big face cat

@chenjunjiedada @stevenzwu hello, Can help me review this PR?

> @huyuanfeng2018: Please fix the checks checkstyle issue done.

> LGTM. just 2 nit comments to be addressed done. @stevenzwu

> Thanks @huyuanfeng2018 for the PR and @stevenzwu for the review! > > @huyuanfeng2018: Could you please backport the changes to Flink 1.17 and 1.16? > > Thanks, Peter ok,...

cc @danny0405

> I think it is good to add reminders, but based on our current operating environment, we need to package two different flink images to distinguish them in the production...

> No, hudi does not package enging specific jars in the uber jar. This will result in a large package

> > flink-table-planner has been removed from the classpath since flink1.15 > > Then how the classes in flink-table-planner is loaded then, can we do similiar loadings in Hudi side?...

``` SingleOutputStreamOperator committed = written.rebalance().transform( GLOBAL_COMMITTER_NAME, new MultiTableCommittableTypeInfo(), new CommitterOperator( streamingCheckpointEnabled, commitUser, createCommitterFactory(), createCommittableStateManager())) .setParallelism(1) .setMaxParallelism(1); ``` This can be avoided by adding rebalance before the commit operator, because the...

> @huyuanfeng2018 this is a good issue, but a chain committer can reduce resource cost. > > maybe we can have an option to control this. +1. Agree