[
https://issues.apache.org/jira/browse/KAFKA-49?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13155286#comment-13155286
]
Jay Kreps commented on KAFKA-49:
--------------------------------
+1 for returning the offset
If we are changing the request format it would also be good to think it through
in some detail to get it right since these kinds of API changes are harder to
rollout then to make. Some questions:
1. Are we trying to maintain compatibility for this change? If so we should
bump up the request id number and ignore the new fields for the old request id.
This is not too hard, but requires a little extra work.
2. Currently we have ProduceRequest and MultiProducerRequest and FetchRequest
and MultiFetchRequest. I recommend we rename MultiProducerRequest to
ProduceRequest and delete the existing ProduceRequest. The current
ProduceRequest has no advantages over MultiProducerRequest and having both
means each change we make has to be done for both. The two variations are just
there for historical reasons--originally there was no multi-* version of the
requests and we added that later. I recommend we do the same for the
FetchRequest/MultiFetchRequest. This will make our lives simpler going forward.
3. Both of the MultiProducerRequest has the format [(topic, partition,
messages), (topic, partition, messages), ...]. This is because it is just a
bunch of repeated ProducerRequests. This is really inefficient, though, as a
common case is that we are producing a bunch of messages for different
partitions under the same topic (i.e. if we are doing the key-based
partitioning). It would be better for the format to be [(topic, [(partition,
messages), ...], topic, [(partition, messages), ...], ...]. This would mean
that each topic is only given once.
I would like to get a quick consensus on the desired format of the produce and
fetch requests up front, then we can break this into appropriate sub tasks so
we don't expand the scope of Prasanth's work too much.
> Add acknowledgement to the produce request.
> -------------------------------------------
>
> Key: KAFKA-49
> URL: https://issues.apache.org/jira/browse/KAFKA-49
> Project: Kafka
> Issue Type: Bug
> Reporter: Jun Rao
>
> Currently, the produce request doesn't get acknowledged. We need to have a
> broker send a response to the producer and have the producer wait for the
> response before sending the next request.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
