Web集群_02

Web集群_01

Keepalived

概述

Keepalived实现了高可用集群
Keepalived最初是为LVS设计 , 专门监控各种服务器节点的状态
Keepalived 后加入了 VRRP 功能 , 防止单点故障
- VRRP ( 虚拟冗余路由协议 )
- VRRP能在不改变网组的情况下 , 将多台路由器虚拟成一个虚拟路由器 , 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

作用

通过 VRRP 实现服务或网络的高可用性
为LVS设计 , 监控集群各个服务节点的状态
某个节点出现故障时 , Keepalived将检测到后将节点从集群系统中剔除
在节点修复后 , 自动将此节点重新加入集群
节点需要手工修复

特点

部署简单 , 只需要配置一个配置文件即可
加入了VRRP , 可以保证业务或网络不间断稳定运行

核心功能

健康检查
- 采用 tcp三次握手 , icmp请求 , http请求 , udp , echo请求等方式
- 对负载均衡后端实际服务器进行保活
故障切换
- 主要应用在配置了主备的服务器上
- 使用ARRP 协议维持主备之间的心跳
- 当主服务器出现问题时 , 由备服务器承载对应业务
- 最大限度减少损失 , 并提供服务的稳定性

原理

keepalived 工作在 TCP/IP五层模型中 , 使用第三层到第五层

网络层 , 传输层 , 应用层

网络层 : 通过 ICMP协议向集群每个节点发送UCMP数据包 , 根据回馈判定节点是否正常运行 , 失效节点则剔除集群
传输层 : 通过 TCP协议的端口连接和扫描技术判断集群节点是否正常 , 若端口无反映则视为失效节点
**应用层 : 用户可以编写shell脚本或playbook剧本来运行keepalived , keepalived根据脚本检测各种程序或服务是否正常 . **

配置Keepalived 高可用web集群

# 在两台web服务器上安装keepalived
[root@pubserver cluster]# vim 07-install-keepalived.yml
---
- name: install keepalived
  hosts: webservers
  tasks:
    - name: install keepalived   # 安装keepalived
      yum:
        name: keepalived
        state: present
        
[root@pubserver cluster]# ansible-playbook 07-install-keepalived.yml

# 修改配置文件
[root@web1 ~]# vim /etc/keepalived/keepalived.conf 
 12    router_id web1    # 设置本机在集群中的唯一识别符
 13    vrrp_iptables     # 自动配置iptables放行规则
 ... ...
 20 vrrp_instance VI_1 {
 21     state MASTER           # 状态，主为MASTER，备为BACKUP
 22     interface eth0         # 网卡
 23     virtual_router_id 51   # 虚拟路由器地址
 24     priority 100           # 优先级
 25     advert_int 1           # 发送心跳消息的间隔
 26     authentication {
 27         auth_type PASS     # 认证类型为共享密码
 28         auth_pass 1111     # 集群中的机器密码相同，才能成为集群
 29     }   
 30     virtual_ipaddress {
 31         192.168.88.80/24    # VIP地址
 32     }   
 33 }
# 删除下面所有行
[root@web1 ~]# systemctl start keepalived

# 等几秒服务完全启动后，可以查看到vip
[root@web1 ~]# ip a s eth0 | grep '88'
    inet 192.168.88.100/24 brd 192.168.88.255 scope global noprefixroute eth0
    inet 192.168.88.80/24 scope global secondary eth0

# 配置web2
[root@web1 ~]# scp /etc/keepalived/keepalived.conf 192.168.88.200:/etc/keepalived/

[root@web2 ~]# vim /etc/keepalived/keepalived.conf 
 12    router_id web2          # 改id
 13    vrrp_iptables
 ... ... 
 20 vrrp_instance VI_1 {
 21     state BACKUP           # 改状态
 22     interface eth0
 23     virtual_router_id 51
 24     priority 80            # 改优先级
 25     advert_int 1
 26     authentication {
 27         auth_type PASS
 28         auth_pass 1111
 29     }
 30     virtual_ipaddress {
 31         192.168.88.80/24
 32     }
 33 }

# 启动服务
[root@web2 ~]# systemctl start keepalived
# 查看地址，eth0不会出现vip
[root@web2 ~]# ip a s | grep '88'
    inet 192.168.88.200/24 brd 192.168.88.255 scope global noprefixroute eth0

