spark 2.1.0 foreachRDD write slowly to HDFS

[446463...@qq.com]

Hi All:
 when  I use spark-streaming to consume kafka topic data to HDFS with sparking 
2.1.0.
I find it's slow .Why?

environment:CDH5.8
scala version:2.11
kafka version 0.10.1.1
spark-streaming-kafka :spark-streaming-kafka-0.8_2.11
should I replace the spark-streaming-kafka with spark-streaming-kafka-0.10_2.11?



446463...@qq.com