Service with respawn, possible?

English Programming/Scripting
1

Hi, everybody!

I,m trying to reach with goals new service :

  1. On start - run two instances of “socat” with parameters; chmod 777 them;
  2. On stop - kill only those two instances of “socat”;
  3. On restart - stop and start service;
  4. On status - check if “socat” instances are running and if not - start them.

Here is my “service”:


#!/bin/bash
    # chkconfig: - 98 02
    # description: socat service
    # processname: socat

    KIND="Socat"
    FsIn='/usr/bin/socat PTY,link=/dev/ttyS0,raw,echo=0 tcp4:192.168.0.188:7000'
    ScIn='/usr/bin/socat PTY,link=/dev/ttyS1,raw,echo=0 tcp4:192.168.0.188:7001'
    Port1=/dev/ttyS0
    Port2=/dev/ttyS1
    start() {
            echo -n $"Starting $KIND services: "
            $FsIn &
            $ScIn &
            chmod 777 $Port1
            chmod 777 $Port2
            echo
    }

    stop() {
            echo -n $"Shutting down $KIND services: "
            killproc $FsIn
            killproc $ScIn
            echo
    }

    restart() {
                echo -n $"Restarting $KIND services: "
                killproc $FsIn
                killproc $ScIn
                $FsIn &
                $ScIn &
                chmod 777 $Port1
                chmod 777 $Port2
                echo
    }

    status() {
     TEST='ps ax | grep -c $FsIn';
     if  "$TEST" = "2" ];
     then
        echo "Alive :)"
     else
        echo "Dead :("
     fi
    }

    case "$1" in
      start)
              start
            ;;
      stop)
              stop
            ;;
      restart)
              restart
            ;;

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

So, “status” section isn’t working :frowning: I have tried a lot of modifications, this is just one of them.
Can service watch for “socat” (maybe some “while” loop) instances and respawn them if they dies?
Thanks in advance!

2

Hi
If you look at /etc/init.d/skeleton for reference, look at touching a
pid file in /var/run and getting status from that.

–
Cheers Malcolm °¿° (Linux Counter #276890)
SUSE Linux Enterprise Desktop 11 (x86_64) Kernel 2.6.27.29-0.1-default
up 16:33, 2 users, load average: 0.07, 0.12, 0.28
GPU GeForce 8600 GTS Silent - CUDA Driver Version: 190.18

3

Thanks for answer. But still no joy. Here is a part of my script:


start)
            echo -n $"Starting socat services: "
            Temp1=/var/run/socat1.pid
            FsIn='/usr/bin/socat PTY,link=/dev/ttyS0,raw,echo=0 tcp4:192.168.1.105:7000'
            Chk1=$(ps -fe | grep "$FsIn" | head -n1|cut -d" " -f 7)
            $FsIn &
            echo $Chk1 > $Temp1
            
          
 # Remember status and be verbose 
 rc_status -v
 ;;

I want to save the PID to /var/run/socat1.pid. If I exec it in console line by line, it works well. But when I’ll try to exec the script, then nothing wrote to socat1.pid file.

P.S. Sorry for English, I’m in a hurry

4

Look at using /etc/inittab. Web search for inittab.

5

I have tried inittab, but in some cases it can’t respawn my process. For example, if I unplug network cable from comp and plug in it in 10 seconds. Only “init q” can help me in that case :frowning:

6

So you write a wrapper that init calls to detect anomalous conditions and exit so that init can restart it. Maybe even create a watchdog process to monitor the work process. Thing is you end up writing your own respawn loop anyway if you write your own SysV init script instead of using inittab.

7

Thanks for answer. I wrote a new script with infinity loop of “init q” and “sleep 30”. In this way all working well by the first look