HOWTO Setup S.M.A.R.T. hard-drive Monitoring
Typically, we use a 3Ware controller, with anywhere from 2 - 16 individual hard-drives attached. Although we use 3Ware's tw_cli tool in both a daily cron-job and also in a Nagios monitor, it's also a good idea to get a daily logwatch-email stanza with drive-statistics and alerts. Belts and suspenders, you know.
In addition to (passive) monitoring, we actively invoke Long Self-Tests on each drive. These tests are scheduled to try to hit low-usage times, and also avoid our tape-backup times.
Install the (Gentoo) package:
hostname # emerge -av smartmontools
Start the smartd monitoring daemon automatically in the default runlevel:
hostname # rc-update add smartd default
Representative configuration file:
hostname # emacs -nw /etc/smartd.conf
# This file: /etc/smartd.conf
# created April 14, 2010, Gordon Pritchard
# Monitor 16 SATA disks connected to a 3ware 9650 controller which
# uses the (built-into-kernel) 3w-9xxx driver.
/dev/twa0 \
-s L/../../0/01 \ # Long Self-tests Sundays between 1-2am
-d 3ware,0 \ # First physical drive, behind 3Ware controller
-I 194 \ # Ignore temperature
-I 9 \ # Ignore power-on hours
-a # Report on:
# SMART health status
# usage failures (-f)
# changes in Prefailure, Usage (-t)
# increases in error (-l error)
# increases in self-test errors (-l selftest)
/dev/twa0 \
-s L/../../0/03 \ # Long Self-tests Sunday between 3-4am
-d 3ware,1 \ # Second physical drive, behind 3Ware controller
-I 194 \ # Ignore temperature
-I 9 \ # Ignore power-on hours
-a # Report on:
# SMART health status (-H)
# usage failures (-f)
# changes in Prefailure, Usage (-t)
# increases in error (-l error)
# increases in self-test errors (-l selftest)
/dev/twa0 \
-s L/../../0/23 \ # Long Self-tests Sundays between 11pm-midnight
-d 3ware,2 \ # Third physical drive, behind 3Ware controller
-I 194 \ # Ignore temperature
-I 9 \ # Ignore power-on hours
-a # Report on:
# SMART health status (-H)
# usage failures (-f)
# changes in Prefailure, Usage (-t)
# increases in error (-l error)
# increases in self-test errors (-l selftest)
/dev/twa0 \
-s L/../../1/01 \ # Long Self-tests Mondays between 1-2am
-d 3ware,3 \ # Fourth physical drive, behind 3Ware controller
-I 194 \ # Ignore temperature
-I 9 \ # Ignore power-on hours
-a # Report on:
# SMART health status (-H)
# usage failures (-f)
# changes in Prefailure, Usage (-t)
# increases in error (-l error)
# increases in self-test errors (-l selftest)
/dev/twa0 \
-s L/../../1/23 \ # Long Self-tests Mondays between 11pm-midnight
-d 3ware,4 \ # Fifth physical drive, behind 3Ware controller
-I 194 \ # Ignore temperature
-I 9 \ # Ignore power-on hours
-a # Report on:
# SMART health status (-H)
# usage failures (-f)
# changes in Prefailure, Usage (-t)
# increases in error (-l error)
# increases in self-test errors (-l selftest)
/dev/twa0 \
-s L/../../2/01 \ # Long Self-tests Tuesdays between 1-2am
-d 3ware,5 \ # Sixth physical drive, behind 3Ware controller
-I 194 \ # Ignore temperature
-I 9 \ # Ignore power-on hours
-a # Report on:
# SMART health status (-H)
# usage failures (-f)
# changes in Prefailure, Usage (-t)
# increases in error (-l error)
# increases in self-test errors (-l selftest)
/dev/twa0 \
-s L/../../2/23 \ # Long Self-tests Tuesdays between 11pm-midnight
-d 3ware,6 \ # Seventh physical drive, behind 3Ware controller
-I 194 \ # Ignore temperature
-I 9 \ # Ignore power-on hours
-a # Report on:
# SMART health status (-H)
# usage failures (-f)
# changes in Prefailure, Usage (-t)
# increases in error (-l error)
# increases in self-test errors (-l selftest)
/dev/twa0 \
-s L/../../3/01 \ # Long Self-tests Wednesdays between 1-2am
-d 3ware,7 \ # Eighth physical drive, behind 3Ware controller
