[
https://issues.apache.org/jira/browse/KAFKA-12643?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17318116#comment-17318116
]
Guozhang Wang commented on KAFKA-12643:
---------------------------------------
Hello [~devano], thanks for reporting this issue, could you check if what you
observed is similar to this scenario?
https://issues.apache.org/jira/browse/KAFKA-12323
> Kafka Streams 2.7 with Kafka Broker 2.6.x regression: bad timestamp in
> transform/process (this.context.schedule function)
> -------------------------------------------------------------------------------------------------------------------------
>
> Key: KAFKA-12643
> URL: https://issues.apache.org/jira/browse/KAFKA-12643
> Project: Kafka
> Issue Type: Bug
> Components: streams
> Affects Versions: 2.7.0
> Reporter: David EVANO
> Priority: Major
> Attachments: Capture d’écran 2021-04-09 à 17.50.05.png
>
>
> During a tranform() or a process() method:
> Define a schedule tyask:
> this.context.schedule(Duration.ofSeconds(1), PunctuationType.WALL_CLOCK_TIME,
> timestamp -> \{...}
> store.put(...) or context.forward(...) produce a record with an invalid
> timestamp.
> For the forward, a workaround is define the timestamp:
> context.forward(entry.key, entry.value.toString(),
> To.all().withTimestamp(timestamp));
> But for state.put(...) or state.delete(...) functions there is no workaround.
> Is it mandatory to have the Kafka broker version aligned with the Kafka
> Streams version?
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
