That box humming on a shelf was holding the accounts and the family photographs at the same time, which is usually how an owner learns it was never a backup at all. Then a bay went red, or a rebuild stopped halfway, or the fans were silent one morning. Synology, QNAP, Buffalo, Netgear and WD My Cloud volumes get put back together here out of the member disks alone. Send the disks. Keep the box.
Every nas external drive job is diagnosed free. One fixed figure follows in writing, agreed before a screwdriver comes out of the drawer.
No fix, no fee all jobs except electronic and mechanical failures, chip level work, DVR and Forensic jobs. Every band is listed on the data recovery cost page.
Matching the symptom to the fault underneath is the first job on any nas external drive, and after twenty-odd years these thirty account for very nearly everything that arrives.
Redundancy has done exactly what it was bought to do. It also means the set is now a single further failure away from real trouble. Treat that light the way you would treat a smoke alarm, not the way you would treat a reminder.
A rebuild reads every sector on every surviving disk, which is exactly the examination a tired one fails. More NAS jobs arrive here through that door than through any other, and the gap is wide.
The array logic has failed while the data underneath it has not. This is reconstruction work. The reset button on the front of the box has nothing whatever to contribute.
One reassuring click resets the configuration, and on some models it starts in on the data as well. The intention was good. It still has to be interrupted.
The flash failed halfway and the unit has not come up since. The array behind it is in perfect health and entirely unaware of the drama. The disks come out and go straight onto imaging bays.
Lose power in the middle of a parity write and the disks end up disagreeing about the same stripe. Nothing will mount while that disagreement stands. Reconciling it from full images costs nothing and risks nothing.
Enclosure electronics give up more often than the disks inside them do. Once the members are imaged, the volume stands up virtually with no box involved at all. The enclosure was always the part you could simply buy again.
Deleted shares and deleted snapshots both come back while nothing has reused their blocks. Time is the only variable that counts here. Move fast and leave the box switched off in the meantime.
A box reachable from the internet makes an appealing target, and an attacker purges whatever can be reached. The snapshot trees sitting underneath frequently escape the sweep altogether.
Pulled out, shuffled, second-guessed, pushed back in. Nobody can now remember which disk came from which bay. The array metadata can, and it will keep that answer safe provided the experimenting stops right now.
DSM stacks storage several layers deep, and all that cleverness tears when the power disappears mid-thought. Full images come first. The layers are then unpicked one at a time, in the order somebody originally built them.
The two-bay and four-bay DiskStation is what homes and small firms across Merseyside genuinely run, so crashed pools on those sit at the very top of the rebuild ledger here.
A thin or flexible volume gives way the moment a member fails during a write. Every disk gets imaged first. Only after that is the volume stood back up virtually.
A bricked single-bay unit still contains a perfectly healthy Linux-formatted disk. Mounted natively, copied out, finished. That really is the whole job on those.
Buffalo boxes mislay their array definitions from time to time. The member disks never do, and the layout gets taken back off them.
Swapped during a clean, or during a panic. The metadata is the witness that establishes which disk belongs where, although a number marked on each caddy beforehand would have spared everybody a good deal of suspense.
A set of ST4000VN006 drives bought together wears at almost identical rates, so the second failure follows hard on the first. The day one of them stops, image every disk in the box.
An interrupted growth leaves the pool half stretched and refusing to mount. Rebuilding it means retracing that expansion step by step, in precisely the order it was attempted, and doing so on images.
WD40EFPX members take too long recovering from an error and the controller drops them for being slow about it. Put on hardware that is willing to wait, the very same drive images cleanly.
Shut into a cupboard with no airflow, a NAS spends years running warm and then drops two disks inside a month. Failures that cluster like that are the signal to image every member, not to swap one and hope.
SMR members cope with everyday duty and then buckle under resilver load, which the whole industry worked out in public a few years back. Pacing that respects the shingling comes first. Reconstruction comes after it.
An archive nobody has ever scrubbed collects bad sectors quietly for years, then hands over every one of them in a single afternoon, at the precise moment the redundancy has been spent. Imaging first takes the sting out of that.
The small operating-system partitions mirrored across every disk have corrupted. The box will not boot as a result, and the data volume sitting behind them is entirely sound. It gets lifted straight out of the images.