-I 194 \ # Ignore temperature
-I 9 \ # Ignore power-on hours
-a # Report on:
# SMART health status
# increases in error (-l error)
# increases in self-test errors (-l selftest)
/dev/twa0 \
-s L/../../3/23 \ # Long Self-tests Wednesdays between 11pm-midnight
-d 3ware,8 \ # Ninth physical drive, behind 3Ware controller
-I 194 \ # Ignore temperature
-I 9 \ # Ignore power-on hours
-a # Report on:
# SMART health status (-H)
# usage failures (-f)
# changes in Prefailure, Usage (-t)
# increases in error (-l error)
# increases in self-test errors (-l selftest)
/dev/twa0 \
-s L/../../4/01 \ # Long Self-tests Thursdays between 1-2am
-d 3ware,9 \ # Tenth physical drive, behind 3Ware controller
-I 194 \ # Ignore temperature
-I 9 \ # Ignore power-on hours
-a # Report on:
# SMART health status (-H)
# usage failures (-f)
# changes in Prefailure, Usage (-t)
# increases in error (-l error)
# increases in self-test errors (-l selftest)
/dev/twa0 \
-s L/../../4/23 \ # Long Self-tests Thursdays between 11pm-midnight
-d 3ware,10 \ # Eleventh physical drive, behind 3Ware controller
-I 194 \ # Ignore temperature
-I 9 \ # Ignore power-on hours
-a # Report on:
# SMART health status (-H)
# usage failures (-f)
# changes in Prefailure, Usage (-t)
# increases in error (-l error)
# increases in self-test errors (-l selftest)
/dev/twa0 \
-s L/../../5/01 \ # Long Self-tests Fridays between 1-2am
-d 3ware,11 \ # Twelfth physical drive, behind 3Ware controller
-I 194 \ # Ignore temperature
-I 9 \ # Ignore power-on hours
-a # Report on:
# SMART health status (-H)
# usage failures (-f)
# changes in Prefailure, Usage (-t)
# increases in error (-l error)
# increases in self-test errors (-l selftest)
/dev/twa0 \
-s L/../../5/23 \ # Long Self-tests Fridays between 11pm-midnight
-d 3ware,12 \ # Thirteenth physical drive, behind 3Ware controller
-I 194 \ # Ignore temperature
-I 9 \ # Ignore power-on hours
-a # Report on:
# SMART health status (-H)
# usage failures (-f)
# changes in Prefailure, Usage (-t)
# increases in error (-l error)
# increases in self-test errors (-l selftest)
/dev/twa0 \
-s L/../../6/01 \ # Long Self-tests Saturdays between 1-2am
-d 3ware,13 \ # Fourteenth physical drive, behind 3Ware controller
-I 194 \ # Ignore temperature
-I 9 \ # Ignore power-on hours
-a # Report on:
# SMART health status (-H)
# usage failures (-f)
# changes in Prefailure, Usage (-t)
# increases in error (-l error)
# increases in self-test errors (-l selftest)
/dev/twa0 \
-s L/../../6/03 \ # Long Self-tests Saturdays between 3-4am
-d 3ware,14 \ # Fifteenth physical drive, behind 3Ware controller
-I 194 \ # Ignore temperature
-I 9 \ # Ignore power-on hours
-a # Report on:
# SMART health status
# usage failures (-f)
# changes in Prefailure, Usage (-t)
# increases in error (-l error)
# increases in self-test errors (-l selftest)
/dev/twa0 \
-s L/../../6/23 \ # Long Self-tests Saturdays between 11pm-midnight
-d 3ware,15 \ # Sixteenth physical drive, behind 3Ware controller
-I 194 \ # Ignore temperature
-I 9 \ # Ignore power-on hours
-a # Report on:
# SMART health status
# usage failures (-f)
# changes in Prefailure, Usage (-t)
# increases in error (-l error)
# increases in self-test errors (-l selftest)
In order to find what parameters are available for monitoring or ignoring, issue the smartctl command:
hostname # smartctl --all /dev/twa0 -d 3ware,0 (of course, change the ,0 to any other 3Ware-connected drive you're interested in)
To learn how long you should allow for a Long SelfTest (typical values are 72min for a 150GB WD Raptor, and 150min for a 500GB WD RE2):
hostname # smartctl -c /dev/twa0 -d 3ware,0 (of course, change the ,0 to any other 3Ware-connected drive you're interested in)