A very particular sort of job comes in from households and small businesses out towards Roslin, Bilston and Loanhead: a Synology or QNAP NAS has died, somebody has pulled one of the drives out of it hoping to grab the family photos or the accounts folder on a laptop, and the moment they plug the disk in, Windows pops up demanding to format it. That is the moment most of the damage happens — not to the drive, but to the person’s nerves. If a NAS has just gone dark on you and you’re staring at that message right now, NAS data recovery in Roslin is what this guide walks through, and the single most important thing you can do is not click the Format button while you read it.
What follows is the workflow we actually run on the data recovery bench for these jobs. It’s built on the fact that a NAS drive looking “blank” in Windows almost never means the data is gone. In the great majority of these callouts the drive is fully intact; Windows just does not speak the language the NAS wrote onto it, and its response is to offer the only tool it has — the format dialog — for a problem that does not exist.
The Message That Panics Everyone
The exact wording is normally one of two variants. Either “You need to format the disk in drive E: before you can use it. Do you want to format it?”, or a smaller pop-up saying the disk is not initialised, offering MBR or GPT with an OK button. Both look, to a non-technical reader, as though Windows is politely offering to fix the drive. It is not. It is offering to overwrite the parts of the drive we most need to keep.
The panic comes from the mismatch between the message and the reality. From the outside it looks as though the drive has forgotten what was on it. From the bench it looks like a perfectly readable NAS disk that Windows happens not to understand. Those are different situations, and they call for opposite responses. Windows wants to write. We want to only ever read.
What Windows Actually Sees When You Plug in a NAS Drive
Windows knows two mainstream filesystems well — NTFS and exFAT — plus FAT32 for older USB sticks and ReFS for very recent Server installs. Everything else is, as far as it’s concerned, an unformatted disk. It scans the first sectors, finds a partition table it does not recognise (or one it thinks is corrupted, because the layout doesn’t match Microsoft’s expectations), and defaults to the format prompt.
Synology drives are usually formatted as Btrfs on top of a Linux mdadm array wrapped in LVM, which is the stack their SHR (Synology Hybrid RAID) layer sits on. QNAP defaults to ext4 or ZFS depending on the model and firmware. Both are perfectly ordinary Linux setups from a technician’s point of view, and both are invisible to a stock Windows install without third-party drivers. The disk is fine. The reader is wrong.
You can confirm this in Disk Management (right-click Start → Disk Management) without touching the drive: the disk will appear at its full advertised size — a 4 TB Synology drive shows up as roughly 3.6 TB — with two or three partitions, all listed as “RAW” or “Unknown”. That partition layout is the fingerprint of a NAS. A truly failed drive tends to either not appear at all or show up as a tiny wrong size.
The One Click That Turns a Straightforward Job Into a Serious One
The most damaging thing on the bench isn’t the drive failure itself. It’s the recovery attempt that came before we saw it.
If you click Format, Windows writes a fresh NTFS or exFAT partition table over the beginning of the disk. That is the exact area holding the map of where the NAS put your files. The file data further down the platters is still physically there, but the roadmap for reassembling it isn’t. Recovery is still possible — deep-scan tools like PhotoRec can pull individual files back by scanning for their signatures — but you lose the folder structure, the filenames, and the neat organisation you had. On a family photo library that is the difference between “here are your holidays from 2019 to last summer” and “here are 84,000 JPEGs called f0000001.jpg”.
The Initialize Disk dialog is worse in one respect: it rewrites just the very first sector, which is short, quiet and easy to miss. Customers often click OK on it without realising anything happened, and it leaves the file data untouched but breaks the array’s ability to recognise its own drive when it goes back in a working NAS. That’s a fixable problem — but only if we know about it before the array is rebuilt on top.
The rule on the bench is simple. If Windows offers to do anything to a NAS drive, the answer is Cancel, every time, until the disk is on a Linux tool that can read it as it is.
SHR, Btrfs, ext4, ZFS — Why the Filesystem Matters
The route in depends on how the NAS was set up, and there are four common cases we see out of the Midlothian units and Roslin households.
Single-drive Synology or QNAP. The easiest one. The whole filesystem — Btrfs or ext4 — sits on that single disk, sometimes wrapped in an mdadm single-disk “array” and LVM volume. A Linux live USB with the right utilities mounts it read-only in a couple of minutes, and everything copies straight off. Most SHR-1 boxes with one drive fitted are this case.
Multi-drive Synology SHR or SHR-2. Here the data is striped across two, four or more drives with parity, and no single drive contains a complete copy of anything. Pulling one disk on its own gives you nothing usable. The array has to be reassembled on a Linux workstation, mdadm has to recognise the members, and the Btrfs volume on top has to be mounted — usually read-only, always with the original drives ideally imaged first. This is bench work, not a home job.
Multi-drive QNAP with ZFS. Newer QNAP models default to ZFS pools. ZFS is unforgiving of missing drives in a striped pool, but conversely it is very good at reading the array back off the surviving disks if they’re all present. As with SHR-2 the drives need to be brought up together on a Linux workstation, and the pool imported read-only.
