Seems to work fine with CentOS 4.4 and Plone 2.5.1. Thanks.

Barry,

Not to be picky, but I think the chkconfig line should read:

# chkconfig: 345 80 20

This way if we turn the service on using "chkconfig zope on" it will start in runlevel 3, which I usually use, especially on old, slow boxes.

Thanks,

Larry Pitcher

Suitably amended.
Thanks for the feedback,
Barry.

Does anyone think it matters that we are starting AND stopping in the order zeoserver->client1->client2 ? (I haven't had any issues with that, but then I don't have a high transaction/ busy site)
Cheers,
Barry

Zeo 'should' come up before starting clients.
Clients 'should' go down before stopping Zeo.

:)

That makes sense --- I've changed the 'stop' order. Thanks for your input!

I adapted and tested the script for Ubuntu Dapper and Plone 2.5.2:
<pre>
#!/bin/sh
# /etc/init.d/zeo
# Startup script for Zope/Plone with ZEOCluster
#
# adapted for Ubuntu paths, 2.5.2 version and LSB functions
#
# chkconfig: 345 80 20
# description: Zope, a web application server
# this works as is for a default universal plone linux install
#
# config: /opt/Plone-2.5.2/zeocluster/client$/etc/zope.conf
# LSB Source function library
. /lib/lsb/init-functions

RETVAL=0
# list zeo clients in the list below
zeoclients="client1 client2"

# this is for the default install path for 2.5.2
clusterpath="/opt/Plone-2.5.2/zeocluster"
prog="ZEOCluster"

start() {
echo -n $"Starting $prog: "
output=`${clusterpath}/server/bin/zeoctl start`
# the return status of zopectl is not reliable, we need to parse
# its output via substring match
if echo $output | grep -q "start"; then
# success
touch /var/lock/$prog
log_success_msg "zeo started successfully"
echo
RETVAL=0
else
# failed
log_failure_msg "zeo failed to start or was already started"
echo
RETVAL=1
fi
for client in $zeoclients
do
echo -n $"Starting $client: "
output=`${clusterpath}/${client}/bin/zopectl start`
# the return status of zopectl is not reliable, we need to parse
# its output via substring match
if echo $output | grep -q "start"; then
# success
touch /var/lock/zope${client}
log_success_msg "$client started successfully"
echo
RETVAL=0
else
# failed
log_failure_msg "$client failed to start or was already started"
echo
RETVAL=1
fi
done
return $RETVAL
}

stop() {

for client in $zeoclients
do
echo -n $"Stopping $client: "
output=`${clusterpath}/${client}/bin/zopectl stop`
# the return status of zopectl is not reliable, we need to parse
# its output via substring match
if echo $output | grep -q "stop"; then
# success
rm /var/lock/zope${client}
log_success_msg "$client stopped successfully"
echo
RETVAL=0
else
# failed
log_failure_msg "$client failed to stop or was already stopped"
echo
RETVAL=1
fi
done
echo -n $"Stopping $prog: "
output=`${clusterpath}/server/bin/zeoctl stop`
# the return status of zopectl is not reliable, we need to parse
# its output via substring match
if echo $output | grep -q "stop"; then
# success
rm /var/lock/$prog
log_success_msg "zeo stopped successfully"
echo
RETVAL=0
else
# failed
log_failure_msg "zeo failed to stop or was already stopped"
echo
RETVAL=1
fi
return $RETVAL
}

restart() {
stop
start
}

case "$1" in
start)
start
;;
stop)
stop
;;
status)
echo "ZEO Server:"
output=`${clusterpath}/server/bin/zeoctl status`
echo $output
for client in $zeoclients
do
echo "Zope Client" $client
output=`${clusterpath}/${client}/bin/zopectl status`
echo $output
done
;;
restart)
restart
;;
condrestart)
[ -e /var/lock/$prog ] && restart
;;
*)
echo $"Usage: $0 {start|stop|status|restart|condrestart}"
RETVAL=2
esac

exit $RETVAL
</pre>

I almost expected that the pre tags would be filtered, but even all the leading spaces are gone. It would be nice to have some formatting preserved in comments. I hope the post might still be useful.

Used it as it was, but removed the extra $'s in front of "Starting" and "Stopping" and changed 2.5.2 to 2.5.3. Here a try at keeping it with some formatting::

#!/bin/sh
# /etc/init.d/zeo
# Startup script for Zope/Plone with ZEOCluster
#
# adapted for Ubuntu paths, 2.5.3 version and LSB functions
#
# chkconfig: 345 80 20
# description: Zope, a web application server
# this works as is for a default universal plone linux install
#
# config: /opt/Plone-2.5.3/zeocluster/client$/etc/zope.conf
# LSB Source function library
. /lib/lsb/init-functions

RETVAL=0
# list zeo clients in the list below
zeoclients="client1 client2"

# this is for the default install path for 2.5.3
clusterpath="/opt/Plone-2.5.3/zeocluster"
prog="ZEOCluster"

start() {
echo -n "Starting $prog: "
output=`${clusterpath}/server/bin/zeoctl start`

# the return status of zopectl is not reliable, we need to parse
# its output via substring match
if echo $output | grep -q "start"; then
# success
touch /var/lock/$prog
log_success_msg "zeo started successfully"
echo
RETVAL=0
else
# failed
log_failure_msg "zeo failed to start or was already started"
echo
RETVAL=1
fi

for client in $zeoclients
do
echo -n "Starting $client: "
output=`${clusterpath}/${client}/bin/zopectl start`

