You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by "Dawid Wysakowicz (JIRA)" <ji...@apache.org> on 2018/06/06 07:18:00 UTC
[jira] [Commented] (FLINK-9533) CEP cant deal with the known
outOfOrder config with event time
[ https://issues.apache.org/jira/browse/FLINK-9533?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16502901#comment-16502901 ]
Dawid Wysakowicz commented on FLINK-9533:
-----------------------------------------
Hi [~aitozi],
I don't think this is a bug at all, neither in Flink nor in Flink CEP library. This is an expected behaviour that follows Watermark semantics. I think the solution proposed in linked stackoverflow thread is the right one to use in such scenario. The solution is to advance the Watermark if stream idleness exceeds the maxOutOfOrderness. Therefore I would vote to close this ticket.
One additional comment, if you create a Jira ticket please make it self-contained. We should not depend on external services to explain issue.
> CEP cant deal with the known outOfOrder config with event time
> --------------------------------------------------------------
>
> Key: FLINK-9533
> URL: https://issues.apache.org/jira/browse/FLINK-9533
> Project: Flink
> Issue Type: Bug
> Components: CEP
> Affects Versions: 1.4.2
> Reporter: aitozi
> Assignee: aitozi
> Priority: Major
>
> As described in the [stackoverflow|https://stackoverflow.com/questions/48694927/flink-cep-pattern-does-not-match-for-first-events-after-starting-job-and-always]
> when set the watermark generate with the maxOutOfOrder, it will run into problem when we extract events from bufferEvent to process, should fix it.
>
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)