Hi All,
We have two 2012 servers that are a part of an NLB cluster using multicast. We are using Cisco networking equipment. Each VMware VM has a single NIC configured.
We are finding that after a few short minutes of inactivity, the NLB's IP address will timeout until the NIC on either VM is disabled/enabled, and then it works again for a few minutes.
This issue does not appear to be reproducible on our 2008 NLB servers that are on the same host/subnet.
Machines on the same subnet are also unable to resolve the cluster IP address, in addition to systems on remote subnets.
The router has a static ARP entry for the cluster IP/MAC address, which is all that we have done in the past for our Windows 2008 servers and never had a problem. We have also tried adding a static entry to the MAC address table to meet the same results.
Did something change with 2012? Please let me know if I can provide additional information.