MHA 数据库高可用+ GTID 同步测试部署
目录
一、部署MHA 环境准备
1.1 虚拟机信息:
192.168.1.57 node57 数据库主节点
192.168.1.58 node58 数据库从节点1
192.168.1.59 node59 数据库从节点2
192.168.1.59 node59 MHA服务
192.168.1.59 vip vip
1.2 安装包信息:良心博主已经下载好rpm离线版
manager
mha4mysql-manager-0.58-0.el7.centos.noarch.rpm
node
mha4mysql-node-0.58-0.el7.centos.noarch.rpm
mysql 监控
perl-DBD-MySQL-4.023-6.el7.x86_64.rpm
epel-release-latest-7.noarch.rpm mha4mysql-manager-0.58-0.el7.centos.noarch.rpm mha4mysql-node-0.58-0.el7.centos.noarch.rpm mysql57-community-release-el7-10.noarch.rpm perl-5.16.3-294.el7_6.x86_64.rpm perl-Class-Load-0.20-3.el7.noarch.rpm perl-Compress-Raw-Bzip2-2.061-3.el7.x86_64.rpm perl-Compress-Raw-Zlib-2.061-4.el7.x86_64.rpm perl-Config-Tiny-2.14-7.el7.noarch.rpm perl-Data-Dumper-2.145-3.el7.x86_64.rpm perl-Data-OptList-0.107-9.el7.noarch.rpm perl-DBD-MySQL-4.023-6.el7.x86_64.rpm perl-DBI-1.627-4.el7.x86_64.rpm perl-devel-5.16.3-294.el7_6.x86_64.rpm perl-IO-Compress-2.061-2.el7.noarch.rpm perl-IO-Socket-IP-0.21-5.el7.noarch.rpm perl-IO-Socket-SSL-1.94-7.el7.noarch.rpm perl-List-MoreUtils-0.33-9.el7.x86_64.rpm perl-MailTools-2.12-2.el7.noarch.rpm perl-Module-Implementation-0.06-6.el7.noarch.rpm perl-Module-Runtime-0.013-4.el7.noarch.rpm perl-Mozilla-CA-20130114-5.el7.noarch.rpm perl-Net-Daemon-0.48-5.el7.noarch.rpm perl-Net-LibIDN-0.12-15.el7.x86_64.rpm perl-Net-SMTP-SSL-1.01-13.el7.noarch.rpm perl-Net-SSLeay-1.55-6.el7.x86_64.rpm perl-Package-DeprecationManager-0.13-7.el7.noarch.rpm perl-Package-Stash-0.34-2.el7.noarch.rpm perl-Package-Stash-XS-0.26-3.el7.x86_64.rpm perl-Params-Util-1.07-6.el7.x86_64.rpm perl-Params-Validate-1.08-4.el7.x86_64.rpm perl-PlRPC-0.2020-14.el7.noarch.rpm perl-Sub-Install-0.926-6.el7.noarch.rpm perl-Sys-Syslog-0.33-3.el7.x86_64.rpm perl-TimeDate-2.30-2.el7.noarch.rpm perl-Try-Tiny-0.12-2.el7.noarch.rpm
二、MHA 基本信息操作
MHA 通俗就是监控数据库状态，实现高可用，即使宕机一台，不影响程序使用，原理:MHA 有检测VIP状态，检测不到时，通过自带的脚本在slave 端选数据库主节点，并启动VIP(VIP漂移)。
2.1 添加IP映射
cat >> /etc/hosts << EOF
############ MHA IP hosts ############ 192.168.1.57 node57
192.168.1.58 node58
192.168.1.59 node59
192.168.1.59 vip
EOF
2.2 安装依赖包
cd /home/backupfile/mha
yum install wget -y
yum install -y perl-DBD-MySQL
perl-Config-Tiny
perl-Log-Dispatch
perl-Parallel-ForkManager
yum install perl-Time-HiRes perl-Parallel-ForkManager -y
wget http://mysql-master-ha.googlecode.com/files/mha4mysql-manager-0.55-0.el6.noarch.rpm
wget http://mysql-master-ha.googlecode.com/files/mha4mysql-node-0.54-0.el6.noarch.rpm
rpm -ivh mha4mysql-node-0.54-0.el6.noarch.rpm
rpm -ivh mha4mysql-manager-0.55-0.el6.noarch.rpm
三、部署mysql 主从参考：mysql搭建以及主从复制
https://www.cnblogs.com/pythonx/p/12054819.html
四、MHA 搭建
安装MHA服务，如果资源允许的话，在两台从节点安装，我这里为减轻mysql主服务压力，吧MHA安装在两台从节点，或者你mysql是一主一从，在主从都可以安装MHA服务。
