CDH安装指南——酒仙网技术

1、集群规划

    说明:由于CDH管理集群很方便,可以动态添加删除主机,动态改变主机主机上的服务,动态监控服务状态;

    1)、主机准备:windows宿主机或linux桌面版宿主机或苹果宿主机(本人使用的debian+xfce桌面作为自己的宿主机

    2)、虚拟机软件准备:使用VMware Workstation12 pro   

        

    3)、虚拟机系统:centos 6.5

    4)、机器规划:hadoopmaster、hadoopslave01、hadoopslave02、hadoopslave(作为基础镜像使用)

        例如:

        注意:所有的虚拟机的账号密码都是统一的,意思就是说分布式看成一个主机

    5)、hadooopmaster 运行 Cloudera Manager

2、资源准备:

    1)、下载postgresql分别:http://yum.postgresql.org/8.4/redhat/rhel-6-x86_64/repoview/

        postgresql84-8.4.22-1PGDG.rhel6.x86_64.rpm  

        postgresql84-libs-8.4.22-1PGDG.rhel6.x86_64.rpm  

        postgresql84-server-8.4.22-1PGDG.rhel6.x86_64.rpm

    2)、下载最新版本的cm资源包压缩包:http://archive-primary.cloudera.com/cm5/repo-as-tarball/5.7.1/   

        

        注意:根据自己的虚拟机下载对应版本的cm,注意cm一定要是最新版本;

    3)、下载最新版本的cm安装包程序:http://archive.cloudera.com/cm5/installer/latest/

        

    4)、安装hadoop、hivingsql、hbase等分布式生态系统软件:http://archive.cloudera.com/cdh5/parcels/5.7.1/

        

        

        注意:分别下载截图中的文件

3、虚拟机配置所有的机器都需要的配置

    1)、修改主机名:vi /etc/sysconfig/network        

  1. NETWORKING=yes
  2. HOSTNAME=hadoopmaster

    2)、修改hosts文件:vi /etc/hosts

  1. #host
  2. 192.168.240.129 hadoopmaster
  3. 192.168.240.130 hadoopslave01
  4. 192.168.240.134 hadoopslave02
  5. #192.168.240.129 archive.cloudera.com

    3)、设置固定IP,根据自己的需求设置:vi /etc/sysconfig/network-scripts/ifcfg-eth0

  1. DEVICE="eth1"
  2. BOOTPROTO="static"
  3. HWADDR="00:50:56:34:61:2E"
  4. IPV6INIT="yes"
  5. NM_CONTROLLED="yes"
  6. ONBOOT="yes"
  7. TYPE="Ethernet"
  8. #UUID="6d6002f2-4f52-49ef-bb07-e2b54637cafb"
  9. IPADDR=192.168.240.134
  10. NETMASK=255.255.255.0
  11. GATEWAY=192.168.240.2
  12. DNS1=192.168.240.2

    4)、重启网络服务:service network restart

4、关机防火墙 和 SELinux

    1)、关闭防火墙

        service iptables stop    (临时关闭防火墙)

        chkconfig iptables off    (重启后生效)

    2)、关闭SELinux

        setenforce 0 (临时设置selinux)

        vi /etc/selinux/config

  1. # This file controls the state of SELinux on the system.
  2. # SELINUX= can take one of these three values:
  3. #     enforcing - SELinux security policy is enforced.
  4. #     permissive - SELinux prints warnings instead of enforcing.
  5. #     disabled - No SELinux policy is loaded.
  6. SELINUX=disabled
  7. # SELINUXTYPE= can take one of these two values:
  8. #     targeted - Targeted processes are protected,
  9. #     mls - Multi Level Security protection.
  10. SELINUXTYPE=targeted

5、集群间服务器使用无密码ssh登陆,再次提醒各个集群中的主机最好使用统一的账号密码

    1)、hadoopmaster主机

  1. ssh-keygen -t rsa (一直按回车即可)
  2. ssh-copy-id -i localhost
  3. ssh-copy-id -i hadoopmaster
  4. ssh-copy-id -i hadoopslave01
  5. ssh-copy-id -i hadoopslave02

    2)、hadoopslave01主机

  1. ssh-keygen -t rsa (一直按回车即可)
  2. ssh-copy-id -i localhost
  3. ssh-copy-id -i hadoopmaster
  4. ssh-copy-id -i hadoopslave01

    3)、hadoopslave02主机

  1. ssh-keygen -t rsa (一直按回车即可)
  2. ssh-copy-id -i localhost
  3. ssh-copy-id -i hadoopmaster
  4. ssh-copy-id -i hadoopslave02

