Configuration fails to take effect about TCP Keep Alive behavior about emqx

[qzhuyan]

you could use following command to check if the value in use is correct

emqx eval 'emqx_config:get_listener_conf(tcp, default, [tcp_options, keepalive], "none")'

you could also enable debug log level , while you client connects, you should see some logs like:
2024-04-11T12:46:15.631182+02:00 [debug] msg: custom_socket_options_successfully, peername: 127.0.0.1:47138, opts: [{keepalive,true},{raw,6,4,<<5,0,0,0>>},{raw,6,5,<<12,0,0,0>>},{raw,6,6,<<3,0,0,0>>}]

if they looks ok. probably your OS/container doesn't allow setting the keep alive socket opts.

[hzms]

@qzhuyan

I can confirm that my configuration was retrieved by emqx, but it doesn't work

probably your OS/container doesn't allow setting the keep alive socket opts => the docker container is emqx:5.5.1, it's official, i didn't build it myself. If you have docker, you can try it yourself very quickly.
If this is a problem with the official docker image, I hope emqx organization can fix it.

[qzhuyan]

Could you enable debug log?

[qzhuyan]

Oh you attached the log in picture I missed it.
No debug log needed for now.

[qzhuyan]

I started the container on ubuntu 22.04 with

docker run -p 1883:1883 -e EMQX_LISTENERS__TCP__DEFAULT__TCP_OPTIONS__KEEPALIVE="5,12,3" emqx/emqx:5.5.1

connect the client with bench tool.

./emqtt_bench conn   -c 1
Start with 6 workers, addrs pool size: 1 and req interval: 60 ms 

1s connect_succ total=1 rate=0.99/sec

tcpdump shows keep alive is sent after 5s idling then another keep alive after 12s.

[qzhuyan]

test.pcap.zip
pcap is attached.