mysql高可用研究(一) 主从+MHA架构

最近在研究mysql的高可用架构，自己想总结下常用的高可用方案都有哪些、有哪些优缺点以及应用的场景？搞得是头昏脑涨，天昏地暗，看了诸多资料，每次都觉得公说公有理婆说婆有理。其实嘛，大家说的都有一定的道理，只不过适合自己的才是最正确的。今天就从比较常用的主从+MHA说起。

学习一种新的架构还是软件，最好还是先从了解它的原理开始，这样才能在应用时扬长避短。

一、【MHA原理】

相较于其它HA软件，MHA的目的在于维持MySQL Replication中Master库的高可用性，其最大特点是可以修复多个Slave之间的差异日志，最终使所有Slave保持数据一致，然后从中选择一个充当新的Master，并将其它Slave指向它。

MHA软件由两部分组成，Manager工具包和Node工具包，具体的说明如下。

Manager工具包主要包括以下几个工具：

masterha_check_ssh              检查MHA的SSH状况
masterha_check_repl             检查MySQL复制状况
masterha_manger                 启动MHA
masterha_check_status           检测当前MHA运行状态
masterha_master_monitor         检测master是否宕机
masterha_master_switch          控制故障转移（自动或者手动）
masterha_conf_host              添加或删除配置的server信息

Node工具包（这些工具通常由MHA Manager的脚本触发，无需人为操作）主要包括以下几个工具：

save_binary_logs                保存和复制master的二进制日志
apply_diff_relay_logs           识别差异的中继日志事件并将其差异的事件应用于其他的slave
filter_mysqlbinlog              去除不必要的ROLLBACK事件（MHA已不再使用这个工具）
purge_relay_logs                清除中继日志（不会阻塞SQL线程）

基本工作流程大致如下：

（1） Manager定期监控Master，监控时间间隔由参数ping_interval决定，缺省为3秒钟一次；可利用其自身的监控功能，也可调用第三方软件来监控；MHA自身提供了两种监控方式：SELECT（执行SELECT 1）和CONNECT（创建连接/断开连接），

主要由ping_type参数决定，默认是select方式。

（2）当监测到Master故障时，调用SSH脚本对所有Node执行一次检查，包括如下几个方面：

――MySQL实例是否可以连接；

――Master服务器是否可以SSH连通；

　　 ――检查SQL Thread的状态；

　　 ――检查哪些Server死掉了，哪些Server是活动的，以及活动的Slave实例；

　　 ――检查Slave实例的配置及复制过滤规则；

　　 ――最后退出监控脚本并返回代表特殊意义代码。

（3）开始Master故障切换，包括如下几个子阶段：

　　 ――Phase 1: Configuration Check Phase

在这个阶段，若某个Slave实例的SQL Thread停止了，则会自动启动它；并再次确认活动的Servers及Slaves。

　　 ――Phase 2: Dead Master Shutdown Phase

在这个阶段，首先调用master_ip_failover_script，若HA是基于VIP实现的，则关闭VIP，若是基于目录数据库实现的，则修改映射记录。然后调用shutdown_script脚本强制关闭主机，以避免服务重启时，发生脑裂。

　　 ――Phase 3: Master Recovery Phase

　　　又包括如下3个子阶段：

　　　　 Phase 3.1: Getting Latest Slaves Phase

　　　　检查各个Slave，获取最近的和最旧的binary log file和position，并检查各个Slave成为Master的优先级，依赖于candidate_master、no_master、 [server_xxx]顺序、binary log差异量等因素。

　　　　 Phase 3.2: Saving Dead Master's Binlog Phase

　　　　若dead master所在服务器依然可以通过SSH连通，则提取dead master的binary log，提取日志的起点就是上一步获取的最新的binary log file和position，直到最后一条事件日志，并在dead master本地的工作目录（由参数remote_workdir决定）中

创建文件保存这些提取到的日志，然后将该文件拷贝到Manager服务器的工作目录下（由参数manager_workdir决定）。若dead master系统就无法连接，也就不存在差异的binary log了。MHA还要对各个Slave节点进行健康检查，主要是SSH连通性。

　　　　Phase 3.3: Determining New Master Phase

　　　　接下来调用apply_diff_relay_logs命令恢复Slave的差异日志，这个差异日志指的是各个Slave之间的relay log。恢复完成后，所有的Slave数据是一致的，此时就可以根据优先级选择New Master了。

　　　　Phase 3.4: New Master Diff Log Generation Phase

　　　　这里是生成dead master和new master之间的差异日志，即将Phase 3.2保存的binary log拷贝到New Master的工作目录中（remote_workdir）。

　　　　Phase 3.5: Master Log Apply Phase

　　　　将上一步拷贝的差异日志恢复到New Master上，若发生错误，也可手动恢复。然后获取New Master的binlog name和position，以便其它Slave从这个新的binlog name和position开始复制。最后会开启New Master的写权限，即将read_only参数设置为0。

　　 ――Phase 4: Slaves Recovery Phase

　　　　Phase 4.1: Starting Parallel Slave Diff Log Generation Phase

　　　　生成Slave与New Slave之间的差异日志，并将该日志拷贝到各Slave的工作目录下，这部分日志dead master和new master之间差异的那部分日志，因为各个Slave在Phase 3.3阶段已经同步了。

　　　　Phase 4.2: Starting Parallel Slave Log Apply Phase

　　　　在各个Slave上应用这部分差异日志，然后通过CHANGE MASTER TO命令将这些Slave指向新的New Master，最后开始复制(start slave)。

　　 ――Phase 5: New master cleanup phase

　　　　清理New Master其实就是重置slave info，即取消原来的Slave信息。至此整个Master故障切换过程完成。

二、【实验部分】

1、【环境说明】：默认三台机器上都已安装mysql5.6,且主从复制已经配置完成。
角色                 主机名      ip地址       功能
主库                    node1   192.168.245.129    (w/r)
candidate_master       node2   192.168.245.131     (r)
从库                    node3    192.168.245.132    (r)
vip： 192.168.245.100

129为主库，对外提供读写服务，而131和132机器对外提供读服务，需要设置为只读状态，不建议将它写入配置文件，因为从库随时会切换为主库。如下：

set global read_only=1

2、配置三台机器之间的信任机制（省）

目的：机器之间能够无需输入密码进行访问。

3、安装mha软件

rpm安装：每台机器都要安装node，manager节点建议安装到单独的一台机器上或者一个不用切换为主的从库上

#安装可能需要的依赖包
[root@node1 software]# yum install perl-DBD-MySQL  
[root@node1 software]# yum install perl-Config-Tiny
[root@node1 software]# yum install perl-Parallel-ForkManager*.rpm                
[root@node1 software]# yum install perl-Mail-Sender*.rpm 
[root@node1 software]# yum install perl-Mail-Sendmail*.rpm 
[root@node1 software]# yum install perl-Log-Dispatch*.rpm 
#安装mha，这里用rpm包安装，默认在/usr/bin
[root@node1 software]# yum install mha4mysql-node-0.56-0.el6.noarch.rpm
[root@node1 software]# yum install mha4mysql-manager-0.56-0.el6.noarch.rpm

二进制包安装

#node安装
wget http://mysql-master-ha.googlecode.com/files/mha4mysql-node-0.53.tar.gz
tar xf mha4mysql-node-0.53.tar.gz
cd mha4mysql-node-0.53
perl Makefile.PL
make && make install

#manager安装
wget http://mysql-master-ha.googlecode.com/files/mha4mysql-manager-0.53.tar.gz
tar xf mha4mysql-manager-0.53.tar.gz 
cd mha4mysql-manager-0.53
perl Makefile.PL
make && make install

4、手工配置主库服务器的vip并测试

这里通过脚本手动创建vip，如下：

[root@node1 scripts]# cat init_vip.sh 
vip="192.168.245.100/32"
/sbin/ip addr add $vip dev eth0

【测试】

绑定完成后，可以用以下命令查看绑定情况：

[root@node2 etc]# ip addr show
1: lo:  mtu 16436 qdisc noqueue state UNKNOWN 
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
    inet6 ::1/128 scope host 
       valid_lft forever preferred_lft forever
2: eth0:  mtu 1500 qdisc pfifo_fast state UP qlen 1000
    link/ether 00:0c:29:d7:0f:4a brd ff:ff:ff:ff:ff:ff
    inet 192.168.245.131/24 brd 192.168.245.255 scope global eth0
    inet 192.168.245.100/32 scope global eth0
    inet6 fe80::20c:29ff:fed7:f4a/64 scope link 
       valid_lft forever preferred_lft forever

到任意从库ssh 192.168.245.100 --看是否连上vip或者mysql -h 192.168.245.100 -udarren -pdarren --是否连上vip数据库. 如果都能够连接上，表示vip设置成功了。

5、配置mha及启动

（1）创建mha监控用户（在主库执行，这样每个服务器都有这个用户了）

mysql> grant all privileges on *.* to 'root'@'%' identified  by '123456';
Query OK, 0 rows affected (0.00 sec)

mysql> flush  privileges;
Query OK, 0 rows affected (0.01 sec)

