You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by "Zhipeng Zhang (Jira)" <ji...@apache.org> on 2023/03/10 07:11:00 UTC
[jira] [Commented] (FLINK-31373) PerRoundWrapperOperator should carry epoch information in watermark
[ https://issues.apache.org/jira/browse/FLINK-31373?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17698766#comment-17698766 ]
Zhipeng Zhang commented on FLINK-31373:
---------------------------------------
As discussed with [~gaoyunhaii] offline, we agree that the watermark is not correctly processed in iteration module.
To avoid the above cases for now, we plan to add a java doc to explain that `flink-m-iteration` module cannot deal with watermarks correctly. We will leave it as a TODO here.
> PerRoundWrapperOperator should carry epoch information in watermark
> -------------------------------------------------------------------
>
> Key: FLINK-31373
> URL: https://issues.apache.org/jira/browse/FLINK-31373
> Project: Flink
> Issue Type: Bug
> Components: Library / Machine Learning
> Affects Versions: ml-2.2.0
> Reporter: Zhipeng Zhang
> Priority: Major
>
> Currently we use PerRoundWrapperOperator to wrap the normal flink operators such that they can be used in iterations.
> We already contained the epoch information in each record so that we know which iteration each record belongs to.
> However, there is no epoch information when the stream element is a watermark. This works in most cases, but fail to address the following use case:
> - In DataStreamUtils#withBroadcast, we will cache the elements (including watermarks) from non-broadcast inputs until the broadcast variables are ready. When the broadcast variables are ready, once we receive a stream element we will process the cached elements first. If the received element is a watermark, the current implementation of iteration module fails (ProxyOutput#collect throws NPE) since there is no epoch information.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)