[
https://issues.apache.org/jira/browse/FLINK-21467?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17300329#comment-17300329
]
Piotr Nowojski commented on FLINK-21467:
----------------------------------------
I don't think there is such issue [~kezhuw]. If you restore to a checkpoint
which one of the inputs has "MAX_WATERMARK", it would also mean that in this
checkpoint the corresponding input has already finished, so there is no room
for non determinism. This source has already finished and that state has been
checkpointed. What I meant is that if you are restoring a non deterministic job
and replaying some records, the outcome might be different for each recovery
attempt - potentially including things like source finishing or not in some of
those inputs.
> Document possible recommended usage of Bounded{One/Multi}Input.endInput and
> emphasize that they could be called multiple times
> ------------------------------------------------------------------------------------------------------------------------------
>
> Key: FLINK-21467
> URL: https://issues.apache.org/jira/browse/FLINK-21467
> Project: Flink
> Issue Type: Improvement
> Components: API / DataStream
> Affects Versions: 1.13.0
> Reporter: Kezhu Wang
> Priority: Major
>
> It is too tempting to use these api, especially {{BoundedOneInput.endInput}},
> to commit final result before FLIP-147 delivered. And this will cause
> re-commit after failover as [~gaoyunhaii] has pointed out in FLINK-21132.
> I have
> [pointed|https://github.com/apache/iceberg/issues/2033#issuecomment-784153620]
> this out in
> [apache/iceberg#2033|https://github.com/apache/iceberg/issues/2033], please
> correct me if I was wrong.
> cc [~aljoscha] [~pnowojski] [~roman_khachatryan]
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
