Signs of a failing hard drive

Most hard drives run quietly for years: Backblaze's drive statistics for the first quarter of 2026 put the annualised failure rate across 341,263 drives at 1.24%. When a drive starts to struggle it usually gives some hint, though not always: Seagate's LaCie knowledge base cautions that drives do not typically give an indication of any problems prior to failure. Any one sign can have an innocent cause, from a tired cable to a fan; what deserves attention is a pattern that repeats or combines with other signs.^[1][2]

New noises: clicking, beeping and grinding

Some noise is ordinary. Seagate notes that every drive makes a certain amount of noise while running, and that a soft chattering or clicking while data is read and written is normal. It also points out that noise from a computer's power-supply cooling fans is often mistaken for hard drive noise. What merits attention is a sound that is new for your drive, louder than before, or continuous rather than tied to activity.^[3]

Clicking divides into two very different sounds. Seagate's LaCie knowledge base calls clicking usually the first indicator of a problem with an external hard disk, then separates soft clicking, which can simply be the normal sound of the drive working, from hard clicking: a very noticeable sound akin to metal-on-metal impacts, which usually indicates physical failure. A drive that has started hard clicking should not be kept running to see what happens next.^[1]

Our clicking hard drive guide walks through what to do, and what not to do

Beeping is usually about power rather than mechanics. Seagate's guidance attributes most external drive beeping to insufficient electrical power, especially with bus-powered drives that have no plug-in adapter, and notes that drives can need a little more power as they age. A different cable, a rear port on a desktop PC or a powered hub often settles it.^[3]

Grinding or clanking is the sound to take seriously. Seagate describes an excessive grinding or clanking noise as something that may point to a physical problem, and recommends replacing a drive that fails its diagnostic test.^[3]

Files open slowly, or the drive feels sluggish

Slowness has boring causes first: a nearly full drive, a large copy or backup in progress, or indexing after a big change. The concerning version is a drive that stays slow once those are ruled out, because it may be re-reading sectors it cannot read first time. How a value changes matters more than any single reading: Backblaze notes that most of the SMART stats it tracks are cumulative, so it watches sector counts over time rather than trusting one snapshot.^[4]

Files corrupting or vanishing

One file that will not open usually means a bad download, an interrupted save or a faulty memory card. Several files, in different folders, failing the same way points at the drive. The mechanism is unglamorous: an uncorrectable sector is one the drive can no longer read reliably, so anything stored on it may open slowly, partly or not at all, which is why Backblaze counts Uncorrectable Sector Count (SMART 198) among the attributes that prompt it to investigate a drive. Files that were fine last month and are broken now deserve a drive question, not a software question.^[4]

The drive has disappeared from the computer

A drive that vanishes from File Explorer or Finder, or never appears at all, is alarming precisely because it looks sudden. Start with the dull explanations: Seagate's LaCie knowledge base lists a failure in the external power supply or power cable among the causes of a clicking drive that is not working, and suggests trying a different one. A different port, cable or power adapter restores many drives to view, and if the drive reappears after a change like that, the drive itself was never the problem.^[1]

Liquid is its own case with its own rules. If a drive has been wet, our water damage FAQ covers what to do first and what to leave switched off.

Can a water damaged hard drive be recovered?

External drive not detected: the full checklist

SMART warnings

SMART (Self-Monitoring, Analysis and Reporting Technology) is the monitoring system built into a drive that reports on attributes of its own state. Backblaze, which watches SMART data across its own fleet, names five attributes it treats as reasons to investigate a drive: Reallocated Sectors Count, Reported Uncorrectable Errors, Command Timeout, Current Pending Sector Count and Uncorrectable Sector Count. A warning against any of those is a reasonable prompt to check your backups, and to look for any of the physical signs above.^[4]

Two numbers from Backblaze's SMART analysis, published in 2016, keep the warnings in proportion. 76.7% of its failed drives had at least one of those five attributes above zero, against 4.2% of drives still running, so a warning really is more common among drives that fail. But 23.3% of failed drives showed no warning at all, which is the other half of the lesson: a clean SMART report is reassurance, not a guarantee.^[4]

One sign or a pattern?

Any single sign can have an innocent cause, and most drives never develop any of them. What warrants action is a pattern: a new sound plus a sluggish drive, a SMART warning plus a file that will not open, or the same sign recurring over several weeks. Outright failure is also the exception rather than the rule: Backblaze's fleet data recorded a quarterly annualised failure rate of 1.24% for the first quarter of 2026, and patterns are how you spot the exceptions.^[2]

If you are not sure which of these your drive is showing, guided help asks a few questions and points you to the right guide for what you are seeing.

Not sure which sign you are seeing? Use guided help