（2）关闭purge_relay_logs：

在修改配置文件前，注意一个知识点，须将从库上的relay log自动清除设置为OFF，因为MySQL数据库主从复制在缺省情况下从库的relay logs会在SQL线程执行完毕后被自动删除，但是对于MHA场景下，对于某些滞后从库的恢复依赖于其他从库的relay log，因此采取禁用自动删除功能以及定期清理的办法。对于清理过多过大的relay log需要注意引起的复制延迟资源开销等。MHA可通过purge_relay_logs脚本及配合cronjob来完成此项任务。

purge_relay_logs的主要功能：

a、为relay日志创建硬链接(最小化批量删除大文件导致的性能问题)

b、SET GLOBAL relay_log_purge=1; FLUSH LOGS; SET GLOBAL relay_log_purge=0;
c、删除relay log（rm –f /path/to/archive_dir/*）

purge_relay_logs的用法及相关参数
1 purge_relay_logs --help
   Usage:
       purge_relay_logs --user=root --password=rootpass --host=127.0.0.1

2 参数描述
--user             用户名，缺省为root
--password         密码
--port             端口号
--host             主机名，缺省为127.0.0.1
--workdir          指定创建relay log的硬链接的位置，默认是/var/tmp，成功执行脚本后，硬链接的中继日志文件被删除,由于系统不同分区创建硬链接文件会失败，故需要执行硬链接具体位置，建议指定为relay log相同的分区
--disable_relay_log_purge 默认情况下，如果参数relay_log_purge=1，脚本不做任何处理，自动退出.设定该参数，脚本会将relay_log_purge设置为0，当清理relay log之后，最后将参数设置为OFF(0)

3 定制清理relay log cronjob
pureg_relay_logs脚本在不阻塞SQL线程的情况下自动清理relay log。对于不断产生的relay log直接将该脚本部署到crontab以实现按天或按小时定期清理。
$ crontab -l 
# purge relay logs at 5am 
0 5 * * * app /usr/bin/purge_relay_logs --user=root --password=PASSWORD --disable_relay_log_purge >> /var/log/masterha/purge_relay_logs.log 2>&1

在我们的两个从库上设置relaylog为OFF:

(product)root@127.0.0.1 [(none)]> set global relay_log_purge=0;
Query OK, 0 rows affected (0.00 sec)

然后通过定时任务，每隔一天定时清除relaylog：

#清除脚本
#!/bin/bash
user=root
passwd=root
port=3306
log_dir='/data/masterha/log'
work_dir='/data'
purge='/usr/bin/purge_relay_logs'

if [ ! -d $log_dir ]
then
   mkdir $log_dir -p
fi

$purge --user=$user --password=$passwd --disable_relay_log_purge --port=$port --workdir=$work_dir >> $log_dir/purge_relay_logs.log 2>&1

#定时任务
crontab -e
#每天早上5点10分执行
10 5 * * * sh /data/scripts/purge_relay_log.sh

（3）修改配置文件：

到manager节点的/etc下面新建masterha目录，并将mha需要的配置初始化文件拷贝到该目录下：

[root@node3 ~]# cd /etc
[root@node3 etc]# mkdir masterha
#创建以下mha日志目录，没有则报错
[root@node3 etc]#mkdir -p /var/log/masterha/app1

[root@node3 mastermha]# ll
total 32
-rw-r--r--. 1 root root   503 Nov  9 01:26 app1.conf
-rwxr-xr-x. 1 root root    55 Nov  9 01:26 drop_vip.sh
-rwxr-xr-x. 1 root root    55 Nov  9 01:26 init_vip.sh
-rw-r--r--. 1 root root   357 Nov  9 01:26 masterha_default.conf
-rwxr-xr-x. 1 root root  3888 Nov  9 01:26 master_ip_failover
-rwxr-xr-x. 1 root root 10298 Nov  9 01:26 master_ip_online_change

然后修改vip的值：在masterha目录下执行grep "vip" *,将会列出所有文件中vip变量，然后一一修改为192.168.245.100。

修改app1.conf文件：

[server default]
manager_log=/var/log/masterha/app1/app1.log
manager_workdir=/var/log/masterha/app1
master_ip_failover_script="/etc/masterha/master_ip_failover"
master_ip_online_change_script="/etc/masterha/master_ip_online_change"
password=root
ping_interval=1
remote_workdir=/var/log/masterha/app1
repl_password=repl4slave
repl_user=repl
report_script="/etc/masterha/send_mail"
shutdown_script=""
ssh_user=root
user=root

[server1]
candidate_master=1
check_repl_delay=0
hostname=192.168.245.129
master_binlog_dir=/data/mysql/mysql_3306/logs

[server3]
hostname=192.168.245.132
port=3306

（4）检查mha环境并启动

#检查MHA Manger到所有MHA Node的SSH连接状态：
[root@node3 masterha]# /usr/bin/masterha_check_ssh --conf=/etc/masterha/app1.conf
Mon Nov 16 01:24:21 2015 - [warning] Global configuration file /etc/masterha_default.cnf not found. Skipping.
Mon Nov 16 01:24:21 2015 - [info] Reading application default configuration from /etc/masterha/app1.conf..
Mon Nov 16 01:24:21 2015 - [info] Reading server configuration from /etc/masterha/app1.conf..
Mon Nov 16 01:24:21 2015 - [info] Starting SSH connection tests..
Mon Nov 16 01:24:24 2015 - [debug] 
Mon Nov 16 01:24:21 2015 - [debug]  Connecting via SSH from root@192.168.245.129(192.168.245.129:22) to root@192.168.245.131(192.168.245.131:22)..
Mon Nov 16 01:24:23 2015 - [debug]   ok.
Mon Nov 16 01:24:23 2015 - [debug]  Connecting via SSH from root@192.168.245.129(192.168.245.129:22) to root@192.168.245.132(192.168.245.132:22)..
Mon Nov 16 01:24:24 2015 - [debug]   ok.
Mon Nov 16 01:24:25 2015 - [debug] 
Mon Nov 16 01:24:22 2015 - [debug]  Connecting via SSH from root@192.168.245.131(192.168.245.131:22) to root@192.168.245.129(192.168.245.129:22)..
Mon Nov 16 01:24:23 2015 - [debug]   ok.
Mon Nov 16 01:24:23 2015 - [debug]  Connecting via SSH from root@192.168.245.131(192.168.245.131:22) to root@192.168.245.132(192.168.245.132:22)..
Mon Nov 16 01:24:25 2015 - [debug]   ok.
Mon Nov 16 01:24:25 2015 - [debug] 
Mon Nov 16 01:24:22 2015 - [debug]  Connecting via SSH from root@192.168.245.132(192.168.245.132:22) to root@192.168.245.129(192.168.245.129:22)..
Mon Nov 16 01:24:24 2015 - [debug]   ok.
Mon Nov 16 01:24:24 2015 - [debug]  Connecting via SSH from root@192.168.245.132(192.168.245.132:22) to root@192.168.245.131(192.168.245.131:22)..
Mon Nov 16 01:24:25 2015 - [debug]   ok.
Mon Nov 16 01:24:25 2015 - [info] All SSH connection tests passed successfully.

