一、 說明
解決數據庫單mysql主節點的單點故障問題。便於數據庫的切換。
二、 原理
從mysql的主從模式延伸而來，兩個mysql節點互爲主從，可讀可寫。
三、 測試環境描述
192.168.0.54（db54） CentOS_6.5x64 Percona_mysql-5.5.38
192.168.0.108 （db108） CentOS_6.5x64 Percona_mysql-5.5.38
四、 配置過程
1. 安裝mysql（可以使用之前寫的一鍵安裝腳本）
2. 檢查兩臺DB是否開啓bin_log
mysql>show variables like 'log_bin';
+---------------+-------+
|Variable_name | Value |
+---------------+-------+
|log_bin | ON |
+---------------+-------+
3. 兩臺服務器修改server-id，並重啓mysql
db54修改爲：server-id = 54
db108修改爲：server-id = 108
重啓：
# /etc/init.d/mysqlrestart
4. 兩臺服務器之間相互開啓3306端口
db54：
-AINPUT -s 192.168.0.108/32 -m state --state NEW -m tcp -p tcp --dport 3306 -jACCEPT
db108：
-AINPUT -s 192.168.0.54/32 -m state --state NEW -m tcp -p tcp --dport 3306 -jACCEPT
5. 以db54爲主庫，db108爲從庫
5.1 主庫db54建立slave同步數據的用戶
mysql>grant replication client,replication slave on *.* to repl@'192.168.0.108'identified by '123456';
mysql>flush privileges;
5.2 清空日誌
mysql>flush master;
5.3 從庫db108指定主服務器
mysql>change master to master_host='192.168.0.54',master_user='repl',master_password='123456';
5.4 啓動從服務器進程並查看運行狀態
mysql>start slave;
mysql>show slave status\G
如果出現如下行，則表明正常啓動：
Slave_IO_Running: Yes
Slave_SQL_Running: Yes
5.5 測試db54到db108的主從同步
5.5.1 主庫db54上創建jibuqi數據庫：
mysql>create database jibuqi;
mysql>showdatabases;
5.5.2 從庫db108查看結果：
mysql>showdatabases;
結果正常。
6. 以db108爲主庫，db54爲從庫
6.1 db54的[mysqld]中配置文件中添加配置：
auto-increment-increment = 2 #整個結構中服務器的總數
auto-increment-offset = 1 #設定數據庫中自動增長的起點，兩個db不能相同，否則主鍵衝突
replicate-do-db = jibuqi #指定同步的數據庫，其他數據庫不同步
重啓mysql：
# /etc/init.d/mysqlrestart
mysql>show master status;
+------------------+----------+--------------+------------------+
|File | Position |Binlog_Do_DB | Binlog_Ignore_DB |
+------------------+----------+--------------+------------------+
|mysql-bin.000003 | 107 | | |
+------------------+----------+--------------+------------------+
6.2 db108創建同步數據的mysql用戶：
mysql>grant replication client,replication slave on *.* to repl@'192.168.0.108'identified by '123456';
mysql>flush privileges;
db108的配置文件[mysqld]中添加配置：
log-bin = mysql-bin
auto-increment-increment = 2
auto-increment-offset = 2 # 與db54不能相同
replicate-do-db = jibuqi
重啓mysql：
# /etc/init.d/mysqlrestart
mysql>show master status;
+------------------+----------+--------------+------------------+
|File | Position |Binlog_Do_DB | Binlog_Ignore_DB |
+------------------+----------+--------------+------------------+
|mysql-bin.000005 | 107 | | |
+------------------+----------+--------------+------------------+
6.3 db54和db108分別指定對方爲自己的主數據庫：
db108服務器的指向：
mysql>stop slave;
mysql>change master to master_host='192.168.0.54',master_user='repl',master_password='123456',master_log_file='mysql-bin.000003',master_log_pos=107;
mysql>start slave;
db54服務器的指向：
mysql>change master to master_host='192.168.0.108',master_user='repl',master_password='123456',master_log_file='mysql-bin.000005',master_log_pos=107;
mysql>start slave;
6.4 測試：
