Merge pull request #6240 from matrix-org/erikj/split_out_persistence_store

Move persist_events out from main data store.

18 files changed, 818 insertions, 667 deletions

diff --git a/changelog.d/6240.misc b/changelog.d/6240.misc
new file mode 100644
index 0000000000..0b3d7a14a1
--- /dev/null
+++ b/changelog.d/6240.misc
@@ -0,0 +1 @@
+Move `persist_events` out from main data store.
diff --git a/synapse/handlers/federation.py b/synapse/handlers/federation.py
index 2da520e6e8..08276fdebf 100644
--- a/synapse/handlers/federation.py
+++ b/synapse/handlers/federation.py
@@ -109,6 +109,7 @@ class FederationHandler(BaseHandler):
         self.hs = hs
 
         self.store = hs.get_datastore()
+        self.storage = hs.get_storage()
         self.federation_client = hs.get_federation_client()
         self.state_handler = hs.get_state_handler()
         self.server_name = hs.hostname
@@ -2664,7 +2665,7 @@ class FederationHandler(BaseHandler):
                 backfilled=backfilled,
             )
         else:
-            max_stream_id = yield self.store.persist_events(
+            max_stream_id = yield self.storage.persistence.persist_events(
                 event_and_contexts, backfilled=backfilled
             )
 
diff --git a/synapse/handlers/message.py b/synapse/handlers/message.py
index 0f8cce8ffe..7908a2d52c 100644
--- a/synapse/handlers/message.py
+++ b/synapse/handlers/message.py
@@ -234,6 +234,7 @@ class EventCreationHandler(object):
         self.hs = hs
         self.auth = hs.get_auth()
         self.store = hs.get_datastore()
+        self.storage = hs.get_storage()
         self.state = hs.get_state_handler()
         self.clock = hs.get_clock()
         self.validator = EventValidator()
@@ -868,7 +869,7 @@ class EventCreationHandler(object):
             if prev_state_ids:
                 raise AuthError(403, "Changing the room create event is forbidden")
 
-        (event_stream_id, max_stream_id) = yield self.store.persist_event(
+        event_stream_id, max_stream_id = yield self.storage.persistence.persist_event(
             event, context=context
         )
 
diff --git a/synapse/server.py b/synapse/server.py
index 1fcc7375d3..0b81af646c 100644
--- a/synapse/server.py
+++ b/synapse/server.py
@@ -95,6 +95,7 @@ from synapse.server_notices.worker_server_notices_sender import (
     WorkerServerNoticesSender,
 )
 from synapse.state import StateHandler, StateResolutionHandler
+from synapse.storage import DataStores, Storage
 from synapse.streams.events import EventSources
 from synapse.util import Clock
 from synapse.util.distributor import Distributor
@@ -196,6 +197,7 @@ class HomeServer(object):
         "account_validity_handler",
         "saml_handler",
         "event_client_serializer",
+        "storage",
     ]
 
     REQUIRED_ON_MASTER_STARTUP = ["user_directory_handler", "stats_handler"]
@@ -224,7 +226,7 @@ class HomeServer(object):
         self.admin_redaction_ratelimiter = Ratelimiter()
         self.registration_ratelimiter = Ratelimiter()
 
-        self.datastore = None
+        self.datastores = None
 
         # Other kwargs are explicit dependencies
         for depname in kwargs:
@@ -233,7 +235,8 @@ class HomeServer(object):
     def setup(self):
         logger.info("Setting up.")
         with self.get_db_conn() as conn:
-            self.datastore = self.DATASTORE_CLASS(conn, self)
+            datastore = self.DATASTORE_CLASS(conn, self)
+            self.datastores = DataStores(datastore, conn, self)
             conn.commit()
         logger.info("Finished setting up.")
 
@@ -266,7 +269,7 @@ class HomeServer(object):
         return self.clock
 
     def get_datastore(self):
-        return self.datastore
+        return self.datastores.main
 
     def get_config(self):
         return self.config
@@ -537,6 +540,9 @@ class HomeServer(object):
     def build_event_client_serializer(self):
         return EventClientSerializer(self)
 
+    def build_storage(self) -> Storage:
+        return Storage(self, self.datastores)
+
     def remove_pusher(self, app_id, push_key, user_id):
         return self.get_pusherpool().remove_pusher(app_id, push_key, user_id)
 
diff --git a/synapse/storage/__init__.py b/synapse/storage/__init__.py
index a249ecd219..a6429d17ed 100644
--- a/synapse/storage/__init__.py
+++ b/synapse/storage/__init__.py
@@ -27,7 +27,24 @@ data stores associated with them (e.g. the schema version tables), which are
 stored in `synapse.storage.schema`.
 """
 
-from synapse.storage.data_stores.main import DataStore  # noqa: F401
+from synapse.storage.data_stores import DataStores
+from synapse.storage.data_stores.main import DataStore
+from synapse.storage.persist_events import EventsPersistenceStorage
+
+__all__ = ["DataStores", "DataStore"]
+
+
+class Storage(object):
+    """The high level interfaces for talking to various storage layers.
+    """
+
+    def __init__(self, hs, stores: DataStores):
+        # We include the main data store here mainly so that we don't have to
+        # rewrite all the existing code to split it into high vs low level
+        # interfaces.
+        self.main = stores.main
+
+        self.persistence = EventsPersistenceStorage(hs, stores)
 
 
 def are_all_users_on_domain(txn, database_engine, domain):
diff --git a/synapse/storage/data_stores/__init__.py b/synapse/storage/data_stores/__init__.py
index 56094078ed..cb184a98cc 100644
--- a/synapse/storage/data_stores/__init__.py
+++ b/synapse/storage/data_stores/__init__.py
@@ -12,3 +12,15 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
+
+
+class DataStores(object):
+    """The various data stores.
+
+    These are low level interfaces to physical databases.
+    """
+
+    def __init__(self, main_store, db_conn, hs):
+        # Note we pass in the main store here as workers use a different main
+        # store.
+        self.main = main_store
diff --git a/synapse/storage/data_stores/main/events.py b/synapse/storage/data_stores/main/events.py
index 067e77ae00..7c3607f308 100644
--- a/synapse/storage/data_stores/main/events.py
+++ b/synapse/storage/data_stores/main/events.py
@@ -17,14 +17,14 @@
 
 import itertools
 import logging
-from collections import Counter as c_counter, OrderedDict, deque, namedtuple
+from collections import Counter as c_counter, OrderedDict, namedtuple
 from functools import wraps
 
 from six import iteritems, text_type
 from six.moves import range
 
 from canonicaljson import json
-from prometheus_client import Counter, Histogram
+from prometheus_client import Counter
 
 from twisted.internet import defer
 
@@ -34,11 +34,9 @@ from synapse.api.errors import SynapseError
 from synapse.events import EventBase  # noqa: F401
 from synapse.events.snapshot import EventContext  # noqa: F401
 from synapse.events.utils import prune_event_dict
-from synapse.logging.context import PreserveLoggingContext, make_deferred_yieldable
 from synapse.logging.utils import log_function
 from synapse.metrics import BucketCollector
 from synapse.metrics.background_process_metrics import run_as_background_process
-from synapse.state import StateResolutionStore
 from synapse.storage._base import make_in_list_sql_clause
 from synapse.storage.background_updates import BackgroundUpdateStore
 from synapse.storage.data_stores.main.event_federation import EventFederationStore
@@ -46,10 +44,8 @@ from synapse.storage.data_stores.main.events_worker import EventsWorkerStore
 from synapse.storage.data_stores.main.state import StateGroupWorkerStore
 from synapse.types import RoomStreamToken, get_domain_from_id
 from synapse.util import batch_iter
-from synapse.util.async_helpers import ObservableDeferred
 from synapse.util.caches.descriptors import cached, cachedInlineCallbacks
 from synapse.util.frozenutils import frozendict_json_encoder
-from synapse.util.metrics import Measure
 
 logger = logging.getLogger(__name__)
 
@@ -60,37 +56,6 @@ event_counter = Counter(
     ["type", "origin_type", "origin_entity"],
 )
 
