Sorry about the late reply. This reply is more specific to the Uber's AthenaX project <https://github.com/uber/AthenaX>.
To trigger the savepoint we need to simply create an instance of YarnClusterClient. This class has implementation to trigger savepoint. To trigger the savepoint for any job running in flink we need the flink_id. When we submit the job to yarn cluster, we also get the corresponding flink id and save it for triggering the savepoint. To generate the client you need the an instance of ApplicationReport. Code for it - ApplicationReport report = cluster.client().getApplicationReport(application-id); Now instantiate the YarnClusterClient and trigger savepoint. -- Sent from: http://apache-flink-user-mailing-list-archive.2336050.n4.nabble.com/
