[
https://issues.apache.org/jira/browse/FLINK-10967?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16773486#comment-16773486
]
Seth Wiesman edited comment on FLINK-10967 at 2/20/19 11:16 PM:
----------------------------------------------------------------
[~aljoscha] There is a bug in the kafka client version 2.0.1 (the version flink
currently uses) and -2.1 (this pr)- which has been resolved in 2.1.1 where the
client fails on rolling broker restarts which lead to flink jobs failing[1].
Depending on the state of the kafka tests and your confidence in them it would
be great to upgrade before 1.8 is released.
[1] https://issues.apache.org/jira/browse/KAFKA-7890
edit: it looks like the author has updated their pr to 2.1.1
was (Author: sjwiesman):
[~aljoscha] There is a bug in the kafka client version 2.0.1 (the version flink
currently uses) and 2.1 (this pr) which has been resolved in 2.1.1 where the
client fails on rolling broker restarts which lead to flink jobs failing[1].
Depending on the state of the kafka tests and your confidence in them it would
be great to upgrade before 1.8 is released.
[1] https://issues.apache.org/jira/browse/KAFKA-7890
> Update kafka dependency to 2.1.0
> --------------------------------
>
> Key: FLINK-10967
> URL: https://issues.apache.org/jira/browse/FLINK-10967
> Project: Flink
> Issue Type: Sub-task
> Components: Kafka Connector
> Reporter: Ismael Juma
> Priority: Major
> Labels: pull-request-available
>
> Apache Kafka 2.1.0 includes an important improvement with regards to producer
> timeouts as described in KIP-91:
> https://cwiki.apache.org/confluence/display/KAFKA/KIP-91+Provide+Intuitive+User+Timeouts+in+The+Producer
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
