2022-09-15  01d5de603aaf0daf0640ba193e0f9844 99+   37 分钟   5.5 k0次访问

关于Linux中Keepalived高可用热备自动化部署的一些笔记

傍晚时分，你坐在屋檐下，看着天慢慢地黑下去，心里寂寞而凄凉，感到自己的生命被剥夺了。当时我是个年轻人，但我害怕这样生活下去，衰老下去。在我看来，这是比死亡更可怕的事。——–王小波

写在前面

  • 今天和小伙伴们分享一些 Keepalived 相关笔记
  • 博文内容涉及：
    • vrrp 协议由来
    • Ansible 方式 Keepalived安装部署
    • Keepalived 脚本方式配置服务检查
    • Keepalived 自动化部署 Ansible 角色编写
  • 食用方式：
    • 需要 Linux 基础知识
    • 需要 Ansible 基础知识
  • 理解不足小伙伴帮忙指正

傍晚时分，你坐在屋檐下，看着天慢慢地黑下去，心里寂寞而凄凉，感到自己的生命被剥夺了。当时我是个年轻人，但我害怕这样生活下去，衰老下去。在我看来，这是比死亡更可怕的事。——–王小波

官网帮助文档：https://www.keepalived.org/manpage.html

关于keepalived是什么，在官网中这样描述：

Keepalived 是一个用 C 语言编写的路由软件。该项目的主要目标是为 Linux 系统和基于 Linux 的基础架构提供简单而强大的负载平衡和高可用性设施。

负载平衡框架依赖于众所周知且广泛使用的Linux 虚拟服务器 (IPVS) 内核模块，提供第 4 层负载平衡。Keepalived 实现了一组检查器，以根据其健康状况动态和自适应地维护和管理负载平衡的服务器池。(本文不涉及)

高可用性是通过 VRRP实现的协议。VRRP 是路由器故障转移的基础。此外，Keepalived 实现了一组与 VRRP 有限状态机的挂钩，提供低级和高速协议交互。为了提供最快的网络故障检测，Keepalived 实现了BFD协议。VRRP 状态转换可以考虑 BFD 提示来驱动快速状态转换。Keepalived 框架可以单独使用，也可以一起使用，以提供弹性基础架构

Keepalived 是免费软件；您可以根据自由软件基金会发布的 GNU 通用公共许可条款重新分发和/或修改它；许可证的第 2 版，或(由您选择)任何更高版本。

今天和小伙伴分享的主要是高可用热备部署，关于负载均衡方面的之后和小伙伴们分享，在部署keepalived之前，需要了解下VRRP协议

vrrp协议由来

当网关路由器出现故障时，本网段内以该设备为网关的主机都不能与 Internet 进行通信。所以需要进行容灾处理，但是通过部署多网关的方式实现网关的备份,存在一些问题:网关间IP地址冲突；主机会频繁切换网络出口。所以为解决网关路由的单点故障，有了VRRP协议。

VRRP即虚拟路由冗余协议,VRRP能够在不改变组网的情况下，从多台网关设备里产生一个虚拟路由器，通过配置虚拟路由器的IP地址为默认网关，实现网关的备份。

对外提供网关服务的是这个虚拟路由器。这样不管是真实路由器哪个出现问题，都不会影响整个网络的运行，提高了网络结构的稳定性。

在这里插入图片描述

VRRP配置方式：

  • 配置VRRP的成员；
  • 配置VRRP的优先级 (默认100)；
  • 查看VRRP信息

VRRP协议通过一种竞选机制来将路由任务交给某个vrrp路由器的。

在VRRP物理结构中，有多个物理的VRRP路由器，其中有一台称为“master”(主节点路由器)，其他的都是“backup”(备节点路由器)

在VRRP虚拟结构中，虚拟路由都是通过MAC+VRID的形式来标识的，如54-89-98-6F-3D-B5-{vrid}
只有master节点才会发送VRRP包(vrrp advertisement message)当master节点宕掉的时候，backup中优先级最高的VRRP设备会抢占并升级为master

