Spark Client CLI

For how to use the Spark CLIs, visit the official website http://archive.apache.org/dist/spark/docs/3.3.1/quick-start.html.

• spark-shell

  It provides an easy way to learn APIs, which is similar to the tool for interactive data analysis. It supports two languages including Scala and Python. In the Spark directory, run the ./bin/spark-shell command to access the interactive interface of Scala, obtain data from HDFS, and then perform the RDD.

  For example: a row of codes can count all words in a file.

  scala> sc.textFile("hdfs://10.96.1.57:9000//wordcount_data.txt").flatMap(l => l.split(" ")).map(w => (w,1)).reduceByKey(_+_).collect()

  You can directly specify the Keytab and Principal in the command line to obtain authentication, and regularly update the keytab and authorized tokens to avoid the authentication expiry. The following command is used as an example.

  spark-shell --principal spark2x/hadoop.<System domain name>@<System domain name> --keytab ${BIGDATA_HOME}/FusionInsight_Spark2x_8.1.0.1/install/FusionInsight-Spark2x-3.1.1/keytab/spark2x/SparkResource/spark2x.keytab --master yarn

• spark-submit

  It is used to submit the Spark application to the Spark cluster for running and return the running results. The class, master, jar and input parameter need to be specified.

  For example: Run the GroupByTest example in the jar. There are four input parameters and the specified running mode of the cluster is local single platform.

  ./bin/spark-submit --class org.apache.spark.examples.GroupByTest --master local[1] examples/jars/spark-examples_2.12-3.1.1-hw-ei-311001-SNAPSHOT.jar 6 10 10 3

  You can directly specify the Keytab and Principal in the command line to obtain authentication, and regularly update the keytab and authorized tokens to avoid the authentication expiry. The following command is used as an example.

  spark-submit --class org.apache.spark.examples.GroupByTest --master yarn --principal spark2x/hadoop.<System domain name>@<System domain name> --keytab ${BIGDATA_HOME}/FusionInsight_Spark2x_8.1.0.1/install/FusionInsight-Spark2x-3.1.1/keytab/spark2x/SparkResource/spark2x.keytab examples/jars/spark-examples_2.12-3.1.1-hw-ei-311001-SNAPSHOT.jar 6 10 10 3

• spark-sql

  It is used to perform the Hive metadata service and query command lines in the local mode. If its logical plan needs to be queried, add the explain extended before the SQL statement.

  For example:

  Select key from src group by key

  You can directly specify the Keytab and Principal in the command line to obtain authentication, and regularly update the keytab and authorized tokens to avoid the authentication expiry. The following command is used as an example.

  spark-sql --principal spark2x/hadoop.<System domain name>@<System domain name> --keytab ${BIGDATA_HOME}/FusionInsight_Spark2x_8.1.0.1/install/FusionInsight-Spark2x-3.1.1/keytab/spark2x/SparkResource/spark2x.keytab --master yarn

• run-example

  It is used to run or debug the default example in the Spark open-source community.

  For example: Run the SparkPi.

  ./run-example SparkPi 100