Ubuntu 14.10 下Spark on yarn安装

浏览数：19 / 时间：2015年06月20日

1 服务器分布

服务器	说明
192.168.1.100	NameNode
192.168.1.101	DataNode
192.168.1.102	DataNode

2 软件环境

　　2.1 安装JDK，添加环境变量

　　2.2 安装Scala，添加环境变量

　　2.3 SSH免密码登陆，A登陆A，A登陆B，可参考http://blog.csdn.net/codepeak/article/details/14447627

ssh-keygen -t rsa -P ‘‘
cat ~./ssh/id_rsa.pub >> ~/.ssh/authorized_keys

scp ~/.ssh/id_rsa.pub username@ipaddress:/location
cat id_rsa.pub >> authorized_keys

　　2.4 主机名设置

sudo nano /etc/hosts

192.168.1.100 cloud001
192.168.1.101 cloud002
192.168.1.102 cloud003

3 Hadoop集群配置（各个机器相同配置）

　　3.1 Hadoop环境安装，环境变量配置

export HADOOP_HOME=/home/hadoop/hadoop-2.2.0
export HADOOP_COMMON_HOME=$HADOOP_HOME
export HADOOP_HDFS_HOME=$HADOOP_HOME
export HADOOP_MAPRED_HOME=$HADOOP_HOME
export HADOOP_YARN_HOME=$HADOOP_HOME
export HADOOP_CONF_DIR=$HADOOP_HOME/etc/hadoop

export SCALA_HOME=/home/hadoop/software/spark/scala-2.11.4
export SPARK_EXAMPLES_JAR=/home/hadoop/software/spark/spark-1.0.0/examples/target/scala-2.11.4/spar$
export SPARK_HOME=/home/hadoop/software/spark/spark-1.0.0
export IDEA_HOME=/home/hadoop/software/dev/idea-IU-139.1117.1

export PATH=$PATH:$SCALA_HOME/bin:$SPARK_HOME/bin:$IDEA_HOME/bin:$HADOOP_HOME/bin:$HADOOP/sbin:$M2_HOME/bin

　　3.2 core.site.xml配置

<configuration>
    <property>
        <name>fs.defaultFS</name>
        <value>hdfs://cloud001:9000</value>
    </property>
    <property>
        <name>hadoop.tmp.dir</name>
        <value>/home/hadoop/hadoop-2.2.0/tmp</value>
    </property>
<!--    <property>
        <name>io.file.buffer.size</name>
        <value>131072</value>
    </property>
    <property>
        <name>hadoop.proxyuser.hadoop.hosts</name>
        <value>*</value>
    </property>
    <property>
        <name>hadoop.proxyuser.hadoop.groups</name>
        <value>*</value>
    </property>-->
</configuration>

　　3.3 hdfs-site.xml 配置

<configuration>
    <property>
        <name>dfs.namenode.secondary.http-address</name>
        <value>cloud001:9001</value>
    </property>
    <property>
        <name>dfs.namenode.name.dir</name>
        <value>file:/home/hadoop/hadoop-2.2.0/dfs/name</value>
    </property>
    <property>
        <name>dfs.datanode.data.dir</name>
        <value>file:/home/hadoop/hadoop-2.2.0/dfs/data</value>
    </property>
    <property>
        <name>dfs.replication</name>
        <value>3</value>
    </property>
    <property>
        <name>dfs.webhdfs.enabled</name>
        <value>true</value>
    </property>
</configuration>

　　3.4 mapred-site.xml 配置

<configuration>
    <property>
        <name>mapreduce.framework.name</name>
        <value>yarn</value>
    </property>
<!--    <property>
        <name>mapreduce.jobhistory.address</name>
        <value>hadoopmaster:10020</value>
    </property>
    <property>
        <name>mapreduce.jobhistory.webapp.address</name>
        <value>hadoopmaster:19888</value>
    </property>-->
</configuration>

　　3.5 yarn-site.xml 配置

<configuration>

<!-- Site specific YARN configuration properties -->
    <property>
        <name>yarn.nodemanager.aux-services</name>
        <value>mapreduce_shuffle</value>
    </property>
<!--    <property>
        <name>yarn.nodemanager.aux-services.mapreduce.shuffle.class</name>
        <value>org.apache.hadoop.mapred.ShuffleHandler</value>
    </property>-->
    <property>
        <name>yarn.resourcemanager.address</name>
        <value>cloud001:8032</value>
    </property>
    <property>
        <name>yarn.resourcemanager.scheduler.address</name>
        <value>cloud001:8030</value>
    </property>
    <property>
        <name>yarn.resourcemanager.resource-tracker.address</name>
        <value>cloud001:8031</value>
    </property>
<!--    <property>
        <name>yarn.resourcemanager.admin.address</name>
        <value>hadoopmaster:8033</value>
    </property>
    <property>
        <name>yarn.resourcemanager.webapp.address</name>
        <value>hadoopmaster:8088</value>
    </property> -->
</configuration>

　　3.6 配置hadoop-env.sh、mapred-env.sh、yarn-env.sh，在开头添加

export JAVA_HOME=/usr/lib/jvm/java-1.7.0-openjdk-amd64

　　3.7 数据节点配置

nano slaves
cloud002
cloud003

4 Spark集群配置（各个机器相同配置）

　　4.1 Spark安装部署

　　下载Spark二进制包，配置环境变量

export SCALA_HOME=/home/hadoop/software/spark/scala-2.11.4
export SPARK_EXAMPLES_JAR=/home/hadoop/software/spark/spark-1.0.0/examples/target/scala-2.11.4/spar$
export SPARK_HOME=/home/hadoop/software/spark/spark-1.0.0

　　配置spark-env.sh，添加如下

export JAVA_HOME=/usr/lib/jvm/java-1.7.0-openjdk-amd64
export SCALA_HOME=/home/hadoop/software/spark/scala-2.11.4
export HADOOP_HOME=/home/hadoop/hadoop-2.2.0

　　配置slaves

cloud002
cloud003

5 集群启动

　　5.1 格式化NameNode节点

hdfs namenode -format

　　5.2 启动Hadoop

sbin/start-all.sh

　　5.3 启动Spark

sbin/start-all.sh

6 测试

　　6.1 本地测试

# bin/run-exampleorg.apache.spark.examples.SparkPi local

　　6.2 普通集群测试

# bin/run-exampleorg.apache.spark.examples.SparkPi spark://cloud001:7077
# bin/run-exampleorg.apache.spark.examples.SparkLR spark://cloud001:7077
# bin/run-exampleorg.apache.spark.examples.SparkKMeans spark://cloud001:7077 file:/usr/local/spark/data/kmeans_data.txt 2 1

　　6.3 结合HDFS的集群模式

# hadoop fs -put README.md .
# MASTER=spark://cloud001:7077bin/spark-shell
scala> val file =sc.textFile("hdfs://cloud001:9000/user/root/README.md")
scala> val count = file.flatMap(line=> line.split(" ")).map(word => (word, 1)).reduceByKey(_+_)

scala> count.collect()

scala> :quit

　　6.4 基于YARN模式

#SPARK_JAR=assembly/target/scala-2.10/spark-assembly_2.10-0.9.0-incubating-hadoop2.2.0.jar

bin/spark-class org.apache.spark.deploy.yarn.Client 

--jar examples/target/scala-2.10/spark-examples_2.10-assembly-0.9.0-incubating.jar 

--class org.apache.spark.examples.SparkPi 

--args yarn-standalone 

--num-workers 3 

--master-memory 4g 

--worker-memory 2g 

--worker-cores 1