You are viewing a plain text version of this content. The canonical link for it is here.

Posted to dev@kafka.apache.org by "Jay Bae (JIRA)" <ji...@apache.org> on 2014/04/02 02:19:14 UTC

[jira] [Commented] (KAFKA-1357) Rack-aware replica assignment

    [ https://issues.apache.org/jira/browse/KAFKA-1357?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13957176#comment-13957176 ] 

Jay Bae commented on KAFKA-1357:
--------------------------------

If you don't have bandwidth to get this done in 0.8.2, I will take it. Let me know.

> Rack-aware replica assignment
> -----------------------------
>
>                 Key: KAFKA-1357
>                 URL: https://issues.apache.org/jira/browse/KAFKA-1357
>             Project: Kafka
>          Issue Type: Improvement
>          Components: replication
>            Reporter: Jay Bae
>            Assignee: Neha Narkhede
>
> Kafka in the cloud should be resilient to zone(rack in data center) outage. ElasticSearch is already doing replica assignment based on rack info, where replicas are not assigned to instances in the same rack.
> We need this feature to make Kafka in AWS resilient in any zone outage. So far, I am doing manually get the broker information, write manual partition assignment json string and do manual assignment for the critical data source. It would be great rack-awareness replica assignment can be done automatically if the broker configuration has a rack info.



--
This message was sent by Atlassian JIRA
(v6.2#6252)