Fix bug where a new writer advances their token too quickly (#16473)

* Fix bug where a new writer advances their token too quickly

When starting a new writer (for e.g. persisting events), the
`MultiWriterIdGenerator` doesn't have a minimum token for it as there
are no rows matching that new writer in the DB.

This results in the the first stream ID it acquired being announced as
persisted *before* it actually finishes persisting, if another writer
gets and persists a subsequent stream ID. This is due to the logic of
setting the minimum persisted position to the minimum known position of
across all writers, and the new writer starts off not being considered.

* Fix sending out POSITIONs when our token advances without update

Broke in #14820

* For replication HTTP requests, only wait for minimal position

changelog.d/16473.bugfix

@@ -0,0 +1 @@
+Fix a long-standing, exceedingly rare edge case where the first event persisted by a new event persister worker might not be sent down `/sync`.

docs/development/synapse_architecture/streams.md

@@ -51,17 +51,24 @@ will be inserted with that ID.
 
 For any given stream reader (including writers themselves), we may define a per-writer current stream ID:
 
-> The current stream ID _for a writer W_ is the largest stream ID such that
+> A current stream ID _for a writer W_ is the largest stream ID such that
 > all transactions added by W with equal or smaller ID have completed.
 
 Similarly, there is a "linear" notion of current stream ID:
 
-> The "linear" current stream ID is the largest stream ID such that
+> A "linear" current stream ID is the largest stream ID such that
 > all facts (added by any writer) with equal or smaller ID have completed.
 
 Because different stream readers A and B learn about new facts at different times, A and B may disagree about current stream IDs.
 Put differently: we should think of stream readers as being independent of each other, proceeding through a stream of facts at different rates.
 
+The above definition does not give a unique current stream ID, in fact there can
+be a range of current stream IDs. Synapse uses both the minimum and maximum IDs
+for different purposes. Most often the maximum is used, as its generally
+beneficial for workers to advance their IDs as soon as possible. However, the
+minimum is used in situations where e.g. another worker is going to wait until
+the stream advances past a position.
+
 **NB.** For both senses of "current", that if a writer opens a transaction that never completes, the current stream ID will never advance beyond that writer's last written stream ID.
 
 For single-writer streams, the per-writer current ID and the linear current ID are the same.

synapse/replication/http/_base.py

@@ -238,7 +238,7 @@ class ReplicationEndpoint(metaclass=abc.ABCMeta):
 
                     data[_STREAM_POSITION_KEY] = {
                         "streams": {
-                            stream.NAME: stream.current_token(local_instance_name)
+                            stream.NAME: stream.minimal_local_current_token()
                             for stream in streams
                         },
                         "instance_name": local_instance_name,

synapse/replication/tcp/streams/_base.py

@@ -33,6 +33,7 @@ from synapse.replication.http.streams import ReplicationGetStreamUpdates
 
 if TYPE_CHECKING:
     from synapse.server import HomeServer
+    from synapse.storage.util.id_generators import AbstractStreamIdGenerator
 
 logger = logging.getLogger(__name__)
 
@@ -107,22 +108,10 @@ class Stream:
     def __init__(
         self,
         local_instance_name: str,
-        current_token_function: Callable[[str], Token],
         update_function: UpdateFunction,
     ):
         """Instantiate a Stream
 
-        `current_token_function` and `update_function` are callbacks which
-        should be implemented by subclasses.
-
-        `current_token_function` takes an instance name, which is a writer to
-        the stream, and returns the position in the stream of the writer (as
-        viewed from the current process). On the writer process this is where
-        the writer has successfully written up to, whereas on other processes
-        this is the position which we have received updates up to over
-        replication. (Note that most streams have a single writer and so their
-        implementations ignore the instance name passed in).
-
         `update_function` is called to get updates for this stream between a
         pair of stream tokens. See the `UpdateFunction` type definition for more
         info.
@@ -133,12 +122,28 @@ class Stream:
             update_function: callback go get stream updates, as above
         """
         self.local_instance_name = local_instance_name
-        self.current_token = current_token_function
         self.update_function = update_function
 
