0224 fix docker entrypoint env overrides

[zmstone]

No description provided.

[Rory-Z]

I think we should add EMQX_NODE_NAME to etc/emqx.conf, because when the user executes bin/emqx ctl in the container, bin/emqx may not get the environment variables from docker-entrypoint.sh

[Rory-Z]

Suggested change

	EMQX_LISTENERS__SSL__default__ENABLE <--> listeners.ssl.default.enable
	EMQX_LISTENERS__SSL__DEFAULT__ENABLE <--> listeners.ssl.default.enable

[zmstone]

this is intentionally left lower case to demo that lower case letters also work.

[Rory-Z]

Cool, maybe add some note, it's easy to get confused

[Rory-Z]

Looks EMQX 4.4 has same issue

[zmstone]

I think we should add EMQX_NODE_NAME to etc/emqx.conf, because when the user executes bin/emqx ctl in the container, bin/emqx may not get the environment variables from docker-entrypoint.sh

The node name is retrieved from parsing the ps -ef command output before executing the ctl command.

[zmstone]

Looks EMQX 4.4 has same issue

For 4.4 the node name is retrieved from the latest vm.<time>.args file.
If the node is never initialized or not running when executing non-boot commands, the command is doomed to fail anyway.

[Rory-Z]

I think we should add EMQX_NODE_NAME to etc/emqx.conf, because when the user executes bin/emqx ctl in the container, bin/emqx may not get the environment variables from docker-entrypoint.sh

The node name is retrieved from parsing the ps -ef command output before executing the ctl command.

When I create a EMQX container like this: docker run -d --name emqx -e EMQX_NODE_NAME="emqx@fake" emqx/emqx:5.0.14
check log, the EMQX looks good:

$ docker logs -f emqx
WARNING: Default (insecure) Erlang cookie is in use.
WARNING: Configure node.cookie in /opt/emqx/etc/emqx.conf or override from environment variable EMQX_NODE__COOKIE
WARNING: Use the same config value for all nodes in the cluster.
EMQX_RPC__PORT_DISCOVERY [rpc.port_discovery]: manual
EMQX_LOG__FILE_HANDLERS__DEFAULT__ENABLE [log.file_handlers.default.enable]: false
EMQX_LOG__CONSOLE_HANDLER__ENABLE [log.console_handler.enable]: true
EMQX_NODE__NAME [node.name]: emqx@fake
Listener ssl:default on 0.0.0.0:8883 started.
Listener tcp:default on 0.0.0.0:1883 started.
Listener ws:default on 0.0.0.0:8083 started.
Listener wss:default on 0.0.0.0:8084 started.
Listener http:dashboard on :18083 started.
EMQX 5.0.14 is running now!

the ps -ef looks is normal

emqx@73c6ca2d0ba5:/opt/emqx$ ps -ef |grep emqx
emqx           1       0  7 06:37 ?        00:00:25 emqx -Bd -spp true -A 4 -IOt 4 -SDio 8 -e 262144 -zdbbl 8192 -Q 1048576 -P 2097152 -- -root /opt/emqx -progname /opt/emqx/bin/emqx -- -home /home/emqx -- -noshell -noshell -noinput -boot /opt/emqx/releases/5.0.14/start -boot_var RELEASE_LIB /opt/emqx/lib -boot_var ERTS_LIB_DIR /opt/emqx/lib -mode embedded -config /opt/emqx/data/configs/app.2023.03.01.06.37.31.config -stdlib restricted_shell emqx_restricted_shell -kernel net_ticktime 120 -shutdown_time 30000 -pa data/patches -mnesia dump_log_write_threshold 5000 -mnesia dump_log_time_threshold 60000 -pa /opt/emqx/releases/5.0.14/consolidated -setcookie emqxsecretcookie -name emqx@fake -mnesia dir "/opt/emqx/data/mnesia/emqx@fake" -- -start_epmd false -epmd_module ekka_epmd -proto_dist ekka -- foreground -emqx_data_dir /opt/emqx/data --

but emqx ctl is not working:

$ docker exec -it emqx bash
emqx@73c6ca2d0ba5:/opt/emqx$ emqx ctl status
Node emqx@fake not responding to pings.
ERROR: node_is_not_running!
emqx@73c6ca2d0ba5:/opt/emqx$ emqx_ctl status
Node emqx@fake not responding to pings.
ERROR: node_is_not_running!

The EMQX opensource 4.4.14 has same issue

[zmstone]

emqx@fake

in this example, emqx ctl actually managed to find the node name.
but failed to ping the node because fake is not a valid hostname.