Open Stack

Hello ops,
I'm following the docs for setting novnc_proxy, and starting to get frustrated with the constant "Server disconnected (code: 1006)" message.

I've setup nova.conf on my controller as:
novncproxy_base_url=http://${public_ip}:6080/vnc_auto.html
novncproxy_host=${public_ip}
novncproxy_port=6080

my compute nodes have
novncproxy_base_url=http://${public_ip_of_the_controller}:6080/vnc_auto.html
vncserver_listen=${internal ip of this compute node}
vncserver_proxyclient_address=${internal ip of this compute node}

If I tail the nova-consoleauth.log while attempting to get a console, I see the correct internal IP for the compute node, and port for the VM.
doing tcpdump on that port on the compute node shows nothing.

if I telnet to the internal IP and vnc port, I get 
Escape character is '^]'.
RFB 003.008
^]close

and finally I see packets on the tcpdump.

This is using ubuntu 12.04/grizzly

nova-novncproxy                     1:2013.1.2-0ubuntu1~cloud0 
novnc                               2012.2~20120906+dfsg-0ubuntu4~cloud0
python-novnc                        2012.2~20120906+dfsg-0ubuntu4~cloud0

Any help or tips would be most appreciated.