在Linux上怎么安装和配置Apache Samza

samza是一个分布式的流式数据处理框架（streaming processing），它是基于Kafka消息队列来实现类实时的流式数据处理的。(准确的说，samza是通过模块化的形式来使用kafka的，因此可以构架在其他消息队列框架上，但出发点和默认实现是基于kafka)

Apache Kafka主要是用来控制发消息的

Apache Hadoop YARN会提供错误信息，隔离处理器，安全和资源管理.

本文将介绍怎么在 Ubuntu 14.04 的32位 系统上安装Samza.

安装准备:

要安装和配置Apache-Samza，需要以下东西

JDK 1.7
maven2
kafka
yarn
zookeeper

#  apt-get install curl gem

下载并设置JDK路径:

我们需要安装JDK并设置好其环境变量.

# cd /usr/java # wget --no-cookies --no-check-certificate --header "Cookie: gpw_e24=http%3A%2F%2Fwww.oracle.com%2F; oraclelicense=accept-securebackup-cookie" "http://download.oracle.com/otn-pub/java/jdk/7u79-b15/jdk-7u79-linux-i586.tar.gz" # tar xzf jdk-7u79-linux-i586.tar.gz

解压并设置好JAVA_HOME路径

# tar -zxvf  jdk-7u79-linux-i586.tar.gz# JAVA_HOME=/usr/java/jdk1.7.0_79# export JAVA_HOME# PATH=$JAVA_HOME/bin:$PATH# export PATH

把上面的加入到 ~/.bashrc 和 /etc/bashrc文件去

安装Maven2:

接下来下载安装maven

#  wget https://launchpad.net/~bneijt/+archive/ubuntu/ppa/+build/2139203/+files/maven3_3.0.1-0~ppa2_all.deb

# dpkg -i maven3_3.0.1-0~ppa2_all.deb

检查maven版本好

# mvn3 -version

Apache Maven 3.0.1 (r1038046; 2010-11-23 16:28:32+0530)
Java version: 1.7.0_79
Java home: /usr/java/jdk1.7.0_79/jre
Default locale: en_IN, platform encoding: UTF-8
OS name: "linux" version: "3.8.0-29-generic" arch: "i386" Family: "unix"

安装Hello-Samza : 


我们就按照在 /usr/local 文件夹下面把

# cd /usr/local

把hello-samza复制进来,

# git clone git://git.apache.org/samza-hello-samza.git hello-samza

本项目中含有一个"grid"的脚本，其中有hello-samza变量，有了这个你可以搞定一切了. 使用它可以安装 Kafka, Yarn和Zookeeper.

执行下面的命令,

# cd /usr/local/hello-samza

root@dev:/usr/local/hello-samza# bin/grid install kafka 

EXECUTING: install kafka
Downloading kafka_2.10-0.8.2.1.tgz...
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
 15 15.4M   15 2406k    0     0   304k      0  0:00:51  0:00:07  0:00:44  443k

root@dev:/usr/local/hello-samza# bin/grid install yarn

EXECUTING: install yarn
Downloading hadoop-2.6.1.tar.gz...
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
 77  187M   77  145M    0     0   239k      0  0:13:23  0:10:22  0:03:01  204k

root@dev:/usr/local/hello-samza# bin/grid install zookeeper

EXECUTING: install zookeeper
Downloading zookeeper-3.4.3.tar.gz...
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
  8 15.4M    8 1324k    0     0   212k      0  0:01:14  0:00:06  0:01:08  266k

现在你会发现所有的包都在hello-samza根目录下面的一个名字叫 “deploy”文件夹里面.

root@dev:/usr/local/hello-samza# cd deployroot@dev:/usr/local/hello-samza/deploy# ls 

kafka  yarn  zookeeper

执行bin/grid bootstrap命令

root@dev:/usr/local/hello-samza# bin/grid bootstrap 

Download http://repo1.maven.org/maven2/org/fusesource/scalate/scalate-util_2.10/1.6.1/scalate-util_2.10-1.6.1.jar
:samza-yarn_2.10:processResources
:samza-yarn_2.10:classes
:samza-yarn_2.10:lesscss
....
....
BUILD SUCCESSFUL

