Case notes · NAS & RAID · EDR-2025-3103
The QNAP Died Overnight, the Tender Was Due Monday.
The practice rang with its triage already done: the QNAP's power supply died overnight, the unit's dead but the disks look fine
, all four members out and numbered one to four as they came out
. The tender was due Monday morning. They wanted two things: a fixed price, and an honest is Monday actually doable
.
Sound familiar? Talk to us.
0800 6890668
In plain English.
The clock changes the order of work, never the amount of it. Skipping the copies and reconstructing on the customer's own disks would have bought a few hours back and staked the whole tender on nothing going wrong — one member faltering mid-read and there is no second chance. The honest route to speed is parallel imaging: all four members copying at once rather than queueing, with the array's geometry being worked out from the earliest finished copies while the later ones are still filling. Nothing waits, and nothing gets left out.
The tools this job took.
See how your case runs →| System | The part it played | Why we rate it |
|---|---|---|
| Atola TaskForce 2 | Ran all four members side by side — where most of the weekend was won | Clones multiple array members at once, shrinking a week's imaging into days |
| PC-3000 Portable PRO | Vendor-level checks on each member without a trip to the main bench | Full firmware access from a unit small enough to sit by the imaging station |
| UFS Explorer RAID Recovery | Proved the disk order, stripe size and parity pattern against the images | Reads NAS volume managers as the layered systems they are, not one flat array |
On the bench.
Four disks copying at once, no queue
All four members were mounted and copying within the hour of arrival, and stayed side by side to the end. Queueing disks one after another is how a three-day array job becomes a week; running the whole set together is what made Friday-to-Monday possible rather than apologetic.
Solve the layout before the last copy finishes
Geometry came out of the metadata on the first images home: candidate values for disk order, stripe size and the parity's rotation, each proved against real file structures rather than taken on trust from the controller's habits. When the final copy landed, the assembly recipe was already written.
Assemble in software, and verify the tender folder first
The volume went together in software over the four images and its filesystem opened first time. Checking ran in the practice's order of need: the tender set — drawings, the cost plan, the appendices — was verified and phoned through as safe while the rest of the volume was still working through validation.
How it ended.
Verified data was back with the practice before the weekend closed, and the tender went in on Monday from their own files. The speed came from running stages together and reporting where things stood after each one — never from leaving a stage out.
More like this in the casebook.
More from RAID & NAS.
Sound like your drive?
Every case here points the same way: power the device down, take the free diagnosis, then make your decision.