-# The number of times we are recalculating the current state
-state_delta_counter = Counter("synapse_storage_events_state_delta", "")
-
-# The number of times we are recalculating state when there is only a
-# single forward extremity
-state_delta_single_event_counter = Counter(
-    "synapse_storage_events_state_delta_single_event", ""
-)
-
-# The number of times we are reculating state when we could have resonably
-# calculated the delta when we calculated the state for an event we were
-# persisting.
-state_delta_reuse_delta_counter = Counter(
-    "synapse_storage_events_state_delta_reuse_delta", ""
-)
-
-# The number of forward extremities for each new event.
-forward_extremities_counter = Histogram(
-    "synapse_storage_events_forward_extremities_persisted",
-    "Number of forward extremities for each new event",
-    buckets=(1, 2, 3, 5, 7, 10, 15, 20, 50, 100, 200, 500, "+Inf"),
-)
-
-# The number of stale forward extremities for each new event. Stale extremities
-# are those that were in the previous set of extremities as well as the new.
-stale_forward_extremities_counter = Histogram(
-    "synapse_storage_events_stale_forward_extremities_persisted",
-    "Number of unchanged forward extremities for each new event",
-    buckets=(0, 1, 2, 3, 5, 7, 10, 15, 20, 50, 100, 200, 500, "+Inf"),
-)
-
 
 def encode_json(json_object):
     """
@@ -102,110 +67,6 @@ def encode_json(json_object):
     return out
 
 
-class _EventPeristenceQueue(object):
-    """Queues up events so that they can be persisted in bulk with only one
-    concurrent transaction per room.
-    """
-
-    _EventPersistQueueItem = namedtuple(
-        "_EventPersistQueueItem", ("events_and_contexts", "backfilled", "deferred")
-    )
-
-    def __init__(self):
-        self._event_persist_queues = {}
-        self._currently_persisting_rooms = set()
-
-    def add_to_queue(self, room_id, events_and_contexts, backfilled):
-        """Add events to the queue, with the given persist_event options.
-
-        NB: due to the normal usage pattern of this method, it does *not*
-        follow the synapse logcontext rules, and leaves the logcontext in
-        place whether or not the returned deferred is ready.
-
-        Args:
-            room_id (str):
-            events_and_contexts (list[(EventBase, EventContext)]):
-            backfilled (bool):
-
-        Returns:
-            defer.Deferred: a deferred which will resolve once the events are
-                persisted. Runs its callbacks *without* a logcontext.
-        """
-        queue = self._event_persist_queues.setdefault(room_id, deque())
-        if queue:
-            # if the last item in the queue has the same `backfilled` setting,
-            # we can just add these new events to that item.
-            end_item = queue[-1]
-            if end_item.backfilled == backfilled:
-                end_item.events_and_contexts.extend(events_and_contexts)
-                return end_item.deferred.observe()
-
-        deferred = ObservableDeferred(defer.Deferred(), consumeErrors=True)
-
-        queue.append(
-            self._EventPersistQueueItem(
-                events_and_contexts=events_and_contexts,
-                backfilled=backfilled,
-                deferred=deferred,
-            )
-        )
-
-        return deferred.observe()
-
-    def handle_queue(self, room_id, per_item_callback):
-        """Attempts to handle the queue for a room if not already being handled.
-
-        The given callback will be invoked with for each item in the queue,
-        of type _EventPersistQueueItem. The per_item_callback will continuously
-        be called with new items, unless the queue becomnes empty. The return
-        value of the function will be given to the deferreds waiting on the item,
-        exceptions will be passed to the deferreds as well.
-
-        This function should therefore be called whenever anything is added
-        to the queue.
-
-        If another callback is currently handling the queue then it will not be
-        invoked.
-        """
-
-        if room_id in self._currently_persisting_rooms:
-            return
-
-        self._currently_persisting_rooms.add(room_id)
-
-        @defer.inlineCallbacks
-        def handle_queue_loop():
-            try:
-                queue = self._get_drainining_queue(room_id)
-                for item in queue:
-                    try:
-                        ret = yield per_item_callback(item)
-                    except Exception:
-                        with PreserveLoggingContext():
-                            item.deferred.errback()
-                    else:
-                        with PreserveLoggingContext():
-                            item.deferred.callback(ret)
-            finally:
-                queue = self._event_persist_queues.pop(room_id, None)
-                if queue:
-                    self._event_persist_queues[room_id] = queue
-                self._currently_persisting_rooms.discard(room_id)
-
-        # set handle_queue_loop off in the background
-        run_as_background_process("persist_events", handle_queue_loop)
-
-    def _get_drainining_queue(self, room_id):
-        queue = self._event_persist_queues.setdefault(room_id, deque())
-
-        try:
-            while True:
-                yield queue.popleft()
-        except IndexError:
-            # Queue has been drained.
-            pass
-
-
 _EventCacheEntry = namedtuple("_EventCacheEntry", ("event", "redacted_event"))
 
 
@@ -241,9 +102,6 @@ class EventsStore(
     def __init__(self, db_conn, hs):
         super(EventsStore, self).__init__(db_conn, hs)
 
-        self._event_persist_queue = _EventPeristenceQueue()
-        self._state_resolution_handler = hs.get_state_resolution_handler()
-
         # Collect metrics on the number of forward extremities that exist.
         # Counter of number of extremities to count
         self._current_forward_extremities_amount = c_counter()
@@ -286,340 +144,106 @@ class EventsStore(
         res = yield self.runInteraction("read_forward_extremities", fetch)
         self._current_forward_extremities_amount = c_counter(list(x[0] for x in res))
 
-    @defer.inlineCallbacks
-    def persist_events(self, events_and_contexts, backfilled=False):
-        """
-        Write events to the database
-        Args:
-            events_and_contexts: list of tuples of (event, context)
-            backfilled (bool): Whether the results are retrieved from federation
-                via backfill or not. Used to determine if they're "new" events
-                which might update the current state etc.
-
-        Returns:
-            Deferred[int]: the stream ordering of the latest persisted event
-        """
-        partitioned = {}
-        for event, ctx in events_and_contexts:
-            partitioned.setdefault(event.room_id, []).append((event, ctx))
-
-        deferreds = []
-        for room_id, evs_ctxs in iteritems(partitioned):
-            d = self._event_persist_queue.add_to_queue(
-                room_id, evs_ctxs, backfilled=backfilled
-            )
-            deferreds.append(d)
-
-        for room_id in partitioned:
-            self._maybe_start_persisting(room_id)
-
-        yield make_deferred_yieldable(
-            defer.gatherResults(deferreds, consumeErrors=True)
-        )
-
-        max_persisted_id = yield self._stream_id_gen.get_current_token()
-
-        return max_persisted_id
-
-    @defer.inlineCallbacks
-    @log_function
-    def persist_event(self, event, context, backfilled=False):
-        """
-
-        Args:
-            event (EventBase):
-            context (EventContext):
-            backfilled (bool):
-
-        Returns:
-            Deferred: resolves to (int, int): the stream ordering of ``event``,
-            and the stream ordering of the latest persisted event
-        """
-        deferred = self._event_persist_queue.add_to_queue(
-            event.room_id, [(event, context)], backfilled=backfilled
-        )
-
-        self._maybe_start_persisting(event.room_id)
-
-        yield make_deferred_yieldable(deferred)
-
-        max_persisted_id = yield self._stream_id_gen.get_current_token()
-        return (event.internal_metadata.stream_ordering, max_persisted_id)
-
-    def _maybe_start_persisting(self, room_id):
-        @defer.inlineCallbacks
-        def persisting_queue(item):
-            with Measure(self._clock, "persist_events"):
-                yield self._persist_events(
-                    item.events_and_contexts, backfilled=item.backfilled
-                )
-
-        self._event_persist_queue.handle_queue(room_id, persisting_queue)
-
     @_retry_on_integrity_error
     @defer.inlineCallbacks
-    def _persist_events(
-        self, events_and_contexts, backfilled=False, delete_existing=False
+    def _persist_events_and_state_updates(
+        self,
+        events_and_contexts,
+        current_state_for_room,
+        state_delta_for_room,
+        new_forward_extremeties,
+        backfilled=False,
+        delete_existing=False,
     ):
-        """Persist events to db
+        """Persist a set of events alongside updates to the current state and
+        forward extremities tables.
 
         Args:
             events_and_contexts (list[(EventBase, EventContext)]):
-            backfilled (bool):
+            current_state_for_room (dict[str, dict]): Map from room_id to the
+                current state of the room based on forward extremities
+            state_delta_for_room (dict[str, tuple]): Map from room_id to tuple
+                of `(to_delete, to_insert)` where to_delete is a list
+                of type/state keys to remove from current state, and to_insert
+                is a map (type,key)->event_id giving the state delta in each
+                room.
+            new_forward_extremities (dict[str, list[str]]): Map from room_id
+                to list of event IDs that are the new forward extremities of
+                the room.
+            backfilled (bool)
             delete_existing (bool):
 
         Returns:
             Deferred: resolves when the events have been persisted
         """
-        if not events_and_contexts:
-            return
-
-        chunks = [
-            events_and_contexts[x : x + 100]
-            for x in range(0, len(events_and_contexts), 100)
-        ]
-
-        for chunk in chunks:
-            # We can't easily parallelize these since different chunks
-            # might contain the same event. :(
-
-            # NB: Assumes that we are only persisting events for one room
-            # at a time.
 
