You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by "Zhilong Hong (Jira)" <ji...@apache.org> on 2021/03/01 03:28:00 UTC
[jira] [Commented] (FLINK-21524) Improve the stability of scheduler
benchmarks
[ https://issues.apache.org/jira/browse/FLINK-21524?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17292580#comment-17292580 ]
Zhilong Hong commented on FLINK-21524:
--------------------------------------
cc [~zhuzh], [~pnowojski], and [~chesnay].
> Improve the stability of scheduler benchmarks
> ---------------------------------------------
>
> Key: FLINK-21524
> URL: https://issues.apache.org/jira/browse/FLINK-21524
> Project: Flink
> Issue Type: Improvement
> Components: Benchmarks, Runtime / Coordination
> Affects Versions: 1.13.0
> Reporter: Zhilong Hong
> Priority: Major
> Fix For: 1.13.0
>
>
> Due to FLINK-21514, we find that when someone modifies interfaces and methods used by scheduler benchmarks, the compilation and execution of flink-benchmark may break.
>
> To improve the stability of scheduler benchmarks, we decide to replace scheduler benchmarks with wrapper/executor classes, and move the current implementations to the flink repository. Also we'll implement several unit tests based on the implementations. In this way, when someone modifies interfaces and methods used by scheduler benchmarks, the unit tests fails, and then they must fix the broken benchmarks.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)