S.M.A.R.T.
S.M.A.R.T. (Self-Monitoring, Analysis, and Reporting Technology) is a supplementary component built into many modern storage devices through which devices monitor, store, and analyze the health of their operation. Statistics are collected (temperature, number of reallocated sectors, seek errors...) which software can use to measure the health of a device, predict possible device failure, and provide notifications on unsafe values.
Smartmontools
The smartmontools package contains two utility programs for analyzing and monitoring storage devices: smartctl and smartd. Install the smartmontools package to use these tools.
SMART support must be available and enabled on each storage device to effectively use these tools. You can use #smartctl to check for and enable SMART support. That done, you can manually #Run a test and #View test results, or you can use #smartd to automatically run tests and email notifications.
smartctl
smartctl is a command-line tool that "controls the Self-Monitoring, Analysis and Reporting Technology (SMART) system built into most ATA/SATA and SCSI/SAS hard drives and solid-state drives."
The -i/--info option prints a variety of information about a device, including whether SMART is available and enabled:
# smartctl --info /dev/sda | grep 'SMART support is:'
SMART support is: Available - device has SMART capability. SMART support is: Enabled
If SMART is available but not enabled, you can enable it:
# smartctl --smart=on /dev/device
You may need to specify a device type. For example, specifying --device=ata tells smartctl that the device type is ATA, and this prevents smartctl from issuing SCSI commands to that device.
Run a test
There are three types of self-tests that a device can execute (all are safe to user data):
- Short: runs tests that have a high probability of detecting device problems,
- Extended or Long: the test is the same as the short check but with no time limit and with complete disk surface examination,
- Conveyance: identifies if damage incurred during transportation of the device.
The -c/--capabilities flag prints which tests a device supports and the approximate execution time of each test. For example:
Use -t/--test=test_name flag to run a test:
# smartctl -t short /dev/device # smartctl -t long /dev/device # smartctl -t conveyance /dev/device
View test results
You can view a device's overall health with the flag. "If the device reports failing health status, this means either that the device has already failed, or that it is predicting its own failure within the next 24 hours. If this happens […] get your data off the disk and to someplace safe as soon as you can."
# smartctl -H /dev/device
You can also view a list of recent test results and detailed information about a device:
# smartctl -l selftest /dev/device # smartctl -a /dev/device
Generate table with attributes of all disks
#!/bin/bash
function drives_csv {
declare -A drive_values
for d in `smartctl --scan -d scsi | cut -d' ' -f1`; do
drive_values["-Drive-----------------"]="${drive_values[-Drive-----------------]},$d"
for l in `smartctl -A $d | grep ATTRIBUTE_NAME -A30 | grep -v ATTRIBUTE_NAME | column -H1,3,4,5,6,7,8,9,11,12,13,14,15 -t -o, | sed 's/ //g'`; do
key=`echo $l | cut -d',' -f1`
value=`echo $l | cut -d',' -f2`
existing=${drive_values["$key"]}
drive_values["${key}"]="${existing},${value}"
#~ echo "${key},${drive_values[$key]}"
done
done
for key in "${!drive_values[@]}"; do
echo "${key}${drive_values[$key]}"
done | sort
}
drives_csv | column -s, -t
smartd
The smartd daemon monitors SMART statuses and emits notifications when something goes wrong. It can be managed with systemd and configured using the configuration file. The configuration file syntax is esoteric, and this wiki page provides only a quick reference. For more complete information, read the examples and comments within the configuration file, or read smartd.conf(5).
daemon management
To start the daemon, check its status, make it auto-start on system boot and read recent log file entries, simply start/enable the systemd unit.
smartd respects all the usual systemctl and journalctl commands.
Define the devices to monitor
To monitor for all possible SMART errors on all disks, the following setting must be added in the configuration file.
/etc/smartd.conf
DEVICESCAN -a
Note this is the default smartd configuration and the parameter, which is the default parameter, may be omitted.
