How to Check SSD and HDD Health in Windows
Checking SSD and HDD health in Windows takes about five minutes, but only if you know which reading to act on. The status Windows shows by default answers a narrower question than most people assume: is the disk visible to the operating system right now. A drive can be accumulating unreadable sectors every hour and still be reported as OK, because the firmware only raises its own failure flag once a vendor threshold has been crossed. I have seen a laptop disk die outright a few weeks after Disk Management showed a clean healthy status. The PowerShell commands come first, then the SMART attributes worth watching, then the symptoms that mean stop using the drive today.
Why a healthy report from Windows is not proof
The answer comes from the drive itself. Every modern disk carries a table of counters that the firmware updates constantly, and Windows shows you one summary of that table. The summary flips to a warning only when a vendor defined threshold is crossed, and until then it reads OK.
The quickest check most people find first is the old Windows Management Instrumentation command. Run it in Command Prompt or PowerShell:
wmic diskdrive get status
The output is one word per physical disk, and in practice it reads OK for almost everything, including drives already throwing errors. WMIC is also on the way out: Microsoft has deprecated the tool and recent Windows 11 builds no longer ship it, so use the PowerShell route below instead.
How to check SSD and HDD health with PowerShell
PowerShell exposes the summary and the counters separately. Right click Start, choose Terminal (Admin) or Windows PowerShell (Admin), and run:
Get-PhysicalDisk | Select-Object FriendlyName, HealthStatus, OperationalStatus, MediaType
You get one line per drive. HealthStatus is the familiar Healthy, Warning, Unhealthy or Unknown. OperationalStatus covers the drive's day job, so OK there with a Warning in HealthStatus means the disk is online and already flagged. MediaType tells you whether it is an SSD or an HDD, and the failure signatures differ. Unknown is common on external drives, where the USB bridge does not pass SMART data through.
For the numbers behind that summary, use the reliability counter. This is the command I run first on any machine that is behaving oddly:
Get-PhysicalDisk | Get-StorageReliabilityCounter | Format-List
It prints properties such as PowerOnHours, Temperature, Wear and ReadErrorsUncorrected. Uncorrected read errors above zero matter most, because those are sectors the drive could not fix and told the host about. PowerOnHours is age in hours, and 8,760 hours is a year of continuous power, so a mechanical disk past 30,000 hours deserves attention whatever the status says.
Two caveats: a USB enclosure or RAID controller often returns no data at all, and you get one block per disk, so check the DeviceId before you trust the numbers.
The SMART attributes that actually matter
Storage Reliability gives totals. For how the drive counts its own wear, read the raw SMART attributes. Attribute 5 counts bad blocks already remapped, attribute 197 counts sectors it could not read on demand, attribute 198 counts sectors that failed and could not be rescued, and attribute 9 is the power on hours counter.
| ID | Attribute | Good value | What a bad reading means |
|---|---|---|---|
| 5 | Reallocated Sectors Count | 0 | Bad blocks the drive has already remapped. A count that keeps climbing means the drive is losing surface area or flash cells. |
| 9 | Power-On Hours | Not a pass or fail | Age in hours. A disk past 30,000 hours on a consumer machine is a candidate for replacement regardless of the other numbers. |
| 187 | Reported Uncorrect | 0 | Errors the drive could not correct and reported to Windows. Any non-zero value is a warning sign. |
| 197 | Current Pending Sector | 0 | Sectors the drive cannot read right now. One or two is a warning, a rising count is an active failure. |
| 198 | Offline Uncorrectable | 0 | Sectors that failed and could not be remapped. A non-zero value usually means the end is close. |
A pending sector count of 1 or 2 means the drive tried to read a sector, failed, and set it aside, so the file that lived there may already be damaged. A rising reallocated sector count is worse, because that number only moves upward and each new entry is a block the drive can no longer use. Treat a reallocated count that grows between checks as a failing drive.
NVMe drives do not use these attribute numbers. They report a health log instead, with fields such as critical warning, percentage used and available spare. The same PowerShell reliability command works on them, and the Wear field maps to percentage used, the closest thing to a straight line of time remaining.
How HDD and SSD failures feel different
A mechanical drive usually tells you it is unhappy before the file system notices. Head arms clicking, grinding or rattling from the drive bay is the drive trying to find a sector it cannot read. The disk then stalls for ten or thirty seconds at a time, Task Manager shows Disk pegged at 100 percent, and Windows logs errors in Event Viewer under Windows Logs and System from the source named disk, often event ID 51 or 153.
An SSD has no moving parts, so it rarely makes a sound. It fails in one of three ways. The controller can drop the drive into read only mode, where the volume shows as RAW and the drive reports that the media is write protected, because the controller refuses to write to worn cells. The drive can vanish from the BIOS or UEFI device list entirely, which means the controller is no longer talking to it. Or it can report healthy while quietly returning corrupt data, which is the worst of the three.
