18.1集群介绍 18.2 keepalived介绍18.3/18.4/18.5 用keepalived配置高可用集群

Posted 2020-10-14

- 18.1 集群介绍
- 18.2 keepalived介绍
- 18.3/18.4/18.5 用keepalived配置高可用集群

- 扩展
- heartbeat和keepalived比较 http://blog.csdn.net/yunhua_lee/article/details/9788433 
- DRBD工作原理和配置 http://502245466.blog.51cto.com/7559397/1298945
- mysql+keepalived http://lizhenliang.blog.51cto.com/7876557/1362313

# 18.1 集群介绍
- Linux集群概述
- 根据功能划分为两大类：高可用和负载均衡
- 高可用集群通常为两台服务器，一台工作，另外一台作为冗余，当提供服务的机器宕机，冗余将接替继续提供服务   //通常
- 对于大企业来说。可用程度达到99.99%或者 是5个9
- 实现高可用的开源软件有：heartbeat、keepalived   //centos6 bug 多，而且很久没有更新了，不建议继续使用  ；keepalived不仅有高可用还有负载均衡
- 负载均衡集群，需要有一台服务器作为分发器，它负责把用户的请求分发给后端的服务器处理，在这个集群里，除了分发器外，就是给用户提供服务的服务器了，这些服务器数量至少为2
- 实现负载均衡的开源软件有LVS、keepalived、haproxy、nginx，商业的有F5、Netscaler








# 18.2 keepalived介绍
- 在这里我们使用keepalived来实现高可用集群，因为heartbeat在centos6上有一些问题，影响实验效果
-  keepalived通过VRRP（Virtual Router Redundancy Protocl 中文为：虚拟路由器冗余协议）来实现高可用。来实现高可用。
-  在这个协议里会将多台功能相同的路由器组成一个小组，这个小组里会有1个master角色和N（N>=1）个backup角色。
-  master会通过组播的形式向各个backup发送VRRP协议的数据包，当backup收不到master发来的VRRP数据包时，就会认为master宕机了。此时就需要根据各个backup的优先级来决定谁成为新的mater。
-  Keepalived要有三个模块，分别是core、check和vrrp。其中core模块为keepalived的核心，负责主进程的启动、维护以及全局配置文件的加载和解析，check模块负责健康检查，vrrp模块是来实现VRRP协议的。








# 18.3 用keepalived配置高可用集群（上）
- 搭建高可用的前提，是先要有一个工具，然后需要有一个服务去让工具实现高可用，这个实验，就是让nginx作为一个服务，让它成为一个高可用的对象；因为nginx在企业里使用量比较大，所以就使用他来做服务对象
- 环境准备
- master ：192.168.202.131 （已经做过lnmp）
- backup：192.168.202.132（lamp，没有nginx服务）
- 两台机器都安装keepalived
- 执行yum install -y keepalived
- 为了方便做实验
- 检查两台机器的selinux，iptables两个防火墙情况，selinux需要关闭，iptables需要关闭firewalld
- 对backup机器安装nginx服务
- yum install -y nginx

