> I found your page very helpful in setting up a raid1 system at my
> work. I do have one unanswered question. How does the software
> controller notify someone that one of the drives has failed?
>
Oh, I have a little script I run daily from a cron file -- actually
two scripts. You could run them as often as you like I suppose
The first (short) is used to manually produce the match template
#######################################################
#!/bin/sh
# get active raid status
# version 1.03 4-18-00
/bin/cat /proc/mdstat | /usr/bin/grep active
########### end
######################################################
The second is the cron task and does the comparison, and
is called when doing the check with the email address of whom to
notify on the command line
######################################################
#! /bin/sh
#
# Version: 1.02 8-7-99 Michael A. Robinton <
# [EMAIL PROTECTED] >
#
# raidcheck.sh This file must be executed periodically by cron
# It will check the raid status and e-mail a message if
# a status error is detected.
#
# USEAGE: raidcheck.sh e-mail@addr {where to send error msg}
#
############ check raid status info ##############
MAILTO=$1
SENDMAILCMD="/usr/lib/sendmail -t -oi -O DeliveryMode=background"
RAIDHOST=`/bin/hostname`
# you must have issued
# raidgetstatus.sh >raidgood
# in the local directory before this routine will work
# capture raid status
RAIDSTATUS=`./raidgetstatus.sh`
# get raid reference
RAIDREF=`cat ./raidgood`
# Test for a clean shutdown with array matching reference
if [ "${RAIDSTATUS}" != "${RAIDREF}" ]; then
${SENDMAILCMD} <<EOF123xxx
From:${RAIDHOST} <root@${RAIDHOST}>
To:${MAILTO}
Reply-To:root@${RAIDHOST}
Errors-To:root@${RAIDHOST}
Sender:root@${RAIDHOST}
Subject:ALERT Raid Device Failure on ${RAIDHOST}
${RAIDHOST} has detected an error in the raid status
The status should be:
${RAIDREF}
The status is:
${RAIDSTATUS}
EOF123xxx
#| ${SENDMAILCMD}
fi
############ end
