ODP集群部署

服务器列表

IP	域名	组件	系统
172.30.20.85	hadoop1.gaojinbo.com	hadoop/hive/spark/Ambari Agent	ubuntu22.04
172.30.20.84	hadoop2.gaojinbo.com	hadoop/hive/spark/Ambari Agent	ubuntu22.04
172.30.32.118	hadoop3.gaojinbo.com	hadoop/hive/spark/Ambari Agent	ubuntu22.04
172.30.32.119	hadoop4.gaojinbo.com	hadoop/hive/spark/Ambari Agent	ubuntu22.04
172.30.58.20	hadoop5.gaojinbo.com	hadoop/hive/spark/Ambari Agent	ubuntu22.04
172.30.32.117	ambari.gaojinbo.com	部署平台ambari server2.7.11+odp1.2.4+pg+mysql	ubuntu22.04

说明

1. ambari和odp操作系统、软件版本必须一致
2. centos7部署失败：缺少依赖包 hadoop_1_2_2_0-native-hdfs
3. 推荐使用ubuntu22.04
4. 推荐使用本地仓库：使用香港代理可能失败
5. ambari需要外部postgresql，hive需要外部mysql

部署步骤

步骤1：安装ambari server

• 安全组打开端口：8080/8440/8441
• 参考 https://gaojinbo.com/docs/%E5%A4%A7%E6%95%B0%E6%8D%AE/ambari/

步骤2：配置ambari到hadoop节点免密登录

• 参考: https://gaojinbo.com/docs/%E6%93%8D%E4%BD%9C%E7%B3%BB%E7%BB%9F/linux/#ssh%E5%85%8D%E5%AF%86%E7%99%BB%E5%BD%95

步骤3：绑定host

1. ambari和hadoop均需添加：

   vi /etc/hosts

   172.30.20.85    hadoop1.gaojinbo.com
   172.30.20.84    hadoop2.gaojinbo.com
   172.30.32.118   hadoop3.gaojinbo.com
   172.30.32.119   hadoop4.gaojinbo.com
   172.30.58.20    hadoop5.gaojinbo.com
   172.30.32.117   ambari.gaojinbo.com

2. 设置FQDN：

   hostnamectl set-hostname hadoop5.gaojinbo.com

步骤4：安装ambari agent

1. 每台hadoop节点执行：

   # 添加密钥
   apt-key adv --recv-keys --keyserver keyserver.ubuntu.com 449E83E81403D1C5
   
   # 配置源
   echo 'deb http://ambari.gaojinbo.com/ambari-release/dist/ubuntu22/1.x/BUILDS/2.7.11.0.0-150/AMBARI AMBARI main' > /etc/apt/sources.list.d/ambari.list
   
   # 禁用systemd-resolved
   systemctl disable systemd-resolved
   systemctl stop systemd-resolved
   echo 'nameserver 114.114.114.114' >> /etc/resolv.conf
   
   # 创建目录
   mkdir /var/lib/livy
   
   # 安装
   apt update -y && sudo apt upgrade -y
   apt install ambari-agent python3-dev -y
   
   # 修改配置
   sed -i 's/localhost/ambari.gaojinbo.com/g' /etc/ambari-agent/conf/ambari-agent.ini
   systemctl enable --now ambari-agent
   ln -s /usr/bin/python3 /usr/bin/python

步骤5：登录ambari后台

• 地址：http://192.168.21.195:18080/#/login
• 账号密码：admin / Qah6h3vGFjnwaC3S

集群配置

1. 建立集群：名称 c1
2. 版本选择：
   • 仓库版本需与选择版本一致：

     https://archive.clemlab.com/ubuntu22/odp-release/1.2.4.0-102/ODP-VDF.xml
     http://ambari.gaojinbo.com/odp-release/dist/ubuntu22/1.x/BUILDS/1.2.4.0-102/ODP-VDF.xml

3. Target Hosts：
   • hadoop1 至 hadoop5
4. SSH Private Key：必须通过文件上传私钥

高级配置

1. HDFS配置：
   • Block replication: 1
   • dfs.permissions.enabled: FALSE
2. 禁用本地53端口：

   systemctl disable systemd-resolved
   systemctl stop systemd-resolved
   echo 'nameserver 114.114.114.114' >> /etc/resolv.conf

错误处理

1. Hive服务启动问题：
   • 修改MySQL配置：

     vi /etc/mysql/mysql.conf.d/mysqld.cnf

     将 127.0.0.1 改为 0.0.0.0，并重启MySQL。
2. Ranger启动问题：
   • 使用低版本MySQL驱动（5.1.49）。
   • 修改用户认证方式：

     ALTER USER rangeradmin IDENTIFIED WITH mysql_native_password BY 'hc123456';

3. Ambari数据库密码修改：

   # 修改配置
   grep "password" /etc/ambari-server/conf/ambari.properties
   echo "bigdata_custom" > /etc/ambari-server/conf/password.dat
   
   # 登录PostgreSQL
   sudo -u postgres psql
   ALTER USER ambari WITH PASSWORD 'bigdata_custom';
   \q
   ambari-server restart

API管理

1. 获取用户信息：

   curl -u admin:admin -X GET http://ambari.gaojinbo.com:8080/api/v1/users/admin

2. 修改用户显示名称：

   curl -u admin:admin -H "X-Requested-By: ambari" -X PUT -d '{"Users/display_name":"admin2"}' http://ambari.gaojinbo.com:8080/api/v1/users/admin

3. 更新密码：

   curl -iv -u admin:admin -H "X-Requested-By: ambari" -X PUT -d '{"AuthenticationSourceInfo/key":"Qah6h3vGFjnwaC3S","AuthenticationSourceInfo/old_key":"admin"}' http://ambari.gaojinbo.com:8080/api/v1/users/admin/sources/1