朝花夕拾之–大數據平台CDH集群離線搭建
標簽: Cloudera-Manager CDH Hadoop 部署 集群
摘要:管理、部署Hadoop集群需要工具,Cloudera Manager便是其一。本文先是簡要對比了當前的類似工具,而后詳細記錄了以離線方式部署CDH集群的步驟。最后對“講究”一詞提出了自己的觀點。
前言
以Apache Hadoop為主導的大數據技術的出現,使得中小型公司對於大數據的存儲與處理也擁有了武器。目前Hadoop有不少發行版:華為發行版 收費、Intel發行版 收費、Cloudera發行版(Cloudera’s Distribution Including Apache Hadoop,簡稱 CDH)免費、Hortonworks發行版(Hortonworks Data Platform
,簡稱 HDP)免費 等,所有這些發行版均是基於Apache Hadoop社區版衍生出來的。
部署、管理擁有數十數百甚至更多節點的Hadoop集群,也需要先進武器。Hortonworks公司的Apache Ambari項目的目的就是通過軟件來配置、監控和管理Hadoop(HDP)集群,以使Hadoop的管理更加簡單。Ambari提供了一個基於它自身RESTful的api實現的直觀的、簡單易用的web界面。Cloudera公司也提供了類似的工具:Cloudera Manager(簡稱 CM)來配置、監控和管理CDH集群。
本文主要內容即是本人早先搭建CDH集群之記錄,故稱作朝花夕拾。需特別注意的是Cloudera Manager與操作系統的版本關系 el7暫不支持,按照官方文檔的要求來,否則安裝會有問題。
注意用戶。
本文是基於操作系統CentOS 6.5, 64-bit;Cloudera Manager 5.3.6;JDK 1.7 版本進行部署的。
部署步驟
網絡配置(所有節點)
[root@cdh-server ~]# vi /etc/sysconfig/network #修改hostname:
NETWORKING=yes
HOSTNAME=cdh-server
[root@cdh-server ~]# vi /etc/hosts #修改ip與主機名的對應關系:
192.168.180.173 cdh-server
192.168.180.175 node175
[root@cdh-server ~]# service network restart #重啟網絡服務生效
安裝JDK(所有節點)
#卸載OpenJDK
[root@cdh-server user1]# rpm -qa | grep java
[root@cdh-server user1]# rpm -e --nodeps java-1.5.0-gcj-1.5.0.0-29.1.el6.x86_64
[root@cdh-server user1]# rpm -e --nodeps java-1.6.0-openjdk-1.6.0.0-1.66.1.13.0.el6.x86_64
[root@cdh-server user1]# rpm -e --nodeps java-1.7.0-openjdk-1.7.0.45-2.4.3.3.el6.x86_64
#安裝JDK
[root@cdh-server user1]# chmod a+x jdk-7u79-linux-x64.rpm
[root@cdh-server user1]# rpm -ivh jdk-7u79-linux-x64.rpm
[root@cdh-server user1]# echo "JAVA_HOME=/usr/java/jdk1.7.0_79/" >>
安裝MySQL(主節點)
[user1@cdh-server]$ cd /home/user1
[user1@cdh-server]$ tar -zxvf mysql-5.6.26-linux-glibc2.5-x86_64.tar.gz
[user1@cdh-server]$ mv mysql-5.6.26-linux-glibc2.5-x86_64 mysql-5.6.26
[user1@cdh-server]$ cd mysql-5.6.26/
[user1@cdh-server]$ vi support-files/my.cnf #新建文件
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
[mysqld]
character-set-server=utf8
default-storage-engine=INNODB
# Uncomment the following if you are using InnoDB tables
innodb_data_home_dir = /home/user1/mysql-5.6.26/data
innodb_data_file_path = ibdata1:10M:autoextend
innodb_log_group_home_dir = /home/user1/mysql-5.6.26/data
# You can set .._buffer_pool_size up to 50 - 80 %
# of RAM but beware of setting memory usage too high
innodb_buffer_pool_size = 16M
innodb_additional_mem_pool_size = 2M
# Set .._log_file_size to 25 % of buffer pool size
innodb_log_file_size = 5M
innodb_log_buffer_size = 8M
innodb_flush_log_at_trx_commit = 1
innodb_lock_wait_timeout = 50
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
初始化MySQL(主節點)
[user1@cdh-server]$ ./scripts/mysql_install_db --defaults-file=/home/user1/mysql-5.6.26/support-files/my.cnf --basedir=/home/user1/mysql-5.6.26 --datadir=/home/user1/mysql-5.6.26/data --user=user1
