Data Recovery Case File · Formatted & Logical Faults · Two Repairs, Opposite Results
Each Repair Writes, and the Second Wrote Over What the First Had Rebuilt
His enquiry records a sequence in which the same kind of action helped and then harmed. A drive with filesystem damage that was "initially invisible, but the start-up automatic repair rendered the drive and some of the files visible" — after which he ran the repair utility deliberately, and "the drive is once again not recognised." Both operations did the same thing: they rebuilt structures by writing new ones, and the second had worse material to work from.
| Media | Desktop hard drive with filesystem damage — presenting partially after an automatic repair, then ceasing to present following a second repair operation |
| Reported situation | Desktop drive suffering filesystem damage · drive initially not visible to the operating system · automatic start-up repair performed by the system · drive and a portion of files becoming visible afterwards · a further repair utility run deliberately · drive no longer recognised following that operation |
| Fault class | Filesystem structures partially rebuilt then overwritten — successive repair operations writing over prior state; content regions substantially retained beneath both |
| Equipment used | Repair sequence reconstructed before any further action was taken · no further repair, check or mount permitted · imaged write-blocked at the block level before any interpretation · structures from before both repairs recovered from residual copies across the volume · content carved by signature and reconciled against the partially visible state |
The decode: what each repair did, and why the order produced this
What a filesystem repair fundamentally is: a rewrite. It reads what structures it can, decides what a consistent filesystem would look like, and writes that — it does not restore anything.
Why the first one appeared to succeed: it found enough. Working from the damaged original, it rebuilt structures consistent enough for the volume to mount and some files to appear.
Why only some files appeared: it discarded what it could not reconcile. Records that could not be made consistent are dropped rather than preserved, so the result is a smaller, valid filesystem.
Why that is the crucial detail: the missing files were not deleted. Their content remained on the drive with nothing pointing at it, which is a recoverable state.
What the second repair then operated on: the first repair's output. It read the rebuilt structures rather than the originals, because the originals had already been overwritten.
Why that made things worse rather than better: the source was already lossy. A repair of a repair works from a reduced description and can produce something the system cannot mount at all.
Why the drive vanishing entirely follows from that: the volume description became invalid. A partition or filesystem header written inconsistently leaves the system with nothing it recognises.
Why this is still recoverable despite two overwrites: both wrote small amounts. Repairs modify structures rather than content, and the great bulk of the drive has not been touched by either operation.
Why the original structures may still exist: filesystems keep duplicate copies elsewhere. Backup records at other positions on the volume frequently survive operations that overwrote the primary ones.
What the lesson is, stated plainly: a repair that helps is a warning rather than an invitation. The moment the drive became partly visible was the moment to copy everything off, not the moment to run the tool again.
On the bench
The repair sequence was reconstructed before any further action was taken — a filesystem repair reading what structures it can, deciding what a consistent filesystem would look like and writing that, discarding records it cannot reconcile. The first therefore produced a smaller valid filesystem with unreferenced content intact; the second operated on that reduced description rather than the originals, producing a volume the system cannot mount. Pre-repair structures were recovered from residual copies.
The outcome
The repair sequence reconstructed first, no further repair permitted, and structures from before both operations recovered from residual copies. 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 decode: both repairs did the same thing. The first rebuilt from the damaged original and worked partly; the second rebuilt from the first one's output, which was already a reduced description.
When a repair has partly worked
Copy everything off immediately and run nothing further. A repair doesn't restore anything — it reads what structures it can, decides what a consistent filesystem would look like, and writes that, discarding whatever it can't reconcile. So a partial success means you're looking at a smaller valid filesystem built over your originals, and running a second repair works from that reduced description rather than the real one. The moment files became visible was the moment to save them.
Run nothing further — call Cambridge Data Recovery on 01223 655015; the sequence reconstructed first, imaged write-blocked, pre-repair structures recovered from residual copies across the volume.
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.