- 准备两台机器aming-01 是131和aming-02 是 132，131作为master，132作为backup
-  两台机器都执行yum install -y keepalived
-  先是master 安装ke

```
[[email protected] ~]# yum install -y keepalived
已加载插件：fastestmirror
Loading mirror speeds from cached hostfile
 * base: mirrors.sohu.com
 * epel: mirrors.ustc.edu.cn



已安装:
  keepalived.x86_64 0:1.3.5-1.el7                                                           

作为依赖被安装:
  lm_sensors-libs.x86_64 0:3.4.0-4.20160601gitf9185e5.el7                                   
  net-snmp-agent-libs.x86_64 1:5.7.2-28.el7                                                 
  net-snmp-libs.x86_64 1:5.7.2-28.el7                                                       

完毕！
[[email protected] ~]# 
                                                           | 3.6 kB  00:00:00     
```
-  backup 机器安装keppalived
```
[[email protected] ~]# yum install -y keepalived
已加载插件：fastestmirror
Loading mirror speeds from cached hostfile



已安装:
  keepalived.x86_64 0:1.3.5-1.el7                                                           

作为依赖被安装:
  lm_sensors-libs.x86_64 0:3.4.0-4.20160601gitf9185e5.el7                                   
  net-snmp-agent-libs.x86_64 1:5.7.2-28.el7                                                 
  net-snmp-libs.x86_64 1:5.7.2-28.el7                                                       

完毕！
[[email protected] ~]# 
```
- 因为之前aming-01 上面做过nginx,所以上面有nginx服务
```
完毕！
[[email protected] ~]# ps aux |grep nginx
root        868  0.0  0.1  45992  1280 ?        Ss   21:55   0:00 nginx: master process /usr/local/nginx/sbin/nginx -c /usr/local/nginx/conf/nginx.conf
nobody      873  0.0  0.4  48480  4180 ?        S    21:55   0:00 nginx: worker process
nobody      874  0.0  0.3  48480  3924 ?        S    21:55   0:00 nginx: worker process
root       2435  0.0  0.0 112684   976 pts/0    S+   22:03   0:00 grep --color=auto nginx
[[email protected] ~]# 
```
- 而第二台aming-02 机器上看下有没有nginx,没有就用yum install -y nginx 安装一个
```
[[email protected] ~]# rpm -qa |grep nginx
[[email protected] ~]# 

[[email protected] ~]# yum install -y nginx




已安装:
  nginx.x86_64 1:1.10.2-2.el7                                                               

作为依赖被安装:
  gd.x86_64 0:2.0.35-26.el7                                                                 
  gperftools-libs.x86_64 0:2.4-8.el7                                                        
  libXpm.x86_64 0:3.5.12-1.el7                                                              
  libunwind.x86_64 2:1.2-2.el7                                                              
  libxslt.x86_64 0:1.1.28-5.el7                                                             
  nginx-all-modules.noarch 1:1.10.2-2.el7                                                   
  nginx-filesystem.noarch 1:1.10.2-2.el7                                                    
  nginx-mod-http-geoip.x86_64 1:1.10.2-2.el7                                                
  nginx-mod-http-image-filter.x86_64 1:1.10.2-2.el7                                         
  nginx-mod-http-perl.x86_64 1:1.10.2-2.el7                                                 
  nginx-mod-http-xslt-filter.x86_64 1:1.10.2-2.el7                                          
  nginx-mod-mail.x86_64 1:1.10.2-2.el7                                                      
  nginx-mod-stream.x86_64 1:1.10.2-2.el7                                                    

完毕！
[[email protected] ~]# 

```
- 在这有俩个nginx 一个是源码包安装的nginx ，一个是yum 安装的nginx ，
- 下面更改配置文件
- master机器配置，服务工具准备好以后，就配置keepalived ，默认的配置文件路径在
- 先把里面内容清空，有个快捷键 >!$  添加如下配置，
```
[[email protected] ~]# ls /etc/keepalived/keepalived.conf
/etc/keepalived/keepalived.conf
[[email protected] ~]# vim /etc/keepalived/keepalived.conf
[[email protected]aming-01 ~]# > !$
> /etc/keepalived/keepalived.conf
[[email protected] ~]# 
```
- 添加如下配置
```
[[email protected] ~]# !vim
vim /etc/keepalived/keepalived.conf

global_defs {
   notification_email {               //邮件
[email protected]
   }
   notification_email_from [email protected]
   smtp_server 127.0.0.1
   smtp_connect_timeout 30
   router_id LVS_DEVEL
}
vrrp_script chk_nginx {
    script "/usr/local/sbin/check_ng.sh"   //检查服务是否正常，通过脚本实现，检查服务健康状态
    interval 3                      //检查时间
}
vrrp_instance VI_1 {
    state MASTER                           //定义master相关
    interface ens33                        //通过那个网站使用vrrp协议，配置时，需注意你的网卡配置文件是否是哪个。因为系统ens并不是固定的。
    virtual_router_id 51                   //定义路由器ID ，配置的时候和从机器一致
    priority 100                              //权重，
    advert_int 1
    authentication {                       //认证相关信息
        auth_type PASS
        auth_pass aminglinux>com   
    }
    virtual_ipaddress {                  //定义一个公有IP（VIP）
        192.168.188.100   //更改为192.168.202.100
    }
    track_script {
        chk_nginx
    }
}
-- 插入 --                                                                30,2         底端
[[email protected] ~]# vim /etc/keepalived/keepalived.conf


