Ansible与haproxy实现负载均衡+高可用

注意：此实验是在前几篇博客的基础上做，需要做好解析，server1安装好ansible，server2和3实现与server1的免密，建立普通用户devops并授权等。

负载均衡

实验环境

主机名（IP）	服务
server1(172.25.11.1)	haproxy
server2(172.25.11.2)	apache
server3(172.25.11.3)	apache
server4（172.25.11.4）	动态加入的apache

server1

• 对普通用户devops授权。

[root@server1 ~]# vim /etc/sudoers
 92 devops  ALL=(ALL)       NOPASSWD: ALL

• 安装haproxy服务（目的是得到作为模版的配置文件haproxy.cnf）

[devops@server1 ansible]$ yum list haproxy
[devops@server1 ansible]$ sudo yum install -y haproxy

• copy并编辑模版文件为如下，添加监控页，修改监听端口以及后端主机2和3：

[devops@server1 ansible]$ vim template/haproxy.cfg.j2
[devops@server1 ansible]$ cp /etc/haproxy/haproxy.cfg template/haproxy.cfg.j2
[devops@server1 ansible]$ vim template/haproxy.cfg.j2

• 编写.yml文件。

[devops@server1 ansible]$ vim playbook.yml
---
 - hosts: webserver    #两个后端主机server2和server3
  vars:
    http_port: 80
  tasks:
    - name: install httpd
      yum:
        name: httpd
        state: present

    - name: copy index.html
      copy:
        content: "{{ ansible_facts['hostname'] }}"   #为了实验效果明显，将发布页的内容设置为各自主机的主机名
        dest: /var/www/html/index.html

    - name: configure httpd
      template:
        src: template/httpd.conf.j2
        dest: /etc/httpd/conf/httpd.conf
        owner: root
        group: root
        mode: 644
      notify: restart httpd

    - name: start httpd and firewalld
      service:
        name: "{{ item }}"
        state: started
      loop:
        - httpd
        - firewalld

    - name: configure firewalld
      firewalld:
        service: http
        permanent: yes
        immediate: yes
        state: enabled

  handlers:
    - name: restart httpd
      service:
        name: httpd
        state: restarted

 - hosts: localhost    #本机安装haproxy
  tasks:
    - name: install haproxy
      yum:
        name: haproxy
        state: present

    - name: configure haproxy    #拷贝模版文件，并重启服务。
      template:
        src: template/haproxy.cfg.j2
        dest: /etc/haproxy/haproxy.cfg
      notify: restart haproxy
    - name: start haproxy
      service:
        name: haproxy
        state: started
  handlers:
    - name: restart haproxy
      service:
        name: haproxy
        state: restarted

• 查看任务列表，并执行，在浏览器中访问调度器server1测试：

[devops@server1 ansible]$ ansible-playbook playbook.yml --list-tasks
[devops@server1 ansible]$ ansible-playbook playbook.yml

• 由上可以看到我们的负载均衡配置成功，实现了轮询。
• 但是上述的负载均衡的实现是静态的，即当我们有新的主机有httpd服务时，他不能主动加入集群进行轮询，我们接下来的配置可以解决这个问题，可以实现主机的动态加入。

server4

• 新添加主机server4，给server4上也新建普通用户devops并授权。

[root@server4 ~]# useradd devops
[root@server4 ~]# passwd devops 
[root@server4 ~]# vim /etc/sudoers
devops  ALL=(ALL)       NOPASSWD: ALL

• server1和server4的免密。

[devops@server1 ansible]$ ssh-copy-id server4
[devops@server1 ansible]$ ssh server4

• 编辑inventory文件，加入server4。

[devops@server1 ansible]$ vim inventory
[test]
server2 http_host=172.25.24.2

[prod]
server3 http_host=172.25.24.3
server4 http_host=172.25.24.4
[webserver:children]
test
prod

• 修改haproxy的模版文件为如下，需要注意的是只有变量采用大括号阔起来，常量要写在括号外：

[devops@server1 ansible]$ vim template/haproxy.cfg.j2 
 75 {% for host in groups['webserver'] %}
 76     server {{ hostvars[host]['ansible_facts']['hostname'] }} {{ hostvars[host    ]['ansible_facts']['eth0']['ipv4']['address'] }}:80 check
 77 {% endfor %}

• 执行并在浏览器中访问测试：

[devops@server1 ansible]$ ansible-playbook playbook.yml

• 如上，我们成功实现了动态加入有服务的节点。
• 查看haproxy的配置文件/etc/haproxy/haproxy.cfg：

[devops@server1 ansible]$ vim /etc/haproxy/haproxy.cfg

高可用（利用roles实现）

实验环境

主机名（IP）	服务
server1(172.25.11.1)	haproxy+keepalived（主）
server2(172.25.11.2)	apache
server3(172.25.11.3)	apache
server4（172.25.11.4）	haproxy+keepalived（备）

server1

• 创建一个keepalived模版

[devops@server1 ansible]$ cd roles/
[devops@server1 roles]$ ls
apache  haproxy
[devops@server1 roles]$ ansible-galaxy init keepalived

[devops@server1 ansible]$ cd roles/
[devops@server1 roles]$ cd keepalived/
[devops@server1 keepalived]$ rm -fr README.md  tests

• 编辑任务的main.yml文件

---
 - name: install keepalived
  yum:
    name: keepalived
    state: present
 - name: configure haproxy
  template:
    src: keepalived.conf.j2     #注意路径
    dest: /etc/keepalived/keepalived.conf
  notify: restart keepalived
 - name: start keepalived
  service:
    name: keepalived
    state: started

• 给主机copy并编辑模版文件

[devops@server1 keepalived]$ cp ~/ansible/templates/keepalived.conf.j2 templates/
[devops@server1 keepalived]$ cd templates/
[devops@server1 templates]$ ls
keepalived.conf.j2
[devops@server1 keepalived]$ vim templates/keepalived.conf.j2

• 编辑触发器handlers的main.yml文件

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

• 给inventory文件中加入变量

[devops@server1 ansible]$ vim inventory
[lb]
server1 STATE=MASTER VRID=11 PRIORITY=100
server4 STATE=BACKUP VRID=11 PRIORITY=50
[test]
server2

[prod]
server3

[webserver:children]
test
prod

• 编辑apache.yml文件

[devops@server1 ansible]$ vim apache.yml
---
 - hosts: all
  tasks:
    - import_role:
        name: apache
      when: ansible_hostname in groups['webserver']
    - import_role:
        name: haproxy
      when: ansible_hostname in groups['lb']
    - import_role:
        name: keepalived
      when: ansible_hostname in groups['lb']

• 执行

[devops@server1 ansible]$ ansible-playbook apache.yml

• 浏览器访问http://172.25.11.100测试（VIP）：
  
  
  
• 当我们关掉server1（master）的keepalived服务时，查看vip已经飘移到server4（backup）上。

[devops@server1 ansible]$ systemctl stop keepalived
[devops@server1 ansible]$ ip a

• 再次开启server1的keepalived服务，发现vip又一次漂移回来了。

[devops@server1 ansible]$ sudo systemctl start keepalived
[devops@server1 ansible]$ ip a

• 综上，我们的负载均衡和高可用部署成功。