mysqlmmm部署.docx

mysqlmmm部署.docx

《mysqlmmm部署.docx》由会员分享，可在线阅读，更多相关《mysqlmmm部署.docx（13页珍藏版）》请在冰豆网上搜索。

mysqlmmm部署

linux下安装mysqlmmm

环境:

OS:

RadHatLinuxAs5

1.安装步骤

1.1环境部署

节点角色

ip

db1

192.168.56.111

db2

192.168.56.112

db3

192.168.56.113

Vip

Role

Description

192.168.56.250

Writer

应用程序通过该ip写入数据

192.168.56.251

Reader

应用程序通过该ip读取数据

192.168.56.252

Reader

应用程序通过该ip读取数据

DB1从正常运行到宕机，再正常运行的整个过程：

DB1宕机后,mmm会将写入的VIP地址飘逸到DB2,DB3从库对应的主库由之前的DB1转变为DB2,当DB1恢复后,DB1和DB2互为主从，DB1宕机期间在DB2上的操作会自动同步到DB1，但是DB3从库对应的主库不会自动指向DB1，同时写入的VIP也不会转移到DB1,这个需要在下次DB2宕机之后才会指向DB1.

1.2mmm下载

mmm下载地址:

1.3建立各主机的等效连接

步骤省略

1.3.1验证等效连接

1.3.1.1名称节点到各节点的等效性

[root@host01~]#sshhost02;host03date

Lastlogin:

MonApr1811:

27:

012016from

[root@host02~]#sshhost01;host03date

Lastlogin:

MonApr1810:

48:

312016from

[root@host03~]#sshhost01;host02date

Lastlogin:

MonApr1814:

52:

302016from

1.4在3台机器上安装mysql

分别在db1、db2、db3上安装mysql

安装步骤可以参考

1.5配置各机器配置文件

1.5.1DB1

vi/db/mmm/mysql/conf/f内容如下

[mysqld]

server_id=1001

log-bin=binlog

log-bin-index=binlog.index

port=3306

basedir=/db/mmm/mysql

datadir=/db/mmm/mysql/data

socket=/db/mmm/mysql/mysql.sock

character-set-server=utf8

max_connections=1500

show_compatibility_56=on

replicate-wild-ignore-table=mysql.%

replicate-wild-ignore-table=information_schema.%

replicate-wild-ignore-table=performance_schema.%

replicate-wild-ignore-table=sys.%

slave-net-timeout=3600

##master-connect-retry=605.1版本以后不支持该参数

##master-retry-count=864005.1版本以后不支持该参数

sync_binlog=1

innodb_flush_log_at_trx_commit=1

1.5.2DB2

vi/db/mmm/mysql/conf/f内容如下

[mysqld]

server_id=1002

log-bin=binlog

log-bin-index=binlog.index

port=3306

basedir=/db/mmm/mysql

datadir=/db/mmm/mysql/data

socket=/db/mmm/mysql/mysql.sock

character-set-server=utf8

max_connections=1500

show_compatibility_56=on

replicate-wild-ignore-table=mysql.%

replicate-wild-ignore-table=information_schema.%

replicate-wild-ignore-table=performance_schema.%

replicate-wild-ignore-table=sys.%

slave-net-timeout=3600

##master-connect-retry=605.1版本以后不支持该参数

##master-retry-count=864005.1版本以后不支持该参数

sync_binlog=1

innodb_flush_log_at_trx_commit=1

1.5.3DB3

vi/db/mmm/mysql/conf/f内容如下

[mysqld]

server_id=1003

read_only=1##设置为只读,跟db1db2的差别

log-bin=binlog

log-bin-index=binlog.index

port=3306

basedir=/db/mmm/mysql

datadir=/db/mmm/mysql/data

socket=/db/mmm/mysql/mysql.sock

character-set-server=utf8

max_connections=1500

show_compatibility_56=on

replicate-wild-ignore-table=mysql.%

replicate-wild-ignore-table=information_schema.%

replicate-wild-ignore-table=performance_schema.%

replicate-wild-ignore-table=sys.%

slave-net-timeout=3600

##master-connect-retry=605.1版本以后不支持该参数

##master-retry-count=864005.1版本以后不支持该参数

1.6启动mysql

启动各DB

[root@host01bin]#./mysqld_safe--defaults-file=/db/mmm/mysql/conf/f--user=mysql

登陆各DB

./mysql-hlocalhost-uroot-p-S/db/mmm/mysql/mysql.sock

1.7主从配置（db1和db2配置为主主,db3配置为db1的从）

1.7.1授权

在db1和db2上执行：

grantreplicationslaveON*.*TO'repl'@'%'identifiedby'mysql';

flushprivileges;

