mha+keepalived实现mysql master高可用

1.mha概述：
MHA是由日本Mysql专家用Perl写的一套Mysql故障切换方案以保障数据库的高可用性，它的功能是能在0-30s之内实现主Mysql故障转移（failover），
MHA故障转移可以很好的帮我们解决从库数据的一致性问题，同时最大化挽回故障发生后的数据。MHA里有两个角色一个是node节点 一个是manager节点，要实现这个MHA，必须最少要三台数据库服务器，一主多备，即一台充当master，一台充当master的备份机，另外一台是从属机，需要说明的是一旦主服务器宕机，备份机即开始充当master提供服务，如果主服务器上线也不会再成为master了，因为如果这样数据库的一致性就被改变了。

2.实验环境概述：
由于资源有限，我这用两台服务器做此实验。
系统版本：
Red Hat Enterprise Linux Server release 5.8 (Tikanga) x86_642.6.18-308.el5

主机名	IP地址	备注
mha-node	192.168.199.113	mysql-master节点
mha-master	192.168.199.115	mysql-slave,master的备节点
VIP:192.168.199.118

主从复制配置在此略过，详见可参考：
http://chocolee.blog.51cto.com/8158455/1415173

拓扑图：
图1




故障后转移：
图2




3.部署MHA：
3.1两台主机安装epel源
rpm-ivh http://mirrors.hustunique.com/epel/5/x86_64/epel-release-5-4.noarch.rpm 3.2 两台主机间配置ssh免密钥登陆
mha-node操作：
ssh-keygen-t dsacd/root/.sshssh-copy-id-i id_dsa.pub 192.168.199.115mha-master操作：
ssh-keygen-t dsacd/root/.sshssh-copy-id-i id_dsa.pub 192.168.199.113ssh-copy-id-i id_dsa.pub 192.168.199.115
3.3安装mha
sed-i 's#gpgcheck=1#gpgcheck=0#g' /etc/yum.conf两台主机都安装node:
http://mysql-master-ha.googlecode.com/files/mha4mysql-node-0.52-0.noarch.rpm
yumlocalinstall mha4mysql-node-0.52-0.noarch.rpm master主机安装master:
http://mysql-master-ha.googlecode.com/files/mha4mysql-manager-0.52-0.noarch.rpm
yumlocalinstall mha4mysql-manager-0.52-0.noarch.rpm
3.4 配置mha
mysql添加root远程登陆用户
mha-node操作：
grantall on *.* to 'root'@'192.168.199.%' identified by '123456';flushprivileges;此用户将同步到mha-master节点上。
mha-master操作：
将mysql-slave设置read-only
setglobal read_only=1;app1.conf配置文件
mkdir/etc/masterha/app1 -pcat>>/etc/masterha/app1/app1.conf <<EOF[serverdefault] manager_workdir=/etc/masterha/app1 manager_log=/etc/masterha/app1/manager.log user=rootpassword=123456ssh_user=rootrepl_user=reprepl_password=123456ping_interval=1shutdown_script=""#master_ip_failover_script="/usr/local/bin/master_ip_failover"master_ip_online_change_script=""report_script=""[server1] hostname=192.168.199.113 master_binlog_dir="/usr/local/mysql/data"candidate_master=1[server2] hostname=192.168.199.115 master_binlog_dir="/usr/local/mysql/data"candidate_master=1EOF3.5 测试
测试ssh连接
masterha_check_ssh--conf=/etc/masterha/app1/app1.conf测试主从复制
masterha_check_repl--conf=/etc/masterha/app1/app1.conf
3.6 启动
masterha_manager--conf=/etc/masterha/app1/app1.conf &
关闭主库，观察failover过程
tail-f /etc/masterha/app1/manager.log-----Failover Report ----- app1:MySQL Master failover 192.168.199.113 to 192.168.199.115 succeeded Master192.168.199.113 is down! CheckMHA Manager logs at mha-master:/etc/masterha/app1/manager.log for details. Startedautomated(non-interactive) failover.Thelatest slave 192.168.199.115(192.168.199.115:3306) has all relay logs for recovery.Selected192.168.199.115 as a new master.192.168.199.115:OK: Applying all logs succeeded.Generatingrelay diff files from the latest slave succeeded.192.168.199.115:Resetting slave info succeeded.Masterfailover to 192.168.199.115(192.168.199.115:3306) completed successfully.从日志中可观察到，已经切换到了192.168.199.115上，目前环境为图2.

将环境还原成图1，配置keepalived.

4.部署keepalived
4.1安装keepalived(两台都要装)
yuminstall openssl* popt* -yln-s /usr/src/kernels/2.6.18-308.el5-x86_64/ /usr/src/linuxll/usr/src/mkdir-p /application/toolscd/application/toolswget http://www.keepalived.org/software/keepalived-1.1.19.tar.gz tarzxf keepalived-1.1.19.tar.gz cdkeepalived-1.1.19./configure--sysconf=/etcmakemakeinstallcp/usr/local/sbin/keepalived /usr/sbin/keepalived /etc/init.d/keepalivedstartps-ef|grep keep/etc/init.d/keepalivedstop4.2 配置keepalived
mha-nodekeepalived配置文件：
[root@mha-nodemysql]# cat /etc/keepalived/keepalived.conf !Configuration File for keepalived global_defs{ router_id MHA_01} vrrp_scriptcheck_mysql { script "sh/etc/keepalived/ck_mysql.sh" interval 2 weight -20} vrrp_instanceVI_1 { state BACKUP interface eth0 virtual_router_id 60 priority 100 advert_int 1 authentication { auth_type PASS auth_pass 1111 } virtual_ipaddress { 192.168.199.118/24 } nopreempt track_script { check_mysql }}mha-masterkeepalived配置文件：
[root@mha-masterapp1]# cat /etc/keepalived/keepalived.conf !Configuration File for keepalived global_defs{ router_id MHA_02} vrrp_scriptcheck_mysql { script "sh/etc/keepalived/ck_mysql.sh" interval 2 weight -20} vrrp_instanceVI_1 { state BACKUP interface eth0 virtual_router_id 60 priority 90 advert_int 1 authentication { auth_type PASS auth_pass 1111 } virtual_ipaddress { 192.168.199.118/24 } track_script { check_mysql }}停止mha-node主机keepalived服务，测试VIP漂移是否正常（略）
ck_mysql.sh脚本：
#!/bin/sh/usr/local/mysql/bin/mysql-utest -p123456 -e "select version()" >/dev/null 2>&1if[ $? -eq 0 ];then exit 0else exit 1fi

5.测试
关闭主库的mysql,测试从库是否被提升为主库，VIP是否漂到到从库上去。
-----Failover Report ----- app1:MySQL Master failover 192.168.199.113 to 192.168.199.115 succeeded Master192.168.199.113 is down! CheckMHA Manager logs at mha-master:/etc/masterha/app1/manager.log for details. Startedautomated(non-interactive) failover.Thelatest slave 192.168.199.115(192.168.199.115:3306) has all relay logs forrecovery.Selected192.168.199.115 as a new master.192.168.199.115:OK: Applying all logs succeeded.Generatingrelay diff files from the latest slave succeeded.192.168.199.115:Resetting slave info succeeded.Master failover to192.168.199.115(192.168.199.115:3306) completed successfully. [root@mha-masterapp1]# ip addr |grep 192.168.199. inet 192.168.199.115/24 brd 192.168.199.255scope global eth0 inet 192.168.199.118/24 scope global secondary eth0