I added ecto to ejabberd, which worked fine in dev mode (run with iex -S mix), but when ejabberd is compiled and started I get this error:
07:02:09.214 [error] Failed to start module <module-name>:
{error,
{shutdown,
{failed_to_start_child,'Elixir.Ejabberd.Repo',
{undef,
[{'Elixir.Ecto.Adapters.Postgres',init,
[[{repo,'Elixir.Ejabberd.Repo'},
{telemetry_prefix,[ejabberd,repo]},
{otp_app,ejabberd},
{timeout,15000},
{pool_size,10}]],
[]},
{'Elixir.Ecto.Repo.Supervisor',init,1,
[{file,"lib/ecto/repo/supervisor.ex"},{line,162}]},
{supervisor,init,1,[{file,"supervisor.erl"},{line,295}]},
{gen_server,init_it,2,[{file,"gen_server.erl"},{line,374}]},
{gen_server,init_it,6,[{file,"gen_server.erl"},{line,342}]},
{proc_lib,init_p_do_apply,3,
[{file,"proc_lib.erl"},{line,249}]}]}}}}
I'm starting Ejabberd.Repo from custom ejabberd module:
def init(:ok) do
children = [
worker(Ejabberd.Repo, [])
]
Supervisor.init(children, strategy: :one_for_one)
end
def start(host, _opts) do
info("Starting ejabberd module #{__MODULE__} at #{host}")
:gen_iq_handler.add_iq_handler(:ejabberd_local, host, @namespace, __MODULE__, :on_iq_received)
Supervisor.start_link(__MODULE__, :ok, [])
end
Ejabberd.Repo is defined in lib/ejabberd/repo.ex:
defmodule Ejabberd.Repo do
use Ecto.Repo,
otp_app: :ejabberd,
adapter: Ecto.Adapters.Postgres
end
connection setup inisde config/config.exs:
config :ejabberd, ecto_repos: [Ejabberd.Repo]
config :ejabberd, Ejabberd.Repo,
adapter: Ecto.Adapters.Postgres,
database: "<database-name>",
username: "<username>",
password: "<password>",
hostname: "<host>"
application setup from mix.exs:
def application do
[mod: {:ejabberd_app, []},
applications: [:kernel, :stdlib, :sasl, :ssl],
included_applications: [:lager, :mnesia, :inets, :p1_utils, :cache_tab,
:fast_tls, :stringprep, :fast_xml, :xmpp, :mqtree,
:stun, :fast_yaml, :esip, :jiffy, :p1_oauth2,
:eimp, :base64url, :jose, :pkix, :os_mon, :ecto, :postgrex, :telemetry]
++ cond_apps()]
end
deps from mix.exs:
defp deps do
[
{:lager, "~> 3.6.0"},
{:p1_utils, "~> 1.0"},
{:fast_xml, "~> 1.1"},
{:xmpp, "~> 1.3.0"},
{:cache_tab, "~> 1.0"},
{:stringprep, "~> 1.0"},
{:fast_yaml, "~> 1.0"},
{:fast_tls, "~> 1.1"},
{:stun, "~> 1.0"},
{:esip, "~> 1.0"},
{:p1_mysql, "~> 1.0"},
{:mqtree, "~> 1.0"},
{:p1_pgsql, "~> 1.1"},
{:jiffy, "~> 0.14.7"},
{:p1_oauth2, "~> 0.6.1"},
{:distillery, "~> 2.0"},
{:pkix, "~> 1.0"},
{:ex_doc, ">= 0.0.0", only: :dev},
{:eimp, "~> 1.0"},
{:base64url, "~> 0.0.1"},
{:jose, "~> 1.8"},
{:ecto_sql, "~> 3.0"},
{:postgrex, ">= 0.0.0"}
]
++ cond_deps()
end
link to official docker repo (https://github.com/processone/docker-ejabberd/tree/master/ecs), only change to docker file is that it is copying files from local version instead cloning it
I guess ejabberd cannot find configuration for ecto, but I don't know where to put it. I've tried adding setup to rel/prod.exs, but it doesn't make a difference.
Am I missing something? How can I make more sense of this error?
credits to @legoscia, release was not including beam file for Elixir.Ecto.Adapters.Postgres, adding ecto_sql to included_applications solved this issue. Application setup in mix.exs:
def application do
[mod: {:ejabberd_app, []},
applications: [:kernel, :stdlib, :sasl, :ssl],
included_applications: [:lager, :mnesia, :inets, :p1_utils, :cache_tab,
:fast_tls, :stringprep, :fast_xml, :xmpp, :mqtree,
:stun, :fast_yaml, :esip, :jiffy, :p1_oauth2,
:eimp, :base64url, :jose, :pkix, :os_mon, :ecto_sql, :ecto, :postgrex, :telemetry]
++ cond_apps()]
end
It seems that applications and included_applications define which applications will be bundled in release