以下操作在 192.168.1.59 操作
4.1 创建MHA工作目录
mkdir /etc/masterha
4.2 创建MHA日志目录
mkdir /etc/masterha/log
4.3 编辑MHA配置文件
slave : 复制用户
manager ： 管理Mysql用户(当时是写root，不知道为什么健康检查一直 是 is NOT OK ，改为其他普通用户就可以了)
master_ip_online_change_script IP漂移故障脚本检查，网上有的添加，我这里直接注释
[root@node59]
vi /etc/masterha/app1.cnf [server default]
#MHA配置文件 remote_workdir=/etc/masterha/app1
#MHA日志 manager_log=/etc/masterha/log/manager.log
#MHA工作目录 manager_workdir=/etc/masterha
#mysql日志目录，根据自己安装目录定义 master_binlog_dir=/home/ap/mysql/data #切换调用的脚本 master_ip_failover_script=/etc/masterha/master_ip_failover report_script=/etc/masterha/send_report #网上说有添加故障脚本，我这里测试没有通过，反而注释就可以
#master_ip_online_change_script=/etc/masterha/master_ip_online_change #MySQL的用户和密码 user=manager password=123456 port=31061 #系统ssh用户 ssh_user=root ssh_port=22 #复制用户 ping_interval=2 repl_user=slave repl_password=123456 port=3306 ssh_port=22
[server1] candidate_master=1 check_repl_delay=0 hostname=192.168.1.57 port=3306 ssh_port=22 [server2] candidate_master=1 check_repl_delay=0 hostname=192.168.1.58 port=3306 ssh_port=22 [server3] candidate_master=1 no_master=1 check_repl_delay=0 hostname=192.168.1.59 ssh_port=22 port=3306
#//设置为候选master，如果设置该参数以后，发生主从切换以后将会将此从库提升为主库，即使这个主库不是集群中事件最新的slave
#//默认情况下如果一个slave落后master 100M的relay logs的话，MHA将不会选择该slave作为一个新的master
#因为对于这个slave的恢复需要花费很长时间，通过设置check_repl_delay=0,
#MHA触发切换在选择一个新的master的时候将会忽略复制延时，这个参数对于设置了candidate_master=1的主机非常有用，
#因为这个候选主在切换的过程中一定是新的master
4.4 添加 vip启动脚本
这里我使用的是ifconfig 命令添加，也可以使用ip 命令
/sbin/ip addr add $vip dev ens192
/sbin/ip addr del $vip dev ens192
[root@node59]
vi init_vip.sh
#!/bin/bash
vip="192.168.1.249/24"
key="2"
/sbin/ifconfig ens192:$key $vip
4.5 添加 vip 删除脚本
[root@node59] vi drop_vip.sh
#!/bin/bash
vip="192.168.1.249/24"
key="2"
/sbin/ifconfig ens192:$key down
4.6 mysqlbinlog设置软连接
我mysql安装/home/ap/mysql
ln -s /home/ap/mysql/bin/mysql /usr/bin/mysql ln -s /home/ap/mysql/bin/mysqlbinlog /usr/bin/mysqlbinlog
五、各个主机配置免密登录，配置互信
以下操作每台都执行
5.1 生成秘钥，如果有秘钥可以忽略此步骤 ssh-keygen 5.2 拷贝公钥 ssh-copy-id node57 ssh-copy-id node58 ssh-copy-id node59 5.3 测试是否可以免密登录，第一次需要输入yes ssh node57 ssh node58 ssh node59
六、MHA服务检查
6.1、主机信任检查
显示All SSH connection tests passed successfully. 为正常。
masterha_check_ssh --conf=/etc/masterha/app1.cnf
Wed Jan 8 18:00:57 2020 - [warning] Global configuration file /etc/masterha_default.cnf not found. Skipping.
Wed Jan 8 18:00:57 2020 - [info] Reading application default configuration from /etc/masterha/app1.cnf..
Wed Jan 8 18:00:57 2020 - [info] Updating application default configuration from /etc/masterha/pm/load_cnf..
Can't exec "/etc/masterha/pm/load_cnf": No such file or directory at /usr/share/perl5/vendor_perl/MHA/Config.pm line 365.
