实验配置环境：
# uname -ro
2.6.32-358.el6.i686 GNU/Linux
1.修改节点名称，集群的每个节点名称够能够相互解析
1.修改节点node1主机名# node1.magedu.com 2.修改节点node1主机名# node2.magedu.com 3.在集群节点node1和node2的/etc/hosts添加如下两行，使其能够相互解析# vim /etc/hosts172.16.0.4 node1.magedu.com node1172.16.0.7 node2.magedu.com node2
2.配置ntp服务器
1.查看是否安装ntp服务的软件包# rpm -qa ntpntp-4.2.4p8-3.el6.centos.i6862.将原来配置文件备份，提供新的编辑配置文件/etc/ntp.conf，# vim /etc/ntp.conf driftfile /var/lib/ntp/driftrestrict default kod nomodify notrap nopeer noqueryrestrict -6 default kod nomodify notrap nopeer noquery restrict 127.0.0.1restrict -6 ::1restrict 172.16.0.0 netmask 255.255.0.0 nomodify notrap server cn.pool.ntp.org preferserver 133.100.11.8 server 127.127.1.0 # local clockfudge 127.127.1.0 stratum 10 includefile /etc/ntp/crypto/pw keys /etc/ntp/keyslogfile /var/log/ntpstats/ntp.log 3.启动nfp服务# service ntpd start 4.验证ntp服务# ntpstatsynchronised to local net at stratum 11 time correct to within 11 ms polling server every 1024 s # ntpstatsynchronised to local net at stratum 11 time correct to within 11 ms polling server every 1024 s[root@localhost ~]# ntpq -p remote refid st t when poll reach delay offset jitter============================================================================== dns1.synet.edu. 202.118.1.46 2 u 134 1024 377 95.428 14611.6 2042.24 133.100.11.8 133.100.53.51 2 u 306 1024 377 151.025 14205.3 2064.42*LOCAL(0) .LOCL. 10 l 30 64 377 0.000 0.000 0.001
3.保证节点一和节点的时间同步
在节点node1和节点node2上配置周期性任务计划，使其时间保持同步# crontab -e*/3 * * * * /usr/sbin/ntpdate 172.16.0.6 &> /dev/null
4.各节点之间基于ssh密钥认证通信
节点node1和节点node2不需要密码进行通信1.将节点node1生成的公钥传到节点node2# ssh-keygen -t rsa -P ''# ssh-copy-id -i .ssh/id_rsa.pub root@node1 2.将节点node2生成的公钥传到节点node1# ssh-keygen -t rsa -P ''# ssh-copy-id -i .ssh/id_rsa.pub root@node2 3.节点node1和node2时间同步# ssh node1 'date';dateSun Apr 20 20:32:00 CST 2014Sun Apr 20 20:32:00 CST 2014
5.安装配置heartbeat v2
1.解决依赖关系（版本不同解决依赖关系所安装的包有所不同）# yum install PyXML -y# yum install libnet -y# yum install net-snmp-libs -y# rpm -ivh heartbeat-2.1.4-12.el6.i686.rpm# rpm -ivh heartbeat-pils-2.1.4-12.el6.i686.rpm# rpm -ivh heartbeat-stonith-2.1.4-12.el6.i686.rpm2.复制heartbeat集群服需要的文件# cp /usr/share/doc/heartbeat-2.1.4/{haresources,ha.cf,authkeys} /etc/ha.d/ # lsauthkeys harc rc.d resource.dha.cf haresources README.config shellfuncs 3.配置authkeys# openssl rand -hex 8cc7cb7ada0d820fb添加如下两行# vim authkeys auth 22 sha1 cc7cb7ada0d820fb 修改authkeys的权限，必须为600或400否则不能启动# chmod 600 authkeys 3.配置核心配置文件ha.cflogfile /var/log/ha-log 启用日志文件#logfacility local0keepalive 800ms 发送心跳信息的间隔deadtime 4 集群故障之后节点退出集群的时间间隔warntime 3 警告时间initdead 90 集群服务启动时的死亡时间间隔udpport 694 监听端口mcast eth0 225.0.19.1 694 1 0 启用多播（确保网卡支持多波)node node1.magedu.comnode node2.magedu.com 添加集群节点只能使用主机名auto_failback on 启用集群节点故障自动转回ping 172.16.0.1 启用ping节点compression bz2 启用压缩功能compression_threshold 2 超过指定的值时则压缩传输，单位KB4.定义集群资源，使用httpd测试 4.1在node1和node2节点中安装httpd# yum install httpd -y# ssh node1 'yum install httpd -y'在node1和node2确保httpd开启不能自启# chkconfig httpd off# chkconfig --list httpdhttpd 0:off 1:off 2:off 3:off 4:off 5:off 6:off# ssh node1 'chkconfig httpd off'# ssh node1 'chkconfig --list httpd'httpd 0:off 1:off 2:off 3:off 4:off 5:off 6:off 4.2分别为node1和node2通过测试页# vim /var/www/html/index.html<h1> Cluster Node2</h1># vim /var/www/html/index.html<h1>Cluter Node1</h1> 4.3 在node1和node2中启动httpd服务，并测试，然后手动停止服务#ssh node1 ' service httpd start'; service httpd start # ssh node1 'curl node2';curl node1 % Total % Received % Xferd Average Speed Time Time Time Current Dload Upload Total Spent Left Speed 0 0 0 0 0 0 0 0 --:--:-- --:--:-- --:--:-- 0<h1> Cluster Node2</h1> 0 24 0 24 0 0 302 0 --:--:-- --:--:-- --:--:-- 400<h1>Cluter Node1</h1> # ssh node1 'service httpd stop';service httpd stop; 4.4 # vim haresources 定义httpd资源node1.magedu.com 172.16.0.19/16/eth1 httpd5.