How to Check Hard Drive Health: 2026 Guide
Hard drives don't send a calendar invite before they fail. One day everything's fine, the next your PC is stuck at a boot loop or your external drive shows up as "RAW" in Disk Management.
I've pulled dying drives out of gaming rigs, NAS boxes, and laptops for years. The pattern is almost always the same: the warning signs were sitting there in plain sight, weeks or months before the actual crash.
This guide walks through exactly how to check hard drive health on Windows, macOS, and Linux, what the numbers actually mean, and which tools are worth your time versus which ones are just noise.
What "Hard Drive Health" Actually Means
Every modern drive, mechanical HDD or solid-state SSD, runs a background diagnostic system called SMART (Self-Monitoring, Analysis, and Reporting Technology). It's built into the drive's firmware, not your operating system.
SMART tracks dozens of internal attributes and reports them as raw values and normalized scores. Reading those numbers correctly is the whole game.
For traditional spinning hard drives, the attributes that matter most relate to mechanical wear:
- Reallocated Sector Count: bad sectors the drive has swapped out for spares
- Current Pending Sector Count: sectors waiting to be reallocated, a strong early warning sign
- Spin Retry Count: how many times the motor struggled to spin up
- Seek Error Rate: how accurately the read/write head is tracking
SSDs fail differently because there are no moving parts. Instead, you're watching flash memory wear out:
- Media Wearout Indicator or Percentage Used: how much of the drive's rated write endurance is consumed
- Total Bytes Written (TBW): cumulative data written over the drive's life
- Uncorrectable Error Count: data errors the controller couldn't fix internally
- Available Spare: the percentage of backup NAND blocks still usable
Both drive types also report temperature, power-on hours, and power cycle count. Temperature especially matters more than most people realize, since heat accelerates wear on both platters and NAND cells.
The Best Tools for Checking Drive Health
You don't need expensive software. Most of the tools that actually matter are free, and a couple are already built into your operating system.
| Tool | Best For | Setup Effort | Depth of Data | Cost |
|---|---|---|---|---|
| CrystalDiskInfo | Windows users wanting fast, readable SMART data | Very low, portable exe available | High, color-coded warnings | Free |
| Windows PowerShell (Get-PhysicalDisk) | Quick health checks without installing anything | Low, built-in | Medium | Free |
| smartctl (Smartmontools) | Linux, macOS, and scripting/automation | Medium, requires install | Very high, raw SMART access | Free |
| Manufacturer tools (Samsung Magician, WD Dashboard, Seagate SeaTools) | Brand-specific extras like firmware updates | Low to medium | High for that brand only | Free |
| HD Sentinel | Long-term monitoring with alerts and predictions | Medium, more configuration | Very high | Freemium |
Pro Tip: Don't trust a single "Good" or "Healthy" verdict from any one tool. CrystalDiskInfo's overall status can say "Good" while a specific attribute like Reallocated Sector Count is already climbing. Always check the individual attribute values, not just the summary badge.
Windows: The Fastest Path
CrystalDiskInfo remains the go-to for most Windows users because it translates raw SMART data into plain language and color codes (blue for healthy, yellow for caution, red for trouble).
If you'd rather skip installing anything, PowerShell already has what you need:
Get-PhysicalDisk | Select-Object DeviceId, FriendlyName, MediaType, HealthStatus, OperationalStatus
For deeper reliability counters on modern drives:
Get-StorageReliabilityCounter -PhysicalDisk (Get-PhysicalDisk) | Select-Object DeviceId, Wear, Temperature, ReadErrorsTotal, WriteErrorsTotal
The older wmic diskdrive get status command still works on most systems, but Microsoft has been phasing wmic out, so I'd lean on PowerShell going forward.
macOS: Built-In but Limited
Disk Utility shows a basic S.M.A.R.T. status (Verified or Failing) at the bottom of the drive info panel. It's a coarse pass/fail, not detailed attribute data.
For real depth on a Mac, install Smartmontools through Homebrew:
brew install smartmontools
sudo smartctl -a /dev/disk0
Linux: Full Control
Smartmontools is the standard here too, and it's usually a single install away:
sudo apt install smartmontools
sudo smartctl -H /dev/sda
sudo smartctl -a /dev/sda
The -H flag gives you the quick pass/fail health assessment. The -a flag dumps every attribute the drive reports, which is where the real diagnostic value lives.
Real-World Testing: How I Actually Vet a Drive
When I buy a used drive, whether it's a refurbished NVMe SSD or a pulled-from-a-server HDD, I run through the same checklist before trusting it with real data.
