A single dead member takes a whole RAID 0 quiet, and quiet is not at all the same thing as erased. Send the complete set: each disk is copied, the geometry of the stripe is proved from those copies rather than assumed, and your files are put back together entirely in software.
A stripe is quoted, never guessed at. Diagnosis costs nothing and one written figure follows it — from £500 + VAT counted by disk, logical recoveries keeping no fix, no fee.
Files in a stripe are cut into pieces and shared across the whole set, with nothing held back against a bad day. The volume therefore lasts exactly as long as its unluckiest disk, and these are the six ways that runs out.
Since every file is distributed in fragments across all the disks, losing any single one of them silences everything at once. The route home is to copy the complete set and put those fragments back together in software.
Nothing is duplicated and nothing is calculated in reserve. That absent safety net is precisely what paid for the speed. A stripe cannot regrow any part of itself, so patient imaging is the only route back worth trusting.
Controllers, bridge boards and enclosures fail on their own account and regularly strand a perfectly healthy set behind them. Each disk is read out, the way the stripe was laid down is worked out, and the volume is raised from the copies.
Rebuilding the array from a menu, or a firmware stumble part way through, can wipe the values that record how the stripe was laid out. Hand a reconstruction the wrong ones and what comes back is convincing noise, so the right ones get proved from the images first.
Interleaving means one cluster of failing sectors on a single member takes a bite out of hundreds of unrelated files at once. Imaging hardware works around the damaged region and collects every sector that still reads.
Aim any of format, initialise or repair at a collapsed stripe and the structures a reconstruction reads get flattened underneath it. Where one of them has already been run, the useful response is to stop there and take the power off.
Nothing in a stripe is duplicated, which means every member holds material that exists nowhere else. Each is copied on a hardware imager, with any physically failed disk repaired at the clean bench first, and the copies then supply the facts a reconstruction needs: which disk sat in which position, how large each block was, and where the striping starts. From there the volume is put together in software, a file system is raised across it, and your files come off that reconstruction. Not one original is written to along the way.
A striped set never repairs itself, and a disk that has started to fail surrenders a little more each time it is asked to spin up. The correct response is therefore a dull one. Turn the whole lot off. Keep away from anything on a menu called initialise or repair, both of which write across the records a reconstruction has to read. Then get the set here in the state you found it.
Editing machines, gaming builds, small servers and NAS units configured for speed rather than safety: striped volumes reach this bench from studios in the Baltic Triangle, from media and design firms near the waterfront, from the university departments at LJMU and from households across the city. Every member travels together to Manchester Data Recovery, Peter House, Oxford Street, Manchester M1 5AN, by tracked insured post or a courier you book yourself. There is more background on RAID recovery and NAS recovery, and every figure is on the data recovery cost page.
Write a document to an ordinary disk and it goes down in one place. Write it to a stripe and it is chopped into blocks, and those blocks are dealt out across the members in rotation like cards. A modest photograph might occupy four blocks on a two-disk set, alternating between them. A video file occupies thousands. The consequence is that no member holds a complete anything, and it is why the advice to send every disk is not a formality: half a stripe reassembles into half of every file, which for most formats is the same as nothing.
It also explains the disproportionate damage a small fault does. A few hundred failing sectors sitting in one place on one member are not a few hundred failing sectors as far as your data is concerned. Because of the interleave, those sectors are scattered through the address space of the reconstructed volume, so the damage lands as a small bite taken out of a very large number of separate files rather than the loss of one folder. Imaging hardware that can read around a bad region, and go back for repeated attempts at the worst of it, is what limits that.
To describe a stripe you need the member order, the block size and the starting offset. Get the order wrong and the blocks come back shuffled. Get the block size wrong and every boundary lands in the wrong place. Get the offset wrong and the whole volume is displaced by a fixed amount. The unpleasant part is that none of those failures announces itself. A wrong reconstruction still produces something that looks like a disk: a partition table appears, directories list, files have names and sizes. They simply do not open, or they open filled with fragments of other people's documents.
So the parameters are established rather than assumed. Known structures are located across the images, their fragments matched, and the candidate descriptions narrowed until only one can be correct. Verification then means opening real files off the reconstruction and confirming they work. It is slower than typing values into a controller and pressing go, and it is the difference between getting your data and getting a volume that merely resembles one.
A great many of these sets never went near a proper controller. Intel RST on a desktop motherboard, Windows Storage Spaces, mdadm on Linux and the striping mode inside consumer NAS boxes all produce striped volumes, and each writes its own metadata in its own place. Some of it lives at the front of each member, some at the end. Move the disks into another machine and the new motherboard may claim the set, offer to initialise it, and mean it.
Which leads to the two habits worth having. Label the disks as they come out, because bay order is the single most useful thing you can hand over and the most easily lost. And never test a member by plugging it into a working computer to see what happens: at best you learn nothing, at worst an operating system helpfully writes a fresh signature onto a disk whose signature was evidence.
It has to be said honestly, because a stripe is the one layout where a total loss is possible. Where a member is damaged past reading, everything crossing the missing blocks is incomplete, and for compressed formats such as video or modern office documents incomplete usually means unusable. What still comes back in that situation is whatever lived entirely within the surviving blocks, plus material that can be carved out of them by signature, which for photographs and plain documents is often a worthwhile haul. The free diagnostic is what tells you which of those two conversations you are having, and it tells you before any money is committed.
Every member copied, the geometry proved rather than guessed, the files handed back. Diagnosis free.
Nearly everything on this bench arrived by tracked, insured post. It is the calmest way to move a failing drive, and a parcel handed in anywhere on Merseyside is normally on the bench the next working day.
Still screwed inside a laptop, tower, MacBook, iMac, server or a CCTV recorder? Take the drive out and post that on its own. We do not strip machines here, and any computer shop will do it in a few minutes. The one job nobody can take on is flash soldered straight to a logic board, as on Apple Silicon Macs and a handful of very slim laptops: if the storage will not unbolt, there is nothing to send.
↓ Print the shipping & booking-in form (PDF)
Address it to Manchester Data Recovery. It is roughly 35 miles from Liverpool along the M62 if you fancy the run, and next working day by tracked post if you do not. We ring or email the moment it is booked in, and the free diagnostic closes 2 working days after that.
Not sure what belongs in the parcel? Call 0800 689 0668 before you tape it shut, or step through the free online diagnostic first.