         # The token from which we last asked for updates
         self.last_token = self.current_token(self.local_instance_name)
 
+    def current_token(self, instance_name: str) -> Token:
+        """This takes an instance name, which is a writer to
+        the stream, and returns the position in the stream of the writer (as
+        viewed from the current process).
+        """
+        # We can't make this an abstract class as it makes mypy unhappy.
+        raise NotImplementedError()
+
+    def minimal_local_current_token(self) -> Token:
+        """Tries to return a minimal current token for the local instance,
+        i.e. for writers this would be the last successful write.
+
+        If local instance is not a writer (or has written yet) then falls back
+        to returning the normal "current token".
+        """
+        raise NotImplementedError()
+
     def discard_updates_and_advance(self) -> None:
         """Called when the stream should advance but the updates would be discarded,
         e.g. when there are no currently connected workers.
@@ -190,6 +195,25 @@ class Stream:
         return updates, upto_token, limited
 
 
+class _StreamFromIdGen(Stream):
+    """Helper class for simple streams that use a stream ID generator"""
+
+    def __init__(
+        self,
+        local_instance_name: str,
+        update_function: UpdateFunction,
+        stream_id_gen: "AbstractStreamIdGenerator",
+    ):
+        self._stream_id_gen = stream_id_gen
+        super().__init__(local_instance_name, update_function)
+
+    def current_token(self, instance_name: str) -> Token:
+        return self._stream_id_gen.get_current_token_for_writer(instance_name)
+
+    def minimal_local_current_token(self) -> Token:
+        return self._stream_id_gen.get_minimal_local_current_token()
+
+
 def current_token_without_instance(
     current_token: Callable[[], int]
 ) -> Callable[[str], int]:
@@ -242,17 +266,21 @@ class BackfillStream(Stream):
         self.store = hs.get_datastores().main
         super().__init__(
             hs.get_instance_name(),
-            self._current_token,
             self.store.get_all_new_backfill_event_rows,
         )
 
-    def _current_token(self, instance_name: str) -> int:
+    def current_token(self, instance_name: str) -> Token:
         # The backfill stream over replication operates on *positive* numbers,
         # which means we need to negate it.
         return -self.store._backfill_id_gen.get_current_token_for_writer(instance_name)
 
+    def minimal_local_current_token(self) -> Token:
+        # The backfill stream over replication operates on *positive* numbers,
+        # which means we need to negate it.
+        return -self.store._backfill_id_gen.get_minimal_local_current_token()
 
-class PresenceStream(Stream):
+
+class PresenceStream(_StreamFromIdGen):
     @attr.s(slots=True, frozen=True, auto_attribs=True)
     class PresenceStreamRow:
         user_id: str
@@ -283,9 +311,7 @@ class PresenceStream(Stream):
             update_function = make_http_update_function(hs, self.NAME)
 
         super().__init__(
-            hs.get_instance_name(),
-            current_token_without_instance(store.get_current_presence_token),
-            update_function,
+            hs.get_instance_name(), update_function, store._presence_id_gen
         )
 
 
@@ -305,13 +331,18 @@ class PresenceFederationStream(Stream):
     ROW_TYPE = PresenceFederationStreamRow
 
     def __init__(self, hs: "HomeServer"):
-        federation_queue = hs.get_presence_handler().get_federation_queue()
+        self._federation_queue = hs.get_presence_handler().get_federation_queue()
         super().__init__(
             hs.get_instance_name(),
-            federation_queue.get_current_token,
-            federation_queue.get_replication_rows,
+            self._federation_queue.get_replication_rows,
         )
 
