Mysql高可用架构MHA搭建及测试故障转移

MHA项目网站

https://code.google.com/p/mysql-master-ha/

一.环境介绍

1.主机部署

manager机:10.10.54.154
master机:10.10.54.156
slave1机:10.10.54.155(备用master)

slave2机:10.10.54.157

2.大致步骤
A.首先用ssh-keygen实现四台主机之间相互免密钥登录
B.安装MHAmha4mysql-node，mha4mysql-manager 软件包
C.建立master，slave1，slave2之间主从复制
D.管理机manager上配置MHA文件
E.masterha_check_ssh工具验证ssh信任登录是否成功
F.masterha_check_repl工具验证mysql复制是否成功
G.启动MHA manager,并监控日志文件
H.测试master(156)宕机后，是否会自动切换

3.说明：下面中括号中的主机名说明了当前操作是在哪台机子上进行的

二.首先用ssh-keygen实现四台主机之间相互免密钥登录

[manager机]
shell> ssh-keygen -t rsa -b 2048
shell> scp-copy-id root@10.10.54.155

shell> scp-copy-id root@10.10.54.156

shell> scp-copy-id root@10.10.54.157

在另外三台机子重复此步骤，使四台机子中的任何两台之间可以免密码登录

三.安装MHAmha4mysql-node，mha4mysql-manager 软件包

1.四台主机上安装MHAmha4mysql-node

1 2 3 4 5

[manager,master,slave1,slave2] shell> yum update shell> yum -y  install  perl-DBD-MySQL ncftp shell> wget http: //mysql-master-ha .googlecode.com /files/mha4mysql-node-0 .53-0.noarch.rpm sehll> rpm -ivh mha4mysql-node-0.53-0.noarch.rpm

2.在manager机子上安装mha4mysql-manager

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

[manager] shell> yum  install  perl shell> yum  install  cpan shell> rpm -ivh mha4mysql-manager-0.53-0.el6.noarch.rpm error: perl(Config::Tiny) is needed by mha4mysql-manager-0.53-0.noarch perl(Log::Dispatch) is needed by mha4mysql-manager-0.53-0.noarch perl(Log::Dispatch::File) is needed by mha4mysql-manager-0.53-0.noarch perl(Log::Dispatch::Screen) is needed by mha4mysql-manager-0.53-0.noarch perl(Parallel::ForkManager) is needed by mha4mysql-manager-0.53-0.noarch perl(Time::HiRes) is needed by mha4mysql-manager-0.53-0.noarch [solution] shell> wget  ftp : //ftp .muug.mb.ca /mirror/centos/5 .10 /os/x86_64/CentOS/perl-5 .8.8-41.el5.x86_64.rpm shell> wget  ftp : //ftp .muug.mb.ca /mirror/centos/6 .5 /os/x86_64/Packages/compat-db43-4 .3.29-15.el6.x86_64.rpm shell> wget http: //downloads .naulinux.ru /pub/NauLinux/6x/i386/sites/School/RPMS/perl-Log-Dispatch-2 .27-1.el6.noarch.rpm shell> wget http: //dl .fedoraproject.org /pub/epel/6/i386/perl-Parallel-ForkManager-0 .7.9-1.el6.noarch.rpm shell> wget http: //dl .fedoraproject.org /pub/epel/6/i386/perl-Mail-Sender-0 .8.16-3.el6.noarch.rpm shell> wget http: //dl .fedoraproject.org /pub/epel/6/i386/perl-Mail-Sendmail-0 .79-12.el6.noarch.rpm shell> wget http: //mirror .centos.org /centos/6/os/x86_64/Packages/perl-Time-HiRes-1 .9721-136.el6.x86_64.rpm shell> rpm -ivh perl-Parallel-ForkManager-0.7.9-1.el6.noarch.rpm perl-Log-Dispatch-2.27-1.el6.noarch.rpm perl-Mail-Sender-0.8.16-3.el6.noarch.rpm perl-Mail-Sendmail-0.79-12.el6.noarch.rpm perl-Time-HiRes-1.9721-136.el6.x86_64.rpm shell> rpm -ivh mha4mysql-manager-0.53-0.el6.noarch.rpm