-            # map room_id->list[event_ids] giving the new forward
-            # extremities in each room
-            new_forward_extremeties = {}
+        # We want to calculate the stream orderings as late as possible, as
+        # we only notify after all events with a lesser stream ordering have
+        # been persisted. I.e. if we spend 10s inside the with block then
+        # that will delay all subsequent events from being notified about.
+        # Hence why we do it down here rather than wrapping the entire
+        # function.
+        #
+        # Its safe to do this after calculating the state deltas etc as we
+        # only need to protect the *persistence* of the events. This is to
+        # ensure that queries of the form "fetch events since X" don't
+        # return events and stream positions after events that are still in
+        # flight, as otherwise subsequent requests "fetch event since Y"
+        # will not return those events.
+        #
+        # Note: Multiple instances of this function cannot be in flight at
+        # the same time for the same room.
+        if backfilled:
+            stream_ordering_manager = self._backfill_id_gen.get_next_mult(
+                len(events_and_contexts)
+            )
+        else:
+            stream_ordering_manager = self._stream_id_gen.get_next_mult(
+                len(events_and_contexts)
+            )
 
-            # map room_id->(type,state_key)->event_id tracking the full
-            # state in each room after adding these events.
-            # This is simply used to prefill the get_current_state_ids
-            # cache
-            current_state_for_room = {}
+        with stream_ordering_manager as stream_orderings:
+            for (event, context), stream in zip(events_and_contexts, stream_orderings):
+                event.internal_metadata.stream_ordering = stream
 
-            # map room_id->(to_delete, to_insert) where to_delete is a list
-            # of type/state keys to remove from current state, and to_insert
-            # is a map (type,key)->event_id giving the state delta in each
-            # room
-            state_delta_for_room = {}
+            yield self.runInteraction(
+                "persist_events",
+                self._persist_events_txn,
+                events_and_contexts=events_and_contexts,
+                backfilled=backfilled,
+                delete_existing=delete_existing,
+                state_delta_for_room=state_delta_for_room,
+                new_forward_extremeties=new_forward_extremeties,
+            )
+            persist_event_counter.inc(len(events_and_contexts))
 
             if not backfilled:
-                with Measure(self._clock, "_calculate_state_and_extrem"):
-                    # Work out the new "current state" for each room.
-                    # We do this by working out what the new extremities are and then
-                    # calculating the state from that.
-                    events_by_room = {}
-                    for event, context in chunk:
-                        events_by_room.setdefault(event.room_id, []).append(
-                            (event, context)
-                        )
-
-                    for room_id, ev_ctx_rm in iteritems(events_by_room):
-                        latest_event_ids = yield self.get_latest_event_ids_in_room(
-                            room_id
-                        )
-                        new_latest_event_ids = yield self._calculate_new_extremities(
-                            room_id, ev_ctx_rm, latest_event_ids
-                        )
-
-                        latest_event_ids = set(latest_event_ids)
-                        if new_latest_event_ids == latest_event_ids:
-                            # No change in extremities, so no change in state
-                            continue
-
-                        # there should always be at least one forward extremity.
-                        # (except during the initial persistence of the send_join
-                        # results, in which case there will be no existing
-                        # extremities, so we'll `continue` above and skip this bit.)
-                        assert new_latest_event_ids, "No forward extremities left!"
-
-                        new_forward_extremeties[room_id] = new_latest_event_ids
-
-                        len_1 = (
-                            len(latest_event_ids) == 1
-                            and len(new_latest_event_ids) == 1
-                        )
-                        if len_1:
-                            all_single_prev_not_state = all(
-                                len(event.prev_event_ids()) == 1
-                                and not event.is_state()
-                                for event, ctx in ev_ctx_rm
-                            )
-                            # Don't bother calculating state if they're just
-                            # a long chain of single ancestor non-state events.
-                            if all_single_prev_not_state:
-                                continue
-
-                        state_delta_counter.inc()
-                        if len(new_latest_event_ids) == 1:
-                            state_delta_single_event_counter.inc()
-
-                            # This is a fairly handwavey check to see if we could
-                            # have guessed what the delta would have been when
-                            # processing one of these events.
-                            # What we're interested in is if the latest extremities
-                            # were the same when we created the event as they are
-                            # now. When this server creates a new event (as opposed
-                            # to receiving it over federation) it will use the
-                            # forward extremities as the prev_events, so we can
-                            # guess this by looking at the prev_events and checking
-                            # if they match the current forward extremities.
-                            for ev, _ in ev_ctx_rm:
-                                prev_event_ids = set(ev.prev_event_ids())
-                                if latest_event_ids == prev_event_ids:
-                                    state_delta_reuse_delta_counter.inc()
-                                    break
-
-                        logger.info("Calculating state delta for room %s", room_id)
-                        with Measure(
-                            self._clock, "persist_events.get_new_state_after_events"
-                        ):
-                            res = yield self._get_new_state_after_events(
-                                room_id,
-                                ev_ctx_rm,
-                                latest_event_ids,
-                                new_latest_event_ids,
-                            )
-                            current_state, delta_ids = res
-
-                        # If either are not None then there has been a change,
-                        # and we need to work out the delta (or use that
-                        # given)
-                        if delta_ids is not None:
-                            # If there is a delta we know that we've
-                            # only added or replaced state, never
-                            # removed keys entirely.
-                            state_delta_for_room[room_id] = ([], delta_ids)
-                        elif current_state is not None:
-                            with Measure(
-                                self._clock, "persist_events.calculate_state_delta"
-                            ):
-                                delta = yield self._calculate_state_delta(
-                                    room_id, current_state
-                                )
-                            state_delta_for_room[room_id] = delta
-
-                        # If we have the current_state then lets prefill
-                        # the cache with it.
-                        if current_state is not None:
-                            current_state_for_room[room_id] = current_state
-
-            # We want to calculate the stream orderings as late as possible, as
-            # we only notify after all events with a lesser stream ordering have
-            # been persisted. I.e. if we spend 10s inside the with block then
-            # that will delay all subsequent events from being notified about.
-            # Hence why we do it down here rather than wrapping the entire
-            # function.
-            #
-            # Its safe to do this after calculating the state deltas etc as we
-            # only need to protect the *persistence* of the events. This is to
-            # ensure that queries of the form "fetch events since X" don't
-            # return events and stream positions after events that are still in
-            # flight, as otherwise subsequent requests "fetch event since Y"
-            # will not return those events.
-            #
-            # Note: Multiple instances of this function cannot be in flight at
-            # the same time for the same room.
-            if backfilled:
-                stream_ordering_manager = self._backfill_id_gen.get_next_mult(
-                    len(chunk)
-                )
-            else:
-                stream_ordering_manager = self._stream_id_gen.get_next_mult(len(chunk))
-
-            with stream_ordering_manager as stream_orderings:
-                for (event, context), stream in zip(chunk, stream_orderings):
-                    event.internal_metadata.stream_ordering = stream
-
-                yield self.runInteraction(
-                    "persist_events",
-                    self._persist_events_txn,
-                    events_and_contexts=chunk,
-                    backfilled=backfilled,
-                    delete_existing=delete_existing,
-                    state_delta_for_room=state_delta_for_room,
-                    new_forward_extremeties=new_forward_extremeties,
+                # backfilled events have negative stream orderings, so we don't
+                # want to set the event_persisted_position to that.
+                synapse.metrics.event_persisted_position.set(
+                    events_and_contexts[-1][0].internal_metadata.stream_ordering
                 )
-                persist_event_counter.inc(len(chunk))
-
-                if not backfilled:
-                    # backfilled events have negative stream orderings, so we don't
-                    # want to set the event_persisted_position to that.
-                    synapse.metrics.event_persisted_position.set(
-                        chunk[-1][0].internal_metadata.stream_ordering
-                    )
-
-                for event, context in chunk:
-                    if context.app_service:
-                        origin_type = "local"
-                        origin_entity = context.app_service.id
-                    elif self.hs.is_mine_id(event.sender):
-                        origin_type = "local"
-                        origin_entity = "*client*"
-                    else:
-                        origin_type = "remote"
-                        origin_entity = get_domain_from_id(event.sender)
-
-                    event_counter.labels(event.type, origin_type, origin_entity).inc()
-
-                for room_id, new_state in iteritems(current_state_for_room):
-                    self.get_current_state_ids.prefill((room_id,), new_state)
-
-                for room_id, latest_event_ids in iteritems(new_forward_extremeties):
-                    self.get_latest_event_ids_in_room.prefill(
-                        (room_id,), list(latest_event_ids)
-                    )
-
-    @defer.inlineCallbacks
-    def _calculate_new_extremities(self, room_id, event_contexts, latest_event_ids):
-        """Calculates the new forward extremities for a room given events to
-        persist.
-
-        Assumes that we are only persisting events for one room at a time.
-        """
-
-        # we're only interested in new events which aren't outliers and which aren't
-        # being rejected.
-        new_events = [
-            event
-            for event, ctx in event_contexts
-            if not event.internal_metadata.is_outlier()
-            and not ctx.rejected
-            and not event.internal_metadata.is_soft_failed()
-        ]
 
