MySQL的MHA实现高可用性

MySQL高可用

（1）MMM: Multi-Master Replication Manager for MySQL，Mysql主主复制管理器是一套灵活的脚本程序，基于perl实现，用来对mysql replication进行监控和故障迁移，并能管理mysql Master-Master复制的配置(同一时间只有一个节点是可写的)

官网： http://www.mysql-mmm.org
         https://code.google.com/archive/p/mysql-master-master/downloads     #国外网站，需要梯子上网
（2）MHA： Master High Availability，对主节点进行监控，可实现自动故障转移至其它从节点；通过提升某一从节点为新的主节点，基于主从复制实现，还需要客户端配合实现，目前MHA主要支持一主多从的架构，要搭建MHA,要求一个复制集群中必须最少有三台数 据库服务器，一主二从，即一台充当master，一台充当备用master，另外一台充当从库，出于机器成本的考虑，淘宝进行了改造，目前淘宝TMHA已经支持一主一从

官网:https://code.google.com/archive/p/mysql-master-ha/
（3）Galera Cluster：wsrep(MySQL extended with the Write Set Replication)
        通过wsrep协议在全局实现复制；任何一节点都可读写，不需要主从复制，实现多主读写

（4）GR（Group Replication）：MySQL官方提供的组复制技术(MySQL 5.7.17引入的技术)，基于原生复制技术Paxos算法

github对应manager下载地址：https://github.com/yoshinorim/mha4mysql-manager/releases/

github对应node下载地址：https://github.com/yoshinorim/mha4mysql-node/releases/

MHA工作原理

1 从宕机崩溃的master保存二进制日志事件（binlog events）
2 识别含有最新更新的slave
3 应用差异的中继日志（relay log）到其他的slave
4 应用从master保存的二进制日志事件（binlog events）
5 提升一个slave为新的master
6 使其他的slave连接新的master进行复制

1、环境准备

A主机： 监控主服务器状态 192.168.34.105

B主机：主服务器         192.168.34.101

C主机： 从服务器1      192.168.34.102

D主机： 从服务器2      192.168.34.103 

2、实现三台服务器的一主多从

注意点：

① 每个节点都需开启二进制和中继日志，因为主会宕机，当主的机器修复完毕，可以作为从继续使用，所以中继日志是必须的；从也会在主宕机的时候，顶为主，所以二进制日志也是必须的

② 各从节点必须显示启用其read-only 属性，并关闭relay_log_purge 清理中继日志的功能

③ 注意每个mysql 的server-id都不能相同

开始配置主从服务器的配置文件

B主服务器上配置

vim /etc/my.cnf
[mysqld]
server-id=1   # 服务ID
log_bin=/data/logbin/mysql-bin  # 指定二进制日志路径
binlog_format=ROW   # 以行的方式保存二进制日志
skip-name-resolve  # 不将IP地址解析成名字
relay_log=relay-log  # 开启中继日志

C从服务器修改配置文件

vim /etc/my.cnf
[mysqld]
server-id=2
read-only
log_bin=/data/logbin/mysql-bin
binlog_format=row
skip-name-resolve
relay_log_purge=0   # 默认是清除中继日志，选择不清除
relay_log=relay-log  # 开启中继日志

D从服务器上修改配置文件　　

[mysqld]
server-id=3
read-only
skip-name-resolve
relay_log_purge=0 # 不清除中继日志
log_bin=/data/logbin/mysql-bin
binlog_format=row
relay_log=relay-log #开启中继日志

配置完主从复制之后，将三个服务器都重启

systemctl restart mariadb

先清除B、C、D之前实验所做的二进制日志文件,生产中不要使用，最好是将主服务器的数据完全备份，然后导入到从服务器，从服务器以最新二进制日志开始复制。

MariaDB [(none)]> reset master;

在B主机授权一个复制账号并实现主从复制

再在B主服务器创建一个赋予复制权限的用户

MariaDB [(none)]> grant replication  slave  on *.* to rpluser@'192.168.34.%' identified by 'centos';