下面为配置的简单Demo

三层交换机SW1上配置, 主路由器(Master)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
#三层交换机SW1上配置, 主路由器(Master)
<Huawei>system-view #进入系统视图
[Huawei]sysname SW1 #修改设备名字
[SW1]
[SW1]undo info-center enable #取消信息提示
#给vlan1配置网关
[SW1]interface Vlanif 1
[SW1-Vlanif1]ip address 192.168.1.252 255.255.255.0
#将三层交换机SW1配置为VRRP的成员，设置虚拟IP地址
[SW1-Vlanif1]vrrp vrid 1 virtual-ip 192.168.1.254
#配置VRRP的优先级，不写默认为100
[SW1-Vlanif1]vrrp vrid 1 priority 105
[SW1-Vlanif1]
#查看VRRP配置信息
[SW1-Vlanif1]display vrrp brief

三层交换机SW2上配置,,备用路由器(backup)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
#三层交换机SW2上配置,,备用路由器(backup)
<Huawei>system-view #进入系统视图
[Huawei]sysname SW2 #修改设备名字
[SW2]
[SW2]undo info-center enable #取消信息提示
#给vlan1配置网关
[SW2]interface Vlanif 1
[SW2-Vlanif1]ip address 192.168.1.253 255.255.255.0
#将三层交换机SW1配置为VRRP的成员，设置虚拟IP地址
[SW2-Vlanif1]vrrp vrid 1 virtual-ip 192.168.1.254
#不用设置优先级，默认为100
#查看VRRP配置信息
[SW2-Vlanif1]display vrrp brief

keepalived 安装部署

回到keepalived中，keepalived 通过VRRP(Virtual Router Redundancy Protocol)虚拟路由冗余协议来实现故障转移。keepalived正常工作时，主节点(master)会不断的发送心跳信息给备节点(backup)

1
2
3
4
5
6
7
8
9
10
11
12
13
┌──[root@vms153.liruilongs.github.io]-[~]
└─$tcpdump -i  ens32 -nn host 224.0.0.18 #组播地址
tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
listening on ens32, link-type EN10MB (Ethernet), capture size 262144 bytes
23:27:36.149062 IP 192.168.26.153 > 224.0.0.18: VRRPv2, Advertisement, vrid 51, prio 100, authtype simple, intvl 1s, length 20
23:27:37.150969 IP 192.168.26.153 > 224.0.0.18: VRRPv2, Advertisement, vrid 51, prio 100, authtype simple, intvl 1s, length 20
23:27:38.152021 IP 192.168.26.153 > 224.0.0.18: VRRPv2, Advertisement, vrid 51, prio 100, authtype simple, intvl 1s, length 20
^C
3 packets captured
3 packets received by filter
0 packets dropped by kernel
┌──[root@vms153.liruilongs.github.io]-[~]
└─$

当备节点在一定时间内没有收到主节点的心跳信息时，备节点会认为主节点宕了，就会接管主节点上的资源，并继续向外提供服务保证其可用性,当主节点恢复时，备节点会自动让出资源并再次自动成为备节点

这里我们通过 ansible 安装配置，下面的两台机器为我们要配置的机器，也就是node组，在master节点操作

  • 192.168.26.153
  • 192.168.26.154
1
2
3
4
5
6
7
┌──[root@vms152.liruilongs.github.io]-[~]
└─$cat inventory
[master]
192.168.26.152
[node]
192.168.26.153
192.168.26.154

安装web服务测试用

1
2
┌──[root@vms152.liruilongs.github.io]-[~]
└─$ansible node -m shell -a "yum -y install httpd"

编写一个小剧本用于环境初始化，这个剧本实现对httpd服务的欢迎页进行内容填充，重启服务，设置防火墙域为trusted，即没有规则

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
┌──[root@vms152.liruilongs.github.io]-[~]
└─$cat httpd.yaml
---
- name: httpd init
  hosts: node
  tasks:
    - name: httpd content
      shell: "echo `hostname` > /var/www/html/index.html"
    - name: Restart service httpd, in all cases
      service:
        name: httpd
        state: restarted
    - name: firewall
      shell: firewall-cmd --set-default-zone=trusted
┌──[root@vms152.liruilongs.github.io]-[~]
└─$

