I use CDH spark(1.5.0-hadoop2.6.0) cluster, and write one spark streaming application, and start spark streaming using following command:
spark-submit --master spark://xxxx:7077 --conf spark.cores.max=4 --num-executors 4 --total-executor-cores 4 --executor-cores 4 --executor-memory 2g --class com.xxxx.KafkaActive streaming-assembly-0.0.1-SNAPSHOT.jar but I found spark node(server) memory is all used out, and the spark streaming executor number is still increasing, and new executor are started but the removed executors(CoarseGrainedExecutorBackend instances) does not exited. what should i do. look the following pics: <http://apache-spark-user-list.1001560.n3.nabble.com/file/t8431/0FBB2E50-FBCA-4AB7-A9B4-9A248082EFD6.png> -- Sent from: http://apache-spark-user-list.1001560.n3.nabble.com/ --------------------------------------------------------------------- To unsubscribe e-mail: user-unsubscr...@spark.apache.org