feat: epmdless deployments

apps/expert/lib/expert/engine_node.ex

@@ -29,17 +29,21 @@ defmodule Expert.EngineNode do
 
     def start(%__MODULE__{} = state, paths, from) do
       this_node = inspect(Node.self())
+      dist_port = Expert.EPMD.dist_port()
 
-      args = [
-        "--name",
-        Project.node_name(state.project),
-        "--cookie",
-        state.cookie,
-        "--no-halt",
-        "-e",
-        "Node.connect(#{this_node})"
-        | path_append_arguments(paths)
-      ]
+      args =
+        [
+          "--erl",
+          "-start_epmd false -erl_epmd_port #{dist_port} -dist_listen false",
+          "--name",
+          Project.node_name(state.project),
+          "--cookie",
+          state.cookie,
+          "--no-halt",
+          "-e",
+          "Node.connect(#{this_node}); IO.puts(\"ok\")"
+          | path_append_arguments(paths)
+        ]
 
       port = Expert.Port.open_elixir(state.project, args: args)
 
@@ -110,7 +114,6 @@ defmodule Expert.EngineNode do
   use GenServer
 
   def start(project) do
-    :ok = ensure_epmd_started()
     start_net_kernel(project)
 
     node_name = Project.node_name(project)
@@ -134,16 +137,6 @@ defmodule Expert.EngineNode do
     :rpc.call(node, Engine, :ensure_apps_started, [])
   end
 
-  defp ensure_epmd_started do
-    case System.cmd("epmd", ~w(-daemon)) do
-      {"", 0} ->
-        :ok
-
-      _ ->
-        {:error, :epmd_failed}
-    end
-  end
-
   if Mix.env() == :test do
     # In test environment, Expert depends on the Engine app, so we look for it
     # in the expert build path.
@@ -278,7 +271,7 @@ defmodule Expert.EngineNode do
 
   @impl true
   def handle_call({:start, paths}, from, %State{} = state) do
-    :ok = :net_kernel.monitor_nodes(true, node_type: :visible)
+    :ok = :net_kernel.monitor_nodes(true, node_type: :all)
     Process.send_after(self(), :maybe_start_timeout, @start_timeout)
     state = State.start(state, paths, from)
     {:noreply, state}

apps/expert/lib/expert/epmd.ex

@@ -0,0 +1,26 @@
+defmodule Expert.EPMD do
+  def dist_port do
+    :persistent_term.get(:expert_dist_port, nil)
+  end
+
+  # EPMD callbacks
+
+  def register_node(name, port), do: register_node(name, port, :inet)
+
+  def register_node(name, port, family) do
+    :persistent_term.put(:expert_dist_port, port)
+
+    # We don't care if EPMD is not running
+    case :erl_epmd.register_node(name, port, family) do
+      {:error, _} -> {:ok, -1}
+      {:ok, _} = ok -> ok
+    end
+  end
+
+  defdelegate start_link(), to: :erl_epmd
+  defdelegate port_please(name, host), to: :erl_epmd
+  defdelegate port_please(name, host, timeout), to: :erl_epmd
+  defdelegate listen_port_please(name, host), to: :erl_epmd
+  defdelegate address_please(name, host, family), to: :erl_epmd
+  defdelegate names(host_name), to: :erl_epmd
+end

apps/expert/rel/remote.vm.args.eex

@@ -0,0 +1,12 @@
+## Customize flags given to the VM: https://www.erlang.org/doc/man/erl.html
+## -mode/-name/-sname/-setcookie are configured via env vars, do not set them here
+
+## Increase number of concurrent ports/sockets
+##+Q 65536
+
+## Tweak GC to run more often
+##-env ERL_FULLSWEEP_AFTER 10
+
+## Enable deployment without epmd
+## (requires changing both vm.args and remote.vm.args)
+-start_epmd false -dist_listen false

apps/expert/rel/vm.args.eex

@@ -0,0 +1,12 @@
+## Customize flags given to the VM: https://www.erlang.org/doc/man/erl.html
+## -mode/-name/-sname/-setcookie are configured via env vars, do not set them here
+
+## Increase number of concurrent ports/sockets
+##+Q 65536
+
+## Tweak GC to run more often
+##-env ERL_FULLSWEEP_AFTER 10
+
+## Enable deployment without epmd
+## (requires changing both vm.args and remote.vm.args)
+-start_epmd false -epmd_module Elixir.XPExpert.EPMD