Cluster startup steps: first start the HDFS system, then start the spark cluster, and finally submit the jar to the spark cluster for execution.
1.hadoop start
cd
/home/***/hadoop-2.7.4/sbin start-all.sh
3.spark start
cd
/home/***/spark-2.2.0/sbin start-all.sh
4.spark提交
cd /home/***/spark-2.2.0/bin
spark-submit --master local --class com.helloworld.kmeans /home/***/xsd11.jar
web ui login view method
1. View hadoop UI
http://192.168.1.***:50070
All Applications 界面(yarn UI:8088)
http://192.168.1.***:8088/cluster
2. View sparkUI (start ./bin/spark-shell first)
Cluster mode: 18088
Stand-alone mode: 4044
Log server
7077
Summary: If you want to view the job execution on the UI interface, you need to remove sc.stop() from the spark program ,spark UI (4040, etc.) interface can be opened