9 Warning Signs Your Hard Drive Is About to Fail

The nine signs, loudest first

Physical decline announces itself through sound and behavior long before the drive stops spinning. The textbook symptom is the click: a rhythmic tick or knock from the case, meaning the head is retrying a read it cannot complete — a single soft click at power-on is normal, a repeating pattern is not. Next come the stalls: opening a folder takes seconds, the disk sits at 100% in Task Manager while copying files that used to fly, and the whole machine freezes the moment it hits a bad sector.

Data-level symptoms are quieter but just as damning: files that refuse to open, arrive corrupted or disappear between reboots; chkdsk reporting bad sectors on every run; the drive occasionally missing from the boot list or File Explorer until a reboot brings it back; and a boot-time warning that SMART status is bad. If two or three of these line up, treat the drive as terminal and stop testing — every hour of full-surface scanning is an hour the heads spend grinding over a platter that is coming apart.

• A repeating click, tick or knock from the case — the head retrying a failed read
• Freezes and 100% disk usage on perfectly ordinary file operations
• Files that vanish, refuse to open, or arrive corrupted
• chkdsk finding bad sectors on every run, not once a year
• The drive intermittently missing from the boot list or File Explorer
• A boot-time warning of bad SMART status — this one you never ignore

What SMART actually tells you

SMART is the drive's own logbook, and Windows shows you almost none of it — the full picture comes from a reader like CrystalDiskInfo or the drive maker's own tool. Three attributes matter most: 05 (Reallocated Sectors Count — the drive has already moved data off damaged areas), C5 (Current Pending Sectors — waiting to be rechecked) and C6 (Uncorrectable Sectors — reads that failed outright). A couple of reallocated sectors on a five-year-old drive is aging; a number that climbs month over month, or any growth in C5/C6, is a countdown. Attribute C7 (UltraDMA CRC errors) is the exception: it usually points to a bad SATA cable, not a bad disk.

The correct order of operations

Backup first, everything else second — and copy the irreplaceable files out before you image anything, because imaging reads every sector including the dying ones. Copy as fast as you can, stop writing new data to the drive, and do not run chkdsk /r, defrag or "repair" utilities on a failing disk: they hammer the very surface you are trying to rescue. Once the data is safe, replace the drive — the SSD upgrade suggests itself here — and only then run diagnostics on the old disk out of curiosity, not for rescue. And if the data is worth real money, stop at the first click and take the drive to a lab: every powered-on hour reduces what they can recover.

Questions and Answers

Read next