1.7.2db3配置为db1的从

changemastertomaster_host='192.168.56.111',

master_port=3306,

master_user='repl',

master_password='mysql',

master_log_file='binlog.000002',

master_log_pos=746;

startslave;

1.7.3db1和db2配置为主主

db1:

changemastertomaster_host='192.168.56.112',

master_port=3306,

master_user='repl',

master_password='mysql',

master_log_file='binlog.000002',

master_log_pos=746;

startslave;

db2:

changemastertomaster_host='192.168.56.111',

master_port=3306,

master_user='repl',

master_password='mysql',

master_log_file='binlog.000002',

master_log_pos=746;

startslave;

1.8安装mmm

db1、db2、db3都要安装mysql-mmm-agent，我们这里db1还需要充当monitor的角色，所以在db1上还需要安装mysql-mmm-monitor.

1.8.1DB1

------------依赖包-----------------------------------

rpm-ivhperl-Algorithm-Diff-1.1902-1.el5.rf.noarch.rpm

rpm-ivhperl-Class-Singleton-1.4-1.el5.rf.noarch.rpm

rpm-ivhperl-Proc-Daemon-0.03-1.2.el5.rf.noarch.rpm

rpm-ivhperl-Net-ARP-1.0-1.el5.rf.x86_64.rpm

rpm-ivhperl-Log-Log4perl-1.33-1.el5.pp.noarch.rpm

rpm-ivhperl-Path-Class-0.16-1.el5.rf.noarch.rpm

------------mmm安装包----------------------------

rpm-ivhmysql-mmm-2.2.1-1.el5.noarch.rpm

rpm-ivhmysql-mmm-tools-2.2.1-1.el5.noarch.rpm

rpm-ivhmysql-mmm-agent-2.2.1-1.el5.noarch.rpm

rpm-ivhmysql-mmm-monitor-2.2.1-1.el5.noarch.rpm

1.8.2DB2

------------依赖包-----------------------------------

rpm-ivhperl-Algorithm-Diff-1.1902-1.el5.rf.noarch.rpm

rpm-ivhperl-Class-Singleton-1.4-1.el5.rf.noarch.rpm

rpm-ivhperl-Proc-Daemon-0.03-1.2.el5.rf.noarch.rpm

rpm-ivhperl-Net-ARP-1.0-1.el5.rf.x86_64.rpm

rpm-ivhperl-Log-Log4perl-1.33-1.el5.pp.noarch.rpm

rpm-ivhperl-Path-Class-0.16-1.el5.rf.noarch.rpm

------------mmm安装包----------------------------

rpm-ivhmysql-mmm-2.2.1-1.el5.noarch.rpm

rpm-ivhmysql-mmm-tools-2.2.1-1.el5.noarch.rpm

rpm-ivhmysql-mmm-agent-2.2.1-1.el5.noarch.rpm

1.8.3DB3

------------依赖包-----------------------------------

rpm-ivhperl-Algorithm-Diff-1.1902-1.el5.rf.noarch.rpm

rpm-ivhperl-Class-Singleton-1.4-1.el5.rf.noarch.rpm

rpm-ivhperl-Proc-Daemon-0.03-1.2.el5.rf.noarch.rpm

rpm-ivhperl-Net-ARP-1.0-1.el5.rf.x86_64.rpm

rpm-ivhperl-Log-Log4perl-1.33-1.el5.pp.noarch.rpm

rpm-ivhperl-Path-Class-0.16-1.el5.rf.noarch.rpm

------------mmm安装包----------------------------

rpm-ivhmysql-mmm-2.2.1-1.el5.noarch.rpm

rpm-ivhmysql-mmm-tools-2.2.1-1.el5.noarch.rpm

rpm-ivhmysql-mmm-agent-2.2.1-1.el5.noarch.rpm

1.9mmm配置

1.9.1授权

在db1db2db3上执行如下授权语句：

grantsuper,replicationclient,processon*.*to'mmm_agent'@'192.168.56.%'identifiedby'mysql';

grantreplicationclienton*.*to'mmm_monitor'@'192.168.56.%'identifiedby'mysql';

flushprivileges;

1.9.2编辑配置文件

1.9.2.1mmm_common.conf

db1、db2、db3上都配置一样，可以在db1上配置好后，拷贝到db2、db3.

