show much either. Replication is working fine between the two connection brokers. We're only running a pilot/evaluation at the moment, but in production with multiple connection servers how could you break it down to troubleshoot these sort of issues? Any ideas what caused these issues?