在C从服务器上启动复制线程

MariaDB [(none)]> CHANGE MASTER TO
    ->   MASTER_HOST='192.168.34.101',
    ->   MASTER_USER='rpluser',
    ->   MASTER_PASSWORD='centos',
    ->   MASTER_PORT=3306,
    ->   MASTER_LOG_FILE='mysql-bin.000001',
    ->   MASTER_LOG_POS=245;

开启C从服务器slave

start slave

在D从服务器上开启复制线程和slave

MariaDB [(none)]> CHANGE MASTER TO  开启复制线程
    ->   MASTER_HOST='192.168.34.101',
    ->   MASTER_USER='rpluser',
    ->   MASTER_PASSWORD='centos',
    ->   MASTER_PORT=3306,
    ->   MASTER_LOG_FILE='mysql-bin.000001',
    ->   MASTER_LOG_POS=245;

Query OK, 0 rows affected (0.42 sec)

MariaDB [(none)]>
MariaDB [(none)]> start slave;  开启slave
Query OK, 0 rows affected (0.00 sec)

在B主机创建一个mha管理监控主从复制主机的账号

在B主服务器上创建一个赋予所有权限的用户，可随时管理主节点的对象用户账号，当主节点宕机后，会选举一个新的主节点

MariaDB [(none)]> grant all on *.* to mhauser@'192.168.34.%' identified by 'magedu';

验证主从复制　

在B主服务器上导入数据库，并查看结果

[root@centos7~]#mysql < hellodb_innodb.sql   导入数据库
MariaDB [(none)]> show databases;
+--------------------+
| Database           |
+--------------------+
| information_schema |
| hellodb            |
| mysql              |
| performance_schema |
| test               |
+--------------------+

在C和D从服务器上查看此时的数据库已经复制过去

MariaDB [(none)]> show databases;
+--------------------+
| Database           |
+--------------------+
| information_schema |
| hellodb            |
| mysql              |
| performance_schema |
| test               |
+--------------------+

3、配置MHA的准备

（1）MHA的安装

MHA软件由两部分组成，Manager工具包和Node工具包

Manager工具包主要包括以下几个工具：
masterha_check_ssh 检查MHA的SSH配置状况
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 故障转移（自动或手动）
masterha_conf_host 添加或删除配置的server信息

需安装2个包 rz传到A主机上，manager包依赖于epel源，需要打开epel源仓库进行安装：

github对应manager下载地址：https://github.com/yoshinorim/mha4mysql-manager/releases/

github对应node下载地址：https://github.com/yoshinorim/mha4mysql-node/releases/

mha4mysql-manager-0.56-0.el6.noarch.rpm

mha4mysql-node-0.56-0.el6.noarch.rpm

在管理节点上安装两个包：

mha4mysql-manager
mha4mysql-node

在被管理节点安装：

mha4mysql-node

在A主机安装MHA包：

[root@centos6~]#yum install  mha4mysql*  -y

在B、C、D主机上安装mha4mysql-node包

[root@node2 ~]# yum install mha4mysql-node-0.56-0.el6.noarch.rpm  -y

（2）实现各个节点都基于KEY验证

在A主机上生成基于Key验证，先生成公私钥对

[root@centos6~]#ssh-keygen

在A主机上生成自己的公钥文件

[root@centos6~]#ssh-copy-id 192.168.34.105

将生成的公钥文件整个目录全部复制到其他服务器上

[root@centos6~]#scp -rp /root/.ssh  192.168.34.101:/root/   复制到B主服务器的
[root@centos6~]#scp -rp /root/.ssh  192.168.34.102:/root/   复制到C从服务器
[root@centos6~]#scp -rp /root/.ssh  192.168.34.103:/root/   复制到D从服务器