四.建立master，slave1，slave2之间主从复制

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

[master:156] 1.shell> vim  /etc/my .cnf #server-id 改为1 server- id =1 log-bin=mysql-bin binlog_format=mixed #授权操作 2.mysql> GRANT ALL PRIVILEGES ON *.* TO  'rep' @ '10.10.54.%'  IDENTIFIED BY  'rep123' ; mysql> flush privileges; 3.mysql> show master status; [slave1,slave2] 4.change master操作 mysql> change master to master_host= '10.10.54.156' , master_port=3306, master_user= 'rep' , master_password= 'rep123' , master_log_file= 'mysql-bin.000001' , master_log_pos=112;

注意：slave1机子上也要授权，因为这个是备用master
[slave1:155]
5.mysql> GRANT ALL PRIVILEGES ON *.* TO 'rep'@'10.10.54.%' IDENTIFIED BY 'rep123';

[master,slave1,slave2]
6.查看主从复制是否成功的一些命令
mysql> start slave;
mysql> stop slave;
mysql> reset slave;
mysql> show slave status\G;

五.所有主机上设置复制权限帐号
mysql> GRANT ALL PRIVILEGES ON *.* TO 'mha_rep'@'10.10.2.10' IDENTIFIED BY '123456';

六.manager上配置MHA文件,管理各个节点

[manager:154]
shell> mkdir -p /masterha/app1
shell> mkdir /etc/masterha
shell> vim /etc/masterha/app1.cnf

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

[server default] user=mha_rep     ##mysql管理用戶名 password=123456 manager_workdir= /masterha/app1    #目录可以任意指定 manager_log= /masterha/app1/manager .log remote_workdir= /masterha/app1 ssh_user=root      #ssh免密钥登录的帐号名 repl_user=rep      #mysql复制帐号，用来在主从机之间同步二进制日志等 repl_password=rep123 ping_interval=1     #ping间隔时间，用来检测master是否正常 [server1] hostname =10.10.54.155 #ssh_port=9999 master_binlog_dir= /data/ndb    #mysql数据库目录 candidate_master=1     #master机宕掉后,优先启用这台作为新master [server2] hostname =10.10.54.156 #ssh_port=9999 master_binlog_dir= /data/ndb candidate_master=1 [server3] hostname =10.10.54.157 #ssh_port=9999 master_binlog_dir= /data/ndb no_master=1      #设置no_master=1使主机不能成为新master

七.验证ssh信任登陆和mysql主从复制是否成功

1.masterha_check_ssh 验证ssh信任登陆

1 2 3 4

[manager:154] shell> masterha_check_ssh --conf= /etc/masterha/app1 .cnf Sun Mar  2 17:45:38 2014 - [debug]   ok. Sun Mar  2 17:45:38 2014 - [info] All SSH connection tests passed successfully.

2.masterha_check_repl 验证mysql复制是否成功

1 2 3 4 5 6 7 8 9 10 11

[manager:154] shell> masterha_check_repl --conf= /etc/masterha/app1 .cnf --------------------------------------------------------- Sun Mar  2 13:16:57 2014 - [info] Slaves settings check  done . Sun Mar  2 13:16:57 2014 - [info] 10.10.54.156 (current master)   +--10.10.54.155   +--10.10.54.157 ... MySQL Replication Health is OK. ---------------------------------------------------------------

八.启动MHA manager,并监控日志文件

1 2 3 4 5 6 7 8 9 10

[manager:154] shell>  nohup  masterha_manager --conf= /etc/masterha/app1 .cnf >  /tmp/mha_manager .log 2>&1 shell>  tail  -f  /masterha/app1/manager .log ---------------------------------------------------------------   10.10.54.156 (current master)   +--10.10.54.155   +--10.10.54.157 ... Sun Mar  2 13:09:25 2014 - [info] Ping(SELECT) succeeded, waiting  until  MySQL doesn't respond.. -----------------------------------------------------------------

监控的manager.log文件表明MHA运行良好，正在 "waiting until MySQL doesn't respond"

九.测试master(156)宕机后，是否会自动切换

1.测试自动切换是否成功

当掉master机子

shell> /etc/init.d/myqld stop
当掉master后，manager上的监控文件/masterha/app1/manager.log显示错误信息，表示不能自动切换：
[error]
-----------------------------------------------------------
Sun Mar  2 13:13:46 2014 - [error][/usr/share/perl5/vendor_perl/MHA/ManagerUtil.pm, ln178] Got ERROR: Use of uninitialized value $msg in scalar chomp at /usr/share/perl5/vendor_perl/MHA/ManagerConst.pm line 90.
-----------------------------------------------------------
解决这个错误是在文件/usr/share/perl5/vendor_perl/MHA/ManagerConst.pm 第90行(chomp $msg)前加入一行：

