Windows Server 2019 – ReFS deduplication

Docker Compose Tutorial for Beginners

While filesystems and deduplication features are technologies that you may not expect to be discussed when it comes to Hyper-V, the improvements in Server 2019 related to ReFS and the deduplication of data carry some huge advantages for Hyper-V servers. In case these are unfamiliar terms, let’s take a minute and define ReFS and deduplication.


Anyone who has worked on computers for a while will recognize FAT, FAT32, and NTFS. These are filesystems that can be used when formatting hard drives. The different versions of filesystems translate into different capabilities of how you can utilize that hard drive. For a number of years, NTFS has been the de facto standard for all hard disks connected to Windows machines.

That is, until Windows Server 2016 came along. We now have a new filesystem option called ReFS. Even if you work in an IT department every day, you may have never heard of ReFS because so far it isn’t getting used all that much. It is primarily used in servers that are involved with Storage Spaces Direct (S2D). If it’s the latest and greatest filesystem from Microsoft, why isn’t it being used as the default option on any new system? Primarily because ReFS is not a bootable filesystem. That immediately cancels out the capability for systems with a single hard disk to be running ReFS on the whole drive. What that implies is that ReFS is for secondary volumes on servers, perhaps volumes intended to hold large amounts of data.

In those instances where you do format a second volume to be ReFS and store data on it, there are some great resiliency and performance advantages to using ReFS instead of NTFS. These advantages were designed to make S2D implementations work better.

Data deduplication

Data deduplication is simply the ability for a computing system to discover multiple bits of data on a drive that are identical, and clean them up. If there were six copies of the exact same file on a system, deduplication could delete five of them, retaining one for the purposes of all six locations. This idea enables some major space savings. Data deduplication itself is not new; we had some capabilities introduced way back in Server 2012 regarding this.

Windows Server 2019 is the first platform where it is possible to enable data deduplication on a volume that is formatted via ReFS.

Why is this important to Hyper-V?

Data deduplication can be incredibly advantageous to run on a volume that stores Hyper-V VM hard drive files because, as you can imagine, there will be a ton of information that is duplicated over and over and over again when you are running dozens of VMs. Think about all of those Windows operating system files that will be identical among all of your VMs running on the Hyper-V host. It’s pretty obvious why it would be beneficial to enable data deduplication on a volume that stored VHDX files.

ReFS has some big resiliency and performance advantages over NTFS, and so it is also obvious that VHDX files would be best served by being stored on an ReFS volume.

Windows Server 2019 is the first platform where you can have your cake, and eat it too. We now have the ability to create an ReFS volume for storing virtual machine hard drives, and also enable data deduplication on that same volume.

Comments are closed.