RDS Host server farm dropping connections due to unexpected connection broker round-robin routing failures

Hi Oliver Smith,

When a user initiates a session, DNS round-robin randomly points the initial RDP client request to any of the four RD Session Host servers in the farm instead of pointing dedicatedly to the RD Connection Broker. If the initial server chosen by DNS is not the target server assigned by the Connection Broker or the one holding the user's disconnected session, the Connection Broker issues an RDP redirection token to the client, but the subsequent redirection sequence fails midway because the client tries to re-authenticate against a changing pool of IP addresses returned by the round-robin query.

To resolve this synchronization drop permanently, you must change the network routing design so that all initial incoming connection requests are explicitly directed to the RD Connection Broker cluster rather than being load-balanced directly across the session hosts via DNS. You need to configure a single dedicated DNS A record pointing to the IP address of your RD Connection Broker, or utilize a dedicated hardware load balancer to manage the initial port 3389 traffic cleanly. On the RD Session Host servers, verify that the redirection settings are properly aligned by checking the deployment properties in Server Manager and ensuring that the registry key IPAddressRedirection under HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Terminal Server\ClusterSettings is correctly managing the token redirection behavior.

Hope this answer has brought you some useful information. If it did, please hit “accept answer”. Should you have any questions, feel free to leave a comment.

Tracy Le.