[
https://issues.apache.org/jira/browse/FLUME-2868?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15525994#comment-15525994
]
Tristan Stevens edited comment on FLUME-2868 at 9/27/16 12:47 PM:
------------------------------------------------------------------
You can specify the key and the Kafka Producer will hash partition based on the
key. There should be a separate JIRA to directly specify the partition.
was (Author: tmgstev):
This isn't resolved, I'm happy to take it if it can be re-opened, if not, we
can raise a new JIRA
> Kafka Channel partition topic by key
> ------------------------------------
>
> Key: FLUME-2868
> URL: https://issues.apache.org/jira/browse/FLUME-2868
> Project: Flume
> Issue Type: Improvement
> Components: Channel
> Affects Versions: v1.6.0
> Reporter: James Kinley
> Fix For: v1.7.0
>
>
> Add the ability to partition a Kafka channel topic by key. Users can
> currently partition topics by key using a file/memory channel with a Kafka
> sink. Although this approach adds unnecessary latency and the Kafka sink
> drops Flume event headers, which can be undesirable.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
