When working with legacy systems like CentOS 6.4, you might encounter the frustrating "daemon: command not found" error while trying to run processes under different users. Let's dive into why this happens and how to properly implement service daemonization.

The daemon function is actually a shell function defined in /etc/init.d/functions, not a standalone executable. The error occurs because:

1. The script isn't properly sourcing the functions file
2. The function might not be available in your shell environment
3. There could be permission issues with the functions file

Here's how your init script should be structured:

#!/bin/bash
#
# mydaemon      Start/Stop my custom daemon
#
# chkconfig: 345 20 80
# description: My custom service

# Source function library
. /etc/init.d/functions

start() {
    echo -n "Starting mydaemon: "
    daemon --user=someuser --pidfile=/var/run/mydaemon.pid \
           /path/to/mydaemon >/dev/null 2>&1 &
    RETVAL=$?
    [ $RETVAL -eq 0 ] && touch /var/lock/subsys/mydaemon
    echo
    return $RETVAL
}

If sourcing the functions file doesn't work, consider these alternatives:

Using start-stop-daemon

start-stop-daemon --start --quiet --background \
    --chuid someuser --exec /path/to/mydaemon \
    --pidfile /var/run/mydaemon.pid

Manual Daemonization

start() {
    echo -n "Starting mydaemon: "
    nohup sudo -u someuser /path/to/mydaemon >/dev/null 2>&1 &
    echo $! > /var/run/mydaemon.pid
    echo "OK"
}

When troubleshooting:

- Verify the functions file exists: ls -l /etc/init.d/functions
- Check file permissions: should be 644 root:root
- Test sourcing manually: . /etc/init.d/functions && type daemon
- Examine environment variables: env | grep PATH
- Verify the user exists: id someuser

Remember that CentOS 6.4 is quite old (released in 2013). For modern systems, consider:

- Upgrading to CentOS 7+ which uses systemd
- Using modern process managers like supervisord
- Containerizing the application with Docker

When working with init scripts on CentOS 6.4, you might encounter the frustrating "daemon: command not found" error, especially when trying to daemonize processes under different users. This commonly occurs when the system can't properly source the necessary functions.

The daemon() function is actually declared in /etc/init.d/functions, but your script might not be loading it properly. Here's what's likely missing:

# Make sure this line exists at the top of your init script
. /etc/init.d/functions

Here's a corrected version of your service script:

#!/bin/sh
#
# mydaemon     Start/Stop the mydaemon service
#
# chkconfig: 2345 90 10
# description: My custom daemon service

. /etc/init.d/functions

start() {
    echo -n "Starting mydaemon: "
    daemon --user someuser --name mydaemon /path/to/mydaemon
    RETVAL=$?
    echo
    return $RETVAL
}

stop() {
    echo -n "Shutting down mydaemon: "
    killproc mydaemon
    RETVAL=$?
    echo
    return $RETVAL
}

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

If you're still having issues, consider these alternatives:

1. Using start-stop-daemon:

start-stop-daemon --start --quiet --pidfile /var/run/mydaemon.pid \
--chuid someuser --exec /path/to/mydaemon

2. Direct nohup approach:

su - someuser -c "nohup /path/to/mydaemon > /dev/null 2>&1 &"

• Verify the script has execute permissions: chmod +x /etc/init.d/mydaemon
• Check that the someuser account exists
• Ensure SELinux isn't blocking the operation
• Test with absolute paths for all commands