-        latest_event_ids = set(latest_event_ids)
-
-        # start with the existing forward extremities
-        result = set(latest_event_ids)
-
-        # add all the new events to the list
-        result.update(event.event_id for event in new_events)
-
-        # Now remove all events which are prev_events of any of the new events
-        result.difference_update(
-            e_id for event in new_events for e_id in event.prev_event_ids()
-        )
-
-        # Remove any events which are prev_events of any existing events.
-        existing_prevs = yield self._get_events_which_are_prevs(result)
-        result.difference_update(existing_prevs)
+            for event, context in events_and_contexts:
+                if context.app_service:
+                    origin_type = "local"
+                    origin_entity = context.app_service.id
+                elif self.hs.is_mine_id(event.sender):
+                    origin_type = "local"
+                    origin_entity = "*client*"
+                else:
+                    origin_type = "remote"
+                    origin_entity = get_domain_from_id(event.sender)
 
-        # Finally handle the case where the new events have soft-failed prev
-        # events. If they do we need to remove them and their prev events,
-        # otherwise we end up with dangling extremities.
-        existing_prevs = yield self._get_prevs_before_rejected(
-            e_id for event in new_events for e_id in event.prev_event_ids()
-        )
-        result.difference_update(existing_prevs)
+                event_counter.labels(event.type, origin_type, origin_entity).inc()
 
-        # We only update metrics for events that change forward extremities
-        # (e.g. we ignore backfill/outliers/etc)
-        if result != latest_event_ids:
-            forward_extremities_counter.observe(len(result))
-            stale = latest_event_ids & result
-            stale_forward_extremities_counter.observe(len(stale))
+            for room_id, new_state in iteritems(current_state_for_room):
+                self.get_current_state_ids.prefill((room_id,), new_state)
 
-        return result
+            for room_id, latest_event_ids in iteritems(new_forward_extremeties):
+                self.get_latest_event_ids_in_room.prefill(
+                    (room_id,), list(latest_event_ids)
+                )
 
     @defer.inlineCallbacks
     def _get_events_which_are_prevs(self, event_ids):
@@ -725,188 +349,6 @@ class EventsStore(
 
