Win 10 / 11 · Maintenance

Run a disk health check before it's too late

Drives fail suddenly. A 10-minute monthly check tells you when to back up and replace before the data is gone.

Before you start

Medium risk. This change is reversible — you can return to the previous state without professional help or data loss.

Don't run chkdsk /r on a drive already reporting bad sectors. The full surface scan hammers a failing disk for hours and is quite capable of finishing it off. Copy the data somewhere safe first, then test.
S.M.A.R.T. reporting 'OK' is not a clean bill of health — it only means no threshold the manufacturer chose has been crossed yet. A reallocated sector count that is climbing at all is the warning, whatever the overall verdict says.
chkdsk tells you very little about an SSD. Wear levelling plus the maker's own tool — Samsung Magician, Crucial Storage Executive — are the checks that mean something there.
Clicking, ticking, or a drive that disappears and comes back: stop testing and copy the data off now. Every extra powered-on hour is spending down what you can still recover.

Task Manager → Performance → Storage → check 'Health' (or use CrystalDiskInfo for full S.M.A.R.T. data)
Any re-allocated sectors, pending sectors, or bad sectors = back everything up NOW
SSDs: check 'Wear Leveling Count' — under 90% is fine, below 50% plan a replacement
Run 'chkdsk C:' from an admin command prompt for a full scan (takes a while, do it overnight)

S.M.A.R.T. health reads Good with no re-allocated, pending or bad sectors, and chkdsk completes without unfixable errors.

Any bad/re-allocated sectors or failing S.M.A.R.T. attributes — back everything up immediately and plan a drive replacement.

Next stop: the related fixes below, or start a guided diagnosis.