Seperate Heartbeat Isolated Network on 2 Hosts

[Darktan]

Okay...

We have a couple ESX 4.0 systems and two clusters. One cluster consists of two hosts with about 6 servers. We're using the HA of ESX to use the failover abilities. These are Exchange 2010 servers and they need a backend private network to communicate heartbeat and replication. The servers are on each host. But we're having issues creating the seperate network. We have one NIC in each host split off and not used for anything. The NIC's are connected to virtual switches in ESX and physically connected to a small switch that is not hooked up to the regular network. We don't use vlans (yet). ANy suggestions on making this work?

Peace

--T

[weinstein5]

Are you using MSCS for your exchange servers? All you will need to do is add another virtual NIC and connect it to the isolated virtual switch and then when configuring the virtual NICs in the VM OS make sure they are on the same subnet -

If you find this or any other answer useful please consider awarding points - 

[a_p_]

This should basically work without issues. Did you setup the port group on the vSwitch as a Virtual Machine port group and add an additional virtual NIC to the Exchange VMs connected to this port group?

Can you please post a screenshot of the virtual nerwork configuration?

André

[Darktan]

One of the problems here is the other group is actually trying to make these corrections. They're the Exchange team and I work on well, everything else it seems like. They are talking that the cluster that they are making are causing issues and screwing things up. So this might not be a VM issue at all. But I wanted to make sure the configuration on the VM side is good.

This is one of the hosts. They failed all the systems to the one host to get it to work at least for now. NIC0 is my vMotion, 1 and 2 are production and NIC3 is supposed to be the private network. From everything I know it should work. But again, it might be Exchange or the MS Clustering thing that is causing the problems.

[Darktan]

Oh and yes, the VM's do have a secondary NIC pointing to the isolated replication network.

[a_p_]

From the virtual network configuration I cannot see anything wrong, except for the VLAN setting you are using on the port group!? You mentioned you don't use VLAN's yet. Is the physical switch port configured as a trunk port in this case?

About the Exchange Cluster. I could be wrong, however if this is set up as an MSCS the Exchange team should really take a look at the MSCS configuration guide about "Cluster across boxes" and the configuration of the physical RDM disks! Maybe there's something wrong with the virtual SCSI controller configuration. (see http://kb.vmware.com/kb/1004617)

André

[Darktan]

The switch is just a dumb switch not connected to anything else. We have a Nexus 7000 sitting in a rack because our network is too lazy to wire it and install all the other switching that we bought to go with it. Once that's done, this will be done with vlans. But it's a good 2 months before he gets off his arse.

I think it's a problem with the cluster too. I'll forward that link along and let them figure it out. They tried all day yesterday to get the NIC's working before I finally heard about it and went in and set it up like it is (took me 10 minutes). I thought I had done it right but with all the problems they are seeing this morning, I wanted to make sure.

Thank you for the help.

Peace

[a_p_]

The switch is just a dumb switch ...

In this case remove the VLAN tagging on the port group. That should do the trick.

André