tillrohrmann commented on issue #8254: [FLINK-12219][runtime] Yarn application can't stop when flink job failed in per-job yarn cluste mode URL: https://github.com/apache/flink/pull/8254#issuecomment-487977504 It is true that Yarn could try to restart the cluster depending on the configuration. Flink would then try to re-execute the job again since it is part of the Yarn application (no problem with the submitted job graph store). Depending on whether the Flink bug is transient or not the failure would happen again until all restarts are depleted or eventually the job will succeed (potentially with producing duplicate results). Given that this is caused by a bug in Flink, I think it is ok to say that we don't give hard guarantees in this case. The important bit is that we report the problem. I would open a PR based on my commit to add this utility.
---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services