# the return status of zopectl is not reliable, we need to parse
# its output via substring match
if echo $output | grep -q "start"; then
# success
touch /var/lock/zope${client}
log_success_msg "$client started successfully"
echo
RETVAL=0
else
# failed
log_failure_msg "$client failed to start or was already started"
echo
RETVAL=1
fi
done

return $RETVAL
}

stop() {

for client in $zeoclients
do
echo -n "Stopping $client: "
output=`${clusterpath}/${client}/bin/zopectl stop`

# the return status of zopectl is not reliable, we need to parse
# its output via substring match
if echo $output | grep -q "stop"; then
# success
rm /var/lock/zope${client}
log_success_msg "$client stopped successfully"
echo
RETVAL=0
else
# failed
log_failure_msg "$client failed to stop or was already stopped"
echo
RETVAL=1
fi
done

echo -n "Stopping $prog: "
output=`${clusterpath}/server/bin/zeoctl stop`

# the return status of zopectl is not reliable, we need to parse
# its output via substring match
if echo $output | grep -q "stop"; then
# success
rm /var/lock/$prog
log_success_msg "zeo stopped successfully"
echo
RETVAL=0
else
# failed
log_failure_msg "zeo failed to stop or was already stopped"
echo
RETVAL=1
fi

return $RETVAL
}

restart() {
stop
start
}

case "$1" in
start)
start
;;
stop)
stop
;;
status)
echo "ZEO Server:"
output=`${clusterpath}/server/bin/zeoctl status`
echo $output
for client in $zeoclients
do
echo "Zope Client" $client
output=`${clusterpath}/${client}/bin/zopectl status`
echo $output
done
;;
restart)
restart
;;
condrestart)
[ -e /var/lock/$prog ] && restart
;;
*)
echo $"Usage: $0 {start|stop|status|restart|condrestart}"
RETVAL=2
esac

exit $RETVAL

All those nice tabs ...

That's my configuration I used in my buildout-environment.
It's for plone3

I think buildout should generate an ready-to-use script somewhere - does anyone know how to implement this?
Not a such easy task because it has to take care of various configuration options, from a single instance to a zeo-cluster.

#!/bin/sh
# /etc/rc.d/init.d/zeo
# Startup script for Zope with ZEOCluster
#
# chkconfig: 345 80 20
# description: Zope, a web application server
# this works as is for a default universal plone linux install
#
# Source function library.
# config: /opt/plone3/parts/client$/etc/zope.conf
. /lib/lsb/init-functions

RETVAL=0
# list zeo clients in the list below
zeoclients="client1 client2"
# this is for the default install path for 3.x
clusterpath="/opt/plone3/parts"
prog="ZEOCluster"

start() {
echo -n $"Starting $prog: "
output=`${clusterpath}/zeoserver/bin/zeoctl start`
# the return status of zopectl is not reliable, we need to parse
# its output via substring match
if echo $output | grep -q "start"; then
# success
touch /var/lock/$prog
log_success_msg "$prog started successfully"
echo
RETVAL=0
else
# failed
log_failure_msg "$prog failed to start or was already started"
echo
RETVAL=1
fi
for client in $zeoclients
do
echo -n $"Starting $client: "
output=`${clusterpath}/${client}/bin/zopectl start`
# the return status of zopectl is not reliable, we need to parse
# its output via substring match
if echo $output | grep -q "start"; then
# success
touch /var/lock/zope${client}
log_success_msg "$prog: $client started successfully"
echo
RETVAL=0
else
# failed
log_failure_msg "$prog: $client failed to start or was already started"
echo
RETVAL=1
fi
done
return $RETVAL
}

stop() {

for client in $zeoclients
do
echo -n $"Stopping $client: "
output=`${clusterpath}/${client}/bin/zopectl stop`
# the return status of zopectl is not reliable, we need to parse
# its output via substring match
if echo $output | grep -q "stop"; then
# success
rm /var/lock/zope${client}
log_success_msg "$prog: $client stopped successfully"
echo
RETVAL=0
else
# failed
log_failure_msg "$prog: $client failed to stop or was already stopped"
echo
RETVAL=1
fi
done
echo -n $"Stopping $prog: "
output=`${clusterpath}/zeoserver/bin/zeoctl stop`
# the return status of zopectl is not reliable, we need to parse
# its output via substring match
if echo $output | grep -q "stop"; then
# success
rm /var/lock/$prog
log_success_msg "$prog stopped successfully"
echo
RETVAL=0
else
# failed
log_failure_msg "$prog failed to stop or was already stopped"
echo
RETVAL=1
fi
return $RETVAL
}

restart() {
stop
start
}

case "$1" in
start)
start
;;
stop)
stop
;;
status)
echo "ZEO Server:"
output=`${clusterpath}/zeoserver/bin/zeoctl status`
echo $output
for client in $zeoclients
do
echo "Zope Client" $client
output=`${clusterpath}/${client}/bin/zopectl status`
echo $output
done
;;
restart)
restart
;;
condrestart)
[ -e /var/lock/$prog ] && restart
;;
*)
echo $"Usage: $0 {start|stop|status|restart|condrestart}"
RETVAL=2
esac

exit $RETVAL

May we have a formatting editor here? *gg*

On Debian you must remember to run update-rc.d command or /etc/init.d changes don't become effective.

http://www.debian-administration.org/articles/28