[
https://issues.apache.org/jira/browse/FLINK-40703?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Martijn Visser resolved FLINK-40703.
------------------------------------
Fix Version/s: kafka-5.1.0
Resolution: Fixed
Fixed in apache/flink-connector-kafka:main
5accb66f79e7427def1ba3ec5810e189bc0ed48c
> Improve correctness coverage: Add ITCase for recovery from a checkpoint whose
> pooled transactional id was reused by a later checkpoint
> --------------------------------------------------------------------------------------------------------------------------------------
>
> Key: FLINK-40703
> URL: https://issues.apache.org/jira/browse/FLINK-40703
> Project: Flink
> Issue Type: Improvement
> Components: Connectors / Kafka
> Reporter: Aleksandr Savonin
> Assignee: Aleksandr Savonin
> Priority: Major
> Labels: pull-request-available
> Fix For: kafka-5.1.0
>
>
> Motivation: improve end-to-end correctness coverage of the exactly-once sink.
> The POOLING recovery path is covered at writer level
> (ExactlyOnceKafkaWriterITCase) and unit level (KafkaCommitterTest,
> TransactionAbortStrategyImplTest, ProducerPoolImplITCase), but no test runs
> the real committer, the backchannel and a MiniCluster restore together.
> The KafkaSinkITCase recovery tests fail their jobs at checkpoint 2, so no
> third checkpoint ever reuses an id.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
