RAID & NAS · case on file · CDF-2025-1103
Four Disks, and Monday Morning at Ten.
The tender for a school extension had to be in by ten o'clock on Monday morning, and the Newport practice had already worked most of it out themselves. Their QNAP vanished off the network on the Thursday, and by Friday there was no power at all — the enclosure is what has failed, not the disks
. All four had been taken out and labelled by bay number before anyone touched them
. They wanted a fixed price and a plain answer to a single question: will it be ready for Monday
.
Does that match yours? Call us.
0800 6890668
Why that happens.
Fitting a job like this into a weekend honestly leaves one option: overlap the stages, never remove one. All four disks go onto imagers together, none of them queuing behind another, and the geometry of the array gets settled using whichever copies complete earliest while the remainder keep reading — so no stage ever sits idle. What we declined to do was assemble the array directly from the practice's own disks. It would have saved a few hours, and it would have staked the whole tender on four faultless reads; if one member had faltered part way in, nothing would have existed underneath it.
What we used on this one.
Every step, in order →| The tools | What it did here | Why it helps |
|---|---|---|
| Atola TaskForce 2 | Imaged all four disks simultaneously, which is what rescued the weekend | Images every member of the set at once, instead of working through them one by one |
| PC-3000 Portable PRO | Interrogated every disk in place for health and firmware state | Manufacturer-level access from a case sitting beside the imagers |
| UFS Explorer RAID Recovery | Established disk order, stripe width and parity rotation from the completed images | Not just the RAID: it also unpicks the volume layers a NAS stacks on top |
How it went.
Four disks on four imagers, nothing queuing
The courier arrived at midday, and inside the hour all four disks were on imagers and reading, where they stayed until the final sector had crossed. Feeding a set through one disk at a time is exactly how a three-day array becomes a week-long one. Handle them together and Friday to Monday remains Friday to Monday.
Prove the array layout while the copying runs on
Metadata came off whichever images completed earliest, and out of it fell the candidate layouts: the position of each disk, the width of the stripe, the direction parity turned. Each candidate then had to hold up against real directory structures sitting on the volume — none was waved through merely because most controllers behave a certain way. The fourth copy finished with the assembly already written down.
Assemble it in software, then verify the tender first
Software assembly across all four images gave a volume whose filesystem came up on the first mount. Verification then ran to the practice's list of priorities instead of the disk's own order: everything in the tender pack — the drawings, the specification, the appendices, the cost plan — was confirmed good on the phone while the other folders were still going through.
Signing it off.
Checked files reached the office before Sunday ended. On the Monday the tender went in, drawn from their own files. What made that possible was stages running over each other, and a plain account at every point of where the job stood — not the removal of a single one.
Pages people often read next.
The RAID & NAS jobs.
Is that what yours is doing?
Power it down, get it to us, and hold off on any decision until the diagnosis says what can still be read.