Wed Jan 8 18:00:57 2020 - [info] Reading server configuration from /etc/masterha/app1.cnf..
Wed Jan 8 18:00:57 2020 - [info] Starting SSH connection tests..
Wed Jan 8 18:00:58 2020 - [debug]
Wed Jan 8 18:00:57 2020 - [debug] Connecting via SSH from root@192.168.1.57(192.168.1.57:22) to root@192.168.1.58(192.168.1.58:22)..
Wed Jan 8 18:00:57 2020 - [debug] ok.
Wed Jan 8 18:00:58 2020 - [debug]
Wed Jan 8 18:00:58 2020 - [debug] Connecting via SSH from root@192.168.1.57(192.168.1.57:22) to root@192.168.1.59(192.168.1.59:22)..
Wed Jan 8 18:00:58 2020 - [debug] ok.
Wed Jan 8 18:00:58 2020 - [debug] Connecting via SSH from root@192.168.1.58(192.168.1.58:22) to root@192.168.1.57(192.168.1.57:22)..
Wed Jan 8 18:00:58 2020 - [debug] ok.
Wed Jan 8 18:00:57 2020 - [debug] Connecting via SSH from root@192.168.1.58(192.168.1.58:22) to root@192.168.1.58(192.168.1.59:22)..
Wed Jan 8 18:00:57 2020 - [debug] ok.
Wed Jan 8 18:00:58 2020 - [debug]
Wed Jan 8 18:00:58 2020 - [debug] Connecting via SSH from root@192.168.1.59(192.168.1.59:22) to root@192.168.1.57(192.168.1.57:22)..
Wed Jan 8 18:00:58 2020 - [debug] ok.
Wed Jan 8 18:00:58 2020 - [debug] Connecting via SSH from root@192.168.1.59(192.168.1.59:22) to root@192.168.1.147(192.168.1.58:22)..
Wed Jan 8 18:00:58 2020 - [debug] ok.
Wed Jan 8 18:00:58 2020 - [info] All SSH connection tests passed successfully.
6.2、mysql健康检查
最后显示 MySQL Replication Health is OK. 为正常。
masterha_check_repl --conf=/etc/masterha/app1.cnf
Sat Jan 11 17:03:41 2020 - [warning] Global configuration file /etc/masterha_default.cnf not found. Skipping.
Sat Jan 11 17:03:41 2020 - [info] Reading application default configuration from /etc/masterha/app1.cnf..
Sat Jan 11 17:03:41 2020 - [info] Reading server configuration from /etc/masterha/app1.cnf..
Sat Jan 11 17:03:41 2020 - [info] MHA::MasterMonitor version 0.58.
Sat Jan 11 17:03:42 2020 - [info] GTID failover mode = 1
Sat Jan 11 17:03:42 2020 - [info] Dead Servers:
Sat Jan 11 17:03:42 2020 - [info] Alive Servers:
Sat Jan 11 17:03:42 2020 - [info] 192.168.1.57(192.168.1.57:3306)
Sat Jan 11 17:03:42 2020 - [info] 192.168.1.58(192.168.1.58:3306)
Sat Jan 11 17:03:42 2020 - [info] 192.168.1.59(192.168.1.59:3306)
Sat Jan 11 17:03:42 2020 - [info] Alive Slaves:
Sat Jan 11 17:03:42 2020 - [info] 192.168.1.58(192.168.1.58:3306) Version=5.7.26-log (oldest major version between slaves) log-bin:enabled
Sat Jan 11 17:03:42 2020 - [info] GTID ON
Sat Jan 11 17:03:42 2020 - [info] Replicating from 192.168.1.57(192.168.1.57:3306)
Sat Jan 11 17:03:42 2020 - [info] Primary candidate for the new Master (candidate_master is set)
Sat Jan 11 17:03:42 2020 - [info] Current Alive Master: 192.168.1.57(192.168.1.57:3306)
Sat Jan 11 17:03:42 2020 - [info] Checking slave configurations..
Sat Jan 11 17:03:42 2020 - [info] Checking replication filtering settings..
Sat Jan 11 17:03:42 2020 - [info] binlog_do_db= , binlog_ignore_db=
Sat Jan 11 17:03:42 2020 - [info] Replication filtering check ok.
Sat Jan 11 17:03:42 2020 - [info] GTID (with auto-pos) is supported. Skipping all SSH and Node package checking.
