Casebook entry · Portable drives · EDR-2025-2319
Six Weeks of Fading, Then Nothing on Tuesday.
For six weeks the volunteer who is tasked with maintaining a village archive in east Devon was watching as his drive faded away: first it takes longer every time I plug it in
, then half the pictures never draw their thumbnails
, and finally last Tuesday it never appeared at all
. He had been storing a decade of scanned photographs and parish newsletters on it, and the borrowed originals had long since returned to their rightful owners.
Sound familiar? Talk to us.
0800 6890668
In plain English.
A drive will never slow the way a hinge becomes stiffened; its slowing is due to re-reads. The sectors which once provided answers on the first request are now providing them on the fifth request or even the fiftieth, and every additional attempt is an added delay you can feel. Thumbnails disappear early, since drawing them means actually fetching the real file data, with all of those retries included. And the morning the hard drive goes missing is the same arithmetic applied to mounting time: the checks it has to pass to be able to be visible now take longer than the computer is willing to wait.
The tools this job took.
See how your case runs →| System | The part it played | Why we rate it |
|---|---|---|
| DeepSpar Disk Imager 4 | Completed the deep passes at SATA level, timeouts tuned around the soft head | Images by head map — timeouts, resets and power-cycling managed per head |
| DeepSpar USB Imager | Took the opening copy through the Hitachi's own socket, no operating system involved | The same careful imaging, run over the USB bridge instead |
| UFS Explorer Professional Recovery | Reassembled the folder tree from the images and staged it onto the new disk | Handles the harder filesystems well — APFS, ReFS, XFS, ZFS and Btrfs |
On the bench.
Stop using it normally, start measuring it
Nobody even opened a single folder when it arrived. A drive that is sinking can only manage a limited number of reads, and 'browsing' uses those up for nothing. Instead measurement was used: the reallocation counters were very high and still increasing, and a survey indicated one out of the four heads answering late (and getting later), its territory containing just the pictures whose thumbnail views had been giving him trouble.
Take the easy majority while it is still easy
The capture ran with confidence first: three solid, sound heads running at full speed, clean zones swept in hours, and the archive's large, healthy bulk banked before the drive was ever even asked a single hard question. It wasn't until then that the settings changed (longer waits, managed resets) and the work turned toward the ground that had been failing.
Coax out what remains, then rebuild on something trustworthy
The soft head's ground was gone over at a slow walking pace, with the unit resting for some time and being power-cycled between stalls. Everything that was recovered was then copied to a new hard drive, with all the files restored in the archive's original folder arrangement, so that when the group plugged in this new disk drive, they were able to simply carry on with their filing as normal.
How it ended.
Almost everything survived: the newsletters, the meeting minutes and the great majority (but not all) of the scan images. A cluster from the earliest scanning sessions was located in the soft head's worst stretch and came back either damaged or not at all. These items are specifically named, file by file, in the report, allowing the group to easily identify the originals they need to ask to borrow again.
More like this in the casebook.
More cases from Portable drives.
Sound like your drive?
Each of the cases here is pointing in the same direction. Shut down the device, take the free diagnosis, and make your decision after that.