#检查主从复制环境
[root@node3 masterha]# /usr/bin/masterha_check_repl --conf=/etc/masterha/app1.conf
Mon Nov 16 01:37:08 2015 - [warning] Global configuration file /etc/masterha_default.cnf not found. Skipping.
Mon Nov 16 01:37:08 2015 - [info] Reading application default configuration from /etc/masterha/app1.conf..
Mon Nov 16 01:37:08 2015 - [info] Reading server configuration from /etc/masterha/app1.conf..
Mon Nov 16 01:37:08 2015 - [info] MHA::MasterMonitor version 0.56.
Mon Nov 16 01:37:09 2015 - [info] GTID failover mode = 0
Mon Nov 16 01:37:09 2015 - [info] Dead Servers:
Mon Nov 16 01:37:09 2015 - [info] Alive Servers:
Mon Nov 16 01:37:09 2015 - [info]   192.168.245.129(192.168.245.129:3306)
Mon Nov 16 01:37:09 2015 - [info]   192.168.245.131(192.168.245.131:3306)
Mon Nov 16 01:37:09 2015 - [info]   192.168.245.132(192.168.245.132:3306)
Mon Nov 16 01:37:09 2015 - [info] Alive Slaves:
Mon Nov 16 01:37:09 2015 - [info]   192.168.245.131(192.168.245.131:3306)  Version=5.6.23-log (oldest major version between slaves) log-bin:enabled
Mon Nov 16 01:37:09 2015 - [info]     Replicating from 192.168.245.129(192.168.245.129:3306)
Mon Nov 16 01:37:09 2015 - [info]     Primary candidate for the new Master (candidate_master is set)
Mon Nov 16 01:37:09 2015 - [info]   192.168.245.132(192.168.245.132:3306)  Version=5.6.21-log (oldest major version between slaves) log-bin:enabled
Mon Nov 16 01:37:09 2015 - [info]     Replicating from 192.168.245.129(192.168.245.129:3306)
Mon Nov 16 01:37:09 2015 - [info] Current Alive Master: 192.168.245.129(192.168.245.129:3306)
Mon Nov 16 01:37:09 2015 - [info] Checking slave configurations..
Mon Nov 16 01:37:09 2015 - [info]  read_only=1 is not set on slave 192.168.245.131(192.168.245.131:3306).
Mon Nov 16 01:37:09 2015 - [info]  read_only=1 is not set on slave 192.168.245.132(192.168.245.132:3306).
Mon Nov 16 01:37:09 2015 - [info] Checking replication filtering settings..
Mon Nov 16 01:37:09 2015 - [info]  binlog_do_db= , binlog_ignore_db= 
Mon Nov 16 01:37:09 2015 - [info]  Replication filtering check ok.
Mon Nov 16 01:37:09 2015 - [info] GTID (with auto-pos) is not supported
Mon Nov 16 01:37:09 2015 - [info] Starting SSH connection tests..
Mon Nov 16 01:37:12 2015 - [info] All SSH connection tests passed successfully.
Mon Nov 16 01:37:12 2015 - [info] Checking MHA Node version..
Mon Nov 16 01:37:13 2015 - [info]  Version check ok.
Mon Nov 16 01:37:13 2015 - [info] Checking SSH publickey authentication settings on the current master..
Mon Nov 16 01:37:13 2015 - [info] HealthCheck: SSH to 192.168.245.129 is reachable.
Mon Nov 16 01:37:14 2015 - [info] Master MHA Node version is 0.56.
Mon Nov 16 01:37:14 2015 - [info] Checking recovery script configurations on 192.168.245.129(192.168.245.129:3306)..
Mon Nov 16 01:37:14 2015 - [info]   Executing command: save_binary_logs --command=test --start_pos=4 --binlog_dir=/data/mysql/mysql_3306/logs --output_file=/var/log/masterha/app1/save_binary_logs_test --manager_version=0.56 --start_file=mysql-bin.000023 
Mon Nov 16 01:37:14 2015 - [info]   Connecting to root@192.168.245.129(192.168.245.129:22).. 
  Creating /var/log/masterha/app1 if not exists..    ok.
  Checking output directory is accessible or not..
   ok.
  Binlog found at /data/mysql/mysql_3306/logs, up to mysql-bin.000023
Mon Nov 16 01:37:15 2015 - [info] Binlog setting check done.
Mon Nov 16 01:37:15 2015 - [info] Checking SSH publickey authentication and checking recovery script configurations on all alive slave servers..
Mon Nov 16 01:37:15 2015 - [info]   Executing command : apply_diff_relay_logs --command=test --slave_user='root' --slave_host=192.168.245.131 --slave_ip=192.168.245.131 --slave_port=3306 --workdir=/var/log/masterha/app1 --target_version=5.6.23-log --manager_version=0.56 --relay_log_info=/data/mysql/mysql_3306/data/relay-log.info  --relay_dir=/data/mysql/mysql_3306/data/  --slave_pass=xxx
Mon Nov 16 01:37:15 2015 - [info]   Connecting to root@192.168.245.131(192.168.245.131:22).. 
  Checking slave recovery environment settings..
    Opening /data/mysql/mysql_3306/data/relay-log.info ... ok.
    Relay log found at /data/mysql/mysql_3306/data, up to relay-bin.000009
    Temporary relay log file is /data/mysql/mysql_3306/data/relay-bin.000009
    Testing mysql connection and privileges..Warning: Using a password on the command line interface can be insecure.
 done.
    Testing mysqlbinlog output.. done.
    Cleaning up test file(s).. done.
Mon Nov 16 01:37:15 2015 - [info]   Executing command : apply_diff_relay_logs --command=test --slave_user='root' --slave_host=192.168.245.132 --slave_ip=192.168.245.132 --slave_port=3306 --workdir=/var/log/masterha/app1 --target_version=5.6.21-log --manager_version=0.56 --relay_log_info=/data/mysql/data/relay-log.info  --relay_dir=/data/mysql/data/  --slave_pass=xxx
Mon Nov 16 01:37:15 2015 - [info]   Connecting to root@192.168.245.132(192.168.245.132:22).. 
  Checking slave recovery environment settings..
    Opening /data/mysql/data/relay-log.info ... ok.
    Relay log found at /data/mysql/data, up to node3-relay-bin.000007
    Temporary relay log file is /data/mysql/data/node3-relay-bin.000007
    Testing mysql connection and privileges..Warning: Using a password on the command line interface can be insecure.
 done.
    Testing mysqlbinlog output.. done.
    Cleaning up test file(s).. done.
Mon Nov 16 01:37:16 2015 - [info] Slaves settings check done.
Mon Nov 16 01:37:16 2015 - [info] 
192.168.245.129(192.168.245.129:3306) (current master)
 +--192.168.245.131(192.168.245.131:3306)
 +--192.168.245.132(192.168.245.132:3306)

Mon Nov 16 01:37:16 2015 - [info] Checking replication health on 192.168.245.131..
Mon Nov 16 01:37:16 2015 - [info]  ok.
Mon Nov 16 01:37:16 2015 - [info] Checking replication health on 192.168.245.132..
Mon Nov 16 01:37:16 2015 - [info]  ok.
Mon Nov 16 01:37:16 2015 - [warning] master_ip_failover_script is not defined.
Mon Nov 16 01:37:16 2015 - [warning] shutdown_script is not defined.
Mon Nov 16 01:37:16 2015 - [info] Got exit code 0 (Not master dead).

MySQL Replication Health is OK.

如果遇到这个报错：

Can't exec "mysqlbinlog": No such file or directory at /usr/share/perl5/vendor_perl/MHA/BinlogManager.pm line 106.
mysqlbinlog version command failed with rc 1:0, please verify PATH, LD_LIBRARY_PATH, and client options
 at /usr/bin/apply_diff_relay_logs line 493
Mon Nov 16 01:32:36 2015 - [error][/usr/share/perl5/vendor_perl/MHA/MasterMonitor.pm, ln205] Slaves settings check failed!
Mon Nov 16 01:32:36 2015 - [error][/usr/share/perl5/vendor_perl/MHA/MasterMonitor.pm, ln413] Slave configuration failed.
Mon Nov 16 01:32:36 2015 - [error][/usr/share/perl5/vendor_perl/MHA/MasterMonitor.pm, ln424] Error happened on checking configurations.  at /usr/bin/masterha_check_repl line 48
Mon Nov 16 01:32:36 2015 - [error][/usr/share/perl5/vendor_perl/MHA/MasterMonitor.pm, ln523] Error happened on monitoring servers.
Mon Nov 16 01:32:36 2015 - [info] Got exit code 1 (Not master dead).

MySQL Replication Health is NOT OK!

解决方法如下，添加软连接（所有节点）

ln -s /usr/local/mysql/bin/mysqlbinlog /usr/local/bin/mysqlbinlog

ln -s /usr/local/mysql/bin/mysql /usr/local/bin/mysql

到此为止，都没问题了，开始启动manager节点进行监控，一般启动我们都使用nohup方式，但是当出现故障成功切换后，manager监控也会关闭，所以建议使用daemontools方式在后台运行。

nohup方式启动

nohup /usr/bin/masterha_manager --conf=/etc/masterha/app1.conf --remove_dead_master_conf --ignore_last_failover &
#关闭mha监控
/usr/bin/masterha_stop --conf=/etc/masterha/app1.conf

daemontools方式启动：

#下载daemontools工具
wget http://cr.yp.to/daemontools/daemontools-0.76.tar.gz
#安装
cd /home/software/daemontools/admin/daemontools-0.76
vim src/conf-cc #在第一行最后空一格后添加 -include /usr/include/errno.h 避免编译时报错
package/install  #安装
#在/etc/init/创建svscan.conf文件并添加如下内容：
vim /etc/init/svscan.conf
start on runlevel [345]
        respawn
        exec /command/svscanboot
#在/etc/init目录下重新加载并启动
[root@node3 init]# initctl reload-configuration
[root@node3 init]# initctl start svscan
svscan start/running, process 17002

使用daemon启动mha监控：

[root@node3 init]# mkdir -p /service/masterha_app
[root@node3 ~]# cd /service/masterha_app/
[root@node3 masterha_app]# pwd
/service/masterha_app
#在/service/masterha_app/创建run脚本并增加执行权限，内容如下
[root@node3 masterha_app]# cat /service/masterha_app/run
#!/bin/bash
 exec /usr/bin/masterha_manager --conf=/etc/masterha/app1.conf --remove_dead_master_conf --ignore_last_failover
[root@node3 masterha_app]# chmod +x /service/masterha_app/run
#停止监控 
svc -d /service/masterha_app 
#开启监控 
svc -u /service/masterha_app

另起一个session，查看启动日志：

[root@node3 app1]# tail -f app1.log 
192.168.245.129(192.168.245.129:3306) (current master)
 +--192.168.245.131(192.168.245.131:3306)
 +--192.168.245.132(192.168.245.132:3306)

Mon Nov 16 01:55:10 2015 - [warning] master_ip_failover_script is not defined.
Mon Nov 16 01:55:10 2015 - [warning] shutdown_script is not defined.
Mon Nov 16 01:55:10 2015 - [info] Set master ping interval 1 seconds.
Mon Nov 16 01:55:10 2015 - [warning] secondary_check_script is not defined. It is highly recommended setting it to check master reachability from two or more routes.
Mon Nov 16 01:55:10 2015 - [info] Starting ping health check on 192.168.245.129(192.168.245.129:3306)..
Mon Nov 16 01:55:10 2015 - [info] Ping(SELECT) succeeded, waiting until MySQL doesn't respond..

