Spark Streaming Overview

As an extension of Spark Core, Spark Streaming is used for high-throughput and fault-tolerant processing of continuous data. Currently supported external input sources include Kafka, Flume, HDFS/S3, Kinesis, Twitter, and TCP socket.

Spark Streaming abstracts continuous data into a Discretized Stream (DStream), which consists of a series of continuous resilient distributed datasets (RDDs). Each RDD contains data generated at a certain time interval. Processing DStream with functions is actually processing these RDDs.

When Spark Streaming is used as data input for Kafka, the following stable and experimental Kafka versions are supported:

Currently, the following versions of CKafka are supported: 0.9.0.x, 0.10.0.x, 0.10.1.x, and 0.10.2.x. The Kafka dependency of version 0.10.2.1 is used in this practice scenario.

In addition, Spark Streaming in EMR also supports direct connection to CKafka. For more information, please see Spark Streaming Connection to Ckafka.

Spark Streaming Connection to CKafka

Applying for a Ckafka instance

Log in to the CKafka Console and create a CKafka instance as instructed in Creating Instances.

Confirm whether the selected network type matches the currently used network.

Creating a topic

Create a topic under the instance as instructed in Creating Topics.
Private IP and port: the bootstrap-server to be used for production and consumption.
A topic named spark_test is created here and used as an example to describe how to produce and consume messages.

CVM environment

Centos 6.8 OS

Production to CKafka

Currently, the following versions of CKafka are supported: 0.9.0.x, 0.10.0.x, 0.10.1.x, and 0.10.2.x. The Kafka dependency of version 0.10.2.1 is used here.

1. Add dependencies to build.sbt:

   name := "Producer Example"
   version := "1.0"
   scalaVersion := "2.11.8"
   libraryDependencies += "org.apache.kafka" % "kafka-clients" % "0.10.2.1"

2. Configure producer_example.scala:

   import java.util.Properties
   import org.apache.kafka.clients.producer._
   object ProducerExample extends App {
    val  props = new Properties()
    props.put("bootstrap.servers", "172.16.16.12:9092") // Private IP and port in the instance information
   
    props.put("key.serializer", "org.apache.kafka.common.serialization.StringSerializer")
    props.put("value.serializer", "org.apache.kafka.common.serialization.StringSerializer")
   
    val producer = new KafkaProducer[String, String](props)
    val TOPIC="test"  // Specify the topic to produce to
    for(i<- 1 to 50){
            val record = new ProducerRecord(TOPIC, "key", s"hello $i") // Produce a message whose `key` is "key" and `value` is "hello i"
            producer.send(record)
    }
    val record = new ProducerRecord(TOPIC, "key", "the end "+new java.util.Date)
    producer.send(record)
    producer.close() // Disconnect at the end
   }

For more information on how to use ProducerRecord, please see ProducerRecord.

Consumption from CKafka

">

DirectStream

1. Add dependencies to build.sbt:

   name := "Consumer Example"
   version := "1.0"
   scalaVersion := "2.11.8"
   libraryDependencies += "org.apache.spark" %% "spark-core" % "2.1.0"
   libraryDependencies += "org.apache.spark" %% "spark-streaming" % "2.1.0"
   libraryDependencies += "org.apache.spark" %% "spark-streaming-kafka-0-10" % "2.1.0"

2. Configure DirectStream_example.scala:

   import org.apache.kafka.clients.consumer.ConsumerRecord
   import org.apache.kafka.common.serialization.StringDeserializer
   import org.apache.kafka.common.TopicPartition
   import org.apache.spark.streaming.kafka010._
   import org.apache.spark.streaming.kafka010.LocationStrategies.PreferConsistent
   import org.apache.spark.streaming.kafka010.ConsumerStrategies.Subscribe
   import org.apache.spark.streaming.kafka010.KafkaUtils
   import org.apache.spark.streaming.kafka010.OffsetRange
   import org.apache.spark.streaming.{Seconds, StreamingContext}
   import org.apache.spark.SparkConf
   import org.apache.spark.SparkContext
   import collection.JavaConversions._
   import Array._
   object Kafka {
    def main(args: Array[String]) {
        val kafkaParams = Map[String, Object](
            "bootstrap.servers" -> "172.16.16.12:9092",
            "key.deserializer" -> classOf[StringDeserializer],
            "value.deserializer" -> classOf[StringDeserializer],
            "group.id" -> "spark_stream_test1",
            "auto.offset.reset" -> "earliest",
            "enable.auto.commit" -> "false"
        )
   
