Case Study — Enterprise NAS
A 16-drive QNAP array, fully recovered.
A university's 100TB QNAP TS-EC1680U-RP arrived with three failed WD Red drives and a botched rebuild on top. Here's how it was imaged, reconstructed and returned in three days.
The case
A 100TB university archive, offline mid-rebuild.
This one arrived as serious as it looked: a 16-bay enterprise NAS, roughly 100TB of RAID 5 data, and damage from an attempted rebuild already done before it reached us.
A QNAP TS-EC1680U-RP — a 16-bay, rack-mounted enterprise NAS — came in from a university client with around 100TB of storage in RAID 5 no longer accessible. At least three of its sixteen WD Red hard drives had failed, and a rebuild attempted on the live array before it reached the lab had likely caused further damage.
Cases like this leave no room to guess twice on a live 16-drive array. Every drive was imaged first, and the RAID was rebuilt from those images alone — nothing further was risked on the originals.
The failure
One batch, sixteen drives, one inevitable failure.
These drives were never going to fail one at a time.
All sixteen 6TB WD Red hard drives in the array were installed in summer 2016, rated for a three-year, 24/7 design life. Because every drive came from the same production batch, once bad media surface started appearing in 2022, further failures were only a matter of time.
Drives began dropping out of the RAID 5 array one after another. By the time the NAS reached the lab, at least three had failed outright — and a rebuild attempted on the live array had likely made the damage worse rather than better.
It's the same pattern behind most serious NAS and RAID failures we see: drives from one batch, ageing together, failing together.
In the lab
Imaging sixteen drives, then rebuilding the array virtually.
- 01
Imaging all sixteen drives
Three drives with bad sectors were imaged using PC-3000 and DDI hardware imagers built for unstable media; the other thirteen were imaged via software. All sixteen drives were fully imaged to new 6TB replacements, with every bad sector corrected, within 48 hours.
- 02
Rebuilding the RAID 5 virtually
All sixteen images were loaded onto a custom 24-bay recovery server with a high-throughput HBA controller running at over 6GB/s, and the array was reconstructed virtually using UFS Explorer Professional Recovery and Reclaim Pro. One drive, corrupted by the earlier failed rebuild, was excluded entirely — RAID 5 tolerated the loss of that single drive.
- 03
Reconstructing the file system
QNAP's underlying LVM2 volume was analysed and, unusually for a QNAP array — which defaults to EXT4 — an NTFS partition of around 82TB was found on it. The damaged NTFS file system was then repaired to restore full access to the volume.
Outcome
Full access restored in three days.
Within three days of the drives arriving, the university had full access to its data again — around 75TB recovered from the failed array. The job was quoted at £10,000 + VAT for a successful recovery of this scale and complexity.
Enterprise RAID and NAS failures like this aren't a job for trial and error. The safest outcome comes from powering the system down the moment something looks wrong, and letting a dedicated RAID data recovery lab image and rebuild every drive before anything else is attempted.
Whatever the brand, drive count or array size, the safest first step is the same: power it down and get an honest assessment before anything else is attempted.