如何在Ubuntu下搭建Spark集群

Posted 2023-03-30

tags:

篇首语：本文由小常识网(cha138.com)小编为大家整理，主要介绍了如何在Ubuntu下搭建Spark集群相关的知识，希望对你有一定的参考价值。

你好，
安装其实就是解压，配置 /etc/profile环境变量
export SPARK_HOME=/data1/spark/spark
export SCALA_HOME=/data1/spark/scala-2.9.3

export PATH=$PATH:$SPARK_HOME/bin:$SCALA_HOME/bin

配置spark的conf下的spark-env.sh
export JAVA_HOME=/usr/java/default
export SCALA_HOME=/data1/spark/scala-2.9.3
export SPARK_MASTER_IP=192.168.0.1

export SPARK_MASTER_WEBUI_PORT=8080
export SPARK_WORKER_WEBUI_PORT=8000

export YARN_CONF_DIR=/data/hadoop/hadoop-2.0/etc/hadoop

配置slaves(ip根据需要修改)
192.168.0.2
192.168.0.3
分发spark目录和scala目录到几台服务器相同路径下
4,启动

进入主节点的spark目录的bin下stop-all.sh是停掉集群，start-all.sh启动集群，jps可以在主节点看到master进程，slave节点看到worker进程
5，运行程序，运行例子进入spark目录下
分布式运行
./run-example org.apache.spark.examples.SparkPi spark://192.168.0.1:7077

./run-example org.apache.spark.examples.SparkLR spark://192.168.0.1:7077
本地运行
./run-example org.apache.spark.examples.SparkPi local

./run-example org.apache.spark.examples.SparkLR local 参考技术A 安装其实就是解压，配置 /etc/profile环境变量
export SPARK_HOME=/data1/spark/spark
export SCALA_HOME=/data1/spark/scala-2.9.3
export PATH=$PATH:$SPARK_HOME/bin:$SCALA_HOME/bin
配置spark的conf下的spark-env.sh
export JAVA_HOME=/usr/java/default
export SCALA_HOME=/data1/spark/scala-2.9.3
export SPARK_MASTER_IP=192.168.0.1

export SPARK_MASTER_WEBUI_PORT=8080
export SPARK_WORKER_WEBUI_PORT=8000
export YARN_CONF_DIR=/data/hadoop/hadoop-2.0/etc/hadoop
配置slaves(ip根据需要修改)
192.168.0.2
192.168.0.3
分发spark目录和scala目录到几台服务器相同路径下
4,启动
进入主节点的spark目录的bin下stop-all.sh是停掉集群，start-all.sh启动集群，jps可以在主节点看到master进程，slave节点看到worker进程
5，运行程序，运行例子进入spark目录下
分布式运行
./run-example org.apache.spark.examples.SparkPi spark://192.168.0.1:7077
./run-example org.apache.spark.examples.SparkLR spark://192.168.0.1:7077
本地运行
./run-example org.apache.spark.examples.SparkPi local
./run-example org.apache.spark.examples.SparkLR local 参考技术B 安装其实就是解压，配置 /etc/profile环境变量
export SPARK_HOME=/data1/spark/spark
export SCALA_HOME=/data1/spark/scala-2.9.3

export PATH=$PATH:$SPARK_HOME/bin:$SCALA_HOME/bin

配置spark的conf下的spark-env.sh
export JAVA_HOME=/usr/java/default
export SCALA_HOME=/data1/spark/scala-2.9.3
export SPARK_MASTER_IP=192.168.0.1

export SPARK_MASTER_WEBUI_PORT=8080
export SPARK_WORKER_WEBUI_PORT=8000

export YARN_CONF_DIR=/data/hadoop/hadoop-2.0/etc/hadoop

配置slaves(ip根据需要修改)
192.168.0.2
192.168.0.3
分发spark目录和scala目录到几台服务器相同路径下
4,启动

进入主节点的spark目录的bin下stop-all.sh是停掉集群，start-all.sh启动集群，jps可以在主节点看到master进程，slave节点看到worker进程
5，运行程序，运行例子进入spark目录下
分布式运行
./run-example org.apache.spark.examples.SparkPi spark://192.168.0.1:7077

./run-example org.apache.spark.examples.SparkLR spark://192.168.0.1:7077
本地运行
./run-example org.apache.spark.examples.SparkPi local