[user1@cdh-server]$ ./bin/mysqld --defaults-file=/home/user1/mysql-5.6.26/support-files/my.cnf --basedir=/home/user1/mysql-5.6.26 --datadir=/home/user1/mysql-5.6.26/data > mysql.log 2>&1 &
[user1@cdh-server]$ ./bin/mysqladmin -u root password '123456'
[user1@cdh-server mysql-5.6.26]$ ./bin/mysql -uroot -p'123456'
#hive
mysql> create database hive DEFAULT CHARSET utf8 COLLATE utf8_general_ci;
Query OK, 1 row affected (0.00 sec)
#Activity Monitor使用
mysql> create database amon DEFAULT CHARSET utf8 COLLATE utf8_general_ci;
Query OK, 1 row affected (0.01 sec)
#Navigator Audit Server使用
mysql> create database audit DEFAULT CHARSET utf8 COLLATE utf8_general_ci;
Query OK, 1 row affected (0.01 sec)
#Navigator Metadata Server
mysql> create database metadata DEFAULT CHARSET utf8 COLLATE utf8_general_ci;
Query OK, 1 row affected (0.01 sec)
mysql> grant all privileges on *.* to 'root'@'localhost' identified by '123456' with grant option;
Query OK, 0 rows affected (0.00 sec)
mysql> grant all privileges on *.* to 'root'@'cdh-server' identified by '123456' with grant option;
Query OK, 0 rows affected (0.00 sec)
#this user scm is for cloudera manager
mysql> grant all privileges on *.* to 'scm'@'localhost' identified by 'scm' with grant option;
Query OK, 0 rows affected (0.00 sec)
mysql> grant all privileges on *.* to 'scm'@'cdh-server' identified by 'scm' with grant option;
Query OK, 0 rows affected (0.00 sec)
mysql> flush privileges;
Query OK, 0 rows affected (0.00 sec)
部署/啟動CM Server(主節點)
[user1@cdh-server ~]$ tar -zxvf cloudera-manager-el6-cm5.3.6_x86_64.tar.gz
[user1@cdh-server ~]$ cp mysql-connector-java-5.1.33-bin.jar ./cm-5.3.6/share/cmf/lib/
[user1@cdh-server ~]$ su - root
[root@cdh-server ~]# cd /home/user1/
[root@cdh-server user1]# cp -rf cloudera /opt
[root@cdh-server user1]# mv CDH-5.3.6-1.cdh5.3.6.p0.11-el6.parcel /opt/cloudera/parcel-repo/CDH-5.3.6-1.cdh5.3.6.p0.11-el6.parcel
[root@cdh-server user1]# mv CDH-5.3.6-1.cdh5.3.6.p0.11-el6.parcel.sha /opt/cloudera/parcel-repo/CDH-5.3.6-1.cdh5.3.6.p0.11-el6.parcel.sha
[root@cdh-server user1]# mv manifest.json /opt/cloudera/parcel-repo/manifest.json
[root@cdh-server user1]# ./cm-5.3.6/share/cmf/schema/scm_prepare_database.sh mysql cm -hlocalhost:3306 -uroot -p123456 --scm-host localhost scm scm scm
[root@cdh-server user1]# ./cm-5.3.6/etc/init.d/cloudera-scm-server start
Starting cloudera-scm-server: [ OK ]
[root@cdh-server user1]# tail -f ./cm-5.3.6/log/cloudera-scm-server/cloudera-scm-server.log
關閉防火牆(所有節點)
#停止iptables
[root@cdh-server user1]# service iptables stop
#通過瀏覽器訪問驗證
http://192.168.180.173:7180/
部署/啟動CM Agent(從節點)
