expected :action/2 to return a Plug.Conn error in Phoenix controller

Question

I am trying to get Hyperledger working in a replicated private pool mode. There are no responses in their forum and hence the question here. Moreover the question is purely about a technical error.

Hyperledger uses Phoenix, Cowboy, Elixir, Ecto, Erlang, Postgre SQL, Node.js, etc. The primary node processes the request, inserts a row into couple of tables and broadcasts a message to the secondary node. The secondary node does something similar and sends a message back to the primary node to continue processing. Here is where the problems start. The primary node now tries to read rows it inserted earlier but gets a nil result. Thereafter the following exception is thrown.

** (exit) an exception was raised:
    ** (RuntimeError) expected :action/2 to return a Plug.Conn
        (hyperledger) web/controllers/log_entry_controller.ex:1: Hyperledger.LogEntryController.phoenix_controller_pipeline/2
        (hyperledger) lib/phoenix/router.ex:297: Hyperledger.Router.dispatch/2
        (hyperledger) lib/phoenix/router.ex:2: Hyperledger.Router.call/2
        (hyperledger) lib/hyperledger/endpoint.ex:1: Hyperledger.Endpoint.phoenix_endpoint_pipeline/1
        (hyperledger) lib/plug/debugger.ex:90: Hyperledger.Endpoint."call (overridable 3)"/2
        (hyperledger) lib/phoenix/endpoint/render_errors.ex:34: Hyperledger.Endpoint.call/2
        (phoenix) lib/phoenix/endpoint/cowboy_handler.ex:50: Phoenix.Endpoint.CowboyHandler.upgrade/4
        (cowboy) src/cowboy_protocol.erl:435: :cowboy_protocol.execute/4

Why is the primary node unable to read something that it inserted before the broadcast was sent to the secondary node. I am suspecting something to do with a stale database connection or similar here. Any help would be greatly appreciated. Please let me know if I should share additional information.

Answer 1

You have an action in your controller that is not returning the conn. All actions must receive conn and return the updated conn.