@PatrickC21
@Fritz2cat
Merci pour vos informations.
En fait ce qui me gêne c'est le fait que la consommation CPU (load average) excessive de raid-check. Il faudrait pouvoir baisser la priorité de raid-check, quitte à allonger la durée de raid-check.
D'après vos réponses, un compromis sécurité/performance serait de ne tourner raid-check qu'une fois par mois, avec automatisation (cron journalier de @Fritz2cat) des contrôles du RAUD (parsing de la ligne "1952986048 blocks [2/2] [UU]", sortie de la commande "/usr/sbin/mdadm --detail /dev/md2").
Question : Quel est le risque si l'on fait seulement une fois tous les 3 mois ? :-). Car si le statut "UU" ne change pas, c'est que le RAID est bon non ? Ou est-ce que le raid-check est utile pour la prévention ?
Bien sûr, en plus il faut aussi des cron pour les alertes SMART en cas de problème hardware des disques.
# Run system wide raid-check once a month on first Sunday at 1am by default
# http://stackoverflow.com/questions/3241086/cron-how-to-schedule-to-run-first-sunday-of-every-month
# min hour day_of_month month day_of_week_ program_to_be_run
0 1 1-7 * * [ "$(date '+\%a')" == "Sun" ] && /usr/sbin/raid-check
# Run system wide raid-check every 3 months (feb, may, aug, nov) on first Sunday at 1am by default
# http://stackoverflow.com/questions/3241086/cron-how-to-schedule-to-run-first-sunday-of-every-month
# min hour day_of_month month day_of_week_ program_to_be_run
0 1 1-7 2,5,8,11 * [ "$(date '+\%a')" == "Sun" ] && /usr/sbin/raid-check
# /usr/sbin/mdadm --detail /dev/md2
Checking RAID1...
Personalities : [raid1]
md2 : active raid1 sdb2[1] sda2[0]
1952986048 blocks [2/2] [UU]
unused devices: <none>
.
.
RAID1 status...
/dev/md2:
Version : 0.90
Creation Time : Mon Oct 19 13:08:16 2015
Raid Level : raid1
Array Size : 1952986048 (1862.51 GiB 1999.86 GB)
Used Dev Size : 1952986048 (1862.51 GiB 1999.86 GB)
Raid Devices : 2
Total Devices : 2
Preferred Minor : 2
Persistence : Superblock is persistent
Update Time : Thu Jul 11 00:10:01 2019
State : active
Active Devices : 2
Working Devices : 2
Failed Devices : 0
Spare Devices : 0
Consistency Policy : resync
UUID : aceb8bc2:20036cf3:a4d2adc2:26fd5302
Events : 0.13713
Number Major Minor RaidDevice State
0 8 2 0 active sync /dev/sda2
1 8 18 1 active sync /dev/sdb2
Edit (ajout):
Pour SMART, j'ai ces settings:
/etc/smartmontools/smartd.conf
# sends a single test email immediately upon smartd startup.
/dev/sda -m root@example.com -M test
/dev/sdb -m root@example.com -M test
# A very silent check. Only report SMART health status if it fails
# But send an email in this case
#--------
/dev/sda -H -C 0 -U 0 -m root@example.com
/dev/sdb -H -C 0 -U 0 -m root@example.com
# Finally, you can schedule tests at your preferred schedule using the "-s" flag and the regular expression in the form of "T/MM/DD/d/HH", where:
#
# T in the regular expression indicates the kind of test:
#
# L: long test
# S: short test
# C: Conveyance test (ATA only)
# O: Offline (ATA only)
# and the remaining characters represent the date and time when the test should be performed:
#
# MM is the month of the year.
# DD is the day of the month.
# HH is the hour of day.
# d is the day of the week (ranging from 1=Monday through 7=Sunday).
# MM, DD, and HH are expressed with two decimal digits.
#--------
# short test every day at 02h, long test every Saturday at 04h
#/dev/sda -a -o on -S on -s (S/../.././02|L/../../6/04)
#/dev/sdb -a -o on -S on -s (S/../.././02|L/../../6/04)
# short test every day at 02h
/dev/sda -a -o on -S on -s (S/../.././02)
/dev/sdb -a -o on -S on -s (S/../.././02)
