官网：
https://cwiki.apache.org/confluence/pages/viewpage.action?pageId=38571133
一、背景
栈的定义可以在源代码树中找到/ambari-server/src/main/resources/stacks。当你安装Ambari Server服务之后，栈的定义可以被发现/var/lib/ambari-server/resources/stacks。
二、结构
一个栈的结构定义如下
|_ stacks |_|_ metainfo.xml |_ hooks |_ repos repoinfo.xml |_ services |_ metainfo.xml metrics.json |_ configuration {configuration files} |_ package {files, scripts, templates}
三、定义一个服务和组件
在service里的metainfo.xml文件描述这个服务，服务的组件和管理脚本用于执行命令。一个组件的服务可以是MASTER,SLAVE或CLIENT类别。这个
Ambari支持不同的命令脚本用PYTHON编写的。类型是用来知道如何执行命令脚本。你也可以创建自定义命令除了default lifecycle commands之外你的组件需要去支持。
例如，YARN Service描述ResourceManager组件配置metainfo.xml如下：
<component> <name>RESOURCEMANAGERname> <category>MASTERcategory> <commandScript> <script>scripts/resourcemanager.pyscript> <scriptType>PYTHONscriptType> <timeout>600timeout> commandScript> <customCommands> <customCommand> <name>DECOMMISSIONname> <commandScript> <script>scripts/resourcemanager.pyscript> <scriptType>PYTHONscriptType> <timeout>600timeout> commandScript> customCommand> customCommands> component>
ResourceManager是一个MASTER组件，并且命令脚本是scripts/resourcemanager.py，可以被找到services/YARN/package目录，PYTHON命令脚本,脚本作为PYTHON方法实现default lifecycle commands。这是默认的安装方法安装命令：
class Resourcemanager(Script): def install(self, env): self.install_packages(env) self.configure(env)
你也可以看到一个自定义的命令定义DECOMMISSION,这意味着还有一个DECOMMISSION方法在python命令脚本:
def decommission(self, env): import params ... Execute(yarn_refresh_cmd, user=yarn_user ) pass
四、使用堆栈继承
栈可以扩展其他堆栈为了分享命令脚本和配置。这样可以减少重复的代码在栈使用以下：
·为子栈定义存储库
·在子栈添加新的服务（不是在父栈）
·覆盖父服务命令脚本
·覆盖父服务的配置
五、例子:实现一个自定义服务
在本例中,我们将创建一个名为“SAMPLESRV”的定制服务,将其添加到现有的栈的定义。
此服务包括MASTER，SLAVE，CLIENT组件。
创建并且添加服务
1.Ambari Server上，浏览到/var/lib/ambari-server/resources/stacks/HDP/2.0.6/services这个目录。在这种情况下，我们将浏览到HDP 2.0的定义。
cd /var/lib/ambari-server/resources/stacks/HDP/2.0.6/services
2.创建一个目录/var/lib/ambari-server/resources/stacks/HDP/2.0.6/services/SAMPLESRV将包含SAMPLESRV的服务定义。
mkdir /var/lib/ambari-server/resources/stacks/HDP/2.0.6/services/SAMPLESRV
cd /var/lib/ambari-server/resources/stacks/HDP/2.0.6/services/SAMPLESRV
3.浏览到新创建的SAMPLESRV目录，创建一个metainfo.xml文件去描述一个新的服务。例如：
xml version="1.0"?>
<metainfo>
<schemaVersion>2.0schemaVersion>
<services>
<service>
<name>SAMPLESRVname>
<displayName>New Sample ServicedisplayName>
<comment>A New Sample Servicecomment>
<version>1.0.0version>
<components>
<component>
<name>SAMPLESRV_MASTERname>
<displayName>Sample Srv MasterdisplayName>
<category>MASTERcategory>
<cardinality>1cardinality>
<commandScript>
<script>scripts/master.pyscript>
<scriptType>PYTHONscriptType>
<timeout>600timeout>
commandScript>
component>
<component>
<name>SAMPLESRV_SLAVEname>
<displayName>Sample Srv SlavedisplayName>
<category>SLAVEcategory>
<cardinality>1+cardinality>
<commandScript>
<script>scripts/slave.pyscript>
<scriptType>PYTHONscriptType>
<timeout>600timeout>
commandScript>
component>
<component>
<name>SAMPLESRV_CLIENTname>
<displayName>Sample Srv ClientdisplayName>
<category>CLIENTcategory>
<cardinality>1+cardinality>
<commandScript>
<script>scripts/sample_client.pyscript>
<scriptType>PYTHONscriptType>
<timeout>600timeout>
commandScript>
component>
components>
<osSpecifics>
<osSpecific>
<osFamily>anyosFamily>
osSpecific>
osSpecifics>
service>
services>
metainfo>
4.在上面，我的服务名是"SAMPLESRV"它包含：
·一个MASTER组件“SAMPLESRV_MASTER”
·一个SLAVE组件“SAMPLESRV_SLAVE”
·一个CLIENT组件“SAMPLESRV_CLIENT”
5.接下来，我们创建命令脚本。为命令脚本创建一个目录/var/lib/ambari-server/resources/stacks/HDP/2.0.6/services/SAMPLESRV/package/scripts我们制定的服务metainfo
1 mkdir -p /var/lib/ambari-server/resources/stacks/HDP/2.0.6/services/SAMPLESRV/package/scripts 2 cd /var/lib/ambari-server/resources/stacks/HDP/2.0.6/services/SAMPLESRV/package/scripts
6.进入目录并创建.py的命令脚本文件。
例如master.py文件：
import sys from resource_management import * class Master(Script): def install(self, env): print 'Install the Sample Srv Master'; def stop(self, env): print 'Stop the Sample Srv Master'; def start(self, env): print 'Start the Sample Srv Master'; def status(self, env): print 'Status of the Sample Srv Master'; def configure(self, env): print 'Configure the Sample Srv Master'; if __name__ == "__main__": Master().execute()
例如slave.py文件：
import sys from resource_management import * class Slave(Script): def install(self, env): print 'Install the Sample Srv Slave'; def stop(self, env): print 'Stop the Sample Srv Slave'; def start(self, env): print 'Start the Sample Srv Slave'; def status(self, env): print 'Status of the Sample Srv Slave'; def configure(self, env): print 'Configure the Sample Srv Slave'; if __name__ == "__main__": Slave().execute()
例如sample_client.py文件：
import sys from resource_management import * class SampleClient(Script): def install(self, env): print 'Install the Sample Srv Client'; def configure(self, env): print 'Configure the Sample Srv Client'; if __name__ == "__main__": SampleClient().execute()
7.重启Ambari Server
ambari-server restart
参考网址：
https://cwiki.apache.org/confluence/pages/viewpage.action?pageId=38571133
https://github.com/hortonworks-gallery/ambari-zeppelin-service#option-2-automated-deployment-of-a-fresh-hdp-cluster-that-includes-zeppelin-via-blueprints
https://github.com/Symantec/ambari-cassandra-service