当看到最后一句“Ping(SELECT) succeeded, waiting until MySQL doesn't respond..”表示mha已经启动起来了。当然，你或许在一台机子上希望监控多套master-salve复制，这非常容易，只要为第二套集群创建一个新的配置文件并启动manager

# masterha_manager --conf=/etc/conf/masterha/app1.cnf  
# masterha_manager --conf=/etc/conf/masterha/app2.cnf
如果你在app1和app2上有一些共有的参数，可在全局配置文件中配置。

启动参数介绍：

--remove_dead_master_conf 该参数表示当发生主从切换后，老的主库的ip将会从配置文件中移除。如果故障机器修复好了，需要手工添加ip信息到配置文件中

--manger_log 日志存放位置

--ignore_last_failover 在缺省情况下，如果MHA检测到连续发生宕机，且两次宕机间隔不足8小时的话，则不会进行Failover，之所以这样限制是为了避免ping-pong效应。该参数代表忽略上次MHA触发切换产生的文件，默认情况下，MHA发生切换后会在日志目录，也就是上面我设置的/data产生app1.failover.complete文件，下次再次切换的时候如果发现该目录下存在该文件将不允许触发切换，除非在第一次切换后收到删除该文件，为了方便，这里设置为--ignore_last_failover。

查看MHA Manager监控是否正常：

[root@node3 app1]#  /usr/bin/masterha_check_status --conf=/etc/masterha/app1.conf
app1 (pid:6795) is running(0:PING_OK), master:192.168.245.129

6、模拟测试

（1）自动Failover测试

首先用sysbench工具模拟主库业务的写入。

#下载sysbench
http://dev.mysql.com/downloads/benchmarks.html
#安装sysbench
yum install libtool -y  #依赖包
tar zxvf sysbench-0.4.12.7.tar.gz
cd sysbench-0.4.12.7
./configure && make && make install

在主库执行，创建sysbench测试表：

(product)root@127.0.0.1 [(none)]> create database sbtest;
Query OK, 1 row affected (0.00 sec)

[root@node1 sysbench-0.4.12.7]# /usr/local/bin/sysbench --test=oltp --oltp-table-size=100000 --oltp-read-only=off --init-rng=on --num-threads=4 --max-requests=0 --oltp-dist-type=uniform --max-time=1800 --mysql-user=root --mysql-socket=/tmp/mysql_3306.sock --mysql-password=root --mysql-host=192.168.245.129 --db-driver=mysql --mysql-table-engine=innodb --oltp-test-mode=complex prepare
sysbench 0.4.12.6:  multi-threaded system evaluation benchmark

Creating table 'sbtest'...
Creating 100000 records in table 'sbtest'...

停止接管主库那个从库的io_thread线程：

(product)root@127.0.0.1 [(none)]> stop slave io_thread;
Query OK, 0 rows affected (0.01 sec)

向主库插入数据：

[root@node1 sysbench-0.4.12.7]# /usr/local/bin/sysbench --test=oltp --oltp-table-size=100000 --oltp-read-only=off --init-rng=on --num-threads=4 --max-requests=0 --oltp-dist-type=uniform --max-time=180 --mysql-user=root --mysql-socket=/tmp/mysql_3306.sock --mysql-password=root --mysql-host=192.168.245.129 --db-driver=mysql --mysql-table-engine=innodb --oltp-test-mode=complex run
sysbench 0.4.12.6:  multi-threaded system evaluation benchmark

Running the test with following options:
Number of threads: 4
Initializing random number generator from timer.

Random number generator seed is 0 and will be ignored


Doing OLTP test.
Running mixed OLTP test
Using Uniform distribution
Using "BEGIN" for starting transactions
Using auto_inc on the id column
Using 1 test tables
Threads started!
Time limit exceeded, exiting...
(last message repeated 3 times)
Done.

OLTP test statistics:
    queries performed:
        read:                            489090
        write:                           174675
        other:                           69870
        total:                           733635
    transactions:                        34935  (194.05 per sec.)
    deadlocks:                           0      (0.00 per sec.)
    read/write requests:                 663765 (3686.93 per sec.)
    other operations:                    69870  (388.10 per sec.)

Test execution summary:
    total time:                          180.0317s
    total number of events:              34935
    total time taken by event execution: 719.7722
    per-request statistics:
         min:                                  3.47ms
         avg:                                 20.60ms
         max:                                444.43ms
         approx.  95 percentile:              28.17ms

Threads fairness:
    events (avg/stddev):           8733.7500/260.92
    execution time (avg/stddev):   179.9430/0.01

然后启动从库上io_thread:

(product)root@127.0.0.1 [(none)]> start slave io_thread;
Query OK, 0 rows affected, 1 warning (0.00 sec)

