Hi Dipanjan,

I think you are right that it's already been supported to submit a job to cluster via SQL client. This is supported in [1]. Besides, I think that it is not configured in the YAML. It's specified in the CLI options when you start up the SQL client CLI.

[1] https://issues.apache.org/jira/browse/FLINK-8852

在 2019年9月23日,上午11:00,Terry Wang <zjuwangg@gmail.com> 写道:

Hi Dipanjan:

I just looked through the Flink SQL client code and got the same conclusion as you.
Look forward to receiving other comments.

Terry Wang

在 2019年9月22日,下午11:53,Dipanjan Mazumder <javahub@yahoo.com.INVALID> 写道:

Hi ,
 Thanks again for responding on my earlier queries..
   I was again going through the Flink SQL client code and came across the default custom command-line , few days back i came to know that Flink sql client is not supported in a full fledged cluster with different resource managers like yarn but this "org.apache.flink.yarn.cli.FlinkYarnSessionCli" class seems like is used by the SQL client to establish session with yarn managed cluster.

Am i wrong in thinking this or is there some other use for this class. Please kindly help on the same.