To monitor for all possible SMART errors on and , and ignore all other devices:
To monitor for all possible SMART errors on externally connected disks (USB-backup disks spring to mind) it is prudent to tell smartd the UUID of the device since the of the drive might change during a reboot.
First, you will have to get the UUID of the disk to monitor: now look for the disk you want to Monitor
I know that my USB disk attached to /dev/sde during boot. Now to tell smartd to monitor that disk simply use the path.
/etc/smartd.conf
/dev/disk/by-uuid/820cdd8a-866a-444d-833c-1edb0f4becac -a
Note that you may additionally need for smartd to work.
Now your USB disk will be monitored even if the path changes during reboot.
Notifying potential problems
To have an email sent when a failure or new error occurs, use the option:
To be able to send the email externally (i.e. not to the root mail account) a MTA (Mail Transport Agent) or a MUA (Mail User Agent) will need to be installed and configured. Common MUAs are msmtp and Postfix, but perhaps the easiest dma will suffice. Common MTAs are sendmail and Postfix. It is enough to simply configure S-nail if you do not want anything else, but you will need to follow these instructions.
The -M test option causes a test email to be sent each time the smartd daemon starts:
Emails can take quite a while to be delivered. To make sure you are warned immediately if your hard drive fails, you may also define a script to be executed in addition to the email sending:
To send an email and a system notification, put something like this into :
#!/bin/sh # Send email echo "$SMARTD_MESSAGE" | mail -s "$SMARTD_FAILTYPE" "$SMARTD_ADDRESS" # Notify user wall "$SMARTD_MESSAGE"
If you are running a desktop environment, you might also prefer having a popup to appear on your desktop. In this case, you can use this script (replace and with the user and userid running X respectively) :
/usr/local/bin/smartdnotify
#!/bin/sh sudo -u ''X_user'' DISPLAY=:0 DBUS_SESSION_BUS_ADDRESS=unix:path=/run/user/''X_userid''/bus notify-send "S.M.A.R.T Error ($SMARTD_FAILTYPE)" "$SMARTD_MESSAGE" --icon=dialog-warning -u critical
This requires and a compatible desktop environment. See Desktop notifications for more details.
You can also put your custom scripts into :
This scripts notifies every logged in users on the system via libnotify.
This script requires and and a compatible desktop environment.
You can execute your custom scripts with
/etc/smartd.conf
DEVICESCAN -m @smartdnotify
Power management
If you use a computer under control of power management, you should instruct smartd how to handle disks in low power mode. Usually, in response to SMART commands issued by smartd, the disk platters are spun up. So if this option is not used, then a disk which is in a low-power mode may be spun up and put into a higher-power mode when it is periodically polled by smartd.
More info on smartmontools wiki.
On some devices the does not work. You get the following error message in syslog:
As an alternative, you can use the -i option of smartd. It controls how often smartd spins the disks up to check their status. Default is 30 minutes. To change it create and edit .
/etc/default/smartmontools
SMARTD_ARGS="-i 10800" Check status every 10800 seconds (3 hours)
For more info see .
Schedule self-tests
smartd can tell disks to perform self-tests on a schedule. The following configuration will start a short self-test every day between 2-3am, and an extended self test weekly on Saturdays between 3-4am:
Alert on temperature changes
smartd can track disk temperatures and alert if they rise too quickly or hit a high limit. The following will log changes of 4 degrees or more, log when temp reaches 35 degrees, and log/email a warning when temp reaches 40:
Complete smartd.conf example
Putting together all of the above gives the following example configuration:
- smartd scans for disks and monitors all it finds
- monitor all attributes
- enable automatic offline data collection
- enable automatic attribute autosave
- do not check if disk is in standby, and suppress log message to that effect so as not to cause a write to disk
-s ...schedule short and long self-tests- monitor temperature
- mail alerts
/etc/smartd.conf
DEVICESCAN -a -o on -S on -n standby,q -s (S/../.././02|L/../../6/03) -W 4,35,40 -m ''username-or-email''
Console applications
- (from ) also provides some disk health metrics: in particular, high values in the f_await column mean that the disk does not respond quickly to requests, and might be failing.