+    def current_token(self, instance_name: str) -> Token:
+        return self._federation_queue.get_current_token(instance_name)
+
+    def minimal_local_current_token(self) -> Token:
+        return self._federation_queue.get_current_token(self.local_instance_name)
+
 
 class TypingStream(Stream):
     @attr.s(slots=True, frozen=True, auto_attribs=True)
@@ -341,20 +372,25 @@ class TypingStream(Stream):
             update_function: Callable[
                 [str, int, int, int], Awaitable[Tuple[List[Tuple[int, Any]], int, bool]]
             ] = typing_writer_handler.get_all_typing_updates
-            current_token_function = typing_writer_handler.get_current_token
+            self.current_token_function = typing_writer_handler.get_current_token
         else:
             # Query the typing writer process
             update_function = make_http_update_function(hs, self.NAME)
-            current_token_function = hs.get_typing_handler().get_current_token
+            self.current_token_function = hs.get_typing_handler().get_current_token
 
         super().__init__(
             hs.get_instance_name(),
-            current_token_without_instance(current_token_function),
             update_function,
         )
 
+    def current_token(self, instance_name: str) -> Token:
+        return self.current_token_function()
 
-class ReceiptsStream(Stream):
+    def minimal_local_current_token(self) -> Token:
+        return self.current_token_function()
+
+
+class ReceiptsStream(_StreamFromIdGen):
     @attr.s(slots=True, frozen=True, auto_attribs=True)
     class ReceiptsStreamRow:
         room_id: str
@@ -371,12 +407,12 @@ class ReceiptsStream(Stream):
         store = hs.get_datastores().main
         super().__init__(
             hs.get_instance_name(),
-            current_token_without_instance(store.get_max_receipt_stream_id),
             store.get_all_updated_receipts,
+            store._receipts_id_gen,
         )
 
 
-class PushRulesStream(Stream):
+class PushRulesStream(_StreamFromIdGen):
     """A user has changed their push rules"""
 
     @attr.s(slots=True, frozen=True, auto_attribs=True)
@@ -387,20 +423,16 @@ class PushRulesStream(Stream):
     ROW_TYPE = PushRulesStreamRow
 
     def __init__(self, hs: "HomeServer"):
-        self.store = hs.get_datastores().main
+        store = hs.get_datastores().main
 
         super().__init__(
             hs.get_instance_name(),
-            self._current_token,
-            self.store.get_all_push_rule_updates,
+            store.get_all_push_rule_updates,
+            store._push_rules_stream_id_gen,
         )
 
-    def _current_token(self, instance_name: str) -> int:
-        push_rules_token = self.store.get_max_push_rules_stream_id()
-        return push_rules_token
 
-
-class PushersStream(Stream):
+class PushersStream(_StreamFromIdGen):
     """A user has added/changed/removed a pusher"""
 
     @attr.s(slots=True, frozen=True, auto_attribs=True)
@@ -418,8 +450,8 @@ class PushersStream(Stream):
 
         super().__init__(
             hs.get_instance_name(),
-            current_token_without_instance(store.get_pushers_stream_token),
             store.get_all_updated_pushers_rows,
+            store._pushers_id_gen,
         )
 
 
@@ -447,15 +479,20 @@ class CachesStream(Stream):
     ROW_TYPE = CachesStreamRow
 
     def __init__(self, hs: "HomeServer"):
-        store = hs.get_datastores().main
+        self.store = hs.get_datastores().main
         super().__init__(
             hs.get_instance_name(),
-            store.get_cache_stream_token_for_writer,
-            store.get_all_updated_caches,
+            self.store.get_all_updated_caches,
         )
 
+    def current_token(self, instance_name: str) -> Token:
+        return self.store.get_cache_stream_token_for_writer(instance_name)
 
-class DeviceListsStream(Stream):
+    def minimal_local_current_token(self) -> Token:
+        return self.current_token(self.local_instance_name)
+
+
+class DeviceListsStream(_StreamFromIdGen):
     """Either a user has updated their devices or a remote server needs to be
     told about a device update.
     """
@@ -473,8 +510,8 @@ class DeviceListsStream(Stream):
         self.store = hs.get_datastores().main
         super().__init__(
             hs.get_instance_name(),
-            current_token_without_instance(self.store.get_device_stream_token),
             self._update_function,
+            self.store._device_list_id_gen,
         )
 
     async def _update_function(
@@ -525,7 +562,7 @@ class DeviceListsStream(Stream):
         return updates, upper_limit_token, devices_limited or signatures_limited
 
