一主两从，manager节点为vm3
master：10.0.0.2 vm1 node
slave1: 10.0.0.3 vm2 node
slave2: 10.0.0.4 vm3 manager

OS verision均为：CentOS release 6.5

1，从https://code.google.com/p/mysql-master-ha/wiki/Downloads?tm=2下载MHA安装包
mha4mysql-node-0.56-0.el6.noarch.rpm
mha4mysql-manager-0.56-0.el6.noarch.rpm

node需要安装mha4mysql-node-0.56-0.el6.noarch.rpm包，
manager节点需要安装mha4mysql-manager-0.56-0.el6.noarch.rpm和mha4mysql-node-0.56-0.el6.noarch.rpm
或者从https://github.com/yoshinorim下载源码包
2 给所有节点配置EPEL yum源。
rpm -ivh http://mirrors.opencas.cn/epel/6/i386/epel-release-6-8.noarch.rpm

yum clean all; yum makecache

3 为所有节点安装mha4mysql-node-0.56-0.el6.noarch.rpm
yum -y install perl-DBD-MySQLncftp
rpm -ivh mha4mysql-node-0.56-0.el6.noarch.rpm

4 为manager节点安装mha4mysql-manager-0.56-0.el6.noarch.rpm
yum -y install perl-Config-Tiny perl-Params-Validate perl-Log-Dispatch perl-Parallel-ForkManager
rpm -ivh mha4mysql-manager-0.56-0.el6.noarch.rpm

5 配置各节点间的ssh公钥信任，实现免密钥登录。（manager to node, node to node）
vm1:
ssh-keygen -t rsa
ssh-copy-id -i /root/.ssh/id_rsa.pub root@10.0.0.3
ssh-copy-id -i /root/.ssh/id_rsa.pub root@10.0.0.4
vm2:
ssh-keygen -t rsa
ssh-copy-id -i /root/.ssh/id_rsa.pub root@10.0.0.2
ssh-copy-id -i /root/.ssh/id_rsa.pub root@10.0.0.4
vm3:
ssh-keygen -t rsa
ssh-copy-id -i /root/.ssh/id_rsa.pub root@10.0.0.2
ssh-copy-id -i /root/.ssh/id_rsa.pub root@10.0.0.3

分别从各个节点ssh登陆其他节点。（必须做，头一次需要输入yes确认）

6 在各mysql节点创建MHA的监控用户
grant all privileges on *.* to 'mha_monitor'@'10.0.0.%' identified by '123456';
mysql> select user,host from user;
+-------------+-----------+
| user| host|
+-------------+-----------+
| mha_monitor | 10.0.0.%|
| repl| 10.0.0.%|

7 配置mha conf文件
mkdir /etc/masterha
mkdir -p /masterha/app1

配置文件app1.cnf内容如下
cat /etc/masterha/app1.cnf
[server default]
manager_workdir=/masterha/app1
manager_log=/masterha/app1/manager.log
ssh_user=root##OS ssh登陆用户

user=mha_monitor ##mha登陆mysql主从的用户
password=123456##mha登陆mysql主从的用户密码

repl_user=repl ##主从复制用户
repl_password=p4ssw0rd ##主从复制用户的密码

ping_interval=1
shutdown_script=""
#master_ip_failover_script="/usr/local/bin/master_ip_failover"
master_ip_online_change_script=""
report_script=""

[server1]
hostname=10.0.0.2
master_binlog_dir="/data/mysql3306"
port=3306
candidate_master=1

[server2]
hostname=10.0.0.3
master_binlog_dir="/data/mysql3306"
port=3306
candidate_master=1

[server3]
hostname=10.0.0.4
master_binlog_dir="/data/mysql3306"
port=3306
no_master=1

8 启动
nohup masterha_manager --conf=/etc/masterha/app1.cnf > /tmp/mha_manager.log< /dev/null 2>&1 &
shutdown 掉master后，可以顺利切换。

