数据中心常常通过 Keepalived 来实现 MySQL 的 cluster。但在公有云上，由于不支持组播和广播，Keepalived 的 cluster IP 并不能被其他主机识别，传统配置方法并不适用。

Google Cloud Computer Engine 具有 Alias IP 功能，该功能可为 VM 增加第二个 IP 地址，其他 VM 可通过这个 IP 地址连接 GCE。

借此，可将 Alias IP 作为 Cluster IP，配合 Keepalived，进而实现高可用的 Cluster：

通过共有云 VPC 系统路由，向其他 VM 宣告该 alias IP 的存在
通过 Keepalived 的服务检测机制和选举机制，实现浮动 IP 对应 Alias IP，同时利用脚本控制 alias IP 所关联的 VM
两台MySQL服务器配置为HA模式，本文不做赘述。

一 Alias IP 功能

如下图所示：

可通过Google Cloud SDK 设置 VM alias IP。

二实现方法

1 Google Cloud VM 创建

创建两台 CentOS7 VM：

鉴于需要控制 alias IP，所以需要 Compuyte Engine 的权限：

其他部分为常规配置。

2 安装 MySQL

yum install -y wget
wget http://repo.mysql.com/mysql-community-release-el7-5.noarch.rpm
yum install mysql-community-release-el7-5.noarch.rpm -y
yum makecache
yum install mysql-community-server -y

systemctl enable mysql
systemctl start mysqld

关闭 Selinux，selinux 会造成 keepalived 脚本运行 gcloud 命令失败。

vim /etc/sysconfig/selinux

SELINUX=disabled

MySQL中创建密码和授权：

GRANT ALL PRIVILEGES ON *.* TO 'root'@'%'      IDENTIFIED BY 'redhat'      WITH GRANT OPTION;
FLUSH PRIVILEGES;

3 安装 keepalived

yum install keepalived -y
cd /etc/keepalived

Cluster-VM1 的配置

vim keepalived.conf

! Configuration File for keepalived

global_defs {
   router_id mysql-1
}

vrrp_script chk_appsvc {
#检查服务状态脚本
    script /etc/keepalived/bin/keepalived-check-appsvc.sh           
    interval 2                              #检查周期
    fall 2
    rise 2
}

vrrp_instance VI_1 {
    state MASTER                          #初始状态
    interface eth0
    virtual_router_id 51
    priority 100                          #优先级
      #keepalived多台机器间通讯采用单播方式，这个是本机地址
      unicast_src_ip  10.1.1.11        
      #keepalived多台机器间通讯采用单播方式，这个是对端地址
      unicast_peer {
        10.1.1.12                         
      }
    track_script {                        #检查本节点的状态
        chk_appsvc
    }

    advert_int 1
    nopreempt
   authentication {                        #keepalived间的认证
        auth_type PASS
        auth_pass 1111
    }
    #Keepalived的浮动IP
    virtual_ipaddress {
        10.1.1.200                                                           
    }
    #Keepalived选举后的脚本
    notify /etc/keepalived/bin/keepalived-action.sh      
    }

Cluster-VM2的配置与VM1类似，只需要修改优先级、本机和对端IP地址。

4 Keepalived 脚本

在 Cluster-VM1 上创建 keepalived 相关脚本：

mkdir bin
cd bin

#在本机keepalived成为Master时，给本机添加Alias IP
touch add_alias.sh 
 
#notification script,根据keepalived的状态执行不同的脚本   
touch keepalived-action.sh                   
#检查本机服务状态      
touch keepalived-check-appsvc.sh      
#在本机keepalived成为Backup或其他状态时，删除Alias IP      
touch rm_alias.sh                              
#log文件，显示状态
touch /tmp/keepalived.log           
      
chmod +x *
chmod 777 /tmp/keepalived.log

Tracking Script:
vim keepalived-check-appsvc.sh

#!/bin/bash
MYSQL_PWD="redhat"                   #MySQL登录信息
MYSQL_USERNAME="root"
MYSQL_HOST="10.1.1.11"
      。
#检查MySQL的databases，如果能够查询到performance_schema，
#说明MySQL工作正常
mysql_status=`mysql -h$MYSQL_HOST -u$MYSQL_USERNAME -p$MYSQL_PWD -s -e "show databases;" 2>&1 | grep performance_schema | wc -l`
      
#如果MySQL正常，正常退出，否则exit 1
if [ "$mysql_status" == "1" ]; then           
        exit 0
else
  exit 1
fi

Notification Script:
vim keepalived-action.sh

#!/bin/bash

TYPE=$1
NAME=$2
STATE=$3

#如果Keepalived状态成为Master，运行add_alias脚本
case $STATE in
"MASTER")                                            /etc/keepalived/bin/add_alias.sh        
    exit 0
    ;;

#如果Keepalived状态成为其他，运行rm_alias脚本
"BACKUP"|"STOP")                            /etc/keepalived/bin/rm_alias.sh $STATE
    exit 0
    ;;
"FAULT")
    /etc/keepalived/bin/rm_alias.sh $STATE
    exit 0
    ;;
*)
    /etc/keepalived/bin/rm_alias.sh $STATE
    exit 1
    ;;
esac

