shield_lock
JinDaGe
← Back to skills
extension
Category: Development & EngineeringNo API key required

smartctl

smartctl queries SMART (Self-Monitoring, Analysis and Reporting Technology) data from hard drives and SSDs to assess health, run self-tests, and detect early signs of failure. Invoked when the user asks about disk health, drive failure risk, bad sectors, reallocated sectors, or wants to run diagnostic tests on a drive. Triggers on: smartctl, SMART, disk health, drive health, disk failure, bad sectors, reallocated sectors, drive test, smartmontools, SMART attributes, NVMe health, pending sectors, uncorrectable sectors, SMART warning.

personAuthor: jakexiaohubgithub
open_in_newRepositorydownloadDownload package

Identity

| Property | Value | |----------|-------| | Binary | smartctl | | Config | /etc/smartd.conf (for the background monitoring daemon smartd) | | Logs | Syslog / journald via smartd; self-test results stored on device | | Type | CLI tool (part of smartmontools) | | Install | apt install smartmontools / dnf install smartmontools |

Key Operations

| Task | Command | |------|---------| | Overall health check (pass/fail summary) | sudo smartctl -H /dev/sda | | All SMART attributes and device info | sudo smartctl -a /dev/sda | | Device identity only (model, serial, firmware) | sudo smartctl -i /dev/sda | | Start a short self-test (~2 minutes) | sudo smartctl -t short /dev/sda | | Start a long self-test (hours, depends on size) | sudo smartctl -t long /dev/sda | | Start a conveyance self-test (after shipping) | sudo smartctl -t conveyance /dev/sda | | Check self-test log / results | sudo smartctl -l selftest /dev/sda | | Check SMART error log | sudo smartctl -l error /dev/sda | | Enable SMART on a drive | sudo smartctl -s on /dev/sda | | NVMe drive health | sudo smartctl -a --device=nvme /dev/nvme0 | | USB drive via ATA passthrough | sudo smartctl -a -d sat /dev/sdb | | USB drive with alternate passthrough | sudo smartctl -a -d usb /dev/sdb |

Key SMART Attributes

| ID | Attribute | Critical Threshold | |----|-----------|-------------------| | 5 | Reallocated Sector Count | Any non-zero value is a warning; rising count indicates imminent failure | | 9 | Power-On Hours | Context only; high hours on a used drive increases risk | | 187 | Reported Uncorrectable Errors | Any non-zero value is serious | | 188 | Command Timeout | Non-zero indicates bus or connector issues | | 196 | Reallocation Event Count | Non-zero: sectors have been remapped | | 197 | Current Pending Sector Count | Non-zero: sectors waiting to be reallocated; read errors occurring | | 198 | Offline Uncorrectable Sectors | Non-zero: sectors that could not be recovered | | 199 | UDMA CRC Error Count | Non-zero: cable or controller signal integrity issue |

Common Failures

| Symptom | Cause | Fix | |---------|-------|-----| | Smartctl open device: failed to open | Requires root | Prefix with sudo | | USB drive returns no SMART data | Enclosure does not pass SMART commands | Try -d sat, -d usb, or -d sat,12 | | "SMART support is: Unavailable" | Drive or firmware does not support SMART | Replace with a SMART-capable drive; no workaround | | Overall health says PASSED but attribute 5 is rising | Overall PASSED is a conservative pass/fail; attributes tell the real story | Non-zero Reallocated Sector Count means the drive is degrading — schedule replacement | | Self-test status shows "in progress" | Test runs in the background | Check again in a few minutes; smartctl -a shows percentage complete | | NVMe shows different attribute names | NVMe uses a different SMART data structure | Use --device=nvme; attributes map to NVMe Log Page 0x02 fields |

Common Failures (Self-Test Results)

| Status | Meaning | |--------|---------| | Completed without error | Test passed — no issues found | | Aborted by host | Test was cancelled before completing | | Interrupted (host reset) | System rebooted mid-test | | Fatal or unknown error | Test itself failed — drive is suspect | | Completed: read failure | Test found a read error at the logged LBA; sector is failing | | Self-test routine in progress | Test is running; check again later |

Pain Points

  • Root required: smartctl opens the raw device file, which requires root. Always use sudo.
  • USB enclosure passthrough: most USB-to-SATA bridge chips do not forward SMART commands by default. Try -d sat (SAT passthrough) or -d usb to work around this. Some cheap enclosures support no passthrough method at all.
  • Tests run in the background: starting a short or long test returns immediately. The test runs on the drive's internal controller. Check results with -l selftest or -a after waiting the estimated time (shown when the test starts).
  • "PASSED" does not mean healthy: the overall health check passes unless a SMART threshold has been crossed by the manufacturer's definition. A drive with dozens of reallocated sectors can still return PASSED. Always read the attribute table, specifically attributes 5, 187, 196, 197, and 198.
  • NVMe attribute names differ: NVMe uses a different log page structure than ATA SMART. Attribute names like "Available Spare" and "Percentage Used" replace the ATA numbering. Always include --device=nvme for NVMe drives.
  • smartd for ongoing monitoring: smartctl is a one-shot query tool. For continuous monitoring, configure smartd in /etc/smartd.conf to run scheduled self-tests and email alerts on threshold crossings.

References

See references/ for:

  • cheatsheet.md — 10 task-organized patterns for common smartctl workflows
  • docs.md — man pages and upstream documentation links