当前位置： 首页 > news >正文

Keepalived双机热备概述

news 2025/7/16 5:50:45

前言：

        在现代IT架构中，服务的高可用性是保障业务连续性的核心需求之一。Keepalived作为一款轻量级的高可用解决方案，通过VRRP协议实现IP漂移和健康检测，能够有效构建双机热备系统，避免单点故障导致的业务中断。无论是Web服务、数据库还是负载均衡集群，Keepalived都能以简洁的配置和高效的故障切换机制，为关键应用提供稳定的冗余保障。本文将简要概述Keepalived双机热备的原理与部署要点，帮助读者快速掌握其核心能力。

目录

概述

一、keepalived初始化部署

二、配置文件

三、 其他配置项说明

四、名词解释

五、高阶使用

1、介绍

2、keepalived主要作用

3、工作在三层、四层和七层原理

4、健康状态检测方式

4.1 HTTP服务状态检测

 ​

4.2 TCP端口状态检测（使用TCP端口服务基本上都可以使用）

4.3 邮件服务器SMTP检测

4.4 用户自定义脚本检测real_server服务状态

5、状态转换通知功能

5.1 实例状态通知

5.2 虚拟服务器检测通知

六、项目实际中配置

总结

概述

Keepalived是一个基于VRRP协议（虚拟冗余路由协议）来实现的LVS服务高可用方案，可以利用其来避免单点故障。一个LVS服务会有2台服务器运行Keepalived，一台为主服务器（MASTER），一台为备份服务器（BACKUP），但是对外表现为一个虚拟IP，主服务器会发送特定的消息（心跳检测，heartbeat）给备份服务器，当备份服务器收不到这个消息的时候，即主服务器宕机的时候， 备份服务器就会接管虚拟IP，继续提供服务，从而保证了高可用性。

Keepalived的作用是检测服务器的状态，如果有一台web服务器死机，或工作出现故障，Keepalived将检测到，并将有故障的服务器从系统中剔除，同时使用其他服务器代替该服务器的工作，当服务器工作正常后Keepalived自动将服务器加入到服务器群中。

一、keepalived初始化部署

安装

[root@web1 ~]# yum install keepalived

配置并启动

 [root@web2 keepalived]# systemctl enable --now  keepalived.service nginx.service

验证

 [root@web1 keepalived]# ip a2: ens33: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP group default qlen 1000link/ether 00:0c:29:c8:dc:33 brd ff:ff:ff:ff:ff:ffinet 192.168.115.111/24 brd 192.168.115.255 scope global ens33valid_lft forever preferred_lft foreverinet 192.168.115.250/32 scope global ens33##停掉master查看backup[root@web1 keepalived]# systemctl stop  keepalived.service[root@web2 keepalived]# ip a2: ens33: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP group default qlen 1000link/ether 00:0c:29:8a:4a:79 brd ff:ff:ff:ff:ff:ffinet 192.168.115.112/24 brd 192.168.115.255 scope global ens33valid_lft forever preferred_lft foreverinet 192.168.115.250/32 scope global ens33

二、配置文件

 global_defs {notification_email {  #指定keepalived在发生切换时需要发送email到的对象，一行一个sysadmin@fire.loc}notification_email_from Alexandre.Cassen@firewall.loc #指定发件人smtp_server localhost #指定smtp服务器地址smtp_connect_timeout 30 #指定smtp连接超时时间router_id LVS_DEVEL #运行keepalived机器的一个标识}vrrp_sync_group VG_1{ #监控多个网段的实例group {inside_network #实例名outside_network}notify_master /path/xx.sh #指定当切换到master时，执行的脚本notify_backup /path/xx.sh #指定当切换到backup时，执行的脚本notify_fault "path/xx.sh VG_1" #故障时执行的脚本notify /path/xx.shsmtp_alert #使用global_defs中提供的邮件地址和smtp服务器发送邮件通知}

Keepalived在转换状态时会依照状态来呼叫：

  • 当进入Master状态时会呼叫notify_master

  • 当进入Backup状态时会呼叫notify_backup

  • 当发现异常情况时进入Fault状态呼叫notify_fault

 vrrp_instance inside_network {state BACKUP #指定那个为master，那个为backup，如果设置了nopreempt这个值不起作用，主备靠priority决定interface eth0 #设置实例绑定的网卡 VRRP心跳包从哪块网卡发出dont_track_primary #忽略vrrp的interface错误（默认不设置）track_interface{ #设置额外的监控，里面那个网卡出现问题都会切换eth1eth2}mcast_src_ip #发送多播包的地址，如果不设置默认使用绑定网卡的primary ipgarp_master_delay #在切换到master状态后，延迟进行gratuitous ARP请求virtual_router_id 50 #VPID标记 相同VRID的LVS属于同一组，根据优先级选举出一个主priority 99 #优先级，高优先级竞选为masteradvert_int 10 #检查间隔，默认1秒 VRRP心跳包（报文）的发送周期，单位为s 组播信息发送间隔，两个节点设置必须一样（实际并不一定完全是10秒，测试结果是小于10秒的随机值）nopreempt #设置为不抢占 注：这个配置只能设置在backup主机上，而且这个主机优先级要比另外一台高

首先nopreemt必须在state为BACKUP的节点上才生效（因为是BACKUP节点决定是否来成为MASTER的），其次要实现类似于关闭auto failback的功能需要将所有节点的state都设置为BACKUP，或者将master节点的priority设置的比BACKUP低。我个人推荐使用将所有节点的state都设置成BACKUP并且都加上nopreempt选项，这样就完成了关于autofailback功能，当想手动将某节点切换为MASTER时只需去掉该节点的nopreempt选项并且将priority改的比其他节点大，然后重新加载配置文件即可（等MASTER切过来之后再将配置文件改回去再reload一下）。

preempt_delay #抢占延时，默认5分钟debug #debug级别authentication { #设置认证auth_type PASS #认证方式auth_pass 111111 #认证密码（密码只识别前8位）}virtual_ipaddress { #设置vip192.168.202.200}}virtual_server 192.168.202.200 23 {delay_loop 6 #健康检查时间间隔（实际并不一定完全是6秒，测试结果是小于6秒的随机值？）lb_algo rr  #lvs调度算法rr|wrr|lc|wlc|lblc|sh|dhlb_kind DR  #负载均衡转发规则NAT|DR|TUNpersistence_timeout 5 #会话保持时间protocol TCP #使用的协议persistence_granularity <NETMASK> #lvs会话保持粒度virtualhost <string> #检查的web服务器的虚拟主机（host：头）   sorry_server<IPADDR> <port> #备用机，所有realserver失效后启用real_server 192.168.200.5 23 {weight 1 #默认为1,0为失效inhibit_on_failure #在服务器健康检查失效时，将其设为0，而不是直接从ipvs中删除notify_up <string> | <quoted-string> #在检测到server up后执行脚本notify_down <string> | <quoted-string> #在检测到server down后执行脚本TCP_CHECK {connect_timeout 3 #连接超时时间nb_get_retry 3 #重连次数delay_before_retry 3 #重连间隔时间connect_port 23  #健康检查的端口的端口bindto <ip>  #检查的IP地址}HTTP_GET | SSL_GET{url{ #检查url，可以指定多个path /digest <string> #检查后的摘要信息status_code 200 #检查的返回状态码，301 302 }connect_port <port>bindto <IPADD>connect_timeout 5nb_get_retry 3delay_before_retry 2}​SMTP_CHECK{host{connect_ip <IP ADDRESS>connect_port <port> #默认检查25端口bindto <IP ADDRESS>}connect_timeout 5retry 3delay_before_retry 2helo_name <string> | <quoted-string> #smtp helo请求命令参数，可选}MISC_CHECK{misc_path <string> | <quoted-string> #外部脚本路径misc_timeout #脚本执行超时时间misc_dynamic #如设置该项，则退出状态码会用来动态调整服务器的权重，返回0 正常，不修改；返回1，检查失败，权重改为0；返回2-255，正常，权重设置为：返回状态码-2}}​
 

三、 其他配置项说明
 

keepalived 的核心就是将IPVS配置成高可用，生成ipvs规则来完成负载均衡效果。
 

  •  
    virtual server （虚拟服务）的定义：
     
  •  
    virtual_server IP port #定义虚拟主机IP地址及其端口
     
  •  
    virtual_server fwmark int #ipvs的防火墙打标，实现基于防火墙的负载均衡集群
     
  •  
    virtual_server group string #将多个虚拟服务器定义成组，将组定义成虚拟服务
     
  •  
    lb_algo{rr|wrr|lc|wlc|lblc|lblcr} #定义LVS的调度算法
     
  •  
    lb_kind {NAT|DR|TUN} #定义LVS的模型
     
  •  
    presitence_timeout #定义支持持久连接的时长
     
  •  
    protocol TCP #规则所能支持的协议
     
  •  
    sorry_server #如果所有real_server都出现故障了，利用此返回信息
     
 

四、名词解释
 

虚拟路由器： 由一个Master路由器和多个Backup路由器组成。主机将虚拟路由器当作默认网关;
 

VRID：虚拟路由器的标识。有相同VRID的一组路由器构成一个虚拟路由器;
 

Master路由器：虚拟路由器中承担报文转发任务的路由器;
 

Backup路由器 ：Master路由器出现故障时，能够代替Master路由器工作的路由器;
 

虚拟IP 地址：虚拟路由器的IP 地址。一个虚拟路由器可以拥有一个或多个IP地址;
 

IP地址拥有者： 接口IP地址与虚拟IP地址相同的路由器被称为IP地址拥有者;
 

虚拟MAC地址： 一个虚拟路由器拥有一个虚拟MAC地址。通常情况下，虚拟路由器回应ARP请求使用的是虚拟MAC地址，只有虚拟路由器做特殊配置的时候，才回应接口的真实MAC地址;
 