Sat Jan 11 17:03:42 2020 - [info] Checking SSH publickey authentication settings on the current master..
Sat Jan 11 17:03:42 2020 - [info] HealthCheck: SSH to 192.168.1.57 is reachable.
Sat Jan 11 17:03:42 2020 - [info]
192.168.1.57(192.168.1.57:3306) (current master)
+--192.168.1.58(192.168.1.58:3306)
Sat Jan 11 17:03:42 2020 - [info] Checking replication health on 192.168.1.58..
Sat Jan 11 17:03:42 2020 - [info] ok.
Sat Jan 11 17:03:42 2020 - [info] Checking master_ip_failover_script status:
Sat Jan 11 17:03:42 2020 - [info] /etc/masterha/master_ip_failover --command=status --ssh_user=root --orig_master_host=192.168.1.57 --orig_master_ip=192.168.1.57 --orig_master_port=3306
IN SCRIPT TEST====/sbin/ifconfig ens192:2 down==/sbin/ifconfig ens192:2 192.168.1.249/24===
Checking the Status of the script.. OK
ssh: Could not resolve hostname cluster1: Name or service not known
Sat Jan 11 17:03:43 2020 - [info] OK.
Sat Jan 11 17:03:43 2020 - [warning] shutdown_script is not defined.
Sat Jan 11 17:03:43 2020 - [info] Got exit code 0 (Not master dead).
MySQL Replication Health is OK.
6.3 确认mysql event_scheduler 是否关闭
如果没有关闭三台数据库都执行以下命令
mysql>show variables like 'event_scheduler'
mysql>set global event_scheduler=off;
6.4、从启动 mha
nohup /usr/bin/masterha_manager --conf=/etc/masterha/app1.cnf --ignore_last_failover < /dev/null > /etc/masterha/log/manager.log 2 >&1 &
6.5、检查状态
masterha_check_status --conf=/etc/masterha/app1.cnf
app1 (pid:20454) is running(0:PING_OK), master:192.168.1.57
七、检查vip 漂移
为了模拟MHA服务是否真正管用，这里测试MHAVIP漂移是否真正漂移，第一测试需要手动在数据库主节点(57) 操作，创建vip，执行sh init_vip.sh
测试步骤，停止主数据库节点(57)，查询vip 是否漂移到从数据库节点上(58)。
使用vip 登录数据库， select @@server_id; 查询是否和从数据库的server_id一致。
然后在启动主数据库节点(57)，这里从数据库节点为主(58)，刚启动旧主数据库节点(57)，在(57)执行以下命令，主从同步。
stop slave; change master to master_host='192.168.1.58',master_port=3306, master_user='slave', master_password='123456', MASTER_AUTO_POSITION=1; reset slave; start slave; show slave status\G;
7.1、192.168.1.57 操作
检查是否有vip ,如果没有执行 sh init_vip.sh
[root@node-57 masterha]#ifconfig
ens192:2: flags=4163<UP,BROADCAST,RUNNING,MULTICAST> mtu 1500
inet 192.168.1.249 netmask 255.255.255.0 broadcast 192.168.1.255
ether 00:0c:29:5b:ff:bb txqueuelen 1000 (Ethernet)
停止主mysql 在 192.168.1.57 操作
[root@node-57 masterha]#/etc/init.d/mysql stop
7.2 在192.168.1.58 操作
[root@node58 masterha]#ifconfig
ens192:2: flags=4163<UP,BROADCAST,RUNNING,MULTICAST> mtu 1500
inet 192.168.1.249 netmask 255.255.255.0 broadcast 192.168.1.255
ether 00:0c:29:5b:ff:bb txqueuelen 1000 (Ethernet)
[root@node58 masterha]# mysql -uroot -h192.168.1.249 -P31061 -p123456 -e "select @@server_id" mysql: [Warning] Using a password on the command line interface can be insecure. +-------------+ | @@server_id | +-------------+ | 10 | +-------------+
7.3 这里也可以使用手动恢复之前环境。
192.168.1.57 主数据库节点
192.168.1.58 主数据库从节点，手动添加节点和删除节点，我这里没有使用，大家可以尝试一下。
#####################################################################################################################################################
masterha_conf_host --command=delete --conf=/etc/masterha/app1.cnf --hostname=192.168.1.147 -block=1
masterha_conf_host --command=add --conf=/etc/masterha/app1.cnf --hostname=192.168.1.147 -block=1 --params="candidate_master=1;check_repl_delay=0;port=31061"
#####################################################################################################################################################
