Automatically establish ssh-tunnel, wait until ssh-tunnel is established, then establish normal VPN connection

Question

I got this script:

#!/usr/bin/env bash

if [ ! "$UID" = 0 ]; then
    if [ `type -P gksu` ]; then
        SUDOAPP="gksu"
    elif [ `type -P kdesu` ]; then
        SUDOAPP="kdesu"
    else
        SUDOAPP="sudo"
    fi
fi

if [ -n "$1" ]; then
    if [ "$1" = "start" ]; then
        $SUDOAPP systemctl start openvpn@******
    elif [ "$1" = "stop" ]; then
        $SUDOAPP systemctl stop openvpn@******
    elif [ "$1" = "restart" ]; then
        $SUDOAPP systemctl restart openvpn@******
    else
        echo "Invalid command"
        exit 1
    fi
else
    echo "Run 'start', 'stop' or 'restart' as an argument to start, stop or restart the ******"
    exit 1
fi

It works fine. However I also need to establish the ssh tunnel. - Before openvpn connects to my VPN. I've got a script which does precisely that:

#!/bin/bash
# --------------------------------------------------------
# ******* | https://******.org | ****************************************
# SSH Client Configuration, Linux/OSX
# ******_*************
# --------------------------------------------------------

chmod 600 /etc/openvpn/sshtunnel.key
while :
do
echo ""; echo "****** SSH Tunnel"
ssh -i /etc/openvpn/sshtunnel.key -L ****:127.0.0.1:**** sshtunnel@**.**.**.* -p ** -N -T -v
read -t 5 -p "Retry? (or wait 5 sec for Y)" yn
if [[ $yn == "n" || $yn == "N" ]]; then break; fi
done

How do I add this to the first script in a way as to make the openvpn part wait until the ssh client is fired up?

Answer 1

The first script can loop checking the tunnel until it succeeds. You can use nc (netcat) to do that and capture the output in a shell variable:

while [[ -z "$nc_output" ]]; do
  read -r nc_output < <(nc -v -d -u localhost openvpn 2>&1)
  sleep 2
done

This checks every 2 seconds whether UDP port "openvpn" (substitute what you're actually tunnelling) can be connected to, relying on the -v option to output text if it succeeds.