Backup Nic configuration For cluster server

[kk5]

We are having a 2 node mscs for exchange. each node has 3 nics. 1St nic is called production nic(this nic has been configured for all communications). 2nd nic is called backup nic(configured for Client access only,) 3rd is heartbeat nic, connecting 2 nodes using a cross over cable(configure for internal cluster communication).
The problem is , when network team runs sniffer, they are able to see the two servers sending UDP/3343 packets to each other through the backup nic.
We have a heartbeat nic exclusively to communicate between the two servers, but the communication is also going on backup nic. how can we resolve the issue.

 

[xmsre]

In a cluster, the private is set to "internal communications only" and the public is set to "all communications". On the properties of the cluster, You set private first in the binding order. The private NICs on each node should be connected via a crossover cable, have netbios disabled, and have no gateway address. If you have a second heartbeat, treat it exactly like the first. Given this, I see a couple issues with your setup:

1. Why can the network folks see any traffic on the heartbeat? It should be going over a crossover cable and not connected to a switch.
2. If heartbeat traffic is going over the public, then the private is not first in the binding order or the private is down.

 

[kk5]

thanks for the reply, let me answer your questions
1. on the properties of cluster, private is first in the binding order- yes it is first

2. private nics are connected via crossover cable and they are using static ip. netbios is disabled and there is no gateway

3. network folks say that every second the two nodes are sending packets to each other through UDP port 3343.

4. Backup nic has been configured for client access and we want to use it only when the backup jobs run.

After the changes were made we never rebooted the server. do you think a server reboot will help. Appreciate your help

 

[xmsre]

I don't believe the changes in binding order will take effect until a reboot. You can do this one node at a time to prevent an extended outage.

 

[kk5]

I did a reboot of both the nodes, after the reboot, network team has run a sniffer on backup nic and still they can see the 2 nodes communicating with each other through backup nic.
On the backup nic i changed the subnet mask and the moment i changed it, i see the cluster going down,
And the reason for that is, for the cluster there is a virtual backup nic in cluster group resources and i didnt change the subnet mask over there. but my question is when i change something for backup nic why the cluster is going down. how would i configure the cluster to use the production nic. Please help me out.

 

[xmsre]

The traffic is obviously going over the backup NIC. Check your configuration on the primary heartbeat to figure out why it is not going across here.

One question. How is it possible to take a sniff on the heartbeat? The two nodes should be directly connected via a crossover cable, not going through a switch.

 

[kk5]

For Heartbeat we have a cross over cable connected between thte 2 nodes, and the heartbeat is working fine,

Backup nic is connected to a switch and network team says that nodes shouldnt communicate with each other through backup nic. they are sending and receivng packets to each other per second basis.

Note: Backup nic is connected to switch and the network team is sniffer on switch