Spark1.2集群环境搭建(Standalone+HA) 4G内存5个节点也是蛮拼的

准备工作：1、笔记本4G内存
，操作系统WIN7

2、工具VMware Workstation

3、虚拟机：CentOS6.4共五台

4、搭建好Hadoop集群( 方便Spark可从HDSF上读取文件，进行实验测试)

实验环境：Hadoop
HA集群：

[align=left]Ip[/align]	hostname	role
192.168.249.130	SY-0130	ActiveNameNode
192.168.249.131	SY-0131	StandByNameNode
192.168.249.132	SY-0132	DataNode1
192.168.249.133	SY-0133	DataNode2

Spark HA集群：

Ip	hostname	role
192.168.249.134	SY-0134	Master
192.168.249.130	SY-0130	StandBy Master
192.168.249.131	SY-0131	worker
192.168.249.132	SY-0132	worker
192.168.249.133	SY-0133	worker

实验环境仅作学习用，4G内存确实蛮拼的，资源非常有限。下周换上几台台式机作集群。

上述SY-0134是新克隆的虚拟机，作为Spark的环境中的Master，原属于Hadoop集群中的4个节点分别作为StandByMaster 和 Worker角色。

关于虚拟机环境设置、网络配置、Hadoop集群搭建参见《Hadoop2.6集群环境搭建》

本文重点关注Spark1.2环境、Zookeeper环境简易搭建，仅作学习与实验原型，且不涉及太多理论知识。

软件安装：(注：用户hadoop登录SY-0134)

1、在节点SY-0134，hadoop用户目录创建toolkit
文件夹，用来保存所有软件安装包，建立labsp文件作为本次实验环境目录。

[hadoop@SY-0134 ~]$ mkdir labsp

[hadoop@SY-0134~]$ mkdir toolkit

我将下载的软件包存放在toolkit中如下

1 2	[hadoop@SY-0134 toolkit]$ ls hadoop-2.5.2.tar.gz hadoop-2.6.0.tar.gz jdk-7u71-linux-i586.gz scala-2.10.3.tgz spark-1.2.0-bin-hadoop2.3.tgz zookeeper-3.4.6.tar.gz

2、这次实验我下载的Spark包是spark-1.2.0-bin-hadoop2.3.tgz
，Scala版本是2.10.3，Zookeeper是3.4.6。这里需要注意的是，Spark和Scala有版本对应关系，可在Spark官网介绍中找到Spark版本支持的Scala版本。

3、JDK安装及环境变量设置

[hadoop@SY-0134 ~]$ mkdir lab

#我将jdk7安装在lab目录

[hadoop@SY-0134 jdk1.7.0_71]$ pwd

/home/hadoop/lab/jdk1.7.0_71

#环境变量设置：

[hadoop@SY-0134 ~]$ vi .bash_profile

1 2 3 4 5 6 7 8

# User specific environment and startup programs export JAVA_HOME=/home/hadoop/lab/jdk1.7.0_71 PATH=$JAVA_HOME/bin PATH HOME/bin export PATH export CLASSPATH=. JAVA_HOME/lib/dt.jar JAVA_HOME/lib/tools.jar

#设置生效

[hadoop@SY-0130 ~]$ source .bash_profile

4、Scala安装及环境变量设置

我将scala解压到/home/hadoop/labsp/scala-2.10.3位置。

修改.bash_profile文件

增加：export SCALA_HOME=/home/hadoop/labsp/scala-2.10.3

修改：PATH=$JAVA_HOME/bin

PATH

HOME/bin

SCALA_HOME/bin

#设置生效

[hadoop@SY-0130 ~]$ source .bash_profile

检验Scala是否安装好：

[hadoop@SY-0134 ~]$ scala

Welcome to Scala version 2.10.3 (Java HotSpot(TM) Client VM, Java 1.7.0_71).

