Home / server / Questions / 713620
Accepted
Arvo Bowen
Asked: 2015-08-13 05:26:16 +0800 CST

After making my VM highly available I get "Could not find Ethernet switch" error when trying to live migrate

  • 772

I normally create new VMs from the Failover Cluster Manager GUI. But this time around I went a different route because I had to (due to time constraints and available resources at the time). For this one VM I had to create it on cluster node1 directly. Then when I got it up and running the cluster had the disk available for me to use so I shut down the VM, then added it as a service/role in the Failover Cluster Manager GUI.

That seemed to come over into the Failover Cluster Manager with no issues at all.

Then I tried to do a Live Migration and got the error "Could not find Ethernet switch 'Broadcom NetXtreme Gigabit Ethernet #3 - Virtual Switch'."

A few things to note...

  • The NICs are named the same exact thing on BOTH nodes.
  • The NICs do seem to have the "device names" mixed up.
    • Node1 NIC's Name:NIC1
    • Node1 NIC's Device Name:'Broadcom NetXtreme Gigabit Ethernet #3 - Virtual Switch'.
    • Node2 NIC's Name:NIC1
    • Node2 NIC's Device Name:'Broadcom NetXtreme Gigabit Ethernet #2 - Virtual Switch'.
  • I have all NICs enable and in use except for ONE, and that's NIC1.
  • I could not find "Refresh the Configuration of a Virtual Machine" on Server 2012 R2 (like there was in Server 2008 R2).
  • I tried refreshing the VM configuration by using 'Update-ClusterVirtualMachineConfiguration -Name "Virtual Machine Configuration A PAPTR"' on both nodes (with no errors, just shows the VM status).
  • I created a new VM via the Failover Cluster Manager GUI and it live migrates with no issues at all.

Any suggestions?

windows-server-2012-r2

1 Answers

  1. Best Answer

    Finally figured it out...

    I kept thinking it was because of the WINDOWS "Device Name" in the Network Connections list. I was completely wrong and it kept me going down the wrong path.

    The real reason the issue was happening was due to the Virtual Ethernet Adaptor created by Hyper-V. Each of your cluster nodes will have one (or should) in order to get your VMs a network connection. To get to the Virtual Ethernet Adaptor you need to access the Virtual Switch Manager. To get to that open up the Hyper-V Manager, on the left right click the server node (you can add all your nodes here to manage them rather than logging on to each one) and select "Virtual Switch Manager...". Once you do that, click on the Virtual Switch already created (below "New virtual network switch") or create a new one. The "Name" field needs to be the same on ALL nodes you plan on live migrating to. In my case I changed all mine to "HyperV Virtual Switch".

    enter image description here

    • 1