[root@cdh-server user1]# tar -zxvf cloudera-manager-el6-cm5.3.6_x86_64.tar.gz
[root@cdh-server user1]# vi cm-5.3.6/etc/cloudera-scm-agent/config.ini
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
# Hostname of the CM server.
#server_host=localhost
server_host=cdh-server
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
[root@cdh-server user1]# useradd -G sys --home=/home/user1/cm-5.3.6/run/cloudera-scm-server --no-create-home --comment "Cloudera SCM User" cloudera-scm
[root@cdh-server user1]# useradd --comment "Cloudera SCM User" cloudera-scm #若上一步執行正確,則此步省略
[root@cdh-server user1]# echo 0 > /proc/sys/vm/swappiness
[root@cdh-server user1]# ./cm-5.3.6/etc/init.d/cloudera-scm-agent start
Starting cloudera-scm-agent: [ OK ]
[root@cdh-server user1]# tail -f ./cm-5.3.6/log/cloudera-scm-agent/cloudera-scm-agent.log
配置CDH
登陸Cloudera Manager http://192.168.180.173:7180/,並新建集群Cluster_user1,進行各服務的配置啟動。
#安裝配置hive出錯時,在hiveServer上:
[root@hive-server user1]# cp mysql-connector-java-5.1.33-bin.jar /opt/cloudera/parcels/CDH-5.3.6-1.cdh5.3.6.p0.11/lib/hive/lib/
#同理:use this jar for Navigator Audit Server and Navigator Metadata Server or Activity Server
[root@cdh-server user1]# cp mysql-connector-java-5.1.33-bin.jar /usr/share/java/mysql-connector-java.jar
其他
停止集群步驟
- 停止Cloudera Management Service和Cluster_user1
- 從節點停止Agent
[root@cdh-server user1]# ./cm-5.3.6/etc/init.d/cloudera-scm-agent stop - 主節點停止Server
[root@cdh-server user1]# ./cm-5.3.6/etc/init.d/cloudera-scm-server stop
啟動集群步驟
- 主節點啟動MySQL
[user1@cdh-server]$ ./bin/mysqld --defaults-file=/home/user1/mysql-5.6.26/support-files/my.cnf --basedir=/home/user1/mysql-5.6.26 --datadir=/home/user1/mysql-5.6.26/data > mysql.log 2>&1 & [user1@cdh-server]$ ps -a | grep mysql - 從節點啟動Agent
[root@cdh-server user1]# ./cm-5.3.6/etc/init.d/cloudera-scm-agent start Starting cloudera-scm-agent: [ OK ] [root@cdh-server user1]# tail -f ./cm-5.3.6/log/cloudera-scm-agent/cloudera-scm-agent.log - 主節點啟動Server
[root@cdh-server user1]# ./cm-5.3.6/etc/init.d/cloudera-scm-server start Starting cloudera-scm-server: [ OK ] [root@cdh-server user1]# tail -f ./cm-5.3.6/log/cloudera-scm-server/cloudera-scm-server.log - 啟動各服務
登陸Cloudera Manager http://192.168.180.173:7180/,進行各服務的檢查啟動。
題外話–論講究
講究 是一種情懷,也是一項技能。當處於賣方市場時,公司可能不需要太多的講究就可以賺得盆滿缽盈,然而如果它有更高的追求也就是情懷,可能會更講究,比如更加注重細節、用戶體驗或者極力完美;而處於買方市場時,供需關系會使得公司不得不講究起來,這對他們來說是一項必須技能。
講究 既是推動人類社會發展的動力,也是人類社會發展的成果。當前很多互聯網公司都處於買方市場,誰能贏得用戶誰就拼得勝利,如履薄冰、小心翼翼、極致體驗之講究將他們推向了社會發展的浪尖,先進技術的發明和使用也是層出不窮。
講究 不光體現於公司,也體現於個人、地域、國家等。愈是講究的國家,愈是發達;講究細節的公司,競爭力就越強。然而由於影響因素眾多,講究的個體有錢、贏錢與否則另當別論了。
作者 @王安琪

aitanjupt@hotmail.com
2015 年 10月 20日