./run-example org.apache.spark.examples.SparkLR local

Hadoop+Spark:集群环境搭建

环境准备：

在虚拟机下，大家三台Linux ubuntu 14.04 server x64 系统（下载地址：http://releases.ubuntu.com/14.04.2/ubuntu-14.04.2-server-amd64.iso）:

192.168.1.200 master

192.168.1.201 node1

192.168.1.202 node2

在Master上安装Spark环境：

Spark集群环境搭建：

搭建hadoop集群使用hadoop版本是hadoop2.6.4(这里hadoop我已经安装完成，具体如何安装hadoop具体请参考我的文章：《Hadoop:搭建hadoop集群》)

搭建spark这里使用spark版本是spark1.6.2（spark-1.6.2-bin-hadoop2.6.tgz）

1、下载安装包到master虚拟服务器：

在线下载：

hadoop@master:~/ wget http://mirror.bit.edu.cn/apache/spark/spark-1.6.2/spark-1.6.2-bin-hadoop2.6.tgz

离线上传到集群：

2、解压spark安装包到master虚拟服务器/usr/local/spark下，并分配权限：

#解压到/usr/local/下
hadoop@master:~$ sudo tar -zxvf spark-1.6.2-bin-hadoop2.6.tgz -C /usr/local/
hadoop@master:~$ cd /usr/local/
hadoop@master:/usr/local$ ls
bin  games   include  man   share                      src
etc  hadoop  lib      sbin  spark-1.6.2-bin-hadoop2.6
#重命名 为spark
hadoop@master:/usr/local$ sudo mv spark-1.6.2-bin-hadoop2.6/ spark/
hadoop@master:/usr/local$ ls
bin  etc  games  hadoop  include  lib  man  sbin  share  spark  src
#分配权限
hadoop@master:/usr/local$ sudo chown -R hadoop:hadoop spark
hadoop@master:/usr/local$

3、在master虚拟服务器/etc/profile中添加Spark环境变量：

　编辑/etc/profile文件

sudo vim /etc/profile

在尾部添加$SPARK_HOME变量，添加后，目前我的/etc/profile文件尾部内容如下：

export JAVA_HOME=/usr/lib/jvm/java-8-oracle
export JRE_HOME=/usr/lib/jvm/java-8-oracle
export SCALA_HOME=/opt/scala/scala-2.10.5
# add hadoop bin/ directory to PATH
export HADOOP_HOME=/usr/local/hadoop
export SPARK_HOME=/usr/local/spark
export PATH=$JAVA_HOME/bin:$JAVA_HOME/jre/bin:$JAVA_HOME:$HADOOP_HOME/bin:$HADOOP_HOME/sbin:$SCALA_HOME/bin:$SPARK_HOME/bin:$PATH
export CLASSPATH=$CLASS_PATH::$JAVA_HOME/lib:$JAVA_HOME/jre/lib

生效：

source /etc/profile

在Master配置Spark：

1、配置master虚拟服务器hadoop-env.sh文件：

sudo vim /usr/local/spark/conf/hadoop-env.sh

注意：默认情况下没有hadoop-env.sh和slaves文件，而是*.template文件，需要重命名：

hadoop@master:/usr/local/spark/conf$ ls
docker.properties.template  metrics.properties.template   spark-env.sh
fairscheduler.xml.template  slaves.template
log4j.properties.template   spark-defaults.conf.template
hadoop@master:/usr/local/spark/conf$ sudo vim spark-env.sh
hadoop@master:/usr/local/spark/conf$ mv  slaves.template slaves

在文件末尾追加如下内容：

export STANDALONE_SPARK_MASTER_HOST=192.168.1.200
export SPARK_MASTER_IP=192.168.1.200
export SPARK_WORKER_CORES=1
#every slave node start work instance count
export SPARK_WORKER_INSTANCES=1
export SPARK_MASTER_PORT=7077
export SPARK_WORKER_MEMORY=1g
export MASTER=spark://${SPARK_MASTER_IP}:${SPARK_MASTER_PORT}
export SCALA_HOME=/opt/scala/scala-2.10.5
export JAVA_HOME=/usr/lib/jvm/java-8-oracle
export SPARK_HISTORY_OPTS="-Dspark.history.fs.logDirectory=hdfs://192.168.1.200:9000/SparkEventLog"
export SPARK_WORKDER_OPTS="-Dspark.worker.cleanup.enabled=true"
export HADOOP_CONF_DIR=/usr/local/hadoop/etc/hadoop