定义MHA 管理配置文件，在A主机上新建一个目录文件并配置相关的配置文件

[root@centos6~]#mkdir /etc/mha/   新建一个mha目录，存放配置文件路径
[root@centos6~]#vim /etc/mha/cluster1.conf  新建一个cluster1,cnf配置文件并在里边配置相关数据
[server default]
user=mhauser    #管理用户
password=magedu  # 管理密码
manager_workdir=/data/mastermha/cluster1/     mha工作路径
manager_log=/data/mastermha/cluster1/manager.log   mha日志文件
remote_workdir=/data/mastermha/cluster1/      每个远程主机的工作目录
master_binlog_dir=/data/logbin/  # 指定数据库的二进制日志文件路径，此路径是主从复制数据库的路径，如果单独存放二进制日志，一定要指定，否则会报错。
ssh_user=root    #使用ssh远程连接
repl_user=rpluser        # master授权的用户名
repl_password=centos     # master授权的密码
ping_interval=1           ping时间时长

[server1]
hostname=192.168.34.101  主服务器IP地址
candidate_master=1
[server2]
hostname=192.168.34.102   候选主服务器IP地址
candidate_master=1   # 开启候选功能
[server3]
hostname=192.168.34.103
candidate_master=1  # 开启候选功能

开启服务前检测

在所有节点实现相互之间ssh key验证　

[root@centos7~]#masterha_check_ssh --conf=/etc/mha/cluster1.conf
Wed Nov 27 10:41:04 2019 - [warning] Global configuration file /etc/masterha_default.cnf not found. Skipping.
Wed Nov 27 10:41:04 2019 - [info] Reading application default configuration from /etc/mha/cluster1.conf..
Wed Nov 27 10:41:04 2019 - [info] Reading server configuration from /etc/mha/cluster1.conf..
Wed Nov 27 10:41:04 2019 - [info] Starting SSH connection tests..
Wed Nov 27 10:41:05 2019 - [debug]
Wed Nov 27 10:41:04 2019 - [debug]  Connecting via SSH from root@192.168.34.101(192.168.34.101:22) to root@192.168.34.102(192.168.34.102:22)..
Wed Nov 27 10:41:05 2019 - [debug]   ok.
Wed Nov 27 10:41:05 2019 - [debug]  Connecting via SSH from root@192.168.34.101(192.168.34.101:22) to root@192.168.34.103(192.168.34.103:22)..
Warning: Permanently added '192.168.34.103' (ECDSA) to the list of known hosts.
Wed Nov 27 10:41:05 2019 - [debug]   ok.
Wed Nov 27 10:41:06 2019 - [debug]
Wed Nov 27 10:41:04 2019 - [debug]  Connecting via SSH from root@192.168.34.102(192.168.34.102:22) to root@192.168.34.101(192.168.34.101:22)..
Wed Nov 27 10:41:05 2019 - [debug]   ok.
Wed Nov 27 10:41:05 2019 - [debug]  Connecting via SSH from root@192.168.34.102(192.168.34.102:22) to root@192.168.34.103(192.168.34.103:22)..
Wed Nov 27 10:41:06 2019 - [debug]   ok.
Wed Nov 27 10:41:06 2019 - [debug]
Wed Nov 27 10:41:05 2019 - [debug]  Connecting via SSH from root@192.168.34.103(192.168.34.103:22) to root@192.168.34.101(192.168.34.101:22)..
Warning: Permanently added '192.168.34.103' (ECDSA) to the list of known hosts.
Wed Nov 27 10:41:05 2019 - [debug]   ok.
Wed Nov 27 10:41:05 2019 - [debug]  Connecting via SSH from root@192.168.34.103(192.168.34.103:22) to root@192.168.34.102(192.168.34.102:22)..
Wed Nov 27 10:41:06 2019 - [debug]   ok.
Wed Nov 27 10:41:06 2019 - [info] All SSH connection tests passed successfully.   验证已经成功

