Misconfigured Dual Stack (IPv6 not responding) causes issues with non-happy-eyeballs clients
ctheune opened this issue · comments
Christian Theune commented
I've started talking to the API using Python's requests
library and noticed that it takes a long time to connect. There is a AAAA
record for v6.db.transport.rest
so this gets talked to with a higher priority until the non-responsiveness falls back to IPv4 after around 60-90 seconds.
I'm not quite sure whether this is broken on your provider's side or the service itself. Routing looks like this and isn't in the target network but close:
chazzzzz (2003:a:745:af0a:3028:8704:ffe1:4d34) -> v6.db.transport.rest (2a01:4f8:c012:91b7::) 2024-01-03T16:46:25+0100
Keys: Help Display mode Restart statistics Order of fields quit
Last 179 pings
1. 2003:a:745:af0a::1 ...?.........
2. 2003:0:1701:a409::1 ....>........
3. 2003:0:1701:e400::2 .>>>.........
4. 2003:0:1902:4000::1 .............
5. 2003:0:f00::209 .............
6. 2a01:4f8:0:3::611 ............
7. (waiting for reply) ????????????