[root@host01mysql-mmm]#more/etc/mysql-mmm/mmm_common.conf

active_master_rolewriter

cluster_interfaceeth0

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

bin_path/usr/libexec/mysql-mmm/

replication_userrepl##复制的账号

replication_passwordmysql##复制账号密码

agent_usermmm_agent##跟上面步骤创建的用户保持一致

agent_passwordmysql

ip192.168.56.111##db1的ip

modemaster

peerdb2

ip192.168.56.112##db2的ip

modemaster

peerdb1

ip192.168.56.113##db3的ip

modeslave

hostsdb1,db2

ips192.168.56.250##设置写入的虚拟IP

modeexclusive

hostsdb1,db2

ips192.168.56.251,192.168.56.252##设置读取的虚拟IP

modebalanced

1.9.2.2mmm_agent.conf

在db1、db2、db3上都需要配置，配置分别如下：

db1

[root@host01mysql-mmm]#more/etc/mysql-mmm/mmm_agent.conf

includemmm_common.conf

#The'this'variablereferstothisserver.Properoperationrequires

#that'this'server（db1bydefault）,aswellasallotherservers,havethe

#properIPaddressessetinmmm_common.conf.

thisdb1

db2

[root@host02mysql-mmm]#moremmm_agent.conf

includemmm_common.conf

#The'this'variablereferstothisserver.Properoperationrequires

#that'this'server（db1bydefault）,aswellasallotherservers,havethe

#properIPaddressessetinmmm_common.conf.

thisdb2

db3

[root@host03mysql-mmm]#moremmm_agent.conf

includemmm_common.conf

#The'this'variablereferstothisserver.Properoperationrequires

#that'this'server（db1bydefault）,aswellasallotherservers,havethe

#properIPaddressessetinmmm_common.conf.

thisdb3

1.9.2.3mmm_mon.conf

仅在monitor的节点上配置，我们这里db1充当monitor的角色，所以只需要在db1上做配置。

[root@host01mysql-mmm]#more/etc/mysql-mmm/mmm_mon.conf

includemmm_common.conf

ip127.0.0.1

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_ips192.168.56.111,192.168.56.112##监控的双主db的ip地址

auto_set_online60

#Thekill_host_bindoesnotexistbydefault,thoughthemonitorwill

#throwawarningaboutitmissing.Seethesection5.10"KillHost

#Functionality"inthePDFdocumentation.

#

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

#

monitor_usermmm_monitor##监控账号

monitor_passwordmysql##监控账号密码

debug0

1.10启动mmm

1.10.1启动agent

在db1db2db3上执行如下命令

[root@host01init.d]#/etc/init.d/mysql-mmm-agentstart

StartingMMMAgentDaemon:

[OK]

1.10.2启动监控

在监控机器上执行如下命令,我们这里是采用db1做监控机，就在该机器上执行即可。

[root@host01init.d]#/etc/init.d/mysql-mmm-monitorstart

StartingMMMMonitorDaemon:

[OK]

1.10.3查看监控状态

在监控机器上执行命令

[root@host01init.d]#mmm_controlshow

db1（192.168.56.111）master/ONLINE.Roles:

reader（192.168.56.251）,writer（192.168.56.250）

db2（192.168.56.112）master/ONLINE.Roles:

reader（192.168.56.252）

db3（192.168.56.113）slave/ONLINE.Roles:

1.11故障模拟

停止db1看下192.168.56.250看下是否能否漂移到db2上去，同时查看DB3的Slave的IP会不会从DB1改到DB2。

停止db1:

[root@host01init.d]#mysqladmin-hlocalhost-uroot-p-S/db/mmm/mysql/mysql.sockshutdown

查看mmm状态

[root@host01init.d]#mmm_controlshow

db1（192.168.56.111）master/HARD_OFFLINE.Roles:

db2（192.168.56.112）master/ONLINE.Roles:

reader（192.168.56.251）,reader（192.168.56.252）,writer（192.168.56.250）

db3（192.168.56.113）slave/ONLINE.Roles:

们可以看到当db1恢复后就充当slave的角色了！

只有当db2挂了以后db1又会担当起主服务器的写入功能

查看db3复制的状态

mysql>showslavestatus\G;

***************************1.row***************************

Slave_IO_State:

Waitingformastertosendevent

Master_Host:

192.168.56.112

Master_User:

repl

Master_Port:

3306

Connect_Retry:

60

Master_Log_File:

binlog.000004

Read_Master_Log_Pos:

692

Relay_Log_File:

host03-relay-bin.000002

Relay_Log_Pos:

317

Relay_Master_Log_File:

binlog.000004

Slave_IO_Running:

Yes

Slave_SQL_Ru