MHA简单部署

MHA是目前比较成熟的mysql高可用集群方式之一.

一、参考文档：
1.官方文档：【 https://github.com/yoshinorim/mha4mysql-manager/wiki 】
2.个人博客： 好多
3.下载地址：
（1）manager:【 https://github.com/yoshinorim/mha4mysql-manager/releases 】
（2）node: 【 https://github.com/yoshinorim/mha4mysql-node/releases/tag/v0.58 】

二、基础环境
1.操作系统：Centos 7.4
2.mysql版本：percona-mysql 5.7.22
3.MHA版本：0.58
4.服务器信息

三、部署流程
1.安装node
# master,slave-01
(1)下载安装包

mkdir -p /opt/upload/mha_packges
cd opt/upload/mha_packges
wget 'https://github.com/yoshinorim/mha4mysql-node/releases/download/v0.58/mha4mysql-node-0.58-0.el7.centos.noarch.rpm'

(2)安装node

yum -y install perl-DBD-MySQL -y
yum -y localinstall mha4mysql-node-0.58-.el7.centos.noarch.rpm

2.安装manager
#172.18.6.14
(1)下载安装包

mkdir -p /opt/upload/mha_packges
cd /opt/upload/mha_packges
wget 'https://github.com/yoshinorim/mha4mysql-node/releases/download/v0.58/mha4mysql-node-0.58-0.el7.centos.noarch.rpm'
wget 'https://github.com/yoshinorim/mha4mysql-manager/releases/download/v0.58/mha4mysql-manager-0.58-0.el7.centos.noarch.rpm'

(2)安装manager及node

yum -y install perl-DBD-MySQL perl-Config-Tiny perl-Log-Dispatch perl-Parallel-ForkManager
yum -y localinstall mha4mysql-node-0.58-.el7.centos.noarch.rpm
yum -y localinstall mha4mysql-manager-0.58-.el7.centos.noarch.rpm

3.添加hosts
#master-01,slave-01,slave-02操作

cat <<EOF>> /etc/hosts
172.18.6.15 mysql_master-
172.18.6.13 mysql_slave-
172.18.6.14 mysql_slave-
EOF

4.配置免密登陆
(1)配置三台服务器免密登陆
(2)测试命令：masterha_check_ssh --conf=/opt/ mha/conf/default.cnf
5.创建manager工作目录
mkdir -p /opt/mha/{manager_data,remote_data,manager_logs,conf}

6.生成manager配置文件

cat <<EOF> /opt/mha/conf/default.cnf
[server default]
# mha_directory #
manager_workdir=/opt/mha/manager_data
manager_log=/opt/mha/manager_logs/manager.log
remote_workdir=/opt/mha/remote_data

# mha_setting #
ping_interval=
shutdown_script=""
master_ip_failover_script="/usr/bin/master_ip_failover"
master_ip_online_change_script="/usr/bin/master_ip_online_change"

# DATABASE #
user=mha
password=dingkai123 #最好不要用特殊字符,mha难以识别

# SSH #
ssh_user=root
ssh_port=
repl_user=repl #mysql主从复制账号
repl_password=dingkai. #mysql主从复制账号密码
master_binlog_dir=/opt/mysql/mysql_log #binlog目录

[server1]
hostname= mysql_master-
port=
master_binlog_dir=/opt/mysql/mysql_log
candidate_master=
check_repl_delay=

[server2]
hostname=mysql_slave-
port=
master_binlog_dir=/opt/mysql/mysql_log
candidate_master=
check_repl_delay=

[server3]
hostname=mysql_slave-
port=
master_binlog_dir=/opt/mysql/mysql_log
no_master=

[binlog1]
hostname=mysql_master-
EOF

5.在master创建mha管理账号
grant all privileges on *.* to mha@172.18.6.% identified by 'MFdb@2018!';
6.修改‘master_ip_failover’ ，'master_ip_online_change'脚本（这两个脚本在manager源码中的 mha4mysql-manager-0.58/samples/scripts 下）
(1)修改master_ip_failover
添加

my $vip = '172.18.6.251/32';
my $key = '0';
my $ssh_port = '10022';
my $ssh_start_vip = "/sbin/ifconfig eth0:$key $vip";
my $ssh_stop_vip = "/sbin/ifconfig eth0:$key down";

(2)修改master_ip_online_change
#添加以下内容

my $vip = '172.18.6.251/32';
my $ssh_port = '10022';
my $if = "eth0";

(3)添加脚本到 /usr/bin/下

cp -a master_ip_failover master_ip_online_change /usr/bin/
chmod +x /usr/bin/master_ip_failover
chmod +x /usr/bin/master_ip_online_change

6.配置虚拟网卡
#master实例操作
ifconfig eth0:0 172.18.6.201/32

7.检查复制情况
#slave-02执行
masterha_check_repl --conf=/opt/mha/conf/default.cnf #检查复制状态
masterha_manager --conf=/opt/mha/conf/default.cnf #启动mha
masterha_check_status --conf=/opt/mha/conf/default.cnf #查看mha运行状态

.使用脚本控制
()创建mha控制脚本
cat <<EOF> /usr/bin/mha.sh
#!/usr/bin/bash
#filename: mha.sh
#用途：启动关闭MHA

#VARIABLES
config_file="/opt/mha/conf/default.cnf "
base_path="/opt/mha "
log_path="/opt/mha/manager_logs"
masterha_manager="/usr/bin/masterha_manager"
masterha_stop="/usr/bin/masterha_stop"
masterha_check_status="/usr/bin/masterha_check_status"
mha_status=$(masterha_check_status --conf=/opt/mha/conf/default.cnf |grep -w 'pid'|awk -F" " '{print $2}'|awk -F":" '{print $2}'|awk -F")" '{print $1}')

#functions
start(){
    if [ -z ${mha_status} ]
    then
      /usr/bin/nohup ${masterha_manager} --conf=${config_file} > ${log_path}/nohup.log >& &
    else
      echo "MHA is Running!"
    fi
}

stop(){
    if [ ! -z ${mha_status} ]
    then
      ${masterha_stop}  --conf=${config_file}
    else
      echo "MHA is not running!"
    fi
}

(2)脚本命令
mha.sh start #启动mha
mha.sh stop #停止mha
mha.sh status #查看mha状态

注意:当master宕机后,mha会自动将一个slave节点提升为主,随后mha自动停止运行,在手动恢复数据库之前,mha都是不可用的