A. Live Migration is the technology that copies the memory of a running Hyper-V guest to a target node while the guest is still running, which means there's no downtime of the guest required to transfer the virtual machine (VM) state from one Hyper-V host to another. To solve the time delay of dismounting and mounting the LUN that holds a VM's VHD files, the system uses Cluster Shared Volumes (CSVs), which allows all nodes in the cluster to access the CSV enabled storage concurrently, so no dismount/mount operation is needed.

If you're using pass-through disks, your pass-through disk can't be part of the CSV namespace, because the Hyper-V host isn't actually mounting the storage. The guest performs I/O directly, which means the storage isn't accessible to all nodes at the same time. In the Failover Clustering configuration of the VM, you'll configure the pass-through disk as a dependency for the resource group, so when you perform a Live Migration the pass-through storage will have to dismount from the current Hyper-V host and mount on the new Hyper-V host. This process will slow down the migration of the VM and probably cause a noticeable pause to clients, or even disconnects.

The answer is, then, that you can still use Live Migration with pass-through disks, but the dismount/mount of the pass-through will take time and may break the TCP connection timeout. Your best option is to use VHDs instead of pass-through, because VHDs can be part of CSVs, and you won't need the dismount/mount operation. VHD and pass-through disk performance is basically equal, so performance isn't a major concern.

Related Reading:

Check out hundreds more useful Q&As like this in John Savill's FAQ for Windows. Also, watch instructional videos made by John at ITTV.net.