[
https://issues.apache.org/jira/browse/FLINK-26416?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Fabian Paul updated FLINK-26416:
--------------------------------
Description:
With the introduction of Sink V2, the operator model of the sink changed
slightly therefore it makes sense to test different upgrade/sanity scenarios.
You can take any of the existing Sinks in the project. I would recommend the
FileSink.
# Run a job with Flink 1.14 and take a savepoint and try to restore and resume
with 1.15
# Run a job with Flink 1.15 and take a savepoint and try to restore and resume
with 1.15
# Run a bounded job with Flink 1.15
In all cases, please verify that all records have been written at the end of
the scenario and there are no duplicates.
was:
With the introduction of Sink V2, the operator model of the sink changed
slightly therefore it makes sense to test different upgrade/sanity scenarios.
You can take any of the existing Sinks in the project. I would recommend the
FileSink.
# Run a job with Flink 1.14 and take a savepoint and try to restore and resume
with 1.15
# Run a job with Flink 1.15 and take a savepoint and try to restore and resume
with 1.15
# Run a bounded job with Flink 1.15
In all cases, please verify that all records have been written at the end of
the scenario and there are no duplicates.
> Release Testing: Sink V2 sanity checks
> --------------------------------------
>
> Key: FLINK-26416
> URL: https://issues.apache.org/jira/browse/FLINK-26416
> Project: Flink
> Issue Type: Improvement
> Components: Connectors / Common
> Affects Versions: 1.15.0
> Reporter: Fabian Paul
> Assignee: Liu
> Priority: Blocker
> Labels: release-testing
> Fix For: 1.15.0
>
>
> With the introduction of Sink V2, the operator model of the sink changed
> slightly therefore it makes sense to test different upgrade/sanity scenarios.
>
> You can take any of the existing Sinks in the project. I would recommend the
> FileSink.
>
> # Run a job with Flink 1.14 and take a savepoint and try to restore and
> resume with 1.15
> # Run a job with Flink 1.15 and take a savepoint and try to restore and
> resume with 1.15
> # Run a bounded job with Flink 1.15
>
> In all cases, please verify that all records have been written at the end of
> the scenario and there are no duplicates.
>
>
>
>
>
>
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