2、配置master虚拟服务器下slaves文件：

sudo vim /usr/local/spark/conf/slaves

在slaves文件中内容如下：

192.168.1.201
192.168.1.202

注意：每行写一个机器的ip。

3、Master虚拟机下/usr/local/spark/目录下创建logs文件夹，并分配777权限：

hadoop@master:/usr/local/spark$ mkdir logs
hadoop@master:/usr/local/spark$ chmod 777 logs

复制Master虚拟服务器上的/usr/loca/spark下文件到所有slaves节点（node1、node2）下：

1、复制Master虚拟服务器上/usr/local/spark/安装文件到各个salves(node1、node2)上：

注意：拷贝钱需要在ssh到所有salves节点（node1、node2）上，创建/usr/local/spark/目录，并分配777权限。

hadoop@master:/usr/local/spark/conf$ cd ~/
hadoop@master:~$ sudo chmod 777 /usr/local/spark
hadoop@master:~$ scp -r /usr/local/spark hadoop@node1:/usr/local
scp: /usr/local/spark: Permission denied
hadoop@master:~$ sudo scp -r /usr/local/spark hadoop@node1:/usr/local
hadoop@node1\'s password:
scp: /usr/local/spark: Permission denied
hadoop@master:~$ sudo chmod 777 /usr/local/spark
hadoop@master:~$ ssh node1
Welcome to Ubuntu 14.04.2 LTS (GNU/Linux 3.16.0-30-generic x86_64)

 * Documentation:  https://help.ubuntu.com/

  System information as of Fri Sep 23 16:40:31 UTC 2016

  System load:  0.08               Processes:           400
  Usage of /:   12.2% of 17.34GB   Users logged in:     0
  Memory usage: 5%                 IP address for eth0: 192.168.1.201
  Swap usage:   0%

  Graph this data and manage this system at:
    https://landscape.canonical.com/

New release \'16.04.1 LTS\' available.
Run \'do-release-upgrade\' to upgrade to it.

Last login: Wed Sep 21 16:19:25 2016 from master
hadoop@node1:~$ cd /usr/local/
hadoop@node1:/usr/local$ sudo mkdir spark
[sudo] password for hadoop:
hadoop@node1:/usr/local$ ls
bin  etc  games  hadoop  include  lib  man  sbin  share  spark  src
hadoop@node1:/usr/local$ sudo chmod 777 ./spark
hadoop@node1:/usr/local$ exit
hadoop@master:~$ scp -r /usr/local/spark hadoop@node1:/usr/local
...........
hadoop@master:~$ ssh node2
Welcome to Ubuntu 14.04.2 LTS (GNU/Linux 3.16.0-30-generic x86_64)

 * Documentation:  https://help.ubuntu.com/

  System information as of Fri Sep 23 16:15:03 UTC 2016

  System load:  0.08               Processes:           435
  Usage of /:   13.0% of 17.34GB   Users logged in:     0
  Memory usage: 6%                 IP address for eth0: 192.168.1.202
  Swap usage:   0%

  Graph this data and manage this system at:
    https://landscape.canonical.com/

Last login: Wed Sep 21 16:19:47 2016 from master
hadoop@node2:~$ cd /usr/local
hadoop@node2:/usr/local$ sudo mkdir spark
[sudo] password for hadoop:
hadoop@node2:/usr/local$ sudo chmod 777 ./spark
hadoop@node2:/usr/local$ exit
logout
Connection to node2 closed.
hadoop@master:~$ scp -r /usr/local/spark hadoop@node2:/usr/local
...........

2、修改所有salves节点(node1、node2)上/etc/profile，追加$SPARK_HOME环境变量:

注意：一般都会遇到权限问题。最好登录到各个salves节点(node1、node2)上手动编辑/etc/profile。

hadoop@master:~$ ssh node1
Welcome to Ubuntu 14.04.2 LTS (GNU/Linux 3.16.0-30-generic x86_64)

 * Documentation:  https://help.ubuntu.com/

  System information as of Fri Sep 23 16:42:44 UTC 2016

  System load:  0.01               Processes:           400
  Usage of /:   12.2% of 17.34GB   Users logged in:     0
  Memory usage: 5%                 IP address for eth0: 192.168.1.201
  Swap usage:   0%

  Graph this data and manage this system at:
    https://landscape.canonical.com/

New release \'16.04.1 LTS\' available.
Run \'do-release-upgrade\' to upgrade to it.

