[
https://issues.apache.org/jira/browse/NIFI-1095?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14989606#comment-14989606
]
Mark Payne commented on NIFI-1095:
----------------------------------
I implemented this functionality as part of NIFI-1095, as 1095 required a
complete rewrite of the Processor. We now allow Partition Strategies of Round
Robin, Random (the Kafka default), and User-Defined Partition. The User-Defined
Partition allows users to partition the data ahead of time with attributes so
that if desired related data can be pushed to the same partition.
> Add Round Robin custom partitioner class to PutKafka
> ----------------------------------------------------
>
> Key: NIFI-1095
> URL: https://issues.apache.org/jira/browse/NIFI-1095
> Project: Apache NiFi
> Issue Type: Improvement
> Components: Core Framework
> Affects Versions: 0.3.0
> Reporter: Andre
> Assignee: Mark Payne
>
> While Kafka offers some flexibility around partitioning and data rebalancing,
> the current PutKafka processor has just two strategies for rebalancing:
> - Random (default?)
> - Key (aka Hashed)
> It would be great if it also had RoundRobin as some of the other Kafka
> producers around there [1][2].
> Implementation can be done with certain ease:
> http://ankitasblogger.blogspot.com.au/2014/12/kafka-round-robin-partition.html
> [1] see partitioner ->
> https://hekad.readthedocs.org/en/v0.9.2/config/outputs/kafka.html
> [2] see roundrobin.py ->
> https://github.com/mumrah/kafka-python/tree/master/kafka/partitioner
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