将slava 变为主， 将 主 变为slava ，恢复之前操作,为了方便下面操作。
两次 YES
[root@node59 masterha]# masterha_master_switch --conf=/etc/masterha/app1.cnf --master_state=alive --orig_master_is_new_slave --running_updates_limit=1000 Sat Jan 11 17:40:54 2020 - [info] MHA::MasterRotate version 0.58. Sat Jan 11 17:40:54 2020 - [info] Starting online master switch.. Sat Jan 11 17:40:54 2020 - [info] Sat Jan 11 17:40:54 2020 - [info] * Phase 1: Configuration Check Phase.. Sat Jan 11 17:40:54 2020 - [info] Sat Jan 11 17:40:54 2020 - [warning] Global configuration file /etc/masterha_default.cnf not found. Skipping. Sat Jan 11 17:40:54 2020 - [info] Reading application default configuration from /etc/masterha/app1.cnf.. Sat Jan 11 17:40:54 2020 - [info] Reading server configuration from /etc/masterha/app1.cnf.. Sat Jan 11 17:40:55 2020 - [info] GTID failover mode = 1 Sat Jan 11 17:40:55 2020 - [info] Current Alive Master: 192.168.1.57(192.168.1.57:3306) Sat Jan 11 17:40:55 2020 - [info] Alive Slaves: Sat Jan 11 17:40:55 2020 - [info] 192.168.1.58(192.168.1.58:3306) Version=5.7.26-log (oldest major version between slaves) log-bin:enabled Sat Jan 11 17:40:55 2020 - [info] GTID ON Sat Jan 11 17:40:55 2020 - [info] Replicating from 192.168.1.57(192.168.1.57:3306) Sat Jan 11 17:40:55 2020 - [info] Primary candidate for the new Master (candidate_master is set) It is better to execute FLUSH NO_WRITE_TO_BINLOG TABLES on the master before switching. Is it ok to execute on 192.168.1.57(192.168.1.57:3306)? (YES/no): YES Sat Jan 11 17:40:58 2020 - [info] Executing FLUSH NO_WRITE_TO_BINLOG TABLES. This may take long time.. Sat Jan 11 17:40:58 2020 - [info] ok. Sat Jan 11 17:40:58 2020 - [info] Checking MHA is not monitoring or doing failover.. Sat Jan 11 17:40:58 2020 - [info] Checking replication health on 192.168.1.58.. Sat Jan 11 17:40:58 2020 - [info] ok. Sat Jan 11 17:40:58 2020 - [info] Searching new master from slaves.. Sat Jan 11 17:40:58 2020 - [info] Candidate masters from the configuration file: Sat Jan 11 17:40:58 2020 - [info] 192.168.1.57(192.168.1.57:3306) Version=5.7.26-log log-bin:enabled Sat Jan 11 17:40:58 2020 - [info] GTID ON Sat Jan 11 17:40:58 2020 - [info] 192.168.1.58(192.168.1.58:3306) Version=5.7.26-log (oldest major version between slaves) log-bin:enabled Sat Jan 11 17:40:58 2020 - [info] GTID ON Sat Jan 11 17:40:58 2020 - [info] Replicating from 192.168.1.57(192.168.1.57:3306) Sat Jan 11 17:40:58 2020 - [info] Primary candidate for the new Master (candidate_master is set) Sat Jan 11 17:40:58 2020 - [info] Non-candidate masters: Sat Jan 11 17:40:58 2020 - [info] Searching from candidate_master slaves which have received the latest relay log events.. Sat Jan 11 17:40:58 2020 - [info] From: 192.168.1.57(192.168.1.57:3306) (current master) +--192.168.1.58(192.168.1.58:3306) To: 192.168.1.58(192.168.1.58:3306) (new master) +--192.168.1.57(192.168.1.57:3306) Starting master switch from 192.168.1.57(192.168.1.57:3306) to 192.168.1.58(192.168.1.58:3306)? (yes/NO): yes Sat Jan 11 17:41:00 2020 - [info] Checking whether 192.168.1.58(192.168.1.58:3306) is ok for the new master.. Sat Jan 11 17:41:00 2020 - [info] ok. Sat Jan 11 17:41:00 2020 - [info] 192.168.1.57(192.168.1.57:3306): SHOW SLAVE STATUS returned empty result. To check replication filtering