swh_storage_backend_config = {'cls': 'local', 'db': 'postgresql://postgres@127.0.0.1:10152/tests', 'journal_writer': {'brokers': ['127.0.0.1:41033'], 'client_id': 'kafka_writer-1', 'cls': 'kafka', 'prefix': 'pkppvgbfmd-1'}, 'objstorage': {'args': {}, 'cls': 'memory'}} kafka_prefix = 'pkppvgbfmd', kafka_consumer_group = 'test-consumer-pkppvgbfmd' kafka_server = '127.0.0.1:41033' @patch("swh.storage.backfill.RANGE_GENERATORS", RANGE_GENERATORS) def test_backfiller( swh_storage_backend_config, kafka_prefix: str, kafka_consumer_group: str, kafka_server: str, ): prefix1 = f"{kafka_prefix}-1" prefix2 = f"{kafka_prefix}-2" journal1 = { "cls": "kafka", "brokers": [kafka_server], "client_id": "kafka_writer-1", "prefix": prefix1, } swh_storage_backend_config["journal_writer"] = journal1 storage = get_storage(**swh_storage_backend_config) # fill the storage and the journal (under prefix1) for object_type, objects in TEST_OBJECTS.items(): method = getattr(storage, object_type + "_add") method(objects) # now apply the backfiller on the storage to fill the journal under prefix2 backfiller_config = { "brokers": [kafka_server], "client_id": "kafka_writer-2", "prefix": prefix2, "storage_dbconn": swh_storage_backend_config["db"], } # Backfilling backfiller = JournalBackfiller(backfiller_config) for object_type in TEST_OBJECTS: backfiller.run(object_type, None, None) # now check journal content are the same under both topics # use the replayer scaffolding to fill storages to make is a bit easier # Replaying #1 sto1 = get_storage(cls="memory") replayer1 = JournalClient( brokers=kafka_server, group_id=f"{kafka_consumer_group}-1", prefix=prefix1, stop_on_eof=True, ) worker_fn1 = functools.partial(process_replay_objects, storage=sto1) > replayer1.process(worker_fn1) .tox/py3/lib/python3.7/site-packages/swh/storage/tests/test_backfill.py:230: _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ .tox/py3/lib/python3.7/site-packages/swh/journal/client.py:264: in process batch_processed, at_eof = self.handle_messages(messages, worker_fn) .tox/py3/lib/python3.7/site-packages/swh/journal/client.py:291: in handle_messages worker_fn(dict(objects)) .tox/py3/lib/python3.7/site-packages/swh/storage/replay.py:61: in process_replay_objects _insert_objects(object_type, objects, storage) .tox/py3/lib/python3.7/site-packages/swh/storage/replay.py:139: in _insert_objects method(object_converter_fn[object_type](o) for o in objects) _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ self = <swh.storage.in_memory.InMemoryStorage object at 0x7fd1d74a2390> origins = <generator object _insert_objects.<locals>.<genexpr> at 0x7fd1d65114f8> def origin_add(self, origins: List[Origin]) -> Dict[str, int]: added = 0 for origin in origins: if origin.url not in self._origins: self.origin_add_one(origin) added += 1 > self._cql_runner.increment_counter("origin", len(origins)) E TypeError: object of type 'generator' has no len() .tox/py3/lib/python3.7/site-packages/swh/storage/in_memory.py:659: TypeError