 
-class ToDeviceStream(Stream):
+class ToDeviceStream(_StreamFromIdGen):
     """New to_device messages for a client"""
 
     @attr.s(slots=True, frozen=True, auto_attribs=True)
@@ -539,12 +576,12 @@ class ToDeviceStream(Stream):
         store = hs.get_datastores().main
         super().__init__(
             hs.get_instance_name(),
-            current_token_without_instance(store.get_to_device_stream_token),
             store.get_all_new_device_messages,
+            store._device_inbox_id_gen,
         )
 
 
-class AccountDataStream(Stream):
+class AccountDataStream(_StreamFromIdGen):
     """Global or per room account data was changed"""
 
     @attr.s(slots=True, frozen=True, auto_attribs=True)
@@ -560,8 +597,8 @@ class AccountDataStream(Stream):
         self.store = hs.get_datastores().main
         super().__init__(
             hs.get_instance_name(),
-            current_token_without_instance(self.store.get_max_account_data_stream_id),
             self._update_function,
+            self.store._account_data_id_gen,
         )
 
     async def _update_function(

synapse/replication/tcp/streams/events.py

@@ -19,10 +19,10 @@ from typing import TYPE_CHECKING, Iterable, Optional, Tuple, Type, TypeVar, cast
 import attr
 
 from synapse.replication.tcp.streams._base import (
-    Stream,
     StreamRow,
     StreamUpdateResult,
     Token,
+    _StreamFromIdGen,
 )
 
 if TYPE_CHECKING:
@@ -139,7 +139,7 @@ _EventRows: Tuple[Type[BaseEventsStreamRow], ...] = (
 TypeToRow = {Row.TypeId: Row for Row in _EventRows}
 
 
-class EventsStream(Stream):
+class EventsStream(_StreamFromIdGen):
     """We received a new event, or an event went from being an outlier to not"""
 
     NAME = "events"
@@ -147,9 +147,7 @@ class EventsStream(Stream):
     def __init__(self, hs: "HomeServer"):
         self._store = hs.get_datastores().main
         super().__init__(
-            hs.get_instance_name(),
-            self._store._stream_id_gen.get_current_token_for_writer,
-            self._update_function,
+            hs.get_instance_name(), self._update_function, self._store._stream_id_gen
         )
 
     async def _update_function(

synapse/replication/tcp/streams/federation.py

@@ -18,6 +18,7 @@ import attr
 
 from synapse.replication.tcp.streams._base import (
     Stream,
+    Token,
     current_token_without_instance,
     make_http_update_function,
 )
@@ -47,7 +48,7 @@ class FederationStream(Stream):
             # will be a real FederationSender, which has stubs for current_token and
             # get_replication_rows.)
             federation_sender = hs.get_federation_sender()
-            current_token = current_token_without_instance(
+            self.current_token_func = current_token_without_instance(
                 federation_sender.get_current_token
             )
             update_function: Callable[
@@ -57,15 +58,21 @@ class FederationStream(Stream):
         elif hs.should_send_federation():
             # federation sender: Query master process
             update_function = make_http_update_function(hs, self.NAME)
-            current_token = self._stub_current_token
+            self.current_token_func = self._stub_current_token
 
         else:
             # other worker: stub out the update function (we're not interested in
             # any updates so when we get a POSITION we do nothing)
             update_function = self._stub_update_function
-            current_token = self._stub_current_token
+            self.current_token_func = self._stub_current_token
 
-        super().__init__(hs.get_instance_name(), current_token, update_function)
+        super().__init__(hs.get_instance_name(), update_function)
+
+    def current_token(self, instance_name: str) -> Token:
+        return self.current_token_func(instance_name)
+
+    def minimal_local_current_token(self) -> Token:
+        return self.current_token(self.local_instance_name)
 
