smartmontools-disk-health

name	smartmontools-disk-health
description	Use when checking disk, SSD, or NVMe S.M.A.R.T. health with root-only smartmontools commands, interpreting smartctl output, running self-tests, diagnosing smartd, or adding NixOS smartd monitoring configuration.

Smartmontools Disk Health

Overview

Use this skill to inspect disk health with smartctl, configure continuous monitoring with smartd, and encode the result declaratively in NixOS. Prefer device evidence from the current machine over assumptions.

Workspace Scope

• Device inspection is machine-scoped and does not require a repository working directory.
• Persistent NixOS configuration changes belong in /home/spreadzhao/workspaces/spreadconfig; when editing configuration, load and follow spreadconfig-nix.
• Do not store health-check runtime output under the spreadconfig skill source unless the user explicitly asks for a reusable skill resource.

Privileges

Run real device checks with root privileges. In non-interactive agent contexts, prefer sudo -n smartctl ... and sudo -n smartd ...; if sudo is unavailable, report that the health check cannot be completed and give the exact command for the user to run.

Do not interpret non-root smartctl --scan empty output, /dev/nvme* "No such device", or systemd bus access failures as disk absence. Those are usually permission or sandbox boundaries.

Workflow

1. Identify whole-disk devices with lsblk -o NAME,TYPE,SIZE,MODEL,SERIAL,WWN,TRAN,ROTA,MOUNTPOINTS and sudo smartctl --scan or sudo smartctl --scan-open. Ignore partitions unless the device type explicitly requires otherwise.
2. For persistent NixOS configuration, prefer /dev/disk/by-id/... paths over /dev/sdX or /dev/nvmeXnY when available.
3. Read current health before changing monitoring:
   • sudo smartctl -i -H -A -l error -l selftest DEVICE
   • use sudo smartctl -a DEVICE or sudo smartctl -x DEVICE when the short report is not enough.
4. Check capability and test duration with sudo smartctl -c DEVICE before scheduling or running self-tests.
5. Run non-captive self-tests only when appropriate:
   • short: sudo smartctl -t short DEVICE
   • long: sudo smartctl -t long DEVICE
   • results: sudo smartctl -l selftest DEVICE
6. Configure NixOS smartd only after deciding whether autodetection is enough or explicit devices are safer.
7. Validate with Nix eval/build checks. Do not switch or boot unless the user asks.

NixOS Pattern

For a simple shared module:

{ pkgs, ... }:

{
  environment.systemPackages = [ pkgs.smartmontools ];

  services.smartd = {
    enable = true;
    autodetect = true;
  };
}

Use explicit services.smartd.devices when USB bridges, RAID controllers, removable disks, or unstable device names make autodetection too broad or too noisy.

Read More

Read references/smartmontools.md when you need option details, smartd directive examples, NixOS explicit-device examples, or interpretation heuristics.