Confirmed immediate downlink

[citruspress]

When sending a confirmed immediate downlink message like the following: '{"data":"04000000000001000000", "confirmed":true, "port": 2, "time":"immediately"}'

We never get a confirmation from the device and all following messages are just "re-up" (red arrows).

Sending unconfirmed immediate doesn't work but doesn't cause any errors on the following messages afterwards.

Sending class A messages works as expected but the device is a class C device so we prefer immediate messages.

The log file is pasted below.

2020-02-28 16:48:02.122 [info] <0.6523.1>@lorawan_gw_forwarder:terminate:131 packet_forwarder interface terminated: {function_clause,[{lorawan_gw_forwarder,build_txpk,[{txq,869.525,<<"SF9BW125">>,<<"4/5">>,immediately,16},undefined,0,<<160,51,204,187,170,0,22,0,2,251,85,206,121,45,214,22,251,28,13,236,174,121,49>>],[{file,"/lorawan-server/src/lorawan_gw_forwarder.erl"},{line,191}]},{lorawan_gw_forwarder,handle_info,2,[{file,"/lorawan-server/src/lorawan_gw_forwarder.erl"},{line,112}]},{gen_server,try_dispatch,4,[{file,"gen_server.erl"},{line,616}]},{gen_server,handle_msg,6,[{file,"gen_server.erl"},{line,686}]},{proc_lib,init_p_do_apply,3,[{file,"proc_lib.erl"},{line,247}]}]} 2020-02-28 16:48:02.122 [error] <0.6523.1>@lorawan_gw_forwarder:build_txpk:191 gen_server lorawan_gw_forwarder terminated with reason: no function clause matching lorawan_gw_forwarder:build_txpk({txq,869.525,<<"SF9BW125">>,<<"4/5">>,immediately,16}, undefined, 0, <<160,51,204,187,170,0,22,0,2,251,85,206,121,45,214,22,251,28,13,236,174,121,49>>) line 191 2020-02-28 16:48:02.123 [error] <0.6523.1>@lorawan_gw_forwarder:build_txpk:191 CRASH REPORT Process lorawan_gw_forwarder with 0 neighbours crashed with reason: no function clause matching lorawan_gw_forwarder:build_txpk({txq,869.525,<<"SF9BW125">>,<<"4/5">>,immediately,16}, undefined, 0, <<160,51,204,187,170,0,22,0,2,251,85,206,121,45,214,22,251,28,13,236,174,121,49>>) line 191 2020-02-28 16:48:02.123 [error] <0.3547.1>@lorawan_gw_forwarder:build_txpk:191 Supervisor {<0.3547.1>,lorawan_gw_sup} had child packet_forwarder started with lorawan_gw_forwarder:start_link([{port,1680}]) at <0.6523.1> exit with reason no function clause matching lorawan_gw_forwarder:build_txpk({txq,869.525,<<"SF9BW125">>,<<"4/5">>,immediately,16}, undefined, 0, <<160,51,204,187,170,0,22,0,2,251,85,206,121,45,214,22,251,28,13,236,174,121,49>>) line 191 in context child_terminated 2020-02-28 16:48:02.123 [debug] <0.3547.1> Supervisor {<0.3547.1>,lorawan_gw_sup} started lorawan_gw_forwarder:start_link([{port,1680}]) at pid <0.6839.1> 2020-02-28 16:50:47.377 [warning] <0.7008.1> node AABBCC33 downlink_missed 2020-02-28 16:50:47.578 [debug] <0.7008.1>@lorawan_handler:retransmit:236 AABBCC33 retransmitting 2 2020-02-28 16:51:03.862 [warning] <0.7016.1> node AABBCC33 downlink_missed 2020-02-28 16:51:04.063 [debug] <0.7016.1>@lorawan_handler:retransmit:236 AABBCC33 retransmitting 3 2020-02-28 16:51:20.513 [warning] <0.7027.1> node AABBCC33 downlink_missed 2020-02-28 16:51:20.714 [debug] <0.7027.1>@lorawan_handler:retransmit:236 AABBCC33 retransmitting 4 2020-02-28 16:51:54.629 [warning] <0.7038.1> node AABBCC33 downlink_missed 2020-02-28 16:51:54.831 [debug] <0.7038.1>@lorawan_handler:retransmit:236 AABBCC33 retransmitting 5 2020-02-28 16:52:39.886 [warning] <0.7053.1> node AABBCC33 downlink_missed 2020-02-28 16:52:40.087 [debug] <0.7053.1>@lorawan_handler:retransmit:236 AABBCC33 retransmitting 6 2020-02-28 16:53:25.615 [warning] <0.7091.1> node AABBCC33 downlink_missed 2020-02-28 16:53:25.816 [debug] <0.7091.1>@lorawan_handler:retransmit:236 AABBCC33 retransmitting 7

[WillooWisp]

I experience the same problem and according to the log there is a crash in the forwarder. Any known issue or tip on how to solve this?

CRASH REPORT Process lorawan_gw_forwarder with 0 neighbours crashed with reason: no function clause matching lorawan_gw_forwarder:build_txpk

[citruspress]

Turns out this is a separate issue. It works as expected on stable.

I thought we were running stable but since the Dockerfile clones the git repository without checking out any branches the stable image is built on the latest branch.

Maybe this is a separate discussion, but I think the Dockerfile should use COPY instead of git clone, is there a reason for cloning the repository again?

Would you be interested in a pull request changing that and looking into adding multi stage build (https://docs.docker.com/develop/develop-images/multistage-build/) to reduce the image size?

[gotthardp]

Yes, pull requests are always welcome. There is no reason for not using COPY, I guess. It's just my poor Docker expertise.