Most drives announce themselves for weeks before they stop — long pauses, stalling copies, freezes that clear on their own. It reads as an old computer and it is usually a dying disk. That phase is when almost everything is still recoverable, often by you, for nothing.
$ cdr diagnose /dev/sda → Device: Seagate ST2000DM006 · 2 TB → Status: NOT DETECTED — service area / firmware corruption → Client: confidential · Cambridge, Cambridgeshire $ cdr engineer-working → Service area: firmware repaired → Drive: re-initialised · detected → Imaging: 1.82 TB / 2.00 TB (91%) $ cdr verify → ✓ documents — 64,820 files → ✓ photos — 51,300 files → ✓ drive visible again — recovered
Mechanical failure is a process rather than an event. These are the signs, in the order they usually appear.
Files take longer to open, the machine hangs briefly on save, boot time creeps up. The drive is retrying bad sectors repeatedly before eventually succeeding.
A new whine, an occasional tick, a sound on start-up that was not there before. Not yet the click of death, but the same components beginning to struggle.
Documents that will not open, photographs with corrupted blocks, folders that appear empty. The file system is intact and the sectors beneath it are not.
Not detected, or detected and unreadable. By this point the warning phase has been running for some time and the recoverable window has narrowed.
Look at the SMART data. On Windows, CrystalDiskInfo reads it without installing anything permanent; on a Mac, DriveDx does the same. Reallocated sector count, pending sectors and uncorrectable errors are the three fields that matter, and any of them climbing means the drive is failing whatever else it says.
Then check whether the drive still mounts. If it does, copy your most important files off immediately — now, before reading further. A failing drive that still reads is a gift and it will not last.
Then stop using it. Not later, not after you have finished what you were doing. Every hour a degrading drive stays in service spends readable sectors you may want back.
Do not run chkdsk or a repair utility. These write to the disk they are checking, and on a drive with failing sectors that write frequently lands somewhere it cannot be completed. It is the most common way a recoverable drive becomes a partial one.
Single drive recovery from £300 +VAT, agreed in writing after the free 48-hour diagnostic and unchanged afterwards. Array members start from £500 +VAT.
Most jobs run no fix, no fee. Where clean-air work or donor parts are needed those costs commit at the start, so a 50% deposit applies with the balance payable only on success.
Unknowable, and that is the honest answer. It could be months or it could be tomorrow. What is certain is that a drive reporting reallocated sectors will not improve. Copy everything off today and replace it.
Sustained slowness with no other cause is frequently a drive retrying bad sectors. Each retry succeeds eventually, which is why you see delay rather than failure. It is an early warning, not a performance problem.
Classic sector-level degradation. The file system still knows where everything is, and the sectors holding some of it can no longer be read. Recoverable, and considerably more so if the drive stops being used now.
If it still mounts and you have a spare disk, a sector-level clone with a tool that skips errors rather than retrying them is reasonable. Do not use a tool that hammers bad sectors — that accelerates the failure.
Urgent. A new mechanical sound means components are moving in ways they should not. Switch it off rather than finishing what you were doing.
Yes, at the free diagnostic and in writing. Fault class, what we expect to recover, and one fixed figure. If the answer is that it is not worth attempting, you get that instead of a quotation.
Behind every repair sits a 10,000-part inventory, a free written diagnostic to open, and no fix no fee on most jobs. Begin your recovery today.