1)、----登录spark
ssh root@133.37.23.88
/home/mr/spark/bin/beeline -u "jdbc:hive2://133.37.23.88:18000/zxvmax" -n mr -p ""
2)、----在spark——shell 新开辟一个空间跑任务,(在跑sql时间过长的情况)
/home/mr/spark/bin/spark-shell --master spark://zdh234:7077 --executor-memory 6g --total-executor-cores 20 --driver-memory 3g
sqlContext.sql("use default").collect;
3)、----在装有spark组件的服务器开一个sparksql
ps -ef | grep watchdog
开启sparksql的脚本python /home/mr/bin/spark/sparksql_watchdog.py start