We're taking on new cases now · Weekdays, 9am–5:30pm Need it back quickly? Call 0800 6890668
MKDR Milton Keynes Data Recovery 0800 6890668 Get it diagnosed
MKDR / Casebook / A Grinding Hum We All Blamed on the Compressor

NAS & RAID · bench record · MKD-2025-4928

A Grinding Hum We All Blamed on the Compressor.

The unit had sat in the back room of a Leighton Buzzard framing shop since 2018 — on top of the guillotine cabinet, out of the way and out of mind — until counter staff began mentioning a grinding hum we all blamed on the compressor. Ten days later the shared folder would not open and the front light held solid red. WD offered a warranty replacement; what was inside it stayed, as ever, the owner’s business.

Confirmed by the customer All cases anonymised

Same fault? Start here.
0800 6890668

Without the jargon.

Strip away the app and the badge and a My Cloud amounts to a single ordinary hard disk with a small computer sitting on top of it. When that disk stops answering, every layer above goes down at the same moment — shares, web page, status light — because all of them rest on the same storage. One disk, one door, and nothing behind it but itself. The noise the shop had half-registered for a fortnight was the disk saying so, and a single-bay box keeps no second copy of anything: whatever that one disk holds is the entire estate.

The kit this job needed.

See how a job runs here →
KitWhy it was usedWhat it gives us
DeepSpar Disk Imager 4Read it head by head from a fresh map once the swap was doneMaps the heads, then images each — resets, timeouts and power cycles controlled
PC-3000 Data ExtractorHeld timeouts and resets short while a worn disk did the readingThe PC-3000 imager, driven from the firmware side of the tool
UFS Explorer RAID RecoveryTook apart the box’s Linux volume stack so the shares returned as sharesUnpicks a NAS layer by layer instead of treating it as one flat array

What happened on the bench.

01

Out of the case, and judged on its own merits

Stripped out of the enclosure, the 3TB disk settled it quickly: its heads were worn beyond dependable reading, and the grinding note was them failing to do the job. The appliance’s software sees only storage that has stopped replying — so shares, dashboard and light went out together.

02

New heads first, then work out what the wear had cost

A donor head stack from a matching drive put the mechanism back in order, and mapping came before any serious reading. One region had paid the price of the shop’s patience: hour after hour the till PC had gone on syncing its order folder to a box whose heads were giving out, and that stretch of surface showed it. Left alone, an appliance has no idea when to stop trying.

03

Read the image as the Linux box it actually is

WD builds the storage as stacked Linux volumes, which is not something a tool expecting NTFS will make any sense of. Read on its own terms, the image opened out into the shop’s shares just as the counter knew them — job folders, customer artwork, the order book — instead of ten thousand files with numbers for names.

The outcome.

Everything on the shop’s list came back on new media, with one honest exception. A run of older customer artwork scans sat inside the worn band and could not be read; the report lists them by name. The box went for its warranty replacement empty.

The lesson: An appliance left unattended will grind a failing disk for weeks. New noises are worth chasing, and a one-disk box is not a backup.

Does that sound like your device?

One lesson runs through the lot: cut the power, let us look for free, then decide with the facts in hand.

0800 6890668