How to check your SSD's health

Every SSD keeps a running record of its own health: how much of its rated life it has used, how much spare flash is left, how much data has been written to it, and whether it has hit errors. Reading it takes a minute and a free tool, and it tells you whether a drive is worth selling, keeping or recycling.

On Windows

• Windows itself, for NVMe drives: Settings → System → Storage → Advanced storage settings → Disks & volumes, then Properties on the drive. Drive health shows the estimated remaining life, available spare and temperature.
• CrystalDiskInfo, free, for any drive: it shows a Health Status with a percentage (Good, Caution or Bad), Total Host Writes, Power On Hours and the full list of health attributes.
• The drive maker’s utility — Samsung Magician, WD Dashboard, Crucial Storage Executive, Kingston SSD Manager — shows the same figures in the maker’s own words, and can update the drive’s firmware.

On Linux and macOS

Install smartmontools (from your distribution’s packages, or brew install smartmontools on a Mac) and run sudo smartctl -a /dev/nvme0 for an NVMe drive or sudo smartctl -a /dev/sda for a SATA one. On Linux, sudo nvme smart-log /dev/nvme0 from nvme-cli gives the NVMe figures on their own.

Reading an NVMe drive’s numbers

NVMe drives all report the same standard fields:

• Percentage Used — the maker’s estimate of how much of the drive’s rated life is gone. 0% is new; it can pass 100% while the drive still works. CrystalDiskInfo’s health percentage is 100 minus this.
• Available Spare — the spare flash left to replace worn blocks, as a percentage, next to the threshold below which the drive warns it is failing.
• Data Units Written — total data written. Each unit is 512,000 bytes, so multiply by 512,000 for bytes; most tools convert it to terabytes for you.
• Media and Data Integrity Errors — errors the drive could not correct. On a healthy drive this is 0.

Reading a SATA drive’s numbers

SATA drives report numbered attributes, and makers use different ones for wear. The ones to look at:

• Wear or life remaining — Samsung reports a Wear Leveling Count, Crucial a Percent Lifetime Remaining, others a Media Wearout Indicator. CrystalDiskInfo turns whichever the drive has into its health percentage.
• Total LBAs Written or Total Host Writes — data written. The unit differs between makers, so read it from a tool that converts it rather than doing the sum yourself.
• Reallocated Sector Count and any pending or uncorrectable sector counts — blocks that have failed. A few reallocations are normal over a long life; a rising count is not.

Data written against rated TBW

Every drive has an endurance rating in terabytes written (TBW), on its datasheet and warranty terms. Dividing the data written so far by that rating tells you how far through its rated life the drive is. Ratings vary a lot between drives of the same size:

Everyday use in a desktop or laptop writes far less each year than these ratings allow, so most drives are replaced long before they reach them. Drives used as a cache, for video capture or for Chia plotting are the exception.

When is a drive worn out?

• Healthy: health well above 50%, no media errors, no growing reallocated count. Sell it as a working drive.
• Heavily written: health low, or data written near or past the rating, but no errors. It still works; at RebootSemi it prices lower than a lightly used drive of the same size.
• Failing: Caution or Bad status with media errors, Available Spare under its threshold, or a drive that has switched itself to read-only. Copy off what you need now. We do not pay for failing drives, but one sent with other parts is physically destroyed and recycled.

Next: wipe the drive before it leaves you, and see what it is worth.