You are viewing a plain text version of this content. The canonical link for it is here.
Posted to jira@kafka.apache.org by "John Roesler (JIRA)" <ji...@apache.org> on 2018/04/02 14:46:00 UTC

[jira] [Commented] (KAFKA-6729) KTable should use user source topics if possible and not create changelog topic

    [ https://issues.apache.org/jira/browse/KAFKA-6729?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16422576#comment-16422576 ] 

John Roesler commented on KAFKA-6729:
-------------------------------------

Maybe worth clarifying, this is a performance regression, not a correctness one.

> KTable should use user source topics if possible and not create changelog topic
> -------------------------------------------------------------------------------
>
>                 Key: KAFKA-6729
>                 URL: https://issues.apache.org/jira/browse/KAFKA-6729
>             Project: Kafka
>          Issue Type: Bug
>          Components: streams
>    Affects Versions: 1.0.0
>            Reporter: Matthias J. Sax
>            Priority: Critical
>             Fix For: 1.2.0
>
>
> With KIP-182 we reworked Streams API largely and introduced a regression into 1.0 code base. If a KTable is populated from a source topic, ie, StreamsBuilder.table() -- the KTable does create its own changelog topic. However, in older releases (0.11 or older), we don't create a changelog topic but use the user specified source topic instead.
> We want to reintroduce this optimization to reduce the load (storage and write) on the broker side for this case.



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