     @staticmethod
     def _stub_current_token(instance_name: str) -> int:

synapse/replication/tcp/streams/partial_state.py

@@ -15,7 +15,7 @@ from typing import TYPE_CHECKING
 
 import attr
 
-from synapse.replication.tcp.streams import Stream
+from synapse.replication.tcp.streams._base import _StreamFromIdGen
 
 if TYPE_CHECKING:
     from synapse.server import HomeServer
@@ -27,7 +27,7 @@ class UnPartialStatedRoomStreamRow:
     room_id: str
 
 
-class UnPartialStatedRoomStream(Stream):
+class UnPartialStatedRoomStream(_StreamFromIdGen):
     """
     Stream to notify about rooms becoming un-partial-stated;
     that is, when the background sync finishes such that we now have full state for
@@ -41,8 +41,8 @@ class UnPartialStatedRoomStream(Stream):
         store = hs.get_datastores().main
         super().__init__(
             hs.get_instance_name(),
-            store.get_un_partial_stated_rooms_token,
             store.get_un_partial_stated_rooms_from_stream,
+            store._un_partial_stated_rooms_stream_id_gen,
         )
 
 
@@ -56,7 +56,7 @@ class UnPartialStatedEventStreamRow:
     rejection_status_changed: bool
 
 
-class UnPartialStatedEventStream(Stream):
+class UnPartialStatedEventStream(_StreamFromIdGen):
     """
     Stream to notify about events becoming un-partial-stated.
     """
@@ -68,6 +68,6 @@ class UnPartialStatedEventStream(Stream):
         store = hs.get_datastores().main
         super().__init__(
             hs.get_instance_name(),
-            store.get_un_partial_stated_events_token,
             store.get_un_partial_stated_events_from_stream,
+            store._un_partial_stated_events_stream_id_gen,
         )

synapse/storage/util/id_generators.py

@@ -133,6 +133,15 @@ class AbstractStreamIdGenerator(metaclass=abc.ABCMeta):
         """
         raise NotImplementedError()
 
+    @abc.abstractmethod
+    def get_minimal_local_current_token(self) -> int:
+        """Tries to return a minimal current token for the local instance,
+        i.e. for writers this would be the last successful write.
+
+        If local instance is not a writer (or has written yet) then falls back
+        to returning the normal "current token".
+        """
+
     @abc.abstractmethod
     def get_next(self) -> AsyncContextManager[int]:
         """
@@ -312,6 +321,9 @@ class StreamIdGenerator(AbstractStreamIdGenerator):
     def get_current_token_for_writer(self, instance_name: str) -> int:
         return self.get_current_token()
 
+    def get_minimal_local_current_token(self) -> int:
+        return self.get_current_token()
+
 
 class MultiWriterIdGenerator(AbstractStreamIdGenerator):
     """Generates and tracks stream IDs for a stream with multiple writers.
@@ -408,6 +420,11 @@ class MultiWriterIdGenerator(AbstractStreamIdGenerator):
         # The maximum stream ID that we have seen been allocated across any writer.
         self._max_seen_allocated_stream_id = 1
 
+        # The maximum position of the local instance. This can be higher than
+        # the corresponding position in `current_positions` table when there are
+        # no active writes in progress.
+        self._max_position_of_local_instance = self._max_seen_allocated_stream_id
+
         self._sequence_gen = PostgresSequenceGenerator(sequence_name)
 
         # We check that the table and sequence haven't diverged.
@@ -427,6 +444,16 @@ class MultiWriterIdGenerator(AbstractStreamIdGenerator):
             self._current_positions.values(), default=1
         )
 
+        # For the case where `stream_positions` is not up to date,
+        # `_persisted_upto_position` may be higher.
+        self._max_seen_allocated_stream_id = max(
+            self._max_seen_allocated_stream_id, self._persisted_upto_position
+        )
+
+        # Bump our local maximum position now that we've loaded things from the
+        # DB.
+        self._max_position_of_local_instance = self._max_seen_allocated_stream_id
+
         if not writers:
             # If there have been no explicit writers given then any instance can
             # write to the stream. In which case, let's pre-seed our own
@@ -545,6 +572,14 @@ class MultiWriterIdGenerator(AbstractStreamIdGenerator):
                     if instance == self._instance_name:
                         self._current_positions[instance] = stream_id
 
