Store state groups separately from events (#2784)

* Split state group persist into seperate storage func

* Add per database engine code for state group id gen

* Move store_state_group to StateReadStore

This allows other workers to use it, and so resolve state.

* Hook up store_state_group

* Fix tests

* Rename _store_mult_state_groups_txn

* Rename StateGroupReadStore

* Remove redundant _have_persisted_state_group_txn

* Update comments

* Comment compute_event_context

* Set start val for state_group_id_seq

... otherwise we try to recreate old state groups

* Update comments

* Don't store state for outliers

* Update comment

* Update docstring as state groups are ints

synapse/storage/engines/sqlite3.py

@@ -16,6 +16,7 @@
 from synapse.storage.prepare_database import prepare_database
 
 import struct
+import threading
 
 
 class Sqlite3Engine(object):
@@ -24,6 +25,11 @@ class Sqlite3Engine(object):
     def __init__(self, database_module, database_config):
         self.module = database_module
 
+        # The current max state_group, or None if we haven't looked
+        # in the DB yet.
+        self._current_state_group_id = None
+        self._current_state_group_id_lock = threading.Lock()
+
     def check_database(self, txn):
         pass
 
@@ -43,6 +49,19 @@ class Sqlite3Engine(object):
     def lock_table(self, txn, table):
         return
 
+    def get_next_state_group_id(self, txn):
+        """Returns an int that can be used as a new state_group ID
+        """
+        # We do application locking here since if we're using sqlite then
+        # we are a single process synapse.
+        with self._current_state_group_id_lock:
+            if self._current_state_group_id is None:
+                txn.execute("SELECT COALESCE(max(id), 0) FROM state_groups")
+                self._current_state_group_id = txn.fetchone()[0]
+
+            self._current_state_group_id += 1
+            return self._current_state_group_id
+
 
 # Following functions taken from: https://github.com/coleifer/peewee