Who This Is For
Beginner.
Think of it like your car's check-engine light. A small warning today beats a breakdown on the highway next month.
What You'll Build
SMART scheduled checks with alerting, or at minimum a documented manual review process.
Prerequisites
- A Linux host with direct or USB-attached disks
Install smartmontools
$sudo apt install smartmontools
Basic Commands
$sudo smartctl -i /dev/sda # device info
$sudo smartctl -H /dev/sda # health status
$sudo smartctl -a /dev/sda # full attribute report
$sudo smartctl -t short /dev/sda
$sudo smartctl -t long /dev/sda
$sudo smartctl -H /dev/sda # health status
$sudo smartctl -a /dev/sda # full attribute report
$sudo smartctl -t short /dev/sda
$sudo smartctl -t long /dev/sda
Key Warning Indicators
- Reallocated sectors, the drive has already had to work around bad sectors
- Pending sectors, sectors waiting to be reallocated
- Uncorrectable sectors, data loss has already occurred in that spot
- Rising temperature trend
- For NVMe: media errors in the health log
A "Passed" Status Is Not a Guarantee
SMART self-tests can pass while individual attributes show a worrying trend. Read the full attribute report, not just the pass/fail summary.
Schedule Tests with smartd
smartd (installed alongside smartmontools) runs scheduled tests and can email or webhook on warnings, configure it in /etc/smartd.conf.
When a Disk Shows Warning Signs
- Confirm backups are current
- Copy off any critical data immediately
- Avoid running further stress tests if failure looks imminent
- Replace the drive
Security & Backup Notes
- SMART is an early-warning system, not a replacement for backups, a drive can fail catastrophically with little or no SMART warning
Troubleshooting
- USB enclosures sometimes don't pass SMART data through correctly, try
-d sator the enclosure-specific device type flag with smartctl
Lab Finish Line
SMART scheduled checks with alerting or a documented review process.