You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by "Stephan Ewen (Jira)" <ji...@apache.org> on 2020/11/09 23:43:00 UTC
[jira] [Resolved] (FLINK-20049) Simplify handling of "request
split".
[ https://issues.apache.org/jira/browse/FLINK-20049?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Stephan Ewen resolved FLINK-20049.
----------------------------------
Resolution: Fixed
Implemented in 1.12.0 (master) via 92ca579942112e1ad7bf897b9f63e70f277145a0
> Simplify handling of "request split".
> -------------------------------------
>
> Key: FLINK-20049
> URL: https://issues.apache.org/jira/browse/FLINK-20049
> Project: Flink
> Issue Type: Improvement
> Components: API / Core
> Reporter: Stephan Ewen
> Assignee: Stephan Ewen
> Priority: Major
> Fix For: 1.12.0
>
>
> This issue is similar to FLINK-19265
> Split assignment events are treated specially by the source API. Users do not create them directly but call methods on the contexts to assign splits.
> The {{RequestSplitEvent}} is in contrast to that a custom user event and needs to be handled like a custom event, when sent by enumerators and received by the readers.
> That seems a bit confusing and inconsistent, given that {{RequestSplitEvent}} is essential for all use cases with pull-based split assignment, which is pretty much any batch use case and various streaming use cases. The event should be on the same level as the AddSplitEvent.
> I suggest that we add a {{SourceReaderContext.requestSplit()}} and {{SplitEnumerator.handleSplitRequest()}}, to have split requests and responses symmetrical.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)