         return existing_prevs
 
-    @defer.inlineCallbacks
-    def _get_new_state_after_events(
-        self, room_id, events_context, old_latest_event_ids, new_latest_event_ids
-    ):
-        """Calculate the current state dict after adding some new events to
-        a room
-
-        Args:
-            room_id (str):
-                room to which the events are being added. Used for logging etc
-
-            events_context (list[(EventBase, EventContext)]):
-                events and contexts which are being added to the room
-
-            old_latest_event_ids (iterable[str]):
-                the old forward extremities for the room.
-
-            new_latest_event_ids (iterable[str]):
-                the new forward extremities for the room.
-
-        Returns:
-            Deferred[tuple[dict[(str,str), str]|None, dict[(str,str), str]|None]]:
-            Returns a tuple of two state maps, the first being the full new current
-            state and the second being the delta to the existing current state.
-            If both are None then there has been no change.
-
-            If there has been a change then we only return the delta if its
-            already been calculated. Conversely if we do know the delta then
-            the new current state is only returned if we've already calculated
-            it.
-        """
-        # map from state_group to ((type, key) -> event_id) state map
-        state_groups_map = {}
-
-        # Map from (prev state group, new state group) -> delta state dict
-        state_group_deltas = {}
-
-        for ev, ctx in events_context:
-            if ctx.state_group is None:
-                # This should only happen for outlier events.
-                if not ev.internal_metadata.is_outlier():
-                    raise Exception(
-                        "Context for new event %s has no state "
-                        "group" % (ev.event_id,)
-                    )
-                continue
-
-            if ctx.state_group in state_groups_map:
-                continue
-
-            # We're only interested in pulling out state that has already
-            # been cached in the context. We'll pull stuff out of the DB later
-            # if necessary.
-            current_state_ids = ctx.get_cached_current_state_ids()
-            if current_state_ids is not None:
-                state_groups_map[ctx.state_group] = current_state_ids
-
-            if ctx.prev_group:
-                state_group_deltas[(ctx.prev_group, ctx.state_group)] = ctx.delta_ids
-
-        # We need to map the event_ids to their state groups. First, let's
-        # check if the event is one we're persisting, in which case we can
-        # pull the state group from its context.
-        # Otherwise we need to pull the state group from the database.
-
-        # Set of events we need to fetch groups for. (We know none of the old
-        # extremities are going to be in events_context).
-        missing_event_ids = set(old_latest_event_ids)
-
-        event_id_to_state_group = {}
-        for event_id in new_latest_event_ids:
-            # First search in the list of new events we're adding.
-            for ev, ctx in events_context:
-                if event_id == ev.event_id and ctx.state_group is not None:
-                    event_id_to_state_group[event_id] = ctx.state_group
-                    break
-            else:
-                # If we couldn't find it, then we'll need to pull
-                # the state from the database
-                missing_event_ids.add(event_id)
-
-        if missing_event_ids:
-            # Now pull out the state groups for any missing events from DB
-            event_to_groups = yield self._get_state_group_for_events(missing_event_ids)
-            event_id_to_state_group.update(event_to_groups)
-
-        # State groups of old_latest_event_ids
-        old_state_groups = set(
-            event_id_to_state_group[evid] for evid in old_latest_event_ids
-        )
-
-        # State groups of new_latest_event_ids
-        new_state_groups = set(
-            event_id_to_state_group[evid] for evid in new_latest_event_ids
-        )
-
-        # If they old and new groups are the same then we don't need to do
-        # anything.
-        if old_state_groups == new_state_groups:
-            return None, None
-
-        if len(new_state_groups) == 1 and len(old_state_groups) == 1:
-            # If we're going from one state group to another, lets check if
-            # we have a delta for that transition. If we do then we can just
-            # return that.
-
-            new_state_group = next(iter(new_state_groups))
-            old_state_group = next(iter(old_state_groups))
-
-            delta_ids = state_group_deltas.get((old_state_group, new_state_group), None)
-            if delta_ids is not None:
-                # We have a delta from the existing to new current state,
-                # so lets just return that. If we happen to already have
-                # the current state in memory then lets also return that,
-                # but it doesn't matter if we don't.
-                new_state = state_groups_map.get(new_state_group)
-                return new_state, delta_ids
-
-        # Now that we have calculated new_state_groups we need to get
-        # their state IDs so we can resolve to a single state set.
-        missing_state = new_state_groups - set(state_groups_map)
-        if missing_state:
-            group_to_state = yield self._get_state_for_groups(missing_state)
-            state_groups_map.update(group_to_state)
-
-        if len(new_state_groups) == 1:
-            # If there is only one state group, then we know what the current
-            # state is.
-            return state_groups_map[new_state_groups.pop()], None
-
-        # Ok, we need to defer to the state handler to resolve our state sets.
-
-        state_groups = {sg: state_groups_map[sg] for sg in new_state_groups}
-
-        events_map = {ev.event_id: ev for ev, _ in events_context}
-
-        # We need to get the room version, which is in the create event.
-        # Normally that'd be in the database, but its also possible that we're
-        # currently trying to persist it.
-        room_version = None
-        for ev, _ in events_context:
-            if ev.type == EventTypes.Create and ev.state_key == "":
-                room_version = ev.content.get("room_version", "1")
-                break
-
-        if not room_version:
-            room_version = yield self.get_room_version(room_id)
-
-        logger.debug("calling resolve_state_groups from preserve_events")
-        res = yield self._state_resolution_handler.resolve_state_groups(
-            room_id,
-            room_version,
-            state_groups,
-            events_map,
-            state_res_store=StateResolutionStore(self),
-        )
-
-        return res.state, None
-
-    @defer.inlineCallbacks
-    def _calculate_state_delta(self, room_id, current_state):
-        """Calculate the new state deltas for a room.
-
-        Assumes that we are only persisting events for one room at a time.
-
-        Returns:
-            tuple[list, dict] (to_delete, to_insert): where to_delete are the
-            type/state_keys to remove from current_state_events and `to_insert`
-            are the updates to current_state_events.
-        """
-        existing_state = yield self.get_current_state_ids(room_id)
-
-        to_delete = [key for key in existing_state if key not in current_state]
-
-        to_insert = {
-            key: ev_id
-            for key, ev_id in iteritems(current_state)
-            if ev_id != existing_state.get(key)
-        }
-
-        return to_delete, to_insert
-
     @log_function
     def _persist_events_txn(
         self,
diff --git a/synapse/storage/persist_events.py b/synapse/storage/persist_events.py
new file mode 100644
index 0000000000..cf66225574
--- /dev/null
+++ b/synapse/storage/persist_events.py
@@ -0,0 +1,652 @@
+# -*- coding: utf-8 -*-
+# Copyright 2014-2016 OpenMarket Ltd
+# Copyright 2018-2019 New Vector Ltd
+# Copyright 2019 The Matrix.org Foundation C.I.C.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import logging
+from collections import deque, namedtuple
+
+from six import iteritems
+from six.moves import range
+
+from prometheus_client import Counter, Histogram
+
+from twisted.internet import defer
+
+from synapse.api.constants import EventTypes
+from synapse.logging.context import PreserveLoggingContext, make_deferred_yieldable
+from synapse.metrics.background_process_metrics import run_as_background_process
+from synapse.state import StateResolutionStore
+from synapse.storage.data_stores import DataStores
+from synapse.util.async_helpers import ObservableDeferred
+from synapse.util.metrics import Measure
+
+logger = logging.getLogger(__name__)
+
+# The number of times we are recalculating the current state
+state_delta_counter = Counter("synapse_storage_events_state_delta", "")
+
+# The number of times we are recalculating state when there is only a
+# single forward extremity
+state_delta_single_event_counter = Counter(
+    "synapse_storage_events_state_delta_single_event", ""
+)
+
+# The number of times we are reculating state when we could have resonably
+# calculated the delta when we calculated the state for an event we were
+# persisting.
+state_delta_reuse_delta_counter = Counter(
+    "synapse_storage_events_state_delta_reuse_delta", ""
+)
+
+# The number of forward extremities for each new event.
+forward_extremities_counter = Histogram(
+    "synapse_storage_events_forward_extremities_persisted",
+    "Number of forward extremities for each new event",
+    buckets=(1, 2, 3, 5, 7, 10, 15, 20, 50, 100, 200, 500, "+Inf"),
+)
+
+# The number of stale forward extremities for each new event. Stale extremities
+# are those that were in the previous set of extremities as well as the new.
+stale_forward_extremities_counter = Histogram(
+    "synapse_storage_events_stale_forward_extremities_persisted",
+    "Number of unchanged forward extremities for each new event",
+    buckets=(0, 1, 2, 3, 5, 7, 10, 15, 20, 50, 100, 200, 500, "+Inf"),
+)
+
+
+class _EventPeristenceQueue(object):
+    """Queues up events so that they can be persisted in bulk with only one
+    concurrent transaction per room.
+    """
+
+    _EventPersistQueueItem = namedtuple(
+        "_EventPersistQueueItem", ("events_and_contexts", "backfilled", "deferred")
+    )
+
+    def __init__(self):
+        self._event_persist_queues = {}
+        self._currently_persisting_rooms = set()
+
+    def add_to_queue(self, room_id, events_and_contexts, backfilled):
+        """Add events to the queue, with the given persist_event options.
+
+        NB: due to the normal usage pattern of this method, it does *not*
+        follow the synapse logcontext rules, and leaves the logcontext in
+        place whether or not the returned deferred is ready.
+
+        Args:
+            room_id (str):
+            events_and_contexts (list[(EventBase, EventContext)]):
+            backfilled (bool):
+
+        Returns:
+            defer.Deferred: a deferred which will resolve once the events are
+                persisted. Runs its callbacks *without* a logcontext.
+        """
+        queue = self._event_persist_queues.setdefault(room_id, deque())
+        if queue:
+            # if the last item in the queue has the same `backfilled` setting,
+            # we can just add these new events to that item.
+            end_item = queue[-1]
+            if end_item.backfilled == backfilled:
+                end_item.events_and_contexts.extend(events_and_contexts)
+                return end_item.deferred.observe()
+
+        deferred = ObservableDeferred(defer.Deferred(), consumeErrors=True)
+
+        queue.append(
+            self._EventPersistQueueItem(
+                events_and_contexts=events_and_contexts,
+                backfilled=backfilled,
+                deferred=deferred,
+            )
+        )
+
+        return deferred.observe()
+
+    def handle_queue(self, room_id, per_item_callback):
+        """Attempts to handle the queue for a room if not already being handled.
+
+        The given callback will be invoked with for each item in the queue,
+        of type _EventPersistQueueItem. The per_item_callback will continuously
+        be called with new items, unless the queue becomnes empty. The return
+        value of the function will be given to the deferreds waiting on the item,
+        exceptions will be passed to the deferreds as well.
+
+        This function should therefore be called whenever anything is added
+        to the queue.
+
+        If another callback is currently handling the queue then it will not be
+        invoked.
+        """
+
+        if room_id in self._currently_persisting_rooms:
+            return
+
+        self._currently_persisting_rooms.add(room_id)
+
+        @defer.inlineCallbacks
+        def handle_queue_loop():
+            try:
+                queue = self._get_drainining_queue(room_id)
+                for item in queue:
+                    try:
+                        ret = yield per_item_callback(item)
+                    except Exception:
+                        with PreserveLoggingContext():
+                            item.deferred.errback()
+                    else:
+                        with PreserveLoggingContext():
+                            item.deferred.callback(ret)
+            finally:
+                queue = self._event_persist_queues.pop(room_id, None)
+                if queue:
+                    self._event_persist_queues[room_id] = queue
+                self._currently_persisting_rooms.discard(room_id)
+
+        # set handle_queue_loop off in the background
+        run_as_background_process("persist_events", handle_queue_loop)
+
+    def _get_drainining_queue(self, room_id):
+        queue = self._event_persist_queues.setdefault(room_id, deque())
+
+        try:
+            while True:
+                yield queue.popleft()
+        except IndexError:
+            # Queue has been drained.
+            pass
+
+
+class EventsPersistenceStorage(object):
+    """High level interface for handling persisting newly received events.
+
+    Takes care of batching up events by room, and calculating the necessary
+    current state and forward extremity changes.
+    """
+
+    def __init__(self, hs, stores: DataStores):
+        # We ultimately want to split out the state store from the main store,
+        # so we use separate variables here even though they point to the same
+        # store for now.
+        self.main_store = stores.main
+        self.state_store = stores.main
+
+        self._clock = hs.get_clock()
+        self.is_mine_id = hs.is_mine_id
+        self._event_persist_queue = _EventPeristenceQueue()
+        self._state_resolution_handler = hs.get_state_resolution_handler()
+
+    @defer.inlineCallbacks
+    def persist_events(self, events_and_contexts, backfilled=False):
+        """
+        Write events to the database
+        Args:
+            events_and_contexts: list of tuples of (event, context)
+            backfilled (bool): Whether the results are retrieved from federation
+                via backfill or not. Used to determine if they're "new" events
+                which might update the current state etc.
+
+        Returns:
+            Deferred[int]: the stream ordering of the latest persisted event
+        """
+        partitioned = {}
+        for event, ctx in events_and_contexts:
+            partitioned.setdefault(event.room_id, []).append((event, ctx))
+
+        deferreds = []
+        for room_id, evs_ctxs in iteritems(partitioned):
+            d = self._event_persist_queue.add_to_queue(
+                room_id, evs_ctxs, backfilled=backfilled
+            )
+            deferreds.append(d)
+
+        for room_id in partitioned:
+            self._maybe_start_persisting(room_id)
+
+        yield make_deferred_yieldable(
+            defer.gatherResults(deferreds, consumeErrors=True)
+        )
+
+        max_persisted_id = yield self.main_store.get_current_events_token()
+
+        return max_persisted_id
+
+    @defer.inlineCallbacks
+    def persist_event(self, event, context, backfilled=False):
+        """
+
+        Args:
+            event (EventBase):
+            context (EventContext):
+            backfilled (bool):
+
+        Returns:
+            Deferred: resolves to (int, int): the stream ordering of ``event``,
+            and the stream ordering of the latest persisted event
+        """
+        deferred = self._event_persist_queue.add_to_queue(
+            event.room_id, [(event, context)], backfilled=backfilled
+        )
+
+        self._maybe_start_persisting(event.room_id)
+
+        yield make_deferred_yieldable(deferred)
+
+        max_persisted_id = yield self.main_store.get_current_events_token()
+        return (event.internal_metadata.stream_ordering, max_persisted_id)
+
+    def _maybe_start_persisting(self, room_id):
+        @defer.inlineCallbacks
+        def persisting_queue(item):
+            with Measure(self._clock, "persist_events"):
+                yield self._persist_events(
+                    item.events_and_contexts, backfilled=item.backfilled
+                )
+
+        self._event_persist_queue.handle_queue(room_id, persisting_queue)
+
+    @defer.inlineCallbacks
+    def _persist_events(
+        self, events_and_contexts, backfilled=False, delete_existing=False
+    ):
+        """Calculates the change to current state and forward extremities, and
