You are viewing a plain text version of this content. The canonical link for it is here.
Posted to jira@kafka.apache.org by "Amanuel Mekonnen (JIRA)" <ji...@apache.org> on 2019/06/06 17:52:00 UTC

[jira] [Commented] (KAFKA-1793) Allow replacing ZooKeeper with a different distributed coordination system

    [ https://issues.apache.org/jira/browse/KAFKA-1793?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16857947#comment-16857947 ] 

Amanuel Mekonnen commented on KAFKA-1793:
-----------------------------------------

I think this issue seems more relevant now Zookeeper is being edged out by Consul and etcd

> Allow replacing ZooKeeper with a different distributed coordination system
> --------------------------------------------------------------------------
>
>                 Key: KAFKA-1793
>                 URL: https://issues.apache.org/jira/browse/KAFKA-1793
>             Project: Kafka
>          Issue Type: Wish
>          Components: replication
>    Affects Versions: 0.8.1.1
>            Reporter: Paul Lorenz
>            Assignee: Neha Narkhede
>            Priority: Minor
>
> It would simplify our deployment infrastructure if we could reuse our current service discovery and distributed configuration system (Consul) instead of requiring a separate ZK installation. 
> I looked through the code and it looks like ZK code (specifically zkClient), is scattered throughout the codebase. 
> It would be nice if any distributed coordination code that used ZK was behind a single (or set of) API. It would then hopefully be straightforward to provide a replacement implementation for that API. The implementing class could be specified in the config.
> One potential benefit would be the ability to provide implementations for different versions of ZK. 
> Is there any interest in having this feature available?



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)