优先级：VRRP根据优先级来确定虚拟路由器中每台路由器的地位;
 

非抢占方式：如果Backup路由器工作在非抢占方式下，则只要Master路由器没有出现故障，Backup路由器即使随后被配置了更高的优先级也不会成为Master路由器;
 

抢占方式：如果Backup路由器工作在抢占方式下，当它收到VRRP报文后，会将自己的优先级与通告报文中的优先级进行比较。如果自己的优先级比当前的Master路由器的优先级高，就会主动抢占成为Master路由器；否则，将保持Backup状态
 

五、高阶使用
 

1、介绍
 

Keeaplived 主要有两种应用场景，一个是通过配置keepalived结合ipvs做到负载均衡（LVS+Keepalived）。另一个是通过自身健康检查、资源接管功能做高可用（双机热备），实现故障转移。
 

以下内容主要针对Keepalived+MySQL双主实现双机热备为根据，主要讲解keepalived的状态转换通知功能，利用此功能可有效加强对MySQL数据库监控。
 

2、keepalived主要作用
 

keepalived采用VRRP（virtual router redundancy protocol），虚拟路由冗余协议，以软件的形式实现服务器热备功能。通常情况下是将两台linux服务器组成一个热备组（master-backup），同一时间热备组内只有一台主服务器（master）提供服务，同时master会虚拟出一个共用IP地址（VIP），这个VIP只存在master上并对外提供服务。如果keepalived检测到master宕机或服务故障，备服务器（backup）会自动接管VIP成为master，keepalived并将master从热备组移除，当master恢复后，会自动加入到热备组，默认再抢占成为master，起到故障转移功能。
 

3、工作在三层、四层和七层原理
 

Layer3： 工作在三层时，keepalived会定期向热备组中的服务器发送一个ICMP数据包，来判断某台服务器是否故障，如果故障则将这台服务器从热备组移除。
 

Layer4： 工作在四层时，keepalived以TCP端口的状态判断服务器是否故障，比如检测mysql 3306端口，如果故障则将这台服务器从热备组移除。
 

 ! Configuration File for keepalivedglobal_defs {notification_email {     example@163.com  }    notification_email_from  example@example.comsmtp_server 127.0.0.1smtp_connect_timeout 30router_id MYSQL_HA}vrrp_instance VI_1 {  state BACKUPinterface eth1virtual_router_id 50      nopreempt                   #当主down时，备接管，主恢复，不自动接管      priority 100      advert_int 1authentication {         auth_type PASS   ahth_pass 123       }      virtual_ipaddress {     192.168.1.200          #虚拟IP地址}virtual_server 192.168.1.200 3306 {            delay_loop 6#    lb_algo rr #    lb_kind NATpersistence_timeout 50    protocol TCP  real_server 192.168.1.201 3306 {       #监控本机3306端口        weight 1   notify_down /etc/keepalived/kill_keepalived.sh   #检测3306端口为down状态就执行此脚本（只有keepalived关闭，VIP才漂移 ）   TCP_CHECK {         #健康状态检测方式，可针对业务需求调整（TTP_GET|SSL_GET|TCP_CHECK|SMTP_CHECK|MISC_CHECK）connect_timeout 3           nb_get_retry 3             delay_before_retry 3     }  #MISC_CHECK {                       ## 使用 MISC_CHECK 方式自定义脚本做健康检查#    misc_path "/etc/keepalived/check.sh"    ## 检测脚本#    misc_timeout   10              ## 执行脚本的超时时间#    misc_dynamic                   ## 根据退出状态码动态调整服务器的权重#    }​}}
 

 ​
 

Layer7：工作在七层时，keepalived根据用户设定的策略判断服务器上的程序是否正常运行，如果故障则将这台服务器从热备组移除。
 

  ! Configuration File for keepalivedglobal_defs {notification_email {example@163.com}notification_email_from  example@example.com      smtp_server 127.0.0.1smtp_connect_timeout 30router_id MYSQL_HA}vrrp_script check_nginx {   script /etc/keepalived/check_nginx.sh    #检测脚本interval 2   #执行间隔时间}vrrp_instance VI_1 {state BACKUPinterface eth1virtual_router_id 50nopreempt                   #当主down时，备接管，主恢复，不自动接管priority 100 advert_int 1authentication {   auth_type PASS        ahth_pass 123     }     virtual_ipaddress {192.168.1.200          #虚拟IP地址   }     track_script {          #在实例中引用脚本    check_nginx      }  }
 

 ​
 

脚本内容如下：
 

         # cat /etc/keepalived/check_nginx.shCount1=`netstat -antp |grep -v grep |grep nginx |wc -l`if [ $Count1 -eq 0 ]; then/usr/local/nginx/sbin/nginxsleep 2  Count2=`netstat -antp |grep -v grep |grep nginx |wc -l`if [ $Count2 -eq 0 ]; then service keepalived stop      elseexit 0fi else   exit 0fi ​#也可以简单如下：#!/bin/bash[[ `ps -C nginx --no-header |wc -l` -eq 0 ]] && exit 1 || exit 0# 如果没有nginx进程 返回错误状态1  #if [ `ps -C nginx --no-header |wc -l` -eq 0 ];then#        echo "$(date) nginx pid not found">>/etc/keepalived/keepalived.log#        #killall keepalived#fi#在keepalived主机中，运行如下脚本，用来检测本机的nginx进行状态vim start_keepalived.sh#!/bin/bashwhile truedoif pgrep nginx &> /dev/null;thensystemctl start keepalivedfisleep 3​done​​​#mysql 检测脚本如下：#!/bin/bash[[ `ps -C mysqld --no-header |wc -l` -eq 0 ]] && exit 1 || exit 0# 如果没有nginx进程 返回错误状态1  if [ `ps -C nginx --no-header |wc -l` -eq 0 ];thenecho "$(date) nginx pid not found">>/etc/keepalived/keepalived.logkillall keepalivedfi
 

 ​
 

4、健康状态检测方式
 

4.1 HTTP服务状态检测
 

 HTTP_GET或SSL_GET {           url {      path /index.html        #检测url，可写多个         digest  24326582a86bee478bac72d5af25089e    #检测效验码        #digest效验码获取方法：genhash -s IP -p 80 -u http://IP/index.html         status_code 200         #检测返回http状态码    }       connect_port 80 #连接端口  connect_timeout 3  #连接超时时间 nb_get_retry 3  #重试次数delay_before_retry 2 #连接间隔时间}
 

 ​
 

4.2 TCP端口状态检测（使用TCP端口服务基本上都可以使用）
 

         TCP_CHECK {             connect_port 80     #健康检测端口，默认为real_server后跟端口    connect_timeout 5      nb_get_retry 3      delay_before_retry 3}
 

 

4.3 邮件服务器SMTP检测
 

 SMTP_CHECK {            #健康检测邮件服务器smtp       host {        connect_ip          connect_port    }           connect_timeout 5       retry 2    delay_before_retry 3   hello_name "mail.domain.com"   }
 

 

4.4 用户自定义脚本检测real_server服务状态
 

  MISC_CHECK {         misc_path /script.sh    #指定外部程序或脚本位置      misc_timeout 3      #执行脚本超时时间  !misc_dynamic       #不动态调整服务器权重（weight），如果启用将通过退出状态码动态调整real_server权重值# misc_dynamic        ## 根据退出状态码动态调整服务器的权重}
 

 ​
 

5、状态转换通知功能
 

keepalived主配置邮件通知功能，默认当real_server宕机或者恢复时才会发出邮件。有时我们更想知道keepalived的主服务器故障切换后，VIP是否顺利漂移到备服务器，MySQL服务器是否正常？那写个监控脚本吧，可以，但没必要，因为keepalived具备状态检测功能，所以我们直接使用就行了。
 

 主配置默认邮件通知配置模板如下：global_defs           # Block id{   notification_email    # To:   {      admin@example1.com ...        }​# From: from address that will be in header    notification_email_from admin@example.comsmtp_server 127.0.0.1   # IP  smtp_connect_timeout 30 # integer, secondsrouter_id my_hostname   # string identifying the machine,                        # (doesn't have to be hostname).
 

5.1 实例状态通知
 

a) notify_master ：节点变为master时执行
 

b) notify_backup ： 节点变为backup时执行
 

c) notify_fault ： 节点变为故障时执行
 

5.2 虚拟服务器检测通知
 

a) notify_up ： 虚拟服务器up时执行
 

b) notify_down ： 虚拟服务器down时执行
 

 示例：! Configuration File for keepalivedglobal_defs { notification_email {example@163.com     }   notification_email_from example@example.com    smtp_server 127.0.0.1 smtp_connect_timeout 30  router_id MYSQL_HA}    vrrp_instance VI_1 {       state BACKUP     interface eth1virtual_router_id 50  nopreempt           #当主down时，备接管，主恢复，不自动接管   priority 100advert_int 1  authentication {          auth_type PASS          ahth_pass 123     }    virtual_ipaddress {      192.168.1.200     }          notify_master /etc/keepalived/to_master.sh          notify_backup /etc/keepalived/to_backup.shnotify_fault /etc/keepalived/to_fault.sh    } virtual_server 192.168.1.200 3306 {     delay_loop 6     persistence_timeout 50    protocol TCP       real_server 192.168.1.201 3306 {            weight 1         notify_up /etc/keepalived/mysql_up.sh         notify_down /etc/keepalived/mysql_down.sh           TCP_CHECK {connect_timeout 3            nb_get_retry 3            delay_before_retry 3           }        }     }
 