db54的jibuqi數據庫導入數據表api_pedometeraccount，檢查db108上是是否有相應table（檢查結果正常）。
db108的jibuqi數據庫導入數據表api_pedometerdevice，檢查db54上是是否有相應table（檢查結果正常）。
至此，雙主同步的模式完成。
五、 mysql-MMM（Master-Master Replication Manager for MySQL）
5.1 簡介
MMM即Master-Master Replication Manager for MySQL（mysql主主複製管理器）關於mysql主主複製配置的監控、故障轉移和管理的一套可伸縮的腳本套件（在任何時候只有一個節點可以被寫入），這個套件也能對居於標準的主從配置的任意數量的從服務器進行讀負載均衡，所以你可以用它來在一組居於複製的服務器啓動虛擬ip，除此之外，它還有實現數據備份、節點之間重新同步功能的腳本。MySQL本身沒有提供replication failover的解決方案，通過MMM方案能實現服務器的故障轉移，從而實現mysql的高可用。MMM不僅能提供浮動IP的功能，更可貴的是如果當前的主服務器掛掉後，會將你後端的從服務器自動轉向新的主服務器進行同步複製，不用手工更改同步配置。這個方案是目前比較成熟的解決方案。詳情請看官網：http://mysql-mmm.org
5.2 結構說明
a. 服務器列表
服務器 | 主機名 | IP | serverID | mysql版本 | 系統 |
master1 | db1 | 192.168.1.19 | 54 | mysql5.5 | Centos 6.5 |
master2 | db2 | 192.168.1.20 | 108 | mysql5.5 | Centos 6.5 |
b. 虛擬IP列表
VIP | Role | description |
192.168.1.190 | write | 應用配置的寫入VIP |
192.168.1.201 | read | 應用配置的讀入VIP |
192.168.1.203 | read | 應用配置的讀入VIP |
5.3 MMM的安裝（安裝在db1中，可以單獨準備一臺服務器來安裝）
a. 升級perl模塊
# yuminstall –y cpan
# cpan-i Algorithm::Diff Class::Singleton DBI DBD::mysql Log::Dispatch Log::Log4perlMail::Send Net::Ping Proc::Daemon Time::HiRes Params::Validate Net::ARP
這兩個File::Basename File::stat模塊好像安裝不上，要升級5.12.2纔可以，不安裝也可以使用；Net:ARP必須要安裝，要不然VIP會出不來的。
b. 下載並安裝mysql-mmm
# wget http://mysql-mmm.org/_media/:mmm2:mysql-mmm-2.2.1.tar.gz-O mysql-mmm-2.2.1.tar.gz
# tar-zxvf mysql-mmm-2.2.1.tar.gz
# cdmysql-mmm-2.2.1
# make;makeinstall
c. mysql-mmm配置
直接修改配置文件
· 修改db1的配置文件：
mmm_agent.conf：
# vim/etc/mysql_mmm/mmm_agent
include mmm_common.conf
this db1
mmm_common.conf：
# vim/etc/mysql_mmm/mmm_common.conf
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 repl
replication_password 123456
agent_user mmm_agent
agent_password mmm_agent
</host>
<host db1>
ip 192.168.1.19
mode master
peer db2
</host>
<host db2>
ip 192.168.1.20
mode master
peer db1
</host>
<role writer>
hosts db1, db2
ips 192.168.1.190
mode exclusive # exclusive是排他模式，任何時候只能一個host擁有該角色
</role>
<role reader>
hosts db1, db2
ips 192.168.1.201,192.168.1.203
mode balanced # balanced均衡模式，可以多個host同時擁有此角色
</role>
mmm_mon.conf
# vim/etc/mysql_mmm/mmm_mon.conf
include mmm_common.conf
<monitor>
ip 127.0.0.1
pid_path /var/run/mmm_mond.pid
bin_path /usr/lib/mysql-mmm/
status_path /var/lib/misc/mmm_mond.status
auto_set_online 5 #自動切換的時間，單位爲秒
ping_ips 192.168.1.19, 192.168.1.20
</monitor>
<host default>
monitor_user mmm_monitor
monitor_password mmm_monitor
</host>
debug 0
· 修改db2的配置文件：
mmm_agent.conf：
# vim /etc/mysql_mmm/mmm_agent
includemmm_common.conf
thisdb2
mmm_common.conf的配置文件內容與db1的相同。
d. 啓動mmm程序：
db1啓動agent和mon：
# /etc/init.d/mysql-mmm-agent start
Daemonbin: '/usr/sbin/mmm_agentd'
Daemonpid: '/var/run/mmm_agentd.pid'
StartingMMM Agent daemon... Ok
# /etc/init.d/mysql-mmm-monitor start
Daemon bin: '/usr/sbin/mmm_mond'
Daemon pid: '/var/run/mmm_mond.pid'
Starting MMM Monitor daemon: Baseclass package "Class::Singleton" is empty.