- Run smartctl or CrystalDiskInfo first and screenshot the full attribute list as a baseline
- Check Power-On Hours against the seller's claimed usage; mismatches are a red flag
- Look at Reallocated Sector Count and Pending Sector Count specifically, since these rarely lie
- For SSDs, check Percentage Used or Media Wearout against the drive's rated TBW
- Run a short SMART self-test (
smartctl -t short /dev/sda, then check results after a few minutes) followed by a long test overnight if the drive passes - Benchmark with CrystalDiskMark or a similar tool and compare results against the drive's published spec sheet
- Monitor temperature under sustained load, not just idle, since thermal throttling and heat-related errors only show up under stress
In my testing, drives with even a handful of reallocated sectors tend to accumulate more within a few months. It's rarely a one-time event.
Pro Tip: Run the SMART self-test rather than relying only on passive attribute readings. Passive SMART data can look clean while a self-test actively exercises the drive's read heads or NAND controller and surfaces problems the idle report misses.
The Honest Limitations Nobody Mentions
SMART is useful, but it's not a crystal ball. It's worth being upfront about where it falls short.
A meaningful share of drive failures happen with no prior SMART warning at all. Controller failures, firmware bugs, and sudden electrical faults don't always show up as gradually worsening attribute values.
Different manufacturers also implement SMART inconsistently. A "raw value" for the same attribute number can mean something different on a Seagate drive versus a Western Digital drive, which is why tools like CrystalDiskInfo maintain manufacturer-specific interpretation tables.
Thresholds are conservative by design too. Drive makers set "failure" thresholds to avoid false positives, so by the time an attribute crosses into red territory, the drive may already be in serious trouble.
External USB enclosures complicate things further. Some bridge chipsets don't pass SMART data through at all, so a perfectly good external drive might just show "no SMART information available" instead of an actual status.
And no health check replaces backups. Even a drive reporting perfect SMART numbers can fail from a power surge, a physical drop, or simple bad luck. A healthy drive does not explain every slowdown; our RAM pricing and upgrade guide includes a workflow for checking memory use before spending.
Setting Up a Practical Monitoring Routine
Checking once isn't enough. Here's the setup I actually recommend for ongoing peace of mind.
- Install CrystalDiskInfo (Windows) or Smartmontools (macOS/Linux) and set it to run at startup
- Enable automatic email or notification alerts if your tool supports it, so you're not manually checking every week
- Schedule a monthly full SMART report export, even just a saved screenshot or text log, so you can compare trends over time
- Run a full disk self-test quarterly on drives holding important data
- Pair health monitoring with an actual backup schedule, not just alerts, since 3-2-1 backup principles matter more than any SMART reading
- Replace any drive showing rising Reallocated Sector Count, Pending Sector Count, or a Percentage Used above 90% on an SSD, regardless of what the overall status badge says
On a NAS running something like TrueNAS or Synology DSM, built-in SMART scheduling and email alerts handle most of this automatically once configured, which is worth setting up the day you deploy the box rather than after a scare.
Frequently Asked Questions
Can a drive show a healthy SMART status and still fail soon? Yes. SMART tracks known wear patterns, but sudden controller failures, firmware bugs, and physical shocks can take out a drive with no warning in the attribute data.
How often should I check hard drive health? A quick check monthly is reasonable for most users, with continuous background monitoring turned on for anything holding data you care about, like a NAS or your primary boot drive.
Does running SMART checks or self-tests wear out the drive faster? No. Reading SMART attributes is passive and adds negligible wear. Even active self-tests use a tiny fraction of the read/write cycles compared to normal daily use.
Which SMART attributes matter most for an SSD versus an HDD? For HDDs, watch Reallocated Sector Count and Pending Sector Count. For SSDs, focus on Percentage Used, Media Wearout Indicator, and Available Spare, since these track flash wear rather than mechanical failure.
Where This Is Heading
Drive health monitoring is getting smarter on its own. NVMe SSDs already expose far richer telemetry than old SATA SMART ever could, including detailed thermal throttling events and per-namespace wear data.
Cloud storage providers and NAS vendors are increasingly layering machine-learning failure prediction on top of raw SMART data, catching patterns that a fixed threshold would miss entirely.
For now, the fundamentals haven't changed: check your drive's attributes regularly, take Reallocated Sectors and rising Percentage Used seriously, and never let a single "Good" badge replace an actual backup. The tools get better every year, but the discipline of actually looking is still on you.
