搭建cdh平台

简介

CDH：全称Cloudera’s Distribution Including Apache Hadoop，是Hadoop众多分支中的一种，由Cloudera维护，基于稳定版本的Apache Hadoop构建。
优点：

版本划分清晰
版本更新速度快
支持Kerberos安全认证
文档清晰
支持多种安装方式（Cloudera Manager方式）

服务器信息

IP	hostname
1.1.1.1	hostname1
2.2.2.2	hostname2
3.3.3.3	hostname3
4.4.4.4	hostname4
5.5.5.5	hostname5

搭建cdh平台

添加用户
adduser cdh
passwd cdh
赋予权限
chmod u+w /etc/sudoers
vim /etc/sudoers
cdh ALL=(ALL) ALL
chmod u-w /etc/sudoers
免密登录

ssh-keygen -t rsa
cat id_rsa.pub >> authorized_keys
ssh-copy-id -i ~/.ssh/id_rsa.pub  username@xxx.xxx.x.xxx ##将本机的公钥复制到远程机器的authorized_keys文件中.
scp authorized_keys clickhouse@10.101.104.139:/home/clickhouse/.ssh
chmod 600 ~/.ssh/authorized_keys

修改主机名
vim /etc/hostname
修改hosts文件

vim /etc/hosts
#添加如下内容
1.1.1.1 hostname1
2.2.2.2 hostname2
3.3.3.3 hostname3
4.4.4.4 hostname4
5.5.5.5 hostname5

关闭防火墙
关闭selinux

#查看selinux的状态
/usr/sbin/sestatus -v
vim /etc/sysconfig/selinux
#修改如下行
SELINUX=disabled

安装jdk8+
安装ntp
CDH要求各个节点时间同步，所以必须安装ntp服务。

#查看ntpd的状态
service ntpd status
#安装时钟服务ntp
yum -y install ntp
#启动ntp服务
service ntpd start
#设置开机自启
chkconfig ntpd on

优化交换分区

echo "vm.swappiness = 10" >> /etc/sysctl.conf
sysctl -p

禁用透明大页面压缩

echo never > /sys/kernel/mm/transparent_hugepage/defrag
echo never > /sys/kernel/mm/transparent_hugepage/enabled

安装mysql5+
mysql用于存储元数据
只需在主节点上安装即可

#安装mysql
//www.greatytc.com/p/276d59cbc529
#连接mysql
mysql -uroot -p
#创建存储元数据的数据库
create database scm default character set utf8 default collate utf8_general_ci;
create database amon default character set utf8 default collate utf8_general_ci;
create database rman default character set utf8 default collate utf8_general_ci;
create database hue default character set utf8 default collate utf8_general_ci;
create database metastore default character set utf8 default collate utf8_general_ci;
create database sentry default character set utf8 default collate utf8_general_ci;
create database nav default character set utf8 default collate utf8_general_ci;
create database navms default character set utf8 default collate utf8_general_ci;
create database oozie default character set utf8 default collate utf8_general_ci;
#创建用户
create user 'scm'@'%' identified with mysql_native_password by '123456';
create user 'amon'@'%' identified with mysql_native_password by '123456';
create user 'rman'@'%' identified with mysql_native_password by '123456';
create user 'hue'@'%' identified with mysql_native_password by '123456';
create user 'hive'@'%' identified with mysql_native_password by '123456';
create user 'sentry'@'%' identified with mysql_native_password by '123456';
create user 'nav'@'%' identified with mysql_native_password by '123456';
create user 'navms'@'%' identified with mysql_native_password by '123456';
create user 'oozie'@'%' identified with mysql_native_password by '123456';
#赋予权限
grant all on scm.* to 'scm'@'%' with grant option;
grant all on amon.* to 'amon'@'%' with grant option;
grant all on rman.* to 'rman'@'%' with grant option;
grant all on hue.* to 'hue'@'%' with grant option;
grant all on metastore.* to 'hive'@'%' with grant option;
grant all on sentry.* to 'sentry'@'%' with grant option;
grant all on nav.* to 'nav'@'%' with grant option;
grant all on navms.* to 'navms'@'%' with grant option;
#刷新权限
flush privileges;

安装mysql-connector

#下载connector
wget https://dev.mysql.com/get/Downloads/Connector-J/mysql-connector-java-5.1.48.tar.gz
#解压
tar -zxvf mysql-connector-java-5.1.48.tar.gz
#
cd mysql-connector-java-5.1.48
#若没有该目录，则创建
mkdir -p /usr/share/java
#复制，改名
cp mysql-connector-java-5.1.48.jar /usr/share/java/mysql-connector-java.jar
#远程复制到集群的其他节点
scp /opt/mysql-connector-java-5.1.48/mysql-connector-java-5.1.48.jar cdh@2.2.2.2:/usr/share/java/mysql-connector-java.jar

安装CM（cloudera manager）
cloudera-manager-daemon 与 cloudera-manager-agent 是要在每台机器进行安装的，cloudera-manager-server 只需要在集群中想作为未来管理节点的机器上安装即可。

