RAID & NAS · case file · HUL-2025-1103
A Tender at Ten, and Four Disks in a Box.
Monday at ten was the deadline for a school extension tender, and the Grimsby practice had most of it written already. The QNAP vanished from the network on the Thursday. A day later there was no life in it at all: the NAS itself is dead, the drives are all right
. All four were out before they rang us, numbered by bay as we lifted them
. They wanted a fixed price and an answer to one thing: can we have it back for Monday
.
Sounds like yours? Ring us.
0800 6890668
What causes it.
Weekend work is won on scheduling. Nothing was taken out of the job. The stages were stacked on top of one another instead. Each disk had an imager to itself and all four started together, so no drive waited in a queue behind another. The array geometry was settled from whatever finished copying first while the slower members carried on reading. Nothing on the bench idles. There was an obvious corner to cut: rebuild on the practice's four original disks and skip the copies. It buys a few hours. It also stakes a tender on four drives all reading perfectly first time, and if one had gone soft partway through there would have been nothing left to go back to.
The tools this one needed.
The steps we follow →| The gear | What this one did | Why we have it |
|---|---|---|
| Atola TaskForce 2 | Took all four images at the same time — that is what saved the weekend | Takes every disk in the set together, instead of one imager working down the row |
| PC-3000 Portable PRO | Checked each disk where it sat, for health and firmware condition | Factory-level access, out of a case stood next to the imagers |
| UFS Explorer RAID Recovery | Worked out drive order, stripe size and how parity rotated, from the finished images | Puts the array back together, then unwinds the volume layers a NAS lays over it |
The stages.
Four disks, four imagers, no waiting in line
The box landed at midday. Within the hour all four were on imagers and reading, and they stayed there until the last sector was across. Put a set through one drive at a time and a three-day array turns into a week. Run them together and the weekend is still just a weekend.
Work out the geometry while copying continues
Metadata came off the images that landed first, and it produced a shortlist of layouts: which drive sat where, how wide the stripe ran, which direction parity turned. Each option was then tested against real folder trees on the volume. Nothing was accepted on the grounds that controllers usually behave one way. By the time the last copy closed, the layout was already on paper.
Build it in software, then check the tender first
Assembled in software from the four images, the volume mounted first time and the filesystem came up clean. Checking ran in the practice's order of priority, not the drive's: drawings, specification, appendices, cost plan. Each item in the tender pack was signed off by phone as it appeared, with plenty of folders still to come.
How it finished.
Verified files were with the office before Sunday was out. They filed the tender on Monday from their own drawings. That happened because the stages ran over each other, and because they got a straight answer whenever they rang to ask where it stood. Nothing was left out to make the date.
What people read after this.
Other RAID & NAS cases.
Is yours doing the same?
Turn it off. Send it to us. The diagnosis comes first. It says which files are still readable and which are not.