+        if self._writers:
+            # If we have explicit writers then make sure that each instance has
+            # a position.
+            for writer in self._writers:
+                self._current_positions.setdefault(
+                    writer, self._persisted_upto_position
+                )
+
         cur.close()
 
     def _load_next_id_txn(self, txn: Cursor) -> int:
@@ -688,6 +723,9 @@ class MultiWriterIdGenerator(AbstractStreamIdGenerator):
             if new_cur:
                 curr = self._current_positions.get(self._instance_name, 0)
                 self._current_positions[self._instance_name] = max(curr, new_cur)
+                self._max_position_of_local_instance = max(
+                    curr, new_cur, self._max_position_of_local_instance
+                )
 
             self._add_persisted_position(next_id)
 
@@ -702,10 +740,26 @@ class MultiWriterIdGenerator(AbstractStreamIdGenerator):
         # persisted up to position. This stops Synapse from doing a full table
         # scan when a new writer announces itself over replication.
         with self._lock:
-            return self._return_factor * self._current_positions.get(
+            if self._instance_name == instance_name:
+                return self._return_factor * self._max_position_of_local_instance
+
+            pos = self._current_positions.get(
                 instance_name, self._persisted_upto_position
             )
 
+            # We want to return the maximum "current token" that we can for a
+            # writer, this helps ensure that streams progress as fast as
+            # possible.
+            pos = max(pos, self._persisted_upto_position)
+
+            return self._return_factor * pos
+
+    def get_minimal_local_current_token(self) -> int:
+        with self._lock:
+            return self._return_factor * self._current_positions.get(
+                self._instance_name, self._persisted_upto_position
+            )
+
     def get_positions(self) -> Dict[str, int]:
         """Get a copy of the current positon map.
 
@@ -774,6 +828,18 @@ class MultiWriterIdGenerator(AbstractStreamIdGenerator):
 
         self._persisted_upto_position = max(min_curr, self._persisted_upto_position)
 
+        # Advance our local max position.
+        self._max_position_of_local_instance = max(
+            self._max_position_of_local_instance, self._persisted_upto_position
+        )
+
+        if not self._unfinished_ids and not self._in_flight_fetches:
+            # If we don't have anything in flight, it's safe to advance to the
+            # max seen stream ID.
+            self._max_position_of_local_instance = max(
+                self._max_seen_allocated_stream_id, self._max_position_of_local_instance
+            )
+
         # We now iterate through the seen positions, discarding those that are
         # less than the current min positions, and incrementing the min position
         # if its exactly one greater.

tests/storage/test_id_generators.py

@@ -259,8 +259,9 @@ class MultiWriterIdGeneratorTestCase(HomeserverTestCase):
 
         id_gen = self._create_id_generator()
 
-        # The table is empty so we expect an empty map for positions
-        self.assertEqual(id_gen.get_positions(), {})
+        # The table is empty so we expect the map for positions to have a dummy
+        # minimum value.
+        self.assertEqual(id_gen.get_positions(), {"master": 1})
 
     def test_single_instance(self) -> None:
         """Test that reads and writes from a single process are handled
@@ -349,15 +350,12 @@ class MultiWriterIdGeneratorTestCase(HomeserverTestCase):
         first_id_gen = self._create_id_generator("first", writers=["first", "second"])
         second_id_gen = self._create_id_generator("second", writers=["first", "second"])
 
-        # The first ID gen will notice that it can advance its token to 7 as it
-        # has no in progress writes...
         self.assertEqual(first_id_gen.get_positions(), {"first": 3, "second": 7})
-        self.assertEqual(first_id_gen.get_current_token_for_writer("first"), 3)
+        self.assertEqual(first_id_gen.get_current_token_for_writer("first"), 7)
         self.assertEqual(first_id_gen.get_current_token_for_writer("second"), 7)
 
