A pair that wore out on one timetable. A resync that copied the stale disk over the live one. An enclosure that died with two perfectly good members sitting inside it. None of those is the end of the story. Each half is copied by itself, the two are compared on contents rather than on a status flag, and your files are lifted off whichever turns out to be the better source.
Nothing about a mirror is estimated. The examination is free and the written number follows it — from £500 + VAT by disk count, and logical work carries no fix, no fee.
A mirror makes a single undertaking and honours it well: the volume keeps working when one disk dies. Everything below goes around that undertaking rather than through it, which is how mirrored pairs end up in a padded box heading for the A14.
Two disks off one order, switched on together and given identical duty every day since, do not fail politely years apart. Weeks is the usual gap. Whatever finished the first one has been at work on its twin the entire time, at the same temperature, in the same cupboard.
In a two-bay unit the part that gives out first is usually the bridge or the controller rather than either disk. What that leaves is two serviceable members trapped in a case with nothing left to talk to, and the only way onwards is to read each of them directly, outside the box.
Return the wrong disk to its bay, or leave the box to reach its own conclusion overnight, and the stale or blank member gets copied over the current one. The mechanism performs perfectly. It simply performs in the wrong direction, at full speed, without asking.
Where a pair has run degraded at separate times, each disk finishes up with a different edition of the same folders. Merged carelessly they spoil each other. Copied apart, each is free to surrender the material only it still carries.
A deletion, a bad save, a botched update, an encryption run. Mirroring exists to reproduce writes the instant they occur, so all four reach the second disk at the same moment as the first and neither member is spared any of it.
An update that fails halfway through, or an embedded system that eats its own boot volume, leaves a dead unit with two working disks trapped inside. The route out is to copy each disk and raise the volume on hardware that still functions.
Mirroring works by writing everything twice, simultaneously, to two disks meant to stay identical down to the sector. The operating system is not told a write has completed until both members have accepted it. That is the entire mechanism, and it is worth noticing what the mechanism is incapable of considering: nothing in it asks whether the write was a good idea.
Emptying a folder is a write. So is an encryption run, a corrupted save, a database shutting down badly, an application update going wrong. Each one is copied obediently onto the second disk in the same instant, at bus speed, with no interval in which anybody could intervene. The catalogue of disasters a mirror genuinely prevents has exactly one entry: one member failing while the other keeps serving.
Held up against a backup the difference states itself. A backup is yesterday, parked somewhere other than today. A mirror is today written twice, inside one chassis, on one power supply, behind one controller. Owning both is sensible and only one of the two gives you a route back. To be clear about what is on offer here, this site sells no backup product of any kind: nothing to sign, nothing billed monthly. It is a bench that retrieves data after the event, and whatever it has to say about where copies should live costs nothing.
Each member carries a small housekeeping record — a counter that steps forward as events occur, a timestamp, a flag saying whether that disk is currently in step with its partner. When the set assembles, the two records are compared and whichever disk shows the higher counter is treated as correct. Its partner is then brought into line, which in a mirror means overwritten.
Nine times in ten that is the right behaviour. A member pulled out for a fortnight really is behind, and copying the live disk onto it is precisely what ought to happen. Trouble starts when the records are absent, damaged or misread, which is common after a controller has failed or an update has gone badly. The comparison still runs, still produces an answer and still acts on it at full speed. Only now the answer is wrong, and the disk being trusted is the stale one.
A recovery repeats none of that arithmetic. Both members are imaged first and the leading copy is then chosen on evidence: which image carries a file system that holds together, which holds the version of a document whose date you recognise, which shows the characteristic pattern of having been overwritten from the outside in. A counter can be read in a second and can be wrong. Reading the disks takes considerably longer and is much harder to fool.
Two-bay Synology, QNAP, Netgear and Buffalo units account for most of them, with small towers under desks making up the rest. Consultancies and practices around Oadby and Wigston, surveyors and architects in Market Harborough, food producers at Melton Mowbray holding traceability records, engineering firms up towards Loughborough and out along the M69 past Hinckley. Both disks go into one parcel bound for Cambridge Data Recovery, Compass House, Vision Park, Chivers Way, Cambridge CB24 9AD, tracked and insured. Parity is covered on RAID 5 data recovery, striping on RAID 0 recovery, the general picture on RAID recovery and NAS recovery, the county on data recovery Leicestershire, and every band under data recovery cost.
A mirror may satisfy any read from either member, which sounds harmless and conceals something slow. Sectors weaken gradually rather than all at once. A box with no scrubbing routine has no reason to visit a quiet corner of a disk for months together, so one member can be rotting away while the front panel reports two green lights. By the time symptoms surface they tend to be strange rather than dramatic: a file that opens on Monday and refuses on Thursday, a checksum that passes and then does not, software complaining about data it produced itself. If that is what brought you here, both disks matter more than usual, because the sound copy of any particular file may exist on only one of them.
The pair is split the moment the box is opened and handled as two independent jobs. Each member goes onto a hardware imager, and any disk with a physical fault is stabilised in the clean-air enclosure before that begins. The two images are then compared against one another for content: which holds more of your material, and which holds the later edition of it. The winner becomes the base for the rebuild and the other is used to fill what the winner lacks. Neither original is written to at any point.
Synology wraps its own volume manager around the standard Linux arrangement and lays a modern file system on top. QNAP does something comparable under other names. Netgear's X-RAID adds expansion logic of its own, and Buffalo has variations again. All of them are readable and none of them is readable by pushing a member into a Windows machine and hoping. What that means at your end is small: send both disks rather than one, write down which bay each came out of, and tell us the make and model of the unit.
A backwards resync takes a fraction more with every write it lands, so the gap between noticing and pulling the plug is the most valuable minute anywhere in this page. Stop it and stop it immediately. After that, the second rule is to leave the disks entirely alone. Trying each member in a desktop machine to see which one works gains you almost nothing: at best you learn something the images would have shown anyway, and at worst Windows offers to initialise the disk and somebody in a hurry clicks yes. That one word turns a routine job into an awkward one.
Two images taken separately, the stronger one used as the base, the weaker one closing the gaps. 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.