LVS+Keepalived+HAProxy双主高可用负载均衡集群

FJW0208142026-03-24 8:21

实验说明

  • LVS+Keeplived(接受四层流量) -> HAProxy(进行七层处理) -> Nginx(web业务或者反向代理应用业务)
  • NFS可作为静态资源的挂载与实现后端业务的统一，可使用DRBD实现NFS的高可用，后续可添加Redis与Mysql来存储更复杂数据或实现会话保持
  • DNS轮询是为了实现一个大流量业务Keepalived双主热备下保证数据的均匀分发，依托双主VIP架构打破传统单主备机闲置弊端，两台负载均衡节点同时在线承载流量，硬件资源利用率最大化

实验架构图

实验环境

主机名 IP地址 OS 所需软件
LB1 172.25.254.50 RockyLinux9.6 Keepalived，HAProxy，LVS
LB2 172.25.254.60 RockyLinux9.6 Keepalived，HAProxy，LVS
RS1 172.25.254.10 RockyLinux9.6 Nginx，nfs-utils
RS2 172.25.254.20 RockyLinux9.6 Nginx，nfs-utils
NFS-DNS 172.25.254.66 RockyLinux9.6 nfs-utils，bind
client 172.25.254.67 RockyLinux9.6

DNS服务部署

bash 复制代码 
[root@nfs-dns ~]# dnf install bind -y

[root@nfs-dns ~]# vim /etc/named.conf
options {
        listen-on port 53 { any; };
...
        allow-query     { any; };		
        forwarders      { 8.8.8.8;}；
...
zone "web.com" IN {
        type master;
        file "named.web.com";
};

[root@nfs-dns ~]# cp -p /var/named/named.localhost /var/named/named.web.com
[root@nfs-dns ~]# cat /var/named/named.web.com
$TTL 1D
@       IN SOA  dns.web.com. admin.web.com. (
                                        0       ; serial
                                        1D      ; refresh
                                        1H      ; retry
                                        1W      ; expire
                                        3H )    ; minimum
        NS      dns.web.com.
dns     A       127.0.0.1
www     A       172.25.254.100
www     A       172.25.254.200

[root@nfs-dns ~]# systemctl enable --now named

#测试
[root@nfs-dns ~]# nslookup www.web.com
Server:         ::1
Address:        ::1#53

Name:   www.web.com
Address: 172.25.254.100
Name:   www.web.com
Address: 172.25.254.200

NFS服务部署

bash 复制代码 
[root@nfs-dns ~]# dnf install nfs-utils -y
[root@nfs-dns ~]# mkdir -p /data/web 
[root@nfs-dns ~]# echo '/data/web *(rw,no_root_squash,no_all_squash,sync)' > /etc/exports

[root@nfs-dns ~]# systemctl enable --now nfs-server.service

#测试
[root@nfs-dns ~]# exportfs -rv
exporting *:/data/web

#web节点进行挂载
[root@RS1 ~]# dnf install nginx -y
[root@RS2 ~]# dnf install nginx -y

[root@RS1 ~]# dnf install nfs-utils -y
[root@RS2 ~]# dnf install nfs-utils -y

[root@RS1 ~]# mount 172.25.254.66:/data/web /usr/share/nginx/html/
[root@RS2 ~]# mount 172.25.254.66:/data/web /usr/share/nginx/html/

[root@RS1 ~]# echo "172.25.254.66:/data/web /usr/share/nginx/html/ nfs defaults 0 0" >> /etc/fstab
[root@RS2 ~]# echo "172.25.254.66:/data/web /usr/share/nginx/html/ nfs defaults 0 0" >> /etc/fstab

HAProxy部署

bash 复制代码 
[root@LB1 ~]# dnf install haproxy -y
[root@LB2 ~]# dnf install haproxy -y

#LB1与LB2
[root@LB1 ~]# vim /etc/haproxy/haproxy.cfg
......
frontend http_front
        bind            *:80
        default_backend web_servers
        
backend web_servers
        balance roundrobin
        server web01 172.25.254.10:80 check inter 3000 fall 3 rise 2
        server web02 172.25.254.20:80 check inter 3000 fall 3 rise 2

[root@LB1 ~]# systemctl enable --now haproxy.service

LVS+Keepalived部署

安装软件

bash 复制代码 
[root@LB1 ~]# dnf install keepalived ipvsadm -y
[root@LB2 ~]# dnf install keepalived ipvsadm -y

DR模式配置VIP

由于使用的是DR模式，所以要给RS配置VIP

bash 复制代码 
[root@RS1 ~]# cd /etc/NetworkManager/system-connections/
[root@RS1 system-connections]# cp -p eth0.nmconnection lo.nmconnection
[root@RS1 system-connections]# cat > lo.nmconnection << EOF
[connection]
id=lo
type=loopback
interface-name=lo

[ipv4]
method=manual
address1=127.0.0.1/8
address2=172.25.254.100/32
address3=172.25.254.200/32
EOF

[root@RS1 system-connections]# nmcli connection reload
[root@RS1 system-connections]# nmcli connection up lo
[root@RS1 system-connections]# ip a			#查看VIP是否添加成功


[root@RS2 ~]# cd /etc/NetworkManager/system-connections/
[root@RS2 system-connections]# cp -p eth0.nmconnection lo.nmconnection
[root@RS2 system-connections]# cat > lo.nmconnection << EOF
[connection]
id=lo
type=loopback
interface-name=lo