# 测试，现在访问88.80，看到是web1上的内容
[root@client1 ~]# curl http://192.168.88.80
Welcome from web1

# 模拟web1出现故障
[root@web1 ~]# systemctl stop keepalived.service 

# 测试，现在访问88.80，看到是web2上的内容
[root@client1 ~]# curl http://192.168.88.80
Welcome from web2

# 在web2上查看vip，可以查看到vip 192.168.88.80
[root@web2 ~]# ip a s | grep '88'
    inet 192.168.88.200/24 brd 192.168.88.255 scope global noprefixroute eth0
    inet 192.168.88.80/24 scope global secondary eth0

监控80端口 , 实现主备切换

实现原理

配置高可用的web集群时，Keepalived只为服务器提供了VIP
Keepalived不知道服务器上运行了哪些服务
MASTER服务器可以通过跟踪脚本监视本机的80端口，一旦本机80端口失效，则将VIP切换至BACKUP服务器
Keepalived对脚本的要求是，退出码为0表示访问成功；退出码为1表示失败。

实施

# 1. 在MASTER上创建监视脚本
[root@web1 ~]# vim /etc/keepalived/check_http.sh
#!/bin/bash
ss -tlnp | grep :80 &> /dev/null && exit 0 || exit 1

[root@web1 ~]# chmod +x /etc/keepalived/check_http.sh

# 2. 修改MASTER配置文件，使用脚本
[root@web1 ~]# vim /etc/keepalived/keepalived.conf 
  1 ! Configuration File for keepalived
  2 
  3 global_defs {
...略...
 18 }
 19 
 20 vrrp_script chk_http_port {  # 定义监视脚本
 21     script "/etc/keepalived/check_http.sh"
 22     interval 2   # 脚本每隔2秒运行一次
 23 }
 24
 25 vrrp_instance VI_1 {
 26     state MASTER
 27     interface eth0
 28     virtual_router_id 51
 29     priority 100
 30     advert_int 1
 31     authentication {
 32         auth_type PASS
 33         auth_pass 1111
 34     }
 35     virtual_ipaddress {
 36         192.168.88.80/24
 37     }
 38     track_script {    # 引用脚本
 39         chk_http_port
 40     }
 41 }

# 3. 重起服务
[root@web1 ~]# systemctl restart keepalived.service 

# 4. 测试，关闭web1的nginx后，VIP将会切换至web2
[root@web1 ~]# systemctl stop nginx.service 
[root@web1 ~]# ip a s | grep 88
    inet 192.168.88.100/24 brd 192.168.88.255 scope global noprefixroute eth0
[root@web2 ~]# ip a s | grep 88
    inet 192.168.88.200/24 brd 192.168.88.255 scope global noprefixroute eth0
    inet 192.168.88.80/24 scope global secondary eth0
    
# 5. 当MASTER的nginx修复后，VIP将会切换回至web1
[root@web1 ~]# systemctl start nginx.service 
[root@web1 ~]# ip a s | grep 88
    inet 192.168.88.100/24 brd 192.168.88.255 scope global noprefixroute eth0
    inet 192.168.88.80/24 scope global secondary eth0

[root@web2 ~]# ip a s | grep 88
    inet 192.168.88.200/24 brd 192.168.88.255 scope global noprefixroute eth0

配置高可用 , 负载均衡web集群

开启web1 , web2 的虚拟网卡 ifup lo:0 192.168.88.15
ifup lo:0
设置开启自动启动
/etc/rc.d/rc.local 内为开机执行命令 , 添加 ifup lo:0

LVS-DR模式

client1：eth0->192.168.88.10
lvs1：eth0->192.168.88.5
lvs2：eth0->192.168.88.6
web1：eth0->192.168.88.100
web2：eth0->192.168.88.200

环境准备

# 关闭2台web服务器上的keepalived，并卸载
[root@pubserver cluster]# vim 08-rm-keepalived.yml
---
- name: remove keepalived
  hosts: webservers
  tasks:
    - name: stop keepalived       # 停服务
      service:
        name: keepalived
        state: stopped
      
    - name: uninstall keepalived   # 卸载
      yum:
        name: keepalived
        state: absent

