Data Recovery Case File · Mac & Apple Ecosystem · The Reinstall That Didn't Hold

Fixed for three days: why a reinstall that doesn't hold is itself the diagnosis — First Aid's clean bill decoded, the physical-versus-corruption question answered properly, and the iMac's data secured first

The enquiry arrived on his father's behalf, structured like an engineer's incident report — the problem, the background, what we suspect — and it deserves answering in the same spirit. The machine: a late-2015 iMac running Big Sur that "will not complete startup; stalls at the white Apple logo on black screen." The background, precisely logged: three days earlier, Office apps had stopped opening after a suspected auto-update, so they reinstalled Office and reinstalled Big Sur via Recovery Mode — which fixed everything, with the Mac "running smoothly" — and "we had also run Disk Utility First Aid over all drives, with 'success' results." Then, three days later: the stall. His asks, exactly right: (a) the data retrieved — a backup of the drive — and (b) a diagnosis: "is this a physical problem or data corruption?" This page answers both — starting with the pattern his own log reveals: a fix that works and then stops holding is not a failed fix. It's a diagnosis.

The decode: the relapse pattern, the clean bill, and the question answered properly

Why a fix that doesn't hold points below the software: the three-day arc is the most informative thing in the log. A reinstall via Recovery is major surgery at the software layer: gigabytes of fresh system written, structures rebuilt from scratch — and it worked, which tells us the software layer was rebuildable and briefly sound. Then the same class of failure returned, unprovoked. When clean software degrades back into the same illness within days, the suspicion moves down a layer: something beneath the software is corrupting what sits on it — most commonly, on a drive of this vintage, degrading media: weak sectors returning bad or unreadable data, so that a freshly-written system decays as its critical regions land on failing ground. The original Office trouble, the "cure," and the relapse then read as one story: the software wasn't the patient. It was the symptom, twice.

First Aid's "success," decoded once more: the clean bill three days before failure fits the same story, for the reason this archive keeps documenting: First Aid audits bookkeeping — whether the filesystem's structures are self-consistent — and a freshly reinstalled volume passes that audit almost by definition. What First Aid cannot see is the ground beneath: sectors that read correctly today and marginally tomorrow. A "success" result is a true statement about the map and no statement at all about the territory — which is why it coexists so comfortably with a drive three days from collapse.

His question (b), answered the only rigorous way: "physical or corruption?" is exactly the right question, and it has exactly one rigorous answer: measurement, not inference. Boot symptoms cannot distinguish the two — both stall at the same logo — but bench diagnostics can: the drive's media and heads read directly, error rates and weak regions mapped as physical fact, the verdict established rather than argued. And the order of operations his ask (a) already implied is the professional one regardless of the verdict: the data comes off first — because if the answer is "physical," every further boot attempt was spending the drive, and if the answer is "corruption," the image loses nothing and the repair happens safely on the copy.

On the bench

The iMac gave up its drive per its construction, and the Atola Insight Forensic answered question (b) by measurement: media and heads assessed directly, the weak regions the relapse had implied mapped and confirmed — the verdict, in writing, that the software's two illnesses had both been symptoms of degrading ground. Question (a) was already underway by then: imaging on the PC-3000 Express under Data Extractor, the failing neighbourhoods worked patiently on their own passes, until the father's volume stood complete in the image. The Mac volume was verified from the copy — documents, photos, the Office files that started it all — and delivered alongside the written diagnosis: both halves of the enquiry answered, in the order that kept them both answerable.

The outcome

The data secured and verified; the physical-versus-corruption question answered by measurement, in writing — degrading media beneath a twice-rebuilt software layer. Free assessment, one fixed written figure including VAT, no recovery, no fee. The pattern, posted for every household that fixes a machine twice: a repair that works and then stops holding has diagnosed itself — the fault lives below the layer you repaired; First Aid's clean bill audits the map, never the ground; and the moment a fix relapses, the priority inverts for good: data off first, and let the verdict arrive by measurement while the files are already safe.

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.