Hi Harsha,

I've now created a new JIRA https://issues.apache.org/jira/browse/KAFKA-1878 which explains the issue in a bit more detail (sorry, I initially thought this change could go linked with that existing JIRA, but you are right it needed a different one). The JIRA includes the exception which is causing the test to fail. I've opened a review request with a proposed fix https://reviews.apache.org/r/30026/diff/

-Jaikiran
On Sunday 18 January 2015 09:37 PM, Harsha wrote:
Jaikiran,
        I can't reproduce the failure of the ProdcuerFailureHandlingTest.
        I ran the single test .  you probably are seeing some errors
        written to console when you use ./gradlew -i -Dsingle.test .
        These errors are expected in some unit tests as some of these
        test failure cases.
If you can reproduce this or even intermittent test failure can you
please open up a new JIRA and attach your patch there.
Your review patch is attached KAFKA-1867 which is a different issue.
Thanks,
Harsha

On Sun, Jan 18, 2015, at 07:16 AM, Jaikiran Pai wrote:
I could reproduce this consistently when that test *method* is run
individually. From what I could gather, the __consumer_offset topic
(being accessed in that test) had 50 partitions (default) which took a
while for each of them to be assigned a leader and do other
initialization and that timed out the metadata update wait during the
producer.send. I increased the metadata fetch timeout specifically for
that producer in that test method and was able to get past this. I've
sent a patch here https://reviews.apache.org/r/30013/


-Jaikiran

On Sunday 18 January 2015 12:30 AM, Manikumar Reddy wrote:
   I am consistently getting these errors. May be transient errors.

On Sun, Jan 18, 2015 at 12:05 AM, Harsha <ka...@harsha.io> wrote:

I don't see any failures in tests with the latest trunk or 0.8.2. I ran
it few times in a loop.
-Harsha

On Sat, Jan 17, 2015, at 08:38 AM, Manikumar Reddy wrote:
ProducerFailureHandlingTest.testCannotSendToInternalTopic is failing on
both 0.8.2 and trunk.

Error on 0.8.2:
kafka.api.ProducerFailureHandlingTest > testCannotSendToInternalTopic
FAILED
      java.util.concurrent.ExecutionException:
org.apache.kafka.common.errors.TimeoutException: Failed to update
metadata
after 3000 ms.
          at

org.apache.kafka.clients.producer.KafkaProducer$FutureFailure.<init>(KafkaProducer.java:437)
          at

org.apache.kafka.clients.producer.KafkaProducer.send(KafkaProducer.java:352)
          at

org.apache.kafka.clients.producer.KafkaProducer.send(KafkaProducer.java:248)
          at

kafka.api.ProducerFailureHandlingTest.testCannotSendToInternalTopic(ProducerFailureHandlingTest.scala:309)
          Caused by:
          org.apache.kafka.common.errors.TimeoutException: Failed to update
metadata after 3000 ms.


Error on Trunk:
kafka.api.test.ProducerFailureHandlingTest >
testCannotSendToInternalTopic
FAILED
      java.lang.AssertionError: null
          at org.junit.Assert.fail(Assert.java:69)
          at org.junit.Assert.assertTrue(Assert.java:32)
          at org.junit.Assert.assertTrue(Assert.java:41)
          at

kafka.api.test.ProducerFailureHandlingTest.testCannotSendToInternalTopic(ProducerFailureHandlingTest.scala:312)


Reply via email to