keepalived热备应用 - 无聊の时光

一、keepalived概述

概述

keepalived起初为lvs设计的一款强大的辅助工具，保证lvs负载调度器的故障切换以及web节点的健康状态检查，后续被应用到很多需要容错的场景下，keepalived自身基于VRRP协议--虚拟路由冗余协议，思科共有协议；

二、设计原理

应用场景
设计模块
- core模块
  为keepalived的核心组件，负责主进程的启动、维护及全局配置文件的加载和解析
- check模块
  负责real server节点池的节点的健康检测
- VRRP模块在master与backup之间执行新调检测
热备实现过程
将多个主机以软件的方式组成一个热备足，通过共有虚拟ip(VIP)地址对外提供服务，同一时刻，热备足中只有一台主机在工作，别的主机冗余状态，当当前在线的主机失效时，其他冗余的主机会自动阶梯虚拟主机ip地址，继续提供服务，以保证架构的稳定性；

三、keepalived实现双机热备

系统类型	IP地址	主机名	所需软件
centos7.8	192.168.100.105	node1.linux.com	keepalived-1.2.13.tar.gz
centos7.8	192.168.100.106	node2.linux.com	keepalived-1.2.13.tar.gz

安装node1节点上的httpd服务

[root@node1 ~]# yum -y install httpd
[root@node1 ~]# cat <<END >>/var/www/html/index.html
192.168.100.105
END
[root@node1 ~]# systemctl start httpd
[root@node1 ~]# systemctl enable httpd
Created symlink from /etc/systemd/system/multi-user.target.wants/httpd.service to /usr/lib/systemd/system/httpd.service.
[root@node1 ~]# netstat -utpln |grep 80
tcp        0      0 0.0.0.0:80              0.0.0.0:*               LISTEN      1217/httpd

安装node2节点上的httpd服务

[root@node1 ~]# yum -y install httpd
[root@node1 ~]# cat <<END >>/var/www/html/index.html
192.168.100.106
END
[root@node1 ~]# systemctl start httpd
[root@node1 ~]# systemctl enable httpd
Created symlink from /etc/systemd/system/multi-user.target.wants/httpd.service to /usr/lib/systemd/system/httpd.service.
[root@node1 ~]# netstat -utpln |grep 80
tcp        0      0 0.0.0.0:80              0.0.0.0:*               LISTEN      1217/httpd

在两台node节点上安装keepalived软件程序（两台安装步骤一致，在此只列出一台）

[root@node1 ~]# yum -y install kernel-devel openssl-devel popt-devel        ##安装内核开发包，popt支持库等工具
[root@node1 ~]# tar -zxvf keepalived-1.2.13.tar.gz -C /usr/src/
[root@node1 ~]# cd /usr/src/keepalived-1.2.13/
[root@node1 keepalived-1.2.13]# ./configure --prefix=/usr/local/keepalived 
[root@node1 keepalived-1.2.13]# make &&make install
[root@node1 keepalived-1.2.13]# cd
[root@node1 ~]# mkdir -p /etc/keepalived        ##程序的主配置目录
[root@node1 ~]# cp /usr/local/keepalived/etc/keepalived/keepalived.conf /etc/keepalived/        ## 复制主配置文件
[root@node1 ~]# cp /usr/local/keepalived/etc/sysconfig/keepalived /etc/sysconfig/       ##复制启动时需要加载的配置文件
[root@node1 ~]# cp /usr/local/keepalived/etc/rc.d/init.d/keepalived /etc/init.d/        ##复制服务的控制脚本
[root@node1 ~]# cp /usr/local/keepalived/sbin/keepalived /usr/sbin/     ##复制keepalived的命令
[root@node1 ~]# chmod 755 /etc/init.d/keepalived        ##为指控脚本指定权限

