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

Posted to issues@flink.apache.org by "Weijie Guo (Jira)" <ji...@apache.org> on 2023/08/11 02:40:00 UTC

[jira] [Commented] (FLINK-32709) Fix the bug of low memory utilization for Hybrid Shuffle

    [ https://issues.apache.org/jira/browse/FLINK-32709?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17753025#comment-17753025 ] 

Weijie Guo commented on FLINK-32709:
------------------------------------

Master(1.18) via 0f841624a388be9ad551ec1ccbc15abb31044d6e.

> Fix the bug of low memory utilization for Hybrid Shuffle
> --------------------------------------------------------
>
>                 Key: FLINK-32709
>                 URL: https://issues.apache.org/jira/browse/FLINK-32709
>             Project: Flink
>          Issue Type: Bug
>          Components: Runtime / Network
>    Affects Versions: 1.18.0
>            Reporter: Yuxin Tan
>            Assignee: Yuxin Tan
>            Priority: Major
>              Labels: pull-request-available
>
> Currently, each subpartition in Disk/Remote has a segment size of 8M. When writing segments to the Disk tier with a parallelism of 1000, only shuffle data exceeding 1000 * 8M can be written to the Memory tier again. However, for most shuffles, the data volume size falls below this limit, significantly impacting Memory tier utilization. 
> For better performance, it is necessary to address this issue to improve the memory tier utilization.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)