A striped volume goes quiet the moment any one member dies, and quiet is not the same as erased. Send the whole set. Each disk is copied first, the layout is worked out from those copies rather than assumed, and the reassembly happens entirely in software, with nothing written back to the hardware you posted.
Nothing about a stripe is assumed. Assessment costs nothing and one fixed number follows it — from £500 + VAT by member count, and logical reassembly carries no fix, no fee.
A striped volume slices every file into blocks and deals them round the members like a hand of cards, holding nothing back against a bad day. It therefore survives exactly as long as its unluckiest disk. These are the six ways that runs out.
Files live as fragments distributed over every disk in the set, so the loss of any single one of them takes the whole volume down at once. The route home is to copy the complete set and rebuild those fragments in software, well away from the box they came out of.
Nothing is duplicated and nothing is held back for later. That missing reserve is precisely what paid for the speed. A stripe has no mechanism for regrowing any part of itself, which leaves careful imaging as the only route back worth having.
Bridge boards, enclosures and controller cards fail on schedules of their own, and quite often one of them goes before any disk does. Each member is read out individually, the arrangement is derived from those reads, and the volume is raised on copies rather than on the hardware that dropped it.
Rebuilding an array from a menu, or a firmware stumble partway through the job, can erase the values that record how the stripe was arranged. Hand a reconstruction the wrong ones and the result looks like a disk and opens like rubbish, so the right ones get proved from the images beforehand.
A cluster of weak sectors in one place on one member does not ruin one file. Threaded through the interleave, it takes a fragment out of hundreds of unrelated files at once, which is why imaging hardware that steps round damage and returns to it repeatedly matters here more than almost anywhere.
Point format, initialise or repair at a stripe that has already collapsed and the records a reconstruction needs are flattened where they lie. If one of the three has run already, the productive thing to do is stop there and pull the power.
On an ordinary disk a document occupies one place. On a stripe it is divided into blocks and those blocks travel round the members in turn, so a photograph might sit in four blocks alternating between two disks while a video file sits in several thousand. Nothing on any member is complete in itself. That is why the instruction to send the entire set is a technical requirement rather than paperwork: half a stripe rebuilds into half of every file, and half of a modern compressed file is worth precisely nothing.
The same arithmetic explains the reach of a small fault. Three hundred bad sectors sitting together on one member are not three hundred bad sectors as far as you are concerned. The interleave sprays them across the address space of the finished volume, so what disappears is not a folder but a bite out of an enormous number of unrelated files. Imaging hardware that skips a bad region, comes back to it later and reads it in smaller pieces is what holds that number down.
Which member sat in which position. How large each block is. How far into every disk the data area starts. Alter any one of those and the reconstruction is wrong: the wrong order shuffles the blocks, the wrong block size puts every boundary out of place, the wrong offset slides the whole volume sideways by a fixed distance. What makes that treacherous rather than merely annoying is that a wrong reconstruction does not announce itself. It mounts. There is a partition table, folders with the correct names, files with believable sizes and dates. Open one and it is noise, or a paragraph from a document you have never seen.
All three therefore come out of the images rather than out of anybody's memory. Structures with a shape already known are found across the members, their fragments lined up against one another, and each candidate arrangement is ruled out in turn until one is left standing. Then comes the step that decides everything: real documents, photographs and video are opened off the finished volume and watched. It is slower than entering numbers into a controller and pressing go, and it is the whole distinction between your data and a convincing imitation of it.
Editing workstations, gaming builds, small servers and NAS boxes deliberately configured for throughput. They reach this bench from design and video work in the Cultural Quarter, from research machines at the university and at De Montfort where speed mattered more than redundancy, from photographers and editors across the county, and from distribution offices at Magna Park and along the M69. Every member travels together to Cambridge Data Recovery, Compass House, Vision Park, Chivers Way, Cambridge CB24 9AD, tracked and insured, or on a courier you book yourself. Parity is on RAID 5 data recovery, mirrors on RAID 1 recovery, background on RAID recovery and NAS recovery, and the figures on data recovery cost.
A great many striped sets were never near dedicated hardware. Intel RST on a desktop board, Storage Spaces under Windows, mdadm under Linux and the performance mode in a consumer NAS all produce stripes, and each records its own description in its own location — the front of the disk, the back of it, sometimes both. Carry those members into a different machine and the new board may decide they are unconfigured, offer to set them up, and then do exactly as offered.
Two habits follow from that. Write the bay number on each disk as you take it out, because the order is the single most useful thing you can pass on and the first thing that goes missing. And resist testing members by plugging them into a working computer to see which ones appear. The best result is learning something the images would have told us anyway; the worst is an operating system stamping a new signature over one that was evidence.
Most pages in this trade avoid this paragraph. RAID 0 is the everyday arrangement in which permanent loss is a realistic outcome, and pretending otherwise wastes your postage. Where a member is damaged past reading, every file whose blocks crossed it comes back with holes in it, and compressed formats do not tolerate holes. Video, camera raw and current office documents are either whole or they are nothing. No laboratory invents the missing blocks, and one that claims to is selling something.
What does survive is worth knowing about in advance. Small files that happened to fit inside the blocks on the surviving members come back complete. Beyond those, material can be carved out by signature, which means recognising the header of a JPEG or a PDF and reading forward from it; that produces a genuine haul for photographs and plain text and very little for anything large. Which of those two conversations applies to your set is established during the free diagnostic, before a figure is agreed and before any money has moved.
Cut the power to the whole set and leave it off. Nothing about a stripe improves with time, and a disk that has started to fail surrenders a little more surface every time it is asked to start. Decline anything offering to initialise, repair or rebuild. Gather every member, including one already written off, and mark each with the bay it occupied. Wrap them individually so they cannot knock against each other in transit. Then send the set as you found it, and let the reconstruction be attempted on copies, where a wrong guess costs an hour of computing rather than a volume.
Each member copied, the layout demonstrated instead of assumed, the volume rebuilt in software. Looking is free.
Very little of this work is handed over in person. It arrives as a parcel, which suits hardware that has already stopped behaving: boxed, padded and insured is a gentler seventy miles than the M1 in a glovebox. Something posted in Leicestershire this afternoon is normally booked in at Cambridge the next working day.
The general rule is the drive travels and the machine stays behind — out of the laptop, out of the tower, out of the iMac, out of the recorder under the counter. This bench does not dismantle equipment, and a repair shop will do it while you wait. Three things are the other way round, and getting them wrong costs you the recovery: an external drive stays sealed in its own case, a NAS comes as a complete unit, and a WD My Passport or My Book travels whole with its cable, because on those the encryption key is held on the bridge board rather than on the disk — separate the two and the data becomes unreadable even to us. A Fusion Mac needs both of its drives, each labelled. The one thing nobody can work round is flash soldered onto the mainboard, as on Apple Silicon machines: if it will not come off, there is nothing to post.
↓ Print the shipping & booking-in form (PDF)
Address it to Cambridge Data Recovery. It is about seventy miles from Leicester if you fancy driving it — M1 south to Junction 19, then the A14 east — and the lab is two minutes off Junction 32 with parking at the door. Posting costs you a stamp and a day instead. Whichever you choose, you hear from us the moment it is booked in, and the free diagnostic closes two working days after that.
Not certain what belongs in the box? Ring 0800 689 0668 before you tape it up, or let the free online diagnostic ask the questions for you.