配置node1上master主节点

[root@node1 ~]#  vi /etc/keepalived/keepalived.conf
global_defs {
    router_id HA_TEST_R1                ##本服务器的名称，若环境中有多个keepalived时，此名称不能一致
}
vrrp_instance VI_1 {                    ##定义VRRP热备实例，每一个keep组都不同
    state MASTER                                ##MASTER表示主服务器
    interface eth0                              ##承载VIP地址的物理接口
    virtual_router_id 1                 ##虚拟路由器的ID号，每一个keep组都不同
    priority 100                                ##优先级，数值越大优先级越高
    advert_int 1                                ##通告检查间隔秒数（心跳频率）
    authentication {                    ##认证信息
        auth_type PASS          ##认证类型
        auth_pass 123456                ##密码字串
    }
    virtual_ipaddress {
  192.168.100.95                                ##指定漂移地址（VIP）
    }
}
virtual_server 192.168.100.95 80 {       #vip配置
    delay_loop 2                                 #每隔2秒检查一次real_server状态
    #lb_algo wrr                    ##指定lvs的调度算法
    #lb_kind DR                 ##lvs集群模式（如若不结合LVS请删掉两个配置）
    persistence_timeout 60              ##会话保持时间
    protocol TCP                                        ##选择协议
    real_server 192.168.100.105 80 {             ##本机地址
        weight :3                               ##服务器的权重
        notify_down /etc/keepalived/check.sh            ##指定节点失效后，采用的脚本，notify_up表示节点正常后，采用的脚本
                ##健康检查方式一共有HTTP_GET|SSL_GET|TCP_CHECK|SMTP_CHECK|MISC_CHECK这些
        TCP_CHECK {
            connect_timeout 10          ##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 80         ##健康检查端口
        }
    }
   }
[root@node1 ~]# vi /etc/keepalived/check.sh
#!/bin/bash
/etc/init.d/keepalived stop
echo -e "$(ip a |grep eth0 |grep inet |awk '{print $2}'|awk -F'/' '{print $1}') (httpd) is down on $(date +%F-%T)" >>/root/check_httpd.log
:wq
[root@node1 ~]# chmod 777 /etc/keepalived/check.sh
[root@node1 ~]# /etc/init.d/keepalived start
Starting keepalived (via systemctl):                       [  确定  ]
[root@node1 ~]# ip a |grep 192.168.100.95
    inet 192.168.100.95/32 scope global eth0

配置node2上backup从节点

[root@node2 ~]# vi /etc/keepalived/keepalived.conf
global_defs {
    router_id HA_TEST_R2            ##本服务器的名称
}
vrrp_instance VI_1 {
    state BACKUP                    ##BACKUP表示从服务器
    interface eth0
    virtual_router_id 1
    priority 99                 ##优先级，低于主服务器
    advert_int 1
    authentication {
        auth_type PASS
        auth_pass 123456
    }
    virtual_ipaddress {
        192.168.100.95
    }
}
virtual_server 192.168.100.95 80 {   ##vip配置
    delay_loop 2                     ##每隔2秒检查一次real_server状态
    #lb_algo wrr
    #lb_kind DR                 ##如若不结合LVS，需要去掉两个配置项
    persistence_timeout 60           ##会话保持时间
    protocol TCP
    real_server 192.168.100.106 80 {    ##本机地址
        weight :3
        notify_down /etc/keepalived/check.sh
        TCP_CHECK {
            connect_timeout 10          ##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 80             ##健康检查端口
        }
    }
   }
:wq
[root@node2 ~]# vi /etc/keepalived/check.sh
#!/bin/bash
/etc/init.d/keepalived stop
echo -e "$(ip a |grep eth0 |grep inet |awk '{print $2}'|awk -F'/' '{print $1}') (httpd) is down on $(date +%F-%T)" >>/root/check_httpd.log
:wq
[root@node2 ~]# chmod 777 /etc/keepalived/check.sh
[root@node2 ~]# /etc/init.d/keepalived start

客户端访问测试双机热备的效果
当node1与node2均正常状态

当node1的httpd服务关闭

[root@node1 ~]# systemctl stop httpd
[root@node1 ~]# /etc/init.d/keepalived status
¡ñ keepalived.service - SYSV: Start and stop Keepalived
   Loaded: loaded (/etc/rc.d/init.d/keepalived; bad; vendor preset: disabled)
   Active: inactive (dead)
     Docs: man:systemd-sysv-generator(8)

