Since arpwatch and spamassassin crashed for some unknown reasons (probably bugs) lately, I needed a tool that would monitor services status on a regular basis.

I found Nagios http://www.nagios.org/ but it was not really what I needed. Indeed, Nagios can only monitor some specific services (usually services opening a tcp port) and it offered just way too many features, thus requiring a lot of dependencies and was not the easiest piece of software to install.

To do the job, I made this (very) little script that runs from a cron on a daily basis :

<code>#!/bin/sh

SERVICES_REGULAR="squid smb 3dm mysqld postfix frox dhcpd sshd spamassassin httpd courier-imap arpwatch xinetd"

echo "Services"
echo "========"
for services_regular in $SERVICES_REGULAR
do
/sbin/service $services_regular status | sed -re 's/(pid|(|)|[0-9]{2,}|...| )//g;s/is/ : /g'
done
/sbin/service named status | grep server | sed -re 's/server/named/g;s/is/:/g'
echo " "</code>

The output is :

Services

squid : running smbd : running nmbd : running 3dmd : running mysqld : running master : running Frox : running dhcpd : running sshd : running spamd : running httpd : running arpwatch : running xinetd : running named : up and running

This should work on many Red Hat flavors, if you take a look at the code, named doesn’t output a regular “service is (running|stopped)”, I needed to make a special line for this service