6、建立cm资源服务器——安装Apache httpd web服务器

    1)、检查httpd服务是否安装:service httpd status

        如果没有安装:yum install -y httpd

  1. yum install -y httpd #安装httpd
  2. service httpd start #启动命令
  3. service httpd stop #关闭命令
  4. service httpd restart #重启命令
  5. service httpd status #查看服务状态
  6. chkconfig httpd on #将httpd加入到启动列表

7、选择hadoopmaster作为cm的资源服务器发布cm

  1. [root@localhost hadoopmaster]# pwd
  2. /home/hadoop/hadoopmaster
  3. [root@localhost hadoopmaster]# ll
  4. total 529076
  5. -rw-r--r--. 1 root root 541768507 Jun 14 09:14 cm5.7.1-centos6.tar.gz
  6. [root@localhost hadoopmaster]# tar -zxvf cm5.7.1-centos6.tar.gz
  7. cm/
  8. cm/hadoopmaster-manager.repo
  9. cm/5
  10. cm/5.7.1/
  11. cm/5.7.1/mirrors
  12. cm/5.7.1/repodata/
  13. cm/5.7.1/repodata/other.xml.gz.asc
  14. cm/5.7.1/repodata/filelists.xml.gz
  15. cm/5.7.1/repodata/primary.xml.gz.asc
  16. cm/5.7.1/repodata/repomd.xml
  17. cm/5.7.1/repodata/other.xml.gz
  18. cm/5.7.1/repodata/repomd.xml.asc
  19. cm/5.7.1/repodata/primary.xml.gz
  20. cm/5.7.1/repodata/filelists.xml.gz.asc
  21. cm/5.7.1/RPMS/
  22. cm/5.7.1/RPMS/x86_64/
  23. cm/5.7.1/RPMS/x86_64/hadoopmaster-manager-daemons-5.7.1-1.cm502.p0.297.el6.x86_64.rpm
  24. cm/5.7.1/RPMS/x86_64/enterprise-debuginfo-5.7.1-1.cm502.p0.297.el6.x86_64.rpm
  25. cm/5.7.1/RPMS/x86_64/hadoopmaster-manager-server-db-2-5.7.1-1.cm502.p0.297.el6.x86_64.rpm
  26. cm/5.7.1/RPMS/x86_64/hadoopmaster-manager-server-5.7.1-1.cm502.p0.297.el6.x86_64.rpm
  27. cm/5.7.1/RPMS/x86_64/hadoopmaster-manager-agent-5.7.1-1.cm502.p0.297.el6.x86_64.rpm
  28. cm/5.7.1/RPMS/x86_64/oracle-j2sdk1.7-1.7.0+update45-1.x86_64.rpm
  29. cm/5.7.1/RPMS/x86_64/jdk-6u31-linux-amd64.rpm
  30. cm/5.7.1/RPMS/noarch/
  31. cm/RPM-GPG-KEY-hadoopmaster
  32. [root@localhost hadoopmaster]# ll
  33. total 529080
  34. drwxrwxr-x. 3 1106  592      4096 Jun 11 11:08 cm
  35. -rw-r--r--. 1 root root 541768507 Jun 14 09:14 cm5.7.1-centos6.tar.gz
  36. [root@localhost hadoopmaster]#

        ——移动加压后的cm文件夹到web目录,并设置权限

  1. [root@localhost html]# pwd
  2. /var/www/html
  3. [root@localhost html]# mkdir -p cm5/redhat/6/x86_64
  4. [root@localhost html]# cd cm5/redhat/6/x86_64
  5. [root@localhost x86_64]# mv /home/soft/cloudera/cm .
  6. [root@localhost x86_64]# chmod -R ugo+rX cm

