Quick answer

What this guide helps you do

Install smartmontools, identify the physical disk, read SATA or NVMe health data, run self-tests and recognise evidence that demands an immediate backup.

SMART is evidence, not a guarantee

A PASS result does not prove a disk will not fail. A failing result, growing media errors, pending sectors or failed self-test is a reason to protect data immediately. Keep independent backups regardless of SMART status.

Identify the physical device

lsblk -d -o NAME,SIZE,MODEL,SERIAL,TRAN,ROTA
lsblk -o NAME,SIZE,FSTYPE,MOUNTPOINTS

Run SMART commands against the physical device, not a mounted directory or ordinary partition. RAID controllers and USB bridges may need device-specific passthrough options.

Install and read information

sudo apt update
sudo apt install smartmontools
sudo smartctl --scan-open
sudo smartctl -i /dev/sdX
sudo smartctl -a /dev/sdX

For NVMe, smartctl normally reports NVMe health fields through the device such as /dev/nvme0.

For SATA disks, review the overall result, reallocated sectors, current pending sectors, offline uncorrectable sectors, UDMA CRC errors, temperature, power-on hours, error log and self-test log. Attribute names and raw encodings can be vendor-specific.

For NVMe, review critical warnings, temperature, available spare, percentage used, media and data integrity errors, and error-log entries.

Run a short self-test

sudo smartctl -t short /dev/sdX
sudo smartctl -l selftest /dev/sdX

The first command reports an expected completion time. Wait that long, then read the log. Do not assume that starting the test means it passed.

Run a long test deliberately

sudo smartctl -c /dev/sdX
sudo smartctl -t long /dev/sdX
sudo smartctl -l selftest /dev/sdX

A long test can take hours and add workload. Schedule it appropriately. If a drive is already showing serious errors, prioritise copying irreplaceable data rather than repeatedly stressing it.

Respond to warnings

Back up immediately, reduce writes, save the complete SMART report, check kernel I/O errors and plan replacement. CRC errors can indicate a cable or connection problem; pending or uncorrectable sectors point more directly at media trouble. Do not clear evidence before recording it.

Verification checklist

  • The physical device and serial are correct.
  • A complete SMART report was saved.
  • Health, attributes, errors and self-tests were reviewed.
  • Short or long tests completed rather than merely started.
  • Kernel I/O logs were checked.
  • Warning trends trigger backup and replacement planning.
  • SMART monitoring complements independent backups.

Interpretation varies by device and transport; this guide does not certify any disk as healthy.

Next: Set a Static IP on Ubuntu Server. Return to Mount a Drive Automatically with fstab.

Official reference: smartmontools.