You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@spark.apache.org by "Joseph Cooper (JIRA)" <ji...@apache.org> on 2019/08/14 21:39:00 UTC
[jira] [Commented] (SPARK-28738) Add ability to include metadata in
CanCommitOffsets API
[ https://issues.apache.org/jira/browse/SPARK-28738?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16907642#comment-16907642 ]
Joseph Cooper commented on SPARK-28738:
---------------------------------------
I think I see why commitAsync might not support this. During the polling loop for offset commits if a higher offset is encountered, a lesser one might get skipped and the metadata for that one wont get committed.
> Add ability to include metadata in CanCommitOffsets API
> -------------------------------------------------------
>
> Key: SPARK-28738
> URL: https://issues.apache.org/jira/browse/SPARK-28738
> Project: Spark
> Issue Type: New Feature
> Components: DStreams
> Affects Versions: 2.4.4
> Reporter: Joseph Cooper
> Priority: Major
>
> It is possible to commit metadata with an offset to Kafka. Currently, the CanCommitOffsets API does not expose this functionality. See [https://github.com/apache/spark/blob/017919b636fd3ce43ccf5ec57f1c1842aa2130db/external/kafka-0-10/src/main/scala/org/apache/spark/streaming/kafka010/DirectKafkaInputDStream.scala#L300]
>
> We could make the commit queue take (OffsetRange, String) instead of just OffsetRange and copy the two existing commitAsync functions and make them take Array[(OffsetRange, String)].
--
This message was sent by Atlassian JIRA
(v7.6.14#76016)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@spark.apache.org
For additional commands, e-mail: issues-help@spark.apache.org