As part of the Windows Server 2016 course we made a setup where we sat up a NLB-server to switch between two IIS sites on hostname1 and hostname2, on the hostname2 IIS we sat up a site on a random port cant remember the port number for now lets call it 1234, this worked when accessing the site directly on http://hostname2:1234 but when we used the http://nlbserver:1234 nothing showed. But after suspending the hostname1 in the NLB, where the port 1234 was not being used on the hostname1 server, you could access the http://nlbserver:1234, and after doing this you where able to re-enable the hostname1 in the NLB, and then you where able to access the http://nlbserver:1234 and then i pointed correctly to the hostname2 every time, but it seemed like you had to get the nlbserver to discover that it had to look at both servers for the 1234 port, otherwise it would just reject the request. I had the exact same experience when recreating a similar setup at home. Shouldn't the NLB try the hostname2 when no response from the hostname1? Or is this by design?
Hi Richard, I actually just deleted the VM's since I'm going through the MCSA 2016 material, but basically if you have access to the Microsoft Learning demo setup for the Lab exercises regarding the NLB, you should see the same error there, since it was during the Lab at a Microsoft Course we ran into the problem, and everyone seemed to experience the problem, but with the workaround, where you suspend the HOSTNAME1 and access the NLBSERVER:1234 seems to resolve the problem. But when I'm done with my Live Migration tests and Replication exercises, on my machine i can try and regenerate the same errors. This is the problem with a computer with only 240GB storage, I have to delete, revert and reinstall a lot while testing and playing around with VM's, but I get pretty good at installing Windows Server 2016 :)