You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by "Chesnay Schepler (Jira)" <ji...@apache.org> on 2022/08/31 08:56:00 UTC

[jira] [Comment Edited] (FLINK-29132) SubtaskMetricStore causes memory leak.

    [ https://issues.apache.org/jira/browse/FLINK-29132?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17598245#comment-17598245 ] 

Chesnay Schepler edited comment on FLINK-29132 at 8/31/22 8:55 AM:
-------------------------------------------------------------------

True, the attempts will continue to pile up until the job/tm terminates.

[~pltbkd] please fix by leveraging the execution history size parameter


was (Author: zentol):
True, the attempts will continue to pile up until the job/tm terminates.

[~pltbkd] please fix

> SubtaskMetricStore causes memory leak.
> --------------------------------------
>
>                 Key: FLINK-29132
>                 URL: https://issues.apache.org/jira/browse/FLINK-29132
>             Project: Flink
>          Issue Type: Bug
>          Components: Runtime / Metrics
>            Reporter: Sitan Pang
>            Priority: Blocker
>         Attachments: dump-attempts.png, dump-metricStore.png, image-2022-08-30-11-56-34-608.png, image-2022-08-30-11-57-59-325.png
>
>
> In [FLINK-28588], MetricStore supports multiple attempts of a subtask. However, `SubtaskMetricStore` doesn't have a clean mechanism.  In failover scenario, `attempts` pile up and cause OOM.



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