Current Environment:
HA Connection Broker servers and about 30 RDS Servers (all virtual). What we have noticed is that when an RDS Host is having response issue (it's still online and running but in a state where it's barely responsive, but apparently responsive enough that Connection broker doesn't see it as offline). Obviously when this happens users trying to log in that have a session on that host the connection broker tries to redirect them to. The normal timeout function never kicks in because the RDS Server is still responding. We have to figure out which RDS Server is not responding and reboot it to resolve the issue.
I have seen the registry keys related to HKLM\SYSTEM\CurrentControlSet\Services\Tssdis\Parameters timeserversilentbeforeping , pingmode, etc.
My questions are:
1 - what is the actual "ping" from the connection broker to the RDS Server, is it an actual ICMP Ping or an application ping of the terminal server service?
2 - any suggestions on how to automatically resolve this issue short of having to find the offending RDS Server and manually rebooting it?
-SMB