验证复制集群的连接配置参数是否OK

[root@openstack ~]# masterha_check_repl  --conf=/etc/mha/cluster1.conf
Sun Mar 29 20:42:03 2020 - [warning] Global configuration file /etc/masterha_default.cnf not found. Skipping.
Sun Mar 29 20:42:03 2020 - [info] Reading application default configuration from /etc/mha/cluster1.conf..
Sun Mar 29 20:42:03 2020 - [info] Reading server configuration from /etc/mha/cluster1.conf..
Sun Mar 29 20:42:03 2020 - [info] MHA::MasterMonitor version 0.56.
Sun Mar 29 20:42:04 2020 - [info] GTID failover mode = 0
Sun Mar 29 20:42:04 2020 - [info] Dead Servers:
Sun Mar 29 20:42:04 2020 - [info] Alive Servers:
Sun Mar 29 20:42:04 2020 - [info]   192.168.7.101(192.168.7.101:3306)
Sun Mar 29 20:42:04 2020 - [info]   192.168.7.102(192.168.7.102:3306)
Sun Mar 29 20:42:04 2020 - [info]   192.168.7.103(192.168.7.103:3306)
Sun Mar 29 20:42:04 2020 - [info] Alive Slaves:
Sun Mar 29 20:42:04 2020 - [info]   192.168.7.102(192.168.7.102:3306)  Version=5.5.64-MariaDB (oldest major version between slaves) log-bin:enabled
Sun Mar 29 20:42:04 2020 - [info]     Replicating from 192.168.7.101(192.168.7.101:3306)
Sun Mar 29 20:42:04 2020 - [info]     Primary candidate for the new Master (candidate_master is set)
Sun Mar 29 20:42:04 2020 - [info]   192.168.7.103(192.168.7.103:3306)  Version=10.3.10-MariaDB-log log-bin:enabled
Sun Mar 29 20:42:04 2020 - [info]     Replicating from 192.168.7.101(192.168.7.101:3306)
Sun Mar 29 20:42:04 2020 - [info]     Primary candidate for the new Master (candidate_master is set)
Sun Mar 29 20:42:04 2020 - [info] Current Alive Master: 192.168.7.101(192.168.7.101:3306)
Sun Mar 29 20:42:04 2020 - [info] Checking slave configurations..
Sun Mar 29 20:42:04 2020 - [warning]  relay_log_purge=0 is not set on slave 192.168.7.102(192.168.7.102:3306).
Sun Mar 29 20:42:04 2020 - [info] Checking replication filtering settings..
Sun Mar 29 20:42:04 2020 - [info]  binlog_do_db= , binlog_ignore_db=
Sun Mar 29 20:42:04 2020 - [info]  Replication filtering check ok.
Sun Mar 29 20:42:04 2020 - [info] GTID (with auto-pos) is not supported
Sun Mar 29 20:42:04 2020 - [info] Starting SSH connection tests..
Sun Mar 29 20:42:06 2020 - [info] All SSH connection tests passed successfully.
Sun Mar 29 20:42:06 2020 - [info] Checking MHA Node version..
Sun Mar 29 20:42:07 2020 - [info]  Version check ok.
Sun Mar 29 20:42:07 2020 - [info] Checking SSH publickey authentication settings on the current master..
Sun Mar 29 20:42:07 2020 - [info] HealthCheck: SSH to 192.168.7.101 is reachable.
Sun Mar 29 20:42:07 2020 - [info] Master MHA Node version is 0.56.
Sun Mar 29 20:42:07 2020 - [info] Checking recovery script configurations on 192.168.7.101(192.168.7.101:3306)..
Sun Mar 29 20:42:07 2020 - [info]   Executing command: save_binary_logs --command=test --start_pos=4 --binlog_dir=/data/logbin/ --output_file=/data/mastermha/cluster1//save_binary_logs_test --manager_version=0.56 --start_file=mysql-bin.000001
Sun Mar 29 20:42:07 2020 - [info]   Connecting to root@192.168.7.101(192.168.7.101:22)..
  Creating /data/mastermha/cluster1 if not exists..    ok.
  Checking output directory is accessible or not..
   ok.
  Binlog found at /data/logbin/, up to mysql-bin.000001
Sun Mar 29 20:42:08 2020 - [info] Binlog setting check done.
Sun Mar 29 20:42:08 2020 - [info] Checking SSH publickey authentication and checking recovery script configurations on all alive slave servers..
Sun Mar 29 20:42:08 2020 - [info]   Executing command : apply_diff_relay_logs --command=test --slave_user='mhauser' --slave_host=192.168.7.102 --slave_ip=192.168.7.102 --slave_port=3306 --workdir=/data/mastermha/cluster1/ --target_version=5.5.64-MariaDB --manager_version=0.56 --relay_log_info=/data/mysql/relay-log.info  --relay_dir=/data/mysql/  --slave_pass=xxx
Sun Mar 29 20:42:08 2020 - [info]   Connecting to root@192.168.7.102(192.168.7.102:22)..
Creating directory /data/mastermha/cluster1/.. done.
  Checking slave recovery environment settings..
    Opening /data/mysql/relay-log.info ... ok.
    Relay log found at /data/mysql, up to centos7-relay-bin.000002
    Temporary relay log file is /data/mysql/centos7-relay-bin.000002
    Testing mysql connection and privileges.. done.
    Testing mysqlbinlog output.. done.
    Cleaning up test file(s).. done.
Sun Mar 29 20:42:08 2020 - [info]   Executing command : apply_diff_relay_logs --command=test --slave_user='mhauser' --slave_host=192.168.7.103 --slave_ip=192.168.7.103 --slave_port=3306 --workdir=/data/mastermha/cluster1/ --target_version=10.3.10-MariaDB-log --manager_version=0.56 --relay_log_info=/data/mysql/relay-log.info  --relay_dir=/data/mysql/  --slave_pass=xxx
Sun Mar 29 20:42:08 2020 - [info]   Connecting to root@192.168.7.103(192.168.7.103:22)..
  Checking slave recovery environment settings..
    Opening /data/mysql/relay-log.info ... ok.
    Relay log found at /data/mysql, up to node2-relay-bin.000002
    Temporary relay log file is /data/mysql/node2-relay-bin.000002
    Testing mysql connection and privileges.. done.
    Testing mysqlbinlog output.. done.
    Cleaning up test file(s).. done.
Sun Mar 29 20:42:09 2020 - [info] Slaves settings check done.
Sun Mar 29 20:42:09 2020 - [info]
192.168.7.101(192.168.7.101:3306) (current master)
 +--192.168.7.102(192.168.7.102:3306)
 +--192.168.7.103(192.168.7.103:3306)

