FlightAware Discussions

Timeout connecting to adept

There’s a few topics about this but none quite address my issue: I’m get a continuous supply of these in my logs:

[piaware] 2021/02/15 16:56:47 Connecting to FlightAware adept server at
[piaware] 2021/02/15 16:56:47 Connection with adept server at established
[piaware] 2021/02/15 16:57:02 Connection attempt with adept server at timed out
[piaware] 2021/02/15 16:57:02 reconnecting in 51 seconds...

I can telnet and connect with nc but the connection always terminates after 20-40s with a timeout. Why does the connection keep failing?

I’m using docker images from mikenye not rPi.

You have a problem with your networking.

I’d never have thought. Any idea what it might be? It pings, traceroutes etc. If I connect then should there be a message or reply or something to test it?

Port 1200 expects TLS, you could use openssl s_client to connect and see if at least the TLS handshake is completing. But if nc is timing out then you’re not even getting that far. Step 1 is to work out why TCP is broken; perhaps a firewall problem?

Thanks, that’s helpful. s_client connects and returns certificates and all that but there is a verification error: Verification error: self signed certificate in certificate chain. Connection remains open for 60s. Is all that expected?

This is on host computer not from docker container. That’s harder to test :frowning:

Here’s a workaround for the reference book:

iptables -I FORWARD -p tcp --tcp-flags SYN,RST SYN -j TCPMSS --clamp-mss-to-pmtu

That’s on docker host. It’s because docker is not handling MTU on interfaces correctly (I think)

Anyway bodge or not, I’m on line again :slight_smile:

That’s normal; we’re using an internal CA chain (shipped with piaware itself) that s_client won’t know about by default.

Ugh, path MTU again, that really is the Achilles heel of TCP.

I’ve spent days trying to sort my network out. I’ve no idea what went wrong but I know a lot more about MTU than I did. C’est la vie. Thanks for your help.