        val sparkConf = new SparkConf()
        sparkConf.setMaster("local")
        sparkConf.setAppName("Kafka")
        val ssc = new StreamingContext(sparkConf, Seconds(5))
        val topics = Array("spark_test")
   
        val offsets : Map[TopicPartition, Long] = Map()
   
        for (i <- 0 until 3){
            val tp = new TopicPartition("spark_test", i)
            offsets.updated(tp , 0L)
        }
        val stream = KafkaUtils.createDirectStream[String, String](
            ssc,
            PreferConsistent,
            Subscribe[String, String](topics, kafkaParams)
        )
        println("directStream")
        stream.foreachRDD{ rdd=>
            // Output the obtained message
            rdd.foreach{iter =>
                val i = iter.value
                println(s"${i}")
            }
            // Get the offset
            val offsetRanges = rdd.asInstanceOf[HasOffsetRanges].offsetRanges
            rdd.foreachPartition { iter =>
                val o: OffsetRange = offsetRanges(TaskContext.get.partitionId)
                println(s"${o.topic} ${o.partition} ${o.fromOffset} ${o.untilOffset}")
            }
        }
   
        // Start the computation
        ssc.start()
        ssc.awaitTermination()
    }
   }

RDD

1. Configure build.sbt in the way as detailed here.
2. Configure RDD_example:

   import org.apache.kafka.clients.consumer.ConsumerRecord
   import org.apache.kafka.common.serialization.StringDeserializer
   import org.apache.spark.streaming.kafka010._
   import org.apache.spark.streaming.kafka010.LocationStrategies.PreferConsistent
   import org.apache.spark.streaming.kafka010.ConsumerStrategies.Subscribe
   import org.apache.spark.streaming.kafka010.KafkaUtils
   import org.apache.spark.streaming.kafka010.OffsetRange
   import org.apache.spark.streaming.{Seconds, StreamingContext}
   import org.apache.spark.SparkConf
   import org.apache.spark.SparkContext
   import collection.JavaConversions._
   import Array._
   object Kafka {
    def main(args: Array[String]) {
        val kafkaParams = Map[String, Object](
            "bootstrap.servers" -> "172.16.16.12:9092",
            "key.deserializer" -> classOf[StringDeserializer],
            "value.deserializer" -> classOf[StringDeserializer],
            "group.id" -> "spark_stream",
            "auto.offset.reset" -> "earliest",
            "enable.auto.commit" -> (false: java.lang.Boolean)
        )
        val sc = new SparkContext("local", "Kafka", new SparkConf())
        val java_kafkaParams : java.util.Map[String, Object] = kafkaParams
        // Pull messages in the corresponding offset range from the partition in order. The request will be blocked if no messages can be pulled, until the specified waiting time elapses or the number of produced new messages reaches the number for messages to be pulled
        val offsetRanges = Array[OffsetRange](
            OffsetRange("spark_test", 0, 0, 5),
            OffsetRange("spark_test", 1, 0, 5),
            OffsetRange("spark_test", 2, 0, 5)
        )
        val range = KafkaUtils.createRDD[String, String](
            sc,
            java_kafkaParams,
            offsetRanges,
            PreferConsistent
        )
        range.foreach(rdd=>println(rdd.value))
        sc.stop()
    }
   }

   For more information on how to use kafkaParams, please see kafkaParams.

Configuring the environment

Installing sbt

1. Download the sbt package from sbt's official website.

2. After decompression, create an sbt_run.sh script with the following content in the sbt directory and add executable permissions:

   #!/bin/bash
   SBT_OPTS="-Xms512M -Xmx1536M -Xss1M -XX:+CMSClassUnloadingEnabled -XX:MaxPermSize=256M"
   java $SBT_OPTS -jar `dirname $0`/bin/sbt-launch.jar "$@"

   chmod u+x ./sbt_run.sh

3. Run the following command:

   ./sbt-run.sh sbt-version

   The display of sbt version indicates a successful installation.

