TODO: intro
- DigitalOcean
- 512 MB RAM
- Ubuntu 14.04
SSH allows you to share ports forwarded to the server with other
remote machines, not just the server itself. By default, SSH will
open ports on the loopback address 127.0.0.1
on the server, but
can be configured to use th wildcard address 0.0.0.0
instead, which
means that we will be able to reach the forwarded port from the internet.
On the server, add the following line to /etc/ssh/sshd_config
:
GatewayPorts yes
Then restart the ssh daemon
sudo service ssh restart
We only allow the following connections:
- Connections present at the time running the setup
- Port 20 SSH
- Port 80 HTTP
On the server run:
sudo iptables -I INPUT -i lo -j ACCEPT
sudo iptables -A INPUT -m conntrack --ctstate ESTABLISHED,RELATED -j ACCEPT
sudo iptables -A INPUT -p tcp --dport ssh -j ACCEPT
sudo iptables -A INPUT -p tcp --dport http -j ACCEPT
sudo iptables -A INPUT -j DROP
sudo apt-get install iptables-persistent
sudo service iptables-persistent start
On your local machine, start a web server to test out forwarding. For example,
cd
to a directory and start a simple webserver using Python:
cd ~/lolcats/
python -m SimpleHTTPServer
# Serving HTTP on 0.0.0.0 port 8000 ...
Then set up remote port forwarding via SSH:
ssh root@<your-server> -R 80:localhost:8000
If everything is ok, ssh should now listen on tcp 0.0.0.0:80 on the server and forward requests to port 8000 on your local machine, where the Python webserver is running. You can check whether SSHD uses the correct address on the server:
netstat -tunelp
It should look like this:
Active Internet connections (only servers)
Proto Recv-Q Send-Q Local Address Foreign Address State User Inode PID/Program name
tcp 0 0 0.0.0.0:22 0.0.0.0:* LISTEN 0 10643 1223/sshd
tcp 0 0 0.0.0.0:80 0.0.0.0:* LISTEN 0 11653 1412/0
tcp6 0 0 :::22 :::* LISTEN 0 10645 1223/sshd
tcp6 0 0 :::80 :::* LISTEN 0 11654 1412/0
Now open a web browser, navigate to http://<your-server>
and you should
see the directory listing produced by the Python webserver. If so, you have
successfully forwarded the port from the remote server to your local machine.
Although I ommitted it in the previous section, because I also wanted to check
some things on the remote server, I would typically add the -N
flag to the
ssh
command in order to prevent it from starting a shell session on the
server, which really isn't necessary.
ssh -N root@<your-server> -R 80:localhost:<local-port
You might also want to add the -f
flag, which runs the command as a background
process. However, I tend to forget those processes and find it tedious to go
through the output of ps aux | grep ssh
to find unterminated port forwardings
in order to kill
them afterwards. Instead, I prefer having a command line
window open during the entire connection time, and when I am done I just hit Ctrl-C.
Maybe that's all you wanted, then you can stop reading now and be happy. However, I really wanted a command on my local machine that will allow me to spin up a new subdomain on my forwarding server with a one liner. Here's the interface I am aiming for:
fwd 1234
# creates random subdomain kh3451o5u3204.example.com on server and
# forwards requests to port 1234 on the local machine
fwd 1234 foobar
# creates random subdomain foobar.example.com on server and
# forwards requests to port 1234 on the local machine
When the command is killed with Ctrl + C, the subdomain should disappear from the server in order not to clutter the configuration.
Install nginx
sudo apt-get install nginx
- Configure nginx
server {
listen 80;
server_name dev.example.com;
location / {
proxy_pass http://127.0.0.1:10001;
}
}
- Add
/usr/local/bin/fwd
on local machine
#! /usr/bin/env ruby
require 'securerandom'
require 'yaml'
require 'net/ssh'
def ssh_exec!(ssh, command)
stdout = ''
stderr = ''
exit_code = nil
ssh.open_channel do |channel|
channel.exec(command) do |*, success|
abort "Could not execute #{command} on server." unless success
channel.on_data do |*, data|
stdout << data
end
channel.on_extended_data do |*, data|
stderr << data
end
channel.on_request('exit-status') do |*, data|
exit_code = data.read_long
end
end
end
ssh.loop
[stdout, stderr, exit_code]
end
local_port = ARGV.shift.to_i
abort 'Must be root to forward ports' if local_port < 1024 && `whoami` != 'root'
config = YAML.load_file "#{File.expand_path '~'}/.fwdrc"
abort 'No host provided in ~/.fwdrc' unless config['ssh_host']
abort 'No user provided in ~/.fwdrc' unless config['ssh_user']
ssh_host = config['ssh_host']
ssh_user = config['ssh_user']
subdomain = ARGV.shift
subdomain_length = 8
subdomain ||= SecureRandom.random_number(32**subdomain_length).to_s(32).rjust(subdomain_length, '0')
url = "http://#{subdomain}.#{ssh_host}"
system "echo '#{url}' | pbcopy"
puts 'URL copied to clipboard.'
Net::SSH.start(ssh_host, ssh_user) do |ssh|
remote_host = '0.0.0.0'
remote_port = nil
10_000.upto(11_000) do |port|
print "Checking port #{port}... "
stdout, * = ssh_exec! ssh, %Q(netstat -anp tcp | awk '$6 ~ "LISTEN" && $4 ~ "#{port}$"')
if stdout.empty?
puts 'ok.'
remote_port = port
break
else
puts 'already in use.'
end
end
abort 'Could not find an open port.' unless remote_port
ssh.forward.remote_to(local_port, 'localhost', remote_port, remote_host)
ssh.loop { !ssh.forward.active_remotes.include? [remote_port, remote_host] }
# create subdomain config
# nginx reload
puts "Now forwarding requests from #{url} to local port #{local_port}."
puts 'Hit Ctrl-C to stop.'
ssh.loop { true }
end
- Make executable
sudo chmod +x /usr/local/bin/fwd
- Add
~/.fwdrc
ssh_host: example.com
ssh_user: root
- This can now be removed again from
/etc/ssh/sshd_config
GatewayPorts yes
- Restart sshd
sudo service ssh restart