状态参数后可以是bash命令，也可以是shell脚本，内容根据自己需求定义，以上示例中所涉及状态脚本如下：
 

1.当服务器改变为主时执行此脚本
 

 yum install -y sendmail mailxsystemctl enabled --now sendmail# cat to_master.sh #!/bin/bashDate=$(date +%F" "%T)IP=$(ifconfig eth0 |grep "inet addr" |cut -d":" -f2 |awk '{print $1}')Mail="z13516052620@163.com"echo "$Date $IP change to master." |mail -s "Master-Backup Change Status" $Mail
 

2.当服务器改变为备时执行此脚本
 

  # cat to_backup.sh#!/bin/bashDate=$(date +%F" "%T)IP=$(ifconfig eth0 |grep "inet addr" |cut -d":" -f2 |awk '{print $1}')Mail="baojingtongzhi@163.com"echo "$Date $IP change to backup." |mail -s "Master-Backup Change Status" $Mail
 

3.当服务器改变为故障时执行此脚本
 

   # cat to_fault.sh#!/bin/bashDate=$(date +%F" "%T)IP=$(ifconfig eth0 |grep "inet addr" |cut -d":" -f2 |awk '{print $1}')Mail="baojingtongzhi@163.com"echo "$Date $IP change to fault." |mail -s "Master-Backup Change Status" $Mail
 

4.当检测TCP端口3306为不可用时，执行此脚本，杀死keepalived，实现切换
 

 # cat mysql_down.sh#!/bin/bashDate=$(date +%F" "%T)IP=$(ifconfig eth0 |grep "inet addr" |cut -d":" -f2 |awk '{print $1}')Mail="baojingtongzhi@163.com"pkill keepalivedecho "$Date $IP The mysql service failure,kill keepalived." |mail -s "Master-Backup MySQL Monitor" $Mail
 

5.当检测TCP端口3306可用时，执行此脚本
 

  # cat mysql_up.sh#!/bin/bashDate=$(date +%F" "%T)IP=$(ifconfig eth0 |grep "inet addr" |cut -d":" -f2 |awk '{print $1}')Mail="baojingtongzhi@163.com"echo "$Date $IP The mysql service is recovery." |mail -s "Master-Backup MySQL Monitor" $Mail
 

六、项目实际中配置
 

Keepalived高可用集群一主一从配置
 

web1、web2主机分别下载keepalived、nginx
 

[root@web1 ~]# yum install keepalived nginx
[root@web2 ~]# yum install keepalived nginx
 

主web1修改配置文件
 

[root@web1 ~]# cd /etc/keepalived/
[root@web1 keepalived]# ls
keepalived.conf.sample
[root@web1 keepalived]# cp keepalived.conf.sample  keepalived.conf
 

 

从web2修改配置文件
 

[root@web2 ~]# cd /etc/keepalived/
[root@web2 keepalived]# ls
keepalived.conf.sample
[root@web2 keepalived]# cp keepalived.conf.sample keepalived.conf
[root@web2 keepalived]# vim keepalived.conf
 

 

切换到主web1，开启keepalived.service
 

[root@web1 keepalived]# systemctl start keepalived.service 
 

keepalived.service启动后ip a命令查看该ens33接口承载了多个虚拟IP(VIP)，表明Keepalived已成功启动并分配了VIP
 

[root@web1 ~]# ip a
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00inet 127.0.0.1/8 scope host lovalid_lft forever preferred_lft foreverinet6 ::1/128 scope host noprefixroute valid_lft forever preferred_lft forever
2: ens33: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000link/ether 00:0c:29:b0:07:74 brd ff:ff:ff:ff:ff:ffinet 192.168.100.1/24 brd 192.168.100.255 scope global noprefixroute ens33valid_lft forever preferred_lft foreverinet 192.168.100.101/32 scope global ens33valid_lft forever preferred_lft foreverinet 192.168.100.102/32 scope global ens33valid_lft forever preferred_lft foreverinet 192.168.100.103/32 scope global ens33valid_lft forever preferred_lft foreverinet6 fe80::20c:29ff:feb0:774/64 scope link noprefixroute valid_lft forever preferred_lft forever
3: ens34: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000link/ether 00:0c:29:b0:07:7e brd ff:ff:ff:ff:ff:ffinet 192.168.72.176/24 brd 192.168.72.255 scope global dynamic noprefixroute ens34valid_lft 1167sec preferred_lft 1167secinet6 fe80::20c:29ff:feb0:77e/64 scope link noprefixroute valid_lft forever preferred_lft forever
 

切换到从web2，开启keepalived.service
 

[root@web2 ~]# systemctl start keepalived
 

当前web2未显示VIP因为是BACKUP节点，
 

[root@web2 ~]# ip a
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00inet 127.0.0.1/8 scope host lovalid_lft forever preferred_lft foreverinet6 ::1/128 scope host noprefixroute valid_lft forever preferred_lft forever
2: ens33: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000link/ether 00:0c:29:59:8d:4f brd ff:ff:ff:ff:ff:ffinet 192.168.100.2/24 brd 192.168.100.255 scope global noprefixroute ens33valid_lft forever preferred_lft foreverinet6 fe80::20c:29ff:fe59:8d4f/64 scope link noprefixroute valid_lft forever preferred_lft forever
3: ens34: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000link/ether 00:0c:29:59:8d:59 brd ff:ff:ff:ff:ff:ffinet 192.168.72.178/24 brd 192.168.72.255 scope global dynamic noprefixroute ens34valid_lft 1600sec preferred_lft 1600secinet6 fe80::20c:29ff:fe59:8d59/64 scope link noprefixroute valid_lft forever preferred_lft forever
 

上述Keepalived高可用集群一主一从配置完成
 

模拟故障：
 

手动停止web1的keepalived，验证web2是否接管VIP
 

切换到主web1
 

[root@web1 ~]# systemctl stop keepalived.service 
 

切换到从web2​​测试故障转移
 

[root@web2 ~]# ip a
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00inet 127.0.0.1/8 scope host lovalid_lft forever preferred_lft foreverinet6 ::1/128 scope host noprefixroute valid_lft forever preferred_lft forever
2: ens33: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000link/ether 00:0c:29:59:8d:4f brd ff:ff:ff:ff:ff:ffinet 192.168.100.2/24 brd 192.168.100.255 scope global noprefixroute ens33valid_lft forever preferred_lft foreverinet 192.168.100.101/32 scope global ens33valid_lft forever preferred_lft foreverinet 192.168.100.102/32 scope global ens33valid_lft forever preferred_lft foreverinet 192.168.100.103/32 scope global ens33valid_lft forever preferred_lft foreverinet6 fe80::20c:29ff:fe59:8d4f/64 scope link noprefixroute valid_lft forever preferred_lft forever
3: ens34: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000link/ether 00:0c:29:59:8d:59 brd ff:ff:ff:ff:ff:ffinet 192.168.72.178/24 brd 192.168.72.255 scope global dynamic noprefixroute ens34valid_lft 1139sec preferred_lft 1139secinet6 fe80::20c:29ff:fe59:8d59/64 scope link noprefixroute valid_lft forever preferred_lft forever
 

web1、web2分别开启防火墙，web1、web2分别关闭keepalived.service
 

[root@web1 ~]# systemctl start firewalld
[root@web1 ~]# systemctl stop keepalived.service[root@web2 ~]# systemctl start firewalld
[root@web2 ~]# systemctl stop keepalived.service
 

web1再次开启keepalived.service，VIP返回
 

[root@web1 ~]# ip a
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00inet 127.0.0.1/8 scope host lovalid_lft forever preferred_lft foreverinet6 ::1/128 scope host noprefixroute valid_lft forever preferred_lft forever
2: ens33: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000link/ether 00:0c:29:b0:07:74 brd ff:ff:ff:ff:ff:ffinet 192.168.100.1/24 brd 192.168.100.255 scope global noprefixroute ens33valid_lft forever preferred_lft foreverinet 192.168.100.101/32 scope global ens33valid_lft forever preferred_lft foreverinet 192.168.100.102/32 scope global ens33valid_lft forever preferred_lft foreverinet 192.168.100.103/32 scope global ens33valid_lft forever preferred_lft foreverinet6 fe80::20c:29ff:feb0:774/64 scope link noprefixroute valid_lft forever preferred_lft forever
3: ens34: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000link/ether 00:0c:29:b0:07:7e brd ff:ff:ff:ff:ff:ffinet 192.168.72.176/24 brd 192.168.72.255 scope global dynamic noprefixroute ens34valid_lft 1507sec preferred_lft 1507secinet6 fe80::20c:29ff:feb0:77e/64 scope link noprefixroute valid_lft forever preferred_lft forever
 

web2也开启keepalived.service，此时可以看到同一个热备组，但是相同的IP，则发生脑裂
 

[root@web2 ~]# ip a
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00inet 127.0.0.1/8 scope host lovalid_lft forever preferred_lft foreverinet6 ::1/128 scope host noprefixroute valid_lft forever preferred_lft forever
2: ens33: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000link/ether 00:0c:29:59:8d:4f brd ff:ff:ff:ff:ff:ffinet 192.168.100.2/24 brd 192.168.100.255 scope global noprefixroute ens33valid_lft forever preferred_lft foreverinet 192.168.100.101/32 scope global ens33valid_lft forever preferred_lft foreverinet 192.168.100.102/32 scope global ens33valid_lft forever preferred_lft foreverinet 192.168.100.103/32 scope global ens33valid_lft forever preferred_lft foreverinet6 fe80::20c:29ff:fe59:8d4f/64 scope link noprefixroute valid_lft forever preferred_lft forever
3: ens34: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000link/ether 00:0c:29:59:8d:59 brd ff:ff:ff:ff:ff:ffinet 192.168.72.178/24 brd 192.168.72.255 scope global dynamic noprefixroute ens34valid_lft 1184sec preferred_lft 1184secinet6 fe80::20c:29ff:fe59:8d59/64 scope link noprefixroute valid_lft forever preferred_lft forever
 

