Find something simple (like the syslog startup script) and hack it into a xymon startup script. For the start command, "su - hobbit -c <wherever>runclient.sh".

Here is an example that I use:

#!/sbin/sh

Greg Hubbard, EDS, January 2006

Made from Solaris syslog startup script

BBHOME=/app/hobbit/client; export BBHOME BBUSER=xymon; export BBUSER HOST="uname -n.domain.com"

case "$1" in 'start') if [ -f $BBHOME/runclient.sh ]; then echo 'Hobbit client starting.' # su - $BBUSER -c "cd $BBHOME; ./runclient.sh --hostname=$HOST start" fi ;;

'stop') if [ -f $BBHOME/runclient.sh ]; then

            su - $BBUSER -c "cd $BBHOME; ./runclient.sh stop"
    fi
    ;;

*) echo "Usage: $0 { start | stop }" exit 1 ;; esac This is for a client. The server startup is similar, except it runs the server startup script. This is not the most clever or elaborate of scripts, but it does work for me.

From: David Peters [mailto:davidp at electronf.com] Sent: Friday, June 05, 2009 5:09 AM To: hobbit at hswn.dk Subject: [hobbit] Hobbit startup script.

Hi,

I understand that hobbit needs to run as the hobbit user, but it is a bit annoying to have the hobbit startup script require the hobbit user to run it.

When I want to run it automagically out of /etc/init.d I have to change the script to not check for hobbit and then do an su to the hobbit user thus:

su -c "/home/xymon/server/bin/hobbitlaunch --config=/home/xymon/server/etc/hobbitlaunch.cfg --env=/home/xymon/server/etc/hob bitserver.cfg --log=/var/log/xymon/hobbitlaunch.log --pidfile=/var/log/xymon/hobbitlaunch.pid"

why is it done this way rather than as above, thus allowing the server to start out of init.d?