Total time: 20 mins 32.855 secs
/usr/local/hello-samza
EXECUTING: install zookeeper
Using previously downloaded file /root/.samza/download/zookeeper-3.4.3.tar.gz
EXECUTING: install yarn
Using previously downloaded file /root/.samza/download/hadoop-2.6.1.tar.gz
EXECUTING: install kafka
Using previously downloaded file /root/.samza/download/kafka_2.10-0.8.2.1.tgz
EXECUTING: start zookeeper
JMX enabled by default
Using config: /usr/local/hello-samza/deploy/zookeeper/bin/../conf/zoo.cfg
Starting zookeeper ... STARTED
EXECUTING: start yarn
starting resourcemanager, logging to /usr/local/hello-samza/deploy/yarn/logs/yarn-root-resourcemanager-dev.out
starting nodemanager, logging to /usr/local/hello-samza/deploy/yarn/logs/yarn-root-nodemanager-dev.out
EXECUTING: start kafka
 

上面的grid执行完后,你就可以验证YARN是否安装好了并在运行，访问URL http://localhost:8088. 看到的就是YARN UI界面.

Build一个Samza工作包:

你需要build下这个包，YARN就是通过这个包来执行grid的. 

注: 比如你build的是hello-samza项目的最新版的话，记得首先执行下下面的命令。

root@dev:/usr/local/hello-samza#./gradlew publishToMavenLocal 

你可以在hello-samza项目中使用这些命令:

root@dev:/usr/local/hello-samza# mvn clean packageroot@dev:/usr/local/hello-samza# mkdir -p deploy/samzaroot@dev:/usr/local/hello-samza# tar -xvf ./target/hello-samza-0.10.0-dist.tar.gz -C deploy/samza

执行Samza任务:

完成build Samza包之后，你就可以在grid使用t run-job.sh 脚本来完成一些任务了

root@dev:/usr/local/hello-samza # deploy/samza/bin/run-job.sh --config-factory=org.apache.samza.config.factories.PropertiesConfigFactory --config-path=file://$PWD/deploy/samza/config/wikipedia-feed.properties 

 

上面的这个任务将会从Wikipedia上把实施反馈编辑撤销掉，会把这些编辑放到一个叫“thelinuxfaq-raw”的主题里面去. 

让这个主题运行几分钟后，你再来看下Kafka 最后面的更新情况:

root@dev:/usr/local/hello-samza#  deploy/kafka/bin/kafka-console-consumer.sh  --zookeeper localhost:2181 --topic thelinuxfaq-raw

再次访问YARN UI界面(http://localhost:8088). 你就看到Samza很正常的运行而不是有错误提示了! 

关闭Samza:

一切都弄好了,你就可以使用grid脚本关闭所有的相关服务器了.

root@dev:/usr/local/hello-samza #  bin/grid stop all 

输出示例:

EXECUTING: stop allEXECUTING: stop kafkaEXECUTING: stop yarnstopping resourcemanagerstopping nodemanagerEXECUTING: stop zookeeperJMX enabled by defaultUsing config: /usr/local/hello-samza/deploy/zookeeper/bin/../conf/zoo.cfgStopping zookeeper ... STOPPED

启动Samza :

同意的,你可以使用grid脚本来启动所有服务,

root@dev:/usr/local/hello-samza #  bin/grid start all 

输出示例:

EXECUTING: start allEXECUTING: start zookeeperJMX enabled by defaultUsing config: /usr/local/hello-samza/deploy/zookeeper/bin/../conf/zoo.cfgStarting zookeeper ... STARTEDEXECUTING: start yarn....EXECUTING: start kafka

 转自：在Linux上怎么安装和配置Apache Samza

原标题：在Linux上怎么安装和配置Apache Samza

关键词：linux