执行剧本并测试填充结果测试

1
2
3
4
5
6
7
8
9
10
11
┌──[root@vms152.liruilongs.github.io]-[~]
└─$ansible-playbook httpd.yaml
........
┌──[root@vms152.liruilongs.github.io]-[~]
└─$ansible node -m shell -a 'hostname;cat /var/www/html/index.html'
192.168.26.154 | CHANGED | rc=0 >>
vms154.liruilongs.github.io
vms154.liruilongs.github.io
192.168.26.153 | CHANGED | rc=0 >>
vms153.liruilongs.github.io
vms153.liruilongs.github.io

安装 keepalived，我们使用的版本为：keepalived-1.3.5-19.el7.x86_64

1
2
3
4
┌──[root@vms152.liruilongs.github.io]-[~]
└─$ansible node -m yum -a 'name=keepalived state=installed'
192.168.26.154 | SUCCESS => {
。。。。。。

编辑配置文件模板，把主备节点配置文件中不一样的，或者希望单独设置的内容做成变量

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
┌──[root@vms152.liruilongs.github.io]-[~]
└─$cat keepalived.conf.j2
! Configuration File for keepalived

global_defs {
   router_id LVS_DEVEL #设置路由ID,可以和主机名相同，也可以随便定义
   vrrp_iptables       #手动添加(禁止设置防火墙规则，keepalved每次启动都会自动添加防火墙拒绝所有的规则)
}

vrrp_instance VI_1 {
    state {{ role }}
    interface ens32 #定义网络接口，根据自己虚拟机上的网卡修改
    virtual_router_id 51  #主备服务器VRID号必须保持一致
    priority {{ priority }} #服务器优先级，优先级高则优先获得浮动IP
    advert_int 1
    authentication {
        auth_type PASS
        auth_pass 1111
    }
    virtual_ipaddress {
        192.168.26.200
    }
}

编写剧本，复制模板文件，然后重启keepalived服务，这里我们通过两个小剧本的方式传递不同的变量

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
┌──[root@vms152.liruilongs.github.io]-[~]
└─$cat keepalived.yaml
---
- name: vms153.liruilongs.github.io config
  hosts: 192.168.26.153
  tags:
    - master
  vars:
    role: MASTER
    priority: 100
  tasks:
    - name: copy keeplived config
      template:
        src: keepalived.conf.j2
        dest: /etc/keepalived/keepalived.conf

    - name: restart keeplived
      service:
        name: keepalived
        state: restarted


- name: vms154.liruilongs.github.io config
  hosts: 192.168.26.154
  tags:
    - backup
  vars:
    role: BACKUP
    priority: 50
  tasks:
    - name: copy keepalived config
      template:
        src: keepalived.conf.j2
        dest: /etc/keepalived/keepalived.conf

    - name: restart keepalived
      service:
        name: keepalived
        state: restarted

运行剧本并测试

1
2
3
┌──[root@vms152.liruilongs.github.io]-[~]
└─$ansible-playbook keepalived.yaml
。。。。。。。。。。。

假设153机器应为某些原因，需要进行停机处理，我们可以直接把 keepalived 干掉,vip自动切到154

1
2
3
4
5
6
7
8
9
10
11
12
┌──[root@vms152.liruilongs.github.io]-[~]
└─$curl 192.168.26.200:80
vms153.liruilongs.github.io
┌──[root@vms152.liruilongs.github.io]-[~]
└─$ansible 192.168.26.153 -m shell -a "systemctl stop keepalived"
192.168.26.153 | CHANGED | rc=0 >>

┌──[root@vms152.liruilongs.github.io]-[~]
└─$curl 192.168.26.200:80
vms154.liruilongs.github.io
┌──[root@vms152.liruilongs.github.io]-[~]
└─$

