Since Windows Server 1709, Data deduplication is supported for ReFS volume. That means that it will be also available in Windows Server 2019. I have updated my S2D lab to Windows Server 2019 to show you how it will be easy to enable deduplication on your S2D volume.

Requirements

To implement data deduplication on S2D volume in Windows Server 2019 you need the following:

• An up and running S2D cluster executed by Windows Server 2019
• (Optional) Windows Admin Center: it will help to implement deduplication
• Install the deduplication feature on each node: Install-WindowsFeature FS-Data-Deduplication

Enable deduplication

Storage Spaces Direct in Windows Server 2019 will be fully manageable from Windows Admin Center (WAC). That means that you can also enable deduplication and compression from WAC. Connect to your hyperconverged cluster from WAC and navigate to Volumes. Select the volume you want and enable Deduplication and compression.

WAC raises an information pop-up to explain you what deduplication and compression is. Click on Start.

Select Hyper-V as deduplication mode and click on Enable deduplication.

Once it is activated, WAC should tell you the percent of saved state. Currently it is not working in WAC 1804.

Get information by using PowerShell

To get deduplication information on S2D node, open a PowerShell prompt. You can get the list of data deduplication command by running Get-Command *Dedup*.

If you run Get-DedupStatus, you should get the following data deduplication summary. As you can see in the following screenshot, I have saved some spaces in my CSV volume.

By running Get-DedupVolume, you can get the saving rate of the data deduplication. In my lab, data deduplication helps me to save almost 50% of storage space. Not bad.

Conclusion

Data deduplication on S2D was expected by many of customers. With Windows Server 2019, the feature will be available. Currently when you deploy 3-way Mirroring for VM, only 33% of raw storage is available. With Data Deduplication we can expect 50%. Thanks to this feature, the average cost of a S2D solution will be reduced.

The post Storage Spaces Direct and deduplication in Windows Server 2019 appeared first on Tech-Coffee.

Moreover ReFS comes with some integrity and protection against corruption. ReFS doesn’t need check disk as we execute on NTFS partition. So when a Cluster Shared Volume is formatted by using ReFS, you’ll no longer be warn to run a CHKDSK on the volume.

Fixed VHDX Creation

On My Hyper-V Host I have two identical SSD (Crucial MX 200): one formatted by using NTFS file system and another with ReFS file system. So I create three fixed VHDX: 10GB, 50GB, and 90GB. Below you can find the time measure of the creation of a fixed VHDX on NTFS partition:

• 10GB: 22 seconds
• 50GB: 106 seconds
• 90GB: 203 seconds

Now I make the same test on the ReFS partition:

• 10GB: 1 second
• 50GB: 1 second
• 90GB: 1 second

Just for laughs, I create a fixed VHDX on the ReFS partition of 500GB. I need just a second.

So now you can create fixed VHDX almost instantly compared to a standard NTFS partition. Pretty cool no?

About Checkpoint

In Windows Server 2012 R2, checkpoints are not supported in production environment. But who has not used checkpoint on his production VM before trying something? If you have already used checkpoints, you should know that we have to limit the checkpoint usage in a small time interval. Otherwise, the checkpoint grows and it can be very difficult to merge the checkpoint with the VM.

Checkpoints benefit the ReFS accelerated VHDX operations because the merging process happens without data being copied.

Thanks to the ReFS Accelerated VHDX operations, I think that all volume storage to store VM in Windows Server 2016 should be formatted by using ReFS. It really improve the efficiency to create fixed VHDX and to merge checkpoint.

The post Why using ReFS with Hyper-V 2016 appeared first on Tech-Coffee.