You are viewing a plain text version of this content. The canonical link for it is here.
Posted to jira@kafka.apache.org by "A. Sophie Blee-Goldman (Jira)" <ji...@apache.org> on 2021/03/19 00:51:00 UTC

[jira] [Commented] (KAFKA-12506) Expand AdjustStreamThreadCountTest

    [ https://issues.apache.org/jira/browse/KAFKA-12506?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17304551#comment-17304551 ] 

A. Sophie Blee-Goldman commented on KAFKA-12506:
------------------------------------------------

cc [~wcarlson5] -- the first two items are probably a prerequisite for the test I described [here|https://github.com/apache/kafka/pull/10356#pullrequestreview-615974924], I'll try and remember to cross them off this list when we have something like that

> Expand AdjustStreamThreadCountTest
> ----------------------------------
>
>                 Key: KAFKA-12506
>                 URL: https://issues.apache.org/jira/browse/KAFKA-12506
>             Project: Kafka
>          Issue Type: Improvement
>          Components: streams, unit tests
>            Reporter: A. Sophie Blee-Goldman
>            Priority: Major
>              Labels: newbie, newbie++
>
> Right now the AdjustStreamThreadCountTest runs a minimal topology that just consumes a single input topic, and doesn't produce any data to this topic. Some of the complex concurrency bugs that we've found only showed up when we had some actual data to process and a stateful topology: [KAFKA-12503|https://issues.apache.org/jira/browse/KAFKA-12503]
> We should try to expand the Streams application in this test to really flex the feature. Imo this would include 
> # Produce some data to the input topic 
> # One or more stateful operators in the topology, with caching enabled
> # A repartition to get multiple subtopologies
> # Write out some results to an output topic
> # Verify that these results show up in the output



--
This message was sent by Atlassian Jira
(v8.3.4#803005)