kill主库的mysqld：

pkill -9 mysqld

最后，查看mha日志，这时已经接管过来了！

Mon Nov 16 02:07:11 2015 - [warning] master_ip_failover_script is not defined.
Mon Nov 16 02:07:11 2015 - [warning] shutdown_script is not defined.
Mon Nov 16 02:07:11 2015 - [info] Set master ping interval 1 seconds.
Mon Nov 16 02:07:11 2015 - [warning] secondary_check_script is not defined. It is highly recommended setting it to check master reachability from two or more routes.
Mon Nov 16 02:07:11 2015 - [info] Starting ping health check on 192.168.245.129(192.168.245.129:3306)..
Mon Nov 16 02:07:11 2015 - [info] Ping(SELECT) succeeded, waiting until MySQL doesn't respond..
Mon Nov 16 03:02:12 2015 - [warning] Got error on MySQL select ping: 2006 (MySQL server has gone away)
Mon Nov 16 03:02:12 2015 - [info] Executing SSH check script: save_binary_logs --command=test --start_pos=4 --binlog_dir=/data/mysql/mysql_3306/logs --output_file=/var/log/masterha/app1/save_binary_logs_test --manager_version=0.56 --binlog_prefix=mysql-bin
Mon Nov 16 03:02:13 2015 - [info] HealthCheck: SSH to 192.168.245.129 is reachable.
Mon Nov 16 03:02:13 2015 - [warning] Got error on MySQL connect: 2013 (Lost connection to MySQL server at 'reading initial communication packet', system error: 111)
Mon Nov 16 03:02:13 2015 - [warning] Connection failed 2 time(s)..
Mon Nov 16 03:02:14 2015 - [warning] Got error on MySQL connect: 2013 (Lost connection to MySQL server at 'reading initial communication packet', system error: 111)
Mon Nov 16 03:02:14 2015 - [warning] Connection failed 3 time(s)..
Mon Nov 16 03:02:15 2015 - [warning] Got error on MySQL connect: 2013 (Lost connection to MySQL server at 'reading initial communication packet', system error: 111)
Mon Nov 16 03:02:15 2015 - [warning] Connection failed 4 time(s)..
Mon Nov 16 03:02:15 2015 - [warning] Master is not reachable from health checker!
Mon Nov 16 03:02:15 2015 - [warning] Master 192.168.245.129(192.168.245.129:3306) is not reachable!
Mon Nov 16 03:02:15 2015 - [warning] SSH is reachable.
Mon Nov 16 03:02:15 2015 - [info] Connecting to a master server failed. Reading configuration file /etc/masterha_default.cnf and /etc/masterha/app1.conf again, and trying to connect to all servers to check server status..
Mon Nov 16 03:02:15 2015 - [warning] Global configuration file /etc/masterha_default.cnf not found. Skipping.
Mon Nov 16 03:02:15 2015 - [info] Reading application default configuration from /etc/masterha/app1.conf..
Mon Nov 16 03:02:15 2015 - [info] Reading server configuration from /etc/masterha/app1.conf..
Mon Nov 16 03:02:15 2015 - [info] GTID failover mode = 0
Mon Nov 16 03:02:15 2015 - [info] Dead Servers:
Mon Nov 16 03:02:15 2015 - [info]   192.168.245.129(192.168.245.129:3306)
Mon Nov 16 03:02:15 2015 - [info] Alive Servers:
Mon Nov 16 03:02:15 2015 - [info]   192.168.245.131(192.168.245.131:3306)
Mon Nov 16 03:02:15 2015 - [info]   192.168.245.132(192.168.245.132:3306)
Mon Nov 16 03:02:15 2015 - [info] Alive Slaves:
Mon Nov 16 03:02:15 2015 - [info]   192.168.245.131(192.168.245.131:3306)  Version=5.6.23-log (oldest major version between slaves) log-bin:enabled
Mon Nov 16 03:02:15 2015 - [info]     Replicating from 192.168.245.129(192.168.245.129:3306)
Mon Nov 16 03:02:15 2015 - [info]     Primary candidate for the new Master (candidate_master is set)
Mon Nov 16 03:02:15 2015 - [info]   192.168.245.132(192.168.245.132:3306)  Version=5.6.21-log (oldest major version between slaves) log-bin:enabled
Mon Nov 16 03:02:15 2015 - [info]     Replicating from 192.168.245.129(192.168.245.129:3306)
Mon Nov 16 03:02:15 2015 - [info] Checking slave configurations..
Mon Nov 16 03:02:15 2015 - [info]  read_only=1 is not set on slave 192.168.245.131(192.168.245.131:3306).
Mon Nov 16 03:02:15 2015 - [info]  read_only=1 is not set on slave 192.168.245.132(192.168.245.132:3306).
Mon Nov 16 03:02:15 2015 - [info] Checking replication filtering settings..
Mon Nov 16 03:02:15 2015 - [info]  Replication filtering check ok.
Mon Nov 16 03:02:15 2015 - [info] Master is down!
Mon Nov 16 03:02:15 2015 - [info] Terminating monitoring script.
Mon Nov 16 03:02:15 2015 - [info] Got exit code 20 (Master dead).
Mon Nov 16 03:02:15 2015 - [info] MHA::MasterFailover version 0.56.
Mon Nov 16 03:02:15 2015 - [info] Starting master failover.
Mon Nov 16 03:02:15 2015 - [info] 
Mon Nov 16 03:02:15 2015 - [info] * Phase 1: Configuration Check Phase..
Mon Nov 16 03:02:15 2015 - [info] 
Mon Nov 16 03:02:16 2015 - [info] GTID failover mode = 0
Mon Nov 16 03:02:16 2015 - [info] Dead Servers:
Mon Nov 16 03:02:16 2015 - [info]   192.168.245.129(192.168.245.129:3306)
Mon Nov 16 03:02:16 2015 - [info] Checking master reachability via MySQL(double check)...
Mon Nov 16 03:02:16 2015 - [info]  ok.
Mon Nov 16 03:02:16 2015 - [info] Alive Servers:
Mon Nov 16 03:02:16 2015 - [info]   192.168.245.131(192.168.245.131:3306)
Mon Nov 16 03:02:16 2015 - [info]   192.168.245.132(192.168.245.132:3306)
Mon Nov 16 03:02:16 2015 - [info] Alive Slaves:
Mon Nov 16 03:02:16 2015 - [info]   192.168.245.131(192.168.245.131:3306)  Version=5.6.23-log (oldest major version between slaves) log-bin:enabled
Mon Nov 16 03:02:16 2015 - [info]     Replicating from 192.168.245.129(192.168.245.129:3306)
Mon Nov 16 03:02:16 2015 - [info]     Primary candidate for the new Master (candidate_master is set)
Mon Nov 16 03:02:16 2015 - [info]   192.168.245.132(192.168.245.132:3306)  Version=5.6.21-log (oldest major version between slaves) log-bin:enabled
Mon Nov 16 03:02:16 2015 - [info]     Replicating from 192.168.245.129(192.168.245.129:3306)
Mon Nov 16 03:02:16 2015 - [info] Starting Non-GTID based failover.
Mon Nov 16 03:02:16 2015 - [info] 
Mon Nov 16 03:02:16 2015 - [info] ** Phase 1: Configuration Check Phase completed.
Mon Nov 16 03:02:16 2015 - [info] 
Mon Nov 16 03:02:16 2015 - [info] * Phase 2: Dead Master Shutdown Phase..
Mon Nov 16 03:02:16 2015 - [info] 
Mon Nov 16 03:02:16 2015 - [info] Forcing shutdown so that applications never connect to the current master..
Mon Nov 16 03:02:16 2015 - [warning] master_ip_failover_script is not set. Skipping invalidating dead master IP address.
Mon Nov 16 03:02:16 2015 - [warning] shutdown_script is not set. Skipping explicit shutting down of the dead master.
Mon Nov 16 03:02:16 2015 - [info] * Phase 2: Dead Master Shutdown Phase completed.
Mon Nov 16 03:02:16 2015 - [info] 
Mon Nov 16 03:02:16 2015 - [info] * Phase 3: Master Recovery Phase..
Mon Nov 16 03:02:16 2015 - [info] 
Mon Nov 16 03:02:16 2015 - [info] * Phase 3.1: Getting Latest Slaves Phase..
Mon Nov 16 03:02:16 2015 - [info] 
Mon Nov 16 03:02:16 2015 - [info] The latest binary log file/position on all slaves is mysql-bin.000023:35416929
Mon Nov 16 03:02:16 2015 - [info] Latest slaves (Slaves that received relay log files to the latest):
Mon Nov 16 03:02:16 2015 - [info]   192.168.245.131(192.168.245.131:3306)  Version=5.6.23-log (oldest major version between slaves) log-bin:enabled
Mon Nov 16 03:02:16 2015 - [info]     Replicating from 192.168.245.129(192.168.245.129:3306)
Mon Nov 16 03:02:16 2015 - [info]     Primary candidate for the new Master (candidate_master is set)
Mon Nov 16 03:02:16 2015 - [info]   192.168.245.132(192.168.245.132:3306)  Version=5.6.21-log (oldest major version between slaves) log-bin:enabled
Mon Nov 16 03:02:16 2015 - [info]     Replicating from 192.168.245.129(192.168.245.129:3306)
Mon Nov 16 03:02:16 2015 - [info] The oldest binary log file/position on all slaves is mysql-bin.000023:35416929
Mon Nov 16 03:02:16 2015 - [info] Oldest slaves:
Mon Nov 16 03:02:16 2015 - [info]   192.168.245.131(192.168.245.131:3306)  Version=5.6.23-log (oldest major version between slaves) log-bin:enabled
Mon Nov 16 03:02:16 2015 - [info]     Replicating from 192.168.245.129(192.168.245.129:3306)
Mon Nov 16 03:02:16 2015 - [info]     Primary candidate for the new Master (candidate_master is set)
Mon Nov 16 03:02:16 2015 - [info]   192.168.245.132(192.168.245.132:3306)  Version=5.6.21-log (oldest major version between slaves) log-bin:enabled
Mon Nov 16 03:02:16 2015 - [info]     Replicating from 192.168.245.129(192.168.245.129:3306)
Mon Nov 16 03:02:16 2015 - [info] 
Mon Nov 16 03:02:16 2015 - [info] * Phase 3.2: Saving Dead Master's Binlog Phase..
Mon Nov 16 03:02:16 2015 - [info] 
Mon Nov 16 03:02:17 2015 - [info] Fetching dead master's binary logs..
Mon Nov 16 03:02:17 2015 - [info] Executing command on the dead master 192.168.245.129(192.168.245.129:3306): save_binary_logs --command=save --start_file=mysql-bin.000023  --start_pos=35416929 --binlog_dir=/data/mysql/mysql_3306/logs --output_file=/var/log/masterha/app
1/saved_master_binlog_from_192.168.245.129_3306_20151116030215.binlog --handle_raw_binlog=1 --disable_log_bin=0 --manager_version=0.56
  Creating /var/log/masterha/app1 if not exists..    ok.
 Concat binary/relay logs from mysql-bin.000023 pos 35416929 to mysql-bin.000023 EOF into /var/log/masterha/app1/saved_master_binlog_from_192.168.245.129_3306_20151116030215.binlog ..
 Binlog Checksum enabled
  Dumping binlog format description event, from position 0 to 120.. ok.
  No need to dump effective binlog data from /data/mysql/mysql_3306/logs/mysql-bin.000023 (pos starts 35416929, filesize 35416929). Skipping.
 Binlog Checksum enabled
 /var/log/masterha/app1/saved_master_binlog_from_192.168.245.129_3306_20151116030215.binlog has no effective data events.
Event not exists.
Mon Nov 16 03:02:18 2015 - [info] Additional events were not found from the orig master. No need to save.
Mon Nov 16 03:02:18 2015 - [info] 
Mon Nov 16 03:02:18 2015 - [info] * Phase 3.3: Determining New Master Phase..
Mon Nov 16 03:02:18 2015 - [info] 
Mon Nov 16 03:02:18 2015 - [info] Finding the latest slave that has all relay logs for recovering other slaves..
Mon Nov 16 03:02:18 2015 - [info] All slaves received relay logs to the same position. No need to resync each other.
Mon Nov 16 03:02:18 2015 - [info] Searching new master from slaves..
Mon Nov 16 03:02:18 2015 - [info]  Candidate masters from the configuration file:
Mon Nov 16 03:02:18 2015 - [info]   192.168.245.131(192.168.245.131:3306)  Version=5.6.23-log (oldest major version between slaves) log-bin:enabled
Mon Nov 16 03:02:18 2015 - [info]     Replicating from 192.168.245.129(192.168.245.129:3306)
Mon Nov 16 03:02:18 2015 - [info]     Primary candidate for the new Master (candidate_master is set)
Mon Nov 16 03:02:18 2015 - [info]  Non-candidate masters:
Mon Nov 16 03:02:18 2015 - [info]  Searching from candidate_master slaves which have received the latest relay log events..
Mon Nov 16 03:02:18 2015 - [info] New master is 192.168.245.131(192.168.245.131:3306)
Mon Nov 16 03:02:18 2015 - [info] Starting master failover..
Mon Nov 16 03:02:18 2015 - [info] 
From:
192.168.245.129(192.168.245.129:3306) (current master)
 +--192.168.245.131(192.168.245.131:3306)
 +--192.168.245.132(192.168.245.132:3306)

