Github user tillrohrmann commented on a diff in the pull request: https://github.com/apache/flink/pull/2257#discussion_r71322624 --- Diff: flink-yarn/src/main/java/org/apache/flink/yarn/YarnFlinkResourceManager.java --- @@ -78,6 +79,9 @@ /** The containers where a TaskManager is starting and we are waiting for it to register */ private final Map<ResourceID, YarnContainerInLaunch> containersInLaunch; + /** The container where a TaskManager has been started and is running in */ + private final Map<ResourceID, Container> containersLaunched; --- End diff -- It is true that it holds the registered resources but it does not hold the launched containers. When a `JobManager` loses its leadership the list of registered workers will be cleared. In order to reconstruct the mapping `ResourceID --> Container`, you need this new map.
--- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---