lamber-ken edited a comment 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-487131615 > Hi @lamber-ken, I think the inner `Exception` is not a problem because it is actually only a `IOException`. Thus, the `archiveJob` method can only throw an `IOException`. > > The NPE is clearly a Flink bug and as such we should rather fix the bug than guard against it. What we could do, though, is to run `jobTerminationFuture.complete(ApplicationStatus.fromJobStatus(archivedExecutionGraph.getState()));` in a `finally` block. But we should not catch the exception there. Per contract `jobReachedGloballyTerminalState` should not throw any exception. If it does, then it violates the contract and Flink gives no guarantees how it behaves afterwards. Hi, @tillrohrmann, thanks for you explain and agree with your point. I had move `jobTerminationFuture.complete(ApplicationStatus.fromJobStatus(archivedExecutionGraph.getState()))` to a finally block, and log the exception for analysis, then throw it continue.
---------------------------------------------------------------- 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: us...@infra.apache.org With regards, Apache Git Services
