[jira] [Commented] (STORM-697) Support for Emitting Kafka Message Offset and Partition

["Ankur C (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/STORM-697?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14941341#comment-14941341 ] 

Ankur C commented on STORM-697:
-------------------------------

Can we merge this into 0.10 release ?

> Support for Emitting Kafka Message Offset and Partition
> -------------------------------------------------------
>
>                 Key: STORM-697
>                 URL: https://issues.apache.org/jira/browse/STORM-697
>             Project: Apache Storm
>          Issue Type: Improvement
>          Components: storm-kafka
>            Reporter: Matthew Tieman
>            Assignee: Matthew Tieman
>   Original Estimate: 4h
>  Remaining Estimate: 4h
>
> It would be nice expose the offset and partition of messages consumed from Kafka to the Scheme generating the Tuples. This is useful for auditing/replaying data from arbitrary points on a Kafka topic. There are a few ways to accomplish this. 
> One is to define a new scheme, say MessageAndMetadataScheme, and add another case in KafkaUtils.generateTuples to determine if the config's schema is an instance of MessageAndMetadataScheme.
> Another is to overload KafkaUtils.generateTuples to accept and instance of MessageAndRealOffset. Next, add a flag to SpoutConfig indicating that the scheme should accept a byte array of MessageAndRealOffset, this flag will be used to determine which generateTuples method to call.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)