When one member of the set dies, the whole striped volume shuts down and goes quiet, however going quiet is a very different thing from being erased. Get a quote first, then send all of the discs, labelled by their bays, and each one will be copied so that the shape of the stripe can be worked out from those copies instead of being guessed at, with the files then reassembled entirely in software.
Nobody guesses at a stripe. The diagnostic is free and one fixed figure follows it, from £500 + VAT calculated by disc, and on logical rebuilds, in most cases you only pay if we recover your data.
Files that are placed onto a striped volume are divided into blocks and then dealt out across the individual member discs, a bit like cards, without anything being held back in reserve for a bad day. The result is that the volume only lasts as long as its least fortunate member disc, and the six ways that this tends to happen are listed below.
The reason is that every file was broken into fragments and spread across all of the discs, so when you lose one disc you lose the whole lot at once. To get back to your data, the complete set needs to be copied and then the fragmented pieces rebuilt using software, as opposed to inside the box itself.
Nothing is duplicated and nothing is held back in reserve, and that missing safety net is exactly what paid for the speed in the first place. A stripe cannot rebuild or regrow any piece of itself, and this is the main reason why patient imaging is the only route back that is really worth trusting.
Every one of the other components in the box (the enclosure, the bridge board, the controller and so on) has its own failure rate as well, and more often than not it is one of these that fails first. The individual members are read out separately, the layout is figured out from reading those members, and then the volume is raised on the copies instead of on the original discs.
When the array is rebuilt from a menu, or the firmware stumbles at some point partway through, the values that record how the stripe was laid out can be erased. If a reconstruction is fed the wrong values, what comes back will look like a disc but will open like rubbish, which is why the correct values are proved from the images first, before anything is rebuilt.
If there is a cluster of weak sectors on one of the members, the loss of data is not limited to just one file, because the way the data is interleaved means the damage takes a fragment out of hundreds of unrelated files at once. That is why it matters so much here to use imaging hardware that reads around the damaged area and then comes back to it again and again.
Point format, initialise or repair at a collapsed stripe and the structures a reconstruction depends on are flattened underneath it. If one of these has already run, the useful thing to do is to stop there and cut the power.
If you save a document to an ordinary disc, it is all stored in one place. If you save it to a stripe, it is cut into blocks, and those blocks are dealt out across the members in turn. An average photograph might take four blocks on a two-disc set, alternating between the pair, while a video file takes thousands. As a result, no single member holds a complete anything, and that is why the instruction to send every disc is not a formality. Half a stripe reassembles into half of every file, and for most modern formats half a file is the same as none of it.
The same arithmetic also explains why even a small fault in a single area can cause so much damage. Three hundred failing sectors sitting together in one place on one member do not stay together from your point of view, because the interleaving spreads them across the entire address space of the rebuilt volume. So rather than losing an entire folder, you end up losing bites out of a huge number of unrelated files, and that is where imaging hardware comes in, stepping around bad areas, coming back to those regions later and reading them in smaller pieces, which is what keeps that number down.
A stripe is described by three values: which member sat in which position, how large each block is, and how far into each disc the data begins, and changing any one of these will produce a reconstruction that is wrong. If you have the wrong order, the blocks are put together in the wrong sequence, if you have the wrong block size, all of the boundaries land in the wrong place, and if you have the wrong offset, then the entire volume is shifted sideways by a fixed amount. This situation can be very serious rather than just annoying, because a wrong reconstruction does not necessarily look wrong at all: it mounts, you see a partition table, you see folders with the right names and files with reasonable-looking sizes and dates, but when you open one of those files it turns out to be random noise, or the middle of some other person's spreadsheet.
All three of the values are worked out from evidence and then verified. Structures whose layouts are already known are located across the images, their fragments are lined up and compared against each other, and the possible descriptions are eliminated until there is only one candidate remaining. At the end, real documents, photos and videos are opened from the finished reconstruction to show that they actually work, and while this process takes much longer than simply entering some numbers into a controller and pressing go, it represents the whole difference between your data and a volume that only resembles it.
