We have encountered a slight issue when using Microsoft NLB in a VI3 cluster.
We have two ESX servers each attached to an independent cisco switch. We are using MS NLB in multicast mode and have configured each switch with a static CAM entry and add an arp entry on the router for the NLB cluster IP/MAC. As expected if all of the VMs in the NLB cluster are running on one ESX server (and subsequently one switch) the NLB cluster is not accessible.
I am wondering if anyone else has come accross this and if there is a solution that does not require an alternative load balancer?
Regards,
Stuart
The easiest solution would be to use an anti-affinity rule to separate one or two machines from each other so you are never in this scenario. Otherwise, I am not a fan of NLB, and prefer the other myriad of software and hardware load balancers. NLB introduces another management problem which I don't typically like. Sorry for the rant, but the rules should fix your immediate issue.
-KjB
Multicast mode should work just fine on the latest ESX Server. Which version of ESX are you using?