故障恢复：
 

web1、web2分别关闭防火墙
 

[root@web1 ~]# systemctl stop firewalld
[root@web2 ~]# systemctl stop firewalld
 

 web1脑裂现象恢复
 

[root@web1 ~]# ip a
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00inet 127.0.0.1/8 scope host lovalid_lft forever preferred_lft foreverinet6 ::1/128 scope host noprefixroute valid_lft forever preferred_lft forever
2: ens33: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000link/ether 00:0c:29:b0:07:74 brd ff:ff:ff:ff:ff:ffinet 192.168.100.1/24 brd 192.168.100.255 scope global noprefixroute ens33valid_lft forever preferred_lft foreverinet 192.168.100.101/32 scope global ens33valid_lft forever preferred_lft foreverinet 192.168.100.102/32 scope global ens33valid_lft forever preferred_lft foreverinet 192.168.100.103/32 scope global ens33valid_lft forever preferred_lft foreverinet6 fe80::20c:29ff:feb0:774/64 scope link noprefixroute valid_lft forever preferred_lft forever
3: ens34: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000link/ether 00:0c:29:b0:07:7e brd ff:ff:ff:ff:ff:ffinet 192.168.72.176/24 brd 192.168.72.255 scope global dynamic noprefixroute ens34valid_lft 1730sec preferred_lft 1730secinet6 fe80::20c:29ff:feb0:77e/64 scope link noprefixroute valid_lft forever preferred_lft forever
 

 web2脑裂现象恢复 
 

[root@web2 ~]# ip a
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00inet 127.0.0.1/8 scope host lovalid_lft forever preferred_lft foreverinet6 ::1/128 scope host noprefixroute valid_lft forever preferred_lft forever
2: ens33: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000link/ether 00:0c:29:59:8d:4f brd ff:ff:ff:ff:ff:ffinet 192.168.100.2/24 brd 192.168.100.255 scope global noprefixroute ens33valid_lft forever preferred_lft foreverinet6 fe80::20c:29ff:fe59:8d4f/64 scope link noprefixroute valid_lft forever preferred_lft forever
3: ens34: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000link/ether 00:0c:29:59:8d:59 brd ff:ff:ff:ff:ff:ffinet 192.168.72.178/24 brd 192.168.72.255 scope global dynamic noprefixroute ens34valid_lft 1567sec preferred_lft 1567secinet6 fe80::20c:29ff:fe59:8d59/64 scope link noprefixroute valid_lft forever preferred_lft forever
 

验证​Keepalived实现高可用以及Nginx流量分发问题：
 

###web1开启nginx，输入内容到index.html文件中
[root@web1 ~]# systemctl start nginx
[root@web1 ~]# echo web1 > /usr/share/nginx/html/index.html###web2开启nginx，输入内容到index.html文件中
[root@web2 ~]# systemctl start nginx
[root@web2 ~]# echo web2 > /usr/share/nginx/html/index.html
 

访问192.168.100.101web站点，只能访问到web1上，因为web1为主持有VIP，而web2为从只能等待接管，当web1正常时，所有请求都流向web1，只有web1故障时，VIP才会漂移到web2
 

 

此时web1关闭keepalived.service模拟宕机
 

[root@web1 ~]# systemctl stop keepalived.service
 

切换到web2查看到已将web1之前持有的VIP接管，表明VIP当前已漂移到web2节点
 

[root@web2 ~]# ip a
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00inet 127.0.0.1/8 scope host lovalid_lft forever preferred_lft foreverinet6 ::1/128 scope host noprefixroute valid_lft forever preferred_lft forever
2: ens33: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000link/ether 00:0c:29:59:8d:4f brd ff:ff:ff:ff:ff:ffinet 192.168.100.2/24 brd 192.168.100.255 scope global noprefixroute ens33valid_lft forever preferred_lft foreverinet 192.168.100.101/32 scope global ens33valid_lft forever preferred_lft foreverinet 192.168.100.102/32 scope global ens33valid_lft forever preferred_lft foreverinet 192.168.100.103/32 scope global ens33valid_lft forever preferred_lft foreverinet6 fe80::20c:29ff:fe59:8d4f/64 scope link noprefixroute valid_lft forever preferred_lft forever
3: ens34: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000link/ether 00:0c:29:59:8d:59 brd ff:ff:ff:ff:ff:ffinet 192.168.72.178/24 brd 192.168.72.255 scope global dynamic noprefixroute ens34valid_lft 1667sec preferred_lft 1667secinet6 fe80::20c:29ff:fe59:8d59/64 scope link noprefixroute valid_lft forever preferred_lft forever
 

访问192.168.100.101web站点，此时可以看到web2接管VIP后所有流量都会指向当前持有VIP的web2节点，此时成功测试了故障转移，web2已接管服务
 

 

此时切换到web1上重新启动keepalived.service，
 

[root@web1 ~]# systemctl start keepalived.service
 

重启web1的keepalived后，因web1主设置的优先级​大于web2从，VIP 192.168.100.101已成功绑定到web1的ens33接口，同时还有其他VIP(192.168.100.102/103)也绑定在该接口，表明web1已恢复为MASTER节点
 

[root@web1 ~]# ip a
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00inet 127.0.0.1/8 scope host lovalid_lft forever preferred_lft foreverinet6 ::1/128 scope host noprefixroute valid_lft forever preferred_lft forever
2: ens33: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000link/ether 00:0c:29:b0:07:74 brd ff:ff:ff:ff:ff:ffinet 192.168.100.1/24 brd 192.168.100.255 scope global noprefixroute ens33valid_lft forever preferred_lft foreverinet 192.168.100.101/32 scope global ens33valid_lft forever preferred_lft foreverinet 192.168.100.102/32 scope global ens33valid_lft forever preferred_lft foreverinet 192.168.100.103/32 scope global ens33valid_lft forever preferred_lft foreverinet6 fe80::20c:29ff:feb0:774/64 scope link noprefixroute valid_lft forever preferred_lft forever
3: ens34: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000link/ether 00:0c:29:b0:07:7e brd ff:ff:ff:ff:ff:ffinet 192.168.72.176/24 brd 192.168.72.255 scope global dynamic noprefixroute ens34valid_lft 1060sec preferred_lft 1060secinet6 fe80::20c:29ff:feb0:77e/64 scope link noprefixroute valid_lft forever preferred_lft forever
[root@web1 ~]# 
 

以上当前状态表明Keepalived高可用集群配置正确，能够按预期进行主备切换
 

Keepalived与Nginx集成配置高可用环境
 

web1修改配置文件
 

[root@web1 keepalived]# vim keepalived.conf
 

 

编写脚本
 

[root@web1 keepalived]# vim /etc/keepalived/check_nginx.sh
[root@web1 keepalived]# cat /etc/keepalived/check_nginx.sh Count1=`netstat -antp |grep -v grep |grep nginx |wc -l`if [ $Count1 -eq 0 ]; thensystemctl restart nginxsleep 2  Count2=`netstat -antp |grep -v grep |grep nginx |wc -l`if [ $Count2 -eq 0 ]; then service keepalived stop      elseexit 0fi 
else   exit 0
fi
 

 

给脚本赋予执行权限
 

[root@web1 keepalived]# chmod +x check_nginx.sh 
[root@web1 keepalived]# ls
check_nginx.sh  keepalived.conf  keepalived.conf.sample
 

将脚本传给web2
 

[root@web1 keepalived]# scp check_nginx.sh 192.168.100.2:/etc/keepalived/
The authenticity of host '192.168.100.2 (192.168.100.2)' can't be established.
ED25519 key fingerprint is SHA256:8qsJ8qXVwW4GcC9bkntVyvJjAgoWaVXcjyQYB6pLCtY.
This key is not known by any other names.
Are you sure you want to continue connecting (yes/no/[fingerprint])? yes
Warning: Permanently added '192.168.100.2' (ED25519) to the list of known hosts.Authorized users only. All activities may be monitored and reported.
root@192.168.100.2's password: 
check_nginx.sh                                             100%  345   398.4KB/s   00:00 
 

切换到web2查看到web1的脚本传输成功并修改web2的配置文件
 

[root@web2 ~]# cd /etc/keepalived/
[root@web2 keepalived]# ls
check_nginx.sh  keepalived.conf  keepalived.conf.sample
[root@web2 keepalived]# vim keepalived.conf
 

 

web1、web2分别重启keepalived.service
 

[root@web1 keepalived]# systemctl restart keepalived.service
[root@web2 keepalived]# systemctl restart keepalived.service
 

web1、web2验证配置正常
 

[root@web1 keepalived]# ip a
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00inet 127.0.0.1/8 scope host lovalid_lft forever preferred_lft foreverinet6 ::1/128 scope host noprefixroute valid_lft forever preferred_lft forever
2: ens33: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000link/ether 00:0c:29:b0:07:74 brd ff:ff:ff:ff:ff:ffinet 192.168.100.1/24 brd 192.168.100.255 scope global noprefixroute ens33valid_lft forever preferred_lft foreverinet 192.168.100.101/32 scope global ens33valid_lft forever preferred_lft foreverinet 192.168.100.102/32 scope global ens33valid_lft forever preferred_lft foreverinet 192.168.100.103/32 scope global ens33valid_lft forever preferred_lft foreverinet6 fe80::20c:29ff:feb0:774/64 scope link noprefixroute valid_lft forever preferred_lft forever
3: ens34: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000link/ether 00:0c:29:b0:07:7e brd ff:ff:ff:ff:ff:ffinet 192.168.72.176/24 brd 192.168.72.255 scope global dynamic noprefixroute ens34valid_lft 1547sec preferred_lft 1547secinet6 fe80::20c:29ff:feb0:77e/64 scope link noprefixroute valid_lft forever preferred_lft forever[root@web2 keepalived]# ip a
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00inet 127.0.0.1/8 scope host lovalid_lft forever preferred_lft foreverinet6 ::1/128 scope host noprefixroute valid_lft forever preferred_lft forever
2: ens33: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000link/ether 00:0c:29:59:8d:4f brd ff:ff:ff:ff:ff:ffinet 192.168.100.2/24 brd 192.168.100.255 scope global noprefixroute ens33valid_lft forever preferred_lft foreverinet6 fe80::20c:29ff:fe59:8d4f/64 scope link noprefixroute valid_lft forever preferred_lft forever
3: ens34: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000link/ether 00:0c:29:59:8d:59 brd ff:ff:ff:ff:ff:ffinet 192.168.72.178/24 brd 192.168.72.255 scope global dynamic noprefixroute ens34valid_lft 1452sec preferred_lft 1452secinet6 fe80::20c:29ff:fe59:8d59/64 scope link noprefixroute valid_lft forever preferred_lft forever
 