[root@pubserver cluster]# ansible-playbook 08-rm-keepalived.yml

# 创建新虚拟机lvs2
[root@myhost ~]# vm clone lvs2

# 为lvs2设置ip地址
[root@myhost ~]# vm setip lvs2 192.168.88.6

# 连接
[root@myhost ~]# ssh 192.168.88.6

配置高可用 , 负载均衡

1. 在2台web服务器的lo上配置vip

2. 在2台web服务器上配置内核参数

3. 删除lvs1上的`eth0`上的VIP地址。因为vip将由keepalived接管

[root@pubserver cluster]# vim 09-del-lvs1-vip.yml
---
- name: del lvs1 vip
  hosts: lvs1
  tasks:
    - name: rm vip
      lineinfile:            # 在指定文件中删除行
        path: /etc/sysconfig/network-scripts/ifcfg-eth0
        regexp: 'IPADDR2='   # 正则匹配
        state: absent
      notify: restart system

  handlers:
    - name: restart system
      shell: reboot
      
[root@pubserver cluster]# ansible-playbook 09-del-lvs1-vip.yml

# 查看结果
[root@lvs1 ~]# ip a s eth0 | grep 88
    inet 192.168.88.5/24 brd 192.168.88.255 scope global noprefixroute eth0

4. 删除lvs1上的lvs规则。因为lvs规则将由keepalived创建

[root@lvs1 ~]# ipvsadm -Ln   # 查看规则
[root@lvs1 ~]# ipvsadm -D -t 192.168.88.15:80

5. 在lvs上配置keepalived

# 在主机清单文件中加入lvs2的说明
[root@pubserver cluster]# vim inventory
...略...
[lb]
lvs1 ansible_host=192.168.88.5
lvs2 ansible_host=192.168.88.6
...略...

# 安装软件包
[root@pubserver cluster]# cp 01-upload-repo.yml 10-upload-repo.yml
---
- name: config repos.d
  hosts: lb
  tasks:
    - name: delete repos.d
      file:
        path: /etc/yum.repos.d
        state: absent

    - name: create repos.d
      file:
        path: /etc/yum.repos.d
        state: directory
        mode: '0755'

    - name: upload local88
      copy:
        src: files/local88.repo
        dest: /etc/yum.repos.d/

[root@pubserver cluster]# ansible-playbook 10-upload-repo.yml 

[root@pubserver cluster]# vim 11-install-lvs2.yml
---
- name: install lvs keepalived
  hosts: lb
  tasks:
    - name: install pkgs    # 安装软件包
      yum:
        name: ipvsadm,keepalived
        state: present

[root@pubserver cluster]# ansible-playbook 11-install-lvs2.yml

[root@lvs1 ~]# vim /etc/keepalived/keepalived.conf 
 12    router_id lvs1       # 为本机取一个唯一的id
 13    vrrp_iptables        # 自动开启iptables放行规则
... ...
 20 vrrp_instance VI_1 {
 21     state MASTER
 22     interface eth0
 23     virtual_router_id 51
 24     priority 100
 25     advert_int 1
 26     authentication {
 27         auth_type PASS
 28         auth_pass 1111
 29     }   
 30     virtual_ipaddress {
 31         192.168.88.15       # vip地址，与web服务器的vip一致
 32     }   
 33 }
 # 以下为keepalived配置lvs的规则
 35 virtual_server 192.168.88.15 80 {   # 声明虚拟服务器地址
 36     delay_loop 6     # 健康检查延迟6秒开始
 37     lb_algo wrr      # 调度算法为wrr
 38     lb_kind DR       # 工作模式为DR
 39     persistence_timeout 50  # 50秒内相同客户端调度到相同服务器
 40     protocol TCP     # 协议是TCP
 41 
 42     real_server 192.168.88.100 80 {   # 声明真实服务器
 43         weight 1          # 权重
 44         TCP_CHECK {       # 通过TCP协议对真实服务器做健康检查
 45             connect_timeout 3 # 连接超时时间为3秒
 46             nb_get_retry 3    # 3次访问失败则认为真实服务器故障
 47             delay_before_retry 3  # 两次检查时间的间隔3秒
 48         }
 49     }
 50     real_server 192.168.88.200 80 {
 51         weight 2
 52         TCP_CHECK {
 53             connect_timeout 3
 54             nb_get_retry 3
 55             delay_before_retry 3
 56         }
 57     }
 58 }

# 以下部分删除

# 启动keepalived服务
[root@lvs1 ~]# systemctl start keepalived

# 验证
[root@lvs1 ~]# ip a s eth0 | grep 88
    inet 192.168.88.5/24 brd 192.168.88.255 scope global noprefixroute eth0
    inet 192.168.88.15/32 scope global eth0