(Perhaps you need to 'use' the module which defines that package first,
or make that module available in @INC (@INC contains:/usr/local/lib64/perl5 /usr/local/share/perl5 /usr/lib64/perl5/vendor_perl/usr/share/perl5/vendor_perl /usr/lib64/perl5 /usr/share/perl5 .).
at/usr/share/perl5/vendor_perl/MMM/Monitor/Agents.pm line 2
BEGIN failed--compilation abortedat /usr/share/perl5/vendor_perl/MMM/Monitor/Agents.pm line 2.
Compilation failed in require at/usr/share/perl5/vendor_perl/MMM/Monitor/Monitor.pm line 15.
BEGIN failed--compilation abortedat /usr/share/perl5/vendor_perl/MMM/Monitor/Monitor.pm line 15.
Compilation failed in require at/usr/sbin/mmm_mond line 28.
BEGIN failed--compilation abortedat /usr/sbin/mmm_mond line 28.
Failed
啓動mysql-mmm-monitor失敗，修復方法：
# perl -MCPAN -e shell
Terminaldoes not support AddHistory.
cpanshell -- CPAN exploration and modules installation (v1.9402)
Enter'h' for help.
cpan[1]> Class::Singleton
Catchingerror: "Can't locate object method \"Singleton\" via package\"Class\" (perhaps you forgot to load \"Class\"?) at/usr/share/perl5/CPAN.pm line 375, <FIN> line 1.\cJ" at /usr/share/perl5/CPAN.pmline 391
CPAN::shell() called at -e line 1
cpan[2]>Class
Unknownshell command 'Class'. Type ? for help.
cpan[3]> install Class::Singleton
CPAN:Storable loaded ok (v2.20)
Goingto read '/root/.cpan/Metadata'
Database was generated on Thu, 27 Nov 201408:53:16 GMT
CPAN:LWP::UserAgent loaded ok (v5.833)
CPAN:Time::HiRes loaded ok (v1.9726)
Warning:no success downloading '/root/.cpan/sources/authors/01mailrc.txt.gz.tmp47425'.Giving up on it. at /usr/share/perl5/CPAN/Index.pm line 225
Fetchingwith LWP:
http://www.perl.org/CPAN/authors/01mailrc.txt.gz
Goingto read '/root/.cpan/sources/authors/01mailrc.txt.gz'
............................................................................DONE
Fetchingwith LWP:
http://www.perl.org/CPAN/modules/02packages.details.txt.gz
Goingto read '/root/.cpan/sources/modules/02packages.details.txt.gz'
Database was generated on Fri, 28 Nov 201408:29:02 GMT
..............
New CPAN.pm version (v2.05) available.
[Currently running version is v1.9402]
You might want to try
install CPAN
reload cpan
to both upgrade CPAN.pm and run the newversion without leaving
the current session.
..............................................................DONE
Fetchingwith LWP:
http://www.perl.org/CPAN/modules/03modlist.data.gz
Goingto read '/root/.cpan/sources/modules/03modlist.data.gz'
DONE
Goingto write /root/.cpan/Metadata
Runninginstall for module 'Class::Singleton'
CPAN:Data::Dumper loaded ok (v2.124)
'YAML'not installed, falling back to Data::Dumper and Storable to read prefs'/root/.cpan/prefs'
Runningmake for S/SH/SHAY/Class-Singleton-1.5.tar.gz
CPAN:Digest::SHA loaded ok (v5.47)
Checksumfor /root/.cpan/sources/authors/id/S/SH/SHAY/Class-Singleton-1.5.tar.gz ok
Scanningcache /root/.cpan/build for sizes
............................................................................DONE
Class-Singleton-1.5/
Class-Singleton-1.5/Changes
Class-Singleton-1.5/lib/
Class-Singleton-1.5/lib/Class/
Class-Singleton-1.5/lib/Class/Singleton.pm
Class-Singleton-1.5/Makefile.PL
Class-Singleton-1.5/MANIFEST
Class-Singleton-1.5/META.yml
Class-Singleton-1.5/README
Class-Singleton-1.5/t/
Class-Singleton-1.5/t/singleton.t
CPAN: File::Temploaded ok (v0.22)
CPAN.pm: Going to buildS/SH/SHAY/Class-Singleton-1.5.tar.gz
Checkingif your kit is complete...