模拟故障：
 

web1停止nginx服务
 

[root@web1 keepalived]# systemctl stop nginx###查看服务状态
[root@web1 keepalived]# systemctl status nginx
○ nginx.service - The nginx HTTP and reverse proxy serverLoaded: loaded (/usr/lib/systemd/system/nginx.service; disabled; preset: disabled)Active: inactive (dead)7月 15 23:26:03 web1 systemd[1]: Starting The nginx HTTP and reverse proxy server...
7月 15 23:26:04 web1 nginx[4712]: nginx: the configuration file /etc/nginx/nginx.conf synt>
7月 15 23:26:04 web1 nginx[4712]: nginx: configuration file /etc/nginx/nginx.conf test is >
7月 15 23:26:04 web1 systemd[1]: Started The nginx HTTP and reverse proxy server.
7月 15 23:26:29 web1 systemd[1]: Stopping The nginx HTTP and reverse proxy server...
7月 15 23:26:29 web1 systemd[1]: nginx.service: Deactivated successfully.
7月 15 23:26:29 web1 systemd[1]: Stopped The nginx HTTP and reverse proxy server.
 

当web1的Nginx停止后，VIP（192.168.100.101/102/103）自动从web1迁移到web2，故障转移成功
 

[root@web1 keepalived]# ip a
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00inet 127.0.0.1/8 scope host lovalid_lft forever preferred_lft foreverinet6 ::1/128 scope host noprefixroute valid_lft forever preferred_lft forever
2: ens33: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000link/ether 00:0c:29:b0:07:74 brd ff:ff:ff:ff:ff:ffinet 192.168.100.1/24 brd 192.168.100.255 scope global noprefixroute ens33valid_lft forever preferred_lft foreverinet6 fe80::20c:29ff:feb0:774/64 scope link noprefixroute valid_lft forever preferred_lft forever
3: ens34: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000link/ether 00:0c:29:b0:07:7e brd ff:ff:ff:ff:ff:ffinet 192.168.72.176/24 brd 192.168.72.255 scope global dynamic noprefixroute ens34valid_lft 1779sec preferred_lft 1779secinet6 fe80::20c:29ff:feb0:77e/64 scope link noprefixroute valid_lft forever preferred_lft forever
 

切换到web2，ip a命令查看输出显示：web2已接管所有VIP（新增192.168.100.101/102/103）web1已释放VIP（仅剩192.168.100.1）
 

[root@web2 ~]# ip a
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00inet 127.0.0.1/8 scope host lovalid_lft forever preferred_lft foreverinet6 ::1/128 scope host noprefixroute valid_lft forever preferred_lft forever
2: ens33: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000link/ether 00:0c:29:59:8d:4f brd ff:ff:ff:ff:ff:ffinet 192.168.100.2/24 brd 192.168.100.255 scope global noprefixroute ens33valid_lft forever preferred_lft foreverinet 192.168.100.101/32 scope global ens33valid_lft forever preferred_lft foreverinet 192.168.100.102/32 scope global ens33valid_lft forever preferred_lft foreverinet 192.168.100.103/32 scope global ens33valid_lft forever preferred_lft foreverinet6 fe80::20c:29ff:fe59:8d4f/64 scope link noprefixroute valid_lft forever preferred_lft forever
3: ens34: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000link/ether 00:0c:29:59:8d:59 brd ff:ff:ff:ff:ff:ffinet 192.168.72.178/24 brd 192.168.72.255 scope global dynamic noprefixroute ens34valid_lft 1621sec preferred_lft 1621secinet6 fe80::20c:29ff:fe59:8d59/64 scope link noprefixroute valid_lft forever preferred_lft forever
 

访问192.168.100.101web站点，此时依然可以看到web2接管VIP后所有流量都会指向当前持有VIP的web2节点
 

 

但脚本中写了关于nginx服务启动二次检测，当检测到web1的nginx服务停止后2s则会进行回切，将web1的nginx服务重新启动，此时web2的接管服务会重新回到web1，此时脚本和高可用生效
 

 

​基于LVS(DR模式)+Keepalived+Nginx的七层高可用
 

lvs1主机修改配置文件
 

[root@lvs1 keepalived]# vim keepalived.conf
[root@lvs1 keepalived]# cat keepalived.conf
! Configuration File for keepalivedglobal_defs {router_id LVS1
}vrrp_instance VI_1 {state MASTERinterface ens33virtual_router_id 51priority 100advert_int 1authentication {auth_type PASSauth_pass 1111}virtual_ipaddress {192.168.100.103}
}virtual_server 192.168.100.103 80 {delay_loop 6lb_algo rrlb_kind DRpersistence_timeout 50protocol TCPreal_server 192.168.100.1 80 {weight 1HTTP_GET {url {path /test.html}connect_timeout 3retry 3delay_before_retry 3}}real_server 192.168.100.2 80 {weight 1HTTP_GET {url {path /test.html}connect_timeout 3retry 3delay_before_retry 3}}
}
 

lvs1将配置文件传给lvs2
 

[root@lvs1 keepalived]# scp keepalived.conf 192.168.100.101:/etc/keepalived/
The authenticity of host '192.168.100.101 (192.168.100.101)' can't be established.
ED25519 key fingerprint is SHA256:8qsJ8qXVwW4GcC9bkntVyvJjAgoWaVXcjyQYB6pLCtY.
This key is not known by any other names.
Are you sure you want to continue connecting (yes/no/[fingerprint])? yes
Warning: Permanently added '192.168.100.101' (ED25519) to the list of known hosts.Authorized users only. All activities may be monitored and reported.
root@192.168.100.101's password: 
keepalived.conf                                                                             100%  910     1.6MB/s  
 

lvs2修改配置文件
 

[root@lvs2 keepalived]# vim keepalived.conf
[root@lvs2 keepalived]# cat keepalived.conf
! Configuration File for keepalivedglobal_defs {router_id LVS2
}vrrp_instance VI_1 {state MASTERinterface ens33virtual_router_id 51priority 90advert_int 1authentication {auth_type PASSauth_pass 1111}virtual_ipaddress {192.168.100.103}
}virtual_server 192.168.100.103 80 {delay_loop 6lb_algo rrlb_kind DRpersistence_timeout 50protocol TCPreal_server 192.168.100.1 80 {weight 1HTTP_GET {url {path /test.html}connect_timeout 3retry 3delay_before_retry 3}}real_server 192.168.100.2 80 {weight 1HTTP_GET {url {path /test.html}connect_timeout 3retry 3delay_before_retry 3}}
}
 

切换到web1主机在html目录下创建新文件并输入内容
 

[root@web1 keepalived]# cd /usr/share/nginx/html/
[root@web1 html]# ls
404.html  50x.html  index.html  nginx-logo.png
[root@web1 html]# vim test.html
[root@web1 html]# cat test.html 
test
 

 再切换到web2主机在html目录下创建新文件并输入内容
 

[root@web2 keepalived]# cd /usr/share/nginx/html/
[root@web2 html]# ls
404.html  50x.html  index.html  nginx-logo.png
[root@web2 html]# vim test.html
[root@web2 html]# cat test.html
test
 

 lvs1、lvs2分别加载内核模块 (ip_vs)。
 

[root@lvs1 keepalived]# modprobe ip_vs
[root@lvs2 keepalived]# modprobe ip_vs 
 

lvs1、lvs2启动Keepalived，查看LVS 规则已生效。
 

[root@lvs1 keepalived]# systemctl start keepalived.service
[root@lvs1 keepalived]# ipvsadm -Ln
IP Virtual Server version 1.2.1 (size=4096)
Prot LocalAddress:Port Scheduler Flags-> RemoteAddress:Port           Forward Weight ActiveConn InActConn
TCP  192.168.100.103:80 rr persistent 50-> 192.168.100.1:80             Route   1      0          0         -> 192.168.100.2:80             Route   1      0          0      [root@lvs2 keepalived]# systemctl start keepalived.service
[root@lvs2 keepalived]# ipvsadm -Ln
IP Virtual Server version 1.2.1 (size=4096)
Prot LocalAddress:Port Scheduler Flags-> RemoteAddress:Port           Forward Weight ActiveConn InActConn
TCP  192.168.100.103:80 rr persistent 50-> 192.168.100.1:80             Route   1      0          0         -> 192.168.100.2:80             Route   1      1          0            
 

流量将通过轮询策略​分发到两个后端 Nginx 节点（192.168.100.1 和 192.168.100.2）
 

[root@lvs1 keepalived]# curl 192.168.100.1/test.html
test
[root@lvs1 keepalived]# curl 192.168.100.2/test.html
test
 

