IT技术互动交流平台

keepalived高可用lvs

作者:佚名  来源:IT165收集  发布日期:2016-11-14 20:23:35

LVS只是做一个负载均衡,通过访问VIP来访问后端的网站程序,一旦LVS宕机,整个网站就访问不了,这就出现了单点。所以要结合keepalive这种高可用软件来保证整个网站的高可用。本文将介绍如何利用keepalive来实现LVS的高可用(以LVS的DR模式为例,生产环境后台的real server 网站内容是一致的,为了看到实验效果,这里是两个不同的页面)。
实验拓扑:

1.在两台Real Server上配置web服务

2.配置两台Real Server
配置脚本如下:
#!/bin/bash
#
 
VIP='10.1.88.88'
NETMASK='255.255.255.255'
IFACE='lo:0'
 
case $1 in
start)
    echo 1 > /proc/sys/net/ipv4/conf/all/arp_ignore
    echo 1 > /proc/sys/net/ipv4/conf/lo/arp_ignore
    echo 2 > /proc/sys/net/ipv4/conf/all/arp_announce
    echo 2 > /proc/sys/net/ipv4/conf/lo/arp_announce
 
    ifconfig $IFACE $VIP broadcast $VIP netmask $NETMASK up
    route add -host $VIP $IFACE
    ;;
stop)
    echo 0 > /proc/sys/net/ipv4/conf/all/arp_ignore
    echo 0 > /proc/sys/net/ipv4/conf/lo/arp_ignore
    echo 0 > /proc/sys/net/ipv4/conf/all/arp_announce
    echo 0 > /proc/sys/net/ipv4/conf/lo/arp_announce
 
    ifconfig $IFACE down
    ;;
*)
    echo "Usage: $(basename $0) {start|stop}"
    exit 1
esac
3.配置keepalived
keepalived master的配置:
! Configuration File for keepalived
 
global_defs {
   notification_email {
       root@localhost
   }
   notification_email_from keepalived@localhost
   smtp_server 127.0.0.1
   smtp_connect_timeout 30
   router_id node1
}
 
vrrp_script chk_down {                #配置脚本
    script "[[ -f /etc/keepalived/down ]] && exit 1 || exit 0"  #脚本内容
    interval 2     #间隔两秒执行一次脚本
    weight -6      #如果脚本执行失败则优先级减6
}
 
vrrp_instance VI_1 {               #配置实例    
    state MASTER                       #状态
    interface eno16777736              #keepalived所配置的接口  
    virtual_router_id 88              #VRID
    priority 100                     #优先级
    advert_int 1                   
    authentication {
        auth_type PASS
        auth_pass 1111
    }
 
    virtual_ipaddress {
        10.1.88.88/16 dev eno16777736 label eno16777736:0    #VIP
    }
 
    track_script {
        chk_down
    }
    
    notify_master "/etc/keepalived/notify.sh master"
    notify_backup "/etc/keepalived/notify.sh backup"
    notify_fault  "/etc/keepalived/notify.sh fault"
}
 
virtual_server 10.1.88.88 80 {    #配置虚拟服务器
    delay_loop 3       
    lb_algo rr
    lb_kind DR
    protocol TCP
 
    sorry_server 127.0.0.1 80        #当所有realserver宕机后,则启用sorry_server
 
    real_server 10.1.68.5 80 {            #realserver配置

        weight 1
        HTTP_GET {        #realserver健康状态检测
            url {
                path /
                status_code 200
            }
 
            connect_timeout 1
            nb_get_retry 3
            delay_before_retry 1
        }
    }
 
    real_server 10.1.68.6 80 {
        weight 1
        HTTP_GET {
            url {
                path /
                status_code 200
            }
 
            connect_timeout 1
            nb_get_retry 3
            delay_before_retry 1
        }
    }
}
keepalived backup配置:
global_defs {
   notification_email {
       root@localhost
   }
   notification_email_from keepalived@localhost
   smtp_server 127.0.0.1
   smtp_connect_timeout 30
   router_id node1
}
 
vrrp_script chk_down {
    script "[[ -f /etc/keepalived/down ]] && exit 1 || exit 0"
    interval 2
    weight -6
}
 
vrrp_instance VI_1 {
    state BACKUP             #配置bakup服务器state为BACKUP
    interface eno16777736
    virtual_router_id 88
    priority 98              #优先级为98,比master低
    advert_int 1
    authentication {
        auth_type PASS
        auth_pass 1111
    }
 
    virtual_ipaddress {
        10.1.88.88/16 dev eno16777736 label eno16777736:0
    }
 
    track_script {
        chk_down
    }
 
    notify_master "/etc/keepalived/notify.sh master"
    notify_backup "/etc/keepalived/notify.sh backup"
    notify_fault  "/etc/keepalived/notify.sh fault"
}
 
virtual_server 10.1.88.88 80 {
    delay_loop 3
    lb_algo rr
    lb_kind DR
    protocol TCP
 
    sorry_server 127.0.0.1 80   
 
    real_server 10.1.68.5 80 {
        weight 1
        HTTP_GET {
            url {
                path /
                status_code 200
            }
 
            connect_timeout 1
            nb_get_retry 3
            delay_before_retry 1
        }
    }
 
    real_server 10.1.68.6 80 {
        weight 1
        HTTP_GET {
            url {
                path /

 

                status_code 200
            }
 
            connect_timeout 1
            nb_get_retry 3
            delay_before_retry 1
        }
    }
}
在两台调度器上面配置httpd,用于sorry_server

在两台前端调度器上同时启动keepalived
service keepalived start    #master
service keepalived start    #backup
此时两台调度器已经同时生成了ipvs规则

此时VIP已经在master的网卡接口上

4.测试

模拟其中一台调度器故障,VIP立即漂移至backup调度器,不影响调度


其中有一台故障可正常调度至另一台realserver

后端realserver全部故障,此时sorry_server服务器响应

 

 

延伸阅读:

Tag标签: keepalived高可用lvs  
  • 专题推荐

About IT165 - 广告服务 - 隐私声明 - 版权申明 - 免责条款 - 网站地图 - 网友投稿 - 联系方式
本站内容来自于互联网,仅供用于网络技术学习,学习中请遵循相关法律法规