-        # ... but the second ID gen doesn't know that.
         self.assertEqual(second_id_gen.get_positions(), {"first": 3, "second": 7})
-        self.assertEqual(second_id_gen.get_current_token_for_writer("first"), 3)
+        self.assertEqual(second_id_gen.get_current_token_for_writer("first"), 7)
         self.assertEqual(second_id_gen.get_current_token_for_writer("second"), 7)
 
         # Try allocating a new ID gen and check that we only see position
@@ -398,6 +396,56 @@ class MultiWriterIdGeneratorTestCase(HomeserverTestCase):
         second_id_gen.advance("first", 8)
         self.assertEqual(second_id_gen.get_positions(), {"first": 8, "second": 9})
 
+    def test_multi_instance_empty_row(self) -> None:
+        """Test that reads and writes from multiple processes are handled
+        correctly, when one of the writers starts without any rows.
+        """
+        # Insert some rows for two out of three of the ID gens.
+        self._insert_rows("first", 3)
+        self._insert_rows("second", 4)
+
+        first_id_gen = self._create_id_generator(
+            "first", writers=["first", "second", "third"]
+        )
+        second_id_gen = self._create_id_generator(
+            "second", writers=["first", "second", "third"]
+        )
+        third_id_gen = self._create_id_generator(
+            "third", writers=["first", "second", "third"]
+        )
+
+        self.assertEqual(
+            first_id_gen.get_positions(), {"first": 3, "second": 7, "third": 7}
+        )
+        self.assertEqual(first_id_gen.get_current_token_for_writer("first"), 7)
+        self.assertEqual(first_id_gen.get_current_token_for_writer("second"), 7)
+        self.assertEqual(first_id_gen.get_current_token_for_writer("third"), 7)
+
+        self.assertEqual(
+            second_id_gen.get_positions(), {"first": 3, "second": 7, "third": 7}
+        )
+        self.assertEqual(second_id_gen.get_current_token_for_writer("first"), 7)
+        self.assertEqual(second_id_gen.get_current_token_for_writer("second"), 7)
+        self.assertEqual(second_id_gen.get_current_token_for_writer("third"), 7)
+
+        # Try allocating a new ID gen and check that we only see position
+        # advanced after we leave the context manager.
+
+        async def _get_next_async() -> None:
+            async with third_id_gen.get_next() as stream_id:
+                self.assertEqual(stream_id, 8)
+
+                self.assertEqual(
+                    third_id_gen.get_positions(), {"first": 3, "second": 7, "third": 7}
+                )
+                self.assertEqual(third_id_gen.get_persisted_upto_position(), 7)
+
+        self.get_success(_get_next_async())
+
+        self.assertEqual(
+            third_id_gen.get_positions(), {"first": 3, "second": 7, "third": 8}
+        )
+
     def test_get_next_txn(self) -> None:
         """Test that the `get_next_txn` function works correctly."""
 
@@ -600,6 +648,70 @@ class MultiWriterIdGeneratorTestCase(HomeserverTestCase):
         with self.assertRaises(IncorrectDatabaseSetup):
             self._create_id_generator("first")
 
