This is also one of the last TechMails that TechRepublic will be publishing that I’ve written. There might be one or two more in the queue yet, I can’t remember if everything I’ve submitted has been published or not. It has been almost 12 years of writing monthly for TechRepublic — initially as full length articles and the last few years as blog-style tips (which are both easier and harder to write; you get less words and more topics to cover in a month). The last year or so I’ve also been writing Mac tips for TechRepublic, as one of the original contributors to the Mac track.

Twelve years is a long time, so it is with mixed feelings that I gave my resignation to TechRepublic last month. I’ve worked with some really great editors: Jack, Sonja, Selena… you guys have been great to work with and I will definitely miss working with you. I think, after a respite from technical writing, that I’ll work on updating some of the documentation I have on the linsec.ca wiki, and hopefully finish off a few more that I’ve started but never completed. At least writing for my own wiki there are no deadlines, no wracking the brain for various topics that appeal to a larger crowd (I can concentrate on the niche security/sysadmin stuff that I enjoy), so it should be a little more relaxed and hobby-ish, rather than feeling like “real work”.

Nagios is one of my favourite things to tinker with, and using NRPE is a bit nicer than using the check_by_ssh plugin that I used to use. I don’t really care if the load average or memory usage is sniffed, so using NRPE is nice and easy (and removes the hassles of key-based auth for SSH, seeing all those SSH connections show up in logwatch reports, etc.). If you have the ability to install NRPE on a remote system to monitor, it’s worth trying out.

This may need some tweaking to work with other Nagios setups, but shouldn’t be too hard to adapt. One of these days I’ll do a writeup on my Nagios configuration. Anyways, the wrapper script is as follows. It could probably be optimized a bit more, but it works well enough. WordPress doesn’t handle the indents very well, so keep that in mind.

#!/bin/sh

usage() {
    echo "${0} -w [warn] -c [crit] [drives]"
}

if [ "${1}" == "-h" -o "${1}" == "--help" ]; then
    usage
    exit 0
fi
if [ "${1}" == "-w" ]; then
    shift
    warn="${1}"
    shift
else
    usage
    exit 1
fi
if [ "${1}" == "-c" ]; then
    shift
    crit="${1}"
    shift
else
    usage
    exit 1
fi
while [ "${1}" != "" ]; do
    drives="${drives} ${1}"
    shift
done
if [ "${drives}" == "" ]; then
    usage
    exit 1
fi

status=0
smsg=""
htemp=0

for drive in ${drives}; do
    msg=""
    stats=`/usr/local/sbin/hddtemp ${drive}`
    model=`echo ${stats} | cut -d ':' -f 2`
    temp=`echo ${stats} | cut -d ':' -f 3 | cut -d ' ' -f 2`
    dev=`echo ${drive}|cut -d '/' -f 3`

    if [ "${temp}" -ge "${warn}" ]; then
        if [ "${status}" != "2" ]; then
            status=1
        fi
    fi

    if [ "${temp}" -ge "${crit}" ]; then
        status=2
    fi

    if [ "${temp}" -gt "${htemp}" ]; then
        htemp="${temp}"
    fi

    smsg="${smsg}${dev}=${temp}C; "
done

case "${status}" in
    2)
        wmsg="CRITICAL"
        ;;
    1)
        wmsg="WARN"
        ;;
    0)
        wmsg="OK"
        ;;
esac

echo "HDDTEMP ${wmsg} - ${smsg}|hddtemp=${htemp};${warn};${crit};0"

The output, in Nagios’ status view looks like:

HDDTEMP OK - hda=22C: sda=24C: sdb=24C:

It’s called as “hddtemp-mon -w 30 -c 35 /dev/hda /dev/sda /dev/sdb”.