Hadoop cluster deployment by Ambari
1. Open http://192.168.242.181:8080 login username / password: admin / admin
2. Click on "LAUNCH INSTALL WIZARD", start creating a cluster
3. a name for the cluster
4. Earlier we built a local repository, where select "Use Local Repository". Delete other OS, leaving only redhat7 the line. And fill in the corresponding address in front of a web service built in the Base URL that column.
5. In the "Target Hosts" fill in hadoop cluster which need to be deployed to the machine. You can fill IP.
SSH Private Key in the selected file is zero-based installation Ambari (1) - Preparation before installation in the landing configuration secret to avoid that machine to other machines of id_rsa this file (ambari host). I use the root account, so this file is in /root/.ssh/ directory.
6. "Confirm Hosts" this step, ambari ambari agent will be installed in the configuration of the above hosts, just to wait.
7. If necessary, select the service. If some services depend on other services, without dependent service choice, when you click "Next", will do the appropriate prompt.
8. Click "Next", similar to the following warnings if you can not control, follow-up, if necessary, can install
9. "Assign Masters" this step, ambari automatically assigned to various services on different machines. You can be manually adjusted.
10. "Assign Slaves and Clients" This step, DataNode node position of the manual distribution of Hadoop, NodeManager the position ...... YARN
其中 NFSGateway 是通过挂载的方式，像访问本地文件系统一样访问 Hadoop 的文件系统。
Phoenix Query Server 是一个HBase的开源SQL引擎。你可以使用标准的JDBC API代替HBase客户端API来创建表，插入数据，查询你的HBase数据。
11. 设置密码。（有一行的 username 是 N/A，比较奇怪，不知道用在哪）
12. 数据库配置。如果选择了安装 hive 或 Ranger，需要输入相关的数据库的信息。此例中没有这一步。
13. 目录配置。配置各个服务需要用到的目录。使用默认值即可。
14. 创建用户。默认情况下， ambari 会为每个服务创建一个 linux 的用户，用不同的用户启动不同的服务。我设成了同一个用户。
15. “ALL CONFIGURATIONS”，这一步可以修改前面的一些配置
16. 点击 “Next”后，出现总结页。可以下载创建这个集群的元数据信息。点击 “DEPLOY” 后，就开始部署了。
17. 效果展示。
此时SmartSense有报错，
这是HST SERVER未安装，HST是ambari的收费子项目，安装在集群子节点，用于集群整体的性能反馈调优。这里没有安装，如何需要安装，可单独进行。