A storage migration that created a new volume on the wrong unit. The array reported itself as fresh and empty within seconds — and every byte of the original data was still on the disks, addressed by a map that had just been replaced.
← All case files · from £500 + VAT
An architectural practice was migrating from an older NAS to a new twelve-bay Synology. During setup, the create-volume operation was run against the wrong unit — the populated one rather than the empty one.
The array reported itself as new and empty. On it were project archives, drawing sets and rendering assets going back nine years, with a backup that covered the current year only.
creating a volume lays down fresh array and file-system structures. It does not erase twelve disks — that would take many hours rather than the moments it actually took.
mirrored pairs then striped, with no parity to compute, so establishing pair membership and stripe order was the whole problem.
a small amount of test data had been copied to the new volume before the mistake was noticed, landing at the beginning of the address space.
RAID 10 helped here. With mirrored pairs there are two physical copies of every block, so where new metadata had overwritten one member of a pair, the other frequently still held the original.
All twelve disks were imaged read-only before anything else. Working from the images, the original layout was derived by searching for the previous array’s structures beneath the newly written ones — stripe boundaries located by structural repetition, and pair membership established by finding disks whose content matched block for block.
The original RAID 10 was then assembled virtually. Where a stripe had been touched by the new metadata on one member, its mirror partner supplied the original content, and the Btrfs file system was rebuilt from the assembled volume.
About 95%, including the complete project archive for all nine years. The shortfall was concentrated where the test writes had landed and where both members of a pair had been touched by the new metadata in the same region.
Drawing files were opened in their application rather than counted, and the folder structure matched against the practice’s own project index before return.
Initialising an array is alarming and it is not erasing. Creating a volume writes new metadata over the old, which takes seconds; genuinely wiping twelve disks would take most of a day. The data is still there until something overwrites it.
Which is why the response matters more than the mistake. Stop immediately, write nothing further to the new volume, and power the unit down — every file copied onto it afterwards lands on top of something. NAS and RAID recovery starts from £500 +VAT after a free 48-hour diagnostic.
Drop the drive at our Oxford Science Park reception, or post it to us — it costs nothing to find out what happened. You get a written figure from the fixed bands before any work begins.