Looksgood
Generatinga Unix-style Makefile
WritingMakefile for Class::Singleton
WritingMYMETA.yml and MYMETA.json
Couldnot read '/root/.cpan/build/Class-Singleton-1.5-42kiLS/MYMETA.yml'. Fallingback to other methods to determine prerequisites
cplib/Class/Singleton.pm blib/lib/Class/Singleton.pm
Manifying1 pod document
SHAY/Class-Singleton-1.5.tar.gz
/usr/bin/make -- OK
Warning(usually harmless): 'YAML' not installed, will not store persistent state
Runningmake test
PERL_DL_NONLAZY=1"/usr/bin/perl" "-MExtUtils::Command::MM""-MTest::Harness" "-e" "undef*Test::Harness::Switches; test_harness(0, 'blib/lib', 'blib/arch')" t/*.t
t/singleton.t.. ok
Alltests successful.
Files=1,Tests=29, 0 wallclock secs ( 0.01usr 0.01 sys + 0.01 cusr 0.00 csys = 0.03 CPU)
Result:PASS
SHAY/Class-Singleton-1.5.tar.gz
/usr/bin/make test -- OK
Warning(usually harmless): 'YAML' not installed, will not store persistent state
Runningmake install
Prepending/root/.cpan/build/Class-Singleton-1.5-42kiLS/blib/arch/root/.cpan/build/Class-Singleton-1.5-42kiLS/blib/lib to PERL5LIB for 'install'
Manifying1 pod document
Installing/usr/local/share/perl5/Class/Singleton.pm
Installing/usr/local/share/man/man3/Class::Singleton.3pm
Appendinginstallation info to /usr/lib64/perl5/perllocal.pod
SHAY/Class-Singleton-1.5.tar.gz
/usr/bin/make install -- OK
Warning(usually harmless): 'YAML' not installed, will not store persistent state
cpan[4]>exit
Terminaldoes not support GetHistory.
Lockfileremoved.
#/etc/init.d/mysql-mmm-monitor start
Daemonbin: '/usr/sbin/mmm_mond'
Daemonpid: '/var/run/mmm_mond.pid'
StartingMMM Monitor daemon: Ok
db2啓動agent：
# /etc/init.d/mysql-mmm-agent start
Daemonbin: '/usr/sbin/mmm_agentd'
Daemonpid: '/var/run/mmm_agentd.pid'
StartingMMM Agent daemon... Ok
e. 修改防火牆，根據情況開放mmm端口（方法略）
f. db1和db2添加監控授權用戶，用於檢測mysql狀態。
mysql> grant super,replicationclient,process on *.* to 'mmm_agent'@'192.168.1.20' identified by 'mmm_agent';
mysql> grant super,replication client,processon *.* to 'mmm_agent'@'192.168.1.19' identified by 'mmm_agent';
mysql > grant super,replicationclient,process on *.* to 'mmm_agent'@'localhost' identified by 'mmm_agent';
mysql > grant super,replicationclient,process on *.* to 'mmm_agent'@'127.0.0.1' identified by 'mmm_agent';
mysql > grant super,replicationclient,process on *.* to 'mmm_monitor'@'192.168.1.20' identified by 'mmm_monitor';
mysql > grant super,replication client,processon *.* to 'mmm_monitor'@'192.168.1.19' identified by 'mmm_monitor’;
mysql > grant super,replicationclient,process on *.* to 'mmm_monitor'@'localhost' identified by 'mmm_ monitor';
mysql > grant super,replicationclient,process on *.* to 'mmm_ monitor'@'127.0.0.1' identified by 'mmm_ monitor';
mysql > flush privileges;
g. 檢查狀態（在monitor所在的機器）：
# mmm_control ping
OK: Pinged successfully!
