(一)keepalived概述
Keepalived通过VRRP(虚拟路由冗余协议)协议实现虚拟IP的漂移。当master故障后，VIP会自动漂移到backup，这时通知下端主机刷新ARP表，如果业务是通过VIP连接到服务器的，则此时依然能够连接到正常运行的主机，RedHat给出的VRRP工作原理如下图：
本来对VIP漂移有一定了解的我，看了上面的图后，越来越懵了。因此只能根据我的个人理解，来对keepalived的VIP漂移做一个解释了，假设我现在有一套这样的环境：
主机A的IP地址为：192.168.10.11
主机B的IP地址为：192.168.10.12
我们再单独定义一个keepalived使用的VIP：192.168.10.10
当2台主机安装了keepalive并正常运行时，keepalive会选择一个节点做为主节点(这里假设为主机A，IP为192.168.10.11)，由于A是主节点，所以主机A上还会生成一个IP地址192.168.10.10，即虚拟IP(Virtual IP，也称VIP)，此时我们使用192.168.10.10访问主机，访问到的主机是A；假如A主机上的keepalived由于某些原因(例如服务器宕机、用户主动关闭…)关闭了，keepalived备用节点会检查与主节点keepalived的通信是否正常，检测到不正常，则会提升一个备节点为主节点，相应的虚拟IP也会在对应的主机上生成，从而实现高可用的目的。
(二)MySQL是如何结合keepalived实现高可用的
在MySQL中，通过搭建MySQL双主复制，保持2台主机上的MySQL数据库一模一样，并在2台主机上安装keepalived软件，启用VIP，用户应用程序通过VIP访问数据库。当包含VIP的主机上的数据库发生故障时，关闭keepalived，从而将VIP漂移到另一个节点，用户依然可以正常访问数据库。 (这里需要注意，虽然MySQL架构双主复制，2个节点都可以写入数据，但是我们在使用的时候，是通过VIP访问其中一个实例，并没有2个数据库实例一起使用)。这里我简单画了一个流程图，来说明keepalive与MySQL实现高可用的过程：
(三)keepalived+MySQL实现高可用过程实现
基础环境规划：
主机名
IP地址
备注
服务器A
hosta
192.168.10.11
keepalive主节点
服务器B
hostb
192.168.10.12
keepalive备节点
192.168.10.10
虚拟IP，会在keepalive启动后分配到上面2台机器的主节点上
(3.1)搭建MySQL双主复制环境
STEP2：配置双主复制参数
服务器A
服务器B
[mysqld]
basedir=/usr/local/mysql
datadir=/mysql/data
server_id = 1
binlog_format=ROW
log_bin=/mysql/binlog/master-bin
auto-increment-increment = 2 #字段变化增量值
auto-increment-offset = 1 #初始字段ID为1
slave-skip-errors = all #忽略所有复制产生的错误
gtid_mode=ON
enforce-gtid-consistency=ON
[mysqld]
basedir=/usr/local/mysql
datadir=/mysql/data
server_id = 2
binlog_format=ROW
log_bin=/mysql/binlog/master-bin
auto-increment-increment = 2 #字段变化增量值
auto-increment-offset = 2 #初始字段ID为2
slave-skip-errors = all #忽略所有复制产生的错误
gtid_mode=ON
enforce-gtid-consistency=ON
STEP3：创建复制用户，2个数据库上都要创建
grant replication slave on *.* to 'rep'@'%' identified by '123';
STEP4：将hosta的数据拷贝到hostb，并应用
[root@hostb ~]# mysqldump -uroot -p123456 -h 192.168.10.11 --single-transaction --all-databases --master-data=2 > hosta.sql
[root@hostb ~]# mysql -uroot -p123456 < hosta.sql
STEP5：hostb上开启复制，以下脚本在hostb上执行
--配置复制
mysql> CHANGE MASTER TO
-> master_host='192.168.10.11',-> master_port=3306,-> master_user='rep',-> master_password='123',-> MASTER_AUTO_POSITION = 1;
Query OK,0 rows affected, 2 warnings (0.01sec)--开启复制
mysql>start slave;
Query OK,0 rows affected (0.00sec)--查看复制状态
mysql>show slave status \G*************************** 1. row ***************************Slave_IO_State: Waitingfor master tosend event
Master_Host:192.168.10.11Master_User: rep
Master_Port:3306Connect_Retry:60Master_Log_File: master-bin.000001Read_Master_Log_Pos:322Relay_Log_File: hostb-relay-bin.000002Relay_Log_Pos:417Relay_Master_Log_File: master-bin.000001Slave_IO_Running: Yes
Slave_SQL_Running: Yes
STEP6：hosta上开启复制，以下脚本在hosta上执行
mysql> CHANGE MASTER TO
