[ 
https://issues.apache.org/jira/browse/FLINK-41310?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18146463#comment-18146463
 ] 

Sergey Nuyanzin commented on FLINK-41310:
-----------------------------------------

Merged as 
[c5f8802be786b7e301f8bcfdd5cc8fcf9b41ba3a|https://github.com/apache/flink/commit/c5f8802be786b7e301f8bcfdd5cc8fcf9b41ba3a]

> BatchExecAdaptiveJoin should take external-buffer-memory from the compiled 
> plan
> -------------------------------------------------------------------------------
>
>                 Key: FLINK-41310
>                 URL: https://issues.apache.org/jira/browse/FLINK-41310
>             Project: Flink
>          Issue Type: Bug
>          Components: Table SQL / Planner
>    Affects Versions: 2.5.0
>            Reporter: Bonnie Varghese
>            Assignee: Bonnie Varghese
>            Priority: Major
>              Labels: pull-request-available
>             Fix For: 2.5.0
>
>
> {\{BatchExecAdaptiveJoin}} computes the join's managed memory budget at 
> translation time from the node's config, but builds the operator at runtime, 
> where it reads \{{table.exec.resource.external-buffer-memory}} from the job 
> config. The buffer is a share of that budget, so both values should come from 
> the same config.
> {\{BatchExecHashJoin}} and \{{BatchExecSortMergeJoin}} read both values at 
> translation time. The adaptive join only decides the join strategy at 
> runtime; its memory budget is already fixed before it runs.
> This should be fixed after FLINK-41263 is merged.



--
This message was sent by Atlassian Jira
(v10.3.26#10030026)

Reply via email to