Microsoft is starting to push ReFS as a "development drive" in Windows 11.
However, I am not liking losing the super-fast searching for files via directly reading the MFT on an NTFS volume.
Is anyone becoming reliant on the new features that ReFS is supposed to bring?
ALSO:
I asked Claude this question, and I'm wondering if it is accurate, or just trying to please me and tell me what I want to hear:
-----------
Using a ReFS volume with Everything 1.5
4:55 PM
Examined ReFS compatibility and Everything 1.5 indexing capabilities
Everything 1.5 has better ReFS support than 1.4, but there are still some important caveats to know:
What works
Everything 1.5 can index ReFS volumes using the USN journal, similar to NTFS. ReFS does implement a USN journal, so real-time change monitoring should function.
Basic file/folder indexing works fine for most ReFS configurations.
Known limitations / gotchas
No MFT direct access — Everything's fast initial scan on NTFS works by reading the MFT directly. ReFS doesn't have an MFT, so Everything falls back to a slower folder-crawl method for the initial index build on ReFS volumes.
USN journal gaps — If the USN journal wraps or is unavailable (e.g., after a dirty shutdown or remount), Everything may need to do a full re-crawl rather than a fast catch-up scan.
ReFS on Storage Spaces / refs dedup — If your ReFS volume uses block cloning or deduplication, there are no known issues with indexing, but those features are transparent to Everything anyway.
Volume shadow copies / snapshots — ReFS on Windows Server supports VSS-backed snapshots differently than NTFS; Everything doesn't index shadow copies on either filesystem, so no change there.
Practical setup tips
In Everything's volume properties (right-click the drive → Properties), confirm it's set to "NTFS" or "Everything" indexing method — if it auto-detected wrong, manually set it to folder monitor + USN journal.
You can force a full re-index with Tools → Rebuild database if the index seems stale after first setup.
Watch the Everything log (Tools → Options → General → Log or the log window) for any journal errors on the ReFS volume.
Bottom line: For a typical ReFS data drive, Everything 1.5 works well enough — you just pay a one-time cost on initial index build (slower crawl vs. NTFS MFT scan), then real-time monitoring via USN journal keeps it current. Is there a specific behavior you're running into?
-----------
Looking for opinions about ReFS
Re: Looking for opinions about ReFS
This is just my opinion..
Avoid ReFS for spinning disks.
It's minor, but File and Directory lookup is located all over the place (many seeks) and has to read a lot.
Use ReFS on SSDs only.
To me ReFS is designed for SSDs, not HDDs, eg: copy on write.
ReFS has bit rot detection.
For NTFS use RAID scrubbing or something like voidhash.
The USN Journal API for Dev Drives is currently broken.
I'm using a workaround to get it going in Everything 1.5.
Avoid ReFS for spinning disks.
It's minor, but File and Directory lookup is located all over the place (many seeks) and has to read a lot.
Use ReFS on SSDs only.
To me ReFS is designed for SSDs, not HDDs, eg: copy on write.
ReFS has bit rot detection.
For NTFS use RAID scrubbing or something like voidhash.
The USN Journal API for Dev Drives is currently broken.
I'm using a workaround to get it going in Everything 1.5.