A block device carved out of NAS space tears whenever the volume beneath it tears. The container file is repaired first, then the file system living inside the container. Two reconstructions, one job.
Bad memory introduces errors on the way to disk, and the file system then certifies them as correct, in perfectly good faith. A volume in that state has to be audited rather than trusted, so auditing is what it gets.
A cache device in read-write mode is holding data the disks have not seen yet. Lose that device, or one half of a cache pair, and nothing mounts, because part of the volume is genuinely absent. The cache gets imaged as a member in its own right.
When a NAS dies, moving the disks into another box sounds like the obvious next step. Vendors lay their arrays out differently, though, and the replacement will very likely offer to initialise the set for you. Say yes to that and there is no way back, so send the disks to a bench instead.
Modern NAS file systems need free space in order to record any change at all, so a volume run to capacity can reach the point where it cannot even delete a file to make room. Snapshots hanging on to that space make matters worse. On an image the whole thing untangles safely, which is not true of a live volume.
Folder-level encryption on Synology and QNAP comes with a key file, and owners are told to store it somewhere safe. Somewhere safe turns out, very often, to be the NAS. No key and no passphrase means the folder stays shut, and that gets said at the diagnostic stage rather than at the end of the job.
The NAS was mirroring a laptop. Somebody tidied the laptop up, and the mirror dutifully deleted the same files at its own end. Nothing broke and nobody attacked anything. Those files now live in snapshots and in unallocated space, and the number that come back falls with every week the box keeps running.
Files on a NAS are cut into stripes and shared out over the disks, either by standard RAID or by whatever the maker chose to call its own arrangement. SHR at Synology, BeyondRAID back when Drobo was selling them. However it is dressed up, the margin is narrower than the box implies. Losing one disk gets handled without much fuss. What follows is the dangerous part, because a rebuild reads every remaining disk from one end to the other for hours at a stretch, and that is precisely the load under which a second disk of the same vintage, bought in the same batch, decides it has had enough. Work here is arranged so your disks never face that. Every one gets copied in full before anything else happens. Stripe size, disk order and the direction parity rotates are then derived from the copies. The volume goes back together in software, and only at that point does a file come off it. Your own hardware is never written to.
Do not touch the rebuild button. Re-initialising is what the wizard puts forward as the obvious next move, and it is the one option on that screen capable of ending the job outright. What to do instead: power the unit down properly, draw the caddies out one at a time, and write the bay number onto each disk with a marker as it appears, because the sequence counts for a great deal when somebody puts the array back together on a bench. Numbered disks go in the parcel and the empty chassis stays with you. A one-bay unit such as a My Cloud just needs its case opening and the bare drive sending. If the case will not come apart, phone first rather than levering at it. Broken clips, and a cracked disk sitting behind them, are a familiar sight here.
More Buffalo units come through this door than the company's market share would predict. LinkStation 210 and 220 boxes out of spare rooms and two-person offices, TeraStation boxes out of firms that needed four bays in something rack-shaped. Both carry the same piece of good news inside them. What Buffalo fits is plain SATA drives, with no special chip anywhere in the chassis holding the shares to ransom and nothing in the enclosure that has to be present before the data can be read. Confusion comes from the formatting. Buffalo lays its volumes down as XFS, or as ext3 and ext4 on older models, and Windows understands none of those, so a LinkStation disk wired straight into a PC gets reported as empty along with a friendly suggestion that you format it. That suggestion is the one thing on the screen to refuse. On this bench each member disk is imaged end to end, the RAID geometry gets reconstructed from those images instead of from a controller that lost the plot, and the XFS volume is mounted in software so the shares come off as ordinary folders. The case that turns up again and again is a rebuild nobody stopped. One drive gets flagged, a replacement goes into the bay, the rebuild then reads every surviving member from start to finish over several hours, and a second drive of the same age out of the same batch quits somewhere in the middle. Now the array is down and the enclosure has run out of ideas. Data like that comes back from the member disks, and it comes back far more reliably when nobody has gone for a third rebuild in the meantime. Shut it down, number the caddies, post the drives, keep the shell. Buffalo jobs are priced alongside everything else on the array bench, from £500 + VAT and going up with the disk count.
A NAS is a RAID set speaking a vendor dialect, and the dialects that matter are all spoken in this building:
Members are cloned in parallel on dedicated hardware, and that happens before anybody offers a theory about the array. Every risk in the rebuild then falls on copies while the originals sit idle.
