Fix not sending events over federation when using sharded event persisters (#8536)

* Fix outbound federaion with multiple event persisters.

We incorrectly notified federation senders that the minimum persisted
stream position had advanced when we got an `RDATA` from an event
persister.

Notifying of federation senders already correctly happens in the
notifier, so we just delete the offending line.

* Change some interfaces to use RoomStreamToken.

By enforcing use of `RoomStreamTokens` we make it less likely that
people pass in random ints that they got from somewhere random.

1 files changed, 7 insertions, 2 deletions

diff --git a/synapse/federation/sender/__init__.py b/synapse/federation/sender/__init__.py
index e33b29a42c..604cfd1935 100644
--- a/synapse/federation/sender/__init__.py
+++ b/synapse/federation/sender/__init__.py
@@ -40,7 +40,7 @@ from synapse.metrics import (
     events_processed_counter,
 )
 from synapse.metrics.background_process_metrics import run_as_background_process
-from synapse.types import ReadReceipt
+from synapse.types import ReadReceipt, RoomStreamToken
 from synapse.util.metrics import Measure, measure_func
 
 logger = logging.getLogger(__name__)
@@ -154,10 +154,15 @@ class FederationSender:
             self._per_destination_queues[destination] = queue
         return queue
 
-    def notify_new_events(self, current_id: int) -> None:
+    def notify_new_events(self, max_token: RoomStreamToken) -> None:
         """This gets called when we have some new events we might want to
         send out to other servers.
         """
+        # We just use the minimum stream ordering and ignore the vector clock
+        # component. This is safe to do as long as we *always* ignore the vector
+        # clock components.
+        current_id = max_token.stream
+
         self._last_poked_id = max(current_id, self._last_poked_id)
 
         if self._is_processing: