Take the 2-minute tour ×
Server Fault is a question and answer site for professional system and network administrators. It's 100% free, no registration required.

I am doing some experiments in NLB using Windows Server 2008 R2 Standard. I have 2 windows machines, each with 2 nics. On each machine, only 1 NIC is used, the other one is not plugged into the network and is disabled at the operating system level.

I have setup a NLB cluster in unicast mode with the 2 machines. According to what I've read on some technical articles, in unicast mode, nodes part of the same cluster should not be able to talk to each other.

What is confusing me is that they actually can. I can ping one from the other, and I can also launch a remote desktop connection from one to the other without any problem.

Am I doing something strange, or had the unicast limitations been lifted in Windows Server 2008 R2?

Thanks in advance.

share|improve this question

1 Answer 1

up vote 0 down vote accepted

Nodes should be able to talk to each other - for the cluster to be working properly you need to maintain a heartbeat between the nodes in the cluster so they must be able to see each other.

share|improve this answer
    
According to support.microsoft.com/kb/556067: "1. Ordinary network communications between cluster hosts not possible. Network traffic intended for any individual computer within the cluster generates additional networking overhead for all computers in the cluster. 2. Further to this, we cannot use Network Load Balancing Manager on this computer to configure and manage NLB nodes." I am either doing something wrong, or the behaviour of NLB is not as documented. –  Guillaume Sep 16 '10 at 9:11
    
That doc refers to pre 2008 servers though. I've only used NLB on 2008 and intra node worked fine. This KB support.microsoft.com/kb/898867 mentions that behavior like it's a bug so I presume the resolved it in 2008 onwards. –  Chris W Sep 16 '10 at 11:04
    
Thanks Chris, I did try to setup MySQL master-slave replication from one cluster node to the other, and it worked fine. As you said, this is a bug which must have been fixed. –  Guillaume Sep 16 '10 at 11:46

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.