Github user vanzin commented on the issue: https://github.com/apache/spark/pull/19399 I'm actually more concerned about the SHS having its own idea of what success and failure means; it sounds like a source for headaches (why is the app failed in the SHS but successful in YARN?). There isn't enough information in event logs today to define success or status, so we should either add that explicitly to the Spark API (e.g. `sc.stop(FAILED)`) or not bother with trying to figure that out.
--- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org