I'm running a MS Windows Failover Cluster with 4 nodes and one SCVMM host. All servers have the latest update of Windows Server 2012 R2.
What I'm trying to do is create some virtual machines on the SCVMM Host and then migrate them onto the cluster and vice versa.

The problem occurs every single time when migrating to or from Node 4.

If I create 4 VMs and migrate them to the 4 nodes, 3 complete successfully and the 4th one fails. The only difference I noticed is one extra step introduced when working with Node 4. The step is Install VM components. 

Successful migration to all other nodes:

Move virtual machine VM-3 from VMMHost to Node-3 using SAN  
Run pre checks for transfer 
Change virtual machine status   
Export Hyper-V virtual machine  
Deploy file (using LAN)  
Deploy file (using LAN)     
Deploy file (using SAN transfer)    
Import Hyper-V virtual machine      
Remove virtual machine

Failed migration with the extra step to Node 4:

Move virtual machine VM-4 from VMMHost to Node-4 using SAN  
Run pre checks for transfer 
Change virtual machine status   
Export Hyper-V virtual machine  
Deploy file (using LAN) 
Deploy file (using LAN)     
Deploy file (using SAN transfer)    
Import Hyper-V virtual machine      
Remove virtual machine      
Install VM components   

The same step occurs even when I attempt the process with 8 VMs instead of 4. Any other number (1,2,3) of VMs does not encounter the additional step.

The issue also occurs when migrating from cluster to host with the same error on the same VM and the same Node.

This is the error shown by VMM:

Error (13206)
Virtual Machine Manager cannot locate the boot or system volume on virtual machine VM-4. The resulting virtual machine might not start or operate properly.

I checked the Roles and Features on all the nodes and they are the same. Also, up do date. SCVMM is also up to date with UR7.

Any idea what causes this extra step to occur only on that particular node? And why does it fail? Also, how could I prevent it?

• Edited by slybloty 11 hours 53 minutes ago

Yes Elton Ji, all the VMs are created from the same template. There is nothing different in the process. 

Yes, that is correct. I can successfully migrate to and from any node except that 4th one. The "problematic" one is the node not the VM. 

The only problem I found after running the cluster validation was this:

- Node 'Node-1' is missing a virtual Ethernet switch 'HP NC382i DP Multifunction Gigabit Server Adapter #53 - Virtual Switch' that is present on at least one other node. Either remove the virtual Ethernet switch from all nodes or ensure that it is present on all nodes. - Node 'Node-2' is missing a virtual Ethernet switch 'HP NC382i DP Multifunction Gigabit Server Adapter #53 - Virtual Switch' that is present on at least one other node. Either remove the virtual Ethernet switch from all nodes or ensure that it is present on all nodes. - Node 'Node-3Node' is missing a virtual Ethernet switch 'Broadcom BCM5709C NetXtreme II GigE (NDIS VBD Client) #53 - Virtual Switch' that is present on at least one other node. Either remove the virtual Ethernet switch from all nodes or ensure that it is present on all nodes. - Node 'Node-4' is missing a virtual Ethernet switch 'HP NC382i DP Multifunction Gigabit Server Adapter #53 - Virtual Switch' that is present on at least one other node. Either remove the virtual Ethernet switch from all nodes or ensure that it is present on all nodes.

Also, I'm not sure how to check integration services state prior to the migration.

• Edited by slybloty 11 hours 53 minutes ago