CentOS 6.9下通过init.d运行PHP脚本异常问题咨询

Hey there, let's work through the two issues you're hitting with running a PHP script via an init.d service on CentOS 6.9. I've run into these exact problems before, so here's a step-by-step fix for each one:

This happens because your PHP script is still attached to the terminal session. When you close the terminal, it sends a SIGHUP signal that kills the process. To fix this, you need to run the script as a daemon (background process detached from the terminal).

Here are two reliable ways to do this in your init.d script:

• Use nohup to ignore the SIGHUP signal, redirect output to a log file, and send the process to the background:

  nohup php /path/to/your/script.php > /var/log/your_script.log 2>&1 &
  

  The 2>&1 part redirects error output to the same log file as standard output, so you can debug issues later.
• Alternatively, use setsid to create a new session for the process, completely detaching it from your terminal:

  setsid php /path/to/your/script.php > /var/log/your_script.log 2>&1
  

This is because your init.d script isn't tracking the process ID (PID) of the running PHP script. By default, init.d thinks the service stops when the foreground command finishes—but since we're running PHP in the background, the script exits immediately, leading to a "stopped" status.

To fix this, you need to add PID file management to your init.d script:

Step 1: Define the PID file path #

At the top of your init.d script, add a line to specify where the PID file will live:

PIDFILE="/var/run/your_script.pid"

Make sure the directory /var/run/ is writable by the user running the PHP script.

Step 2: Update the start function #

When starting the script, write the process's PID to the file:

start() {
    echo -n "Starting your PHP script: "
    # Run the script in the background and save its PID
    nohup php /path/to/your/script.php > /var/log/your_script.log 2>&1 &
    echo $! > $PIDFILE
    echo "done"
}

The $! variable holds the PID of the last background process.

Step 3: Add a stop function to kill the process using the PID #

stop() {
    echo -n "Stopping your PHP script: "
    if [ -f $PIDFILE ]; then
        # Kill the process using the PID from the file
        kill $(cat $PIDFILE)
        # Clean up the PID file
        rm -f $PIDFILE
        echo "done"
    else
        echo "not running"
    fi
}

Step 4: Add a status function to check if the process is active #

status() {
    if [ -f $PIDFILE ]; then
        PID=$(cat $PIDFILE)
        # Check if the PID corresponds to a running process
        if ps -p $PID > /dev/null; then
            echo "PHP script is running with PID $PID"
        else
            echo "PID file exists but process is not running"
            # Clean up stale PID file
            rm -f $PIDFILE
        fi
    else
        echo "PHP script is not running"
    fi
}

Step 5: Update the case statement to handle the status command #

Make sure your script's main case block includes the status option:

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

• Don't forget to make your init.d script executable:

  chmod +x /etc/init.d/your_script_name
  

• If you want the service to start on boot, register it with chkconfig:

  chkconfig --add your_script_name
  chkconfig your_script_name on
  

• Double-check permissions on your log file to ensure the PHP process can write to it:

  chown your_script_user:your_script_group /var/log/your_script.log
  

  (Replace your_script_user and your_script_group with the actual user/group running your PHP script—often apache or www-data.)

内容的提问来源于stack exchange，提问作者Som