Fix lost events on replication reconnection

changelog.d/7427.feature

@@ -0,0 +1 @@
+Add support for running replication over Redis when using workers.

synapse/replication/tcp/handler.py

@@ -81,9 +81,6 @@ def __init__(self, hs):
         self._instance_id = hs.get_instance_id()
         self._instance_name = hs.get_instance_name()
 
-        # Set of streams that we've caught up with.
-        self._streams_connected = set()  # type: Set[str]
-
         self._streams = {
             stream.NAME: stream(hs) for stream in STREAMS_MAP.values()
         }  # type: Dict[str, Stream]
@@ -99,9 +96,13 @@ def __init__(self, hs):
         # The factory used to create connections.
         self._factory = None  # type: Optional[ReconnectingClientFactory]
 
-        # The currently connected connections.
+        # The currently connected connections. (The list of places we need to send
+        # outgoing replication commands to.)
         self._connections = []  # type: List[AbstractConnection]
 
+        # For each connection, the incoming streams that are coming from that connection
+        self._streams_by_connection = {}  # type: Dict[AbstractConnection, Set[str]]
+
         LaterGauge(
             "synapse_replication_tcp_resource_total_connections",
             "",
@@ -257,9 +258,11 @@ async def on_RDATA(self, conn: AbstractConnection, cmd: RdataCommand):
         #   2. so we don't race with getting a POSITION command and fetching
         #      missing RDATA.
         with await self._position_linearizer.queue(cmd.stream_name):
-            if stream_name not in self._streams_connected:
-                # If the stream isn't marked as connected then we haven't seen a
-                # `POSITION` command yet, and so we may have missed some rows.
+            # make sure that we've processed a POSITION for this stream *on this
+            # connection*. (A POSITION on another connection is no good, as there
+            # is no guarantee that we have seen all the intermediate updates.)
+            sbc = self._streams_by_connection.get(conn)
+            if not sbc or stream_name not in sbc:
                 # Let's drop the row for now, on the assumption we'll receive a
                 # `POSITION` soon and we'll catch up correctly then.
                 logger.debug(
@@ -302,21 +305,25 @@ async def on_POSITION(self, conn: AbstractConnection, cmd: PositionCommand):
             # Ignore POSITION that are just our own echoes
             return
 
-        stream = self._streams.get(cmd.stream_name)
+        logger.info("Handling '%s %s'", cmd.NAME, cmd.to_line())
+
+        stream_name = cmd.stream_name
+        stream = self._streams.get(stream_name)
         if not stream:
-            logger.error("Got POSITION for unknown stream: %s", cmd.stream_name)
+            logger.error("Got POSITION for unknown stream: %s", stream_name)
             return
 
         # We protect catching up with a linearizer in case the replication
         # connection reconnects under us.
-        with await self._position_linearizer.queue(cmd.stream_name):
+        with await self._position_linearizer.queue(stream_name):
             # We're about to go and catch up with the stream, so remove from set
             # of connected streams.
-            self._streams_connected.discard(cmd.stream_name)
+            for streams in self._streams_by_connection.values():
+                streams.discard(stream_name)
 
             # We clear the pending batches for the stream as the fetching of the
             # missing updates below will fetch all rows in the batch.
-            self._pending_batches.pop(cmd.stream_name, [])
+            self._pending_batches.pop(stream_name, [])
 
             # Find where we previously streamed up to.
             current_token = stream.current_token()
@@ -326,6 +333,12 @@ async def on_POSITION(self, conn: AbstractConnection, cmd: PositionCommand):
             # between then and now.
             missing_updates = cmd.token != current_token
             while missing_updates:
+                logger.info(
+                    "Fetching replication rows for '%s' between %i and %i",
+                    stream_name,
+                    current_token,
+                    cmd.token,
+                )
                 (
                     updates,
                     current_token,
@@ -341,16 +354,18 @@ async def on_POSITION(self, conn: AbstractConnection, cmd: PositionCommand):
 
                 for token, rows in _batch_updates(updates):
                     await self.on_rdata(
-                        cmd.stream_name,
+                        stream_name,
                         cmd.instance_name,
                         token,
                         [stream.parse_row(row) for row in rows],
                     )
 
+            logger.info("Caught up with stream '%s' to %i", stream_name, cmd.token)
+
             # We've now caught up to position sent to us, notify handler.
-            await self._replication_data_handler.on_position(cmd.stream_name, cmd.token)
+            await self._replication_data_handler.on_position(stream_name, cmd.token)
 