#创建目录
mkdir /opt/cloudera-manager
cd /opt/cloudera-manager
#下载cm包
wget http://archive.cloudera.com/cm5/installer/5.14.4/cloudera-manager-installer.bin
wget http://archive.cloudera.com/cm5/redhat/7/x86_64/cm/cloudera-manager.repo
wget http://archive.cloudera.com/cm5/redhat/7/x86_64/cm/5.14.4/RPMS/x86_64/cloudera-manager-agent-5.14.4-1.cm5144.p0.3.el7.x86_64.rpm
wget http://archive.cloudera.com/cm5/redhat/7/x86_64/cm/5.14.4/RPMS/x86_64/cloudera-manager-daemons-5.14.4-1.cm5144.p0.3.el7.x86_64.rpm
wget http://archive.cloudera.com/cm5/redhat/7/x86_64/cm/5.14.4/RPMS/x86_64/cloudera-manager-server-5.14.4-1.cm5144.p0.3.el7.x86_64.rpm
wget http://archive.cloudera.com/cm5/redhat/7/x86_64/cm/5.14.4/RPMS/x86_64/cloudera-manager-server-db-2-5.14.4-1.cm5144.p0.3.el7.x86_64.rpm
wget http://archive.cloudera.com/cm5/redhat/7/x86_64/cm/5.14.4/RPMS/x86_64/enterprise-debuginfo-5.14.4-1.cm5144.p0.3.el7.x86_64.rpm
#复制cloudera-manager.repo到yum仓库
cp -v /opt/cloudera-manager/cloudera-manager.repo /etc/yum.repos.d/
#远程复制两个包到集群的其它节点下
scp cloudera-manager-agent-5.14.4-1.cm5144.p0.3.el7.x86_64.rpm cdh@2.2.2.2:/opt/cloudera-manager/
scp cloudera-manager-daemons-5.14.4-1.cm5144.p0.3.el7.x86_64.rpm root@2.2.2.2:/opt/cloudera-manager/
#在主节点上安装三个包
yum --nogpgcheck -y localinstall cloudera-manager-daemons-*.rpm
yum --nogpgcheck -y localinstall cloudera-manager-agent-*.rpm
yum --nogpgcheck -y localinstall cloudera-manager-server-*.rpm
#在远程节点上安装两个包
ssh -t cdh@2.2.2.2 'yum --nogpgcheck -y localinstall cloudera-manager-daemons-*.rpm'
ssh -t cdh@2.2.2.2 'yum --nogpgcheck -y localinstall cloudera-manager-agent-*.rpm'
#修改agent文件
vim /etc/cloudera-scm-agent/config.ini
#修改内容如下：hostname1为主节点名称
server_host=hostname1
#复制文件到集群的其他节点上
scp /etc/cloudera-scm-agent/config.ini cdh@2.2.2.2:/etc/cloudera-scm-agent/
#主节点创建目录
mkdir /var/lib/cloudera-scm-server
chown cloudera-scm:cloudera-scm /var/lib/cloudera-scm-server
#初始化
/usr/share/cmf/schema/scm_prepare_database.sh -hhostname1 -uroot -p123456 --scm-host hostname1 --force mysql scm scm 123456

安装cdh
只在主节点上执行

#创建目录
mkdir -p /opt/cloudera/parcel-repo
cd /opt/cloudera/parcel-repo
#下载cdh包
wget http://archive.cloudera.com/cdh5/parcels/5.14.4/CDH-5.14.4-1.cdh5.14.4.p0.3-el7.parcel
wget http://archive.cloudera.com/cdh5/parcels/5.14.4/CDH-5.14.4-1.cdh5.14.4.p0.3-el7.parcel.sha1
wget http://archive.cloudera.com/cdh5/parcels/5.14.4/manifest.json
#重命名
mv CDH-5.14.4-1.cdh5.14.4.p0.3-el7.parcel.sha1 CDH-5.14.4-1.cdh5.14.4.p0.3-el7.parcel.sha
chown -R cloudera-scm:cloudera-scm /opt/cloudera/parcel-repo

启动cm

#启动server
#在主节点上启动
systemctl start cloudera-scm-server
#启动主节点上的agent
systemctl start cloudera-scm-agent
#远程启动集群其它服务器的agent
ssh -t cdh@2.2.2.2 'systemctl start cloudera-scm-agent'
#查看日志
tail -f /var/log/cloudera-scm-server/cloudera-scm-server.log
#查看端口
netstat -tpnl | grep 7180

连接cm
在浏览器输入地址：http://1.1.1.1:7180/
安装组件
。。。
安装spark2
默认cdh安装的spark版本是1.6.0，但是版本太低，可以额外安装spark2，两个版本可以共存。

cd /opt/cloudera
mkdir csd
cd csd
#下载csd包
wget SPARK2_ON_YARN-2.3.0.cloudera3.jar
cd ../parcel-repo
mv manifest.json manifest.json.old
#下载parcels
wget http://archive.cloudera.com/spark2/parcels/2.3.0.cloudera3/SPARK2-2.3.0.cloudera3-1.cdh5.13.3.p0.458809-el7.parcel
wget http://archive.cloudera.com/spark2/parcels/2.3.0.cloudera3/SPARK2-2.3.0.cloudera3-1.cdh5.13.3.p0.458809-el7.parcel.sha1
wget http://archive.cloudera.com/spark2/parcels/2.3.0.cloudera3/manifest.json
#重命名
mv SPARK2-2.3.0.cloudera3-1.cdh5.13.3.p0.458809-el7.parcel.sha1 SPARK2-2.3.0.cloudera3-1.cdh5.13.3.p0.458809-el7.parcel.sha
#重启agent和server
service cloudera-scm-server restart
service cloudera-scm-agent restart
#之后通过cm页面安装spark2

卸载cm
//www.greatytc.com/p/79d1411aaa42

END