```
- 配置文件编辑完了之后还需要定义check脚本
- 定义一个check的脚本
```
[[email protected] ~]# vim /usr/local/sbin/check_ng.sh

#!/bin/bash
#时间变量，用于记录日志
d=`date --date today +%Y%m%d_%H:%M:%S`
#计算nginx进程数量
n=`ps -C nginx --no-heading|wc -l`
#如果进程为0，则启动nginx，并且再次检测nginx进程数量，
#如果还为0，说明nginx无法启动，此时需要关闭keepalived
if [ $n -eq "0" ]; then
        /etc/init.d/nginx start
        n2=`ps -C nginx --no-heading|wc -l`
        if [ $n2 -eq "0"  ]; then
                echo "$d nginx down,keepalived will stop" >> /var/log/check_ng.log
                systemctl stop keepalived
        fi
fi
~                                                                                           
~                                                                                           
                                                                                        
~                                                                                           
~                                                                                           
-- 插入 --                                                                15,3         全部

[[email protected] ~]# vim /usr/local/sbin/check_ng.sh
[[email protected] ~]# 

```






# 18.4 用keepalived配置高可用集群（中）
- 写完脚本之后，还需要对它做一个权限更改，如果没有权限加载这个脚本的话，它就没有办法启动keepalived服务。
```
[[email protected] ~]# chmod 755 /usr/local/sbin/check_ng.sh
[[email protected] ~]# 
```
- 启动keepalived 服务
```
[[email protected] ~]# systemctl start keepalived
[[email protected] ~]# 

[[email protected] ~]# ps aux |grep keep
root       2498 12.9  0.1 120720  1468 ?        Ss   22:35   0:07 /usr/sbin/keepalived -D
root       2499  0.0  0.2 120720  2752 ?        S    22:35   0:00 /usr/sbin/keepalived -D
root      19605  0.0  0.0 112680   980 pts/0    R+   22:36   0:00 grep --color=auto keep
[[email protected] ~]# ps aux |grep nginx
root        868  0.0  0.1  45992  1280 ?        Ss   21:55   0:00 nginx: master process /usr/local/nginx/sbin/nginx -c /usr/local/nginx/conf/nginx.conf
nobody      873  0.0  0.4  48480  4180 ?        S    21:55   0:00 nginx: worker process
nobody      874  0.0  0.3  48480  3924 ?        S    21:55   0:00 nginx: worker process
root      30104  0.0  0.0 112680   980 pts/0    R+   22:37   0:00 grep --color=auto nginx
[[email protected] ~]# 
```
- 先停止nginx 看看是否会自动启动,结果是会的，
```
[[email protected] ~]# /etc/init.d/nginx stop
Stopping nginx (via systemctl):                            [  确定  ]
[[email protected] ~]# 

[[email protected] ~]# /etc/init.d/nginx stop
Stopping nginx (via systemctl):                            [  确定  ]
[[email protected] ~]# ps aux |grep nginx
root     115314  0.0  0.1  45992  1296 ?        Ss   22:50   0:00 nginx: master process /usr/local/nginx/sbin/nginx -c /usr/local/nginx/conf/nginx.conf
nobody   115318  0.0  0.3  48480  3940 ?        S    22:50   0:00 nginx: worker process
nobody   115319  0.0  0.3  48480  3940 ?        S    22:50   0:00 nginx: worker process
root     115345  0.0  0.0 112680   980 pts/0    S+   22:51   0:00 grep --color=auto nginx
[[email protected] ~]# 

[[email protected] ~]# date
2017年 11月 08日 星期三 22:51:34 CST
[[email protected] ~]# 

