I know this error is infuriating—you're writing to your mirrored or RAID-5 array and suddenly Windows slaps you with ERROR_FT_WRITE_RECOVERY. It's scary because it sounds like your data is toast, but it's often just a drive that tripped and needs a nudge.
Quick Fix: Reseat and Rebuild
First, try the direct approach. Shut down the PC, open the case or NAS, and reseat both drives—or all drives if it's a RAID-5 set. Cables loosen over time, and a flaky SATA or power connector causes exactly this kind of write recovery failure. While you're in there, swap SATA cables if you have spares. I've seen a worn cable produce this exact code on a workgroup server.
Boot back up. If the error persists, head to Disk Management (right-click Start → Disk Management). Look for the volume marked with a warning icon. If it says Failed Redundancy, right-click the volume and select Reactivate Disk—if that option's there, you're in luck. For a mirror, that often clears the error immediately. For RAID-5 arrays on a hardware controller, use the controller's management tool (like Intel Rapid Storage or the vendor's utility) to check the status and reactivate the disk.
Why This Works
Windows logs the write error when it can't simultaneously write to all members of a fault-tolerant volume. A drive that's momentarily offline—due to a loose connection or a transient error—triggers the fault-tolerant write recovery mechanism. Reactivating the disk forces a resync, which is basically the OS copying the good data over to the recovered drive. That's why reseating and reactivating fixes it in most cases: you're clearing the error state and letting the system rebuild parity or mirror data.
The reason the fix is often this simple is that modern drives are good at hiding their own errors. A loose cable or a failing port can cause a split-second timeout that the driver reports as a write fault. Once the connection is solid again, the array rebuilds without drama.
If the Fix Doesn't Stick
If the error reappears after a reboot or during heavy writes, the drive itself is likely failing. Run CHKDSK on the volume, but understand that CHKDSK doesn't repair hardware. Check the S.M.A.R.T. status of each physical disk. Use WMIC from an elevated command prompt:
wmic diskdrive get model,status
Any drive that shows "Pred Fail" is done. Replace it. If you're on Windows Storage Spaces, you can remove the failed drive and add a new one through the Storage Spaces control panel. For hardware RAID, use the controller BIOS or management software to mark the drive as failed and rebuild to a new one.
Less Common Variations
Sometimes you see this error after a system crash or power outage. The array is healthy, but Windows is still holding a stale write I/O. In that case, a clean reboot usually clears it—no reseating needed.
Another variation: you get the error when the volume is nearly full. Fault-tolerant writes need temporary space for recovery data. If you're under 10% free space, free some up and retry.
On older Windows Server 2003 or 2008 boxes, this error can appear if the disk cache is misconfigured. Disable write caching on the physical disks via Disk Management → Properties → Policies, and see if that stops the error. It's a band-aid, not a cure, but it works.
Prevention
Don't wait for the error to strike. Keep your SATA cables latched properly—use locking cables if you have them. And don't cheap out on cables; a $3 cable can cost you a weekend.
Monitor drive health monthly. Windows doesn't tell you when a drive is heading south, so use a tool like CrystalDiskInfo or the smartmontools package. If S.M.A.R.T. reports reallocated sectors or pending errors, back up and swap the drive before it fails.
Finally, always keep a spare drive of the same or larger size for your array. When one dies, you can rebuild immediately rather than waiting for a replacement to ship. That's the difference between a 30-minute fix and a 3-day outage.
If you've tried all this and the error won't go away, it's time to check the RAID controller itself. A dying controller can throw this error on a healthy array. Test by booting to a Linux live USB and checking if the drive reads/writes normally. If it does, the controller's your problem.