If you folks have a stacktrace, can you please share it?
Also, does it happen in both OTP 27 and 28?
josevalim
Creator of Elixir
stevegrossi
Sure thing, I’ve confirmed it happens on both elixir 1.19.1-otp-28/erlang 28.1.1 as well as elixir 1.19.1-otp-27/erlang 27.3.4.4
This is the stacktrace (actual application name replaced with my_app):
Server: dev.localhost:4000 (http)
Request: GET /campaigns
** (exit) an exception was raised:
** (RuntimeError) could not lookup Ecto repo MyApp.Repo because it was not started or it does not exist
(ecto 3.13.4) lib/ecto/repo/registry.ex:22: Ecto.Repo.Registry.lookup/1
(ecto 3.13.4) lib/ecto/repo/supervisor.ex:176: Ecto.Repo.Supervisor.tuplet/2
(my_app 0.0.1) lib/my_app/repo.ex:4: MyApp.Repo.get_by/3
(my_app 0.0.1) lib/my_app/teams/teams.ex:697: MyApp.Teams.load_team_from_subdomain/1
(my_app 0.0.1) lib/my_app_web/plugs/load_team_from_subdomain.ex:16: MyAppWeb.Plugs.LoadTeamFromSubdomain.call/2
(my_app 0.0.1) MyAppWeb.Router.with_team/2
(my_app 0.0.1) lib/my_app_web/router.ex:1: MyAppWeb.Router.__pipe_through4__/1
(phoenix 1.8.1) lib/phoenix/router.ex:407: Phoenix.Router.__call__/5
(my_app 0.0.1) lib/my_app_web/endpoint.ex:1: MyAppWeb.Endpoint.plug_builder_call/2
(my_app 0.0.1) lib/my_app_web/endpoint.ex:1: MyAppWeb.Endpoint."call (overridable 3)"/2
(my_app 0.0.1) deps/plug/lib/plug/debugger.ex:155: MyAppWeb.Endpoint."call (overridable 4)"/2
(my_app 0.0.1) lib/my_app_web/endpoint.ex:1: MyAppWeb.Endpoint.call/2
(phoenix 1.8.1) lib/phoenix/endpoint/sync_code_reload_plug.ex:22: Phoenix.Endpoint.SyncCodeReloadPlug.do_call/4
(plug_cowboy 2.7.4) lib/plug/cowboy/handler.ex:11: Plug.Cowboy.Handler.init/2
(cowboy 2.14.2) /Users/steve/Work/monorepo/my_app/deps/cowboy/src/cowboy_handler.erl:37: :cowboy_handler.execute/2
(cowboy 2.14.2) /Users/steve/Work/monorepo/my_app/deps/cowboy/src/cowboy_stream_h.erl:310: :cowboy_stream_h.execute/3
(cowboy 2.14.2) /Users/steve/Work/monorepo/my_app/deps/cowboy/src/cowboy_stream_h.erl:299: :cowboy_stream_h.request_process/3
(stdlib 7.1) proc_lib.erl:333: :proc_lib.init_p_do_apply/3
file=application_controller.erl line=2129 [notice] Application my_app exited: shutdown
0
jozuas
Ooh, I tried over and over again, but I cannot replicate with OTP27. OTP27 still shows that there’s a rebuild after iex start, but it does not result in 700+ files being recompiled unlike with OTP28:
iex --name "tp@127.0.0.1" --cookie dev -S mix run
Erlang/OTP 27 [erts-15.2.7.2] [source] [64-bit] [smp:12:12] [ds:12:12:10] [async-threads:1] [jit]
20:36:43.734 pid=<0.274.0> module=Appsignal function=initialize/0 line=89 [info] AppSignal disabled.
20:36:44.181 pid=<0.649.0> module=Bandit function=start_link/1 line=321 [info] Running TpWeb.Endpoint with Bandit 1.8.0 at 0.0.0.0:4000 (http)
20:36:44.185 pid=<0.471.0> module=Phoenix.Endpoint.Supervisor function=log_access_url/2 line=435 [info] Access TpWeb.Endpoint at http://localhost:4000
Interactive Elixir (1.19.0) - press Ctrl+C to exit (type h() ENTER for help)
../priv/static/assets/app.js 5.0mb ⚠️
⚡ Done in 107ms
Rebuilding...
20:36:44.637 pid=<0.1120.0> module=Phoenix.LiveReloader.Channel function=handle_info/2 line=48 [debug] Live reload: priv/static/assets/app.js
Done in 786ms.
20:36:45.755 pid=<0.1139.0> module=Phoenix.Logger function=phoenix_router_dispatch_start/4 line=284 [debug] Processing with TpWeb.PageController.landing_page/2
Parameters: %{}
Pipelines: [:browser, :public_layout]
iex(tp@127.0.0.1)1>
Elixir 1.19 OTP28 stacktrace:
iex --name "tp@127.0.0.1" --cookie dev -S mix run
Erlang/OTP 28 [erts-16.1] [source] [64-bit] [smp:12:12] [ds:12:12:10] [async-threads:1] [jit]
20:40:51.180 pid=<0.274.0> module=Appsignal function=initialize/0 line=89 [info] AppSignal disabled.
20:40:51.631 pid=<0.649.0> module=Bandit function=start_link/1 line=321 [info] Running TpWeb.Endpoint with Bandit 1.8.0 at 0.0.0.0:4000 (http)
20:40:51.636 pid=<0.471.0> module=Phoenix.Endpoint.Supervisor function=log_access_url/2 line=435 [info] Access TpWeb.Endpoint at http://localhost:4000
Interactive Elixir (1.19.0) - press Ctrl+C to exit (type h() ENTER for help)
../priv/static/assets/app.js 5.0mb ⚠️
⚡ Done in 108ms
Rebuilding...
