配置Ipython Nodebook 运行 Python Spark 程序

配置Ipython Nodebook 运行 Python Spark 程序

1.1、安装Anaconda

Anaconda的官网是https://www.anaconda.com,下载对应的版本；

1.1.1、下载Anaconda

$ cd /opt/local/src/
$ wget -c https://repo.anaconda.com/archive/Anaconda3-5.2.0-Linux-x86_64.sh

1.1.2、安装Anaconda

# 参数 -b 表示 batch -p 表示指定安装目录
$ bash Anaconda3-5.2.0-Linux-x86_64.sh -p /opt/local/anaconda -b

1.1.3、配置Anaconda相关环境变量

• 配置环境变量

$ tail -n 8 ~/.bashrc

# Anaconda3
export ANACONDA_PATH=/opt/local/anaconda
export PATH=$ANACONDA_PATH/bin:$PATH

# PySpark
export PYSPARK_DRIVER_PYTHON=$ANACONDA_PATH/bin/ipython
export PYSPARK_PYTHON=$ANACONDA_PATH/bin/python

• 启用环境变量

$ source ~/.bashrc

• 验证

$ python --version
Python 3.6.5 :: Anaconda, Inc.

1.2、在Ipython Notebook 使用pySpark

1.2.1、创建工作目录

$ mkdir  ~/ipynotebook
$ cd ~/ipynotebook

1.2.2、Ipython Notebook 运行pySpark

• 运行Ipython Notebook

$ PYSPARK_DRIVER_PYTHON=ipython PYSPARK_DRIVER_PYTHON_OPTS="notebook" pyspark
[TerminalIPythonApp] WARNING | Subcommand `ipython notebook` is deprecated and will be removed in future versions.
[TerminalIPythonApp] WARNING | You likely want to use `jupyter notebook` in the future
[I 14:21:56.030 NotebookApp] JupyterLab beta preview extension loaded from /opt/local/anaconda/lib/python3.6/site-packages/jupyterlab
[I 14:21:56.030 NotebookApp] JupyterLab application directory is /opt/local/anaconda/share/jupyter/lab
[I 14:21:56.037 NotebookApp] Serving notebooks from local directory: /home/hadoop/ipynotebook
[I 14:21:56.037 NotebookApp] 0 active kernels
[I 14:21:56.037 NotebookApp] The Jupyter Notebook is running at:
[I 14:21:56.037 NotebookApp] http://localhost:8888/?token=5b68718fdabe4488decf07703a3bd76bf46d5dc733a6617d
[I 14:21:56.037 NotebookApp] Use Control-C to stop this server and shut down all kernels (twice to skip confirmation).
[C 14:21:56.040 NotebookApp] 

    Copy/paste this URL into your browser when you connect for the first time,
    to login with a token:
        http://localhost:8888/?token=5b68718fdabe4488decf07703a3bd76bf46d5dc733a6617d&token=5b68718fdabe4488decf07703a3bd76bf46d5dc733a6617d
[I 14:21:56.683 NotebookApp] Accepting one-time-token-authenticated connection from 127.0.0.1

会自动通过默认的浏览器打开http://localhost:8888 页面

• 在IPython Notebook 上编写程序

1.2.3、Ipython Notebook 在Hadoop Yarn 运行pySpark

• 运行Ipython Notebook

$ PYSPARK_DRIVER_PYTHON=ipython PYSPARK_DRIVER_PYTHON_OPTS="notebook" HADOOP_CONF_DIR=/opt/local/hadoop/etc/hadoop MASTER=yarn-client pyspark
[TerminalIPythonApp] WARNING | Subcommand `ipython notebook` is deprecated and will be removed in future versions.
[TerminalIPythonApp] WARNING | You likely want to use `jupyter notebook` in the future
[I 14:50:48.149 NotebookApp] JupyterLab beta preview extension loaded from /opt/local/anaconda/lib/python3.6/site-packages/jupyterlab
[I 14:50:48.149 NotebookApp] JupyterLab application directory is /opt/local/anaconda/share/jupyter/lab
[I 14:50:48.157 NotebookApp] Serving notebooks from local directory: /home/hadoop/ipynotebook
[I 14:50:48.157 NotebookApp] 0 active kernels
[I 14:50:48.157 NotebookApp] The Jupyter Notebook is running at:
[I 14:50:48.157 NotebookApp] http://localhost:8888/?token=8fe2c599dc39a23104dd6a058a0e05de3d9e88cfeda71b45
[I 14:50:48.157 NotebookApp] Use Control-C to stop this server and shut down all kernels (twice to skip confirmation).
[C 14:50:48.161 NotebookApp] 

