stuvusit / smartd

Ansible role to install and configure smartmontools

smartd

This role installs and configures smartd. In the default configuration, all devices will be tested and observed, but no mails will be sent.

Requirements

Debian, Ubuntu, Arch Linux, or Fedora /usr/bin/mail has to exist and be configured to be able to send report mails if that is desired.

Role Variables

This role only uses one top-level var called smartd_devices, which is a dict of dicts to specify device monitoring. Each key corresponds to one drive path, while DEVICESCAN (the only default entry) may be used to scan for devices. After a DEVICESCAN entry, all further lines are ignored by smartd, so you probably want to insert it at the end or not at all.

All possible options inside the dict are briefly described here. You can find a detailed wall-of-text for each one on the smartd man page.

Name	Mandatory / Default	Description
type	auto	The type of device, e.g. ata, scsi, marvell, etc.
nocheck	standby,15,q	Powermode setting. By default, devices in standby are only woken up every 15th attempt and skipped check attempts are not logged.
check_type	normal	Use normal or permissive to force SMART checking even if it is not advertised.
ata_offline_testing		Set to on or off to enable or disable automatic ATA offline testing by the device itself. Note that you have to quote the value because of ansible's interpretation of off.
attribute_autosave	True	Enables or disables attribute autosave on startup
check_health_return_status	True	Check device health status using the SMART return status
report_error_types	[error,xerror,selftest,offlinests,"scterc,0,0"]	List of error types to report. By default, SMART errors as well as failed tests are logged and TLER will be disabled.
non_smart_settings	["lookahead,on"]	List of non-SMART options to set
test_schedule	(L/../../6/01|S/../.././02)	REGEXP to specify self-test schedules. By default, a short test is executed daily after 1am. A long test is executed each sunday after 2am.
mail_recipients	[]]	List of mail addresses to report to
mail_frequency	diminishing	Frequency of report emails. once, daily and diminishing are possible values.
mail_script		Path to a script that will be executed in addition to sending report mails
report_ata_failure	True	Report failure of any ATA usage attributes
report_ata_prefail	True	Report change of any ATA prefail attributes
report_ata_usage	False	Report any time a usage attribute has changed
ignore_ata_failure_ids	[]	List of SMART IDs to ignore when checking for failures
ignore_ata_usage_ids	[]	List of SMART IDs to ignore when tracking usage value changes
report_raw_ids	[]	List of IDs to force reporting raw values for. Append ! to an ID to consider changes critical.
ata_pending_id	197+	ID specifying pending sectors. Append + to only report increasage instead of non-zero-ness.
ata_uncorrectable_id	198+	ID specifying pending sectors. Append + to only report increasage instead of non-zero-ness.
temperature_report_diff	10	Difference in temperature to report. Use 0 to ignore.
temperature_report_info	45	Temperature threshold to start sending informational reports. Use 0 to ignore.
temperature_report_crit	50	Temperature threshold to start sending critical reports. Use 0 to ignore.
firmware_bugs	[]	List of known firmware bugs SMARTD should circumvent.
vendor_formats	[]	List of options to rewrite interpretation of raw SMART values and their interpretation.
preset_mode	use	Set to ignore if you don't want to use known presets for a detected drive.

Example Playbook

- hosts: storage
  roles:
    - role: smartd
      smartd_devices:
        /dev/sda:
          check_type: permissive
          test_schedule: L/../../7/04
          temperature_report_diff: 5
        DEVICESCAN:
          non_smart_settings:
           - lookahead,on
           - wcache,off
           - standby,off
           - apm,254

This configuration will force SMART checking even if it is not advertised as implemented on /dev/sda. The device will only be tested after 4am every sunday and already temperature changes of 5 degrees will be reported. All other drives will be monitored using the default settings, except for the listed options set additionally.

License

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.

Author Information

• Michel Weitbrecht (SlothOfAnarchy) [email protected]

smartd's Issues

smartd_devices should be a list not dict

It's not possible to specify a 3ware raid controller with the current config definition since the device path is always the same

given

      smartd_devices:
        /dev/twa0:
          type: "3ware,0"
        /dev/twa0:
          type: "3ware,1"

will create a warning

[WARNING]: While constructing a mapping from play.example.yml, line x, column y, found a duplicate dict
key (/dev/twa0). Using last defined value only.

i would suggest a list instead

      smartd_devices:
        - name: "/dev/twa0"
          type: "3ware,0"
        - name: "/dev/twa0"
          type: "3ware,1"

so the template code would become

{% for config in smartd_devices | d([]) %}
        {{- config.name -}}

i know this breaks backward compatibility but maybe with an extra check in the template both variants could be supported?

thanks for the role. i might submit a patch when i find a minute

Warning Emails

Hello,

is there any information on how to make email notifications for warnings work?

typo in day of the week of long test

• (L/../../6/01|S/../.././02)

• (L/../../7/01|S/../.././02)

sunday is 7th day of the week
d is the day of the week, expressed with one decimal digit. The range is from 1 (Monday) to 7 (Sunday) inclusive.