Rethinking Storage With Microsoft's Resilient File System

If you want to move to Hyper-V 2016, you must check this out.

• By Rick Vanover
• 08/24/2016

Over the years, we've seen a number of improvements from Microsoft in the area of new storage technologies. They include SMB 3.0, Shared VHDX, ODX, Storage Spaces Direct, Deduplication and now Resilient File System (ReFS).

ReFS in particular has been moving very fast in terms of innovation. Paul Ferrill has covered the v2 release of ReFS then in the technical preview stage of Windows Server 2016; now the third version of ReFS is in the latest technical preview.

After looking it over, I've become convinced that ReFS 3.0 should be your storage technology if you'll be using Hyper-V 2016. That's because ReFS 3.0 is an intelligent file system, and it's going to help Hyper-V environments in scores of ways. Everything from VM provisioning, to backup, to migration and deletion can benefit from ReFS on the Windows Server 2016 platform.

In fact, I'd go so far as to say that you'd really be doing Hyper-V 2016 wrong if you stayed with NTFS. This is obvious in the "single server with direct-attached storage" environment; but environments using shared storage also need to see the light of ReFS as well. This approach, Storage Spaces Direct, will also leverage ReFS as a Clustered Shared Volume resource.

While selecting the right or best storage arrangement is up for debate, and should be based on your organization's requirements, I think it's imperative that Hyper-V 2016 testing also include the latest storage technologies leveraging ReFS.

Part of that means considering that it may be time to re-think your Hyper-V storage strategy. I've seen, a number of times, the situation occur where an upgrade doesn't change bad decisions from the past, or fix issues with the virtual infrastructure "plumbing." I've seen this in Hyper-V as well as with other virtualization platforms.

My top tip here is to consider your upgrade to Hyper-V 2016 an opportunity to migrate to a new cluster, built completely correct. The new cluster would have the storage, networking and clustering configuration correct for the platform. Replication is a good way to migrate completely to a new cluster, yet the old environment is there in case things don't go as planned.

The final task is to ensure the VM is updated with the Update-VMVersion cmdlet. This will ensure that the VMs are able to use the newest features of Hyper-V 2016.