rules, temporarily executing CHANGE MASTER to a dummy host. Sat Jan 11 17:41:00 2020 - [info] 192.168.1.57(192.168.1.57:3306): Resetting slave pointing to the dummy host. Sat Jan 11 17:41:00 2020 - [info] ** Phase 1: Configuration Check Phase completed. Sat Jan 11 17:41:00 2020 - [info] Sat Jan 11 17:41:00 2020 - [info] * Phase 2: Rejecting updates Phase.. Sat Jan 11 17:41:00 2020 - [info] master_ip_online_change_script is not defined. If you do not disable writes on the current master manually, applications keep writing on the current master. Is it ok to proceed? (yes/NO): yes Sat Jan 11 17:41:04 2020 - [info] Locking all tables on the orig master to reject updates from everybody (including root): Sat Jan 11 17:41:04 2020 - [info] Executing FLUSH TABLES WITH READ LOCK.. Sat Jan 11 17:41:04 2020 - [info] ok. Sat Jan 11 17:41:04 2020 - [info] Orig master binlog:pos is master-bin.000014:8073944. Sat Jan 11 17:41:04 2020 - [info] Waiting to execute all relay logs on 192.168.1.58(192.168.1.58:3306).. Sat Jan 11 17:41:04 2020 - [info] master_pos_wait(master-bin.000014:8073944) completed on 192.168.1.58(192.168.1.58:3306). Executed 0 events. Sat Jan 11 17:41:04 2020 - [info] done. Sat Jan 11 17:41:04 2020 - [info] Getting new master's binlog name and position.. Sat Jan 11 17:41:04 2020 - [info] slave-bin.000009:7714282 Sat Jan 11 17:41:04 2020 - [info] All other slaves should start replication from here. Statement should be: CHANGE MASTER TO MASTER_HOST='192.168.1.58', MASTER_PORT=3306, MASTER_AUTO_POSITION=1, MASTER_USER='slave', MASTER_PASSWORD='xxx'; Sat Jan 11 17:41:04 2020 - [info] Setting read_only=0 on 192.168.1.58(192.168.1.58:3306).. Sat Jan 11 17:41:04 2020 - [info] ok. Sat Jan 11 17:41:04 2020 - [info] Sat Jan 11 17:41:04 2020 - [info] * Switching slaves in parallel.. Sat Jan 11 17:41:04 2020 - [info] Sat Jan 11 17:41:04 2020 - [info] Unlocking all tables on the orig master: Sat Jan 11 17:41:04 2020 - [info] Executing UNLOCK TABLES.. Sat Jan 11 17:41:04 2020 - [info] ok. Sat Jan 11 17:41:04 2020 - [info] Starting orig master as a new slave.. Sat Jan 11 17:41:04 2020 - [info] Resetting slave 192.168.1.57(192.168.1.57:3306) and starting replication from the new master 192.168.1.58(192.168.1.58:3306).. Sat Jan 11 17:41:04 2020 - [info] Executed CHANGE MASTER. Sat Jan 11 17:41:04 2020 - [info] Slave started. Sat Jan 11 17:41:04 2020 - [info] All new slave servers switched successfully. Sat Jan 11 17:41:04 2020 - [info] Sat Jan 11 17:41:04 2020 - [info] * Phase 5: New master cleanup phase.. Sat Jan 11 17:41:04 2020 - [info] Sat Jan 11 17:41:04 2020 - [info] 192.168.1.58: Resetting slave info succeeded. Sat Jan 11 17:41:04 2020 - [info] Switching master to 192.168.1.58(192.168.1.58:3306) completed successfully.
7.4 恢复之前环境还有一种办法
192.168.1.57 主 ----->停止mysql服务
192.168.1.58 变为了主节点，vip 也漂移到58上。
启动 192.168.1.57 mysql 服务，执行主从同步 192.168.1.57 是从节点。192.168.1.58 是主节点。
stop slave; change master to master_host='192.168.1.58',master_port=3306, master_user='slave', master_password='123456', MASTER_AUTO_POSITION=1; reset slave; start slave; show slave status\G;
停止 现在的主数据库节点192.168.1.58， 然而 192.168.1.57 从节点会变为数据库的主节点，vip 会漂移到57 服务上
在启动192.168.1.58mysql 服务，执行主从同步，环境还原。
MHA常见问题:
https://www.cnblogs.com/pythonx/protected/p/12177588.html
问题:
slava 没有权限change
change master to master_auto_position=0;192.168.1.57