-> master_host='192.168.10.12',-> master_port=3306,-> master_user='rep',-> master_password='123',-> MASTER_AUTO_POSITION = 1;
Query OK,0 rows affected, 2 warnings (0.01sec)
mysql>start slave;
Query OK,0 rows affected (0.01sec)
mysql>show slave status \G;*************************** 1. row ***************************Slave_IO_State: Waitingfor master tosend event
Master_Host:192.168.10.12Master_User: rep
Master_Port:3306Connect_Retry:60Master_Log_File: master-bin.000001Read_Master_Log_Pos:154Relay_Log_File: hosta-relay-bin.000002Relay_Log_Pos:369Relay_Master_Log_File: master-bin.000001Slave_IO_Running: Yes
Slave_SQL_Running: Yes
STEP7：测试双主复制
在hosta上创建数据库testdb，到hostb服务器上查看数据库是否已经创建
--hosta上创建数据库
create databasetestdb；--hostb上查看数据库，发现已经创建
mysql>show databases;+--------------------+
| Database |
+--------------------+
| information_schema |
| db1 |
| lijiamandb |
| mysql |
| performance_schema |
| sbtest |
| sys |
| testdb |
+--------------------+
8 rows in set (0.01 sec)
在hostb的testdb数据库上创建表t1,并插入数据，到hosta上查看是否复制过来
--在hostb上创建表并插入数据
mysql> usetestdbDatabasechanged
mysql> create table t1(id int,name varchar(20));
Query OK,0 rows affected (0.01sec)
mysql> insert into t1 values(1,'a');
Query OK,1 row affected (0.01sec)--在hosta上查看数据，数据已经过来
mysql> select * fromtestdb.t1;+------+------+
| id | name |
+------+------+
| 1 | a |
+------+------+
1 row in set (0.00 sec)
到这，双主复制已经搭建完成，接下来安装配置keepalived。
(3.2)安装配置keepalived
(3.2.1)keepalived的安装与管理
keepalived可以使用源码安装，也可以使用yum在线安装，这里直接使用yum在线安装:
[root@hosta data]# yum install -y keepalived
使用如下命令查看安装路径:
[root@hosta data]# rpm -ql keepalived/etc/keepalived/etc/keepalived/keepalived.conf/etc/sysconfig/keepalived/usr/bin/genhash/usr/lib/systemd/system/keepalived.service/usr/libexec/keepalived/usr/sbin/keepalived/usr/share/doc/keepalived-1.3.5
… 略
使用如下命令管理keepalived
# 开启keepalived
systemctl start keepalived 或者 service keepalived start
# 关闭keepalived
systemctl stop keepalived 或者 service keepalived stop
# 查看keepalived运行状态
systemctl status keepalived 或者 service keepalived status
# 重新启动keepalived
systemctl restart keepalived 或者 service keepalived restart
(3.2.2)keepalived的配置
keepalived的配置文件为：/etc/keepalived/keepalived.conf，我的配置文件如下：
【hosta主机的配置文件】
[root@hosta keepalived]# catkeepalived.conf! Configuration File forkeepalived
global_defs {
notification_email {
ops@wangshibo.cn
tech@wangshibo.cn
}
notification_email_from ops@wangshibo.cn
smtp_server127.0.0.1smtp_connect_timeout30router_id MASTER-HA
}
vrrp_script chk_mysql_port { #检测mysql服务是否在运行。有很多方式，比如进程，用脚本检测等等
script"/mysql/chk_mysql.sh"#这里通过脚本监测
interval2#脚本执行间隔，每2s检测一次
weight –5
#脚本结果导致的优先级变更，检测失败(脚本返回非0)则优先级 -5
fall2
#检测连续2次失败才算确定是真失败。会用weight减少优先级(1-255之间)
rise1#检测1次成功就算成功。但不修改优先级
}
vrrp_instance VI_1 {
