cloudera cloudera离线安装

Last updated on September 15, 2024 pm

🧙 Questions

离线安装cdh6.2.0

☄️ Ideas

前提

关闭防火墙

systemctl disable firewalld
systemctl stop firewalld
systemctl status firewalld

修改hostname

#172.16.215.83 iZ8vbgxsdbuxmnqr4qd0ykZ iZ8vbgxsdbuxmnqr4qd0ykZ
172.16.215.83 isxcode
如果有多台服务器,都需要配置

hostnamectl set-hostname isxcode
vim /etc/hosts

关闭selinux

setenforce 0
getenforce

关闭swap分区

swapoff -a
free -m

优化系统

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

挂载磁盘

挂载磁盘,绑定/data

mkdir -p /data

上传资源

需要资源邮箱找我

scp -r /Users/ispong/OneDrive/Downloads/cdh root@47.92.80.91:/tmp

创建用户

useradd cdh
passwd cdh
vim /etc/sudoers
#/ Allow root to run any commands anywhere 
# cdh   ALL =(ALL) NOPASSWD: ALL

安装基础软件

最好每台服务器都装一下
如果不行,使用–force

sudo yum remove python3 -y
cd /tmp/cdh/rpm/python3
sudo rpm -ivh ./* --nosignature

cd /tmp/cdh/rpm/createrepo
sudo rpm -ivh ./* --nosignature

cd /tmp/cdh/rpm/bind-utils
sudo rpm -ivh ./* --nosignature

cd /tmp/cdh/rpm/unzip
sudo rpm -ivh ./* --nosignature

# 这个一定要装
cd /tmp/cdh/rpm/httpd
sudo rpm -ivh ./* --nosignature

# 这个一定要装
cd /tmp/cdh/rpm/postgres
sudo rpm -ivh ./* --nosignature

# 如果包中有冲突文件,直接删除
sudo yum remove java-1.8.0-openjdk-headless -y
cd /tmp/cdh/rpm/openjdk
sudo rpm -ivh ./* --nosignature

安装时区同步

每台都要做

cd /tmp/cdh/rpm/ntp
rpm -ivh ./* --nosignature
systemctl enable ntpd
systemctl start ntpd
vim /etc/ntp.conf
# server 0.centos.pool.ntp.org iburst
# server 1.centos.pool.ntp.org iburst
# server 2.centos.pool.ntp.org iburst
# server 3.centos.pool.ntp.org iburst
server 10.244.18.169 iburst
systemctl restart ntpd
systemctl status ntpd
ntpq -p
timedatectl status

安装mysql驱动

cp /tmp/cdh/jdbc/mysql-connector-java-8.0.23.jar /usr/share/java/mysql-connector-java.jar

安装scm

主节点安装 按顺序运行
卸载: yum remove -y cloudera-manager-daemons

cd /tmp/cdh/cdh6.2.0/cloudera-repos-6.2.0
yum install -y cloudera-manager-daemons-6.2.0-968826.el7.x86_64.rpm 
yum install -y cloudera-manager-server-6.2.0-968826.el7.x86_64.rpm
yum install -y cloudera-manager-server-db-2-6.2.0-968826.el7.x86_64.rpm

创建数据库

CREATE DATABASE cdh_scm DEFAULT CHARACTER SET utf8 DEFAULT COLLATE utf8_general_ci;
CREATE DATABASE cdh_hive DEFAULT CHARACTER SET utf8 DEFAULT COLLATE utf8_general_ci;
CREATE DATABASE cdh_monitor DEFAULT CHARACTER SET utf8 DEFAULT COLLATE utf8_general_ci;
GRANT ALL PRIVILEGES ON cdh_scm.* TO 'ispong'@'%';
GRANT ALL PRIVILEGES ON cdh_hive.* TO 'ispong'@'%';
GRANT ALL PRIVILEGES ON cdh_monitor.* TO 'ispong'@'%';
Flush Privileges;

初始化数据库

/opt/cloudera/cm/schema/scm_prepare_database.sh -h ${host} -P ${port} mysql ${dbName} ${username} ${password}
会生成 /etc/cloudera-scm-server/db.properties文件

/opt/cloudera/cm/schema/scm_prepare_database.sh -h isxcode -P 30102 mysql cdh_scm root admin123

安装httpd

Note: 如果报错先不用管,可能因为80端口被占用

systemctl start httpd
vim /etc/httpd/conf/httpd.conf
# / Listen
#Listen 12.34.56.78:80
Listen 30108
systemctl restart httpd
systemctl status httpd

上传cdh6_parcel文件

mkdir -p /var/www/html
mkdir -p /data/httpd/cdh6_parcel
ln -s /data/httpd/cdh6_parcel /var/www/html
cp /tmp/cdh/cdh6.2.0/parcel-6.2.0/* /var/www/html/cdh6_parcel

mkdir -p /data/httpd/cdh6_parcel/cm6/
cp /tmp/cdh/cdh6.2.0/cloudera-repos-6.2.0/* /var/www/html/cdh6_parcel/cm6/

创建本地镜像仓库

cd /var/www/html/cdh6_parcel/cm6/
createrepo .