-            self._streams_connected.add(cmd.stream_name)
+            self._streams_by_connection.setdefault(conn, set()).add(stream_name)
 
     async def on_REMOTE_SERVER_UP(
         self, conn: AbstractConnection, cmd: RemoteServerUpCommand
@@ -408,6 +423,12 @@ def new_connection(self, connection: AbstractConnection):
     def lost_connection(self, connection: AbstractConnection):
         """Called when a connection is closed/lost.
         """
+        # we no longer need _streams_by_connection for this connection.
+        streams = self._streams_by_connection.pop(connection, None)
+        if streams:
+            logger.info(
+                "Lost replication connection; streams now disconnected: %s", streams
+            )
         try:
             self._connections.remove(connection)
         except ValueError:

synapse/replication/tcp/redis.py

@@ -18,7 +18,7 @@
 
 import txredisapi
 
-from synapse.logging.context import PreserveLoggingContext
+from synapse.logging.context import make_deferred_yieldable
 from synapse.metrics.background_process_metrics import run_as_background_process
 from synapse.replication.tcp.commands import (
     Command,
@@ -41,17 +41,23 @@
 class RedisSubscriber(txredisapi.SubscriberProtocol, AbstractConnection):
     """Connection to redis subscribed to replication stream.
 
-    Parses incoming messages from redis into replication commands, and passes
-    them to `ReplicationCommandHandler`
+    This class fulfils two functions:
+
+    (a) it implements the twisted Protocol API, where it handles the SUBSCRIBEd redis
+    connection, parsing *incoming* messages into replication commands, and passing them
+    to `ReplicationCommandHandler`
+
+    (b) it implements the AbstractConnection API, where it sends *outgoing* commands
+    onto outbound_redis_connection.
 
     Due to the vagaries of `txredisapi` we don't want to have a custom
     constructor, so instead we expect the defined attributes below to be set
     immediately after initialisation.
 
     Attributes:
         handler: The command handler to handle incoming commands.
-        stream_name: The *redis* stream name to subscribe to (not anything to
-            do with Synapse replication streams).
+        stream_name: The *redis* stream name to subscribe to and publish from
+            (not anything to do with Synapse replication streams).
         outbound_redis_connection: The connection to redis to use to send
             commands.
     """
@@ -61,13 +67,23 @@ class RedisSubscriber(txredisapi.SubscriberProtocol, AbstractConnection):
     outbound_redis_connection = None  # type: txredisapi.RedisProtocol
 
     def connectionMade(self):
+        logger.info("Connected to redis")
         super().connectionMade()
-        logger.info("Connected to redis instance")
-        self.subscribe(self.stream_name)
-        self.send_command(ReplicateCommand())
-
+        run_as_background_process("subscribe-replication", self._send_subscribe)
         self.handler.new_connection(self)
 
+    async def _send_subscribe(self):
+        # it's important to make sure that we only send the REPLICATE command once we
+        # have successfully subscribed to the stream - otherwise we might miss the
+        # POSITION response sent back by the other end.
+        logger.info("Sending redis SUBSCRIBE for %s", self.stream_name)
+        await make_deferred_yieldable(self.subscribe(self.stream_name))
+        logger.info(
+            "Successfully subscribed to redis stream, sending REPLICATE command"
+        )
+        await self._async_send_command(ReplicateCommand())
+        logger.info("REPLICATE successfully sent")
+
     def messageReceived(self, pattern: str, channel: str, message: str):
         """Received a message from redis.
         """
@@ -120,8 +136,8 @@ async def handle_command(self, cmd: Command):
             logger.warning("Unhandled command: %r", cmd)
 
     def connectionLost(self, reason):
+        logger.info("Lost connection to redis")
         super().connectionLost(reason)
-        logger.info("Lost connection to redis instance")
         self.handler.lost_connection(self)
 
     def send_command(self, cmd: Command):
@@ -130,6 +146,10 @@ def send_command(self, cmd: Command):
         Args:
             cmd (Command)
         """
+        run_as_background_process("send-cmd", self._async_send_command, cmd)
+
+    async def _async_send_command(self, cmd: Command):
+        """Encode a replication command and send it over our outbound connection"""
         string = "%s %s" % (cmd.NAME, cmd.to_line())
         if "\n" in string:
             raise Exception("Unexpected newline in command: %r", string)
@@ -140,15 +160,9 @@ def send_command(self, cmd: Command):
         # remote instances.
         tcp_outbound_commands_counter.labels(cmd.NAME, "redis").inc()
 
-        async def _send():
-            with PreserveLoggingContext():
-                # Note that we use the other connection as we can't send
-                # commands using the subscription connection.
-                await self.outbound_redis_connection.publish(
-                    self.stream_name, encoded_string
-                )
-
-        run_as_background_process("send-cmd", _send)
+        await make_deferred_yieldable(
+            self.outbound_redis_connection.publish(self.stream_name, encoded_string)
+        )
 
 
 class RedisDirectTcpReplicationClientFactory(txredisapi.SubscriberFactory):