如果这个时候153机器恢复，那么我们可以重新拉起keepalived服务，vip回到153

1
2
3
4
5
6
7
┌──[root@vms152.liruilongs.github.io]-[~]
└─$ansible 192.168.26.153 -m shell -a "systemctl start keepalived"
192.168.26.153 | CHANGED | rc=0 >>

┌──[root@vms152.liruilongs.github.io]-[~]
└─$curl 192.168.26.200:80
vms153.liruilongs.github.io

上面的操作，我们可以整合到一个剧本里

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
┌──[root@vms152.liruilongs.github.io]-[~]
└─$cat keepalived.yaml
---
- name:  keepalived init
  hosts: node
  tasks:
    - name: install
      yum:
        name:
          - httpd
          - keepalived
        state: installed

    - name: httpd content
      shell: "echo `hostname` > /var/www/html/index.html"
    - name: Restarted  httpd
      service:
        name: httpd
        state: restarted 

    - name: firewall clons
      shell: firewall-cmd --set-default-zone=trusted
# 主机配置
- name: vms153.liruilongs.github.io config
  hosts: 192.168.26.153
  tags:
    - master
  vars:
    role: MASTER
    priority: 100
    vip: 192.168.26.200
    interface: ens32
  tasks:
    - name: copy keeplived config
      template:
        src: keepalived.conf.j2
        dest: /etc/keepalived/keepalived.conf
    - name: restart keepalived
      service: 
        name: keepalived 
        state: restarted
# 备机配置
- name: vms154.liruilongs.github.io config
  hosts: 192.168.26.154
  tags:
    - backup
  vars:
    role: BACKUP
    priority: 90
    vip: 192.168.26.200
    interface: ens32
  tasks:
    - name: copy keepalived config
      template:
        src: keepalived.conf.j2
        dest: /etc/keepalived/keepalived.conf
    - name: restart keepalived
      service: 
        name: keepalived 
        state: restarted

对于配置文件也可以更灵活一点

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
┌──[root@vms152.liruilongs.github.io]-[~]
└─$cat keepalived.conf.j2
! Configuration File for keepalived

global_defs {
   router_id LVS_DEVEL #设置路由ID,可以和主机名相同，也可以随便定义
   vrrp_iptables       #手动添加(禁止设置防火墙规则，keepalved每次启动都会自动添加防火墙拒绝所有的规则)
}

vrrp_instance VI_1 {
    state {{ role }}
    interface {{ interface }} #定义网络接口，根据自己虚拟机上的网卡修改
    virtual_router_id 51  #主备服务器VRID号必须保持一致
    priority {{ priority }} #服务器优先级，优先级高则优先获得浮动IP
    advert_int 1
    authentication {
        auth_type PASS
        auth_pass 1111
    }
    virtual_ipaddress {
        {{ vip }}
    }
}

有小伙伴会讲，这也太水了，每次故障都需要自己去进行主备切换，其实上面的配置为keepalived的最简单配置，没有使用keepalived 的检查配置，，告警等其他的功能。

如果是IPVS使用keepalived，可以对后端RealServer进行健康状况检查，支持网络层、传输层、应用层进行健康检查。

配置文件解析

熟悉下配置文件，keepalived的配置文件主要由3部分构造，ipvs配置，全局配置，VRRP配置。

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
# 全局配置(全局配置有Global definitions和Static routes/address，全局定义和静态路由)
     global_defs                
         {
            ...
         }
        
# 配置vrrp实例(VRRP实例和VRRP同步组)
# vrrp instance                  # 虚拟路由器，VRRP实例
     vrrp_instance  NAME {
          ...
     }
# vrrp synchronization group     # VRRP同步组
     vrrp_sync_group  NAME  {
          ...
     }
# ipvs的相关配置
# LVS CONFIGURATION：
     # 集群服务，服务内的RS
     Virtual server groups 
     Virtual server               #ipvs集群的vs和rs