Failures are not limited to the built-in disk. A failing USB drive usually shows up in Disk Management with a capacity of 0 bytes and a status of no media, which looks very different from a drive that is simply not plugged in. If you are chasing a drive that disappeared, fixing a USB drive that is not showing up has the recovery steps, and the same warning applies: copy what you can read first.
Firmware updates and TRIM
Firmware is worth checking, and the safe route is the manufacturer's own software. Each brand ships a utility that reports health and applies firmware: Samsung Magician, Crucial Storage Executive and Kingston SSD Manager are common examples, and Seagate and Western Digital do the same for their drives. Firmware fixes genuine bugs, including early models that could lose data when the machine went idle. Never take drive firmware from a driver download site, because those pages repackage old versions.
Back up before you flash firmware, close every program using the drive, and keep the machine on mains power. A firmware update interrupted by a power cut can brick the drive.
TRIM is the other setting worth confirming, because it keeps an SSD's write performance steady as it fills up. Run this from a Command Prompt opened as administrator:
fsutil behavior query DisableDeleteNotify
A result of DisableDeleteNotify = 0 means TRIM is switched on, which is normal on Windows 10 and Windows 11. A value of 1 means TRIM is off, usually because a driver or a disk tool changed it. With TRIM off, deleted files are never marked, so the drive cannot reuse those blocks and the write slowdown on a full SSD becomes permanent. Set it back with the same tool, then query again to confirm.
Match the symptom to the urgency
This is the triage sheet I use on a support call.
| Symptom | Usual cause | How urgent |
|---|---|---|
| Clicking, grinding or rattling from the drive bay | Head or spindle damage on a mechanical disk | Urgent |
| Disk sits at 100 percent in Task Manager for minutes | Long sector retries or a mechanical stall | High |
| A folder will not open and Windows says the file or directory is corrupted and unreadable | Uncorrectable sectors on the volume | High |
| Reallocated sectors higher than they were last month | Progressive wear or surface damage | Medium to high |
| Pending sectors of 1 or 2 | The drive already struggles to read some sectors | Medium |
| The SSD reports that the media is write protected, or shows 0 bytes in Disk Management | The controller has locked a worn drive | Urgent |
| The drive is gone from the BIOS or UEFI list | The controller no longer detects the drive | Urgent |
| Everything works and SMART says OK | Nothing observable yet | Low, but keep checking |
Everything in the top half of that table means the same first move: copy the data off the drive before touching anything else. Use a Windows backup that you have actually tested, starting with the files that would hurt most to lose. If you need room, find the large files eating your disk space first.
Two things to avoid on a sick drive. Do not run chkdsk D: /f /r on a disk with unreadable sectors, because the surface scan takes hours and can push borderline sectors into permanent failure. Do not run Disk Cleanup over it either, since that tool rewrites free space on the drive.
Frequently asked questions
Does Windows 11 have a built-in drive health test?
No. Windows can read the health status and the SMART counters of a drive, but it does not ship a full surface test. The chkdsk command is a file system check rather than a hardware test, and third party tools that run surface scans put a lot of wear on a drive that is already struggling.
Why does Windows report OK on a drive that is about to fail?
The status is based on the drive's own SMART thresholds, and the drive only raises a failure flag once one of those thresholds is crossed. Reallocated and pending sectors often build up for months before any threshold is reached, so an OK status describes the drive today, not the drive next month.
Is one pending sector worth worrying about?
A pending sector count of 1 or 2 is a warning sign rather than an emergency. It means the drive could not read some sectors on demand and put them aside. Check the value again in a week and, if it climbs, treat the drive as failing and copy your files off it.
Can a failing SSD be repaired?
Not in any way that makes it safe to keep using. A drive that has gone read only cannot be written to, and a drive that no longer appears in the BIOS or UEFI list has failed at the hardware level. Copying what you can read is always the first move.
How often should I check drive health?
Monthly for a machine you depend on, and always before a big upgrade or a system move. Write the numbers down each time, because a rising reallocated sector count matters far more than any single reading.
What to remember
- Windows reporting OK means the disk is online, nothing more.
- The real numbers come from Get-PhysicalDisk | Get-StorageReliabilityCounter | Format-List in a PowerShell window opened as administrator.
- Watch attribute 5 for reallocated sectors, attribute 197 for pending sectors and attribute 9 for the age of the drive.
- Pending sectors of 1 or 2 are a warning, and a reallocated count that climbs means the drive is failing.
- Noise from a hard disk, or an SSD that goes read only or vanishes from the BIOS, is an emergency.
- No SMART reading tells you the day a drive will fail, so keep a backup you have tested.
Run the two PowerShell commands now and write the numbers next to the drive serial number, then run them again in a month. If the reallocated count has moved by even one sector, start copying your important files to another drive while you still have a choice.