Github user yaooqinn commented on the issue: https://github.com/apache/spark/pull/19121 @jerryshao thanks for replying if a user explicitly set `SPARK_USER` but not wrapping the `SPARK_USER` UGI doAs for SparkContext initialization, there might be two different users communicating with Hadoop in driver and executors.
--- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org