下面为具体的参数解释

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
┌──[root@vms152.liruilongs.github.io]-[~]
└─$cat /etc/keepalived/keepalived.conf
! Configuration File for keepalived
# 全局配置(全局配置有Global definitions和Static routes/address，全局定义和静态路由)
global_defs {
  # 全局部分定义邮件报警系统，定义邮件发送目标，收件人邮箱地址 
   notification_email {  
     acassen@firewall.loc
     failover@firewall.loc
     sysadmin@firewall.loc
   }
   # 定义发件人邮箱地址
   notification_email_from Alexandre.Cassen@firewall.loc
   # 定义邮件发送服务器IP，本地发送写localhost
   smtp_server 192.168.200.1
   # 定义邮件服务器建立连接的超时时长
   smtp_connect_timeout 30
   # 标识keepalived服务器的字符串，物理节点的标识符；
   router_id LVS_DEVEL 
   # 如果通告与接收的上一个通告来自相同的master路由器，则不执行检查
   vrrp_skip_check_adv_addr
   # 严格遵守VRRP协议，这一项最好关闭(加感叹号)，若不关闭，可用vip无法被ping通
   ! vrrp_strict 
   # 在一个接口发送的两个免费ARP之间的延迟，可以精确到毫秒级(默认是0)
   vrrp_garp_interval 0
   vrrp_gna_interval 0
   # IPV4多播地址，默认224.0.0.18
   vrrp_mcast_group4 225.0.0.18 
}

# 检查调用
vrrp_script  <SCRIPT_NAME>  {          # 脚本名，后面要基于脚本名来进行调用
        script  "/etc/keepalived/chk_script.sh"   # 执行的命令或脚本
        interval  INT          # 每隔多少时间，这个监控脚本要执行一次
        weight   -INT          # 失败了，当前节点的权重要减去多少,对于“weight”值的设置，有一个简单的标准，即“weight”值的绝对值要大于Master和Backup节点“priority”值之差
}


# 虚拟路由器，VRRP实例
vrrp_instance VI_1 {
    # 定义实例的角色状态是master还是backup，在当前VRRP实例中此节点的初始状态
    state MASTER
    # 定义vrrp绑定的接口，即接收或发送心跳通告的接口，即HA监测接口
    interface eth0
    # 虚拟路由标识(VRID)，同一实例该数值必须相同，即master和backup中该值要相同
    virtual_router_id 51
    # 该vrrp实例中本机的keepalived的优先级，优先级最高的为master(可用范围0-255)
    # 该选项的优先级高于state选项，
    # 即若state指定的是backup，但这里设置的值最高，则仍为master
    priority 100
    # 心跳信息发送和接收时间间隔，单位为秒
    advert_int 1
     # 认证方式，同一实例中这个配置必须完全一样才可通过认证，只建议使用PASS认证
    authentication {
        # 使用简单字符认证的方式
        auth_type PASS
        # 最多支持8字符，超过8字符将只取前8字符
        auth_pass 1111
    }
    # 设置的VIP，当master出现故障后，VIP会故障转移到backup
    virtual_ipaddress {
       # 一般情况下我们只设置一个VIP地址，也可以设置多个
       # 这些vip默认配置在interface指定的接口别名上，可使用dev选项来指定网卡: 192.168.200.19/24 dev eth1
       # 使用ip add的方式添加，若要被ifconfig查看，在IP地址后加上label即可
        192.168.200.16
        192.168.200.17
        192.168.200.18
    }

    # 调用检查    
    track_script {
        chk_service
    }

    # 使用非抢占模式
    nopreempt   
    #非抢占式：如果backup路由器工作在此模式下，则若Master路由器没有出现故障，backup即使随后被配置了更高的优先级也不会成为Master     
    
    # 使用延迟抢占模式
    preempt_delay  TIME 
    #抢占式：如果backup路由器工作在抢占方式下，当它收到VRRP报文后，会将主机的优先级与通告报文中的优先级进行比较，如果主机的优先级比当前的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

    notify_master  <STRING>|<QUOTED-STRING>  # 当切换到master模式时，执行此脚本
    notify_backup  <STRING>|<QUOTED-STRING>  # 当切换到backup模式时，执行此脚本
    notify_fault  <STRING>|<QUOTED-STRING>   # 当切换到fault模式时，执行此脚本
    notify  <STRING>|<QUOTED-STRING> 

