How to Check Hard Drive Health (Windows, Mac & SSDs)
Most drives warn you before they die — but only if you know where to look. Every modern hard drive and SSD tracks its own condition through a system called S.M.A.R.T. (Self-Monitoring, Analysis and Reporting Technology). Reading it takes a few minutes and costs nothing. This guide shows you how on every platform, and — more importantly — how to tell which numbers are a real warning.
If your drive is already clicking, beeping, or disappearing, skip the tests. Running scans on a failing drive adds stress and can destroy what’s left. Back up what you can, power it down, and read our clicking-drive guide.
The fastest check on Windows: CrystalDiskInfo
CrystalDiskInfo is a free, widely used tool that reads SMART data and turns it into a simple verdict.
- Download it from the official Crystal Dew World site (choose the standard edition).
- Open it. Each drive appears as a tab at the top.
- Read the Health Status box:
- Good (blue) — no SMART warnings.
- Caution (yellow) — one or more warning attributes are non-zero. Back up now.
- Bad (red) — the drive has crossed its own failure threshold. Back up immediately and replace it.
Don’t stop at the color. Scroll the attribute table and check the specific values described below — a drive can read “Good” while a key counter is quietly climbing.
Built-in Windows options (no download)
- Windows 11 Settings: Settings → System → Storage → Advanced storage settings → Disks & volumes → select the drive → Properties. NVMe SSDs show Drive health, including estimated remaining life, available spare, and temperature.
- PowerShell: open PowerShell as administrator and run
Get-PhysicalDisk. The HealthStatus column reads Healthy, Warning, or Unhealthy. For more detail, runGet-PhysicalDisk | Get-StorageReliabilityCounter | Format-List. - CHKDSK (
chkdsk C: /scan) checks the file system, not the hardware. It’s useful for logical errors, but a clean CHKDSK does not mean the drive is physically healthy.
On a Mac
- System Information: hold Option, click the Apple menu → System Information → Storage (or NVMExpress / SATA). Look for S.M.A.R.T. Status: Verified. “Failing” means back up immediately.
- Disk Utility → First Aid checks and repairs the file system. Like CHKDSK, it isn’t a hardware test.
- For the full picture: install smartmontools (via Homebrew:
brew install smartmontools) and runsmartctl -a disk0in Terminal to see every SMART attribute.
On Linux, the same smartctl -a /dev/sda command works, and smartctl -t short /dev/sda runs the drive’s built-in short self-test.
Manufacturer tools
Drive makers ship their own diagnostics, which can run the drive’s internal self-tests and read vendor-specific data:
- Seagate: SeaTools
- Western Digital / SanDisk: WD Dashboard
- Samsung SSDs: Samsung Magician
- Crucial SSDs: Crucial Storage Executive
- Kingston SSDs: Kingston SSD Manager
- HP, Dell, Lenovo laptops: the built-in hardware diagnostics at boot (on HP, a failed test shows as Hard Drive Short DST Check Failed)
These also show firmware versions — worth checking, because some SSD failures trace back to known firmware bugs that a vendor update fixes.
The SMART attributes that actually matter
Large-scale drive studies (Backblaze’s public reliability data is the best known) found a handful of SMART attributes strongly correlated with failure. These are the ones to watch:
| ID | Attribute | What a non-zero value means |
|---|---|---|
| 05 | Reallocated Sectors Count | The drive has found bad sectors and swapped them for spares. Any growth is a warning. Explained here. |
| C5 (197) | Current Pending Sector Count | Sectors the drive can’t currently read — data there is at risk right now. Explained here. |
| C6 (198) | Uncorrectable Sector Count | Sectors that failed even after retries. |
| BB (187) | Reported Uncorrectable Errors | Read errors the drive couldn’t fix with error correction. |
| C7 (199) | UltraDMA CRC Error Count | Data corrupted in transit — usually a bad cable or port, not the drive itself. |
For SSDs, add these:
- Percentage Used / Wear Leveling Count / Media Wearout Indicator — how much of the flash’s rated write endurance is consumed.
- Available Spare (NVMe) — reserve blocks left. A drop below the threshold is serious.
- Total Host Writes / Data Units Written — compare against the drive’s rated TBW.
A single reallocated sector on an old drive isn’t an emergency. A count that’s rising week to week, or any pending or uncorrectable sectors, means the drive is degrading — treat it as a countdown.
Signs you can see and hear
SMART isn’t the only signal. Take these seriously even if the software says “Good”:
- Clicking, grinding, or repetitive beeping (see hard drive beeping)
- Freezes when opening or saving files
- Files that won’t open or come back corrupted
- Errors like Data error (cyclic redundancy check) or an I/O device error
- An external drive that keeps disconnecting
Our 7 warning signs guide covers these in more depth.
What to do with the results
Healthy: great — now make sure you have a real backup (the 3-2-1 rule is the simplest that works) and re-check every few months.
Caution / rising counts: back up your most important files first, before doing anything else. Don’t run defragmenters, long surface scans, or “repair” utilities — they add hours of stress to a drive that’s already weakening. Then replace the drive.
Bad / failing / not detected: stop using it. If it still mounts, copy only what matters most. If it doesn’t — or if it’s making noise — power it down. That’s the point where professional recovery has the best odds, and every extra power cycle lowers them.
The bottom line
Checking drive health takes five minutes: CrystalDiskInfo on Windows, System Information or smartctl on a Mac, plus your drive maker’s tool. Ignore the overall color and watch the counters — reallocated, pending, and uncorrectable sectors tell you how much time you have left.
If a check came back bad and the drive is already struggling, our hard drive recovery service recovers failing drives for one flat price, with no charge unless we get your data back.