One box on a shelf. The company files on it, and the family photographs too. That is how most owners find out it was never a backup in the first place. Then a disk drops out. Or a rebuild halts. Or the whole thing is silent one Monday. Volumes off Synology and QNAP units, WD My Cloud, Buffalo and Netgear boxes, all get stood back up here from the bare disks. Send the disks. Keep the box.
Every nas external drive job is diagnosed free. The fixed figure reaches you in writing first, before any tools come out.
No fix, no fee all jobs except electronic and mechanical failures, chip level work, DVR and Forensic jobs. Full pricing is on the data recovery cost page.
First job on any nas external drive is matching the symptom to the fault. Twenty-odd years in, these twenty-five cover nearly every one that reaches this bench.
Designed for, survivable, and simultaneously the point at which your array stands one hiccup away from real trouble. Treat that light as a fire alarm, not as something to look at next week.
Resilver load is what kills off a tired survivor. Comfortably the commonest NAS disaster booked into this workshop, and not by a narrow margin either.
Array logic has failed. The data underneath it has not. That is reconstruction work. The restore button on the front of the box contributes nothing whatsoever.
A single reassuring click reinitialises the configuration, and on certain models it makes an early start on your data too. Good intentions all round, and it still needs stopping immediately.
Firmware expired mid-flash around an array in perfect condition. The drives neither know nor care about any of it. They come out and go into the imaging bays.
Lose power during a parity write and the members end up disagreeing, with the volume refusing to mount. Both get reconciled from complete images, where disagreement is something to examine rather than something to fear.
Box electronics fail around untouched disks far more often than disks fail on their own. Imaged members stand the volume back up virtually, and the enclosure was only ever the replaceable part.
Both carve back for as long as their blocks stay unclaimed, so time is the only variable that counts. Move quickly and touch nothing at all in the interval.
Anything reachable from the internet makes an appealing target. Attackers purge what they can see, and snapshot trees underneath frequently survive the sweep.
Pulled, shuffled, second-guessed, reinserted in a hurry. Array metadata still knows the true order even when nobody in the room does, provided the experimenting stops right now.
SHR blends unequal disks into one pool, mdadm and LVM sit beneath, btrfs or ext4 rides on top. Lose power mid-thought and that stack tears. On a DS220+ or a DS923+ we unpick it from full images, one layer at a time, in build order.
Flexible volumes collapse when a member fails mid-write and the pool loses track of its own allocation. Every disk gets imaged first. Only afterwards does the volume get stood up virtually.
Bricked single-bay units still contain a completely healthy Linux-formatted disk. Mount it natively, copy the contents out whole, and that genuinely is the job finished.
Buffalo boxes lose their array definitions periodically. The member disks never lose theirs, so the layout gets read off the drives rather than coaxed out of the enclosure.
Buy ST4000VN006 drives as one batch and they wear at very similar rates, so failure number two treads on the heels of failure number one. The day one goes, image the whole set rather than replacing a single disk.
An interrupted capacity expansion leaves the pool half stretched and refusing to mount. Reconstructed from images by retracing that expansion exactly as it was originally attempted, step for step.
WD40EFPX members take too long recovering from an error and the controller drops them for it. Put the same disk on hardware willing to wait and it images cleanly.
Locked inside an airless cabinet, the box runs warm for years and then sheds drives within weeks of one another. That clustering says image the whole set. It does not say replace one drive.
SMR members cope with daily duty and then buckle when a resilver starts, as the industry found out in public a few years back. Imaged using shingle-aware pacing before reconstruction begins.
An archive nobody ever scrubbed accumulates hidden bad sectors for years, then presents them all in a single afternoon, exactly when the redundancy has already been spent. Imaging first defuses it.
The small operating-system partitions mirrored across every drive corrupt, so the box refuses to start, while the data volume underneath is perfectly sound. Extracted from the images directly.
Put a cache device into read-write mode and it starts holding data the disks have not received yet. Lose that device, or half of a cache pair, and the volume refuses to mount, because a piece of it really is missing. We treat that cache as a full member of the set and image it accordingly.
When a NAS dies, the sensible-sounding move is to put the disks in a different one. Vendors write their arrays differently, so the new box normally offers to initialise them. Accepting is the point of no return, and the drives want a bench before that happens.
Modern NAS file systems require free space simply to record a change, so a volume filled to capacity can arrive at a state where deleting files to make room is itself impossible. Snapshots occupying that space compound it. Untangled from an image, without the risk of working live.
Folder-level encryption is offered by Synology and QNAP alike, using a key file owners are told to store somewhere safe. Somewhere safe turns out, remarkably often, to be the NAS itself. No key and no passphrase means the folder stays shut, and we say so at the diagnostic stage.
RAID, or a vendor's own dialect of it, spreads files across the disks in a NAS. Synology sells SHR, Drobo sold BeyondRAID, and none of them has as much margin as the packaging implies. Losing one disk is survivable. The hazard is what follows. A rebuild reads every surviving disk flat out for hours, which is precisely when its neighbours — same age, same production batch — decide to give up. Your disks are kept well away from that here. Each one is imaged in full before anything else happens. Order, stripe size and parity rotation are then derived from the images. The volume is constructed in software. Files come out of the construction. Not one byte is written back to your hardware.
