Data Recovery Case File · Formatted & Logical Faults · Twenty Hours, One Third

Failing Drive: Why Scan and Repair Ran for Twenty Hours

His enquiry recorded an experiment nobody should repeat, and recorded it precisely enough to be useful. His father's WD My Passport Ultra 3TB "recently displayed a fatal hardware error. I tried it in my PC to see what the message was, and Windows suggested scan and repair. I left it to do so for about 20 hours and it only got a third of the way through. I cancelled it." Three things to say. The error message was telling him something specific and it was not a software problem. The scan's slowness was itself the diagnosis — twenty hours to cover a third of a drive is not a slow computer, it is a drive struggling to answer. And cancelling was the right call, made later than ideal but made. This page explains why the tool Windows offered was the wrong one for the fault it was offered for, and what those twenty hours are likely to have cost.

The decode: the message, the wrong tool, and what twenty hours cost

What a fatal hardware error means: it is Windows relaying something the drive reported. The system asked for data, the drive failed to supply it and returned a hardware-level error rather than a filesystem one. That single word — hardware — is the whole diagnosis, and it is the point at which the correct response diverges completely from what happens next in most households.

Why scan and repair was the wrong tool: Windows responds to a drive it cannot read cleanly by offering to scan and repair the filesystem, and the offer is well-meant and almost always wrong in this situation, for three reasons. It is aimed at the wrong layer: it repairs filesystem structures, and a hardware read failure is not a filesystem problem, so it is treating a symptom on a floor above the fault. It writes: repair means modifying structures on a drive that is already failing to read reliably, which can turn recoverable inconsistency into real damage. And it retries: faced with a sector that will not read, it tries again and again, because that is sensible behaviour on a healthy drive with an occasional glitch. On a degrading drive it means hours of sustained hammering over precisely the weakest regions.

Why the slowness was the answer: this is the part worth generalising. A filesystem check on a healthy 3TB drive is a matter of hours at most. Twenty hours to reach a third means the tool was spending nearly all its time waiting on reads that were not completing — hundreds of thousands of retries. So the progress bar was not reporting a big job; it was reporting a struggling drive, in real time, for anyone able to read it that way. Any repair utility that runs dramatically slower than expected should be stopped rather than waited out, because its slowness is the drive telling you it cannot do what is being asked.

What the twenty hours cost, honestly: stated without alarm, because it cannot be undone and because the outcome is usually still good. Sustained retry loads on failing regions accelerate deterioration — heads passing repeatedly over weak areas, mechanisms working continuously at exactly the wrong task. Some sectors readable at hour one may not be readable at hour twenty. It also consumed a great deal of the drive's remaining margin on a scan that produced nothing. That is a real cost and it is why the instruction matters for the next reader. But he cancelled, he stopped there, and he asked — and a drive that has had one bad session is in a very different position from one that has had a dozen.

On the bench

The Atola Insight Forensic established the mechanism's condition first, mapping where the surface was strongest and where the scan had been labouring — a picture the progress bar had been drawing for twenty hours without anybody being able to read it. Imaging then ran on the PC-3000 Express under Data Extractor with per-sector timeouts enforced, which is the crucial difference from what the drive had already endured: a failing sector is given a strict budget and then deferred rather than retried indefinitely, so the pass keeps moving and the drive is not held over its weakest ground. Healthy regions were captured completely first, weak neighbourhoods revisited last, and the filesystem was rebuilt on the completed image. His father's data was verified by opening and delivered on fresh media.

The outcome

The drive imaged under timeout control and the filesystem rebuilt from the copy, with the contents verified and delivered. Free assessment, one fixed written figure including VAT; where a drive has to be opened, 50% of parts and labour is payable upfront with the balance only on success — otherwise no recovery, no fee. The lesson, put plainly for the next person offered a repair scan: a hardware error is the drive reporting its own failure, and a filesystem repair tool is aimed at the wrong layer, writes to a drive that cannot read reliably, and retries endlessly on exactly the regions that are failing; if a check is running dramatically slower than it should, that slowness is the diagnosis and the right response is to stop it, not to wait it out. Cancelling was right — and cancelling sooner would have been righter.

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.