Search code examples
apache-kafkamesos

Kafka consumer Failed to find leader for Set and Fetching topic metadata with correlation id 4 for topics failed

So I'm getting this error with a single kafka 0.8.2.0 broker and a java consumer running in mesos. I have to say it used to work, and AFAIK nothing has changed.

Here's the error when the consumer runs in mesos:

16/03/17 05:44:23 WARN ConsumerFetcherManager$LeaderFinderThread: [alertConsumer_secasprddb01-1-1458193432645-387abff0-leader-finder-thread], Failed to find leader for Set([events,1], [events,14], [events,2], [events,8], [events,7], [events,10], [events,3], [events,11], [events,5], [events,0], [events,12], [events,13], [events,6], [events,4], [events,9]) kafka.common.KafkaException: fetching topic metadata for topics [Set(events)] from broker [ArrayBuffer(id:0,host:prodMessage01-1.blah.com,port:9092)] failed at kafka.client.ClientUtils$.fetchTopicMetadata(ClientUtils.scala:72)
at kafka.client.ClientUtils$.fetchTopicMetadata(ClientUtils.scala:93)
at kafka.consumer.ConsumerFetcherManager$LeaderFinderThread.doWork(ConsumerFetcherManager.scala:66)
at kafka.utils.ShutdownableThread.run(ShutdownableThread.scala:60) Caused by: java.net.SocketTimeoutException

I even went as far as completely deleting the topic and recreating it form scratch and still the same error.

If I run the script kafka-console-consumer.sh on the same box as the broker, I'm still getting the error:

Failed to find leader for Set([events,1], [events,14], [events,2], [events,8], [events,7], [events,10], [events,3], [events,11], [events,5], [events,0], [events,12], [events,13], [events,6], [events,9], [events,4]) (kafka.consumer.ConsumerFetcherManager$LeaderFinderThread) kafka.common.KafkaException: fetching topic metadata for topics [Set(events)] from broker [ArrayBuffer(id:0,host:prodMessage01-1.blah.com,port:9092)] failed
    at kafka.client.ClientUtils$.fetchTopicMetadata(ClientUtils.scala:72)
    at kafka.client.ClientUtils$.fetchTopicMetadata(ClientUtils.scala:93)
    at kafka.consumer.ConsumerFetcherManager$LeaderFinderThread.doWork(ConsumerFetcherManager.scala:66)
    at kafka.utils.ShutdownableThread.run(ShutdownableThread.scala:60) Caused by: java.net.SocketTimeoutException

So even when everyting is local to the broker, it still is failing. How is this even possible?

I don't see how I can get past this if I've actually deleted and recreated the topic.

If I do a topic describe (kafka-topics.sh -- describe), it works fine: Topic:events PartitionCount:15 ReplicationFactor:1 Configs: Topic: events Partition: 0 Leader: 0 Replicas: 0 Isr: 0 Topic: events Partition: 1 Leader: 0 Replicas: 0 Isr: 0 Topic: events Partition: 2 Leader: 0 Replicas: 0 Isr: 0 Topic: events Partition: 3 Leader: 0 Replicas: 0 Isr: 0 Topic: events Partition: 4 Leader: 0 Replicas: 0 Isr: 0 Topic: events Partition: 5 Leader: 0 Replicas: 0 Isr: 0 Topic: events Partition: 6 Leader: 0 Replicas: 0 Isr: 0 Topic: events Partition: 7 Leader: 0 Replicas: 0 Isr: 0 Topic: events Partition: 8 Leader: 0 Replicas: 0 Isr: 0 Topic: events Partition: 9 Leader: 0 Replicas: 0 Isr: 0 Topic: events Partition: 10 Leader: 0 Replicas: 0 Isr: 0 Topic: events Partition: 11 Leader: 0 Replicas: 0 Isr: 0 Topic: events Partition: 12 Leader: 0 Replicas: 0 Isr: 0 Topic: events Partition: 13 Leader: 0 Replicas: 0 Isr: 0 Topic: events Partition: 14 Leader: 0 Replicas: 0 Isr: 0

Solution

  • See the above comments on how I resolved the issue. But it's not desirable for a production environment.