 切换到web1
 

[root@web1 html]# vim /etc/sysctl.conf
[root@web1 html]# sysctl -p
kernel.sysrq = 0
net.ipv4.ip_forward = 0
net.ipv4.conf.all.send_redirects = 0
net.ipv4.conf.default.send_redirects = 0
net.ipv4.conf.all.accept_source_route = 0
net.ipv4.conf.default.accept_source_route = 0
net.ipv4.conf.all.accept_redirects = 0
net.ipv4.conf.default.accept_redirects = 0
net.ipv4.conf.all.secure_redirects = 0
net.ipv4.conf.default.secure_redirects = 0
net.ipv4.icmp_echo_ignore_broadcasts = 1
net.ipv4.icmp_ignore_bogus_error_responses = 1
net.ipv4.conf.all.rp_filter = 1
net.ipv4.conf.default.rp_filter = 1
net.ipv4.tcp_syncookies = 1
kernel.dmesg_restrict = 1
net.ipv6.conf.all.accept_redirects = 0
net.ipv6.conf.default.accept_redirects = 0
net.ipv4.conf.all.arp_ignore = 1
net.ipv4.conf.all.arp_announce = 2
net.ipv4.conf.all.rp_filter = 0
net.ipv4.conf.all.rp_filter = 0
net.ipv4.conf.all.arp_filter = 0
net.ipv4.conf.all.rp_filter = 0
net.ipv4.conf.default.arp_filter = 0
net.ipv4.conf.default.rp_filter = 0
net.ipv4.conf.ens33.arp_filter = 0
net.ipv4.conf.ens33.rp_filter = 0
net.ipv4.conf.lo.arp_filter = 0
net.ipv4.conf.lo.rp_filter = 0###添加临时IP
[root@web1 ~]# ifconfig lo:0 192.168.100.103/32
[root@web1 ~]# ifconfig
ens33: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500inet 192.168.100.1  netmask 255.255.255.0  broadcast 192.168.100.255inet6 fe80::20c:29ff:feb0:774  prefixlen 64  scopeid 0x20<link>ether 00:0c:29:b0:07:74  txqueuelen 1000  (Ethernet)RX packets 19682  bytes 1662245 (1.5 MiB)RX errors 0  dropped 0  overruns 0  frame 0TX packets 35410  bytes 29300791 (27.9 MiB)TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0ens34: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500inet 192.168.72.176  netmask 255.255.255.0  broadcast 192.168.72.255inet6 fe80::20c:29ff:feb0:77e  prefixlen 64  scopeid 0x20<link>ether 00:0c:29:b0:07:7e  txqueuelen 1000  (Ethernet)RX packets 537  bytes 39137 (38.2 KiB)RX errors 0  dropped 0  overruns 0  frame 0TX packets 61  bytes 7656 (7.4 KiB)TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0lo: flags=73<UP,LOOPBACK,RUNNING>  mtu 65536inet 127.0.0.1  netmask 255.0.0.0inet6 ::1  prefixlen 128  scopeid 0x10<host>loop  txqueuelen 1000  (Local Loopback)RX packets 114  bytes 5568 (5.4 KiB)RX errors 0  dropped 0  overruns 0  frame 0TX packets 114  bytes 5568 (5.4 KiB)TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0lo:0: flags=73<UP,LOOPBACK,RUNNING>  mtu 65536inet 192.168.100.103  netmask 0.0.0.0loop  txqueuelen 1000  (Local Loopback)###写路由条目
[root@web1 ~]# route add -host 192.168.100.103 dev lo:0
[root@web1 ~]# route -n
Kernel IP routing table
Destination     Gateway         Genmask         Flags Metric Ref    Use Iface
0.0.0.0         192.168.72.2    0.0.0.0         UG    101    0        0 ens34
192.168.72.0    0.0.0.0         255.255.255.0   U     101    0        0 ens34
192.168.100.0   0.0.0.0         255.255.255.0   U     100    0        0 ens33
192.168.100.103 0.0.0.0         255.255.255.255 UH    0      0        0 lo
 

切换到web2
 

[root@web2 html]# vim /etc/sysctl.conf
[root@web2 html]# sysctl -p
kernel.sysrq = 0
net.ipv4.ip_forward = 0
net.ipv4.conf.all.send_redirects = 0
net.ipv4.conf.default.send_redirects = 0
net.ipv4.conf.all.accept_source_route = 0
net.ipv4.conf.default.accept_source_route = 0
net.ipv4.conf.all.accept_redirects = 0
net.ipv4.conf.default.accept_redirects = 0
net.ipv4.conf.all.secure_redirects = 0
net.ipv4.conf.default.secure_redirects = 0
net.ipv4.icmp_echo_ignore_broadcasts = 1
net.ipv4.icmp_ignore_bogus_error_responses = 1
net.ipv4.conf.all.rp_filter = 1
net.ipv4.conf.default.rp_filter = 1
net.ipv4.tcp_syncookies = 1
kernel.dmesg_restrict = 1
net.ipv6.conf.all.accept_redirects = 0
net.ipv6.conf.default.accept_redirects = 0
net.ipv4.conf.all.arp_ignore = 1
net.ipv4.conf.all.arp_announce = 2
net.ipv4.conf.all.rp_filter = 0
net.ipv4.conf.all.rp_filter = 0
net.ipv4.conf.all.arp_filter = 0
net.ipv4.conf.all.rp_filter = 0
net.ipv4.conf.default.arp_filter = 0
net.ipv4.conf.default.rp_filter = 0
net.ipv4.conf.ens33.arp_filter = 0
net.ipv4.conf.ens33.rp_filter = 0
net.ipv4.conf.lo.arp_filter = 0
net.ipv4.conf.lo.rp_filter = 0###添加临时IP
[root@web2 html]# ifconfig lo:0 192.168.100.103/32
[root@web2 html]# ifconfig
ens33: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500inet 192.168.100.2  netmask 255.255.255.0  broadcast 192.168.100.255inet6 fe80::20c:29ff:fe59:8d4f  prefixlen 64  scopeid 0x20<link>ether 00:0c:29:59:8d:4f  txqueuelen 1000  (Ethernet)RX packets 18102  bytes 1658904 (1.5 MiB)RX errors 0  dropped 0  overruns 0  frame 0TX packets 35844  bytes 31147681 (29.7 MiB)TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0ens34: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500inet 192.168.72.178  netmask 255.255.255.0  broadcast 192.168.72.255inet6 fe80::20c:29ff:fe59:8d59  prefixlen 64  scopeid 0x20<link>ether 00:0c:29:59:8d:59  txqueuelen 1000  (Ethernet)RX packets 566  bytes 41317 (40.3 KiB)RX errors 0  dropped 0  overruns 0  frame 0TX packets 68  bytes 8132 (7.9 KiB)TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0lo: flags=73<UP,LOOPBACK,RUNNING>  mtu 65536inet 127.0.0.1  netmask 255.0.0.0inet6 ::1  prefixlen 128  scopeid 0x10<host>loop  txqueuelen 1000  (Local Loopback)RX packets 0  bytes 0 (0.0 B)RX errors 0  dropped 0  overruns 0  frame 0TX packets 0  bytes 0 (0.0 B)TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0lo:0: flags=73<UP,LOOPBACK,RUNNING>  mtu 65536inet 192.168.100.103  netmask 0.0.0.0loop  txqueuelen 1000  (Local Loopback)###写路由条目
[root@web2 html]# route add -host 192.168.100.103 dev lo:0
[root@web2 html]# route -n
Kernel IP routing table
Destination     Gateway         Genmask         Flags Metric Ref    Use Iface
0.0.0.0         192.168.72.2    0.0.0.0         UG    101    0        0 ens34
192.168.72.0    0.0.0.0         255.255.255.0   U     101    0        0 ens34
192.168.100.0   0.0.0.0         255.255.255.0   U     100    0        0 ens33
192.168.100.103 0.0.0.0         255.255.255.255 UH    0      0        0 lo
 

验证流量分发
 

 

 

故障模拟： 
 

故障1
 

切换到web1，到html目录移走test.html文件，让lvs1、lvs2无法探测到test.html
 

[root@web1 html]# ls
404.html  50x.html  index.html  nginx-logo.png  test.html
[root@web1 html]# mv test.html /opt/
[root@web1 html]# ls
404.html  50x.html  index.html  nginx-logo.png
 

此时回到lvs1，查看messages日志，日志显示 192.168.100.1:80 的 HTTP 检查失败，Keepalived 已将其从 LVS 规则中移除
 

[root@lvs1 ~]# tail -f /var/log/messages
 

 

当前 ipvsadm -Ln 输出仅剩 192.168.100.2:80，验证了节点摘除操作
 

[root@lvs2 ~]# ipvsadm -Ln
IP Virtual Server version 1.2.1 (size=4096)
Prot LocalAddress:Port Scheduler Flags-> RemoteAddress:Port           Forward Weight ActiveConn InActConn
TCP  192.168.100.103:80 rr persistent 50-> 192.168.100.2:80             Route   1      0          0   
 

此时打开浏览器访问192.168.100.103web站点VIP分发到了web2，实现故障转移
 

 

 
 

切换到web2，到html目录移走test.html文件，让lvs1、lvs2无法探测到test.html
 

[root@web2 ~]# cd /usr/share/nginx/html/
[root@web2 html]# ls
404.html  50x.html  index.html  nginx-logo.png  test.html
[root@web2 html]# mv test.html /opt/
[root@web2 html]# ls
404.html  50x.html  index.html  nginx-logo.png
 

此时回到lvs1，查看messages日志，日志显示 192.168.100.2:80 的 HTTP 检查失败，Keepalived 已将其从 LVS 规则中移除
 

[root@lvs1 ~]# tail -f /var/log/messages
 

 