       ——设置完成后访问下地址:http://hadoopmaster/cm5/redhat/6/x86_64/cm/


8、安装Cloudera Manager 5

    1)、安装postgresql

        是否安装postgresql:rpm -qa|grep postgres,如果已经安装,但版本小于8.4,则使用rpm -e卸载。

  1. [root@localhost postgresql84]# rpm -ivh postgresql84-libs-8.4.21-1PGDG.rhel6.x86_64.rpm
  2. warning: postgresql84-libs-8.4.21-1PGDG.rhel6.x86_64.rpm: Header V4 DSA/SHA1 Signature, key ID 442df0f8: NOKEY
  3. Preparing...                ########################################### [100%]
  4.   1:postgresql84-libs      ########################################### [100%]
  5. [root@localhost postgresql84]# rpm -ivh postgresql84-8.4.21-1PGDG.rhel6.x86_64.rpm
  6. warning: postgresql84-8.4.21-1PGDG.rhel6.x86_64.rpm: Header V4 DSA/SHA1 Signature, key ID 442df0f8: NOKEY
  7. Preparing...                ########################################### [100%]
  8.   1:postgresql84           ########################################### [100%]
  9. [root@localhost postgresql84]# rpm -ivh postgresql84-server-8.4.21-1PGDG.rhel6.x86_64.rpm
  10. warning: postgresql84-server-8.4.21-1PGDG.rhel6.x86_64.rpm: Header V4 DSA/SHA1 Signature, key ID 442df0f8: NOKEY
  11. Preparing...                ########################################### [100%]
  12.   1:postgresql84-server    ########################################### [100%]
  13. [root@localhost postgresql84]#

    2)、修改安装cm的客户端的配置

        cd /etc/yum.repos.d/

        mv /etc/yum.repos.d/Centos-* /home/hadoop/soft/

        vi /etc/yum.repos.d/myrepo.repo

  1. [myrepo]
  2. name=myrepo
  3. baseurl=http://archive.cloudera.com/cm5/redhat/6/x86_64/cm/5/
  4. enabled=1
  5. gpgcheck=0

        vi /etc/hosts

  1. #host
  2. 192.168.240.129 hadoopmaster
  3. 192.168.240.130 hadoopslave01
  4. 192.168.240.134 hadoopslave02
  5. 192.168.240.129 archive.cloudera.com

    3)、安装CM5

  1. [root@localhost cloudera]# chmod +x cloudera-manager-installer.bin
  2. [root@localhost cloudera]# ./cloudera-manager-installer.bin

        注意:查看安装过程日志:/var/log/cloudera-manager-installer/

    4)、安装过程中截图     

        a、安装过程比较漫长,耐心等地

        

        b、安装完成,查看安装地址:http://192.168.240.134:7180/

        

        c、登陆界面,默认账号:admin 密码:admin

        

        d、同意条款

        

        e、选择免费版——cloudera manager express

        

        f、设置ip端,例如:192.168.240.[1-255] ,然后点击搜索    

        

        g、选择主机节点

        

9、集群安装hadoop、hivingsql、hbase、spark、storm等,即CDH

    1)、安装cdh

        a、集群安装

        

        b、使用”更多选项“使用”本地 Parcel 存储库路径“,位置:/opt/cloudera/parcel-repo        

        

       

        注意(每台机器执行):yum install -y portmap python-psycopg2 cyrus-sasl-gssapi mod_ssl openssl-devel MySQL-python


         1)、将前面下载的CDH parcels文件放到 /opt/cloudera/parcel-repo 目录下

  1. [root@hadoopmaster parcel-repo]# mv /home/hadoop/Documents/parcel/* ./
  2. [root@hadoopmaster parcel-repo]# ls
  3. CDH-5.7.1-1.cdh5.7.1.p0.11-el6.parcel       manifest.json
  4. CDH-5.7.1-1.cdh5.7.1.p0.11-el6.parcel.sha1

        2)、生成对应的sha文件

            manifest.json 文件中按照版本“CDH-5.7.1-1.cdh5.7.1.p0.53-el6.parcel” -(根据自己的版本选定) 找到对应的 hash值  

        "hash": "67fc4c86b260eeba15c339f1ec6be3b59b4ebe30"

        3)、hash值保存在sha文件中

        echo '67fc4c86b260eeba15c339f1ec6be3b59b4ebe30'>CDH-5.1.0-1.cdh5.1.0.p0.53-el6.parcel.sha 根据自己的版本定

  1. [root@hadoopmaster parcel-repo]# ls
  2. CDH-5.7.1-1.cdh5.7.1.p0.11-el6.parcel       manifest.json
  3. CDH-5.7.1-1.cdh5.7.1.p0.11-el6.parcel.sha1
  4. [root@hadoopmaster parcel-repo]# echo 'e24f56882ecdcf6837636abc8298c629c0f9d109' > CDH-5.7.1-1.cdh5.7.1.p0.11-el6.parcel.sha1

    2)、验证

        根据自己的使用情况进行验证;

        



文章摘自: http://blog.csdn.net/feixeyes/article/details/38084403

                    http://blog.csdn.net/yhl27/article/details/37726639








原文地址:https://www.cnblogs.com/linuxone/p/5577035.html