+        persists the given events and with those updates.
+
+        Args:
+            events_and_contexts (list[(EventBase, EventContext)]):
+            backfilled (bool):
+            delete_existing (bool):
+
+        Returns:
+            Deferred: resolves when the events have been persisted
+        """
+        if not events_and_contexts:
+            return
+
+        chunks = [
+            events_and_contexts[x : x + 100]
+            for x in range(0, len(events_and_contexts), 100)
+        ]
+
+        for chunk in chunks:
+            # We can't easily parallelize these since different chunks
+            # might contain the same event. :(
+
+            # NB: Assumes that we are only persisting events for one room
+            # at a time.
+
+            # map room_id->list[event_ids] giving the new forward
+            # extremities in each room
+            new_forward_extremeties = {}
+
+            # map room_id->(type,state_key)->event_id tracking the full
+            # state in each room after adding these events.
+            # This is simply used to prefill the get_current_state_ids
+            # cache
+            current_state_for_room = {}
+
+            # map room_id->(to_delete, to_insert) where to_delete is a list
+            # of type/state keys to remove from current state, and to_insert
+            # is a map (type,key)->event_id giving the state delta in each
+            # room
+            state_delta_for_room = {}
+
+            if not backfilled:
+                with Measure(self._clock, "_calculate_state_and_extrem"):
+                    # Work out the new "current state" for each room.
+                    # We do this by working out what the new extremities are and then
+                    # calculating the state from that.
+                    events_by_room = {}
+                    for event, context in chunk:
+                        events_by_room.setdefault(event.room_id, []).append(
+                            (event, context)
+                        )
+
+                    for room_id, ev_ctx_rm in iteritems(events_by_room):
+                        latest_event_ids = yield self.main_store.get_latest_event_ids_in_room(
+                            room_id
+                        )
+                        new_latest_event_ids = yield self._calculate_new_extremities(
+                            room_id, ev_ctx_rm, latest_event_ids
+                        )
+
+                        latest_event_ids = set(latest_event_ids)
+                        if new_latest_event_ids == latest_event_ids:
+                            # No change in extremities, so no change in state
+                            continue
+
+                        # there should always be at least one forward extremity.
+                        # (except during the initial persistence of the send_join
+                        # results, in which case there will be no existing
+                        # extremities, so we'll `continue` above and skip this bit.)
+                        assert new_latest_event_ids, "No forward extremities left!"
+
+                        new_forward_extremeties[room_id] = new_latest_event_ids
+
+                        len_1 = (
+                            len(latest_event_ids) == 1
+                            and len(new_latest_event_ids) == 1
+                        )
+                        if len_1:
+                            all_single_prev_not_state = all(
+                                len(event.prev_event_ids()) == 1
+                                and not event.is_state()
+                                for event, ctx in ev_ctx_rm
+                            )
+                            # Don't bother calculating state if they're just
+                            # a long chain of single ancestor non-state events.
+                            if all_single_prev_not_state:
+                                continue
+
+                        state_delta_counter.inc()
+                        if len(new_latest_event_ids) == 1:
+                            state_delta_single_event_counter.inc()
+
+                            # This is a fairly handwavey check to see if we could
+                            # have guessed what the delta would have been when
+                            # processing one of these events.
+                            # What we're interested in is if the latest extremities
+                            # were the same when we created the event as they are
+                            # now. When this server creates a new event (as opposed
+                            # to receiving it over federation) it will use the
+                            # forward extremities as the prev_events, so we can
+                            # guess this by looking at the prev_events and checking
+                            # if they match the current forward extremities.
+                            for ev, _ in ev_ctx_rm:
+                                prev_event_ids = set(ev.prev_event_ids())
+                                if latest_event_ids == prev_event_ids:
+                                    state_delta_reuse_delta_counter.inc()
+                                    break
+
+                        logger.info("Calculating state delta for room %s", room_id)
+                        with Measure(
+                            self._clock, "persist_events.get_new_state_after_events"
+                        ):
+                            res = yield self._get_new_state_after_events(
+                                room_id,
+                                ev_ctx_rm,
+                                latest_event_ids,
+                                new_latest_event_ids,
+                            )
+                            current_state, delta_ids = res
+
+                        # If either are not None then there has been a change,
+                        # and we need to work out the delta (or use that
+                        # given)
+                        if delta_ids is not None:
+                            # If there is a delta we know that we've
+                            # only added or replaced state, never
+                            # removed keys entirely.
+                            state_delta_for_room[room_id] = ([], delta_ids)
+                        elif current_state is not None:
+                            with Measure(
+                                self._clock, "persist_events.calculate_state_delta"
+                            ):
+                                delta = yield self._calculate_state_delta(
+                                    room_id, current_state
+                                )
+                            state_delta_for_room[room_id] = delta
+
+                        # If we have the current_state then lets prefill
+                        # the cache with it.
+                        if current_state is not None:
+                            current_state_for_room[room_id] = current_state
+
+            yield self.main_store._persist_events_and_state_updates(
+                chunk,
+                current_state_for_room=current_state_for_room,
+                state_delta_for_room=state_delta_for_room,
+                new_forward_extremeties=new_forward_extremeties,
+                backfilled=backfilled,
+                delete_existing=delete_existing,
+            )
+
+    @defer.inlineCallbacks
+    def _calculate_new_extremities(self, room_id, event_contexts, latest_event_ids):
+        """Calculates the new forward extremities for a room given events to
+        persist.
+
+        Assumes that we are only persisting events for one room at a time.
+        """
+
+        # we're only interested in new events which aren't outliers and which aren't
+        # being rejected.
+        new_events = [
+            event
+            for event, ctx in event_contexts
+            if not event.internal_metadata.is_outlier()
+            and not ctx.rejected
+            and not event.internal_metadata.is_soft_failed()
+        ]
+
+        latest_event_ids = set(latest_event_ids)
+
+        # start with the existing forward extremities
+        result = set(latest_event_ids)
+
+        # add all the new events to the list
+        result.update(event.event_id for event in new_events)
+
+        # Now remove all events which are prev_events of any of the new events
+        result.difference_update(
+            e_id for event in new_events for e_id in event.prev_event_ids()
+        )
+
+        # Remove any events which are prev_events of any existing events.
+        existing_prevs = yield self.main_store._get_events_which_are_prevs(result)
+        result.difference_update(existing_prevs)
+
+        # Finally handle the case where the new events have soft-failed prev
+        # events. If they do we need to remove them and their prev events,
+        # otherwise we end up with dangling extremities.
+        existing_prevs = yield self.main_store._get_prevs_before_rejected(
+            e_id for event in new_events for e_id in event.prev_event_ids()
+        )
+        result.difference_update(existing_prevs)
+
+        # We only update metrics for events that change forward extremities
+        # (e.g. we ignore backfill/outliers/etc)
+        if result != latest_event_ids:
+            forward_extremities_counter.observe(len(result))
+            stale = latest_event_ids & result
+            stale_forward_extremities_counter.observe(len(stale))
+
+        return result
+
+    @defer.inlineCallbacks
+    def _get_new_state_after_events(
+        self, room_id, events_context, old_latest_event_ids, new_latest_event_ids
+    ):
+        """Calculate the current state dict after adding some new events to
+        a room
+
+        Args:
+            room_id (str):
+                room to which the events are being added. Used for logging etc
+
+            events_context (list[(EventBase, EventContext)]):
+                events and contexts which are being added to the room
+
+            old_latest_event_ids (iterable[str]):
+                the old forward extremities for the room.
+
+            new_latest_event_ids (iterable[str]):
+                the new forward extremities for the room.
+
+        Returns:
+            Deferred[tuple[dict[(str,str), str]|None, dict[(str,str), str]|None]]:
+            Returns a tuple of two state maps, the first being the full new current
+            state and the second being the delta to the existing current state.
+            If both are None then there has been no change.
+
+            If there has been a change then we only return the delta if its
+            already been calculated. Conversely if we do know the delta then
+            the new current state is only returned if we've already calculated
+            it.
+        """
+        # map from state_group to ((type, key) -> event_id) state map
+        state_groups_map = {}
+
+        # Map from (prev state group, new state group) -> delta state dict
+        state_group_deltas = {}
+
+        for ev, ctx in events_context:
+            if ctx.state_group is None:
+                # This should only happen for outlier events.
+                if not ev.internal_metadata.is_outlier():
+                    raise Exception(
+                        "Context for new event %s has no state "
+                        "group" % (ev.event_id,)
+                    )
+                continue
+
+            if ctx.state_group in state_groups_map:
+                continue
+
+            # We're only interested in pulling out state that has already
+            # been cached in the context. We'll pull stuff out of the DB later
+            # if necessary.
+            current_state_ids = ctx.get_cached_current_state_ids()
+            if current_state_ids is not None:
+                state_groups_map[ctx.state_group] = current_state_ids
+
+            if ctx.prev_group:
+                state_group_deltas[(ctx.prev_group, ctx.state_group)] = ctx.delta_ids
+
+        # We need to map the event_ids to their state groups. First, let's
+        # check if the event is one we're persisting, in which case we can
+        # pull the state group from its context.
+        # Otherwise we need to pull the state group from the database.
+
+        # Set of events we need to fetch groups for. (We know none of the old
+        # extremities are going to be in events_context).
+        missing_event_ids = set(old_latest_event_ids)
+
+        event_id_to_state_group = {}
+        for event_id in new_latest_event_ids:
+            # First search in the list of new events we're adding.
+            for ev, ctx in events_context:
+                if event_id == ev.event_id and ctx.state_group is not None:
+                    event_id_to_state_group[event_id] = ctx.state_group
+                    break
+            else:
+                # If we couldn't find it, then we'll need to pull
+                # the state from the database
+                missing_event_ids.add(event_id)
+
+        if missing_event_ids:
+            # Now pull out the state groups for any missing events from DB
+            event_to_groups = yield self.state_store._get_state_group_for_events(
+                missing_event_ids
+            )
+            event_id_to_state_group.update(event_to_groups)
+
+        # State groups of old_latest_event_ids
+        old_state_groups = set(
+            event_id_to_state_group[evid] for evid in old_latest_event_ids
+        )
+
+        # State groups of new_latest_event_ids
+        new_state_groups = set(
+            event_id_to_state_group[evid] for evid in new_latest_event_ids
+        )
+
+        # If they old and new groups are the same then we don't need to do
+        # anything.
+        if old_state_groups == new_state_groups:
+            return None, None
+
+        if len(new_state_groups) == 1 and len(old_state_groups) == 1:
+            # If we're going from one state group to another, lets check if
+            # we have a delta for that transition. If we do then we can just
+            # return that.
+
+            new_state_group = next(iter(new_state_groups))
+            old_state_group = next(iter(old_state_groups))
+
+            delta_ids = state_group_deltas.get((old_state_group, new_state_group), None)
+            if delta_ids is not None:
+                # We have a delta from the existing to new current state,
+                # so lets just return that. If we happen to already have
+                # the current state in memory then lets also return that,
+                # but it doesn't matter if we don't.
+                new_state = state_groups_map.get(new_state_group)
+                return new_state, delta_ids
+
+        # Now that we have calculated new_state_groups we need to get
+        # their state IDs so we can resolve to a single state set.
+        missing_state = new_state_groups - set(state_groups_map)
+        if missing_state:
+            group_to_state = yield self.state_store._get_state_for_groups(missing_state)
+            state_groups_map.update(group_to_state)
+
+        if len(new_state_groups) == 1:
+            # If there is only one state group, then we know what the current
+            # state is.
+            return state_groups_map[new_state_groups.pop()], None
+
+        # Ok, we need to defer to the state handler to resolve our state sets.
+
+        state_groups = {sg: state_groups_map[sg] for sg in new_state_groups}
+
+        events_map = {ev.event_id: ev for ev, _ in events_context}
+
+        # We need to get the room version, which is in the create event.
+        # Normally that'd be in the database, but its also possible that we're
+        # currently trying to persist it.
+        room_version = None
+        for ev, _ in events_context:
+            if ev.type == EventTypes.Create and ev.state_key == "":
+                room_version = ev.content.get("room_version", "1")
+                break
+
+        if not room_version:
+            room_version = yield self.main_store.get_room_version(room_id)
+
+        logger.debug("calling resolve_state_groups from preserve_events")
+        res = yield self._state_resolution_handler.resolve_state_groups(
+            room_id,
+            room_version,
+            state_groups,
+            events_map,
+            state_res_store=StateResolutionStore(self.main_store),
+        )
+
+        return res.state, None
+
+    @defer.inlineCallbacks
+    def _calculate_state_delta(self, room_id, current_state):
+        """Calculate the new state deltas for a room.
+
+        Assumes that we are only persisting events for one room at a time.
+
+        Returns:
+            tuple[list, dict] (to_delete, to_insert): where to_delete are the
+            type/state_keys to remove from current_state_events and `to_insert`
+            are the updates to current_state_events.
+        """
+        existing_state = yield self.main_store.get_current_state_ids(room_id)
+
+        to_delete = [key for key in existing_state if key not in current_state]
+
+        to_insert = {
+            key: ev_id
+            for key, ev_id in iteritems(current_state)
+            if ev_id != existing_state.get(key)
+        }
+
+        return to_delete, to_insert
diff --git a/tests/crypto/test_keyring.py b/tests/crypto/test_keyring.py
index c4f0bbd3dd..8efd39c7f7 100644
--- a/tests/crypto/test_keyring.py
+++ b/tests/crypto/test_keyring.py
@@ -178,7 +178,7 @@ class KeyringTestCase(unittest.HomeserverTestCase):
         kr = keyring.Keyring(self.hs)
 
