ReFS, as it is popular known, is a file system first introduced in WS2012 but was less popular due to various limitation. ReFS in WS2016 is vastly improved and focused on Virtualization. ReFS are the file system built on NTFS. Therefore, lots of features offered by NTFS are present in ReFS.
What is Unique about ReFS
ReFS have features which not only can detect data corruption but as well can fix those corruptions. Corruption detection and remediation occurs online. So the biggest advantage is you do not need to take server or volume offline to run chkdisk command.
Corruption detections are achieved using Integrity-Streams. Integrity-Streams by default only does checksums on metadata and optionally also gives you an option to enable checksums on file data.
In ReFS, Proactive error detection & Correction is done by Scrubber. Scrubber periodically scans the volume for checking the metadata and integrity data streams. The scrubber prevents parts of volume getting corrupted by matching the checksums. If checksums do not match, corrupted copies are attempted to be replaced with original copies. In the case of attempts fail Salvaging process starts which remove corrupted file/directory from the namespace of course, without requiring a volume to be taken offline. Scrubbing process is transparent to the user.
ReFS is tightly integrated with Storage Space. Integration with Storage Spaces enables automatic repairing of corrupted data occurs using redundant copies of data. ReFS provide real-time tier optimization by dividing volumes into two logical groups. These two logical groups are referred as tiers can be created based on the disk and resiliency types. Performance and capacity are the two tiers available. All writes occur in performance tier, and eventually, data is moved to capacity tier over a period. In summary, ReFS uses Tiers to write hot data on performance tier and cold data on the capacity tier.
ReFS for Virtualization
Sparse VDL: VDL stands for Valid Data Length. VDL is the actual amount of data written on the disk against the provisioned size. Sparse VDL zeroes out a disk in seconds irrespective of the size of the disk you provision for the VM. Sparse VDL feature is of great use for provisioning VMs as disk allocation happens quickly irrespective of disk size.
Block Cloning: As the term says blocks are cloned. VM Cloning and Checkpoint merging operations make the best use of this feature.
What are the Use Cases of ReFS?
Primarily ReFS is focused on Virtualization. There are major enhancements which makes Hyper-V highly attractive in terms of performance, data integrity. At the same time, it can be used as Backup target for archiving data and finally where NTFS limits the file, volume or directory growth
Virtualization
- The biggest benefits of ReFS can be leveraged by Hyper-V especially features like
Sparse VDL and Block Cloning as explained earlier aids creation, expansion, and merging of VHDX easier - Scalability: ReFS provides features and functionality which efficiently scales and support massive data sets and therefore is considered highly scalable file system. NTFS can create maximum file size of 256 TB, in contrast, ReFS maximum file size is 4.7 Zettabytes which means you can have large volumes, a file which is typically required for hundreds of Virtual Machines
- Real Time Tier operations with Storage Spaces provides high performance as explained above
Data Archival & User Documents
Data integrity of data at rest is very critical. Data in archival is hardly used therefore there is no way to find out if data is corrupted. Scrubbing, Integrity-stream ensures data integrity is intact. Scalability of ReFS file system is also best suited for archived data as the size of data grows in TB’s over a period. It is strongly recommended to use storage spaces to get advantages of ReFS features particularly automatic and online repair. With Storage Spaces integration, redundant data copies help automatic detection and correction of corruption within metadata and data. Scalability of ReFS, therefore, is also applicable to application where data growth is exponential e.g. Exchange and MS SQL Server.
Things to Consider While Planning ReFS
Below things must be considered before deploying ReFS
- You cannot boot from ReFS volume
- You cannot build this file system on removable media
- ReFS does not have below features
- Compression
- Encryption
- Quota management
Many of the best features are available only when used in conjunction with Storage Spaces
Difference Between NTFS and ReFS
In the figure below, it can be seen that Quota tab is not in ReFS volume (right side). Likewise, you do not see compress check box seen on NTFS Volume (left side bottom)
Furthermore, encryption feature available on folder created on NTFS (on the left side) is not available in the folder set up in ReFS which is visible by the absence of advance settings.