上述显示安装成功。

5、Spark安装及环境配置

我将spark解压到/home/hadoop/labsp/spark1.2_hadoop2.3位置。下载的这个包是预编译包。

修改.bash_profile文件

增加：export SPARK_HOME=/home/hadoop/labsp/spark1.2_hadoop2.3

修改：PATH=$JAVA_HOME/bin

PATH

HOME/bin

SCALA_HOME/bin:$SPARK_HOME/bin
#设置生效

[hadoop@SY-0130 ~]$ source .bash_profile

#修改spark-env.sh

[hadoop@SY-0134 conf]$ pwd

/home/hadoop/labsp/spark1.2_hadoop2.3/conf

[hadoop@SY-0134 conf]$vi spark-env.sh

核心配置：

export JAVA_HOME=/home/hadoop/lab/jdk1.7.0_71

export SCALA_HOME=/home/hadoop/labsp/scala-2.10.3

export SPARK_DAEMON_JAVA_OPTS="-Dspark.deploy.recoveryMode=ZOOKEEPER -Dspark.deploy.zookeeper.url=SY-0134:2181,SY-0130:2181,SY-0131:2181,SY-0132:2181,SY-0133:2181
-Dspark.deploy.zookeeper.dir=/spark"

至此JDK，Scala，Spark 安装及环境变量设置好，当然上述配置步骤也可一次修改完成。

6、Zookeeper安装

我将zookeeper解压到/home/hadoop/labsp/zookeeper-3.4.6位置。

#配置zoo.cfg文件

[hadoop@SY-0134 zookeeper-3.4.6]$ pwd

/home/hadoop/labsp/zookeeper-3.4.6

[hadoop@SY-0134 zookeeper-3.4.6]$ mkdir data

[hadoop@SY-0134 zookeeper-3.4.6]$ mkdir datalog

[hadoop@SY-0134 zookeeper-3.4.6]$ cd conf

[hadoop@SY-0134 conf]$ cp zoo_sample.cfg zoo.cfg

[hadoop@SY-0134 conf]$ vi zoo.cfg

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34

# The number of milliseconds of each tick tickTime=2000 # The number of ticks that the initial # synchronization phase can take initLimit=10 # The number of ticks that can pass between # sending a request and getting an acknowledgement syncLimit=5 # the directory where the snapshot is stored. # do not use /tmp for storage, /tmp here is just # example sakes. dataDir=/home/hadoop/labsp/zookeeper-3.4.6/data dataLogDir=/home/hadoop/labsp/zookeeper-3.4.6/datalog # the port at which the clients will connect clientPort=2181 # the maximum number of client connections. # increase this if you need to handle more clients #maxClientCnxns=60 # # Be sure to read the maintenance section of the # administrator guide before turning on autopurge. # # http://zookeeper.apache.org/doc/current/zookeeperAdmin.html#sc_maintenance # # The number of snapshots to retain in dataDir #autopurge.snapRetainCount=3 # Purge task interval in hours # Set to "0" to disable auto purge feature #autopurge.purgeInterval=1 server.1=SY-0134:2888:3888 server.2=SY-0130:2888:3888 server.3=SY-0131:2888:3888 server.4=SY-0132:2888:3888 server.5=SY-0133:2888:3888

