In an effort to make my web front end severs a little more redundant, I introduced Windows Server 2008 R2 Network Load Balancing to my environment. After configuring the environment as per Microsoft documentation, I was not able to connect to the virtual IP I'd setup for my test site collection. After banging my head against the wall and rebuilding the servers a couple of times, I opened a ticket with Microsoft.
It turns out that you can't use NLB in the recommended unicast configuration if your cluster nodes are on two separate VMWare ESX hosts as per this article: http://kb.vmware.com/selfservice/microsites/search.do?cmd=displayKC&docType=kc&externalId=1556&sliceId=1&docTypeID=DT_KB_1_1&dialogID=26310418&stateId=0 0 26304946
To learn more about the potential problems you might encouter if you use Cisco switches, have a look at the following:
And from Microsoft:
In the end, I reconfigured the NLB nodes to use Multicast mode and then utilized the Cisco documentation to create a static entry for the MAC of the NLB cluster virtual IP at the local gateway.