```
- keepalived,它的日志在哪里
```
[[email protected] ~]# less /var/log/messages




Nov  8 22:48:27 aming-01 Keepalived[2498]: VRRP child process(102776) died: Respawning
Nov  8 22:48:27 aming-01 Keepalived[2498]: VRRP child process(102777) died: Respawning
Nov  8 22:48:27 aming-01 Keepalived[2498]: VRRP child process(102778) died: Respawning
Nov  8 22:48:27 aming-01 Keepalived[2498]: VRRP child process(102779) died: Respawning
Nov  8 22:48:27 aming-01 rsyslogd-2177: imjournal: begin to drop messages due to rate-limiting
(END)
```
- 来看下它的IP地址,要用ip add ，用ifconfig 看不到的
```
[[email protected] ~]# ip add
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN qlen 1
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
       valid_lft forever preferred_lft forever
    inet6 ::1/128 scope host 
       valid_lft forever preferred_lft forever
2: ens33: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
    link/ether 00:0c:29:2e:28:f2 brd ff:ff:ff:ff:ff:ff
    inet 192.168.202.131/24 brd 192.168.202.255 scope global ens33
       valid_lft forever preferred_lft forever
    inet 192.168.202.100/32 scope global ens33
       valid_lft forever preferred_lft forever
    inet 192.168.202.151/24 brd 192.168.202.255 scope global secondary ens33:0
       valid_lft forever preferred_lft forever
    inet6 fe80::ecdd:28b7:612b:cb7/64 scope link 
       valid_lft forever preferred_lft forever
3: ens37: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
    link/ether 00:0c:29:2e:28:fc brd ff:ff:ff:ff:ff:ff
    inet 192.168.202.134/24 brd 192.168.202.255 scope global dynamic ens37
       valid_lft 1259sec preferred_lft 1259sec
    inet6 fe80::707c:946e:3252:cf7f/64 scope link 
       valid_lft forever preferred_lft forever
[[email protected] ~]# 
```

-  使用ifconfig
```
[[email protected] ~]# ifconfig
ens33: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500
        inet 192.168.202.131  netmask 255.255.255.0  broadcast 192.168.202.255
        inet6 fe80::ecdd:28b7:612b:cb7  prefixlen 64  scopeid 0x20<link>
        ether 00:0c:29:2e:28:f2  txqueuelen 1000  (Ethernet)
        RX packets 7277  bytes 6239399 (5.9 MiB)
        RX errors 0  dropped 0  overruns 0  frame 0
        TX packets 4120  bytes 413974 (404.2 KiB)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

ens33:0: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500
        inet 192.168.202.151  netmask 255.255.255.0  broadcast 192.168.202.255
        ether 00:0c:29:2e:28:f2  txqueuelen 1000  (Ethernet)

ens37: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500
        inet 192.168.202.134  netmask 255.255.255.0  broadcast 192.168.202.255
        inet6 fe80::707c:946e:3252:cf7f  prefixlen 64  scopeid 0x20<link>
        ether 00:0c:29:2e:28:fc  txqueuelen 1000  (Ethernet)
        RX packets 86  bytes 9064 (8.8 KiB)
        RX errors 0  dropped 0  overruns 0  frame 0
        TX packets 20  bytes 3096 (3.0 KiB)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

lo: flags=73<UP,LOOPBACK,RUNNING>  mtu 65536
        inet 127.0.0.1  netmask 255.0.0.0
        inet6 ::1  prefixlen 128  scopeid 0x10<host>
        loop  txqueuelen 1  (Local Loopback)
        RX packets 10  bytes 876 (876.0 B)
        RX errors 0  dropped 0  overruns 0  frame 0
        TX packets 10  bytes 876 (876.0 B)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

[[email protected] ~]# 
```
- 配置从之前，查看下主服务器 有没有防火墙, 以及selinux，从 也是一样
```
[[email protected] ~]# iptables -nvL
Chain INPUT (policy ACCEPT 0 packets, 0 bytes)
 pkts bytes target     prot opt in     out     source               destination         

Chain FORWARD (policy ACCEPT 0 packets, 0 bytes)
 pkts bytes target     prot opt in     out     source               destination         

Chain OUTPUT (policy ACCEPT 0 packets, 0 bytes)
 pkts bytes target     prot opt in     out     source               destination         
[[email protected] ~]# 

[[email protected] ~]# getenforce
Permissive
[[email protected] ~]# 
```
- 检查下从服务器,如果有停止掉
```
[[email protected] ~]# systemctl stop firewalld
[[email protected] ~]# iptables -nvL
Chain INPUT (policy ACCEPT 0 packets, 0 bytes)
 pkts bytes target     prot opt in     out     source               destination         

Chain FORWARD (policy ACCEPT 0 packets, 0 bytes)
 pkts bytes target     prot opt in     out     source               destination         

Chain OUTPUT (policy ACCEPT 0 packets, 0 bytes)
 pkts bytes target     prot opt in     out     source               destination         
[[email protected] ~]# 

[[email protected] ~]# getenforce
Enforcing
[[email protected] ~]# setenforce 0
[[email protected] ~]# getenforce
Permissive
[[email protected] ~]# 
```
- 也需要配置下keepalived,先需要情况里面的原始的配置文件内容，再添加如下内容
```
[[email protected] ~]# > /etc/keepalived/keepalived.conf
[[email protected] ~]# vi !$
vi /etc/keepalived/keepalived.conf
[[email protected] ~]# vi /etc/keepalived/keepalived.conf

     [email protected]
   }
   notification_email_from [email protected]
   smtp_server 127.0.0.1
   smtp_connect_timeout 30
   router_id LVS_DEVEL
}
vrrp_script chk_nginx {
    script "/usr/local/sbin/check_ng.sh"
    interval 3
}
vrrp_instance VI_1 {
    state BACKUP
    interface ens33
    virtual_router_id 51
    priority 90
    advert_int 1
    authentication {
        auth_type PASS
        auth_pass aminglinux>com
    }
    virtual_ipaddress {
        192.168.202.100                这个网段需要和主一样
    }
    track_script {
        chk_nginx
    }
}
:wq

[[email protected] ~]# vi /etc/keepalived/keepalived.conf
[[email protected] ~]# 