To:
192.168.245.131(192.168.245.131:3306) (new master)
 +--192.168.245.132(192.168.245.132:3306)
Mon Nov 16 03:02:18 2015 - [info] 
Mon Nov 16 03:02:18 2015 - [info] * Phase 3.3: New Master Diff Log Generation Phase..
Mon Nov 16 03:02:18 2015 - [info] 
Mon Nov 16 03:02:18 2015 - [info]  This server has all relay logs. No need to generate diff files from the latest slave.
Mon Nov 16 03:02:18 2015 - [info] 
Mon Nov 16 03:02:18 2015 - [info] * Phase 3.4: Master Log Apply Phase..
Mon Nov 16 03:02:18 2015 - [info] 
Mon Nov 16 03:02:18 2015 - [info] *NOTICE: If any error happens from this phase, manual recovery is needed.
Mon Nov 16 03:02:18 2015 - [info] Starting recovery on 192.168.245.131(192.168.245.131:3306)..
Mon Nov 16 03:02:18 2015 - [info]  This server has all relay logs. Waiting all logs to be applied.. 
Mon Nov 16 03:02:20 2015 - [info]   done.
Mon Nov 16 03:02:20 2015 - [info]  All relay logs were successfully applied.
Mon Nov 16 03:02:20 2015 - [info] Getting new master's binlog name and position..
Mon Nov 16 03:02:20 2015 - [info]  mysql-bin.000013:35416929
Mon Nov 16 03:02:20 2015 - [info]  All other slaves should start replication from here. Statement should be: CHANGE MASTER TO MASTER_HOST='192.168.245.131', MASTER_PORT=3306, MASTER_LOG_FILE='mysql-bin.000013', MASTER_LOG_POS=35416929, MASTER_USER='repl', MASTER_PASSWOR
D='xxx';
Mon Nov 16 03:02:20 2015 - [warning] master_ip_failover_script is not set. Skipping taking over new master IP address.
Mon Nov 16 03:02:20 2015 - [info] ** Finished master recovery successfully.
Mon Nov 16 03:02:20 2015 - [info] * Phase 3: Master Recovery Phase completed.
Mon Nov 16 03:02:20 2015 - [info] 
Mon Nov 16 03:02:20 2015 - [info] * Phase 4: Slaves Recovery Phase..
Mon Nov 16 03:02:20 2015 - [info] 
Mon Nov 16 03:02:20 2015 - [info] * Phase 4.1: Starting Parallel Slave Diff Log Generation Phase..
Mon Nov 16 03:02:20 2015 - [info] 
Mon Nov 16 03:02:20 2015 - [info] -- Slave diff file generation on host 192.168.245.132(192.168.245.132:3306) started, pid: 10508. Check tmp log /var/log/masterha/app1/192.168.245.132_3306_20151116030215.log if it takes time..
Mon Nov 16 03:02:20 2015 - [info] 
Mon Nov 16 03:02:20 2015 - [info] Log messages from 192.168.245.132 ...
Mon Nov 16 03:02:20 2015 - [info] 
Mon Nov 16 03:02:20 2015 - [info]  This server has all relay logs. No need to generate diff files from the latest slave.
Mon Nov 16 03:02:20 2015 - [info] End of log messages from 192.168.245.132.
Mon Nov 16 03:02:20 2015 - [info] -- 192.168.245.132(192.168.245.132:3306) has the latest relay log events.
Mon Nov 16 03:02:20 2015 - [info] Generating relay diff files from the latest slave succeeded.
Mon Nov 16 03:02:20 2015 - [info] 
Mon Nov 16 03:02:20 2015 - [info] * Phase 4.2: Starting Parallel Slave Log Apply Phase..
Mon Nov 16 03:02:20 2015 - [info] 
Mon Nov 16 03:02:20 2015 - [info] -- Slave recovery on host 192.168.245.132(192.168.245.132:3306) started, pid: 10510. Check tmp log /var/log/masterha/app1/192.168.245.132_3306_20151116030215.log if it takes time..
Mon Nov 16 03:02:20 2015 - [info] 
Mon Nov 16 03:02:20 2015 - [info] Log messages from 192.168.245.132 ...
Mon Nov 16 03:02:20 2015 - [info] 
Mon Nov 16 03:02:20 2015 - [info] Starting recovery on 192.168.245.132(192.168.245.132:3306)..
Mon Nov 16 03:02:20 2015 - [info]  This server has all relay logs. Waiting all logs to be applied.. 
Mon Nov 16 03:02:20 2015 - [info]   done.
Mon Nov 16 03:02:20 2015 - [info]  All relay logs were successfully applied.
Mon Nov 16 03:02:20 2015 - [info]  Resetting slave 192.168.245.132(192.168.245.132:3306) and starting replication from the new master 192.168.245.131(192.168.245.131:3306)..
Mon Nov 16 03:02:20 2015 - [info]  Executed CHANGE MASTER.
Mon Nov 16 03:02:20 2015 - [info]  Slave started.
Mon Nov 16 03:02:20 2015 - [info] End of log messages from 192.168.245.132.
Mon Nov 16 03:02:20 2015 - [info] -- Slave recovery on host 192.168.245.132(192.168.245.132:3306) succeeded.
Mon Nov 16 03:02:20 2015 - [info] All new slave servers recovered successfully.
Mon Nov 16 03:02:20 2015 - [info] 
Mon Nov 16 03:02:20 2015 - [info] * Phase 5: New master cleanup phase..
Mon Nov 16 03:02:20 2015 - [info] 
Mon Nov 16 03:02:20 2015 - [info] Resetting slave info on the new master..
Mon Nov 16 03:02:20 2015 - [info]  192.168.245.131: Resetting slave info succeeded.
Mon Nov 16 03:02:20 2015 - [info] Master failover to 192.168.245.131(192.168.245.131:3306) completed successfully.
Mon Nov 16 03:02:20 2015 - [info] Deleted server1 entry from /etc/masterha/app1.conf .
Mon Nov 16 03:02:20 2015 - [info] 

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

app1: MySQL Master failover 192.168.245.129(192.168.245.129:3306) to 192.168.245.131(192.168.245.131:3306) succeeded

Master 192.168.245.129(192.168.245.129:3306) is down!

Check MHA Manager logs at node3:/var/log/masterha/app1/app1.log for details.

Started automated(non-interactive) failover.
The latest slave 192.168.245.131(192.168.245.131:3306) has all relay logs for recovery.
Selected 192.168.245.131(192.168.245.131:3306) as a new master.
192.168.245.131(192.168.245.131:3306): OK: Applying all logs succeeded.
192.168.245.132(192.168.245.132:3306): This host has the latest relay log events.
Generating relay diff files from the latest slave succeeded.
192.168.245.132(192.168.245.132:3306): OK: Applying all logs succeeded. Slave started, replicating from 192.168.245.131(192.168.245.131:3306)
192.168.245.131(192.168.245.131:3306): Resetting slave info succeeded.
Master failover to 192.168.245.131(192.168.245.131:3306) completed successfully.

)

（2）修复故障master

通常情况下自动切换以后，待原master主机修复后，如果数据完整的情况下，可能想把原来master重新作为新主库的slave，这时我们可以借助当时自动切换时刻的MHA日志来完成对原master的修复。下面是提取相关日志的命令：

[root@node3 masterha]#  grep -i "All other slaves should start" /var/log/masterha/app1/app1.log
Mon Nov 16 03:02:20 2015 - [info]  All other slaves should start replication from here. Statement should be: CHANGE MASTER TO MASTER_HOST='192.168.245.131', MASTER_PORT=3306, MASTER_LOG_FILE='mysql-bin.000013', MASTER_LOG_POS=35416929, MASTER_USER='repl', MASTER_PASSWORD='xxx';

将change master to拿到原master上执行，那么就作为新master的slave了。

（3）在线切换测试

在许多情况下，需要将现有的主服务器迁移到另外一台服务器上。比如主服务器硬件故障，RAID 控制卡需要重建，将主服务器移到性能更好的服务器上等等。 MHA 提供快速切换和优雅的阻塞写入，这个切换过程只需要 0.5-2s 的时间，这段时间内数据是无法写入的。在很多情况下，0.5-2s 的阻塞写入是可以接受的。因此切换主服务器不需要计划分配维护时间窗口。

MHA在线切换的大概过程：
1.检测复制设置和确定当前主服务器
2.确定新的主服务器
3.阻塞写入到当前主服务器
4.等待所有从服务器赶上复制
5.授予写入到新的主服务器
6.重新设置从服务器

注意，在线切换的时候应用架构需要考虑以下两个问题：

1.自动识别master和slave的问题（master的机器可能会切换），如果采用了vip的方式，基本可以解决这个问题。

2.负载均衡的问题（可以定义大概的读写比例，每台机器可承担的负载比例，当有机器离开集群时，需要考虑这个问题）

为了保证数据完全一致性，在最快的时间内完成切换，MHA的在线切换必须满足以下条件才会切换成功，否则会切换失败。

1.所有slave的IO线程都在运行

2.所有slave的SQL线程都在运行

3.所有的show slave status的输出中Seconds_Behind_Master参数小于或者等于running_updates_limit秒，如果在切换过程中不指定running_updates_limit,那么默认情况下running_updates_limit为1秒。

4.在master端，通过show processlist输出，没有一个更新花费的时间大于running_updates_limit秒。

在线切换步骤如下：

首先，停掉MHA监控：

[root@node3 app1]# svc -d /service/masterha_app
[root@node3 app1]# checkstatus 
app1 is stopped(2:NOT_RUNNING).