Sun Mar 29 20:42:09 2020 - [info] Checking replication health on 192.168.7.102..
Sun Mar 29 20:42:09 2020 - [info]  ok.
Sun Mar 29 20:42:09 2020 - [info] Checking replication health on 192.168.7.103..
Sun Mar 29 20:42:09 2020 - [info]  ok.
Sun Mar 29 20:42:09 2020 - [warning] master_ip_failover_script is not defined.
Sun Mar 29 20:42:09 2020 - [warning] shutdown_script is not defined.
Sun Mar 29 20:42:09 2020 - [info] Got exit code 0 (Not master dead).

MySQL Replication Health is OK.  # 此检测主要是检测二进制日志文件是否可以查到。

启动MHA

（1）开启mha服务,前台执行监控主从复制节点不安全，我们以后台方式启动

[root@openstack ~]# nohup masterha_manager -conf=/etc/mha/cluster1.conf &> /data/mastermha/cluster1/manager.log &

（2）检查此时mha的状态

[root@centos7~]#tail /data/mastermha/cluster1/manager.log  -f
192.168.34.101(192.168.34.101:3306) (current master)
 +--192.168.34.102(192.168.34.102:3306)
 +--192.168.34.103(192.168.34.103:3306)

Wed Nov 27 11:00:13 2019 - [warning] master_ip_failover_script is not defined.
Wed Nov 27 11:00:13 2019 - [warning] shutdown_script is not defined.
Wed Nov 27 11:00:13 2019 - [info] Set master ping interval 1 seconds.
Wed Nov 27 11:00:13 2019 - [warning] secondary_check_script is not defined. It is highly recommended setting it to check master reachability from two or more routes.
Wed Nov 27 11:00:13 2019 - [info] Starting ping health check on 192.168.34.101(192.168.34.101:3306)..
Wed Nov 27 11:00:13 2019 - [info] Ping(SELECT) succeeded, waiting until MySQL doesn't respond..   可以看到此时的状态是成功的，直至mysql不响应