         key1 = signedjson.key.generate_signing_key(1)
-        r = self.hs.datastore.store_server_verify_keys(
+        r = self.hs.get_datastore().store_server_verify_keys(
             "server9",
             time.time() * 1000,
             [("server9", get_key_id(key1), FetchKeyResult(get_verify_key(key1), 1000))],
@@ -209,7 +209,7 @@ class KeyringTestCase(unittest.HomeserverTestCase):
         )
 
         key1 = signedjson.key.generate_signing_key(1)
-        r = self.hs.datastore.store_server_verify_keys(
+        r = self.hs.get_datastore().store_server_verify_keys(
             "server9",
             time.time() * 1000,
             [("server9", get_key_id(key1), FetchKeyResult(get_verify_key(key1), None))],
diff --git a/tests/replication/slave/storage/_base.py b/tests/replication/slave/storage/_base.py
index 104349cdbd..4f924ce451 100644
--- a/tests/replication/slave/storage/_base.py
+++ b/tests/replication/slave/storage/_base.py
@@ -41,6 +41,7 @@ class BaseSlavedStoreTestCase(unittest.HomeserverTestCase):
     def prepare(self, reactor, clock, hs):
 
         self.master_store = self.hs.get_datastore()
+        self.storage = hs.get_storage()
         self.slaved_store = self.STORE_TYPE(self.hs.get_db_conn(), self.hs)
         self.event_id = 0
 
diff --git a/tests/replication/slave/storage/test_events.py b/tests/replication/slave/storage/test_events.py
index a368117b43..b68e9fe082 100644
--- a/tests/replication/slave/storage/test_events.py
+++ b/tests/replication/slave/storage/test_events.py
@@ -234,7 +234,9 @@ class SlavedEventStoreTestCase(BaseSlavedStoreTestCase):
             type="m.room.member", sender=USER_ID_2, key=USER_ID_2, membership="join"
         )
         msg, msgctx = self.build_event()
-        self.get_success(self.master_store.persist_events([(j2, j2ctx), (msg, msgctx)]))
+        self.get_success(
+            self.storage.persistence.persist_events([(j2, j2ctx), (msg, msgctx)])
+        )
         self.replicate()
 
         event_source = RoomEventSource(self.hs)
@@ -290,10 +292,12 @@ class SlavedEventStoreTestCase(BaseSlavedStoreTestCase):
 
         if backfill:
             self.get_success(
-                self.master_store.persist_events([(event, context)], backfilled=True)
+                self.storage.persistence.persist_events(
+                    [(event, context)], backfilled=True
+                )
             )
         else:
-            self.get_success(self.master_store.persist_event(event, context))
+            self.get_success(self.storage.persistence.persist_event(event, context))
 
         return event
 
diff --git a/tests/storage/test_redaction.py b/tests/storage/test_redaction.py
index 427d3c49c5..4561c3e383 100644
--- a/tests/storage/test_redaction.py
+++ b/tests/storage/test_redaction.py
@@ -39,6 +39,7 @@ class RedactionTestCase(unittest.HomeserverTestCase):
 
     def prepare(self, reactor, clock, hs):
         self.store = hs.get_datastore()
+        self.storage = hs.get_storage()
         self.event_builder_factory = hs.get_event_builder_factory()
         self.event_creation_handler = hs.get_event_creation_handler()
 
@@ -73,7 +74,7 @@ class RedactionTestCase(unittest.HomeserverTestCase):
             self.event_creation_handler.create_new_client_event(builder)
         )
 
-        self.get_success(self.store.persist_event(event, context))
+        self.get_success(self.storage.persistence.persist_event(event, context))
 
         return event
 
@@ -95,7 +96,7 @@ class RedactionTestCase(unittest.HomeserverTestCase):
             self.event_creation_handler.create_new_client_event(builder)
         )
 