添加 Alias IP 脚本：
vim add_alias.sh

#!/bin/bash
time=$(TZ=Asia/Shanghai date "+%Y-%m-%d %H:%M:%S")

#记录keepalived状态
echo $time "vm1 become master" >> /tmp/keepalived.log     

#删除Cluster-VM2的Alias IP
/usr/bin/gcloud compute instances network-interfaces update cluster-vm2 --zone asia-east1-b --aliass ""

#添加Alias IP给Cluster-VM1
/usr/bin/gcloud compute instances network-interfaces update cluster-vm1 --zone asia-east1-b --aliass "10.1.1.200/32"

#检测是否成功
adda=`/usr/bin/gcloud compute instances describe cluster-vm1 --zone asia-east1-b | grep ipCidrRange | wc -l`
rma=`/usr/bin/gcloud compute instances describe cluster-vm2 --zone asia-east1-b | grep ipCidrRange | wc -l`

#添加Log
if [ "$adda" == "1" ]; then
  echo $time "add alias IP to vm1 success" >> /tmp/keepalived.log
else
  echo $time "add alias IP to vm1 fail" >> /tmp/keepalived.log
fi

if [ "$rma" == "0" ]; then
  echo $time "remove alias IP from vm2 success" >> /tmp/keepalived.log
else
  echo $time "remove alias IP from vm2 fail" >> /tmp/keepalived.log
fi

删除 Alias IP 脚本
vim rm_alias.sh

#!/bin/bash
time=$(TZ=Asia/Shanghai date "+%Y-%m-%d %H:%M:%S")

#记录keepalived状态
echo $time "vm1 become" $1 >> /tmp/keepalived.log 
   
#删除本机Alias IP
/usr/bin/gcloud compute instances network-interfaces update cluster-vm1 --zone asia-east1-b --aliass ""

#检查状态
adda=`/usr/bin/gcloud compute instances describe cluster-vm2 --zone asia-east1-b | grep ipCidrRange | wc -l`
rma=`/usr/bin/gcloud compute instances describe cluster-vm1 --zone asia-east1-b | grep ipCidrRange | wc -l`

#添加Log
if [ "$adda" == "1" ]; then
  echo $time "add alias IP to vm2 success" >> /tmp/keepalived.log
else
  echo $time "add alias IP to vm2 fail" >> /tmp/keepalived.log
fi

if [ "$rma" == "0" ]; then
  echo $time "remove alias IP from vm1 success" >> /tmp/keepalived.log
else
  echo $time "remove alias IP from vm1 fail" >> /tmp/keepalived.log
fi

Cluster-VM2 脚本与 Cluster-VM1 类似，只需将 IP 和 VM 名称交换。

5 启动 Keepalived

systemctl enable keepalived
systemctl start keepalived

启动 keepalived 后，通过以下命令查看状态：

ip a                          #查看Keepalived的floating IP是否添加成功
ip route list table local     #查看Alias IP是否添加成功
cat /tmp/keepalived.log       #查看keepalived的日志状态

6 检查

通过 ping 以及 mysql 客户端连接 alias IP：

[root@testcentos7 bin]# ping 10.1.1.200
PING 10.1.1.200 (10.1.1.200) 56(84) bytes of data.
64 bytes from 10.1.1.200: icmp_seq=1 ttl=64 time=1.72 ms
64 bytes from 10.1.1.200: icmp_seq=2 ttl=64 time=0.326 ms
^[c64 bytes from 10.1.1.200: icmp_seq=3 ttl=64 time=0.308 ms
^C
--- 10.1.1.200 ping statistics ---
3 packets transmitted, 3 received, 0% packet loss, time 2009ms
rtt min/avg/max/mdev = 0.308/0.786/1.725/0.664 ms

[root@testcentos7 bin]# mysql -h10.1.1.200 -uroot -predhat
Warning: Using a password on the command line interface can be insecure.
Welcome to the MySQL monitor.  Commands end with ; or \g.

Type 'help;' or '\h' for help. Type '\c' to clear the current input statement.
mysql> show databases;
+--------------------+
| Database           |
+--------------------+
| information_schema |
| mysql              |
| performance_schema |
+--------------------+
3 rows in set (0.00 sec)

mysql> show variables where Variable_name like '%hostname%';
+---------------+-------------+
| Variable_name | Value       |
+---------------+-------------+
| hostname      | cluster-vm2 |
+---------------+-------------+
1 row in set (0.00 sec)

在 VM2 上关闭 mysqld:

systemctl stop mysqld

会看到 VM2 状态转为 Fail：

在 VM1 上，keepalived 转为 Master 状态：

同时在客户端重新连接 MySQL 服务器，并检测：

可以看到，已经切换至 VM1 上了。

三总结

Keepalived 和Google Cloud Alias IP 配合，可以非常便捷地实现主备 Cluster。

文章版权归作者所有，未经许可不得转载

创作场景

GCE+ 辅助 IP，实现基于 Keepalived 和 MySQL 的高可用集群