当前 ipvsadm -Ln 输出仅剩 192.168.100.1:80，验证了节点摘除操作
 

[root@lvs1 ~]# ipvsadm -Ln
IP Virtual Server version 1.2.1 (size=4096)
Prot LocalAddress:Port Scheduler Flags-> RemoteAddress:Port           Forward Weight ActiveConn InActConn
TCP  192.168.100.103:80 rr persistent 50-> 192.168.100.1:80             Route   1      0          0     
 

此时打开浏览器访问192.168.100.103web站点VIP分发到了web1
 

 

故障恢复：
 

web1、web2将文件移回html目录 
 

[root@web1 html]# mv /opt/test.html ./
[root@web1 html]# ls
404.html  50x.html  index.html  nginx-logo.png  test.html[root@web2 html]# mv /opt/test.html ./
[root@web2 html]# ls
404.html  50x.html  index.html  nginx-logo.png  test.html
[root@web2 html]# 
 

切换到lvs1，查看messages日志，IP为192.168.100.200的客户端通过SSH登录lvs1，执行命令后立即退出日志，显示无报错
 

 

 查看LVS规则节点添回，实现高可用
 

[root@lvs1 ~]# ipvsadm -Ln
IP Virtual Server version 1.2.1 (size=4096)
Prot LocalAddress:Port Scheduler Flags-> RemoteAddress:Port           Forward Weight ActiveConn InActConn
TCP  192.168.100.103:80 rr persistent 50-> 192.168.100.1:80             Route   1      0          0         -> 192.168.100.2:80             Route   1      1          0    
 

故障2
 

此时在lvs1停止keepalived
 

[root@lvs1 ~]# systemctl stop keepalived.service
 

切换到lvs2，查看到
 

[root@lvs2 ~]# ip a
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00inet 127.0.0.1/8 scope host lovalid_lft forever preferred_lft foreverinet6 ::1/128 scope host noprefixroute valid_lft forever preferred_lft forever
2: ens33: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000link/ether 00:0c:29:9b:f9:1b brd ff:ff:ff:ff:ff:ffinet 192.168.100.101/24 brd 192.168.100.255 scope global noprefixroute ens33valid_lft forever preferred_lft foreverinet 192.168.100.103/32 scope global ens33valid_lft forever preferred_lft foreverinet6 fe80::20c:29ff:fe9b:f91b/64 scope link noprefixroute valid_lft forever preferred_lft forever
3: ens34: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000link/ether 00:0c:29:9b:f9:25 brd ff:ff:ff:ff:ff:ffinet 192.168.72.183/24 brd 192.168.72.255 scope global dynamic noprefixroute ens34valid_lft 1716sec preferred_lft 1716secinet6 fe80::20c:29ff:fe9b:f925/64 scope link noprefixroute valid_lft forever preferred_lft forever
 

再次访问192.168.100.103web网站，实现故障转移
 

 

 

 
 

配置sorry_server
 

新开一台CentOS7.9
 

[root@ding ~]# yum install -y nginx
###开启nginx
[root@ding ~]# systemctl start nginx
###在html目录中index.html文件中写入内容sorry
[root@ding ~]# echo sorry > /usr/share/nginx/html/index.html [root@ding ~]# ip a
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00inet 127.0.0.1/8 scope host lovalid_lft forever preferred_lft foreverinet6 ::1/128 scope host valid_lft forever preferred_lft forever
2: ens33: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP group default qlen 1000link/ether 00:0c:29:b6:ec:db brd ff:ff:ff:ff:ff:ffinet 192.168.100.20624 brd 192.168.72.255 scope global noprefixroute ens33valid_lft forever preferred_lft foreverinet6 fe80::74bb:b5be:6bf1:18e9/64 scope link noprefixroute valid_lft forever preferred_lft forever
3: ens37: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP group default qlen 1000link/ether 00:0c:29:b6:ec:e5 brd ff:ff:ff:ff:ff:ffinet 192.168.72.139/24 brd 192.168.72.255 scope global noprefixroute dynamic ens37valid_lft 1422sec preferred_lft 1422secinet6 fe80::8ce3:19df:bd0:6a62/64 scope link noprefixroute valid_lft forever preferred_lft forever
4: virbr0: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc noqueue state DOWN group default qlen 1000link/ether 52:54:00:a2:49:3b brd ff:ff:ff:ff:ff:ffinet 192.168.122.1/24 brd 192.168.122.255 scope global virbr0valid_lft forever preferred_lft forever
5: virbr0-nic: <BROADCAST,MULTICAST> mtu 1500 qdisc pfifo_fast master virbr0 state DOWN group default qlen 1000link/ether 52:54:00:a2:49:3b brd ff:ff:ff:ff:ff:ff

 

 切换到lvs1主机验证流量分发正常
 

[root@lvs1 ~]# curl 192.168.100.206
sorry
 

此时lvs1主机修改配置文件，添加一条sorry_server
 

[root@lvs1 ~]# vim /etc/keepalived/keepalived.conf
 

 

lvs2主机修改配置文件，同时也添加一条sorry_server
 

[root@lvs2 ~]# vim /etc/keepalived/keepalived.conf
 

 

lvs1、lvs2重启keepalived服务
 

[root@lvs2 ~]# systemctl restart keepalived.service
 

切换到192.168.100.206主机添加临时网卡，并修改sysctl.conf文件添加ARP 抑制并生效
 

[root@ding ~]# vim /etc/sysctl.conf
[root@ding ~]# sysctl -p
kernel.sysrq = 0
net.ipv4.ip_forward = 0
net.ipv4.conf.all.send_redirects = 0
net.ipv4.conf.default.send_redirects = 0
net.ipv4.conf.all.accept_source_route = 0
net.ipv4.conf.default.accept_source_route = 0
net.ipv4.conf.all.accept_redirects = 0
net.ipv4.conf.default.accept_redirects = 0
net.ipv4.conf.all.secure_redirects = 0
net.ipv4.conf.default.secure_redirects = 0
net.ipv4.icmp_echo_ignore_broadcasts = 1
net.ipv4.icmp_ignore_bogus_error_responses = 1
net.ipv4.conf.all.rp_filter = 1
net.ipv4.conf.default.rp_filter = 1
net.ipv4.tcp_syncookies = 1
kernel.dmesg_restrict = 1
net.ipv6.conf.all.accept_redirects = 0
net.ipv6.conf.default.accept_redirects = 0
net.ipv4.conf.all.arp_ignore = 1
net.ipv4.conf.all.arp_announce = 2
net.ipv4.conf.all.rp_filter = 0
net.ipv4.conf.all.rp_filter = 0
net.ipv4.conf.all.arp_filter = 0
net.ipv4.conf.all.rp_filter = 0
net.ipv4.conf.default.arp_filter = 0
net.ipv4.conf.default.rp_filter = 0
net.ipv4.conf.ens33.arp_filter = 0
net.ipv4.conf.ens33.rp_filter = 0
net.ipv4.conf.lo.arp_filter = 0
net.ipv4.conf.lo.rp_filter = 0###添加临时IP
[root@ding ~]# ifconfig lo:0 192.168.100.103/32###添加路由条目
[root@ding ~]# route add -host 192.168.100.103 dev lo:0
 

故障模拟
 

分别在web1、web2上从html目录中移走test.html文件
 

[root@web1 html]# ls
404.html  50x.html  index.html  nginx-logo.png  test.html
[root@web1 html]# mv test.html /opt/
[root@web1 html]# ls
404.html  50x.html  index.html  nginx-logo.png[root@web2 html]# ls
404.html  50x.html  index.html  nginx-logo.png  test.html
[root@web2 html]# mv test.html /opt/
[root@web2 html]# ls
404.html  50x.html  index.html  nginx-logo.png
 

此时切换到lvs1主机上查看日志，成功将sorry_server添加进LVS
 

[root@lvs1 ~]# tail -f /var/log/messages
 

 

此时查看LVS规则，查到添加了sorry_server的IP
 

[root@lvs1 ~]# ipvsadm -Ln
IP Virtual Server version 1.2.1 (size=4096)
Prot LocalAddress:Port Scheduler Flags-> RemoteAddress:Port           Forward Weight ActiveConn InActConn
TCP  192.168.100.103:80 rr persistent 50-> 192.168.100.206:80           Route   1      0          0   
 

此时访问192.168.100.103web网站，显示出sorry字段，实现故障切换
 

 

 

