Do not delete ZK offset when auto_offset_reset is set to smallest about jruby-kafka

[joekiller]

Email replies seem delayed so please pardon a double post...

This implementation follows how Kafka does it in their code as well. A zookeeper znode can only be deleted if no one else has it opened so it gracefully fails if another consumer is registered.

See here: https://github.com/apache/kafka/blob/0.8.1/core/src/main/scala/kafka/consumer/ConsoleConsumer.scala#L181

[joekiller]

This implementation follows how Kafka does it in their code as well. A
zookeeper znode can only be deleted if no one else has it opened so it
gracefully fails if another consumer is registered.

See here:
https://github.com/apache/kafka/blob/0.8.1/core/src/main/scala/kafka/consumer/ConsoleConsumer.scala#L181

We should be consistent with Kafka consumer behavior defined in
http://kafka.apache.org/documentation.html#consumerconfigs

when auto.offset.reset = smallest.

The documentation mentions setting is applied only when ZK path is
missing or when an offset requested is not found in ZK. In contrary, we
delete the offset when this setting is applied which is dangerous.

See
https://github.com/joekiller/jruby-kafka/blob/master/lib/jruby-kafka/group.rb#L161

Happy to provide a patch for this

—
Reply to this email directly or view it on GitHub
#7.

[suyograo]

@joekiller Thanks for the clarification.