[root@lvs1 ~]# ipvsadm -Ln    # 出现规则
IP Virtual Server version 1.2.1 (size=4096)
Prot LocalAddress:Port Scheduler Flags
  -> RemoteAddress:Port           Forward Weight ActiveConn InActConn
TCP  192.168.88.15:80 wrr persistent 50
  -> 192.168.88.100:80            Route   1      0          0     
  -> 192.168.88.200:80            Route   2      0          0    

# 客户端连接测试
[root@client1 ~]# for i in {1..6}; do curl http://192.168.88.15/; done
Welcome from web2
Welcome from web2
Welcome from web2
Welcome from web2
Welcome from web2
Welcome from web2

# 为了效率相同的客户端在50秒内分发给同一台服务器。为了使用同一个客户端可以看到轮询效果，可以注释配置文件中相应的行后，重启keepavlied。
[root@lvs1 ~]# vim +39 /etc/keepalived/keepalived.conf
...略...
    # persistence_timeout 50
...略...
[root@lvs1 ~]# systemctl restart keepalived.service 

# 在客户端验证
[root@client1 ~]# for i in {1..6}; do curl http://192.168.88.15/; done
Welcome from web2
Welcome from web1
Welcome from web2
Welcome from web2
Welcome from web1
Welcome from web2

# 配置LVS2
[root@lvs1 ~]# scp /etc/keepalived/keepalived.conf 192.168.88.6:/etc/keepalived/

[root@lvs2 ~]# vim /etc/keepalived/keepalived.conf 
 12    router_id lvs2
 21     state BACKUP
 24     priority 80

[root@lvs2 ~]# systemctl start keepalived
[root@lvs2 ~]# ipvsadm -Ln   # 出现规则
IP Virtual Server version 1.2.1 (size=4096)
Prot LocalAddress:Port Scheduler Flags
  -> RemoteAddress:Port           Forward Weight ActiveConn InActConn
TCP  192.168.88.15:80 wrr
  -> 192.168.88.100:80            Route   1      0          0     
  -> 192.168.88.200:80            Route   2      0          0

6.验证

# 1. 验证真实服务器健康检查
[root@web1 ~]# systemctl stop nginx
[root@lvs1 ~]# ipvsadm -Ln   # web1在规则中消失
[root@lvs2 ~]# ipvsadm -Ln

[root@web1 ~]# systemctl start nginx
[root@lvs1 ~]# ipvsadm -Ln   # web1重新出现在规则中
[root@lvs2 ~]# ipvsadm -Ln

# 2. 验证lvs的高可用性
[root@lvs1 ~]# shutdown -h now    # 关机
[root@lvs2 ~]# ip a s | grep 88     # 可以查看到vip
    inet 192.168.88.6/24 brd 192.168.88.255 scope global noprefixroute eth0
    inet 192.168.88.15/32 scope global eth0
# 客户端访问vip依然可用
[root@client1 ~]# for i in {1..6}; do curl http://192.168.88.15/; done
Welcome from web1
Welcome from web2
Welcome from web2
Welcome from web1
Welcome from web2
Welcome from web2

HAProxy

也是一款实现负载均衡的调度器
适用于负载特别大的web站点
HAProxy的工作模式：
- mode http：只适用于web服务
- mode tcp：适用于各种服务
- mode health：仅做健康检查，很少使用

初始化配置

client1：eth0 -> 192.168.88.10
HAProxy：eth0 -> 192.168.88.5
web1：eth0 -> 192.168.88.100
web2：eth0 -> 192.168.88.200

# 关闭192.168.88.6
[root@lvs2 ~]# shutdown -h now

# 配置192.168.88.5为haproxy服务器
[root@pubserver cluster]# vim 12-config-haproxy.yml
---
- name: config haproxy
  hosts: lvs1
  tasks:
    - name: rm lvs keepalived     # 删除软件包
      yum:
        name: ipvsadm,keepalived
        state: absent
        
    - name: rename hostname       # 修改主机名
      shell: hostnamectl set-hostname haproxy1
     
    - name: install haproxy       # 安装软件包
      yum:
        name: haproxy
        state: present

[root@pubserver cluster]# ansible-playbook 12-config-haproxy.yml
# web服务器，不需要配置vip，不需要改内核参数。但是存在对haproxy也没有影响。