测试MHA 测试故障转移

（1）在A主机上先创建大文件，并让其他从服务器进行复制，从服务器不复制的状态下，当主服务器宕机后，不能推举一个集群从服务器当新主。

create table testlog (id int auto_increment primary key,name char(10),age int default 20);

delimiter $$

create procedure  pro_testlog()
begin
declare i int;
set i = 1;
while i <= 100000
do  insert into testlog(name,age) values (concat('wang',i),i);
set i = i +1;
end while;
end$$

delimiter ;
call pro_testlog;

（2）将数据库停掉，模拟宕机实验。

# systemctl stop mariadb

（3）在A服务器上查看此时监控状态，此时当前MHA的监控退出，提升一个从服务器作为主服务器

[root@centos7~]#masterha_manager --conf=/etc/mha/cluster1.conf
Wed Nov 27 11:00:07 2019 - [warning] Global configuration file /etc/masterha_default.cnf not found. Skipping.
Wed Nov 27 11:00:07 2019 - [info] Reading application default configuration from /etc/mha/cluster1.conf..
Wed Nov 27 11:00:07 2019 - [info] Reading server configuration from /etc/mha/cluster1.conf..
  Creating /data/mastermha/cluster1 if not exists..    ok.
  Checking output directory is accessible or not..
   ok.
  Binlog found at /var/lib/mysql, up to mariadb-bin.000001
Wed Nov 27 11:22:28 2019 - [warning] Global configuration file /etc/masterha_default.cnf not found. Skipping.
Wed Nov 27 11:22:28 2019 - [info] Reading application default configuration from /etc/mha/cluster1.conf..
Wed Nov 27 11:22:28 2019 - [info] Reading server configuration from /etc/mha/cluster1.conf..

（4）此时查看排错日志，可以看到将哪个从服务器作为主服务器

[root@centos7~]#tail /data/mastermha/cluster1/manager.log

Started automated(non-interactive) failover.
The latest slave 192.168.34.102(192.168.34.102:3306) has all relay logs for recovery.
Selected 192.168.34.102(192.168.34.102:3306) as a new master.      可以看到此时的新主服务器是192.168.34.102
192.168.34.102(192.168.34.102:3306): OK: Applying all logs succeeded.
192.168.34.103(192.168.34.103:3306): This host has the latest relay log events.
Generating relay diff files from the latest slave succeeded.
192.168.34.103(192.168.34.103:3306): OK: Applying all logs succeeded. Slave started, replicating from 192.168.34.102(192.168.34.102:3306)
192.168.34.102(192.168.34.102:3306): Resetting slave info succeeded.
Master failover to 192.168.34.102(192.168.34.102:3306) completed successfully.

查看日志里边的内容，也可以确认此时的192.168.34.102是新主服务器

[root@centos7~]#cat  /data/mastermha/cluster1/manager.log
Wed Nov 27 11:22:32 2019 - [info] New master is 192.168.34.102(192.168.34.102:3306)

可以查看此时新的主服务器的read-only已经关闭

MariaDB [(none)]> show variables like 'read_only';
+---------------+-------+
| Variable_name | Value |
+---------------+-------+
| read_only     | OFF   |
+---------------+-------+

查看D从服务器指向的是新主服务器（192.168.34.102）

MariaDB [(none)]> show slave status\G;
*************************** 1. row ***************************
               Slave_IO_State: Waiting for master to send event
                  Master_Host: 192.168.34.102
                  Master_User: rpluser
                  Master_Port: 3306
                Connect_Retry: 60
              Master_Log_File: mariadb-bin.000001
          Read_Master_Log_Pos: 245
               Relay_Log_File: mariadb-relay-bin.000002
                Relay_Log_Pos: 531
        Relay_Master_Log_File: mariadb-bin.000001
             Slave_IO_Running: Yes
            Slave_SQL_Running: Yes