1	$msg =  ""  unless($msg);

好了，错误解决了，下面我们再次重复上面步骤:
master上mysql服务：shell> /etc/init.d/mysqld stop
再次查看manager机子上监控文件内容

shell> tail -f tail -f /masterha/app1/manager.log
日志文件显示：
-----------------------------------------------------------
----- Failover Report -----
app1: MySQL Master failover 10.10.54.156 to 10.10.54.155 succeeded
Master 10.10.54.156 is down!
Check MHA Manager logs at mycentos4:/masterha/app1/manager.log for details.
Started automated(non-interactive) failover.
The latest slave 10.10.54.155(10.10.54.155:3306) has all relay logs for recovery.
Selected 10.10.54.155 as a new master.
10.10.54.155: OK: Applying all logs succeeded.
10.10.54.157: This host has the latest relay log events.
Generating relay diff files from the latest slave succeeded.
10.10.54.157: OK: Applying all logs succeeded. Slave started, replicating from 10.10.54.155.
10.10.54.155: Resetting slave info succeeded.
Master failover to 10.10.54.155(10.10.54.155:3306) completed successfully.
--------------------------------------------------------

2.切换成功后，检查replication状态
[master:156]
shell> /etc/init.d/mysqld start
[manager:154]
shell> masterha_check_repl --conf=/etc/masterha/app1.cnf
--------------------------------------------------------------
Sun Mar  2 13:22:11 2014 - [info] Slaves settings check done.
Sun Mar  2 13:22:11 2014 - [info] 
10.10.54.155 (current master)
+--10.10.54.156
+--10.10.54.157
...
MySQL Replication Health is OK.
---------------------------------------------------------------
上面的"10.10.54.155 (current master)" 这句表明master成功切换到155机子上

十.上一步测试之后，新master机为155，宕掉155机子，再次测试故障转移

1.启动管理节点
shell> nohup masterha_manager --conf=/etc/masterha/app1.cnf > /tmp/mha_manager.log 2>&1
2.启动日志检测，然后当掉新master(155)，然后查看监控文件变化
shell> tail -f /masterha/app1/manager.log
3.当掉155机子(即新的master)
shell> /etc/init.d/mysqld stop
4.查看manager主机上的监控文件变化
[error][/usr/share/perl5/vendor_perl/MHA/MasterFailover.pm, ln295] Last failover was done at 2014/03/02 13:02:47. Current time is too early to do failover again. If you want to do failover, manually remove /masterha/app1/app1.failover.complete and run this script again.
错误解决办法
1.日志文件提示切换master过快，需要删除/masterha/app1/app1.failover.complete
1.删除app1.failover.complete
shell> rm /masterha/app1/app1.failover.complete
5.重新测试：
master转移成功，重新转为156机子
--------------------------------------------------------
Master 10.10.54.155 is down!
Check MHA Manager logs at mycentos4:/masterha/app1/manager.log for details.
Started automated(non-interactive) failover.
The latest slave 10.10.54.156(10.10.54.156:3306) has all relay logs for recovery.
Selected 10.10.54.156 as a new master.
10.10.54.156: OK: Applying all logs succeeded.
10.10.54.157: This host has the latest relay log events.
Generating relay diff files from the latest slave succeeded.
10.10.54.157: OK: Applying all logs succeeded. Slave started, replicating from 10.10.54.156.
10.10.54.156: Resetting slave info succeeded.
Master failover to 10.10.54.156(10.10.54.156:3306) completed successfully.
-----------------------------------------------------------

//附:故障转移后，用命令恢复原来的master

[manager:154]
1.在旧master上执行
mysql> reset master;
mysql> change master to master_host='10.10.54.155', master_port=3306, master_user='rep', master_password='rep123', master_log_file='mysql-bin.000031', master_log_pos=112;
mysql> start slave;      #暂时先把旧master变为从
2.然后在manager节点上：
[manager:154]
shell> masterha_master_switch --master_state=alive --conf=/etc/masterha/app1.cnf
##master成功切换回