其次，进行在线切换操作（模拟在线切换主库操作，原主库192.168.245.129变为slave，192.168.245.131提升为新的主库）

[root@node3 masterha]# /usr/bin/masterha_master_switch --conf=/etc/masterha/app1.conf --master_state=alive --new_master_host=192.168.245.131 --new_master_port=3306 --orig_master_is_new_slave --running_updates_limit=10000
Mon Nov 16 21:47:24 2015 - [info] MHA::MasterRotate version 0.56.
Mon Nov 16 21:47:24 2015 - [info] Starting online master switch..
Mon Nov 16 21:47:24 2015 - [info] 
Mon Nov 16 21:47:24 2015 - [info] * Phase 1: Configuration Check Phase..
Mon Nov 16 21:47:24 2015 - [info] 
Mon Nov 16 21:47:24 2015 - [warning] Global configuration file /etc/masterha_default.cnf not found. Skipping.
Mon Nov 16 21:47:24 2015 - [info] Reading application default configuration from /etc/masterha/app1.conf..
Mon Nov 16 21:47:24 2015 - [info] Reading server configuration from /etc/masterha/app1.conf..
Mon Nov 16 21:47:24 2015 - [info] GTID failover mode = 0
Mon Nov 16 21:47:24 2015 - [info] Current Alive Master: 192.168.245.129(192.168.245.129:3306)
Mon Nov 16 21:47:24 2015 - [info] Alive Slaves:
Mon Nov 16 21:47:24 2015 - [info]   192.168.245.132(192.168.245.132:3306)  Version=5.6.21-log (oldest major version between slaves) log-bin:enabled
Mon Nov 16 21:47:24 2015 - [info]     Replicating from 192.168.245.129(192.168.245.129:3306)

It is better to execute FLUSH NO_WRITE_TO_BINLOG TABLES on the master before switching. Is it ok to execute on 192.168.245.129(192.168.245.129:3306)? (YES/no): yes
Mon Nov 16 21:47:35 2015 - [info] Executing FLUSH NO_WRITE_TO_BINLOG TABLES. This may take long time..
Mon Nov 16 21:47:35 2015 - [info]  ok.
Mon Nov 16 21:47:35 2015 - [info] Checking MHA is not monitoring or doing failover..
Mon Nov 16 21:47:35 2015 - [info] Checking replication health on 192.168.245.132..
Mon Nov 16 21:47:35 2015 - [info]  ok.
Mon Nov 16 21:47:35 2015 - [error][/usr/share/perl5/vendor_perl/MHA/ServerManager.pm, ln1218] 192.168.245.131 is not alive!
Mon Nov 16 21:47:35 2015 - [error][/usr/share/perl5/vendor_perl/MHA/MasterRotate.pm, ln232] Failed to get new master!
Mon Nov 16 21:47:35 2015 - [error][/usr/share/perl5/vendor_perl/MHA/ManagerUtil.pm, ln177] Got ERROR:  at /usr/bin/masterha_master_switch line 53

这里怎么会报错呢？意思是131这个机器not alive，可是我去检查下没有问题啊，主从也ok的，后来发现了问题所在，原因是：之前启动的manager脚本中加上了--remove_dead_master_conf参数，导致appl.conf中没有131机器的配置。

然后到app1.conf中加上131机器的配置信息，再次执行：

[root@node3 masterha]# /usr/bin/masterha_master_switch --conf=/etc/masterha/app1.conf --master_state=alive --new_master_host=192.168.245.129 --new_master_port=3306 --orig_master_is_new_slave --running_updates_limit=10000
Mon Nov 16 22:34:36 2015 - [info] MHA::MasterRotate version 0.56.
Mon Nov 16 22:34:36 2015 - [info] Starting online master switch..
Mon Nov 16 22:34:36 2015 - [info] 
Mon Nov 16 22:34:36 2015 - [info] * Phase 1: Configuration Check Phase..
Mon Nov 16 22:34:36 2015 - [info] 
Mon Nov 16 22:34:36 2015 - [warning] Global configuration file /etc/masterha_default.cnf not found. Skipping.
Mon Nov 16 22:34:36 2015 - [info] Reading application default configuration from /etc/masterha/app1.conf..
Mon Nov 16 22:34:36 2015 - [info] Reading server configuration from /etc/masterha/app1.conf..
Mon Nov 16 22:34:37 2015 - [info] GTID failover mode = 0
Mon Nov 16 22:34:37 2015 - [info] Current Alive Master: 192.168.245.131(192.168.245.131:3306)
Mon Nov 16 22:34:37 2015 - [info] Alive Slaves:
Mon Nov 16 22:34:37 2015 - [info]   192.168.245.129(192.168.245.129:3306)  Version=5.6.23-log (oldest major version between slaves) log-bin:enabled
Mon Nov 16 22:34:37 2015 - [info]     Replicating from 192.168.245.131(192.168.245.131:3306)
Mon Nov 16 22:34:37 2015 - [info]     Primary candidate for the new Master (candidate_master is set)
Mon Nov 16 22:34:37 2015 - [info]   192.168.245.132(192.168.245.132:3306)  Version=5.6.21-log (oldest major version between slaves) log-bin:enabled
Mon Nov 16 22:34:37 2015 - [info]     Replicating from 192.168.245.131(192.168.245.131:3306)

It is better to execute FLUSH NO_WRITE_TO_BINLOG TABLES on the master before switching. Is it ok to execute on 192.168.245.131(192.168.245.131:3306)? (YES/no): yes
Mon Nov 16 22:34:38 2015 - [info] Executing FLUSH NO_WRITE_TO_BINLOG TABLES. This may take long time..
Mon Nov 16 22:34:38 2015 - [info]  ok.
Mon Nov 16 22:34:38 2015 - [info] Checking MHA is not monitoring or doing failover..
Mon Nov 16 22:34:38 2015 - [info] Checking replication health on 192.168.245.129..
Mon Nov 16 22:34:38 2015 - [info]  ok.
Mon Nov 16 22:34:38 2015 - [info] Checking replication health on 192.168.245.132..
Mon Nov 16 22:34:38 2015 - [info]  ok.
Mon Nov 16 22:34:38 2015 - [info] 192.168.245.129 can be new master.
Mon Nov 16 22:34:38 2015 - [info] 
From:
192.168.245.131(192.168.245.131:3306) (current master)
 +--192.168.245.129(192.168.245.129:3306)
 +--192.168.245.132(192.168.245.132:3306)

To:
192.168.245.129(192.168.245.129:3306) (new master)
 +--192.168.245.132(192.168.245.132:3306)
 +--192.168.245.131(192.168.245.131:3306)

