Liu created FLINK-23300: --------------------------- Summary: Job fails very slow because of no notifyAllocationFailure for DeclarativeSlotManager Key: FLINK-23300 URL: https://issues.apache.org/jira/browse/FLINK-23300 Project: Flink Issue Type: Improvement Components: Runtime / Task Affects Versions: 1.13.1 Reporter: Liu
When container is killed, flink on yarn can detect the problem very quickly. But when using default DeclarativeSlotManager, notifyAllocationFailure is not called and the task is not failed until heartbeat is timeout. So the failover will be very slow. -- This message was sent by Atlassian Jira (v8.3.4#803005)