Data Recovery Case File · NAS · The Layer Beneath the Mount
The superblock that "couldn't be found" was never where he was looking: a knocked-over My Cloud, decoded for its Linux-literate owner
This enquiry came from someone whose rescue attempt was, step for step, what a professional would try first from a desk: drive out of the dead My Cloud, powered SATA-to-USB adapter, a Linux machine, and a methodical read of the results — the file manager saw nothing, the disks utility saw the device, and mount refused with the classic verdict: superblock can't be found. His conclusion that he'd hit his limit was right. His tools weren't the problem, though — his target was.
| Device | WD My Cloud 2TB single-bay NAS; internal 3.5″ drive; data volume ext4; <1TB in use |
| Reported events | Unit believed knocked over during building work; constant red LED, unreachable on the network; drive extracted and examined on Linux via powered adapter — device visible, volume unmountable, superblock not found |
| Fault class | Impact-degraded reads on the drive, beneath an appliance disk layout that hides the filesystem from a whole-disk mount |
| Equipment used | ACE Lab PC-3000 Express + Data Extractor (imaging; appliance-volume assembly and extraction) |
Two reasons the mount was always going to fail
The first reason is pure geography, and it catches Linux-capable owners precisely because they know what they're doing. A My Cloud doesn't lay its ext4 volume across the raw disk: the appliance carves the drive into several partitions — its own operating system, its housekeeping, and only then the data volume, which it additionally wraps in a software-RAID container even with a single disk, mirror-of-one style. Point mount at the whole device and the kernel looks for a filesystem's opening signature where the appliance keeps its partition table instead — and reports, with complete accuracy and total unhelpfulness, that no superblock lives there. The second reason was the one that made this a lab job either way: the tumble. Assessment found the drive reading unreliably — an impact taken while the unit was powered had left the heads marginal — so even the correctly-targeted partition, inside the correctly-assembled container, was sitting on hardware no longer safe to explore by trial and error. His restraint at the "superblock" wall, rather than escalating to filesystem-repair commands against a mystery layout on a wounded drive, kept every option open.
The recovery
The order of operations resolved both problems in sequence. Hardware first: on the PC-3000, with the drive's self-maintenance retired, the disk was imaged with the care its post-impact condition demanded — the strong surfaces early and quick, the strained regions negotiated patiently at the end. Layout second, entirely on the image: the appliance's partitions enumerated, the single-member software-RAID container assembled exactly as the My Cloud's own firmware would have done, and — inside it, at last — the ext4 volume, its superblock precisely where it had been all along, several layers below where any whole-disk mount could reach. The filesystem opened cleanly; his sub-terabyte of data extracted with its folder structure intact, verified, and delivered on a plain drive formatted so his Linux machine mounts it first time, no archaeology required.
Outcome
Full recovery — and a page written partly as the answer his mount command deserved. For every technical reader who's met "superblock can't be found" on a NAS drive: before believing it, ask what the appliance actually built — partitions within the disk, containers within the partitions, the filesystem innermost — because on appliance drives, "not found" almost always means "not here," and "not here" is a treasure map, not a verdict. And for everyone whose storage lives at floor level during building work: appliances running 24/7 are running while the ladder swings. A shelf above shoulder height is cheap insurance.
For Linux hands examining NAS drives
Read the partition table before attempting any mount, and expect appliance layouts: system partitions first, the data volume last and usually inside a software-RAID wrapper even on single-bay units. Never run filesystem repair against a layout you haven't fully mapped, and stop entirely at any sign the drive reads unreliably — imaging outranks exploration on wounded hardware. Your tools are the right ones; make sure the patient's anatomy is on the table first.
The data's usually a layer deeper — call Bristol Data Recovery on 0117 332 1137 for a free assessment.
Request a quote online →
Our case files are drawn from genuine enquiries received by our laboratory over the past ten years, anonymised to protect client confidentiality. Each one describes the diagnostic and recovery procedure our engineers apply to that fault, using the equipment listed.