How can I run a shell script as a daemon under Redhat?
I found a script at http://www.linuxforums.org/forum/programming-scripting/190279-daemon-etc-init-d-functions-does-not-return-launching-process.html#post897522 which I was able to modify to suit my needs. It manually tracks the PID and creates a PID file using pidof
. I ended up having to modify this to use pgrep
as pidof
was unable to see the PID of my script. After that modification, it worked fine. *Note, pgrep seems to only work if the full script name is under 15 characters long
Here's what I ended up with:
#!/bin/bash
#
#
#
# Start on runlevels 3, 4 and 5. Start late, kill early.
# chkconfig: 345 95 05
#
#
#!/bin/bash
# absolute path to executable binary
progpath='/usr/local/bin/script.sh'
# arguments to script
opts=''
# binary program name
prog=$(basename $progpath)
# pid file
pidfile="/var/run/${prog}.pid"
# make sure full path to executable binary is found
! [ -x $progpath ] && echo "$progpath: executable not found" && exit 1
eval_cmd() {
local rc=$1
if [ $rc -eq 0 ]; then
echo '[ OK ]'
else
echo '[FAILED]'
fi
return $rc
}
start() {
# see if running
local pids=$(pgrep $prog)
if [ -n "$pids" ]; then
echo "$prog (pid $pids) is already running"
return 0
fi
printf "%-50s%s" "Starting $prog: " ''
$progpath $opts &
# save pid to file if you want
echo $! > $pidfile
# check again if running
pgrep $prog >/dev/null 2>&1
eval_cmd $?
}
stop() {
# see if running
local pids=$(pgrep $prog)
if [ -z "$pids" ]; then
echo "$prog not running"
return 0
fi
printf "%-50s%s" "Stopping $prog: " ''
rm -f $pidfile
kill -9 $pids
eval_cmd $?
}
status() {
# see if running
local pids=$(pgrep $prog)
if [ -n "$pids" ]; then
echo "$prog (pid $pids) is running"
else
echo "$prog is stopped"
fi
}
case $1 in
start)
start
;;
stop)
stop
;;
status)
status
;;
restart)
stop
sleep 1
start
;;
*)
echo "Usage: $0 {start|stop|status|restart}"
exit 1
esac
exit $?