Passing argument to chkconfig and having multiple process instances

NotionCommotion · 05-20-2017, 01:33 PM

Note that this is NOT a php or ssl question.

The following script will start a sockets server. The first argument passed to server.php will specify the port to be listened to, and the second argument will specify whether standard tcp or ssl is used. I start the system upon boot using chkconfig --levels 235 socketserver on and can start, stop, etc it using something like /etc/init.d/socketserver start.

Now, I want to have the server listen in both modes at the same time. Furthermore, I don't wish to hardcode the port. I was thinking of just doing /etc/init.d/socketserver start 2222 0 or /etc/init.d/socketserver start 2223 1, and using $2 and $3 in the script to send to server.php. But this won't work as it will have the same /var/run/socket_server.pid. Furthermore, I am not sure how I could pass the arguments to chkconfig --levels 235 socketserver on.

How can this be best accomplished?

filename: /etc/init.d/socketserver

Code:

#!/bin/sh
### BEGIN INIT INFO
# Provides:          Socket Server
# Required-Start:    $local_fs $network $named $time $syslog
# Required-Stop:     $local_fs $network $named $time $syslog
# Default-Start:     2 3 4 5
# Default-Stop:      0 1 6
# Short-Description: Start Socket Server
# Description:       Start Socket Server
### END INIT INFO

SCRIPT="/usr/bin/php /var/www/socket/server.php 2222 0"
//SCRIPT="/usr/bin/php /var/www/socket/server.php 2223 1"
RUNAS=Michael
PIDFILE=/var/run/socket_server.pid
LOGFILE=/var/log/socket_server.log
PROG="Socket Server"

start() {
  if [ -f "$PIDFILE" ] && kill -0 $(cat "$PIDFILE"); then
    echo 'Service already running' >&2
    return 1
  fi
  echo 'Starting service…' >&2
  #redirct I/O so non-root user doesn't need special write position to log file.
  local CMD="$SCRIPT >&3 2>&1 & echo \$!"
  su -c "$CMD" $RUNAS  3>"$LOGFILE" >"$PIDFILE"
  echo 'Service started' >&2
}

stop() {
  if [ ! -f "$PIDFILE" ] || ! kill -0 $(cat "$PIDFILE"); then
    echo 'Service not running' >&2
    return 1
  fi
  echo 'Stopping service…' >&2
  kill -15 $(cat "$PIDFILE") && rm -f "$PIDFILE"
  echo 'Service stopped' >&2
}

status() {
    if [ -f "$PIDFILE" ] && ps -p $(cat $PIDFILE) >/dev/null;
    #if [ -d /proc/$(<$PIDFILE) ];
    then
       echo "$PROG is running"
    else
       echo "$PROG is not running"
    fi
    RETVAL=$?
    return $RETVAL
}

case "$1" in
  start)
    start
    ;;
  stop)
    stop
    ;;
  restart)
    stop
    start
    ;;
  status)
    status
    ;;
  *)
    echo "Usage: $0 {start|stop|restart|status}"
esac

rknichols · 05-20-2017, 02:40 PM

In SysV init, there is no mechanism for passing additional arguments to the script during the automatic start/stop for each runlevel. When manually controlling a service with "service {name} start|stop|whatever" you can include additional arguments, but the automatic actions will never pass anything but a single "start" or "stop" argument.

The usual way to include settings that you do not wish to hard-code in the script is to source a file that contains those settings, such as the various network-related service scripts that all contain a line, ". /etc/sysconfig/network". You can code your script to start or stop multiple service instances automatically based on your config file, and still retain the ability to manually start or stop then individually based on additional arguments passed via "service socketserver {start|stop|restart|status} [arg, ...]".

michaelk · 05-20-2017, 02:47 PM

Basically "chkconfig --levels xyz myservice on" command configures what run levels the desired service automatically starts and stops. With the old init system chkconfig would create the appropriate links in the /etc/rcx.d directories i.e. KXYmy_service and/or SXYmy_service that point to the actual init script in /etc/init.d.

You can not pass actual daemon arguments via chkconfig. Instead of hard coding options most services use a configuration file of some kind and normally init scripts do not use command line arguments. I see no reason why your start up script could not start multiple instances so just create a pid/log files for each if desired.

To late again.

r3sistance · 05-20-2017, 04:36 PM

a bit hacky (but I find sysvinit is like that anyways), the only simple way I could see to do this is to create a symbolic link and use the value of $0 to determine which configuration to use... but this really is something that systemd handles way better tho.