diff --git a/changelog.d/4792.bugfix b/changelog.d/4792.bugfix new file mode 100644 index 0000000000..b127b6254f --- /dev/null +++ b/changelog.d/4792.bugfix @@ -0,0 +1 @@ +Handle batch updates in worker replication protocol. \ No newline at end of file diff --git a/synapse/replication/tcp/protocol.py b/synapse/replication/tcp/protocol.py index 49ae5b3355..55630ba9a7 100644 --- a/synapse/replication/tcp/protocol.py +++ b/synapse/replication/tcp/protocol.py @@ -451,7 +451,7 @@ class ServerReplicationStreamProtocol(BaseReplicationStreamProtocol): @defer.inlineCallbacks def subscribe_to_stream(self, stream_name, token): - """Subscribe the remote to a streams. + """Subscribe the remote to a stream. This invloves checking if they've missed anything and sending those updates down if they have. During that time new updates for the stream @@ -478,11 +478,36 @@ class ServerReplicationStreamProtocol(BaseReplicationStreamProtocol): # Now we can send any updates that came in while we were subscribing pending_rdata = self.pending_rdata.pop(stream_name, []) + updates = [] for token, update in pending_rdata: - # Only send updates newer than the current token - if token > current_token: + # If the token is null, it is part of a batch update. Batches + # are multiple updates that share a single token. To denote + # this, the token is set to None for all tokens in the batch + # except for the last. If we find a None token, we keep looking + # through tokens until we find one that is not None and then + # process all previous updates in the batch as if they had the + # final token. + if token is None: + # Store this update as part of a batch + updates.append(update) + continue + + if token <= current_token: + # This update or batch of updates is older than + # current_token, dismiss it + updates = [] + continue + + updates.append(update) + + # Send all updates that are part of this batch with the + # found token + for update in updates: self.send_command(RdataCommand(stream_name, token, update)) + # Clear stored updates + updates = [] + # They're now fully subscribed self.replication_streams.add(stream_name) except Exception as e: