[
https://issues.apache.org/jira/browse/FLINK-35462?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
yuanfenghu updated FLINK-35462:
-------------------------------
Description:
When using the flinksql connector, we can set the parallelism of the sink.
However, when we modify the parallelism of the sink to be the same as the
global one, the global chaining strategy may cause the two operators to be
linked together, For checkpoint or savepoint recovery, which will make the task
unable to proceed. we should allow explicit specification of the chaining mode
of the connector.
Related issues:
[https://github.com/apache/iceberg/issues/10371]
was:
When using the flinksql connector, we can set the parallelism of the sink.
However, when we modify the parallelism of the sink to be the same as the
global one, the global chaining strategy may cause the two operators to be
linked together, which will make the task unable to proceed. For checkpoint or
savepoint recovery, we should allow explicit specification of the chaining mode
of the connector.
Related issues:
https://github.com/apache/iceberg/issues/10371
> Introduce a parameter to control the chaining mode between a Flink SQL
> connector sink or source and its adjacent operator.
> --------------------------------------------------------------------------------------------------------------------------
>
> Key: FLINK-35462
> URL: https://issues.apache.org/jira/browse/FLINK-35462
> Project: Flink
> Issue Type: Improvement
> Components: Table SQL / Planner
> Reporter: yuanfenghu
> Priority: Major
>
> When using the flinksql connector, we can set the parallelism of the sink.
> However, when we modify the parallelism of the sink to be the same as the
> global one, the global chaining strategy may cause the two operators to be
> linked together, For checkpoint or savepoint recovery, which will make the
> task unable to proceed. we should allow explicit specification of the
> chaining mode of the connector.
>
> Related issues:
> [https://github.com/apache/iceberg/issues/10371]
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