}

服务检查

在上面的Demo的基础上，我们添加一个检查脚本,使用下面的检查策略。来检查httpd服务是否可用

1
2
3
4
5
6
7
8
9
10
11
12
┌──[root@vms153.liruilongs.github.io]-[/etc/keepalived]
└─$systemctl status httpd > /dev/null
┌──[root@vms153.liruilongs.github.io]-[/etc/keepalived]
└─$echo $?
0
┌──[root@vms153.liruilongs.github.io]-[/etc/keepalived]
└─$systemctl stop httpd.service
┌──[root@vms153.liruilongs.github.io]-[/etc/keepalived]
└─$systemctl status httpd > /dev/null
┌──[root@vms153.liruilongs.github.io]-[/etc/keepalived]
└─$echo $?
3

也可以這樣

1
2
3
4
5
6
7
8
9
10
11
12
┌──[root@vms153.liruilongs.github.io]-[/etc/keepalived]
└─$systemctl is-active  httpd  -q
┌──[root@vms153.liruilongs.github.io]-[/etc/keepalived]
└─$echo $?
3
┌──[root@vms153.liruilongs.github.io]-[/etc/keepalived]
└─$systemctl start  httpd
┌──[root@vms153.liruilongs.github.io]-[/etc/keepalived]
└─$systemctl is-active  httpd  -q
┌──[root@vms153.liruilongs.github.io]-[/etc/keepalived]
└─$echo $?
0

或者這樣

1
2
3
4
5
6
┌──[root@vms153.liruilongs.github.io]-[/etc/keepalived]
└─$systemctl show httpd -p ActiveState
ActiveState=active
┌──[root@vms153.liruilongs.github.io]-[/etc/keepalived]
└─$systemctl show httpd -p ActiveState | sed 's/ActiveState=//g'
active

是否运行和是否活跃是两个概念，对于某些一次性服务可以使用下面的方式验证

1
2
3
4
5
6
7
8
┌──[root@vms153.liruilongs.github.io]-[/etc/keepalived]
└─$systemctl show httpd -p ActiveState | cut -d'=' -f2
active
┌──[root@vms153.liruilongs.github.io]-[/etc/keepalived]
└─$systemctl show httpd -p SubState | cut -d'=' -f2
running
┌──[root@vms153.liruilongs.github.io]-[/etc/keepalived]
└─$

keepalived部署服务健康检查剧本

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
┌──[root@vms152.liruilongs.github.io]-[~]
└─$cat keepaliveds.yaml
---
- name:  keepalived init
  hosts: node
  tasks:
    - name: install
      yum:
        name:
          - httpd
          - keepalived
        state: installed

    - name: httpd content
      shell: "echo `hostname` > /var/www/html/index.html"

    - name: Restarted  httpd
      service:
        name: httpd
        state: restarted

    - name: firewall clons
      shell: firewall-cmd --set-default-zone=trusted
# 主机配置
- name: vms153.liruilongs.github.io config
  hosts: 192.168.26.153
  tags:
    - master
  vars:
    role: MASTER
    priority: 100
  tasks:
    - name: copy keeplived config
      template:
        src: keepalived.conf.j2
        dest: /etc/keepalived/keepalived.conf
    - name: copy che_service
      copy:
        content: "#!/bin/sh\nsystemctl is-active  httpd  -q"
        dest: /etc/keepalived/che_service.sh
        backup: yes
        mode: '0755'
    - name: restart keepalived
      service:
        name: keepalived
        state: restarted

