目前需要做一个项目需要高可用? ? 项目采用Keepalived+VIP（虚拟IP）进行一个故障转移。Keepalived和VIP的相关知识就不在这里废话介绍了? 介绍一下当前的高可用环境实现思路。
目前有三台服务器。分别为三台服务器上部署keepalived 进行故障转移，当用户访问虚拟IP时，其keepalived会判断其优先级，主要根据监控的相关的服务端口? 如
nginx 监听的端口 80
数据库服务端口? 3306?
php-fpm的9000端口? ?
将优先级的进行虚拟IP（VIP）的地址漂移。使用ip a 可以查看到其绑定的虚拟IP在哪一台服务器上，或通过相关日志进行查看。
服务器相关IP配置
| 资源 | IP? |
| 服务器1 | 192.168.1.101 |
| 服务器1 | 192.168.1.102 |
| 服务器1 | 192.168.1.103 |
步骤一? ? 先从部署或交付那边拿到配置好的虚拟IP? ? 如：192.168.1.86? ?
步骤二? ?启动关键的端口服务? 如? ngngx php-fpm? mysql? ? 使用lsof -i:80? ?命令进行查看
步骤三? 修改或新增一个keepalived的配置? ?如? keepalived.conf
下面贴出我的配置文件? ?
! Configuration File for keepalived
global_defs {
notification_email {
yuluolishang@qq.com
}
notification_email_from yuluolishang@yisa.com
smtp_server 127.0.0.1
smtp_connect_timeout 30
router_id LVS_DEVEL
script_user root
enable_script_security
vrrp_skip_check_adv_addr
#vrrp_strict
#vrrp_iptables
vrrp_garp_interval 0
vrrp_gna_interval 0
}
vrrp_script chk_mysql_port {
script "</dev/tcp/127.0.0.1/3306" # connects and exits
interval 1 # check every second
weight -20 # default priority: -20 if connect fails
}
vrrp_script chk_php_port {
script "</dev/tcp/127.0.0.1/9000" # connects and exits
interval 1 # check every second
weight -20 # default priority: -20 if connect fails
}
vrrp_script chk_nginx_port {
script "</dev/tcp/127.0.0.1/80" # connects and exits
interval 1 # check every second
weight -20 # default priority: -20 if connect fails
}
vrrp_instance VI_1 {
state BACKUP
interface eth0
virtual_router_id 66
priority 100
advert_int 1
authentication {
auth_type PASS
auth_pass Feipeng$
}
unicast_src_ip 192.168.1.101
unicast_peer {
192.168.1.102
192.168.1.103
}
virtual_ipaddress {
192.168.1.86/22
}
track_script {
chk_mysql_port
chk_php_port
chk_nginx_port
}
}配置相关说明? ：
state BACKUP? ?三台服务器都被BACKUP 角色
interface? ?etho? ?这个为该服务器的网卡名称
virtual_router_id? ?虚拟路由? 这个三台服务器为同一个? 必须同一个
priority? ?100? ?优先级? 三台服务器需一次填写100 99 98?
authentication? 认证密码? 三台服务器密码必须设置相同? 否则故障转移不成功。
unicast_src_ip? ? 广播地址? ?即本机IP地址??
unicast_peer? 广播接收地址? 即另外两台服务器的IP地址
virtual_ipaddress? ?虚拟IP(VIP)地址? ?后面的/22 代表子网掩码? 一般与三台服务器的真实IP在同一网段中。
track_script? ?里面为各监控的端口脚本名称? ?一行写一个检测。
步骤四? ?启动keepalived 进程? ?启动故障转移? ? 本keepalived 为离线部署? 离线包目录结构如下：
启动命令? ?/keepalived/sbin/keepalived -f /keepalived/etc/keepalived/keepalived.conf??
/keepalived/sbin/keepalived -f /keepalived/etc/keepalived/keepalived.conf??步骤五? ?验证启动? ??
查看keepalived 进程是否有启动? ?输入 ps aux|grep keepalived
ps aux|grep keepalived会输出如下：
ps aux|grep keepalived
root 182037 0.0 0.0 44096 888 ? Ss 17:13 0:00 /keepalived/sbin/keepalived -f /keepalive/etc/keepalived/keepalived.conf
root 182038 0.1 0.0 44096 1308 ? S 17:13 0:00 /keepalived/sbin/keepalived -f /keepalive/etc/keepalived/keepalived.conf
root 182343 0.0 0.0 112812 968 pts/0 S+ 17:13 0:00 grep --color=auto keepalive
? ? 输入ip a?查看当前漂移VIP 绑定情况
步骤六? ?模拟故障? ?比如说nginx 挂掉了? ?
输入命令? ?pkill nginx
然后输入lsof -i:80? ?此时可以得出? nginx 服务被停掉了??
pkill nginx
lsof -i:80此时? 刚才绑定的VIP的所在服务器上? 此VIP不在其服务器了，而是跑到另一台服务器上绑定了VIP虚拟IP?
至此? ?keepalived 故障转移部署成功！！
坑：
1.同一组故障转移集群? ?virtual_router_id? 虚拟路由IP 必须一样? 要不然部署不成功。
2.interface? ?etho? 网卡名称一定要填写当前服务器IP的所在网卡名称
3.若同一个局域网内，有多个keepalived 实例，keepalived两台机器同时出现vip问题！keepalived默认为广播，应改为单播