Is this drive about to fail?
What warns you, and what only looks like a warning.

First language: what you can hear

These are the signs you notice without any software, and they are the ones worth acting on immediately.

If none of that is happening, you are in the ordinary case, and the rest of this page is about reading the second language properly — which is where most people scare themselves for nothing.

Second language: what the drive counts

Every drive keeps dozens of internal counters. Most of them are for the people who repair drives. Three of them say almost everything to the person who just works on it.

The counter	What it actually tells you
Damaged sectors	Pieces of the drive that can no longer be read, which the drive has quietly swapped for spares. This is the reliable one. Zero is normal. A handful that appeared years ago and never moved is history. A number that is climbing is the clearest warning a drive can give. What they are, in detail →
Hours of use	Age, not danger. Sixty thousand hours is about seven years of continuous running, and plenty of drives pass that in perfect health. It does not say the drive will fail — it says this is no longer the place for your only copy of something.
Errors on the cable	Not the drive at all: the connection. The computer simply reads the data again and nothing is lost. Three errors in seven years are nothing. Three this week are a cable giving up, and a new cable fixes it.

What looks alarming and is not

This is where most of the worry comes from, and almost all of it is misplaced. Half the tools that read these numbers show them raw and let you panic on your own.

What you see	Why it is not the problem
A wear counter at 98 out of 100	It measures expected consumption, not a fault. With zero damaged sectors, a drive is not failing because a number dropped by two.
A frightening temperature	Almost always the highest ever recorded, not the temperature now. One hot afternoon three years ago is written down forever.
Enormous numbers — 38,799,312 “errors”	Manufacturers pack several counters into one number. Read literally it means nothing. On some drives that particular field is not an error count at all.
“S.M.A.R.T. status: OK”	The opposite trap: this is reassuring and should not be. That self-check keeps passing until the last useful moment. It is information, not a guarantee.
The drive is seven years old	Age alone predicts nothing about tomorrow. It changes what you should keep on it, not how likely it is to stop tonight.

A warning that goes off for nothing is worse than no warning at all. After the second false alarm nobody looks — and then the one that mattered goes unread too.

So what do I actually do?

Three situations, and only one of them is urgent.

Notice that in two cases out of three the right answer is do nothing and stop worrying. That is the part no one tells you, and it is the reason the third case is worth believing when it comes.

Where a program helps, and where it does not

You can read all of this yourself with a free tool, once, today. The difficult part is not reading it: it is knowing which numbers matter, and noticing when one of them starts moving.

NanoTracker reads the same counters and says what they mean in plain words — and says not readable instead of inventing a verdict when a drive enclosure will not pass the question through. It also takes a reading a day, so it can compare today with last month rather than showing you a single instant. The difference between “eight damaged sectors” and “eight since last month” is the whole difference.

What it does not do is predict anything, and it will not pretend otherwise. How it reports drive health →

Common questions

Can any software tell me a drive is about to fail?

No. Every tool, including ours, reads counters the drive itself keeps. Those counters describe what has already happened, not what will happen. A drive that has started replacing damaged sectors is telling you something real and worth acting on — but that is a measurement, not a prediction, and anyone selling it as a prediction is selling you fear.

My external drive makes a clicking noise. Is that bad?

A repeated click, a drive that disappears and comes back, or a computer that freezes when you open a folder on it: treat all three as the drive asking you to stop. Copy off what you cannot lose, from the most important file down, and do not keep testing it. Every extra hour of use is a chance you do not need to take.

The drive says its S.M.A.R.T. status is OK. Does that settle it?

No. That self-check keeps saying OK until the last useful moment: it is designed to report a drive that has already failed its own thresholds, not one that is deteriorating. It is information, not a guarantee, which is why the counters underneath matter more than the verdict on top.

How old is too old for an external drive?

There is no age at which a drive becomes dangerous. Sixty thousand hours is about seven years of continuous running, and a drive can pass that with nothing wrong. What age changes is not the risk of this drive failing tomorrow, it is how sensible it is to keep your only copy of something on it.

Does a high temperature mean the drive is dying?

Usually you are looking at the highest temperature ever recorded, not the current one. One hot afternoon three years ago stays written down forever. The temperature now is worth a glance; the peak from the past is history.