8ÔÂ 24 21:28:57 node1.linux.com Keepalived_healthcheckers[2860]: Netlink reflector reports IP 192.168.100.95 added
8ÔÂ 24 21:29:02 node1.linux.com Keepalived_vrrp[2861]: VRRP_Instance(VI_1) Sending gratuitous ARPs on eth0 for 192.168.100.95
8ÔÂ 24 21:36:44 node1.linux.com Keepalived_healthcheckers[2860]: TCP connection to [192.168.100.105]:80 failed !!!
8ÔÂ 24 21:36:44 node1.linux.com Keepalived_healthcheckers[2860]: Removing service [192.168.100.105]:80 from VS [192.168.100.95]:80
8ÔÂ 24 21:36:44 node1.linux.com Keepalived_healthcheckers[2860]: IPVS: Service not defined
8ÔÂ 24 21:36:44 node1.linux.com Keepalived_healthcheckers[2860]: Executing [/etc/keepalived/check.sh] for service [192.168.100.105]:80 in VS [192.168.100.95]:80
8ÔÂ 24 21:36:44 node1.linux.com Keepalived_healthcheckers[2860]: Lost quorum 1-0=1 > 0 for VS [192.168.100.95]:80
8ÔÂ 24 21:36:44 node1.linux.com systemd[1]: Stopping SYSV: Start and stop Keepalived...
8ÔÂ 24 21:36:44 node1.linux.com keepalived[2926]: Stopping keepalived: [  确定 ]
8ÔÂ 24 21:36:44 node1.linux.com systemd[1]: Stopped SYSV: Start and stop Keepalived.

将node1的httpd和keepalived服务重新启动

[root@node1 ~]# systemctl start httpd
[root@node1 ~]# /etc/init.d/keepalived start
Starting keepalived (via systemctl):                       [  确定  ]
[root@node1 ~]# /etc/init.d/keepalived status
¡ñ keepalived.service - SYSV: Start and stop Keepalived
   Loaded: loaded (/etc/rc.d/init.d/keepalived; bad; vendor preset: disabled)
   Active: active (running) since Èý 2022-08-24 21:38:10 CST; 4s ago
     Docs: man:systemd-sysv-generator(8)
  Process: 2979 ExecStart=/etc/rc.d/init.d/keepalived start (code=exited, status=0/SUCCESS)
 Main PID: 2986 (keepalived)
   CGroup: /system.slice/keepalived.service
           ©À©¤2986 keepalived -D
           ©À©¤2988 keepalived -D
           ©¸©¤2989 keepalived -D

8ÔÂ 24 21:38:11 node1.linux.com Keepalived_healthcheckers[2988]: IPVS: Scheduler or persistence engine not found
8ÔÂ 24 21:38:11 node1.linux.com Keepalived_healthcheckers[2988]: IPVS: No such process
8ÔÂ 24 21:38:11 node1.linux.com Keepalived_healthcheckers[2988]: Using LinkWatch kernel netlink reflector...
8ÔÂ 24 21:38:11 node1.linux.com Keepalived_healthcheckers[2988]: Activating healthchecker for service [192.168.100.105]:80
8ÔÂ 24 21:38:11 node1.linux.com Keepalived_vrrp[2989]: VRRP_Instance(VI_1) Transition to MASTER STATE
8ÔÂ 24 21:38:11 node1.linux.com Keepalived_vrrp[2989]: VRRP_Instance(VI_1) Received lower prio advert, forcing new election
8ÔÂ 24 21:38:12 node1.linux.com Keepalived_vrrp[2989]: VRRP_Instance(VI_1) Entering MASTER STATE
8ÔÂ 24 21:38:12 node1.linux.com Keepalived_vrrp[2989]: VRRP_Instance(VI_1) setting protocol VIPs.
8ÔÂ 24 21:38:12 node1.linux.com Keepalived_vrrp[2989]: VRRP_Instance(VI_1) Sending gratuitous ARPs on eth0 for 192.168.100.95
8ÔÂ 24 21:38:12 node1.linux.com Keepalived_healthcheckers[2988]: Netlink reflector reports IP 192.168.100.95 added

将node1节点的httpd和keepalived服务重新启动后，发现node1节点重新抢占了master主节点，如若不想实现node节点自动抢占，可在主从节点配置文件中修改如下：