Renaming party
First let’s rename the “Cluster Disk 1” to its iSCSI LUN name “HyperStorage”. This makes it easier to identify in case more storage is added and you need to end-to-end troubleshoot.
If you take a closer look at the three cluster networks, the installer actually grabbed the correct subnets for each NIC. Now I only have to rename the networks to reflect their purpose.
As you can see, I’ve named the NIC’s on my N1BC node (and not yet on ther other two) to see if the wizard grabs the names…. and so it does! Now I’ll name the others accordingly.
Cluster configuration
Select the properties of the TTG Storage Network and disable cluster network communication for this network.
For some reason the virtual switches need to be created on each node separately AND they have to be named exactly the same to be used throughout the cluster!
I’ll create the Virtual:DMZ switch on each node now. <change pica’s ! >
Enable the CSV
Microsoft want you to know that you only use the Cluster Shared Volume for Hyper-V. Well, kinda hard not to agree to it at this stage 😉
And then we’re off…
Now a new option appears in the list under Failover Cluster Manager: Cluster Shared Volumes. Click on this and in the rightside window click on Add Storage.
Select the HyperStorage disk and click OK. This will add the storage and complete the CSV configuration.
It seems that the CSV is stuck on “Redirected Access” instead of the normal status of “online“. After a quick iSCSI connectivity check I found that on reboot of N2BC it did not re-attach the HyperStorage LUN automagically. I’ve added it through the iscsicpl command to the favorite targets and re-attached it manually.
Now I could select Turn off redirected access .. and all is well!
Heartbeat Network
For the failovercluster to work over the correct network, in my case the TTG Heartbeat/Live Migration (the 13-network), it must know how to use that. For this you need to use powershell on a node. On Hyper-V Server this works a bit differently than on a normal Server 2008 as the powershell here does not start with all the modules loaded.
- On a node, start powershell from the cmdline
- Enter the following command: Import-Module FailoverClusters
- Enter the following command: Get-ClusterNetwork | ft Name, Metric, AutoMetric, Role
As you can see, during the creation of the cluster networks the wizard grabbed the correct network. How can you tell? Because it has the lowest Metric value and the network with the lowest Metric value is the preferred network for CSV failover.
Just to be sure, check this also on the two other nodes. Nice, all the same!
Live Migration Network
As far as I can see this is configured through the VM configuration within Failover Cluster Manager. It only needs to be set one time and it does not matter which VM is used to set this parameter. Weird, I know but hey it works.
Just deselect all networks and only leave the Heartbeat LM (Live Migration) network selected. This now applies to all VM’s.