https://www.cnblogs.com/wwcom123/p/10561806.html
spark submit jobs, yarn-cluster pattern examples:
./bin/spark-submit\
--class com.ww.rdd.wordcount \
--master yarn \
--deploy-mode cluster \
--executor-memory 4G \
--num-executors 50 \
--driver-memory 1G \
--conf spark.default.parallelism=1000 \
--conf spark.memory.fraction=0.75 \
--conf spark.memory.storageFraction=0.5 \
/home/spark/wordcount.jar \
1000 # to the Senate