[
https://issues.apache.org/jira/browse/KAFKA-6323?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16299273#comment-16299273
]
Guozhang Wang commented on KAFKA-6323:
--------------------------------------
My concern of doing alignment with wall-clock time is that we would
intentionally trigger {{puncuate(T2)}} where the passed in parameter {{T}} is
actually not the current system wall-clock time {{NOW}}, but would be smaller
to it. For user punctuation logics that rely on the accuracy of the passed in
system time that might be a real problem (on the other hand, for stream time I
consider this a much less of an issue since it is data drive anyways and hence
stream time is defined on record timestamps only but not on when it is being
processed).
That being said, I do not feel strong against aligning with wall-clock time,
just throwing my two cents here. If people are in favor of doing this I'm OK as
well. Just remind that we need to document the behavior clearly in javadoc.
> punctuate with WALL_CLOCK_TIME triggered immediately
> ----------------------------------------------------
>
> Key: KAFKA-6323
> URL: https://issues.apache.org/jira/browse/KAFKA-6323
> Project: Kafka
> Issue Type: Bug
> Components: streams
> Affects Versions: 1.0.0
> Reporter: Frederic Arno
> Assignee: Frederic Arno
> Fix For: 1.1.0, 1.0.1
>
>
> When working on a custom Processor from which I am scheduling a punctuation
> using WALL_CLOCK_TIME. I've noticed that whatever the punctuation interval I
> set, a call to my Punctuator is always triggered immediately.
> Having a quick look at kafka-streams' code, I could find that all
> PunctuationSchedule's timestamps are matched against the current time in
> order to decide whether or not to trigger the punctuator
> (org.apache.kafka.streams.processor.internals.PunctuationQueue#mayPunctuate).
> However, I've only seen code that initializes PunctuationSchedule's timestamp
> to 0, which I guess is what is causing an immediate punctuation.
> At least when using WALL_CLOCK_TIME, shouldn't the PunctuationSchedule's
> timestamp be initialized to current time + interval?
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
