[
https://issues.apache.org/jira/browse/FLINK-18796?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17171766#comment-17171766
]
Chesnay Schepler commented on FLINK-18796:
------------------------------------------
Isn't the same true for the FlinkKinesisProducer? The KinesisProducer
internally uses a ThreadPoolExecutor to complete the callbacks, which creates
threads lazily when required, so until we start processing data (i.e., call
invoke), no thread has been created yet.
> FlinkKinesisProducer.backpressureLatch should be volatile
> ---------------------------------------------------------
>
> Key: FLINK-18796
> URL: https://issues.apache.org/jira/browse/FLINK-18796
> Project: Flink
> Issue Type: Bug
> Components: Connectors / Kinesis
> Affects Versions: 1.11.1
> Reporter: Roman Khachatryan
> Assignee: Roman Khachatryan
> Priority: Minor
> Labels: pull-request-available
> Fix For: 1.12.0
>
>
> (confirm first)
>
> cc: [~rmetzger]
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