Stripe size, member order, rotation and offset are all derived from the images, and only then does the array get stood up virtually. Geometry has to be demonstrated here, never assumed.
A dead member turns into a mechanical case in its own right, with heads, firmware or board work as needed, until it can be imaged alongside the rest of the set.
Entropy testing and parity checks confirm the solved layout before one file is lifted off it. Nothing here proceeds on a hunch.
SHR and X-RAID, BeyondRAID, QNAP flexible volumes, mdadm beneath LVM, btrfs and ext4. Vendor layering translated back into plain folders.
A sealed home unit gets opened cleanly and its Linux-formatted disk is mounted natively, so the shares come off complete rather than in pieces.
Two-bay DiskStations and the TS-x53 line from QNAP are what actually sit humming in stockrooms, studios and back offices from the Baltic Triangle out to Speke, and over the water in Birkenhead too. They lead the rebuild ledger accordingly. Drobo closing its doors left every BeyondRAID owner with lab work as the only remaining route. Unfortunate, and true. The question owners ask above all others is about disk order, and the honest answer is that the metadata written on the members holds it. A marker pen applied to each caddy before anything is pulled still makes every step after that quicker and safer. Array, NAS and server work starts at £500 + VAT and climbs with the number of disks. Whatever badge is on the front, the sequence is the same: shut it down, number the disks, post the disks, hang on to the shell.
DSM is friendly on the surface and deep underneath. SHR merges disks of different sizes into a single pool. Below that sit mdadm and LVM. Above it sits btrfs, or ext4 on the older units. That depth is what defeats general-purpose software: a crashed volume on a DS220+ reads as gibberish because whatever tool you point at it sees one layer, while the answer lies spread across three. Here the stack is solved from the bottom up, in the order it was assembled, and always from complete images of the members. Establish the SHR arrangement, mount the file system virtually, walk the shared folders off intact. A small unit out of a spare bedroom and a rackmount out of a server cupboard follow an identical route, whatever mixture of capacities they happen to contain.
Buffalo builds a plain box, and plain suits the owner. One or two ordinary SATA disks go into a LinkStation 210 or 220. Four or six go into a TeraStation, arranged as RAID 1, 5 or 6. The enclosure itself contributes nothing whatever to reading those disks again. Current models format the volume as XFS and older ones as ext3 or ext4, and on every model a small partition at the head of each disk carries the box's own operating system. That last detail explains why a Buffalo with damaged firmware can appear thoroughly dead while the shares behind it are untouched. Image each disk, solve the array from those images, mount the volume in software, walk the folders off it. Two habits are worth knowing. Buffalo tends to fit disks from a single batch, so a rebuild leaning on the survivors makes a second failure genuinely likely. And the LinkStation repair wizard is quite happy to re-initialise a volume it does not understand. Shut it down, number the caddies, send the disks.
Add capacity to a ReadyNAS and X-RAID quietly reshapes the volume behind the scenes. Do that a few times over the years and you have expansion zones stacked up like sediment. General-purpose tools read the stacking as damage and report it with real conviction. Rebuilt in the order those zones were laid down, the volume mounts just as it always did in the box. The cheap single-bay boxes set a simpler trap. Their disk carries ext4 or XFS, Windows recognises nothing at all on it, declares it blank and offers to format it. Refuse. Everything is still sitting on there, and both file systems mount natively on a bench here.
Shut the unit down properly before anything else happens. Take the disks out one bay at a time, and as each one comes free, write its bay number on the drive itself in marker pen. That numbering earns its keep several times over during reconstruction. Only the disks need to make the trip. A NAS is one of the few things that can be posted whole if you would rather not open it up, and a sealed single-bay unit generally has to come that way. Post it tracked and insured to Manchester Data Recovery, Peter House, Oxford Street, Manchester M1 5AN, or book a courier yourself. Collection is not a service offered anywhere in this network, so the journey has to be arranged at your end.
Almost every job on this bench came in by tracked, insured post. It is the calmest way to move a drive that is already struggling, and a parcel handed in anywhere on Merseyside usually reaches the bench the following 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.
Diagnosis free, one figure written down, and £300 + VAT covers a single drive or SSD. Start online or ring.