ReFS, Microsoft’s Resilient File System, is designed to look after itself. It checks its own metadata, can repair damage automatically when it sits on a mirrored Storage Spaces pool, and rarely needs any attention. That is exactly why it is so alarming when an ReFS volume suddenly shows as RAW in Disk Management, Windows says “The volume does not contain a recognised file system”, or a Windows Server share simply refuses to open. This guide explains what is usually going on, which familiar fixes do not apply to ReFS, and what to do first so the data stays recoverable.
Where you are likely to meet ReFS
ReFS is not the default file system for most PCs, so many people only discover they were using it when something goes wrong. Common places include:
- Windows Server data volumes, particularly backup repositories and Hyper-V storage
- Storage Spaces and Storage Spaces Direct pools, where ReFS is often chosen for mirrored or parity virtual disks
- Windows 11 Dev Drive volumes, which use ReFS
- Workstations running Windows editions that allow ReFS volumes to be created
If you are not sure, the Properties window of a healthy volume, or the File System column in Disk Management, will tell you. A volume that has already turned RAW will no longer say, so check documentation, backup software settings or a colleague who set it up.
Why an ReFS volume can suddenly show as RAW
“RAW” simply means Windows could not recognise a file system it understands on that partition. With ReFS, there are several typical reasons:
- The disks were moved to an older version of Windows. ReFS has its own on-disk version. Newer versions of Windows can upgrade a volume’s ReFS version when they mount it, and an older Windows installation may then be unable to read it and show it as RAW. This often happens when disks are moved between servers or a machine is rebuilt with an older image.
- A Windows update problem. There have been updates in the past that caused some ReFS volumes to appear as RAW until Microsoft issued a fix. If the problem started straight after updating and the disks are healthy, this is worth investigating before anything else.
- Metadata damage after a power cut or crash. ReFS is robust, but it is not immune to sudden power loss, faulty RAID controllers or storage that lies about completing writes.
- A failing disk or a degraded pool. If the underlying hard drive or SSD is developing bad sectors, or a Storage Spaces virtual disk has lost too many members, ReFS may have nothing consistent to mount.
Why chkdsk is not the answer
With NTFS, the reflex is to run chkdsk /f. ReFS does not work that way. It does not use chkdsk for repair in the way NTFS does, and running general-purpose “fix my disk” tools against a RAW ReFS volume will not rebuild it. Worse, some third-party utilities and the Windows format prompt will happily offer to create a brand new file system on top, which overwrites the very structures a recovery would rely on.
Microsoft does provide a command-line tool, refsutil, on recent versions of Windows and Windows Server. Its salvage function attempts to copy recoverable files from a damaged ReFS volume to a different, healthy volume. It is a legitimate option for a technically confident administrator, but it should only ever be run against healthy hardware or, better still, against an image of the disks, and it needs somewhere else with enough space to write the files out.
What to do first
- Do not format or initialise anything. Ignore any Windows prompt to format the disk, and do not create a new volume or partition in Disk Management.
- Stop writes to the pool. Pause backup jobs, virtual machines and anything else pointing at the storage.
- Write down what changed. A recent update, a migration to different hardware, a power cut, a disk warning: this information often points straight to the cause.
- Check the hardware health. Look at the Storage Spaces status in Server Manager or PowerShell, and at any disk warnings in the event logs. Do not remove, retire or “repair” disks in a degraded pool until you understand which disks hold good data.
- If it was moved to an older OS, go back. Reattaching the disks to the newer Windows version that last mounted them, in the same configuration, is often all that is needed.
- Image before you experiment. If you decide to try refsutil or any other tool, take sector-level images of every disk first and work on those.
When to stop and get help
Stop experimenting if a disk is clicking, disappearing or reporting reallocated or pending sectors, if a Storage Spaces virtual disk shows as detached or incomplete, or if the data is business-critical and there is no recent backup. Each further attempt on unhealthy hardware adds risk, and pool configuration changes can be very hard to undo.
Recovery from ReFS is specialised work because the file system layout differs considerably from NTFS, and on Storage Spaces the data is spread across several disks in slabs. The safe approach is the same as for any server: image every member disk, reconstruct the pool layout from those images, and then extract the files without writing to the originals. You can read more about our approach on our server data recovery page, or see our wider data recovery services across the UK.
Talk to Data Doctor
If an ReFS volume or Storage Spaces pool has turned RAW and the data matters, leave the disks as they are and call Data Doctor on 020 7394 2529 for a free initial consultation. Our lab is at Unit 1, The Willows, 80 Willow Walk, London SE1 5SY, with a clean room for drives that need physical work.