Compiling 757 files (.ex)
20:22:50.233 pid=<0.1019.0> module=gen_server function=error_info/7 line=2785 \[error\] GenServer {Oban.Registry, {Oban, Oban.Met.Examiner}} terminating
\*\* (UndefinedFunctionError) function TpData.Repo.query/3 is undefined (module TpData.Repo is not available)
(tp 1.0.0) TpData.Repo.query(“SELECT pg_notify($1, payload) FROM json_array_elements_text($2::json) AS payload”, \[“public.oban_gossip”, \[“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”\]\], \[prefix: “public”, log: false, oban: true, telemetry_options: \[oban_conf: %Oban.Config{dispatch_cooldown: 5, engine: Oban.Pro.Engines.Smart, get_dynamic_repo: nil, insert_trigger: true, log: false, name: Oban, node: “tp@127.0.0.1”, notifier: {Oban.Notifiers.Postgres, [ ]}, peer: {Oban.Peers.Database, [ ]}, plugins: \[{Oban.Pro.Plugins.DynamicQueues, \[sync_mode: :automatic, queues: \[default: \[limit: 10\], file_to_text: \[limit: 1\], document_search: \[limit: 1\], docx: \[limit: 10\], credit_intel: \[limit: 5\]\]\]}\], prefix: “public”, queues: [ ], repo: TpData.Repo, shutdown_grace_period: 15000, stage_interval: 1000, testing: :disabled}\]\])
(oban 2.20.1) lib/oban/notifiers/postgres.ex:101: Oban.Notifiers.Postgres.notify/3
(oban 2.20.1) lib/oban/notifier.ex:204: anonymous fn/4 in Oban.Notifier.notify/3
(telemetry 1.3.0) /Users/juozas/dev/terrapin-app/deps/telemetry/src/telemetry.erl:324: :telemetry.span/3
(oban_met 1.0.3) lib/oban/met/examiner.ex:115: Oban.Met.Examiner.handle_info/2
(stdlib 7.1) gen_server.erl:2434: :gen_server.try_handle_info/3
(stdlib 7.1) gen_server.erl:2420: :gen_server.handle_msg/3
(stdlib 7.1) proc_lib.erl:333: :proc_lib.init_p_do_apply/3
Last message: :check
0
josevalim
Creator of Elixir
@jozuas What is inside def application in your mix.exs?
@stevegrossi try running your app with elixir --logger-sasl-reports true -S mix phx.server. It will include supervion reports and it should show if apps or supervision trees are shutting down (which would explain the missing repo).
0
jozuas
Sure, here it is:
def application do
[
extra_applications: extra_applications(Mix.env()),
mod: {Tp.Application, []}
]
end
defp extra_applications(:dev) do
[
:logger,
:xmerl,
:wx,
:observer,
:runtime_tools,
:tools
]
end
defp extra_applications(_env) do
[
:logger,
:xmerl,
:runtime_tools
]
end
0
josevalim
Creator of Elixir
Thank you. Can you run the v1.19 branch from Elixir and see if you spot anything different? I have a theory this is related to regexes in .app files, but the issue is certainly not coming from your .app file.
0
vkryukov
Not sure if it’s related, but there was a thread recently with similar symptoms where kino dependency was found to be a culprit: Recompilation triggered after request in dev even without code changes (iex -S mix phx.server) - #9 by axelson
3
jozuas
I cannot reproduce the stacktrace after running my app with this commit in elixir 1.19 branch. However, now each invocation of iex goes through the same analyzing / compiling steps even though there are no code changes:
iex --name "tp@127.0.0.1" --cookie dev -S mix run
Erlang/OTP 28 [erts-16.1] [source] [64-bit] [smp:12:12] [ds:12:12:10] [async-threads:1] [jit]
mix deps.compile running across 6 OS processes
5> ===> Analyzing applications...
4> ===> Analyzing applications...
3> ===> Analyzing applications...
6> ===> Analyzing applications...
1> ===> Analyzing applications...
2> ===> Analyzing applications...
5> ===> Compiling telemetry
4> ===> Compiling mimerl
1> ===> Compiling sleeplocks
3> ===> Compiling erlsom
6> ===> Compiling mochiweb
2> ===> Compiling unicode_util_compat
3> ===> Analyzing applications...
5> ===> Analyzing applications...
4> ===> Analyzing applications...
1> ===> Analyzing applications...
2> ===> Analyzing applications...
4> ===> Compiling certifi
5> ===> Compiling telemetry_poller
3> ===> Compiling metrics
1> ===> Compiling parse_trans
2> ===> Compiling idna
2> ===> Analyzing applications...
2> ===> Compiling hackney
Compiling 758 files (.ex)
Generated tp app
16:36:24.148 pid=<0.5020.0> module=Appsignal function=initialize/0 line=89 [info] AppSignal disabled.
16:36:24.691 pid=<0.5395.0> module=Bandit function=start_link/1 line=321 [info] Running TpWeb.Endpoint with Bandit 1.8.0 at 0.0.0.0:4000 (http)
16:36:24.699 pid=<0.5217.0> module=Phoenix.Endpoint.Supervisor function=log_access_url/2 line=435 [info] Access TpWeb.Endpoint at http://localhost:4000
Interactive Elixir (1.19.1) - press Ctrl+C to exit (type h() ENTER for help)
../priv/static/assets/app.js 5.0mb ⚠️
⚡ Done in 133ms
Rebuilding...
Done in 785ms.
iex(tp@127.0.0.1)1>
0
jozuas
Yep, that was it. I removed kino from my deps, and with Elixir 1.19 OTP 28 I no longer get the stacktrace.
2
nicanor
Same here.
I removed Kino from mix.exs and I’m not experiencing the error anymore.
Thanks for your help!
2