注意：如果之前down掉的主服务器恢复，其他从服务器也不会再指向此服务器，只能将恢复的服务器作为从服务器更好，以免在操作过程中出现其他错误。

将宕机的主服务器做为从服务器

原有 master 节点故障后，需要重新准备好一个新的 MySQL 节点。基于来自于master 节点的备份恢复数据后，将其配置为新的 master 的从节点即可。注意，新加入的节点如果为新增节点，其 IP 地址要配置为原来 master 节点的 IP ，否则，还需要修改 cluster1.conf 中相应的 ip 地址。随后再次启动 manager ，并再次检测其状态。

1、在新成立的主服务器上完全备份数据库

[root@centos7 ~]# mysqldump  -A  > mysql-all-backup.sql
[root@centos7 ~]# scp mysql-all-backup.sql 192.168.7.101:  # 将完全备份的数据库传到之前宕机的主服务器上　　

2、在修复的主机上导入备份的数据

[root@centos7 ~]# mysql < mysql-all-backup.sql

修改配置文件，设置为只读

[mysqld]
server-id=4 # 与前面主从服务器的ID不要一样
read-only
skip-name-resolve  # 不反向解析为域名
relay_log_purge=0 # 不清除中继日志
log_bin=/data/logbin/mysql-bin
binlog_format=row
relay_log=relay-log  # 开启中继日志

3、在新主上查看二进制日志起始位置

（1）在新主上查看此时的二进制日志的起始位置，将修复的主服务器作为从进行复制

[root@centos7 ~]# mysql
Welcome to the MariaDB monitor.  Commands end with ; or \g.
Your MariaDB connection id is 81
Server version: 5.5.64-MariaDB MariaDB Server

Copyright (c) 2000, 2018, Oracle, MariaDB Corporation Ab and others.

Type 'help;' or '\h' for help. Type '\c' to clear the current input statement.

MariaDB [(none)]> show master logs;
+------------------+-----------+
| Log_name         | File_size |
+------------------+-----------+
| mysql-bin.000001 |       245 |
+------------------+-----------+

4、在旧主上开始变为从服务器进行复制

CHANGE MASTER TO # 开始复制
  MASTER_HOST='192.168.7.102', #新主的IP地址
  MASTER_USER='rpluser',
  MASTER_PASSWORD='centos',
  MASTER_PORT=3306,
  MASTER_LOG_FILE='mysql-bin.000001', # 二进制起始名称
  MASTER_LOG_POS=245;   # 二进制日志文件起始大小

MariaDB [(none)]> start salve ;

MariaDB [(none)]> show slave status\G;  # 查看从服务器的状态。
*************************** 1. row ***************************
                Slave_IO_State: Waiting for master to send event
                   Master_Host: 192.168.7.102
                   Master_User: rpluser
                   Master_Port: 3306
                 Connect_Retry: 60
               Master_Log_File: mysql-bin.000001
           Read_Master_Log_Pos: 245
                Relay_Log_File: centos7-relay-bin.000002
                 Relay_Log_Pos: 544
         Relay_Master_Log_File: mysql-bin.000001
              Slave_IO_Running: Yes
             Slave_SQL_Running: Yes

5、在MHA管理主机上验证从服务器加入后状态

如果此时修复的主机显示OK，说明正常。

masterha_check_repl  --conf=/etc/mha/cluster1.conf

注意：

(1) 在生产环境中，当你的主节点挂了后，一定要在从节点上做一个备份，拿着备份文件把主节点手动提升为从节点，并指明从哪一个日志文件的位置开始复制

(2) 每一次自动完成转换后，每一次的(replication health ) 检测不ok 始终都是启动不了必须手动修复主节点，除非你改配置文件

(3) 手动修复主节点提升为从节点后，再次运行检测命令

masterha_check_repl --conf=/etc/mha/cluster1.conf

(4) 再次运行起来就恢复成功了

masterha_manager --conf=/etc/mha/cluster1.conf