LVS-DR双主模式配置 
 

切换到lvs1主机修改配置文件
 

[root@lvs1 ~]# vim /etc/keepalived/keepalived.conf
[root@lvs1 ~]# cat /etc/keepalived/keepalived.conf
! Configuration File for keepalivedglobal_defs {router_id LVS1
}vrrp_instance VI_1 {state MASTERinterface ens33virtual_router_id 51priority 100advert_int 1authentication {auth_type PASSauth_pass 1111}virtual_ipaddress {192.168.100.103}
}virtual_server 192.168.100.103 80 {delay_loop 6lb_algo rrlb_kind DRpersistence_timeout 50protocol TCPsorry_server 192.168.100.206 80real_server 192.168.100.1 80 {weight 1HTTP_GET {url {path /test.html}connect_timeout 3retry 3delay_before_retry 3}}real_server 192.168.100.2 80 {weight 1HTTP_GET {url {path /test.html}connect_timeout 3retry 3delay_before_retry 3}}
}
vrrp_instance VI_2 {state BACKUPinterface ens33virtual_router_id 52priority 90advert_int 1authentication {auth_type PASSauth_pass 1111}virtual_ipaddress {192.168.100.104}
}virtual_server 192.168.100.104 80 {delay_loop 6lb_algo rrlb_kind DRpersistence_timeout 50protocol TCPsorry_server 192.168.100.206 80real_server 192.168.100.1 80 {weight 1HTTP_GET {url {path /test.html}connect_timeout 3retry 3delay_before_retry 3}}real_server 192.168.100.2 80 {weight 1HTTP_GET {url {path /test.html}connect_timeout 3retry 3delay_before_retry 3}}
}
 

lvs2主机修改配置文件
 

[root@lvs2 ~]# vim /etc/keepalived/keepalived.conf
[root@lvs2 ~]# cat /etc/keepalived/keepalived.conf
! Configuration File for keepalivedglobal_defs {router_id LVS2
}vrrp_instance VI_1 {state BACKUPinterface ens33virtual_router_id 51priority 90advert_int 1authentication {auth_type PASSauth_pass 1111}virtual_ipaddress {192.168.100.103}
}virtual_server 192.168.100.103 80 {delay_loop 6lb_algo rrlb_kind DRpersistence_timeout 50protocol TCPsorry_server 192.168.100.206 80real_server 192.168.100.1 80 {weight 1HTTP_GET { url {path /test.html}connect_timeout 3retry 3delay_before_retry 3}}real_server 192.168.100.2 80 {weight 1HTTP_GET {url {path /test.html}connect_timeout 3retry 3delay_before_retry 3}}
}
vrrp_instance VI_2 {state MASTERinterface ens33virtual_router_id 52priority 100advert_int 1authentication {auth_type PASSauth_pass 1111}virtual_ipaddress {192.168.100.104}
}virtual_server 192.168.100.104 80 {delay_loop 6lb_algo rrlb_kind DRpersistence_timeout 50protocol TCPsorry_server 192.168.100.206 80real_server 192.168.100.1 80 {weight 1HTTP_GET { url {path /test.html}connect_timeout 3retry 3delay_before_retry 3}}real_server 192.168.100.2 80 {weight 1HTTP_GET {url {path /test.html}connect_timeout 3retry 3delay_before_retry 3}}
}
 

lvs1、lvs2重启keepalived
 

[root@lvs1 ~]# systemctl restart keepalived
[root@lvs2 ~]# systemctl restart keepalived
 

lvs1 输入ip a命令输出查看ens33接口VIP绑定192.168.100.103
 

[root@lvs1 ~]# ipvsadm -Ln
IP Virtual Server version 1.2.1 (size=4096)
Prot LocalAddress:Port Scheduler Flags-> RemoteAddress:Port           Forward Weight ActiveConn InActConn
TCP  192.168.100.103:80 rr persistent 50-> 192.168.100.1:80             Route   1      0          0         -> 192.168.100.2:80             Route   1      0          0         
TCP  192.168.100.104:80 rr persistent 50-> 192.168.100.1:80             Route   1      0          0         -> 192.168.100.2:80             Route   1      0          0         
[root@lvs1 ~]# ip a
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00inet 127.0.0.1/8 scope host lovalid_lft forever preferred_lft foreverinet6 ::1/128 scope host noprefixroute valid_lft forever preferred_lft forever
2: ens33: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000link/ether 00:0c:29:c2:74:d2 brd ff:ff:ff:ff:ff:ffinet 192.168.100.100/24 brd 192.168.100.255 scope global noprefixroute ens33valid_lft forever preferred_lft foreverinet 192.168.100.103/32 scope global ens33valid_lft forever preferred_lft foreverinet6 fe80::20c:29ff:fec2:74d2/64 scope link noprefixroute valid_lft forever preferred_lft forever
3: ens36: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000link/ether 00:0c:29:c2:74:dc brd ff:ff:ff:ff:ff:ffinet 192.168.72.181/24 brd 192.168.72.255 scope global dynamic noprefixroute ens36valid_lft 1648sec preferred_lft 1648secinet6 fe80::20c:29ff:fec2:74dc/64 scope link noprefixroute valid_lft forever preferred_lft forever
 

lvs2 输入ip a命令输出查看ens33接口VIP绑定192.168.100.104
 

[root@lvs2 ~]# ipvsadm -Ln
IP Virtual Server version 1.2.1 (size=4096)
Prot LocalAddress:Port Scheduler Flags-> RemoteAddress:Port           Forward Weight ActiveConn InActConn
TCP  192.168.100.103:80 rr persistent 50-> 192.168.100.1:80             Route   1      0          0         -> 192.168.100.2:80             Route   1      0          0         
TCP  192.168.100.104:80 rr persistent 50-> 192.168.100.1:80             Route   1      0          0         -> 192.168.100.2:80             Route   1      0          0         
[root@lvs2 ~]# ip a
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00inet 127.0.0.1/8 scope host lovalid_lft forever preferred_lft foreverinet6 ::1/128 scope host noprefixroute valid_lft forever preferred_lft forever
2: ens33: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000link/ether 00:0c:29:9b:f9:1b brd ff:ff:ff:ff:ff:ffinet 192.168.100.101/24 brd 192.168.100.255 scope global noprefixroute ens33valid_lft forever preferred_lft foreverinet 192.168.100.104/32 scope global ens33valid_lft forever preferred_lft foreverinet6 fe80::20c:29ff:fe9b:f91b/64 scope link noprefixroute valid_lft forever preferred_lft forever
3: ens34: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000link/ether 00:0c:29:9b:f9:25 brd ff:ff:ff:ff:ff:ffinet 192.168.72.183/24 brd 192.168.72.255 scope global dynamic noprefixroute ens34valid_lft 1194sec preferred_lft 1194secinet6 fe80::20c:29ff:fe9b:f925/64 scope link noprefixroute valid_lft forever preferred_lft forever
 

打开windows本机进入到C:\Windows\System32\drivers\etc目录中找到hosts文件拖到桌面进行修改
 

 

修改完后保存文件再拖入etc目录
 

 

通过相同的域名 www.c2505.com 访问，浏览器书签分别显示 web1 和 web2，验证请求已成功通过不同 LVS 节点（如 lvs1 和 lvs2）分发到后端 web1 和 web2，此时LVS 的轮询（rr）调度策略生效，实现高可用性
 

 

 

 

总结：
 

        Keepalived通过主备节点间的状态监控与自动切换，实现了服务的高可用性，其优势在于配置灵活、资源占用低且与Linux生态无缝集成。双机热备架构不仅提升了系统的容错能力，还能结合Nginx、LVS等组件构建更复杂的负载均衡方案。然而，实际部署时需注意脑裂风险、网络延迟等潜在问题，通过合理的健康检测策略和日志监控，方能最大化发挥Keepalived的价值。在业务云化与分布式趋势下，Keepalived仍是传统架构中保障关键服务稳定性的可靠选择

				
				

					
					http://www.dtcms.com/a/280410.html
					
				

                

                  
相关文章：

                    
                      
  • 
                        基于深度学习的LSTM、GRU对大数据交通流量分析与预测的研究
                      
    • 
                    
                      
  • 
                        omniparser v2 本地部署及制作docker镜像（20250715）
                      
    • 
                    
                      
  • 
                        从浏览器到服务器：TCP 段的网络传输之旅
                      
    • 
                    
                      
  • 
                        设计模式二：策略模式 (Strategy Pattern)
                      
    • 
                    
                      
  • 
                        云计算如何提高企业的数据安全性和隐私保护
                      
    • 
                    
                      
  • 
                        我会秘书长杨添天带队赴杭州融量农业发展有限公司考察调研
                      
    • 
                    
                      
  • 
                        NQTT-基础知识
                      
    • 
                    
                      
  • 
                        CSS :root伪类详解：实现动态主题切换的关键所在
                      
    • 
                    
                      
  • 
                        7.15 Java基础|大小写转换、数组、ArrayList类
                      
    • 
                    
                      
  • 
                        基于Langchain4j开发AI编程助手
                      
    • 
                    
                      
  • 
                        Python_1
                      
    • 
                    
                      
  • 
                        高等数学强化——导学
                      
    • 
                    
                      
  • 
                        【Python练习】044. 编写一个函数，实现快速排序算法
                      
    • 
                    
                      
  • 
                        第十三讲 | map和set的使用
                      
    • 
                    
                      
  • 
                        JavaDemo——使用CGLIB动态代理
                      
    • 
                    
                      
  • 
                        I3C通信驱动开发注意事项
                      
    • 
                    
                      
  • 
                        【雅思播客016】New Year Resolution 新年决心
                      
    • 
                    
                      
  • 
                        docker搭建freeswitch实现点对点视频，多人视频
                      
    • 
                    
                      
  • 
                        极致cms多语言建站|设置主站默认语言与设置后台固定语言为中文
                      
    • 
                    
                      
  • 
                        嵌入式学习-PyTorch（4）-day21
                      
    • 
                    
                      
  • 
                        多相机depth-rgb图组完整性分拣器_MATLAB实现
                      
    • 
                    
                      
  • 
                        @[TOC](模拟) # 1.替换所有的问号（easy）
                      
    • 
                    
                      
  • 
                        学C++做游戏，先搞懂这些基础要点
                      
    • 
                    
                      
  • 
                        《大数据技术原理与应用》实验报告六 Flink编程实践
                      
    • 
                    
                      
  • 
                        使用JS编写用户信息采集表单
                      
    • 
                    
                      
  • 
                        【Python3-Django】快速掌握DRF：ModelViewSet实战指南
                      
    • 
                    
                      
  • 
                        OneCode 3.0 从0到1干货——AIGC及MCP注解驱动开发物联网AI决策应用
                      
    • 
                    
                      
  • 
                        全新 Python 项目托管到 Gitee 私有仓库完整流程（带详细命令注释）
                      
    • 
                    
                      
  • 
                        OpenVINO initialization error: Failed to find plugins.xml file
                      
    • 
                    
                      
  • 
                        uv 使用指导文档
                      
    •