INTRUST GROUP - Managed Services Provider

Technical Blog

Moving from Hyper-V to Failover Clustering

25
Jan 2012
25 Jan 2012

Highly Available (HA) resource environments are truly amazing, when they work.

Recently, I worked with a client who wanted a complete infrastructure refresh. They were moving from 7+ year old equipment to a HA Windows Hyper-V Virtualized environment. Once the core infrastructure was set up with the SAN and two hosts, I launched Hyper-V Management Console and created a new VM and Virtual Network. The server came up and I built it to the client’s specifications.

Once the machine was set up, I configured Failover Clustering to make the hosts and Virtual Machines (VMs) Highly Available. This is where I encountered a problem. When testing LiveMigration and Failover, the VMs refused to move between hosts unless they were completely shut down and moved manually.

Looking at the logs, it reported that there was a network resource unavailable. I double-checked all of the network settings and ensured that all adapter names were identical between nodes. I tested Failover from Node1 to Node2 and vice versa, to no avail. Neither direction would allow the VM to Live Migrate or Failover.

Loading the Failover Cluster Manager (FCM), I navigated to the Virtual Machine settings. Under the VM settings, I simply removed and re-added the Virtual NIC binding from the Virtual Network that was created in the Hyper-V Manager.

I tested Failover and Live Migration again.

Success! The trick to remember is that if you have any VMs set up prior to implementing Failover Clustering and Cluster Shared Volumes, you must modify their settings from within FCM after you move them from Hyper-V Management, or it will not allow your VMs to be Highly Available.

Cale Plasket

Cale Plasket