We had a RAID failure that went unnoticed for too long on a physical Windows Server 2016 machine. We replaced the failed drives, and the Storage Space is online and browsable again. We can create new files and folders, but existing data cannot be copied or moved to local storage, SMB shares, or cloud storage while we prepare to decommission the server.
Windows Explorer appears to start a copy, but the paste window never opens. TeraCopy reports that files are offline and repeatedly retries, while Robocopy shows similar failures. Reboots and updates have not helped.
The event logs contain errors such as:
- The device \Device\Harddisk35\DR35 has a bad block.
- An error was detected on device \Device\Harddisk45\DR45 during a paging operation.
Storage Spaces reports two physical disks as having Lost Communication, Warning health, and Retired usage. Attempts to remove those disks from the pool either fail with "The requested object could not be found" or freeze indefinitely. What is the safest way to recover or evacuate the data, and should the missing disks be removed from the pool?
3 Answers
Those errors strongly suggest the Storage Space is still missing data or has unreadable extents, even though the volume can be mounted and browsed. First map Harddisk35 and Harddisk45 back to the actual physical disks using Get-Disk, Get-PhysicalDisk, serial numbers, enclosure information, and the storage controller’s inventory. Then check the storage pool and virtual disk health with Get-StoragePool, Get-VirtualDisk, and Get-PhysicalDisk.
Do not remove disks based only on their Windows disk numbers, and do not repeatedly force removal while the command is hanging. If the pool is degraded or data is offline, removing additional members can make recovery impossible. Stop unnecessary writes, make sure you have another copy of anything recoverable, and involve the storage vendor or a data-recovery specialist if the data is important.
Before trying more copy tools, test a few small, known files and record exactly which paths fail. You can use Robocopy with logging and restartable mode to identify readable files, but it cannot recover files whose underlying Storage Spaces extents are unavailable. Avoid running CHKDSK or initializing, formatting, or removing pool members until you have a backup or a recovery plan, since repair operations may alter metadata and reduce the chances of recovery.
A browsable volume does not mean the data is healthy. Storage Spaces can still create new metadata and folders while reads of older files fail because the required mirror/parity copies are unavailable. The “file is offline” behavior is consistent with missing or inaccessible blocks rather than an Explorer or Robocopy problem.
Check the virtual disk’s resiliency and repair status, and look for repair or read errors in the Storage Spaces and system event logs. If the virtual disk is repairable, replace the genuinely failed hardware and allow the repair to complete before attempting a large copy. If it is not repairable, prioritize a sector-aware recovery or imaging approach instead of normal file copies.
The missing disks both show Lost Communication and Retired. Removing them from the pool either returns an object-not-found error or hangs, so I’m going to stop trying that and inventory the physical disks and pool state first.