Hard-drive and SSD health: Reading SMART values

In short: SMART is self-monitoring of hard drives and SSDs; values such as reallocated sectors or wear show early when a drive should be replaced.

In more detail: Drives fail at some point, often without warning. S.M.A.R.T. (Self-Monitoring, Analysis and Reporting Technology) collects measurements in the drive that can be read out. SMART is no guarantee: some drives fail without any anomaly. Hence always: backup.

In Depth

Tools

• CrystalDiskInfo (Windows), smartmontools (smartctl -a /dev/sda, Linux/macOS), GSmartControl (graphical), manufacturer tools.
• NVMe SSDs: smartctl -a /dev/nvme0 or nvme smart-log. They show status such as “Good”, “Caution” or “Bad” and the individual values.

Important values for HDDs

• Reallocated Sector Count: replaced defective sectors. If the value rises, the disk is dying.
• Current Pending Sectors and Uncorrectable Sectors: unstable sectors.
• Spin Retry Count, Seek Error Rate, temperature.
• Power-On Hours: operating hours. Every increase in the critical values is an alarm signal.

Important values for SSDs

• Percentage Used / Wear Leveling Count: wear of the cells.
• Available Spare: reserve cells.
• Media and Data Integrity Errors.
• Data Units Written (TBW): amount of data written, compare with the manufacturer limit. Modern SSDs last many years in everyday use; on warnings back up data and replace.

What to do on a warning?

Back up immediately the important data, then replace the drive. Do not keep writing to it, no continuous operation for very old HDDs. With clicking noises switch off at once. For data recovery ask professionals; trying it yourself can destroy data for good. Keep temperatures low and do not knock the computer.

See also: Wear levelling on SSDs, SSDs, Backup strategy, RAID