1. 上传文件到hdfs

# 前提挂载了 -v ~/bilibili/input_files:/input_files

# hdfs创建input文件夹
docker exec namenode hdfs dfs -mkdir /input

# 将容器内input_files文件夹下的1.txt上传到 hdfs的 /input下
docker exec namenode hdfs dfs -put /input_files/1.txt /input

查看浏览器中是否有指定文件

2. 运行wordCount

2.1 spark-shell运行

# 进入spark master容器内的spark-shell
docker exec -ti master spark-shell --executor-memory 1024M --total-executor-cores 2

# 执行wordCount
sc.textFile("hdfs://namenode:8020/input/1.txt").flatMap(line => line.split(" ")).map(word => (word, 1)).reduceByKey(_ + _).collect

2.2 spark-submit运行example(stand-alone)

这里使用官方默认的example jar运行

docker exec master \
bin/spark-submit \
--class org.apache.spark.examples.SparkPi \
--master 'spark://master:7077' \
--executor-memory 1G \
--total-executor-cores 2 \
./examples/jars/spark-examples_2.12-3.2.1.jar \
10

2.3 spark-submit运行example(yarn)

docker exec master \
bin/spark-submit \
--class org.apache.spark.examples.SparkPi \
--master yarn \
--deploy-mode cluster \
./examples/jars/spark-examples_2.12-3.2.1.jar \
10

2.4 spark-submit运行自定义的jar包(stand-alone)

2.4.1 自定义spark任务

package com.lzw.bigdata.spark.core.wordcount

import org.apache.spark.rdd.RDD
import org.apache.spark.{SparkConf, SparkContext}

object Spark03_WordCount {
  def main(args: Array[String]): Unit = {
    println("---------------start word_count----------------")

    // Spark框架步骤
    // 1. 建立和Spark框架的链接
    val sparkConfig: SparkConf = new SparkConf()
      //      .setMaster("local")
      .setAppName("WordCount")
    val ctx = new SparkContext(sparkConfig)

    println("----------------new SparkContext done---------------")


    // 2. 执行业务逻辑
    // 2.1 读取文件，获取一行一行的数据
    val inputPath: String = args(0)
    val lines: RDD[String] = ctx.textFile(inputPath)
    //    val lines: RDD[String] = ctx.textFile("hdfs://namenode:9000/input")

    println("----------------ctx.textFile done---------------")


    // 2.2 分词，此处按照空格spilt
    val words: RDD[String] = lines.flatMap(line => line.split(" "))

    val wordToOne: RDD[(String, Int)] = words.map(word => (word, 1))

    // Spark框架提供了更多的功能，可以将分组和聚合使用一个方法实现
    // 相同的key会对value做reduce
    val tuple: RDD[(String, Int)] = wordToOne.reduceByKey((t1, t2) => t1 + t2)
    val coll: Array[(String, Int)] = tuple.collect()
    coll.foreach(println)
    println(s"""----------------tuple.foreach(println) done len ${coll.length}---------------""")

    // 3. 关闭连接
    ctx.stop()

    println("----------------ctx.stop() done---------------")

  }
}

2.4.2 生成jar包

参考: https://www.cnblogs.com/moonlightml/p/9312192.html

2.4.3 复制到挂载的jars文件夹内

2.4.4 运行spark-submit

docker exec master \
bin/spark-submit \
--class com.lzw.bigdata.spark.core.wordcount.Spark03_WordCount \
--master 'spark://master:7077' \
--executor-memory 1G \
--total-executor-cores 2 \
/jars/spark_core.jar \
'hdfs://namenode:9000/input'

原文地址：http://www.cnblogs.com/baoshu/p/spark_3.html