如果存在多个服务器,则都要

cd /etc/yum.repos.d
vim cloudera-manager.repo
[cloudera-manager]
name=Cloudera Manager 6.2.0
baseurl=http://isxcode:30108/cdh6_parcel/cm6
gpgcheck=0
enabled=1

离线模式需要删除其他的库,否则cdh无法拉取软件库

cd /etc
mkdir yum.repos.d_bak
mv yum.repos.d/CentOS-* yum.repos.d_bak/
yum-config-manager --enable cloudera-manager
yum clean all && yum makecache

启动scm

systemctl enable cloudera-scm-server
systemctl start cloudera-scm-server
systemctl status cloudera-scm-server

查看scm日志

tail -f /var/log/cloudera-scm-server/cloudera-scm-server.log

修改scm访问端口

INSERT INTO CONFIGS (CONFIG_ID, ATTR, VALUE, CONFIG_CONTAINER_ID) VALUES (4, ‘http_port’, ‘${customPort}’, 2);

INSERT INTO cdh_scm.CONFIGS (CONFIG_ID, ATTR, VALUE, CONFIG_CONTAINER_ID) VALUES (4, 'http_port', '30107', 2);

sudo systemctl stop cloudera-scm-server
sudo systemctl stop cloudera-scm-agent
sudo systemctl restart cloudera-scm-agent

systemctl restart cloudera-scm-server

测试

默认端口号: 7180

安装集群

20240730145812

20240730145840

20240730145947

20240730150015

Cloudera Manager Agent: http://isxcode:30108/cdh6_parcel/cm6
CDH and other software: http://isxcode:30108/cdh6_parcel

20240730150250

一定要修改路径,不要使用映射的路径,给个新的路径

20240730150352

20240730150404

20240730150424

使用之前创建的cdh账号

20240730150453

20240730150541

有可能存在缺少依赖,去/cdh-rpm包中寻找

20240730150806

20240730150855

安装服务

20240730150927

20240730151054

只需要安装hive、hdfs、yarn。推荐安装zookeeper

20240730151209

使用之前创建的cdh_hive和cdh_monitor

20240730151802

修改安装路径
一定不要提前创建目录!!! 会有权限问题
/data/dfs
/data/yarn
/data/cloudera-host-monitor
/data/cloudera-service-monitor

20240730152011

20240730152432

20240730152523

20240730154116

tee -a /etc/profile <<-'EOF'
export HADOOP_HOME=/opt/cloudera/parcels/CDH/lib/hadoop
export HADOOP_CONF_DIR=/opt/cloudera/parcels/CDH/lib/hadoop/etc/hadoop
EOF
source /etc/profile
  • Hive链接信息:
  • url: jdbc:hive2://47.92.37.247:10000/ispong_db
  • hive.metastore.uris: thrift://isxcode:9083
  • username: cdh
  • yarn-web: http://47.92.37.247:8088/cluster

优化配置

  1. 根据提示优化配置
  2. 修改hive的连接数

    hive.server2.thrift.max.worker.threads: 2000

SELECT hive_open_connections,hive_open_operations FROM ENTITY_DATA;
show variables like '%max_connections%';
  1. yarn的资源配置

    yarn.nodemanager.resource.cpu-vcores: 8
    yarn.nodemanager.resource.memory-mb: 16GB
    yarn.scheduler.minimum-allocation-vcores: 1
    yarn.scheduler.minimum-allocation-mb: 2GB
    yarn.scheduler.maximum-allocation-vcores: 4
    yarn.scheduler.maximum-allocation-mb: 4GB

  2. 赋予用户操作权限

    groupadd supergroup
    usermod -a -G supergroup cdh
    newgrp supergroup
  3. 创建用户目录

    hdfs dfs -mkdir /user/cdh
    hdfs dfs -chown cdh:supergroup /user/cdh

cloudera cloudera离线安装
https://ispong.isxcode.com/hadoop/cloudera/cloudera cdh离线安装/
Author
ispong
Posted on
July 27, 2024
Licensed under