The same benefit applies to extending an existing VHD. The same Metadata
manipulation applies when merging differencing disks into the parent; for
example, when merging a checkpoint. Instead of copying the blocks from the
differencing disk into the parent, the Metadata of the parent is updated to point to
the existing blocks on disk to which the differencing disk had written.
Checkpoints are now used for backups with Hyper-V in Windows Server 2016,
which means that backup would also benefit. These changes make the mentioned
VHD operations hundreds, if not thousands, of times faster than with NTFS,
meaning faster provisioning and less I/O on the underlying storage.
Does this mean that ReFS is everywhere in Windows Server 2016? At the time of
this writing, the answer is no, although this could certainly change. Currently,
ReFS cannot be the Windows boot volume, it does not support the deduplication
feature, and its broad use has not been a focus for Microsoft. The focus for ReFS
usage is with Storage Spaces Direct and where high levels of resiliency are
required. Outside of those scenarios, NTFS is still the recommended filesystem.
Nonetheless, when you consider the benefits it brings to Hyper-V and VHD files,
it is certainly worth evaluating beyond just Storage Spaces Direct, provided you do
not need data deduplication. Additionally, when ReFS is used with Cluster Shared
Volumes, the CSV will run in filesystem redirected mode, which means all I/O is
sent over the cluster network to the coordinator node. In NTFS, all nodes can
perform Direct I/O to the shared storage. This is a big reason to carry on using
NTFS for CSV, except for Storage Spaces Direct, which does not have shared
storage anyway, and all access is already redirected. What Microsoft has is a
maturing filesystem that is now in its third version with Windows Server 2016,
and that is ready for the big time when a need requires its particular talents, such
as Storage Spaces Direct.Using Storage Spaces Direct is simple, but it must be enabled at a cluster level, and
once it’s enabled, you will not be able to use clustered Storage Spaces. Management is
done using PowerShell or System Center Virtual Machine 2016. To enable Storage
Spaces Direct, use the following command:
Enable-ClusterS2D [-S2DPoolFriendlyName
This command not only enables the feature but also adds all usable storage to a new
Storage Spaces Direct pool with tiers based on media type, and uses the highest
performing storage for the cache when more than one type of storage is present. You
can specify a name for the pool if the default nomenclature of S2D on
is not desirable.
If you have different types of SSD (as opposed to NVMe and SSD, which will
automatically be used in the best way) and you want to specify the model of device
that should be used for caching, you can add this to the Enable-ClusterS2D command
via the –S2DCacheDeviceModel "
created with the names of Capacity and Performance, and the cache will use the