-        self.get_success(self.store.persist_event(event, context))
+        self.get_success(self.storage.persistence.persist_event(event, context))
 
         return event
 
@@ -116,7 +117,7 @@ class RedactionTestCase(unittest.HomeserverTestCase):
             self.event_creation_handler.create_new_client_event(builder)
         )
 
-        self.get_success(self.store.persist_event(event, context))
+        self.get_success(self.storage.persistence.persist_event(event, context))
 
         return event
 
@@ -263,7 +264,7 @@ class RedactionTestCase(unittest.HomeserverTestCase):
             )
         )
 
-        self.get_success(self.store.persist_event(event_1, context_1))
+        self.get_success(self.storage.persistence.persist_event(event_1, context_1))
 
         event_2, context_2 = self.get_success(
             self.event_creation_handler.create_new_client_event(
@@ -282,7 +283,7 @@ class RedactionTestCase(unittest.HomeserverTestCase):
                 )
             )
         )
-        self.get_success(self.store.persist_event(event_2, context_2))
+        self.get_success(self.storage.persistence.persist_event(event_2, context_2))
 
         # fetch one of the redactions
         fetched = self.get_success(self.store.get_event(redaction_event_id1))
diff --git a/tests/storage/test_room.py b/tests/storage/test_room.py
index 1bee45706f..3ddaa151fe 100644
--- a/tests/storage/test_room.py
+++ b/tests/storage/test_room.py
@@ -62,6 +62,7 @@ class RoomEventsStoreTestCase(unittest.TestCase):
         # Room events need the full datastore, for persist_event() and
         # get_room_state()
         self.store = hs.get_datastore()
+        self.storage = hs.get_storage()
         self.event_factory = hs.get_event_factory()
 
         self.room = RoomID.from_string("!abcde:test")
@@ -72,7 +73,7 @@ class RoomEventsStoreTestCase(unittest.TestCase):
 
     @defer.inlineCallbacks
     def inject_room_event(self, **kwargs):
-        yield self.store.persist_event(
+        yield self.storage.persistence.persist_event(
             self.event_factory.create_event(room_id=self.room.to_string(), **kwargs)
         )
 
diff --git a/tests/storage/test_roommember.py b/tests/storage/test_roommember.py
index 447a3c6ffb..9ddd17f73d 100644
--- a/tests/storage/test_roommember.py
+++ b/tests/storage/test_roommember.py
@@ -44,6 +44,7 @@ class RoomMemberStoreTestCase(unittest.HomeserverTestCase):
         # We can't test the RoomMemberStore on its own without the other event
         # storage logic
         self.store = hs.get_datastore()
+        self.storage = hs.get_storage()
         self.event_builder_factory = hs.get_event_builder_factory()
         self.event_creation_handler = hs.get_event_creation_handler()
 
@@ -70,7 +71,7 @@ class RoomMemberStoreTestCase(unittest.HomeserverTestCase):
             self.event_creation_handler.create_new_client_event(builder)
         )
 
-        self.get_success(self.store.persist_event(event, context))
+        self.get_success(self.storage.persistence.persist_event(event, context))
 
         return event
 
diff --git a/tests/storage/test_state.py b/tests/storage/test_state.py
index 5c2cf3c2db..d573a3e07b 100644
--- a/tests/storage/test_state.py
+++ b/tests/storage/test_state.py
@@ -34,6 +34,7 @@ class StateStoreTestCase(tests.unittest.TestCase):
         hs = yield tests.utils.setup_test_homeserver(self.addCleanup)
 
         self.store = hs.get_datastore()
+        self.storage = hs.get_storage()
         self.event_builder_factory = hs.get_event_builder_factory()
         self.event_creation_handler = hs.get_event_creation_handler()
 
@@ -63,7 +64,7 @@ class StateStoreTestCase(tests.unittest.TestCase):
             builder
         )
 
-        yield self.store.persist_event(event, context)
+        yield self.storage.persistence.persist_event(event, context)
 
         return event
 
diff --git a/tests/test_federation.py b/tests/test_federation.py
index a73f18f88e..d1acb16f30 100644
--- a/tests/test_federation.py
+++ b/tests/test_federation.py
@@ -36,7 +36,8 @@ class MessageAcceptTests(unittest.TestCase):
         # Figure out what the most recent event is
         most_recent = self.successResultOf(
             maybeDeferred(
-                self.homeserver.datastore.get_latest_event_ids_in_room, self.room_id
+                self.homeserver.get_datastore().get_latest_event_ids_in_room,
+                self.room_id,
             )
         )[0]
 
@@ -75,7 +76,8 @@ class MessageAcceptTests(unittest.TestCase):
         self.assertEqual(
             self.successResultOf(
                 maybeDeferred(
-                    self.homeserver.datastore.get_latest_event_ids_in_room, self.room_id
+                    self.homeserver.get_datastore().get_latest_event_ids_in_room,
+                    self.room_id,
                 )
             )[0],
             "$join:test.serv",
@@ -97,7 +99,8 @@ class MessageAcceptTests(unittest.TestCase):
         # Figure out what the most recent event is
         most_recent = self.successResultOf(
             maybeDeferred(
-                self.homeserver.datastore.get_latest_event_ids_in_room, self.room_id
+                self.homeserver.get_datastore().get_latest_event_ids_in_room,
+                self.room_id,
             )
         )[0]
 
@@ -137,6 +140,6 @@ class MessageAcceptTests(unittest.TestCase):
 
         # Make sure the invalid event isn't there
         extrem = maybeDeferred(
-            self.homeserver.datastore.get_latest_event_ids_in_room, self.room_id
+            self.homeserver.get_datastore().get_latest_event_ids_in_room, self.room_id
         )
         self.assertEqual(self.successResultOf(extrem)[0], "$join:test.serv")
diff --git a/tests/test_visibility.py b/tests/test_visibility.py
index 18f1a0035d..6ae1ea9b04 100644
--- a/tests/test_visibility.py
+++ b/tests/test_visibility.py
@@ -36,6 +36,7 @@ class FilterEventsForServerTestCase(tests.unittest.TestCase):
         self.event_creation_handler = self.hs.get_event_creation_handler()
         self.event_builder_factory = self.hs.get_event_builder_factory()
         self.store = self.hs.get_datastore()
+        self.storage = self.hs.get_storage()
 
         yield create_room(self.hs, TEST_ROOM_ID, "@someone:ROOM")
 
@@ -137,7 +138,7 @@ class FilterEventsForServerTestCase(tests.unittest.TestCase):
         event, context = yield self.event_creation_handler.create_new_client_event(
             builder
         )
-        yield self.hs.get_datastore().persist_event(event, context)
+        yield self.storage.persistence.persist_event(event, context)
         return event
 
     @defer.inlineCallbacks
@@ -159,7 +160,7 @@ class FilterEventsForServerTestCase(tests.unittest.TestCase):
             builder
         )
 
-        yield self.hs.get_datastore().persist_event(event, context)
+        yield self.storage.persistence.persist_event(event, context)
         return event
 
     @defer.inlineCallbacks
@@ -180,7 +181,7 @@ class FilterEventsForServerTestCase(tests.unittest.TestCase):
             builder
         )
 
-        yield self.hs.get_datastore().persist_event(event, context)
+        yield self.storage.persistence.persist_event(event, context)
         return event
 
     @defer.inlineCallbacks
diff --git a/tests/utils.py b/tests/utils.py
index 8cced4b7e8..7dc9bdc505 100644
--- a/tests/utils.py
+++ b/tests/utils.py
@@ -325,10 +325,16 @@ def setup_test_homeserver(
         if homeserverToUse.__name__ == "TestHomeServer":
             hs.setup_master()
     else:
+        # If we have been given an explicit datastore we probably want to mock
+        # out the DataStores somehow too. This all feels a bit wrong, but then
+        # mocking the stores feels wrong too.
+        datastores = Mock(datastore=datastore)
+
         hs = homeserverToUse(
             name,
             db_pool=None,
             datastore=datastore,
+            datastores=datastores,
             config=config,
             version_string="Synapse/tests",
             database_engine=db_engine,
@@ -646,7 +652,7 @@ def create_room(hs, room_id, creator_id):
         creator_id (str)
     """
 
-    store = hs.get_datastore()
+    persistence_store = hs.get_storage().persistence
     event_builder_factory = hs.get_event_builder_factory()
     event_creation_handler = hs.get_event_creation_handler()
 
@@ -663,4 +669,4 @@ def create_room(hs, room_id, creator_id):
 
     event, context = yield event_creation_handler.create_new_client_event(builder)
 
-    yield store.persist_event(event, context)
+    yield persistence_store.persist_event(event, context)