Hi Till, Thanks for the information. We are using the session cluster and is working quite good, but we would like to benefit from the new approach of per-job mode in order to have a better control over the jobs that are running on the cluster.
I took a look to the YarnJobClusterEntrypoint and I see now how this planned to be done, but if I understand correctly, in the current state there is not possible to start a Job cluster on kubernetes as there is only concrete implementation for Yarn and mesos? The objective being to have a Flink cluster running on per-job mode and able to execute several self-contained jobs, I imagine the idea would be also to have a Kubernetes specific implementation of the ResourceManager that would be initialized along the TaskManagers and would be listening for the "self-contained jobs" to join, assign resources and start the execution of the specific job, each one with its own JobManager? Is my understanding correct? Is the per-job mode on kubernetes planned to be included on 1.5 ? Regards, Edward -- Sent from: http://apache-flink-user-mailing-list-archive.2336050.n4.nabble.com/