但是masterha_check_ssh会报错，只是不影响功能：
[root@vm3 ~]# masterha_check_ssh --conf=/etc/masterha/app1.cnf
Wed Apr6 15:56:05 2016 - [warning] Global configuration file /etc/masterha_default.cnf not found. Skipping.
Wed Apr6 15:56:05 2016 - [info] Reading application default configuration from /etc/masterha/app1.cnf..
Wed Apr6 15:56:05 2016 - [info] Reading server configuration from /etc/masterha/app1.cnf..
Wed Apr6 15:56:05 2016 - [info] Starting SSH connection tests..
Wed Apr6 15:56:05 2016 - [debug]
Wed Apr6 15:56:05 2016 - [debug]Connecting via SSH from root@vm1(10.0.0.2:22) to root@vm2(10.0.0.3:22)..
Wed Apr6 15:56:05 2016 - [debug] ok.
Wed Apr6 15:56:05 2016 - [debug]Connecting via SSH from root@vm1(10.0.0.2:22) to root@10.0.0.4(10.0.0.4:22)..
Wed Apr6 15:56:05 2016 - [debug] ok.
Wed Apr6 15:56:06 2016 - [debug]
Wed Apr6 15:56:05 2016 - [debug]Connecting via SSH from root@vm2(10.0.0.3:22) to root@vm1(10.0.0.2:22)..
Wed Apr6 15:56:06 2016 - [debug] ok.
Wed Apr6 15:56:06 2016 - [debug]Connecting via SSH from root@vm2(10.0.0.3:22) to root@10.0.0.4(10.0.0.4:22)..
Wed Apr6 15:56:06 2016 - [debug] ok.
Wed Apr6 15:56:06 2016 - [error][/usr/share/perl5/vendor_perl/MHA/SSHCheck.pm, ln63]
Wed Apr6 15:56:06 2016 - [debug]Connecting via SSH from root@10.0.0.4(10.0.0.4:22) to root@vm1(10.0.0.2:22)..
Warning: Permanently added '10.0.0.4' (RSA) to the list of known hosts.
Permission denied (publickey,gssapi-keyex,gssapi-with-mic,password).
Wed Apr6 15:56:06 2016 - [error][/usr/share/perl5/vendor_perl/MHA/SSHCheck.pm, ln111] SSH connection from root@10.0.0.4(10.0.0.4:22) to root@vm1(10.0.0.2:22) failed!
SSH Configuration Check Failed!
at /usr/bin/masterha_check_ssh line 44中间

9 最后shutdown掉master，模拟failover:

----- Failover Report -----

app1: MySQL Master failover vm1(10.0.0.2:3306) to vm2(10.0.0.3:3306) succeeded

Master vm1(10.0.0.2:3306) is down!

Check MHA Manager logs at vm3:/masterha/app1/manager.log for details.

Started automated(non-interactive) failover.
Selected vm2(10.0.0.3:3306) as a new master.
vm2(10.0.0.3:3306): OK: Applying all logs succeeded.
10.0.0.4(10.0.0.4:3306): OK: Slave started, replicating from vm2(10.0.0.3:3306)
vm2(10.0.0.3:3306): Resetting slave info succeeded.

主从切换成功。

##########################################
##########################################
关于masterha_check_ssh报错的问题，MHA官网上发现类似问题的帖子有人给的解释如下：
#3 dba.abde...@gmail.com As mentioned here in the tutorial:
https://code.google.com/p/mysql-master-ha/wiki/Tutorial
You shouldn't add an entry for the manager host. If you have 4 hosts, 1, 2, 3 and 4 (manager), you should add host entries for the first 3 ones only (1, 2 & 3) then it should work without any problems.

他的结构是：
master:10.10.105.9
slave1:10.10.105.10
slave2:10.10.105.11
slave3+mha_manage:10.10.105.12
和我的结构类似。