机器简介：

monitor 192.168.146.131 监控服务器

master-001 192.168.146.129 读写主机001

master-002 192.168.146.130 读写主机002

slave-001 192.168.146.131 只读主机001

虚拟IP分配

master-001 192.168.146.135 writer

master-002 192.168.146.136 reader

slave-001  192.168.146.137 reader

三台相同操作：

关闭防火墙

一，安装mysql

二，mysql 权限赋值

GRANT REPLICATION SLAVE ON *.* TO 'myrepl'@'192.168.146.%' IDENTIFIED BY '123456';

GRANT REPLICATION CLIENT                 ON *.* TO 'mmm_monitor'@'192.168.146.%' IDENTIFIED BY 'monitor';

GRANT SUPER, REPLICATION CLIENT, PROCESS ON *.* TO 'mmm_agent'@'192.168.146.%'   IDENTIFIED BY 'agent';

二，安装MMM

安装mysql-MMM

安装依赖包：

yum install epel-release

yum install  perl-Log-Log4perl-Appender-Socket-UNIX

rpm 安装

rpm -ivh mysql-mmm-2.2.1-2.el6.noarch.rpm

安装mysql-agent

安装依赖包：

yum install -y perl-Algorithm-Diff perl-Class-Singleton perl-Proc-Daemon perl-Net-ARP perl-DBD-mysql

rpm 安装

rpm -ivh mysql-mmm-agent-2.2.1-2.el6.noarch.rpm

chkconfig mysql-mmm-agent on

（131上）monitor 机器上安装monitor包

rpm -ivh mysql-mmm-monitor-2.2.1-2.el6.noarch.rpm

三.配置MMM

修改my.cnf

db1

server-id= 1

gtid_mode= on

log_slave_updates=1           #当一个主故障，另一个立即接管

enforce_gtid_consistency= 1

auto-increment-increment = 2

auto-increment-offset = 1

replicate-ignore-db=mysql,information_schema #不同步的数据库，多个写多行

db2

server-id= 2

gtid_mode= on

log_slave_updates=1           #当一个主故障，另一个立即接管

enforce_gtid_consistency= 1

auto-increment-increment = 2

auto-increment-offset = 2

replicate-ignore-db=mysql,information_schema #不同步的数据库，多个写多行

db3

server-id= 3

gtid_mode= on

log_slave_updates=1           #当一个主故障，另一个立即接管

enforce_gtid_consistency= 1

replicate-ignore-db=mysql,information_schema #不同步的数据库，多个写多行

read_only=1

配置主从

130 和131 上操作

CHANGE MASTER TO MASTER_HOST='192.168.146.129',MASTER_USER='myrepl',MASTER_PASSWORD='123456',MASTER_PORT=3306,MASTER_LOG_FILE='mysql-bin.000022', MASTER_CONNECT_RETRY=311;

start slave;

129上操作

130 和131 上操作

CHANGE MASTER TO MASTER_HOST='192.168.146.130',MASTER_USER='myrepl',MASTER_PASSWORD='123456',MASTER_PORT=3306,MASTER_LOG_FILE='mysql-bin.000032', MASTER_CONNECT_RETRY=234;

start slave;

配置/etc/mysql-mmm/mmm_common.conf #三机器统一

active_master_role      writer

<host default>

cluster_interface       eth0

pid_path                /var/run/mysql-mmm/mmm_agentd.pid

bin_path                /usr/libexec/mysql-mmm/

replication_user        myrepl

replication_password    123456

agent_user              mmm_agent

agent_password          agent

</host>

<host db1>

ip      192.168.146.129

mode    master

peer    db2

</host>

<host db2>

ip      192.168.146.130

mode    master

peer    db1

</host>

<host db3>

ip      192.168.146.131

mode    slave

</host>

<role writer>

hosts   db1, db2

ips     192.168.146.135

mode    exclusive

</role>

<role reader>

hosts   db1, db2, db3

ips     192.168.146.136, 192.168.146.137, 192.168.146.138

mode    balanced

</role>

配置/etc/mysql-mmm/mmm_agent.conf

this db1 #是谁就写谁

管理机上配置 mmm_mon.conf

include mmm_common.conf

<monitor>

ip                  127.0.0.1

pid_path            /var/run/mysql-mmm/mmm_mond.pid

bin_path            /usr/libexec/mysql-mmm

status_path         /var/lib/mysql-mmm/mmm_mond.status

ping_ips            192.168.146.129,192.168.146.130  #其他机器IP

auto_set_online     60

# The kill_host_bin does not exist by default, though the monitor will

# throw a warning about it missing.  See the section 5.10 "Kill Host

# Functionality" in the PDF documentation.

#

# kill_host_bin     /usr/libexec/mysql-mmm/monitor/kill_host

#

</monitor>

<host default>

monitor_user        mmm_monitor

monitor_password    monitor

</host>

debug 0

加入自启动：

chkconfig mysql-mmm-monitor on

chkconfig mysql-mmm-agent on

开启服务：

启动 mmm_agent

service mysql-mmm-agent start

启动 mmm_monitor

service mysql-mmm-monitor start

管理命令

[root@mysqlDB03 mysql-mmm]# mmm_control --help

Invalid command '--help'

Valid commands are:

help                              - show this message

ping                              - ping monitor

show                              - show status

checks [<host>|all [<check>|all]] - show checks status

set_online <host>                 - set host <host> online

set_offline <host>                - set host <host> offline

mode                              - print current mode.

set_active                        - switch into active mode.

set_manual                        - switch into manual mode.

set_passive                       - switch into passive mode.

