Data Recovery Case File · Desktop Drives · Case One Hundred

Three terabytes reporting four gigabytes: a famous drive's familiar exit, caught mid-backup

There's a rueful comedy this trade hears weekly, and this enquiry delivered it with good grace: "I had been in the process of backing the data up — typical that it gives up the ghost before I'm finished!" The drive was an ST3000DM001, a 3TB Seagate from a production run whose elevated failure rates later became a matter of public statistical record — and its exit was the model's signature move: still present in Device Manager, absent from This PC, and reporting its capacity as three or four gigabytes, depending on what asked. Her instincts were flawless: no recovery software, no experiments beyond checking the cables. Just the right question, asked in time.

How a drive forgets its own size

A drive announcing a few gigabytes hasn't shrunk any more than the zero-capacity drives elsewhere in this archive have vanished — it has lost the ability to describe itself. Deep in every drive's on-platter working records lives the translator: the master conversion between the addresses computers ask for and the physical sectors that answer, and the structure that tells the world how big the drive is. When the service area holding it degrades — this model family's best-documented failing — the drive boots into a stunted fallback self, reporting whatever fragment of identity still parses: here, a 3TB estate introducing itself as a USB stick. The reassuring mechanics underneath: her three terabytes of documents sat physically untouched; the drive had lost its index to itself, not its contents. And her half-finished backup deserves its sympathy — the marathon read of a full backup is exactly the workload that surfaces a marginal service area, which is why "it died while I was backing it up" is less cruel irony than cause and effect. The lesson is never don't back up; it's start sooner than the drive's first strange day.

The recovery

Firmware-class fault, firmware-class fix. On the PC-3000, the drive's degraded service structures were repaired from their redundant copies and the translator regenerated — the drive relearning, from its own deeper records, that it was and always had been three terabytes — with its self-maintenance stood down so nothing could re-corrupt what had just been rebuilt. Restored to its full self, it was imaged in Data Extractor with the model's reputation kept firmly in mind: promptly, thoroughly, healthy regions first, nothing left for a second session it might not offer. The NTFS volume mounted from the image complete; her work documents — the very files the interrupted backup had been racing to save — were extracted, verified, and delivered, alongside the answers her enquiry had asked for and received up front: a fixed written figure after free assessment, and a turnaround stated in days and met.

Outcome — and a closing word from case one hundred

Full recovery, the backup finished by other means, the famous drive retired with its statistics. It makes a fitting hundredth entry, because it contains this archive's whole philosophy in one small story: a fault that looked like disappearance and was really description; a customer whose restraint — decline the prompts, skip the software, ask early — preserved every option; and an ending built from the same honest mechanics as the ninety-nine before it. Drives fail. Models earn reputations. Backups start late. And nearly always, beneath the alarming symptom, the data is still there — waiting for the problem to be understood rather than fought. That's the work. It's documented across a hundred cases now, and the number is only where we paused to write it down.

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.