[Bug] Deutexrium is non-functional since 36037

[F0bes]

Hi there, I've managed to bisect this issue to commit #36037

03:08:32.994 [info]  channel-XYZ server: loading
 
03:08:32.994 [info]  channel-XYZ server: creating new meta
 
03:08:32.995 [error] Task #PID<0.704.0> started from #PID<0.675.0> terminating
** (stop) exited in: GenServer.call({:via, Registry, {Registry.Server, {:channel, {XYZ, ABC}}}}, {:generate, :neutral, nil}, 15000)
    ** (EXIT) no process: the process is not alive or there's no process currently associated with the given name, possibly because its application isn't started
    (elixir 1.13.4) lib/gen_server.ex:1019: GenServer.call/3
    (deutexrium 1.3.3) lib/deutexrium.ex:69: Deutexrium.handle_event/1
    (elixir 1.13.4) lib/task/supervised.ex:89: Task.Supervised.invoke_mfa/2
    (stdlib 3.14) proc_lib.erl:226: :proc_lib.init_p_do_apply/3
Function: #Function<7.75347062/0 in Deutexrium.start_link/1>
    Args: []

This is the result whenever a message is received. Thanks.

[portasynthinca3]

Interesting. I will take a look at it, thanks for the report

[portasynthinca3]

Is it possible that access permissions for the parent directory and/or the possibly existing meta_ file are set up incorrectly? I'm able to recreate this issue when artificially restricting my access to an existing meta_ file. Judging by the code, that's the only possible reason a channel GenServer could fail right after or during startup, causing the subsequent immediate call to the just-crashed server to fail too:

deutexrium/lib/deutexrium/server/channel.ex

Lines 53 to 76 in 36037c9

	def init({id, guild}) do
	# load model and meta
	Logger.info("channel-#{id} server: loading")
	meta = try do
	Meta.load!(id)
	rescue
	_ ->
	Logger.info("channel-#{id} server: creating new meta")
	Meta.dump!(id, %Meta{})
	%Meta{}
	end
	model = try do
	Model.load!(id)
	rescue
	_ ->
	Logger.info("channel-#{id} server: creating new model")
	Model.dump!(id, %Model{})
	%Model{}
	end
	Logger.info("channel-#{id} server: loaded")
	timeout = Application.fetch_env!(:deutexrium, :channel_unload_timeout)
	{:ok, {{id, guild}, meta, model, timeout}, timeout}
	end

The crash itself is not logged because this process is out of any supervisor's field of vision by design.

[F0bes]

I usually run the bot as a systemd system job (That way it just runs on my raspberry pi automatically).
I've given the /var/deutex_data (and the meta_ files) full permissions to every user, I've given the folder where I execute it full permissions, and I've ran it as the root user. I still face the same issue unfortunately.