Installing Protobuf

1. Download an appropriate version of Protobuf.
2. Decompress and enter the directory.

   ./configure
   make && make install

   You should install gcc-g++ in advance, and the root permission may be required during installation.
3. Log in again and enter the following on the command line:

   protoc --version

4. The display of Protobuf version indicates a successful installation.

Installing Hadoop

1. Download the required version at Hadoop's official website.
2. Add a Hadoop user.

   useradd -m hadoop -s /bin/bash

3. Grant admin permissions.

   visudo

4. Add the following in a new line under root ALL=(ALL) ALL:
   hadoop ALL=(ALL) ALL
   Save and exit.
5. Use Hadoop for operations.

   su hadoop

6. Configure SSH password-free login.

   cd ~/.ssh/                     # If there is no such directory, please run `ssh localhost` first
   ssh-keygen -t rsa              # There will be prompts. Simply press Enter
   cat id_rsa.pub >> authorized_keys  # Add authorization
   chmod 600 ./authorized_keys    # Modify file permission

7. Install Java.

   sudo yum install java-1.8.0-openjdk java-1.8.0-openjdk-devel

8. Configure ${JAVA_HOME}.

   vim /etc/profile

   Add the following at the end:

   export JAVA_HOME=/usr/lib/jvm/java-1.8.0-openjdk-1.8.0.121-0.b13.el6_8.x86_64/jre
   export PATH=$PATH:$JAVA_HOME

   Modify the corresponding path based on the installation information.
9. Decompress Hadoop and enter the directory.

   ./bin/hadoop version

   The display of version information indicates a successful installation.
10. Configure the pseudo-distributed mode (so that you can build different forms of clusters as needed).

    vim /etc/profile

    Add the following at the end:

    export HADOOP_HOME=/usr/local/hadoop
    export PATH=$HADOOP_HOME/bin:$PATH

    Modify the corresponding path based on the installation information.
11. Modify /etc/hadoop/core-site.xml.

    <configuration>
    <property>
        <name>hadoop.tmp.dir</name>
        <value>file:/usr/local/hadoop/tmp</value>
        <description>Abase for other temporary directories.</description>
    </property>
    <property>
        <name>fs.defaultFS</name>
        <value>hdfs://localhost:9000</value>
    </property>
    </configuration>

12. Modify /etc/hadoop/hdfs-site.xml.

    <configuration>
    <property>
        <name>dfs.replication</name>
        <value>1</value>
    </property>
    <property>
        <name>dfs.namenode.name.dir</name>
        <value>file:/usr/local/hadoop/tmp/dfs/name</value>
    </property>
    <property>
        <name>dfs.datanode.data.dir</name>
        <value>file:/usr/local/hadoop/tmp/dfs/data</value>
    </property>
    </configuration>

13. Change JAVA_HOME in /etc/hadoop/hadoop-env.sh to the Java path.

    export JAVA_HOME=/usr/lib/jvm/java-1.8.0-openjdk-1.8.0.121-0.b13.el6_8.x86_64/jre

14. Format the NameNode.

    ./bin/hdfs namenode -format

    The display of Exitting with status 0 indicates a success.
15. Start Hadoop.

    ./sbin/start-dfs.sh

    NameNode, DataNode and SecondaryNameNode processes will exist upon successful startup.

Installing Spark

Download the required version at Spark's official website.
As Hadoop has already been installed, select Pre-build with user-provided Apache Hadoop here.
This example also uses the hadoop user for operations.

1. Decompress and enter the directory.
2. Modify the configuration file.

   cp ./conf/spark-env.sh.template ./conf/spark-env.sh
   vim ./conf/spark-env.sh

   Add the following in the first line:

   export SPARK_DIST_CLASSPATH=$(/usr/local/hadoop/bin/hadoop classpath)

   Modify the path based on the Hadoop installation information.
3. Run the example.

   bin/run-example SparkPi

   The display of an approximate value of π output by the program indicates a successful installation.