When I set up the network_test application on my local machine and forward port 3000 to it on my router, I can hit the app from any computer on my LAN, but hitting the FB page does not connect (The connection has timed out).
Running tcpdump shows that the FB request is getting to my local machine, but the handshake never completes, so my app never sees the request. Here’s a sample tcpdump of one request (xx-xx-xxx-xxx represents my router’s external IP address, 192.168.1.103 is my dev machine’s IP address as assigned by the router):
06:19:12.263186 IP 192.168.1.103.53094 > xx-xx-xxx-xxx.lightspeed.gdrpmi.sbcglobal.net.hbci: S 1725153101:1725153101(0) win 65535 <mss 1460,nop,wscale 0,nop,nop,timestamp 133766008 0> 06:19:12.264451 IP 192.168.1.103.53094 > 192.168.1.103.hbci: S 1725153101:1725153101(0) win 65535 <mss 1460,nop,wscale 0,nop,nop,timestamp 133766008 0> 06:19:14.860129 IP 192.168.1.103.53094 > xx-xx-xxx-xxx.lightspeed.gdrpmi.sbcglobal.net.hbci: S 1725153101:1725153101(0) win 65535 <mss 1460,nop,wscale 0,nop,nop,timestamp 133766013 0> 06:19:14.861081 IP 192.168.1.103.53094 > 192.168.1.103.hbci: S 1725153101:1725153101(0) win 65535 <mss 1460,nop,wscale 0,nop,nop,timestamp 133766013 0> 06:19:17.860543 IP 192.168.1.103.53094 > xx-xx-xxx-xxx.lightspeed.gdrpmi.sbcglobal.net.hbci: S 1725153101:1725153101(0) win 65535 <mss 1460,nop,wscale 0,nop,nop,timestamp 133766019 0> 06:19:17.861613 IP 192.168.1.103.53094 > 192.168.1.103.hbci: S 1725153101:1725153101(0) win 65535 <mss 1460,nop,wscale 0,nop,nop,timestamp 133766019 0> 06:19:20.861001 IP 192.168.1.103.53094 > xx-xx-xxx-xxx.lightspeed.gdrpmi.sbcglobal.net.hbci: S 1725153101:1725153101(0) win 65535 <mss 1460> 06:19:20.861914 IP 192.168.1.103.53094 > 192.168.1.103.hbci: S 1725153101:1725153101(0) win 65535 <mss 1460> 06:19:23.861448 IP 192.168.1.103.53094 > xx-xx-xxx-xxx.lightspeed.gdrpmi.sbcglobal.net.hbci: S 1725153101:1725153101(0) win 65535 <mss 1460> 06:19:23.862332 IP 192.168.1.103.53094 > 192.168.1.103.hbci: S 1725153101:1725153101(0) win 65535 <mss 1460> 06:19:26.861963 IP 192.168.1.103.53094 > xx-xx-xxx-xxx.lightspeed.gdrpmi.sbcglobal.net.hbci: S 1725153101:1725153101(0) win 65535 <mss 1460> 06:19:26.863280 IP 192.168.1.103.53094 > 192.168.1.103.hbci: S 1725153101:1725153101(0) win 65535 <mss 1460> 06:19:32.862722 IP 192.168.1.103.53094 > xx-xx-xxx-xxx.lightspeed.gdrpmi.sbcglobal.net.hbci: S 1725153101:1725153101(0) win 65535 <mss 1460> 06:19:32.863634 IP 192.168.1.103.53094 > 192.168.1.103.hbci: S 1725153101:1725153101(0) win 65535 <mss 1460> 06:19:44.864158 IP 192.168.1.103.53094 > xx-xx-xxx-xxx.lightspeed.gdrpmi.sbcglobal.net.hbci: S 1725153101:1725153101(0) win 65535 <mss 1460> 06:19:44.865169 IP 192.168.1.103.53094 > 192.168.1.103.hbci: S 1725153101:1725153101(0) win 65535 <mss 1460> 06:20:08.866709 IP 192.168.1.103.53094 > xx-xx-xxx-xxx.lightspeed.gdrpmi.sbcglobal.net.hbci: S 1725153101:1725153101(0) win 65535 <mss 1460> 06:20:08.867661 IP 192.168.1.103.53094 > 192.168.1.103.hbci: S 1725153101:1725153101(0) win 65535 <mss 1460>
What seems to be happening is that the incoming request’s IP is being converted back to its internal value (192.168.1.103), which is what my dev machine tries to ack, but since the browser is listening for an ack on the external IP, the browser never hears it.
Switching the app configuration to FBML corrects the problem, since FB does the request itself and the IP address isn’t converted, but I’m concerned that the rest of the examples will require iframe.
I’m running network_test on my local OSX machine, which sits behind my router, which is forwarding port 3000 to it. I’m hitting the FB page from the same machine, but the behavior is the same with all the machines on my LAN.
Any suggestions? Do I need to set up the tunneling solution on page 29? I removed the tunnel code from my facebook.yml file because I thought I didn’t need it.
Thanks,
Bill