Rely on selector to detect completed connection attempts

Author: dpkp

kafka/client_async.py

@@ -267,9 +267,9 @@ def _conn_state_change(self, node_id, sock, conn):
                 if node_id not in self._connecting:
                     self._connecting.add(node_id)
                 try:
-                    self._selector.register(sock, selectors.EVENT_WRITE)
+                    self._selector.register(sock, selectors.EVENT_WRITE, conn)
                 except KeyError:
-                    self._selector.modify(sock, selectors.EVENT_WRITE)
+                    self._selector.modify(sock, selectors.EVENT_WRITE, conn)
 
                 if self.cluster.is_bootstrap(node_id):
                     self._last_bootstrap = time.time()
@@ -624,6 +624,9 @@ def _poll(self, timeout):
                 self._clear_wake_fd()
                 continue
             elif not (events & selectors.EVENT_READ):
+                conn = key.data
+                if conn.node_id in self._connecting:
+                    self._maybe_connect(conn.node_id)
                 continue
             conn = key.data
             processed.add(conn)

kafka/conn.py

@@ -748,16 +748,16 @@ def connection_delay(self):
         """
         Return the number of milliseconds to wait, based on the connection
         state, before attempting to send data. When disconnected, this respects
-        the reconnect backoff time. When connecting, returns 0 to allow
-        non-blocking connect to finish. When connected, returns a very large
-        number to handle slow/stalled connections.
+        the reconnect backoff time. When connecting or connected, returns a very
+        large number to handle slow/stalled connections.
         """
         time_waited = time.time() - (self.last_attempt or 0)
         if self.state is ConnectionStates.DISCONNECTED:
             return max(self._reconnect_backoff - time_waited, 0) * 1000
-        elif self.connecting():
-            return 0
         else:
+            # When connecting or connected, we should be able to delay
+            # indefinitely since other events (connection or data acked) will
+            # cause a wakeup once data can be sent.
             return float('inf')
 
     def connected(self):

kafka/producer/sender.py

@@ -157,7 +157,7 @@ def run_once(self):
         # difference between now and its linger expiry time; otherwise the
         # select time will be the time difference between now and the
         # metadata expiry time
-        self._client.poll(poll_timeout_ms)
+        self._client.poll(timeout_ms=poll_timeout_ms)
 
     def initiate_close(self):
         """Start closing the sender (won't complete until all data is sent)."""

test/test_client_async.py

@@ -94,7 +94,7 @@ def test_conn_state_change(mocker, cli, conn):
     sock = conn._sock
     cli._conn_state_change(node_id, sock, conn)
     assert node_id in cli._connecting
-    sel.register.assert_called_with(sock, selectors.EVENT_WRITE)
+    sel.register.assert_called_with(sock, selectors.EVENT_WRITE, conn)
 
     conn.state = ConnectionStates.CONNECTED
     cli._conn_state_change(node_id, sock, conn)

test/test_conn.py

@@ -85,7 +85,7 @@ def test_connection_delay(conn):
         conn.last_attempt = 1000
         assert conn.connection_delay() == conn.config['reconnect_backoff_ms']
         conn.state = ConnectionStates.CONNECTING
-        assert conn.connection_delay() == 0
+        assert conn.connection_delay() == float('inf')
         conn.state = ConnectionStates.CONNECTED
         assert conn.connection_delay() == float('inf')