    Copy/paste this URL into your browser when you connect for the first time,
    to login with a token:
        http://localhost:8888/?token=8fe2c599dc39a23104dd6a058a0e05de3d9e88cfeda71b45&token=8fe2c599dc39a23104dd6a058a0e05de3d9e88cfeda71b45

• 在IPython Notebook 上编写程序

• 在YARN查看任务

$ yarn application -list
18/06/24 14:53:06 INFO client.RMProxy: Connecting to ResourceManager at node/192.168.20.10:8032
Total number of applications (application-types: [] and states: [SUBMITTED, ACCEPTED, RUNNING]):1
                Application-Id      Application-Name        Application-Type          User       Queue               State         Final-State         Progress                        Tracking-URL
application_1529805293111_0001          PySparkShell                   SPARK        hadoop     default             RUNNING           UNDEFINED              10%                    http://node:4040

1.2.4、Ipython Notebook 在Spark Stand Alone 运行pySpark

• 启动Spark Stand Alone

$ /opt/local/spark/sbin/start-master.sh

$ /opt/local/spark/sbin/start-slaves.sh

$ jps
13249 Jps
13027 Master
13188 Worker

• 运行Ipython Notebook

$ PYSPARK_DRIVER_PYTHON=ipython PYSPARK_DRIVER_PYTHON_OPTS="notebook" MASTER=spark://node:7077 pyspark --num-executors 1 --total-executor-cores 1 --executor-memory 512m
[TerminalIPythonApp] WARNING | Subcommand `ipython notebook` is deprecated and will be removed in future versions.
[TerminalIPythonApp] WARNING | You likely want to use `jupyter notebook` in the future
[I 15:11:59.211 NotebookApp] JupyterLab beta preview extension loaded from /opt/local/anaconda/lib/python3.6/site-packages/jupyterlab
[I 15:11:59.212 NotebookApp] JupyterLab application directory is /opt/local/anaconda/share/jupyter/lab
[I 15:11:59.230 NotebookApp] Serving notebooks from local directory: /home/hadoop/ipynotebook
[I 15:11:59.230 NotebookApp] 0 active kernels
[I 15:11:59.230 NotebookApp] The Jupyter Notebook is running at:
[I 15:11:59.230 NotebookApp] http://localhost:8888/?token=1972eb523fea28d541985df7ed2ce55cc2bfada7e31eb9ea
[I 15:11:59.230 NotebookApp] Use Control-C to stop this server and shut down all kernels (twice to skip confirmation).
[C 15:11:59.233 NotebookApp] 

    Copy/paste this URL into your browser when you connect for the first time,
    to login with a token:
        http://localhost:8888/?token=1972eb523fea28d541985df7ed2ce55cc2bfada7e31eb9ea&token=1972eb523fea28d541985df7ed2ce55cc2bfada7e31eb9ea
[I 15:12:02.594 NotebookApp] Accepting one-time-token-authenticated connection from 127.0.0.1

• 在IPython Notebook 上编写程序

• 查看Spark Standalone Web UI 界面
  

1.3、总结

启动启动Ipython Notebook，首先进入Ipython Notebook的工作目录，如~/ipynotebook这个根据实际的情况确定；

1.3.1、Local 启动Ipython Notebook

PYSPARK_DRIVER_PYTHON=ipython PYSPARK_DRIVER_PYTHON_OPTS="notebook" pyspark
#### 或者
PYSPARK_DRIVER_PYTHON=ipython PYSPARK_DRIVER_PYTHON_OPTS="notebook" pyspark --master local[*]

1.3.2、Hadoop YARN 启动Ipython Notebook

PYSPARK_DRIVER_PYTHON=ipython PYSPARK_DRIVER_PYTHON_OPTS="notebook" HADOOP_CONF_DIR=/opt/local/hadoop/etc/hadoop MASTER=yarn-client pyspark
#### 或者
PYSPARK_DRIVER_PYTHON=ipython PYSPARK_DRIVER_PYTHON_OPTS="notebook" HADOOP_CONF_DIR=/opt/local/hadoop/etc/hadoop pyspark --master yarn --deploy-mode client

1.3.2、Spark Stand Alone 启动Ipython Notebook

PYSPARK_DRIVER_PYTHON=ipython PYSPARK_DRIVER_PYTHON_OPTS="notebook" MASTER=spark://node:7077 pyspark --num-executors 1 --total-executor-cores 1 --executor-memory 512m 

原文地址：http://blog.51cto.com/balich/2132206