move_role [--force] <role> <host> - move exclusive role <role> to host <host>

(Only use --force if you know what you are doing!)

set_ip <ip> <host>                - set role with ip <ip> to host <host>

[root@mysqlDB03 mysql-mmm]# mmm_control show

db1(192.168.146.129) master/ONLINE. Roles: reader(192.168.146.138), writer(192.168.146.135)

db2(192.168.146.130) master/ONLINE. Roles: reader(192.168.146.137)

db3(192.168.146.131) slave/ONLINE. Roles: reader(192.168.146.136

[root@mysqlDB03 mysql-mmm]# mmm_control checks all

db2  ping         [last change: 2016/11/10 19:18:54]  OK

db2  mysql        [last change: 2016/11/10 19:18:54]  OK

db2  rep_threads  [last change: 2016/11/10 19:18:54]  OK

db2  rep_backlog  [last change: 2016/11/10 19:18:54]  OK: Backlog is null

db3  ping         [last change: 2016/11/10 19:18:54]  OK

db3  mysql        [last change: 2016/11/10 19:19:17]  OK

db3  rep_threads  [last change: 2016/11/10 19:19:17]  OK

db3  rep_backlog  [last change: 2016/11/10 19:19:17]  OK: Backlog is null

db1  ping         [last change: 2016/11/10 19:18:54]  OK

db1  mysql        [last change: 2016/11/10 19:18:54]  OK

db1  rep_threads  [last change: 2016/11/10 19:18:54]  OK

db1  rep_backlog  [last change: 2016/11/10 19:18:54]  OK: Backlog is null

测试转换:

DB1上操作：

[root@mysqlDB01 mysql-mmm]# mysqladmin -uroot -p -S /tmp/mysql.sock shutdown

Enter password:

管理机器上查看：

Bye

[root@mysqlDB03 mysql-mmm]# mmm_control checks all

db2  ping         [last change: 2016/11/10 19:18:54]  OK

db2  mysql        [last change: 2016/11/10 19:18:54]  OK

db2  rep_threads  [last change: 2016/11/10 19:18:54]  OK

db2  rep_backlog  [last change: 2016/11/10 19:18:54]  OK: Backlog is null

db3  ping         [last change: 2016/11/10 19:18:54]  OK

db3  mysql        [last change: 2016/11/10 19:39:20]  OK

db3  rep_threads  [last change: 2016/11/10 19:39:20]  OK

db3  rep_backlog  [last change: 2016/11/10 19:19:17]  OK: Backlog is null

db1  ping         [last change: 2016/11/10 19:18:54]  OK

db1  mysql        [last change: 2016/11/10 19:31:34]  ERROR: Connect error (host = 192.168.146.129:3306, user = mmm_monitor)! Lost connection to MySQL server at 'reading initial communication packet', system error: 111

db1  rep_threads  [last change: 2016/11/10 19:18:54]  OK

db1  rep_backlog  [last change: 2016/11/10 19:18:54]  OK: Backlog is null

[root@mysqlDB03 mysql-mmm]# mmm_control show

db1(192.168.146.129) master/HARD_OFFLINE. Roles:

db2(192.168.146.130) master/ONLINE. Roles: reader(192.168.146.137), writer(192.168.146.135)

db3(192.168.146.131) slave/ONLINE. Roles: reader(192.168.146.136), reader(192.168.146.138)

启动DB1 等待60s再看:

[root@mysqlDB03 mysql-mmm]# mmm_control show

db1(192.168.146.129) master/ONLINE. Roles: reader(192.168.146.136)

db2(192.168.146.130) master/ONLINE. Roles: reader(192.168.146.137), writer(192.168.146.135)

db3(192.168.146.131) slave/ONLINE. Roles: reader(192.168.146.138)

我们想把writer角色切回到db1上：

[root@mysqlDB03 mysql-mmm]# mmm_control move_role writer db1

OK: Role 'writer' has been moved from 'db2' to 'db1'. Now you can wait some time and check new roles info!

[root@mysqlDB03 mysql-mmm]# mmm_control show

db1(192.168.146.129) master/ONLINE. Roles: reader(192.168.146.136), writer(192.168.146.135)

db2(192.168.146.130) master/ONLINE. Roles: reader(192.168.146.137)

db3(192.168.146.131) slave/ONLINE. Roles: reader(192.168.146.138)

遇到两个问题

1. 监控程序服务器无法启动

在管理服务器上，一切都完成后，通过mmm_control查看状态，得到下面的错误信息：

ERROR: Cant connect to monitor daemon! 通过编辑/etc/mysql-mmm/mmm_mon.conf文件将debug 0改为debug 1，打开监控程序的debug状态。重新启动监控程序（service mysql-mmm-monitor restart），就会看到详细的错误信息，找不到Perl Time HiRes库。执行yum -y install perl-Time-HiRes*就可以解决。

2. 防火墙问题导致Warning: agent on host db1 is not reachable.

控制台程序正确启动后，再次执行mmm_control show，却看到下面的输出：

# Warning: agent on host db1 is not reachable

# Warning: agent on host db2 is not reachable

db1(192.168.146.129) master/ONLINE. Roles:

db2(192.168.146.130) master/ONLINE. Roles:

db3(192.168.146.131) slave/ONLINE. Roles: reader(192.168.146.138)

再次打开debug，查看错误信息

通过telnet 192.168.146.130 9989下面检查网络连接，得到了No route to host的错误信息。登录db1，关闭iptables（这不是一个好主意）。同样，关闭db2、db3上的防火墙，再次重启监控程序，一切回到正常状态！