#配置myid文件

[hadoop@SY-0134 data]$ pwd

/home/hadoop/labsp/zookeeper-3.4.6/data

输入1进入SY-0134的zookeeper中的myid文件

echo "1"> home/hadoop/labsp/zookeeper-3.4.6/data/myid

7、SSH免密码登录

虽然在Hadoop集群中，SY-0130，能够免密码登录到SY-0131，SY-0132，SY-0133 。

但是在本次Spark集群中，Master为SY-0134 ，他需要能够免密码登录到SY-0130，SY-0131，SY-0132，SY-0133。

#我是先在SY-0134中，生成公钥。

[hadoop@SY-0134 ~]$ ssh-keygen -t rsa

[hadoop@SY-0134 ~]$ cd .ssh

[hadoop@SY-0134 .ssh]$ ls

id_rsa id_rsa.pub known_hosts

#将id_rsa.pub文件拷贝给SY-0130

[hadoop@SY-0134 .ssh]$ scp id_rsa.pub hadoop@SY-0130:~/.ssh/authorized_keys

#在SY-0130中，生成公钥。

[hadoop@SY-0130 ~]$ ssh-keygen -t rsa

[hadoop@SY-0130 ~]$ cd .ssh

[hadoop@SY-0130 .ssh]$ ls

authorized_keys id_rsa id_rsa.pub known_hosts

#将id_rsa.pub文件的内容追加写入到authorized_keys中。稍微有点特殊。

[hadoop@SY-0130 .ssh]$ cat id_rsa.pub >>authorized_keys

#将SY-0130下的authorized_keys文件使用SCP命令复制到SY-0131，SY-0132，SY-0133 。

8、其他节点Spark，Scala,
zookeeper安装

上述7步仅完成了SY-0134 ，Spark，Scala, Zookeeper的安装，须将三个安装文件目录SCP命令拷贝到SY-0130,SY-0131,SY-132,SY-0133目录，并且同样设置环境变量。

[hadoop@SY-0134 labsp]$ ls

scala-2.10.3 spark1.2_hadoop2.3 zookeeper-3.4.6

另外一点，Zoookeeper 的Server在不同节点上，myid文件内容不一样。

echo "1"> home/hadoop/labsp/zookeeper-3.4.6/data/myid #SY-0134

echo "2"> home/hadoop/labsp/zookeeper-3.4.6/data/myid #SY-0130

echo "3"> home/hadoop/labsp/zookeeper-3.4.6/data/myid #SY-0131

echo "4"> home/hadoop/labsp/zookeeper-3.4.6/data/myid #SY-0132

echo "5"> home/hadoop/labsp/zookeeper-3.4.6/data/myid #SY-0133

集群启动测试：1、在5个节点上分别启动zookeeper
.

[hadoop@SY-0134 zookeeper-3.4.6]$ bin/zkServer.sh start

[hadoop@SY-0130 zookeeper-3.4.6]$ bin/zkServer.sh start

[hadoop@SY-0131 zookeeper-3.4.6]$ bin/zkServer.sh start

[hadoop@SY-0132 zookeeper-3.4.6]$ bin/zkServer.sh start

[hadoop@SY-0133 zookeeper-3.4.6]$ bin/zkServer.sh start

2、在SY-0134启动
Spark Master

1 2 3 4 5

[hadoop@SY-0134 spark1.2_hadoop2.3]$ sbin/start-all.sh starting org.apache.spark.deploy.master.Master, logging to /home/hadoop/labsp/spark1.2_hadoop2.3/sbin/../logs/spark-hadoop-org.apache.spark.deploy.master.Master-1-SY-0134.out SY-0133: starting org.apache.spark.deploy.worker.Worker, logging to /home/hadoop/labsp/spark1.2_hadoop2.3/sbin/../logs/spark-hadoop-org.apache.spark.deploy.worker.Worker-1-SY-0133.out SY-0132: starting org.apache.spark.deploy.worker.Worker, logging to /home/hadoop/labsp/spark1.2_hadoop2.3/sbin/../logs/spark-hadoop-org.apache.spark.deploy.worker.Worker-1-SY-0132.out SY-0131: starting org.apache.spark.deploy.worker.Worker, logging to /home/hadoop/labsp/spark1.2_hadoop2.3/sbin/../logs/spark-hadoop-org.apache.spark.deploy.worker.Worker-1-SY-0131.out

3、在SY-0130启动
Standby Spark Master

[hadoop@SY-0130 spark1.2_hadoop2.3]$ sbin/start-master.sh

starting org.apache.spark.deploy.master.Master, logging to /lab/labsp/spark1.2_hadoop2.3/sbin/../logs/spark-hadoop-org.apache.spark.deploy.master.Master-1-SY-0130.out





有了这样的实验环境，就可以继续深入学习Spark运行架构、SparkSQL等知识啦.

转自：http://www.cnblogs.com/xiejin/p/4213082.html