SMART

Package: smartmontools

smartctl vs smartd

• smartctl is meant for interactive use
• smartd continuously monitors

smartctl

Info on a block device

sudo smartctl -i /dev/sda

• Includes whether SMART support is enabled.
• Does not show all attributes.
• Use -x to show all information. Very long. grep.

If smartctl -i fails, let smartctl know the type

sudo smartctl -d [ata/scsi/sat] -i /dev/sda

• If -d is necessary for examining, it is also necessary for all other commands.

Turning on SMART

sudo smartctl -s on -o on -S on /dev/sda

• -s on: turn SMART on.
• -o on: turn on offline data collection.
• -S on: This enables “autosave of device vendor-specific Attributes”.

Offline data collection could have some performance impact. In practice, this has little effect.

Health check

sudo smartctl -H /dev/sda

Time estimates for self-tests

sudo smartctl -c /dev/sda

Short self-test

Start:

sudo smartctl -t short /dev/sda

Results:

sudo smartctl -l selftest /dev/sda

Long self-test

Start:

sudo smartctl -t long /dev/sda

Results:

sudo smartctl -l selftest /dev/sda

Current disk temperature

smartctl -A /dev/sda | grep Temperature_Celsius

smartd

• Start/enable the daemon.
• Config file: /etc/smartd.conf or /etc/smartmontools/smartd.conf
• man smartd and man smartd.conf

DEVICESCAN

• This line determines what is monitored and what actions to take after SMART scans.

DEVICESCAN options

• -a: equivalent to -Hft -l error -l selftest -l selftests-sts -C 197 -U 198. Default when no other options are set. Gives a lot of info. Perhaps too much info.

• -H: [ATA] Check the health status of the disk with the SMART RETURN STATUS command.

• -m [EMAIL] Send a warning email to the email address if the ‘-H’, ‘-l’, ‘-f’, ‘-C’, or ‘-O’ directives detect a failure or a new error, or if a SMART command to the disk fails.

• -M: Modifiers for-m. Requires -m. Lots of options. Notably the ability to run an alternative, rather than default, mail program.

• -M exec /usr/libexec/smartmontools/smartdnotify

• -n standby,10,q:

  • Prevent a disk from being checked if it is in a certain state. Alternatives to standby are: never, sleep or idle.

  • (Optional) The integer argument determines the maximum number of checks that can be skipped.

  • (Optional) The q prevents smartd writing a log message when checks are skipped. A disk will spin-up if the log message is written.

• -s (S/../.././02|L/../../6/03): schedule a short and/or long self-test. month of the year/day of the month/day of the week/hour of day.

  • S/../.././02: Short-test between 2 - 3 am every morning.
  • L/../../6/03: Long-test between 3 - 4 am on Satuday.

• -W 4,35,40: alerts for temperature changes.

  • Argument 1: report if the temperature has changed by more than X since the last report.
  • Argument 2: Report if the temperature has reached this bound.
  • Argument 3: Send a temperature critical message at this bound.

• man smartd.conf for all options.

• DEVICESCAN -H -m root -M exec /usr/libexec/smartmontools/smartdnotify

• DEVICESCAN -a -o on -S on -m root -M exec /usr/libexec/smartmontools/smartdnotify -s (S/../.././10|L/../../5/11)