## yarn client hdfs文件 spark-submit --master yarn --deploy-mode cluster --driver-memory 1g --num-executors 3 --executor-memory 1g --executor-cores 1 --archives hdfs://hadoop102:8020/user/xxx/python_zips/pyspark_env.zip#my_env --conf spark.pyspark.driver.python=./my_env/pyspark_env/bin/python3.8 --conf spark.pyspark.python=./my_env/pyspark_env/bin/python3.8 /tmp/pycharm_project_455/helloworld.py
--archives hdfs://hadoop102:8020/user/xxx/python_zips/pyspark_env.zip#my_env （一定要有#my_env，分发到yarn 容器中解压生成的目录名）
python环境的zip包 上传到hdfs的用户目录下
zip -r pyspark_env.zip pyspark_env/
hdfs://hadoop102:8020/user/xxx/python_zips/pyspark_env.zip
