On Sat, Oct 07, 2006 at 03:04:25PM -0500, marlowe wrote: > When attempting to run Tor as a middleman server (middleman exit policy > and SocksPort 0), I consistently receive the following error message > > Oct 07 09:26:39.096 [warn] second_elapsed_callback(): Your server > (68.101.42.157:9001) has not managed to confirm that its ORPort is > reachable. Please check your firewalls, ports, address, /etc/hosts file, > etc. > > I am running Tor 0.1.1.24 installed from RPM on a CentOS 4.3 machine > behind a Linksys access point with the server as the DMZ host with all > traffic forwarded to it. The Tor recognizes my DirPort (9030) > immediately, but never the ORPort (9001). Both ports are allowed > inbound through the firewall and I have validated connection from other > Tor nodes with "netstat -an". I can connect to the ORPort on the local > machine with the following command:
When the Linksys box passes the connection on, what IP address does your Tor server think it's getting a connection from? We've got a check in there to make sure we don't consider ourselves reachable too eagerly -- if it's a connection from a local IP address, we don't buy it. But perhaps this is biting us in the case of people who port-forward. Can you look at info-level or debug-level logs and hunt through until it tells you what IP address it's getting a connection from? (If it's all gobbledygook to you, you can also mail me a debug-level log of the first 15 minutes of operation, and I'll look through it.) Thanks, --Roger
