The noise tells us a good deal about what has failed and how fast it is getting worse. A steady click is the heads failing to find their bearings; a scrape is contact with the platter itself. Both deteriorate every rotation, which makes the time between the first noise and the power switch the most important factor in the result.
$ cdr diagnose /dev/sda → Device: Seagate ST2000DM008 · 2 TB → Status: CLICKING — head stack failure → Client: confidential · Newmarket, Cambridgeshire $ cdr engineer-working → Donor heads matched: from stock · 0 days wait → Head-stack swap: complete → Imaging: 1.79 TB / 1.81 TB $ cdr verify → ✓ wedding_photos — 142,318 files → ✓ tax_records — 4,201 files → ✓ everything else — back home
Not a general warning of ill health. A precise mechanical report, and each sound points at a different part.
Everything else about this recovery was decided by the fault. This part was decided by you, and it matters more than anything we do afterwards.
A drive with failing heads has a finite number of rotations left before the damage becomes permanent. Each time it is powered on, the heads load, seek, fail and retry — and on a drive already making contact with the platter, that cycle removes magnetic coating that cannot be replaced.
The difference between a drive switched off at the first click and one that ran for a fortnight while somebody hoped it would settle is frequently the difference between a complete recovery and a partial one. There is no technique that restores a scored platter. The data that was on that track is gone in the physical sense.
So the answer to what should I do right now is always the same: switch it off, leave it off, and do not power it up again to check whether it is still doing it.
The freezer does not work. It is thirty-year-old advice from an era of different drive engineering, and on a modern disk it introduces condensation into a sealed assembly with sub-micron head clearances. It has ended recoveries that would otherwise have been routine.
The most misunderstood part of mechanical work, and the thing that most affects how long a job takes.
Replacing a failed head stack means taking one from a working drive. People reasonably assume that means buying the same model number, and it does not. Two drives with identical labels can carry different head configurations, different platter counts and different firmware revisions, and a mismatch on any of those means the replacement stack cannot read the surface it sits above.
A usable donor matches on family, head map and firmware generation. We hold stock of the families this region sends most — Barracuda and WD Blue generations, the 2.5-inch Toshiba MQ range, and a deliberate reserve of older IDE and early SATA families because fenland industrial equipment keeps producing them.
Where a donor has to be sourced, that changes the timescale rather than the outcome, and you are told at the diagnostic rather than discovering it in week three.
A clicking or grinding drive is mechanical-class work from £300 +VAT, agreed in writing after the free 48-hour diagnostic. Drives forming part of an array start from £500 +VAT.
Because clean-air work and donor parts commit real cost the moment they begin, mechanical jobs carry a 50% deposit with the balance payable only on success. You are told that applies before agreeing to anything.
No, it is the same fault. The drive attempts to initialise, fails, and gives up rather than continuing to retry. The heads have still made contact and the outcome depends on how many times that cycle has run.
Please do not. Software scanning a mechanically failing drive keeps it powered and seeking for hours, which is precisely the load that turns a partial fault into a total one. If it still mounts, copy your most important files immediately and stop — do not run a scan.
Honestly, that depends on how long it ran. A drive switched off promptly usually recovers substantially complete. One that ran for weeks may come back partial, and we will tell you which we expect at the diagnostic rather than afterwards.
No. A drive fitted with donor parts is a vehicle for reading the data off once. It is not a disk to put back into service, and anyone offering that is describing something that does not exist.
Five to ten working days where a donor is in stock. Longer where one has to be sourced, and you will know which applies at the diagnostic.
Frequently yes. Age affects donor availability rather than recoverability, and we keep older families in stock precisely because this region sends industrial and laboratory equipment running them.
Every repair is backed by a 10,000-piece donor store, opens with a free written diagnostic, and on most jobs is no fix, no fee. Begin today.