[
https://issues.apache.org/jira/browse/FLINK-10543?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16716027#comment-16716027
]
Fabian Hueske commented on FLINK-10543:
---------------------------------------
Hi [~sunjincheng121], [~hequn8128], I had a quick look at the PR and it seems
that the clean up behavior of some operators was changed.
1. Some {{ValueState}} were removed from join operators, which breaks savepoint
compatibility.
2. The clean up behavior of joins was changed. It seems that all records are
removed in case of a firing clean up timer while before, non-expired records
were kept.
We only backport bug fixes to release branches that do not break compatibility
(unless it is absolutely necessary).
Hence, the commit on the release-1.7 branch must be reverted.
I think we should also discuss the changes on the master branch.
This issue was about leveraging the new timer deletion feature to improve the
performance and not changing the clean up semantics of the operators.
I'll have a closer look at the changes in the next days to figure out what
exactly changed.
Thanks, Fabian
> Leverage efficient timer deletion in relational operators
> ---------------------------------------------------------
>
> Key: FLINK-10543
> URL: https://issues.apache.org/jira/browse/FLINK-10543
> Project: Flink
> Issue Type: Improvement
> Components: Table API & SQL
> Reporter: Fabian Hueske
> Assignee: Hequn Cheng
> Priority: Minor
> Labels: pull-request-available
> Fix For: 1.8.0, 1.7.1
>
>
> FLINK-9423 added support for efficient timer deletions. This feature is
> available since Flink 1.6 and should be used by the relational operator of
> SQL and Table API.
> Currently, we use a few workarounds to handle situations when deleting timers
> would be the better solution.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