Editing workstations, gaming builds, small servers and NAS boxes set up for speed: stripes reach this bench from post-production and design firms along the Thames Valley, from software houses on the M4 corridor, from photographers and video editors working out of Windsor, Henley and Ascot, and by courier from every district on the London page. Once you have your quote, either by calling us or using our free online diagnostic, every member, labelled by bay, travels together to Guildford Data Recovery, Building 2, Ground Floor, Guildford Business Park, Guildford GU2 8XH, by tracked and insured post or on a courier you book yourself, or you can bring the set directly into the lab, Monday to Friday, 9am to 5:30pm, without an appointment. The background is on RAID recovery and NAS recovery, and the figures are listed on the data recovery cost page.
Plenty of striped sets were never connected to a dedicated controller at all. Intel RST on a desktop board, Storage Spaces on Windows, mdadm under Linux and the fast mode in a consumer NAS all make stripes, and each of them writes its own description somewhere of its own, usually at the front of the disc, at the back, or sometimes in both places. Move those members into a different machine and the new motherboard may identify them as unconfigured, offer to set them up for you, and then carry that offer out to the letter.
Two habits follow from this. Write down the bay number on each disc as it comes out, because the order is the most valuable single thing you can pass along, and it's also the first thing likely to get lost. And don't try to test the members by plugging them into a working computer to see which of them shows up: the best possible outcome is learning something the images would have shown anyway, and the worst is an operating system writing a fresh signature over one that was evidence.
RAID 0 deserves a straight answer here, since it is the one everyday layout in which real loss is possible. Where a member is damaged beyond reading, every file that crosses its missing blocks is incomplete, and for compressed formats such as video, camera raw and current office documents, incomplete usually means unusable. Nothing can invent the missing pieces. What does still come back is whatever happened to sit entirely inside the surviving blocks, together with material carved out by file signature, which for photographs and plain text is often a worthwhile haul. Which of those two conversations you are in is settled by the free diagnostic, before any money has changed hands.
Switch the whole set off and leave it off. A striped volume never mends itself, and any disc which has begun to fail gives up a little bit more each time it is asked to spin up. Avoid anything on a menu labelled initialise or repair. Collect every member (including any already written off) and mark the bays. Then get a quote over the phone or with our free online diagnostic, pack them so that they cannot knock together in the post, and send the set exactly as you found it by tracked, insured post, or bring it into the Guildford lab, Monday to Friday, 9am to 5:30pm, without needing an appointment, and let the reconstruction happen against copies, where a wrong assumption costs an hour rather than a whole volume.
All members will be copied, the geometry is confirmed rather than assumed and all files will be restored, and there is no charge for the diagnosis.
Hardly anyone comes walking through the door here, as almost everything arrives in a padded box. Once you have a quote from us by telephone (or from our free online diagnostic), the least stressful way to move storage that has already given up is a fully tracked and insured parcel, and one handed over at a post office in Bracknell before the last collection will usually be delivered to Guildford the next working day.
Is the hard drive still attached inside a machine, whether that is a laptop, a tower, an iMac, a MacBook, a rack server or a DVR under the till? After getting a quote, take the drive out of the machine and send us only the bare drive (never the whole machine): from a server the disks come out labelled by bay, and from a DVR only the disk needs sending. Our lab does not strip down hardware, but most repair shops on the high street can take a drive out in less than ten minutes. There is one exception where we will always have to say no, which is memory chips soldered directly onto the mainboard, as in T2 and Apple Silicon Macs and certain ultrabooks, and that storage is not taken on. If you are unsure whether you have removable storage, please call us before sending anything, because where the storage cannot be unbolted there is no parcel to make up.
↓ Print the shipping & booking-in form (PDF)
Once you have your quote, the name to put on the parcel is Guildford Data Recovery. It will take you around forty minutes to drive to Guildford from Bracknell via the A322 and then the A3, and there is no requirement for an appointment, whereas posting the parcel to us will cost you a stamp and one working day. In either case, we will contact you as soon as the parcel has been logged onto our system, and two working days later the diagnostic will have been completed.
If you are unsure about what items should be included in the box, please call 0800 689 0668 while the box is still open, or alternatively use the free online diagnostic, which will guide you.