state BACKUP #这里所有节点都定义为BACKUP
interface ens34 #指定虚拟ip的网卡接口
mcast_src_ip192.168.10.11 #本地IPvirtual_router_id51#路由器标识，MASTER和BACKUP必须是一致的
priority101#定义优先级，数字越大，优先级越高，在同一个vrrp_instance下，MASTER的优先级必须大于BACKUP的优先级。
advert_int1nopreempt #不抢占模式，在优先级高的机器上设置即可，优先级低的机器可不设置
authentication {
auth_type PASS
auth_pass1111}
virtual_ipaddress {192.168.10.10 #虚拟IP}
track_script {
chk_mysql_port
}
}
【hostb主机的配置文件】
[root@hostb keepalived]# catkeepalived.conf! Configuration File forkeepalived
global_defs {
notification_email {
ops@wangshibo.cn
tech@wangshibo.cn
}
notification_email_from ops@wangshibo.cn
smtp_server127.0.0.1smtp_connect_timeout30router_id MASTER-HA
}
vrrp_script chk_mysql_port {
script"/mysql/chk_mysql.sh"interval2weight-5fall2rise1}
vrrp_instance VI_1 {
state BACKUP
interface ens34
mcast_src_ip192.168.10.12virtual_router_id51priority99advert_int1authentication {
auth_type PASS
auth_pass1111}
virtual_ipaddress {192.168.10.10}
track_script {
chk_mysql_port
}
}
需要特别注意：nopreempt这个参数只能用于state为BACKUP的情况，所以在配置的时候要把master和backup的state都设置成BACKUP，这样才会实现keepalived的非抢占模式！
【判断MySQL数据库允许状态的文件】
[root@hosta ~]# cat /mysql/chk_mysql.sh#!/bin/bash
counter=$(netstat -na|grep "LISTEN"|grep "3306"|wc -l)if [ "${counter}" -eq 0 ]; thensystemctl stop keepalivedfi
在配置完成之后，启动MySQL数据库和keepalive，需要注意，先启动MySQL，再启动keepalive，因为keepalive启动后会检测MySQL的运行状态，如果MySQL运行异常，keepalive会自动关闭。
(3.3)高可用测试
时间轴
(时间递增)
hosta操作
hostb操作
时间1
# 数据库运行正常
[root@hosta ~]# service mysqld status
SUCCESS! MySQL running (8530)
#keepalived运行正常
[root@hosta ~]# service keepalived status
Active: active (running) since Mon 2020-08-03 22:29:10 CST; 1min 32s ago
# 此时虚拟IP在hosta上
[root@hosta ~]# ip addr
3: ens34: mtu 1500 qdisc pfifo_fast state UP qlen 1000
link/ether 00:0c:29:fb:62:0a brd ff:ff:ff:ff:ff:ff
inet 192.168.10.11/24 brd 192.168.10.255 scope global ens34
valid_lft forever preferred_lft forever
inet 192.168.10.10/32 scope global ens34
valid_lft forever preferred_lft forever
inet6 fe80::20c:29ff:fefb:620a/64 scope link
valid_lft forever preferred_lft forever
# 数据库运行正常
[root@hostb ~]# service mysqld status
SUCCESS! MySQL running (8022)
#keepalived运行正常
[root@hostb ~]# service keepalived status
Active: active (running) since Mon 2020-08-03 22:31:09 CST; 8s ago
[root@hostb ~]# ip addr
3: ens34: mtu 1500 qdisc pfifo_fast state UP qlen 1000
link/ether 00:0c:29:92:3d:5a brd ff:ff:ff:ff:ff:ff
inet 192.168.10.12/24 brd 192.168.10.255 scope global ens34
valid_lft forever preferred_lft forever
inet6 fe80::20c:29ff:fe92:3d5a/64 scope link
valid_lft forever preferred_lft forever
时间2
# 通过VIP访问数据库，访问到的都是hosta上的实例
[root@hosta keepalived]# mysql -uroot -p123456 -h192.168.10.10 -e "select @@hostname"
mysql: [Warning] Using a password on the command line interface can be insecure.
