Good day.

We have 2 TMG 2010 servers working in NLB cluster. When we are opening "Network Load Balancing Manager" on any TMG server we receive the error message "the RPC server is unavailable" (it's about another node in cluster).

I've tried to add access rule which allows RPC (all protocols) between TMG nodes. But it didn't help. My next step should be disable RPC filter, but I can't understand which negative effect could be happened from this action.

Can anybody explain (better give the official link from technet) what could be wrong after disabling RPC Filter?

Perhaps there are another possibilities to solve this problem (we don't see another node in NLB snap-in, only local node)? 

Good day.

We have 2 TMG 2010 servers working in NLB cluster. When we are opening "Network Load Balancing Manager" on any TMG server we receive the error message "the RPC server is unavailable" (it's about another node in cluster).

I've tried to add access rule which allows RPC (all protocols) between TMG nodes. But it didn't help. My next step should be disable RPC filter, but I can't understand which negative effect could be happened from this action.

Can anybody explain (better give the official link from technet) what could be wrong after disabling RPC Filter?

Perhaps there are another possibilities to solve this problem (we don't see another node in NLB snap-in, only local node)? 

If you don't have issues and your configured NLB in 'unicast' mode you can ignore this warning. It is by design because NLB in unicast mode will give your interfaces another unicast MAC Address, which is the same on both servers. But rest asured, NLB is working properly. Read the following link for more information:

"RPC Server Is Unavailable" Error Message When You Connect to NLB Cluster Host Through NLB Manager
https://support.microsoft.com/en-us/kb/291959

There is a solution though if you have an Intra-Array interface. If you configure the C:\Windows\System32\hosts file on each TMG Server, and configure the FQDN to point to the Intra-Array interface IP Address, it does work properly with the NLB console.

NOTE: I always prefer to use NLB in IGMP Multicast mode. But that is a switch-assited NLB method, which requires you to configure the switch/router.

Thank you.

I don't care about this warning, because cluster works well. The main problem is soft disactivate the node while installing updates OS. Another node is invisible in NLB and couldn't be paused correctly.  

Why should we edit hosts if we already DNS which succesefully resolve any host (as TMG servers and IP array)? Does you solution really work? 

We've already enable multicast on the NLB cluster.

• Edited by Alexey Klimenko 17 hours 32 minutes ago

If you can't pause am NLB node correctly, than something is not configured/working properly. Just to be sure, make sure you drain the node via TMG.

If you are using NLB in multicast mode you should be able to see each NLB node. And yes, if you configure the hosts file so that TMG02.yourdomain.local resolves to the intra-address you make sure it uses the Intra-Array interface. Also when you open NLB Manager. The only thing is, not everybody does it. It is not well documented. I always configure it. For example, just a quick search on google shows me the following blog from someone who configures the HOSTS file as well.

Workgroup Deployment with Forefront TMG Enterprise Edition Part 2: Creating the Standalone Array
http://blog.msedge.org.uk/2010/05/workgroup-deployment-with-forefront-tmg_18.html

But the bottom line is. NLB should be configured properly and in TMG you should configure the Intra-Array communication IP Address or hostname correctly.

Thank you.

I don't care about this warning, because cluster works well. The main problem is soft disactivate the node while installing updates OS. Another node is invisible in NLB and couldn't be paused correctly.  

Why should we edit hosts if we already DNS which succesefully resolve any host (as TMG servers and IP array)? Does you solution really work? 

We've already enable multicast on the NLB cluster.

• Edited by Alexey Klimenko Wednesday, May 27, 2015 1:53 PM

Thank you.

I don't care about this warning, because cluster works well. The main problem is soft disactivate the node while installing updates OS. Another node is invisible in NLB and couldn't be paused correctly.  

Why should we edit hosts if we already DNS which succesefully resolve any host (as TMG servers and IP array)? Does you solution really work? 

We've already enable multicast on the NLB cluster.

• Edited by Alexey Klimenko Wednesday, May 27, 2015 1:53 PM

We' have configured the host file (include infomation about another node and IP cluster array) on both servers. But nothing change unfortunately.