Github user carsonwang commented on the issue: https://github.com/apache/spark/pull/22907 What if there is a FetchFailure and Spark reruns some tasks in the previous succeeded shuffle map stage? That will be a new ShuffleMapStage and we will still double counting the accumulators, right?
--- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org