[jira] [Commented] (FLINK-31305) KafkaWriter doesn't wait for errors for in-flight records before completing flush

["Mason Chen (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/FLINK-31305?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17695932#comment-17695932 ] 

Mason Chen commented on FLINK-31305:
------------------------------------

Hi all, this is a critical regression for deployments who have migrated from the FlinkKafkaProducer to KafkaSink. Can someone with permissions assign this issue to me? Thanks in advance!

> KafkaWriter doesn't wait for errors for in-flight records before completing flush
> ---------------------------------------------------------------------------------
>
>                 Key: FLINK-31305
>                 URL: https://issues.apache.org/jira/browse/FLINK-31305
>             Project: Flink
>          Issue Type: Improvement
>          Components: Connectors / Kafka
>    Affects Versions: 1.17.0, 1.16.1
>            Reporter: Mason Chen
>            Priority: Major
>             Fix For: 1.17.0
>
>
> The KafkaWriter flushing needs to wait for all in-flight records to send successfully. This can be achieved by tracking requests and returning a response from the registered callback from the producer#send() logic.
> There is potential for data loss since the checkpoint does not accurately reflect that all records have been sent successfully, to preserve at least once semantics.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)