# 备机配置
- name: vms154.liruilongs.github.io config
  hosts: 192.168.26.154
  tags:
    - backup
  vars:
    role: BACKUP
    priority: 90
  tasks:
    - name: copy keepalived config
      template:
        src: keepalived.conf.j2
        dest: /etc/keepalived/keepalived.conf
    - name: copy che_service
      copy:
        content: "#!/bin/sh\nsystemctl is-active  httpd  -q"
        dest: /etc/keepalived/che_service.sh
        mode: '0755'
        backup: yes
    - name: restart keepalived
      service:
        name: keepalived
        state: restarted

运行测试

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
┌──[root@vms152.liruilongs.github.io]-[~]
└─$curl 192.168.26.200
vms153.liruilongs.github.io
┌──[root@vms152.liruilongs.github.io]-[~]
└─$ansible 192.168.26.153 -m service -a 'name=httpd state=stopped'
192.168.26.153 | CHANGED => {
    "ansible_facts": {
        "discovered_interpreter_python": "/usr/bin/python"
    },
    "changed": true,
    "name": "httpd",
    "state": "stopped",
  .............
┌──[root@vms152.liruilongs.github.io]-[~]
└─$curl 192.168.26.200
vms154.liruilongs.github.io
┌──[root@vms152.liruilongs.github.io]-[~]
└─$

配置文件模板

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
┌──[root@vms152.liruilongs.github.io]-[~]
└─$cat keepalived.conf.j2
! Configuration File for keepalived

global_defs {
   router_id LVS_DEVEL #设置路由ID,可以和主机名相同，也可以随便定义
   vrrp_iptables       #手动添加(禁止设置防火墙规则，keepalved每次启动都会自动添加防火墙拒绝所有的规则)
}

vrrp_script chk_service {
    script /etc/keepalived/che_service.sh
    interval 2
}

vrrp_instance VI_1 {
    state {{ role }}
    interface ens32 #定义网络接口，根据自己虚拟机上的网卡修改
    virtual_router_id 51  #主备服务器VRID号必须保持一致
    priority {{ priority }} #服务器优先级，优先级高则优先获得浮动IP
    advert_int 1
    authentication {
        auth_type PASS
        auth_pass 1111
    }
    virtual_ipaddress {
        192.168.26.200
    }

    track_script {
        chk_service
    }
}

这个剧本调了好些时间，需要注意的是，配置文件不能有其他的非空格的字符，检查脚本要记得授权，如果没有执行检查脚本，可以看下 /var/log/messages 日志文件

1
2
┌──[root@vms153.liruilongs.github.io]-[/etc/keepalived]
└─$cat /var/log/messages | grep -C 10 track

keepalived 角色编写

我们也可以把上面的剧本编写为角色，需要把handlers和和其他的东西抽出来

1
2
3
4
5
6
7
8
9
┌──[root@vms152.liruilongs.github.io]-[~]
└─$ansible-galaxy init keepalived --init-path=./roles
- Role keepalived was created successfully
┌──[root@vms152.liruilongs.github.io]-[~]
└─$ansible-galaxy list
# /root/roles
- keepalived, (unknown version)
┌──[root@vms152.liruilongs.github.io]-[~]
└─$

编写task

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
┌──[root@vms152.liruilongs.github.io]-[~/roles/keepalived]
└─$cat tasks/main.yml
---
# tasks file for keepalived

# 安装 keepalived
- name: Install keepalived
  yum:
    name: keepalived
    state: latest
  tags: keepalived
  notify: restart keepalived

# copy 配置文件
- name: Keepalived configuration
  template:
    src: keepalived.conf.j2
    dest: /etc/keepalived/keepalived.conf
  notify: restart keepalived
# copy 检查脚本
- when:  check_service_name | default(False)
  name: Install check script
  copy:
    content: "#!/bin/sh\nsystemctl is-active  {{ check_service_name }}  -q"
    dest: /etc/keepalived/che_service.sh
    backup: yes
    mode: 0755
    owner: root
    group: root
  notify: restart keepalived

# 启动 服务
- name: Start keepalived
  service:
    name: keepalived
    state: started
    enabled: yes

编写handlers

1
2
3
4
5
6
7
8
9
10
11
┌──[root@vms152.liruilongs.github.io]-[~/roles/keepalived]
└─$cat handlers/main.yml
---
# handlers file for keepalived

