[
https://issues.apache.org/jira/browse/FLINK-35436?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
hk__lrzy updated FLINK-35436:
-----------------------------
Affects Version/s: cdc-3.1.0
> Job can't launch when setting the option schema.change.behavior to IGNORE
> -------------------------------------------------------------------------
>
> Key: FLINK-35436
> URL: https://issues.apache.org/jira/browse/FLINK-35436
> Project: Flink
> Issue Type: Improvement
> Components: Flink CDC
> Affects Versions: cdc-3.1.0
> Reporter: hk__lrzy
> Priority: Major
> Labels: pull-request-available
>
> Now in the 3.0 pipeline, *SchemaOperator* already was necessary operator in
> the Flink DAG, both *PrePartitionOperator* and *DataSinkWriterOperator* have
> connection with the *SchemaRegister* according the *schemaEvolutionClient,*
> but when we set the option schema.change.behavior to ignore or exception, the
> pipeline will add a filter operator instead of the *Schema Operator,* final
> cause the job fail.
> I think we still need keep the option for the schema.change.behavior to meet
> the difference cases, so i advice to move schema.change.behavior to the
> *SchemaRegister* to let *SchemaOperator* will be always in the DAG, and let
> to *SchemaRegister* decided to apply the schema change or not.{*}{*}
>
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
