Restart of docker does not correctly restart kafka

[blawlor]

Hi,
For reasons I don't yet understand, when I do a docker restart container_id, I can no longer send messages to the topics previously created (the kafka client gets the following: org.apache.kafka.common.errors.TimeoutException: Failed to update metadata after 60000 ms.).

If I connect to the running container and just stop the kafka process directly, supervisord successfully restarts it and I don't have any issues, so I'm assuming this probably is docker-specific.

[blawlor]

Tried with the latest version of kafka and it didn't help.

[tobr]

Same Problem here using boot2docker and docker-machine. Using 'docker run' everything works fine, but when I stop the container and start it again with 'docker start' the queue does not receive the messages and I get the timeout metadata error.

[amarcionek]

This is a known issue with Kafka or Zookeeper. See:

https://issues.apache.org/jira/browse/KAFKA-1387
https://issues.apache.org/jira/browse/KAFKA-1451
https://issues.apache.org/jira/browse/KAFKA-1494

Even though some of them say fixed, there is clearly an issue as we have it as well.

[Sreekiranar]

I also had the same issue and i solved it by adding a --net=somename to the command.
Here is my run command
" docker run --net=host -p 2181:2181 -p 9092:9092 --env ADVERTISED_HOST= --env ADVERTISED_PORT=9092 --env LOG_RETENTION_HOURS=1 --env LOG_RETENTION_BYTES=250000000 spotify/kafka "
Please try it.