Data Recovery Case File · Portable Drives · A Backup With No Backup

The Backup Drive Is the Single Point of Failure

Her enquiry describes exemplary discipline resting on one device. A 6TB drive receiving every automatic backup from her desktop: "it now contains all records back to 2014. It is refusing to do another backup. I believe there is 1.5TB of backups in it covering 2014 till today. It's getting old and can be flaky about whether or not it" mounts. She has backed up faithfully for a decade. And the entire decade lives on one ageing drive that nobody backs up — which is the failure mode that catches the most careful people, precisely because they are being careful.

The decode: what refusing to write means, and the order to act in

What a backup drive refusing to accept backups is telling you: that it has reached the end of its usefulness as a destination. Combined with intermittent mounting, that is a drive in decline — and every further backup attempt is sustained write load applied to it, which is the opposite of what it needs. The attempts should stop.

The order that matters, and it is not obvious: the desktop machine still holds the live data. That is the original and it is currently fine. Protecting the original comes first — a second copy of the current files, to different media, today. It is quick, it removes the immediate exposure, and it means the backup history becomes a nice-to-have rather than the only thing standing between her and loss.

Then the history: ten years of versions is genuinely valuable — old drafts, deleted files, things nobody remembers having. Getting it off the failing drive is worth doing, and it needs handling differently from ordinary files.

Why it cannot simply be copied: a backup history of this kind is not a folder of independent snapshots. Each backup stores only what changed and refers back to earlier ones for everything else, so the whole decade is one interlinked structure. Copying it file by file breaks those links and produces something enormously larger and no longer usable as a history. It has to be transferred at block level, as a whole object, onto a destination at least as large.

Why that argues for acting now rather than later: a block-level transfer reads the entire drive, which is exactly the sustained load a declining drive tolerates least. Doing it while the drive still mounts is very different from attempting it after it stops.

The arrangement worth putting in place afterwards, stated plainly: the new drive receives the transferred history and is then retired from active use, and current backups go to a different device. That gives two independent copies of different vintages instead of one continuous chain. A backup that has no backup is one copy on hardware nobody checks — and a decade is a long time for a single drive to be the whole of an arrangement.

The check worth adding: open a file from the backup occasionally. Backup software reports success; only opening something confirms it.

On the bench

Further backup attempts were discontinued, each applying sustained write load to a declining drive. Live machine data was secured to independent media first — the desktop holding the current originals, so removing that exposure took precedence over the history. The drive was imaged write-blocked at block level to preserve the interlinked history, since backups of this kind store only changes and refer back to earlier ones, so file-level copying severs those links. Backup structure was validated after capture and the original retained until verification was complete.

The outcome

The live data secured first, the backup history captured at block level as a whole object and validated after transfer. 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: a drive refusing to accept backups has finished as a destination, and every further attempt is write load it does not need. Protect the live data on your machine first. Then move the history at block level, because each backup refers back to earlier ones and file-level copying breaks the chain.

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.