Data Recovery Case File · Apple Mac · Empty on the Outside

A hollow volume over a full disk: the catalogue decode, First Aid's defeat read as evidence — and a straight answer on reformatting afterwards

His theory arrived pre-formed and, for once in this archive, substantially correct. The WD My Passport for Mac 2TB "has corrupted," he believed, "due to being ejected without using the proper procedure" — and the presentation is the genre's strangest: the Mac reports it unable to repair the disk, yet the drive mounts, opens — "and I can see nothing inside, but there is data on the HDD": 168GB of it, by the volume's own accounting. His request had two parts, and the second is asked less often than it should be: retrieve the data, "and reformat the drive if possible" — the reuse question, which gets the honest, fault-class-specific answer this page ends on.

The decode: how a drive opens onto nothing — and what First Aid's failure really said

A Mac volume keeps its knowledge in layers: a header that lets it mount, space accounting that knows how full it is, and — the layer his eject caught mid-sentence — the catalogue: the master index of every name, folder and location, constantly rewritten as files change. "Ejected without the proper procedure" is precisely the act that interrupts a catalogue mid-update — the eject ritual exists to let the drive finish its bookkeeping before power vanishes — and a catalogue damaged at its root produces his exact paradox: the header mounts the volume (so it opens), the space accounting still counts 168GB (so the data is "definitely there"), and the listing comes up empty (because the index of names won't read). Hollow outside, full inside — his theory, confirmed in the mechanism. First Aid's defeat, then, was information rather than doom: Disk Utility's repair patches modest inconsistencies, and "unable to repair" means it found damage beyond patch scope — a finding, honestly reported by the tool — with the quiet mercy that its failure meant it also declined to rewrite structures it couldn't reconcile (this archive's repair cases document what half-successful repairs prune). His stopping there — one First Aid pass, no third-party fixers, no format — left the drive in the recoverable state the decode promises: because a Mac volume's catalogue is reconstructible, from the journal that logs its recent updates and the alternate structures the filesystem keeps precisely for this day.

The recovery — and the reformat verdict

The drive was imaged first, write-blocked — even logical cases get the copy-then-work order — and the reconstruction ran on the image: the journal replayed to recover the catalogue's interrupted state, alternate structures reconciled where the journal's reach ended, and the index of names rebuilt until the hollow volume filled back in: folders, filenames, dates — the 168GB of music collection and family photographs listing exactly as it had before the yanked cable. Verification opened files across the set; delivery went out on new media. Then his second question, answered by fault class rather than reflex: reuse depends entirely on why a drive failed. Hardware casualties — the failing controllers, degrading platters and lost translators elsewhere in this archive — are retired regardless of how recoverable their data proved, because the hardware's character is disclosed. His was the other kind: a logical injury on healthy hardware — the drive did nothing wrong; the eject did — and the post-recovery verification pass confirmed it: mechanism sound, surfaces clean, health telemetry unremarkable. Verdict, in writing: yes — reformat and reuse with a clear conscience, freshly formatted here as requested, with the one condition that converts the whole episode into a lesson learned: it goes back into service as a second copy of something, never again the only one.

Outcome

The hollow volume refilled, the drive honestly cleared for duty — and the page's filings, one per question he asked. On the fault: a drive that opens onto emptiness while its space reads used has lost its catalogue, not its contents — stop at one First Aid attempt, skip the third-party "fixers" (catalogue rebuilds belong on an image, not the original), and treat the eject ritual as the two seconds of bookkeeping it is; his suspicion about the improper eject was correct, and the habit is the entire prevention. On reuse: ask it the way he did, and expect the answer to come with reasons — logical fault on verified-healthy hardware: reuse; hardware fault of any kind: retire — because a lab that clears or condemns a drive by fault class, with the verification to back it, is pricing your future risk honestly rather than selling you either fear or false comfort. Two straight questions, two straight answers, one drive back at work — demoted, by mutual agreement, to copy number two.

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.