将haresources,ha.cf,authkeys，原封不动的复制到node1中# scp -p authkeys ha.cf haresources node1:/etc/ha.d/ # ssh node1 'ls -l /etc/ha.d/'total 48-rw-------. 1 root root 657 Apr 20 22:00 authkeys-rw-r--r--. 1 root root 10591 Apr 20 22:00 ha.cf-rwxr-xr-x. 1 root root 745 Sep 10 2013 harc-rw-r--r--. 1 root root 5963 Apr 20 22:00 haresourcesdrwxr-xr-x. 2 root root 4096 Apr 20 20:48 rc.d-rw-r--r--. 1 root root 692 Sep 10 2013 README.configdrwxr-xr-x. 2 root root 4096 Apr 20 20:48 resource.d-rw-r--r--. 1 root root 7862 Sep 10 2013 shellfuncs
6.启动node1和node2的集群服务
# ssh node1 'service heartbeat start';service heartbeat start # ss -untl | grep 80tcp LISTEN 0 128 :::80 :::*
7.验证集群服务：
关闭node1中的web服务，web服务正常启动，且vip自动转到inode1
8.基于heartbeat-gui配置Mariadb高可用集群服务
所有heartbeat的配置和上面的相同1.安装heartbeat-guirpm -ivh heartbeat-gui-2.1.4-12.el6.i686.rpm 2.在node1和node2的ha.cf添加如下一行# vim /etc/ha.d/ha/cfcrm on 3.在node1和node2节点中安装Mariadb（具体安装和配置可以参考以前的博客）4.在node1和node2确保httpd开启不能自启# chkconfig mysqld off# chkconfig --list mysqldmysqld 0:off 1:off 2:off 3:off 4:off 5:off 6:off# ssh node2 'chkconfig mysqld off'# ssh node2 'chkconfig --list mysqld'mysqld 0:off 1:off 2:off 3:off 4:off 5:off 6:off 4.关闭node1和node2中的heartbeat，保证mysqld服务关闭# ssh node2 'service heartbeat stop';service heartbeat stop # ssh node2 'service mysqld status';service mysqld statusMySQL is not running[FAILED]MySQL is not running [FAILED] 5.重启node1和node2中的heartbeat服务，确保保证httpd服务关闭# ssh node2 'service heartbeat start';service heartbeat start# ssh node2 'service mysqld status';service mysqld statusMySQL is not running[FAILED]MySQL is not running [FAILED] 6.验证是否启用heartbeat-gui# ss -untl | grep 5560tcp LISTEN 0 10 *:5560 *:* 7.启用heartbeat-gui配置# hb_gui &
9. 配置Mariadb高可用
10.添加资源
添加虚拟ip首先在Resources选项中选择添加本地资源
添加mysqld，直接点击add就能够添加Mariadb
定义资源约束
添加位置约束使vip和Mariadb更倾向node1节点
定义顺序约束，指定资源启动和关闭的顺序
定义Mariadb和vip必须运行于同一个节点
11.启动Mariadb高可用服务
12.验证配置的Mariadb高可用服务，使node1离线并且能够实现故障转移
总结：
heartbeat v1和v2必须在每一个节点都必须在每一节点上都进行相同的配置，配
置高可用主要是实现服务的高可用，一般用几个9来衡量。配置高可用集群必须修改每
个节点的名称，使其能够解析，并且每个集群节点的时间必须保持同步。当出现如下错
误时：使用hostname命令修改为响应的主机即可。
service heartbeat start
Starting High-Availability services:
Heartbeat failure [rc=6]. Failed.
heartbeat: udpport setting must precede media statementsheartbeat[12088]: 2014/04/23_23:04:40 info: Version 2 support: on
heartbeat[12088]: 2014/04/23_23:04:40 WARN: File /etc/ha.d/haresources exists.
heartbeat[12088]: 2014/04/23_23:04:40 WARN: This file is not used because crm is enabled
heartbeat: baudrate setting must precede media statementsheartbeat[12088]: 2014/04/23_23:04:40 ERROR: Current node [node1.magedu.com ] not in configuration!
heartbeat[12088]: 2014/04/23_23:04:40 info: By default, cluster nodes are named by `uname -n` and must be declared with a 'node' directive in the ha.cf file.
heartbeat[12088]: 2014/04/23_23:04:40 info: See also: http://linux-ha.org/ha.cf/NodeDirective
heartbeat[12088]: 2014/04/23_23:04:40 WARN: Logging daemon is disabled --enabling logging daemon is recommended
heartbeat[12088]: 2014/04/23_23:04:40 ERROR: Configuration error, heartbeat not started.
本文出自 “Linux之旅” 博客，请务必保留此出处http://openlinuxfly.blog.51cto.com/7120723/1403247