Do not touch the rebuild button. Re-initialising is presented by the wizard as a sensible next step and it is the most damaging control on the page. What to do instead: shut the unit down properly, pull one caddy at a time, and write the bay number on each disk as it comes out. Bay order carries real weight when the array is put back together. Then post the numbered disks and hold on to the empty enclosure. A single-disk unit such as a My Cloud only needs its shell opened and the drive sent bare. If the shell will not come apart, ring rather than force it. Plastic clips lose that argument, and so does the disk behind them.
NAS recovery is RAID recovery spoken in vendor dialects, and the ones that matter are spoken fluently here:
Members get cloned simultaneously on dedicated hardware before anybody offers an opinion about layout. Every risk after that lands on copies. The originals sit still.
Stripe size, order, rotation and offset solved from the images, then the array stood up virtually. Geometry gets proved on the bench, never assumed from paperwork.
A dead member is a full mechanical job in its own right. Heads, boards, firmware, whatever it needs, until it images alongside the rest of the set.
Parity arithmetic and entropy tests have to agree with the solved layout before any file is lifted off it. Hunches do not survive that stage. That is what it is for.
SHR and X-RAID, BeyondRAID, QNAP flexible volumes, mdadm beneath LVM, plus btrfs and ext4. Proprietary layering converted back into plain named folders.
Home single-drive units opened without damage, with the Linux-formatted disk inside mounted natively and its contents copied out whole rather than in fragments.
Look in the stockrooms, studies and back offices around Aberdeen and the shire and it is two-bay DiskStations and the QNAP TS-x53 range humming away, so those head the rebuild ledger. Open one up and you generally find IronWolf ST4000VN006 or WD Red Plus WD40EFPX members purchased together on a single afternoon, which explains neatly why they tire in step and then fail in step. Drobo closing its doors left BeyondRAID owners with lab work as the sole remaining route, which is unfortunate and also true. Drive order is what owners ask about, more than any other question. Honest answer: the array metadata does still know it, but a marker pen used on every caddy before anything is pulled makes each later stage quicker and safer. Badge on the front makes no odds to the drill. Power it down, number the disks, post the disks, keep the shell.
DSM wears a friendly face on top of a very tall stack. Disks of unequal size get blended into a single pool by SHR, mdadm and LVM sit underneath that, and either btrfs or ext4 rides above the lot. So when a volume crashes on a DS220+, general-purpose software reads it as nonsense: the tool inspects one layer while the answer lies spread over three. Here the stack gets solved in build order, always working from complete member images. First the SHR arrangement is worked out, then the file system mounts virtually, then your shared folders walk out whole. A two-bay unit from a spare bedroom and a rackmount from a server cupboard both take that identical route, whatever mixture of capacities happened to be fitted.
Every time capacity gets added, X-RAID quietly reshapes a ReadyNAS volume, and the result is expansion zones stacked up like sediment. General-purpose tools interpret that stacking as damage, and they say so with great confidence. Put the zones back in the order they were created and the volume mounts precisely as the box always meant it to. Cheap single-bay units lay a simpler trap. The disk inside is formatted ext4 or XFS, Windows sees nothing it knows, pronounces the disk blank and offers a format. Turn that offer down. Your shares are still there in full, and this workshop mounts both of those file systems natively.
Shut the unit down properly. Take the disks out one at a time and write each bay number onto the disk itself, first, second, third, before any of them goes anywhere near a box. That numbering repays the effort several times over at reconstruction. Only the disks travel. Your enclosure stays with you, unless it happens to be a sealed single-bay nobody can open. Post them tracked and insured, book a courier yourself, or bring them to reception at our Edinburgh location, Edinburgh Data Recovery, 4 Redheughs Rigg, Westpoint, South Gyle, Edinburgh EH12 9DQ, open Monday to Friday, 9am to 5:30pm, two and a half hours down the A90. Collection is not a service run from here, so that leg falls to you.
Nearly everything on this bench came by tracked, insured post. It is the safest way to shift a failing drive, and a parcel handed in at a North-East post office is usually here the next working day.
Still bolted into a laptop, desktop, MacBook, iMac, server or CCTV / DVR box? Get the hard drive or SSD out first. The bare drive travels on its own. Stripping machines down is not a job this lab takes on. Flash soldered to a motherboard, as on Apple Silicon Macs and a couple of very thin laptops, is our one flat no: if it will not unbolt, it cannot be worked on.
↓ Print the shipping & booking-in form (PDF)
Address it for the attention of Edinburgh Data Recovery. From Aberdeen that is roughly 130 miles and two and a half hours down the A90, or next working day by tracked post if you would rather not make the run. You get a call the moment it is booked onto the bench.
Not sure what belongs in the box? Ring 0800 689 0668 before you tape it shut, or work through the free online diagnostic.
Free diagnosis, one number written down, no fix no fee on the bulk of jobs. Start online or phone it in.