+    def test_minimal_local_token(self) -> None:
+        self._insert_rows("first", 3)
+        self._insert_rows("second", 4)
+
+        first_id_gen = self._create_id_generator("first", writers=["first", "second"])
+        second_id_gen = self._create_id_generator("second", writers=["first", "second"])
+
+        self.assertEqual(first_id_gen.get_positions(), {"first": 3, "second": 7})
+        self.assertEqual(first_id_gen.get_minimal_local_current_token(), 3)
+
+        self.assertEqual(second_id_gen.get_positions(), {"first": 3, "second": 7})
+        self.assertEqual(second_id_gen.get_minimal_local_current_token(), 7)
+
+    def test_current_token_gap(self) -> None:
+        """Test that getting the current token for a writer returns the maximal
+        token when there are no writes.
+        """
+        self._insert_rows("first", 3)
+        self._insert_rows("second", 4)
+
+        first_id_gen = self._create_id_generator(
+            "first", writers=["first", "second", "third"]
+        )
+        second_id_gen = self._create_id_generator(
+            "second", writers=["first", "second", "third"]
+        )
+
+        self.assertEqual(second_id_gen.get_current_token_for_writer("first"), 7)
+        self.assertEqual(second_id_gen.get_current_token_for_writer("second"), 7)
+        self.assertEqual(second_id_gen.get_current_token(), 7)
+
+        # Check that the first ID gen advancing causes the second ID gen to
+        # advance (as the second ID gen has nothing in flight).
+
+        async def _get_next_async() -> None:
+            async with first_id_gen.get_next_mult(2):
+                pass
+
+        self.get_success(_get_next_async())
+        second_id_gen.advance("first", 9)
+
+        self.assertEqual(second_id_gen.get_current_token_for_writer("first"), 9)
+        self.assertEqual(second_id_gen.get_current_token_for_writer("second"), 9)
+        self.assertEqual(second_id_gen.get_current_token(), 7)
+
+        # Check that the first ID gen advancing doesn't advance the second ID
+        # gen when the second ID gen has stuff in flight.
+        self.get_success(_get_next_async())
+
+        ctxmgr = second_id_gen.get_next()
+        self.get_success(ctxmgr.__aenter__())
+
+        second_id_gen.advance("first", 11)
+
+        self.assertEqual(second_id_gen.get_current_token_for_writer("first"), 11)
+        self.assertEqual(second_id_gen.get_current_token_for_writer("second"), 9)
+        self.assertEqual(second_id_gen.get_current_token(), 7)
+
+        self.get_success(ctxmgr.__aexit__(None, None, None))
+
+        self.assertEqual(second_id_gen.get_current_token_for_writer("first"), 11)
+        self.assertEqual(second_id_gen.get_current_token_for_writer("second"), 12)
+        self.assertEqual(second_id_gen.get_current_token(), 7)
+
 
 class BackwardsMultiWriterIdGeneratorTestCase(HomeserverTestCase):
     """Tests MultiWriterIdGenerator that produce *negative* stream IDs."""
@@ -712,8 +824,8 @@ class BackwardsMultiWriterIdGeneratorTestCase(HomeserverTestCase):
 
         self.get_success(_get_next_async())
 
-        self.assertEqual(id_gen_1.get_positions(), {"first": -1})
-        self.assertEqual(id_gen_2.get_positions(), {"first": -1})
+        self.assertEqual(id_gen_1.get_positions(), {"first": -1, "second": -1})
+        self.assertEqual(id_gen_2.get_positions(), {"first": -1, "second": -1})
         self.assertEqual(id_gen_1.get_persisted_upto_position(), -1)
         self.assertEqual(id_gen_2.get_persisted_upto_position(), -1)
 
@@ -822,11 +934,11 @@ class MultiTableMultiWriterIdGeneratorTestCase(HomeserverTestCase):
         second_id_gen = self._create_id_generator("second", writers=["first", "second"])
 
         self.assertEqual(first_id_gen.get_positions(), {"first": 3, "second": 6})
-        self.assertEqual(first_id_gen.get_current_token_for_writer("first"), 3)
-        self.assertEqual(first_id_gen.get_current_token_for_writer("second"), 6)
+        self.assertEqual(first_id_gen.get_current_token_for_writer("first"), 7)
+        self.assertEqual(first_id_gen.get_current_token_for_writer("second"), 7)
         self.assertEqual(first_id_gen.get_persisted_upto_position(), 7)
 
         self.assertEqual(second_id_gen.get_positions(), {"first": 3, "second": 7})
-        self.assertEqual(second_id_gen.get_current_token_for_writer("first"), 3)
+        self.assertEqual(second_id_gen.get_current_token_for_writer("first"), 7)
         self.assertEqual(second_id_gen.get_current_token_for_writer("second"), 7)
         self.assertEqual(second_id_gen.get_persisted_upto_position(), 7)