Last login: Fri Sep 23 16:40:52 2016 from master
hadoop@node1:~$ sudo vim /etc/profile
[sudo] password for hadoop:
hadoop@node1:~$ exit
logout
Connection to node1 closed.
hadoop@master:~$ ssh node2
Welcome to Ubuntu 14.04.2 LTS (GNU/Linux 3.16.0-30-generic x86_64)

 * Documentation:  https://help.ubuntu.com/

  System information as of Fri Sep 23 16:44:42 UTC 2016

  System load:  0.0                Processes:           400
  Usage of /:   13.0% of 17.34GB   Users logged in:     0
  Memory usage: 5%                 IP address for eth0: 192.168.1.202
  Swap usage:   0%

  Graph this data and manage this system at:
    https://landscape.canonical.com/

New release \'16.04.1 LTS\' available.
Run \'do-release-upgrade\' to upgrade to it.

Last login: Fri Sep 23 16:43:31 2016 from master
hadoop@node2:~$ sudo vim /etc/profile
[sudo] password for hadoop:
hadoop@node2:~$ exit
logout
Connection to node2 closed.
hadoop@master:~$

修改后的所有salves上/etc/profile文件与master节点上/etc/profile文件配置一致。

在Master启动spark并验证是否配置成功：

1、启动命令：

一般要确保hadoop已经启动，之后才启动spark

hadoop@master:~$ cd /usr/local/spark/
hadoop@master:/usr/local/spark$ ./sbin/start-all.sh

2、验证是否启动成功：

方法一、jps

hadoop@master:/usr/local/spark$ ./sbin/start-all.sh
starting org.apache.spark.deploy.master.Master, logging to /usr/local/spark/logs/spark-hadoop-org.apache.spark.deploy.master.Master-1-master.out
192.168.1.201: starting org.apache.spark.deploy.worker.Worker, logging to /usr/local/spark/logs/spark-hadoop-org.apache.spark.deploy.worker.Worker-1-node1.out
192.168.1.202: starting org.apache.spark.deploy.worker.Worker, logging to /usr/local/spark/logs/spark-hadoop-org.apache.spark.deploy.worker.Worker-1-node2.out
hadoop@master:/usr/local/spark$ jps
1650 NameNode
1875 SecondaryNameNode
3494 Jps
2025 ResourceManager
3423 Master
hadoop@master:/usr/local/spark$ cd ~/
hadoop@master:~$ ssh node1
Welcome to Ubuntu 14.04.2 LTS (GNU/Linux 3.16.0-30-generic x86_64)

 * Documentation:  https://help.ubuntu.com/

  System information as of Fri Sep 23 17:33:40 UTC 2016

  System load:  0.06               Processes:           402
  Usage of /:   13.9% of 17.34GB   Users logged in:     0
  Memory usage: 21%                IP address for eth0: 192.168.1.201
  Swap usage:   0%

  Graph this data and manage this system at:
    https://landscape.canonical.com/

New release \'16.04.1 LTS\' available.
Run \'do-release-upgrade\' to upgrade to it.

Last login: Fri Sep 23 17:33:10 2016 from master
hadoop@node1:~$ jps
1392 DataNode
2449 Jps
2330 Worker
2079 NodeManager
hadoop@node1:~$ exit
logout
Connection to node1 closed.
hadoop@master:~$ ssh node2
Welcome to Ubuntu 14.04.2 LTS (GNU/Linux 3.16.0-30-generic x86_64)

 * Documentation:  https://help.ubuntu.com/

  System information as of Fri Sep 23 17:33:40 UTC 2016

  System load:  0.07               Processes:           404
  Usage of /:   14.7% of 17.34GB   Users logged in:     0
  Memory usage: 20%                IP address for eth0: 192.168.1.202
  Swap usage:   0%

  Graph this data and manage this system at:
    https://landscape.canonical.com/

New release \'16.04.1 LTS\' available.
Run \'do-release-upgrade\' to upgrade to it.

Last login: Fri Sep 23 16:51:36 2016 from master
hadoop@node2:~$ jps
2264 Worker
2090 NodeManager
1402 DataNode
2383 Jps
hadoop@node2:~$

方法二、web方式http://192.168.1.200:8080看是否正常:

以上是关于如何在Ubuntu下搭建Spark集群的主要内容，如果未能解决你的问题，请参考以下文章

在Ubuntu下搭建Spark群集

Hadoop+Spark:集群环境搭建

如何基于Jupyter notebook搭建Spark集群开发环境

spark学习--ubuntu14.04集群搭建配置（jdk）

使用 Docker 搭建 Hadoop 集群和 Spark On Yarn

spark集群详细搭建过程及遇到的问题解决