[ipv4]
method=manual
address1=127.0.0.1/8
address2=172.25.254.100/32
address3=172.25.254.200/32
EOF

[root@RS2 system-connections]# nmcli connection reload
[root@RS2 system-connections]# nmcli connection up lo
[root@RS2 system-connections]# ip a

RS配置arp抑制

bash 复制代码 
[root@RS1 system-connections]# cat >> /etc/sysctl.conf <<EOF
net.ipv4.conf.all.arp_ignore = 1
net.ipv4.conf.lo.arp_ignore = 1
net.ipv4.conf.all.arp_announce = 2
net.ipv4.conf.lo.arp_announce = 2
EOF
[root@RS1 system-connections]# sysctl -p

[root@LB2 system-connections]# cat >> /etc/sysctl.conf <<EOF
net.ipv4.conf.all.arp_ignore = 1
net.ipv4.conf.lo.arp_ignore = 1
net.ipv4.conf.all.arp_announce = 2
net.ipv4.conf.lo.arp_announce = 2
EOF
[root@LB2 system-connections]# sysctl -p

配置Keepalived与LVS

bash 复制代码 
[root@LB1 ~]# vim /etc/keepalived/keepalived.conf
! Configuration File for keepalived

global_defs {
   router_id LB1
}

vrrp_instance VI_1 {
    state MASTER
    interface eth0
    virtual_router_id 51
    priority 150
    advert_int 1
    authentication {
        auth_type PASS
        auth_pass 1111
    }
    virtual_ipaddress {
        172.25.254.100/24
    }
}

vrrp_instance VI_2 {
    state BACKUP
    interface eth0
    virtual_router_id 52
    priority 100
    advert_int 1
    authentication {
        auth_type PASS
        auth_pass 1111
    }
    virtual_ipaddress {
        172.25.254.200/24
    }
}

virtual_server 172.25.254.100 80 {
    delay_loop 6
    lb_algo rr
    lb_kind DR
    protocol TCP

    real_server 127.0.0.1 80 {
        weight 1
        TCP_CHECK {
            connect_timeout 3
            retry 3
            delay_before_retry 3
            connect_port 80
        }
    }
}

virtual_server 172.25.254.200 80 {
    delay_loop 6
    lb_algo rr
    lb_kind DR
    protocol TCP

    real_server 127.0.0.1 80 {
        weight 1
        TCP_CHECK {
            connect_timeout 3
            retry 3
            delay_before_retry 3
            connect_port 80
        }
    }
}

systemctl enable --now keepalived

lb02（主 VIP2，备 VIP1）

只需要改：

  • VI_1：state BACKUP priority 100
  • VI_2：state MASTER priority 150
  • router_id LB2

测试

测试业务能否通过DNS轮询解析到域名访问

bash 复制代码 
[root@client ~]# cat /etc/resolv.conf
# Generated by NetworkManager
nameserver 172.25.254.66

[root@client ~]# nslookup www.web.com
Server:         172.25.254.66
Address:        172.25.254.66#53

Name:   www.web.com
Address: 172.25.254.100
Name:   www.web.com
Address: 172.25.254.200

[root@nfs-dns ~]# echo "Welcome to Web Server" > /data/web/index.html
[root@client ~]# curl www.web.com
Welcome to Web Server

测试VIP是否漂移

bash 复制代码 
[root@LB2 ~]# systemctl stop keepalived.service
相关推荐
乘云数字DATABUFF2 天前
5分钟部署开源APM Databuff：OpenTelemetry全链路追踪入门实战
运维·后端
荣--4 天前
一键部署不是为了省时间 —— 它是把"买来的 PaaS"变成"自己的平台"的拐点
运维·zabbix·工程化·一键部署·平台化·边界设计
江华森4 天前
动手实战学 Docker — 从零到集群编排完全指南
运维
Avan_菜菜4 天前
FRP 内网穿透完整实战：从 HTTP 映射到 HTTPS 自签代理
运维·nginx·https
SelectDB5 天前
Litefuse 开源并推出单进程轻量模式，25 秒就能跑起来的 Agent 可观测与评估平台
运维·后端·自动化运维
XIAOHEZIcode7 天前
Linux系统鼠标偏移常见原因以及修复方案
linux·运维·游戏
用户0328472220707 天前
如何搭建本地yum源(上)
运维
大树8810 天前
金刚石散热越强，管路越先见顶
大数据·运维·服务器·人工智能·ai
摇滚侠10 天前
Linux CentOS7 rpm 安装 MySQL 5.7
linux·运维·mysql
霸道流氓气质10 天前
领域驱动设计（DDD）在 Spring Boot 微服务中的实践指南
运维·spring boot·微服务
热门推荐
012026年6月AI大模型全景报告：GPT-5.6、Claude Opus 4.8、Gemini 3.5，中美AI三足鼎立谁主沉浮？022026年6月AI行业全景：从百模大战到Agent元年，这30天发生了什么？032026 年 AI 编程工具终极横评：Cursor vs Claude Code vs Copilot vs Windsurf04飞书长连接_事件订阅（接收消息，审批任务状态变更）05Trae国际版与国内版深度测评：AI原生IDE的双生花06GitHub 镜像站点07【AI】2026 年具身智能模型和世界模型总结08Claude Code、Codex、Cursor三分天下：2026年AI编程Agent生态全景剖析092026年AI架构实战：彻底解决OpenAI接口超时与封号，Python调用GPT-5.2/Sora2企业级架构详解（附源码+压测报告）10【AI总结】2026年6月 主流国内外大模型总结