+------------+
| @@hostname |
+------------+
| hosta |
+------------+
# 通过VIP访问数据库，访问到的都是hosta上的实例
[root@hostb ~]# mysql -uroot -p123456 -h192.168.10.10 -e "select @@hostname"
mysql: [Warning] Using a password on the command line interface can be insecure.
+------------+
| @@hostname |
+------------+
| hosta |
+------------+
时间3
# 主节点关闭MySQL
[root@hosta keepalived]# service mysqld stop
Shutting down MySQL............ SUCCESS!
时间4
# keepalived检测到MySQL关闭后，会自动关闭
[root@hosta keepalived]# service keepalived status
Active: inactive (dead)
# 此时VIP已经不存在
[root@hosta keepalived]# ip addr
3: ens34: mtu 1500 qdisc pfifo_fast state UP qlen 1000
link/ether 00:0c:29:fb:62:0a brd ff:ff:ff:ff:ff:ff
inet 192.168.10.11/24 brd 192.168.10.255 scope global ens34
valid_lft forever preferred_lft forever
inet6 fe80::20c:29ff:fefb:620a/64 scope link
valid_lft forever preferred_lft forever
# hostb上的keepalived运行正常
[root@hostb ~]# service keepalived status
Active: active (running) since Mon 2020-08-03 22:31:09 CST; 6min ago
# 发现VIP已经切换到了hostb
[root@hostb ~]# ip addr
3: ens34: mtu 1500 qdisc pfifo_fast state UP qlen 1000
link/ether 00:0c:29:92:3d:5a brd ff:ff:ff:ff:ff:ff
inet 192.168.10.12/24 brd 192.168.10.255 scope global ens34
valid_lft forever preferred_lft forever
inet 192.168.10.10/32 scope global ens34
valid_lft forever preferred_lft forever
inet6 fe80::20c:29ff:fe92:3d5a/64 scope link
valid_lft forever preferred_lft forever
时间5
# 通过VIP访问数据库，访问到的都是hosta上的实例
[root@hosta keepalived]# mysql -uroot -p123456 -h192.168.10.10 -e "select @@hostname"
mysql: [Warning] Using a password on the command line interface can be insecure.
+------------+
| @@hostname |
+------------+
| hostb |
+------------+
# 通过VIP访问数据库，访问到的都是hosta上的实例
[root@hostb ~]# mysql -uroot -p123456 -h192.168.10.10 -e "select @@hostname"
mysql: [Warning] Using a password on the command line interface can be insecure.
+------------+
| @@hostname |
+------------+
| hostb |
+------------+
时间6
# 重启MySQL
[root@hosta keepalived]# service mysqld start
Starting MySQL. SUCCESS!
# 重启keepalived
[root@hosta keepalived]# service keepalived start
Redirecting to /bin/systemctl start keepalived.service
[root@hosta keepalived]#
[root@hosta keepalived]# service keepalived status
Redirecting to /bin/systemctl status keepalived.service
Active: active (running) since Mon 2020-08-03 22:40:08 CST; 6s ago
# 因为使用了非抢占模式，VIP不会漂回来
# 如果要把VIP漂回来，可以关闭hostb上的数据库或者直接关闭hostb服务器
[root@hosta keepalived]# ip addr
3: ens34: mtu 1500 qdisc pfifo_fast state UP qlen 1000
link/ether 00:0c:29:fb:62:0a brd ff:ff:ff:ff:ff:ff
inet 192.168.10.11/24 brd 192.168.10.255 scope global ens34
valid_lft forever preferred_lft forever
inet6 fe80::20c:29ff:fefb:620a/64 scope link
valid_lft forever preferred_lft forever
【完】