Resync state after partial-state join (#12394)

We work through all the events with partial state, updating the state at each
of them. Once it's done, we recalculate the state for the whole room, and then
mark the room as having complete state.

synapse/storage/databases/main/room.py

@@ -1077,6 +1077,37 @@ class RoomWorkerStore(CacheInvalidationWorkerStore):
             get_rooms_for_retention_period_in_range_txn,
         )
 
+    async def clear_partial_state_room(self, room_id: str) -> bool:
+        # this can race with incoming events, so we watch out for FK errors.
+        # TODO(faster_joins): this still doesn't completely fix the race, since the persist process
+        #   is not atomic. I fear we need an application-level lock.
+        try:
+            await self.db_pool.runInteraction(
+                "clear_partial_state_room", self._clear_partial_state_room_txn, room_id
+            )
+            return True
+        except self.db_pool.engine.module.DatabaseError as e:
+            # TODO(faster_joins): how do we distinguish between FK errors and other errors?
+            logger.warning(
+                "Exception while clearing lazy partial-state-room %s, retrying: %s",
+                room_id,
+                e,
+            )
+            return False
+
+    @staticmethod
+    def _clear_partial_state_room_txn(txn: LoggingTransaction, room_id: str) -> None:
+        DatabasePool.simple_delete_txn(
+            txn,
+            table="partial_state_rooms_servers",
+            keyvalues={"room_id": room_id},
+        )
+        DatabasePool.simple_delete_one_txn(
+            txn,
+            table="partial_state_rooms",
+            keyvalues={"room_id": room_id},
+        )
+
 
 class _BackgroundUpdates:
     REMOVE_TOMESTONED_ROOMS_BG_UPDATE = "remove_tombstoned_rooms_from_directory"