[
https://issues.apache.org/jira/browse/KAFKA-4673?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Ewen Cheslack-Postava resolved KAFKA-4673.
------------------------------------------
Resolution: Fixed
Fix Version/s: 0.10.2.0
0.10.1.2
0.10.3.0
Issue resolved by pull request 2427
[https://github.com/apache/kafka/pull/2427]
> Python VerifiableConsumer service has thread-safety bug for event_handlers
> --------------------------------------------------------------------------
>
> Key: KAFKA-4673
> URL: https://issues.apache.org/jira/browse/KAFKA-4673
> Project: Kafka
> Issue Type: Bug
> Components: system tests
> Affects Versions: 0.10.1.1
> Reporter: Ewen Cheslack-Postava
> Assignee: Ewen Cheslack-Postava
> Fix For: 0.10.3.0, 0.10.1.2, 0.10.2.0
>
> Original Estimate: 1h
> Remaining Estimate: 1h
>
> From
> http://confluent-kafka-0-10-1-system-test-results.s3-us-west-2.amazonaws.com/2017-01-17--001.1484653357--apache--0.10.1--d436fa2/report.html
> {quote}
> ====================================================================================================
> test_id:
> 2017-01-17--001.kafkatest.tests.client.consumer_test.OffsetValidationTest.test_consumer_bounce.clean_shutdown=True.bounce_mode=rolling
> status: FAIL
> run time: 1 minute 42.663 seconds
> dictionary changed size during iteration
> Traceback (most recent call last):
> File
> "/var/lib/jenkins/workspace/system-test-kafka-0.10.1/kafka/venv/local/lib/python2.7/site-packages/ducktape-0.5.3-py2.7.egg/ducktape/tests/runner.py",
> line 106, in run_all_tests
> data = self.run_single_test()
> File
> "/var/lib/jenkins/workspace/system-test-kafka-0.10.1/kafka/venv/local/lib/python2.7/site-packages/ducktape-0.5.3-py2.7.egg/ducktape/tests/runner.py",
> line 162, in run_single_test
> return self.current_test_context.function(self.current_test)
> File
> "/var/lib/jenkins/workspace/system-test-kafka-0.10.1/kafka/venv/local/lib/python2.7/site-packages/ducktape-0.5.3-py2.7.egg/ducktape/mark/_mark.py",
> line 331, in wrapper
> return functools.partial(f, *args, **kwargs)(*w_args, **w_kwargs)
> File
> "/var/lib/jenkins/workspace/system-test-kafka-0.10.1/kafka/tests/kafkatest/tests/client/consumer_test.py",
> line 137, in test_consumer_bounce
> self.await_all_members(consumer)
> File
> "/var/lib/jenkins/workspace/system-test-kafka-0.10.1/kafka/tests/kafkatest/tests/verifiable_consumer_test.py",
> line 83, in await_all_members
> self.await_members(consumer, self.num_consumers)
> File
> "/var/lib/jenkins/workspace/system-test-kafka-0.10.1/kafka/tests/kafkatest/tests/verifiable_consumer_test.py",
> line 80, in await_members
> err_msg="Consumers failed to join in a reasonable amount of time")
> File
> "/var/lib/jenkins/workspace/system-test-kafka-0.10.1/kafka/venv/local/lib/python2.7/site-packages/ducktape-0.5.3-py2.7.egg/ducktape/utils/util.py",
> line 31, in wait_until
> if condition():
> File
> "/var/lib/jenkins/workspace/system-test-kafka-0.10.1/kafka/tests/kafkatest/tests/verifiable_consumer_test.py",
> line 78, in <lambda>
> wait_until(lambda: len(consumer.joined_nodes()) == num_consumers,
> File
> "/var/lib/jenkins/workspace/system-test-kafka-0.10.1/kafka/tests/kafkatest/services/verifiable_consumer.py",
> line 317, in joined_nodes
> return [handler.node for handler in self.event_handlers.itervalues()
> RuntimeError: dictionary changed size during iteration
> {quote}
> It looks like the background thread is incorrectly inserting elements into
> self.event_handlers during this iteration. This is the first time I've seen
> this, so I suspect it's rarely hit. Looking at the code, most access is
> protected by a lock except for the additions at the beginning of the _worker
> method. Looking through the stacktrace, this looks like the likely culprit
> since we start the consumer and immediately start calling
> `await_all_members`, which ultimately iterates over that list. If that call
> happens faster than the thread gets started we could hit a bad interleaving.
> Fix is probably easy -- just add the proper locking around use of that object.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