Older Netgear ReadyNAS or LinkStation. Ext4 with LVM, sometimes on top of a proprietary RAID layer. Same principles, different tooling — but perfectly recoverable if the drives themselves are healthy.
One Drive vs. Two — When It’s an Array and When It Isn’t
A quick sanity check saves a lot of pain. Open the NAS itself (with it unplugged from the mains) and count the drives that were fitted. If there is only one, the disk in front of you is almost certainly the whole story — a single-drive recovery, which is the straightforward end of the job.
If there are two drives and they were set up as SHR, SHR-1, RAID 1 or JBOD, one drive usually holds a complete copy of everything and can be read on its own. If they were set up as RAID 0 or a striped SHR, one drive has half the data and none of it is readable alone. That’s the case where the mistake of “let me just try the other drive in the PC” can do lasting damage, because it wears a working disk that we need to bring the array back up.
If you can see four drives in the NAS, stop, don’t remove any more, and label the bays with a marker pen before anything else. The drive order matters and the array will need to be reassembled with the members in exactly the slots they were in when it worked.
A Recent Callout Off Rosslyn Chapel Road
A family a short walk from Rosslyn Chapel called us out after a lightning storm took out the mains one evening and their Synology DS220j never came back. Two 4 TB WD Reds inside, set up as SHR-1 — effectively RAID 1, so a mirrored pair, meaning each drive held a full copy of everything. Ten years of family photos, a self-employed design portfolio, and their tax paperwork.
They had done what most people do: pulled one drive out, put it in a USB-to-SATA caddy, plugged it into a Windows laptop, and been shown the format prompt. To their enormous credit they closed the dialog, unplugged the caddy, and put it back down before ringing us. That single act — not clicking — was the entire reason the job took ninety minutes rather than a week.
On the bench we imaged both drives first with ddrescue onto a fresh workshop disk (belt and braces, in case a bad sector showed up mid-copy). Then we booted a Linux live environment on our recovery workstation, assembled the mdadm array from the images, mounted the Btrfs volume read-only, and copied the entire “photo” and “work” shares onto a new external drive for them. The whole thing came back with filenames, folders and modification dates intact. The Synology itself turned out to have a blown power supply, which is a separate and far simpler problem — but the panic that day was about the drives, and the drives were never in trouble.
When It Really Is the Board, Not the Filesystem
Occasionally the drive itself has actually failed and the “please format” message is a symptom of that rather than the filesystem confusion above. The tells are different: the drive clicks or beeps on power-up, spins up and immediately spins down, shows up at the wrong size in Disk Management, or fails to appear at all in Device Manager. Modern SMR helium drives that lose their control board can behave like this even when the platters are pristine.
Those cases go a different way. We image what we can with ddrescue — sometimes across several passes with different retry settings — and if the drive won’t co-operate the last route in is board-level work: a donor PCB with the ROM swapped from the original, and in the hardest cases microsoldering the memory chips off entirely for reading on a specialist programmer. That’s the same workflow we walk through in our guide to the drive that clicks and keeps trying to spin up, and it’s worth reading if the drive out of your NAS is making noises it didn’t use to make. That work happens at the Edinburgh bench and typically covers households across Roslin, Loanhead and Bonnyrigg, along with data recovery in Dalkeith and the wider Midlothian belt.
What To Do Right Now if Your NAS Just Died
- Cancel every format and initialise dialog Windows offers. Every time. Unplug the drive if you have to.
- Label the bays before removing more drives. Slot order matters for multi-drive arrays. A strip of masking tape and a marker is enough.
- Don’t click chkdsk, don’t run recovery utilities, and don’t Google “convert RAW to NTFS”. Those instructions were written for an entirely different failure mode — a broken Windows drive — and they will destroy a NAS layout.
- Establish whether it was a single-drive or a multi-drive NAS and, if multi-drive, whether it was mirrored (safe on any one disk) or striped (needs all disks together).
- Bring the drives in on the same day if you can. Fewer plug-ins and unplug-ins are always better; every mount is another chance for a bad sector to be hit.
- If it’s a business NAS with customer or accounts data, treat it as a stopped-clock incident and book the recovery through our business IT support desk so we can also confirm the backup that was supposed to be running actually was.
- If the NAS is inconveniently placed — a loft, a locked comms cupboard, a mounted bracket — a home or office callout handles the removal, transport and reinstall as one visit.
The pattern that runs through all of this is the same pattern we described for a USB stick that starts asking to be formatted, in our guide to a memory stick that suddenly wants formatting: a filesystem mismatch is not a broken drive. A drive that can’t be soldered off, as on the modern laptops in our soldered laptop storage guide, is a fundamentally different situation — a removable NAS drive is one of the easiest recovery scenarios in the trade, provided you don’t click the wrong dialog first.
Short version: if a NAS has died and Windows is offering to format the disk you pulled out of it, close the dialog, put the drive back down, and bring it in. The photos, the accounts file, the ten years of scans — on this sort of job they’re almost always still there.
Last updated: 7 September 2026