- name: restart keepalived
  service:
    name: keepalived
    state: restarted
┌──[root@vms152.liruilongs.github.io]-[~/roles/keepalived]
└─$

编写template

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
┌──[root@vms152.liruilongs.github.io]-[~/roles/keepalived]
└─$cat templates/keepalived.conf.j2
! Configuration File for keepalived

global_defs {
   router_id LVS_DEVEL #设置路由ID,可以和主机名相同，也可以随便定义
   vrrp_iptables       #手动添加(禁止设置防火墙规则，keepalved每次启动都会自动添加防火墙拒绝所有的规则)
}

vrrp_script chk_service {
    script /etc/keepalived/che_service.sh
    interval 2
}

vrrp_instance VI_1 {
    state {{ keep_role }}
    interface {{ keep_interface }} #定义网络接口，根据自己虚拟机上的网卡修改
    virtual_router_id 51  #主备服务器VRID号必须保持一致
    {% if keep_role.lower() == 'master' %}
    priority {{ keep_priority }}
    {% else %}
    priority {{ keep_backup_priority }}
    {% endif %}
    advert_int 1
    authentication {
        auth_type PASS
        auth_pass 1111
    }
    virtual_ipaddress {
        {{ virtual_ipaddress }} dev {{ keep_interface }}
    }

    track_script {
        chk_service
    }
}
┌──[root@vms152.liruilongs.github.io]-[~/roles/keepalived]
└─$

编写defaults变量

1
2
3
4
5
6
7
8
9
10
11
┌──[root@vms152.liruilongs.github.io]-[~/roles/keepalived]
└─$cat defaults/main.yaml
---

keep_role: "master"
keep_priority: 100
keep_backup_priority: 50
keep_interface: "ens32"
virtual_ipaddress: "192.168.26.200"
check_service_name: httpd

编写调用剧本

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
┌──[root@vms152.liruilongs.github.io]-[~/roles/keepalived]
└─$cat tests/test.yml
---
- hosts: 192.168.26.153
  vars:
    keep_role: MASTER
  roles:
    - keepalived

- hosts: 192.168.26.154
  vars:
    keep_role: BACKUP
  roles:
    - keepalived
┌──[root@vms152.liruilongs.github.io]-[~/roles/keepalived]
└─$

博文参考

https://www.keepalived.org/manpage.html

https://www.cnblogs.com/hgzero/p/13718516.html

https://unix.stackexchange.com/questions/396630/the-proper-way-to-test-if-a-service-is-running-in-a-script

https://github.com/tcomerma/ansible-keepalived/

https://github.com/demis-svenska/aws-echis/tree/master/src/commcare_cloud/ansible/roles/keepalived

关于Linux中Keepalived高可用热备自动化部署的一些笔记

https://liruilongs.github.io/2022/09/15/Linux/关于Linux中Keepalived高可用热备部署的一些笔记/

发布于

2022-09-15

更新于

2024-10-08

许可协议

# 推荐文章
  1.《Kubernetes权威指南:从Docker到Kubernetes实践全接触》读书笔记
  2.关于 Git 重写历史的一些笔记
  3.关于Git分支变基操作的一些笔记
  4.关于Git分支基础知识的一些笔记
  5.关于Git分支工作流的一些笔记
  6.关于Git分支高级合并的一些笔记整理

喜欢这篇文章？打赏一下作者吧

评论

目录

关注我

链接

最新评论

加载中，最新评论有1分钟缓存...

最新文章

分类

归档

标签

Kubernetes97
Linux50
Ansible25
Python15
Ceph11
Docker9
Git8
JAVA8
华为云8
CPU7
BPF6
Go6
OpenShift6
DNS5
OKD5
摄影曝光5
程序人生5
selenium4
容器4
摄影构图4
查看全部>>

订阅更新

Your browser is out-of-date!

Update your browser to view this website correctly.&npsb;Update my browser now

×

×