Fix bug in account data replication stream. (#7656)

* Ensure account data stream IDs are unique.

The account data stream is shared between three tables, and the maximum
allocated ID was tracked in a dedicated table. Updating the max ID
happened outside the transaction that allocated the ID, leading to a
race where if the server was restarted then the same ID could be
allocated but the max ID failed to be updated, leading it to be reused.

The ID generators have support for tracking across multiple tables, so
we may as well use that instead of a dedicated table.

* Fix bug in account data replication stream.

If the same stream ID was used in both global and room account data then
the getting updates for the replication stream would fail due to
`heapq.merge(..)` trying to compare a `str` with a `None`. (This is
because you'd have two rows like `(534, '!room')` and `(534, None)` from
the room and global account data tables).

Fix is just to order by stream ID, since we don't rely on the ordering
beyond that. The bug where stream IDs can be reused should be fixed now,
so this case shouldn't happen going forward.

Fixes #7617

3 files changed, 19 insertions, 1 deletions

diff --git a/synapse/storage/data_stores/main/account_data.py b/synapse/storage/data_stores/main/account_data.py
index f9eef1b78e..b58f04d00d 100644
--- a/synapse/storage/data_stores/main/account_data.py
+++ b/synapse/storage/data_stores/main/account_data.py
@@ -297,7 +297,13 @@ class AccountDataWorkerStore(SQLBaseStore):
 class AccountDataStore(AccountDataWorkerStore):
     def __init__(self, database: Database, db_conn, hs):
         self._account_data_id_gen = StreamIdGenerator(
-            db_conn, "account_data_max_stream_id", "stream_id"
+            db_conn,
+            "account_data_max_stream_id",
+            "stream_id",
+            extra_tables=[
+                ("room_account_data", "stream_id"),
+                ("room_tags_revisions", "stream_id"),
+            ],
         )
 
         super(AccountDataStore, self).__init__(database, db_conn, hs)
@@ -387,6 +393,10 @@ class AccountDataStore(AccountDataWorkerStore):
             # doesn't sound any worse than the whole update getting lost,
             # which is what would happen if we combined the two into one
             # transaction.
+            #
+            # Note: This is only here for backwards compat to allow admins to
+            # roll back to a previous Synapse version. Next time we update the
+            # database version we can remove this table.
             yield self._update_max_stream_id(next_id)
 
             self._account_data_stream_cache.entity_has_changed(user_id, next_id)
@@ -405,6 +415,10 @@ class AccountDataStore(AccountDataWorkerStore):
             next_id(int): The the revision to advance to.
         """
 
+        # Note: This is only here for backwards compat to allow admins to
+        # roll back to a previous Synapse version. Next time we update the
+        # database version we can remove this table.
+
         def _update(txn):
             update_max_id_sql = (
                 "UPDATE account_data_max_stream_id"
diff --git a/synapse/storage/data_stores/main/tags.py b/synapse/storage/data_stores/main/tags.py
index 2aa1bafd48..4219018302 100644
--- a/synapse/storage/data_stores/main/tags.py
+++ b/synapse/storage/data_stores/main/tags.py
@@ -233,6 +233,9 @@ class TagsStore(TagsWorkerStore):
             self._account_data_stream_cache.entity_has_changed, user_id, next_id
         )
 
+        # Note: This is only here for backwards compat to allow admins to
+        # roll back to a previous Synapse version. Next time we update the
+        # database version we can remove this table.
         update_max_id_sql = (
             "UPDATE account_data_max_stream_id"
             " SET stream_id = ?"
diff --git a/synapse/storage/prepare_database.py b/synapse/storage/prepare_database.py
index b95434f031..9cc3b51fe6 100644
--- a/synapse/storage/prepare_database.py
+++ b/synapse/storage/prepare_database.py
@@ -33,6 +33,7 @@ logger = logging.getLogger(__name__)
 # schema files, so the users will be informed on server restarts.
 # XXX: If you're about to bump this to 59 (or higher) please create an update
 # that drops the unused `cache_invalidation_stream` table, as per #7436!
+# XXX: Also add an update to drop `account_data_max_stream_id` as per #7656!
 SCHEMA_VERSION = 58
 
 dir_path = os.path.abspath(os.path.dirname(__file__))