Starting master switch from 192.168.245.131(192.168.245.131:3306) to 192.168.245.129(192.168.245.129:3306)? (yes/NO): yes
Mon Nov 16 22:34:39 2015 - [info] Checking whether 192.168.245.129(192.168.245.129:3306) is ok for the new master..
Mon Nov 16 22:34:39 2015 - [info]  ok.
Mon Nov 16 22:34:39 2015 - [info] 192.168.245.131(192.168.245.131:3306): SHOW SLAVE STATUS returned empty result. To check replication filtering rules, temporarily executing CHANGE MASTER to a dummy host.
Mon Nov 16 22:34:39 2015 - [info] 192.168.245.131(192.168.245.131:3306): Resetting slave pointing to the dummy host.
Mon Nov 16 22:34:39 2015 - [info] ** Phase 1: Configuration Check Phase completed.
Mon Nov 16 22:34:39 2015 - [info] 
Mon Nov 16 22:34:39 2015 - [info] * Phase 2: Rejecting updates Phase..
Mon Nov 16 22:34:39 2015 - [info] 
Mon Nov 16 22:34:39 2015 - [info] Executing master ip online change script to disable write on the current master:
Mon Nov 16 22:34:39 2015 - [info]   /etc/masterha/master_ip_online_change --command=stop --orig_master_host=192.168.245.131 --orig_master_ip=192.168.245.131 --orig_master_port=3306 --orig_master_user='root' --orig_master_password='root' --new_master_host=192.168.245.129 --new_master_ip=192.168.245.129 --new_master_port=3306 --new_master_user='root' --new_master_password='root' --orig_master_ssh_user=root --new_master_ssh_user=root   --orig_master_is_new_slave
Unknown option: orig_master_password
Unknown option: new_master_password
Unknown option: orig_master_ssh_user
Unknown option: new_master_ssh_user
Unknown option: orig_master_is_new_slave
Mon Nov 16 22:34:39 2015 501641 Set read_only on the new master.. ok.
Mon Nov 16 22:34:39 2015 508208 Set read_only=1 on the orig master.. ok.
Disabling the VIP on old master: 192.168.245.131 
Mon Nov 16 22:34:39 2015 876127 Killing all application threads..
Mon Nov 16 22:34:39 2015 876272 done.
Mon Nov 16 22:34:39 2015 - [info]  ok.
Mon Nov 16 22:34:39 2015 - [info] Locking all tables on the orig master to reject updates from everybody (including root):
Mon Nov 16 22:34:39 2015 - [info] Executing FLUSH TABLES WITH READ LOCK..
Mon Nov 16 22:34:39 2015 - [info]  ok.
Mon Nov 16 22:34:39 2015 - [info] Orig master binlog:pos is mysql-bin.000014:120.
Mon Nov 16 22:34:39 2015 - [info]  Waiting to execute all relay logs on 192.168.245.129(192.168.245.129:3306)..
Mon Nov 16 22:34:39 2015 - [info]  master_pos_wait(mysql-bin.000014:120) completed on 192.168.245.129(192.168.245.129:3306). Executed 0 events.
Mon Nov 16 22:34:39 2015 - [info]   done.
Mon Nov 16 22:34:39 2015 - [info] Getting new master's binlog name and position..
Mon Nov 16 22:34:39 2015 - [info]  mysql-bin.000024:120
Mon Nov 16 22:34:39 2015 - [info]  All other slaves should start replication from here. Statement should be: CHANGE MASTER TO MASTER_HOST='192.168.245.129', MASTER_PORT=3306, MASTER_LOG_FILE='mysql-bin.000024', MASTER_LOG_POS=120, MASTER_USER='repl', MASTER_PASSWORD='xxx';
Mon Nov 16 22:34:39 2015 - [info] Executing master ip online change script to allow write on the new master:
Mon Nov 16 22:34:39 2015 - [info]   /etc/masterha/master_ip_online_change --command=start --orig_master_host=192.168.245.131 --orig_master_ip=192.168.245.131 --orig_master_port=3306 --orig_master_user='root' --orig_master_password='root' --new_master_host=192.168.245.129 --new_master_ip=192.168.245.129 --new_master_port=3306 --new_master_user='root' --new_master_password='root' --orig_master_ssh_user=root --new_master_ssh_user=root   --orig_master_is_new_slave
Unknown option: orig_master_password
Unknown option: new_master_password
Unknown option: orig_master_ssh_user
Unknown option: new_master_ssh_user
Unknown option: orig_master_is_new_slave
Mon Nov 16 22:34:40 2015 165977 Set read_only=0 on the new master.
Enabling the VIP - 192.168.245.100/32 on the new master - 192.168.245.129 
Mon Nov 16 22:34:40 2015 - [info]  ok.
Mon Nov 16 22:34:40 2015 - [info] 
Mon Nov 16 22:34:40 2015 - [info] * Switching slaves in parallel..
Mon Nov 16 22:34:40 2015 - [info] 
Mon Nov 16 22:34:40 2015 - [info] -- Slave switch on host 192.168.245.132(192.168.245.132:3306) started, pid: 29022
Mon Nov 16 22:34:40 2015 - [info] 
Mon Nov 16 22:34:40 2015 - [info] Log messages from 192.168.245.132 ...
Mon Nov 16 22:34:40 2015 - [info] 
Mon Nov 16 22:34:40 2015 - [info]  Waiting to execute all relay logs on 192.168.245.132(192.168.245.132:3306)..
Mon Nov 16 22:34:40 2015 - [info]  master_pos_wait(mysql-bin.000014:120) completed on 192.168.245.132(192.168.245.132:3306). Executed 0 events.
Mon Nov 16 22:34:40 2015 - [info]   done.
Mon Nov 16 22:34:40 2015 - [info]  Resetting slave 192.168.245.132(192.168.245.132:3306) and starting replication from the new master 192.168.245.129(192.168.245.129:3306)..
Mon Nov 16 22:34:40 2015 - [info]  Executed CHANGE MASTER.
Mon Nov 16 22:34:40 2015 - [info]  Slave started.
Mon Nov 16 22:34:40 2015 - [info] End of log messages from 192.168.245.132 ...
Mon Nov 16 22:34:40 2015 - [info] 
Mon Nov 16 22:34:40 2015 - [info] -- Slave switch on host 192.168.245.132(192.168.245.132:3306) succeeded.
Mon Nov 16 22:34:40 2015 - [info] Unlocking all tables on the orig master:
Mon Nov 16 22:34:40 2015 - [info] Executing UNLOCK TABLES..
Mon Nov 16 22:34:40 2015 - [info]  ok.
Mon Nov 16 22:34:40 2015 - [info] Starting orig master as a new slave..
Mon Nov 16 22:34:40 2015 - [info]  Resetting slave 192.168.245.131(192.168.245.131:3306) and starting replication from the new master 192.168.245.129(192.168.245.129:3306)..
Mon Nov 16 22:34:40 2015 - [info]  Executed CHANGE MASTER.
Mon Nov 16 22:34:41 2015 - [info]  Slave started.
Mon Nov 16 22:34:41 2015 - [info] All new slave servers switched successfully.
Mon Nov 16 22:34:41 2015 - [info] 
Mon Nov 16 22:34:41 2015 - [info] * Phase 5: New master cleanup phase..
Mon Nov 16 22:34:41 2015 - [info] 
Mon Nov 16 22:34:41 2015 - [info]  192.168.245.129: Resetting slave info succeeded.
Mon Nov 16 22:34:41 2015 - [info] Switching master to 192.168.245.129(192.168.245.129:3306) completed successfully.

其中参数的意思：

--orig_master_is_new_slave 切换时加上此参数是将原 master 变为 slave 节点，如果不加此参数，原来的 master 将不启动

--running_updates_limit=10000,故障切换时,候选master 如果有延迟的话， mha 切换不能成功，加上此参数表示延迟在此时间范围内都可切换（单位为s），但是切换的时间长短是由recover 时relay 日志的大小决定

（4）切换时发送邮件：

需要创建一个send_mail脚本，然后将脚本路径写入app1.conf中即可。

#!/usr/bin/perl

#  Copyright (C) 2011 DeNA Co.,Ltd.
#
#  This program is free software; you can redistribute it and/or modify
#  it under the terms of the GNU General Public License as published by
#  the Free Software Foundation; either version 2 of the License, or
#  (at your option) any later version.
#
#  This program is distributed in the hope that it will be useful,
#  but WITHOUT ANY WARRANTY; without even the implied warranty of
#  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
#  GNU General Public License for more details.
#
#  You should have received a copy of the GNU General Public License
#   along with this program; if not, write to the Free Software
#  Foundation, Inc.,
#  51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA

## Note: This is a sample script and is not complete. Modify the script based on your environment.

use strict;
use warnings FATAL => 'all';
use Mail::Sender;
use Getopt::Long;

#new_master_host and new_slave_hosts are set only when recovering master succeeded
my ( $dead_master_host, $new_master_host, $new_slave_hosts, $subject, $body );
my $smtp='smtp.163.com';
my $mail_from='xxxx';
my $mail_user='xxxxx';
my $mail_pass='xxxxx';
my $mail_to=['xxxx','xxxx'];
GetOptions(
  'orig_master_host=s' => \$dead_master_host,
  'new_master_host=s'  => \$new_master_host,
  'new_slave_hosts=s'  => \$new_slave_hosts,
  'subject=s'          => \$subject,
  'body=s'             => \$body,
);

mailToContacts($smtp,$mail_from,$mail_user,$mail_pass,$mail_to,$subject,$body);

sub mailToContacts {
    my ( $smtp, $mail_from, $user, $passwd, $mail_to, $subject, $msg ) = @_;
    open my $DEBUG, "> /tmp/monitormail.log"
        or die "Can't open the debug      file:$!\n";
    my $sender = new Mail::Sender {
        ctype       => 'text/plain; charset=utf-8',
        encoding    => 'utf-8',
        smtp        => $smtp,
        from        => $mail_from,
        auth        => 'LOGIN',
        TLS_allowed => '0',
        authid      => $user,
        authpwd     => $passwd,
        to          => $mail_to,
        subject     => $subject,
        debug       => $DEBUG
    };

    $sender->MailMsg(
        {   msg   => $msg,
            debug => $DEBUG
        }
    ) or print $Mail::Sender::Error;
    return 1;
}



# Do whatever you want here

exit 0;

至此，mysql mha部分就搞定了，下面将会结合mha这个架构加入代理层，从而实现读写分离功能，主要采用360公司的Atlas，敬请期待。。。

MySQL运维和实战

mysql高可用研究(一) 主从+MHA架构

相关

标签