I've been hunting for a few months to try to get an answer to this question and haven't found any articles confirming my claim.
The product that I am integrating with involves Remote Desktop Services. Take the following scenario into consideration:
Two servers with Remote Desktop Services are installed and configured. I'd like to configure a "poor mans" load balancing to these two RDS servers using a Round Robin DNS entry. I've read somewhere in the past that in an end user's Remote Desktop Client v6.1 or above, the RDS Client is capable of being able to provide transparent fault tolerance of an RDP connection to a round robin DNS alias.
For example, I create a RRDNS entry called rds.company.com that entry points to both RDSHOST1 and RDSHOST2.
- With both servers online, I attempt to connect via Remote Desktop Client to rds.company.com
- I am connected to RDSHOST1.
- I close my connection and attempt another connection to the alias and am connected back to RDSHOST1.
- I then power off RDSHOST1 and attempt another connection to the alias, I then get connected to RDSHOST2 (there is no error message or anything displayed by the Remote Desktop Client to the end user showing that it may have failed connection to RDSHOST1 initially).
- I then attempt a few more connections to the alias to confirm that I am still being connected to RDSHOST2 without any error messages being shown in the client.
- I power off RDSHOST2 and power on RDSHOST1.
- Attempt another connection to the alias and am connected to RDSHOST1 without seeing any connection failure messages in the RDS Client.
Does this mean the RDS Client has some built in ability to take a RRDNS alias and intelligently attempt to connect to one address returned from the alias and if it fails it automatically attempts a different entry in the alias until some predefined timeout? I'm trying to understand why this works the way it does and whether this is a "feature" of the RDS Client that seems to be undocumented.
Any info on this would be greatly appreciated.
Thanks,
Carl