Had some problems with the original and the check-script posted by John on 15 Jan 2010. Got a fix that worked, though:
Replaced:

nc -w1 localhost $tunnel_entrance_port
with:
lsof -i :$tunnel_entrance_port

The first passed an error every time, so the original script always tried to re-create the tunnel(s). lsof -i checks for a process listening to a particular port. I figured that if the port is already open, ssh won’t be able to create the tunnel anyway.

I’m on my laptop, I want to run GUI that is on serverB.
I do not have direct access to serverB. I do, however, have access to serverA and from there can get to Server B.

For comman line stuff…I use putty to get to ServerA, and ssh to ServerB. How do I set up tunnel to be able to use VNCclient on my laptop to run vncserver on ServerB.

I know how to create putty tunnel to serverA in order to run vncserver on serverA. But how to do it when I have to get to ServerB thru serverA.

Anyone?

#!/bin/bash

tunnel_entrance_port=$1
tunnel_end=$2
destination=$3
destination_port=$4

## use nc to connect to tunnel entrance port. If it returns non-zero, then create a new connection
nc -w1 localhost $tunnel_entrance_port
if [[ $? -ne 0 ]]; then
echo Creating new tunnel connection
/usr/bin/ssh -f -N -L $tunnel_entrance_port:$destination:$destination_port $tunnel_end
if [[ $? -eq 0 ]]; then
echo Tunnel to $destination through $tunnel_end created successfully
else
echo An error occurred creating a tunnel to $destination through $tunnel_end was $?
fi
fi

Thanks for the information

It really works for me ..

Another issue that´s confusing me is that u don´t use the -i option in your ssh-commands:
/usr/bin/ssh -p 19922 tunnel@localhost ls
AND
/usr/bin/ssh -f -N -L13306:hostb:3306 -L19922:hostb:22 tunnel@hostb
Shouldn´t it be for example: “/usr/bin/ssh -p 19922 – i /home/tunnel/.ssh/id_rsa tunnel@localhost ls” ???

Greets from spain