# mmm_control show
db1(192.168.1.19) master/ONLINE. Roles: reader(192.168.1.203),writer(192.168.1.190)
db2(192.168.1.20) master/ONLINE. Roles: reader(192.168.1.201)
# mmm_control checks
db2 ping [last change:2014/12/01 13:49:47] OK
db2 mysql [last change:2014/12/01 13:49:47] OK
db2 rep_threads [last change:2014/12/01 13:49:47] OK
db2 rep_backlog [last change:2014/12/01 13:49:47] OK: Backlog is null
db1 ping [last change:2014/12/01 13:49:47] OK
db1 mysql [last change:2014/12/01 13:49:47] OK
db1 rep_threads [last change:2014/12/01 13:52:19] OK
db1 rep_backlog [last change:2014/12/01 13:49:47] OK: Backlog is null
# mmm_control 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>
(Onlyuse --force if you know what you are doing!)
set_ip <ip><host> - set rolewith ip <ip> to host <host>
h. 測試切換
DB1上的信息如下：
#mmm_control show
db1(192.168.1.19) master/ONLINE. Roles:reader(192.168.1.203), writer(192.168.1.190)
db2(192.168.1.20) master/ONLINE. Roles:reader(192.168.1.201)
# ip a
1: lo:<LOOPBACK,UP,LOWER_UP> mtu 16436 qdisc noqueue state UNKNOWN
link/loopback 00:00:00:00:00:00 brd00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
2:eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast stateUP qlen 1000
link/ether 00:0c:29:93:d2:50 brdff:ff:ff:ff:ff:ff
inet 192.168.1.19/23 brd 192.168.1.255scope global eth0
inet 192.168.1.203/32 scope global eth0
inet 192.168.1.190/32 scope global eth0
inet6 fe80::20c:29ff:fe93:d250/64 scopelink
valid_lft forever preferred_lft forever
DB2上的信息:
# ip a
1: lo:<LOOPBACK,UP,LOWER_UP> mtu 16436 qdisc noqueue state UNKNOWN
link/loopback 00:00:00:00:00:00 brd00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
2:eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast stateUP qlen 1000
link/ether 00:0c:29:9f:7c:c6 brdff:ff:ff:ff:ff:ff
inet 192.168.1.20/23 brd 192.168.1.255scope global eth0
inet 192.168.1.201/32 scope global eth0
inet6 fe80::20c:29ff:fe9f:7cc6/64 scopelink
valid_lft forever preferred_lft forever
停掉DB1上的mysql應用，看mmm是否會把所有vips切換到DB2：
DB1上：
#/etc/init.d/mysql stop
Shuttingdown MySQL (Percona Server)..... SUCCESS!
# ip a
1: lo:<LOOPBACK,UP,LOWER_UP> mtu 16436 qdisc noqueue state UNKNOWN
link/loopback 00:00:00:00:00:00 brd00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
2:eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast stateUP qlen 1000
link/ether 00:0c:29:93:d2:50 brdff:ff:ff:ff:ff:ff
inet 192.168.1.19/23 brd 192.168.1.255scope global eth0
inet6 fe80::20c:29ff:fe93:d250/64 scopelink
valid_lft forever preferred_lft forever
DB2上：
# ip a
1: lo:<LOOPBACK,UP,LOWER_UP> mtu 16436 qdisc noqueue state UNKNOWN
link/loopback 00:00:00:00:00:00 brd00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
2:eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast stateUP qlen 1000
link/ether 00:0c:29:9f:7c:c6 brdff:ff:ff:ff:ff:ff
inet 192.168.1.20/23 brd 192.168.1.255scope global eth0
inet 192.168.1.201/32 scope global eth0
inet 192.168.1.203/32 scope global eth0
inet 192.168.1.190/32 scope global eth0
inet6 fe80::20c:29ff:fe9f:7cc6/64 scopelink
valid_lft forever preferred_lft forever
vip切換成功。mmm的切換功能成功。
如果想實現mysql的讀寫分離，可以通過mysql_proxy實現。