```
- 脚本也需要去写下
```
[[email protected] ~]# vim /usr/local/sbin/check_ng.sh

#时间变量，用于记录日志
d=`date --date today +%Y%m%d_%H:%M:%S`
#计算nginx进程数量
n=`ps -C nginx --no-heading|wc -l`
#如果进程为0，则启动nginx，并且再次检测nginx进程数量，
#如果还为0，说明nginx无法启动，此时需要关闭keepalived
if [ $n -eq "0" ]; then
        systemctl start nginx
        n2=`ps -C nginx --no-heading|wc -l`
        if [ $n2 -eq "0"  ]; then
                echo "$d nginx down,keepalived will stop" >> /var/log/check_ng.log
                systemctl stop keepalived
        fi
fi
~                                                                                           
~                                                                                           
                                                                                        
~                                                                                           
~                                                                                           
:wq     
[[email protected] ~]# vim /usr/local/sbin/check_ng.sh
[[email protected] ~]# 
```
- 也需要改变权限，不然就启动不了 keepalived
```
[[email protected] ~]# chmod 755 /usr/local/sbin/check_ng.sh
[[email protected] ~]# 

```
- 启动keepalived
```
[[email protected] ~]# ps aux |grep keep
root       2535  0.0  0.1 120720  1404 ?        Ss   23:16   0:00 /usr/sbin/keepalived -D
root       2536  0.0  0.3 127460  3344 ?        S    23:16   0:00 /usr/sbin/keepalived -D
root       2537  0.0  0.3 131588  3024 ?        S    23:16   0:00 /usr/sbin/keepalived -D
root       2601  0.0  0.0 112680   980 pts/0    S+   23:16   0:00 grep --color=auto keep
[[email protected] ~]# 
```
- 主从都启动了 keepalived，现在俩边 主从 都有 nginx，那你怎么去区分这个nginx？
- 先来看下主的ip
```
[[email protected] ~]# ifconfig
ens33: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500
        inet 192.168.202.131  netmask 255.255.255.0  broadcast 192.168.202.255
        inet6 fe80::ecdd:28b7:612b:cb7  prefixlen 64  scopeid 0x20<link>
        ether 00:0c:29:2e:28:f2  txqueuelen 1000  (Ethernet)
        RX packets 7410  bytes 6251067 (5.9 MiB)
        RX errors 0  dropped 0  overruns 0  frame 0
        TX packets 5672  bytes 512782 (500.7 KiB)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

ens33:0: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500
        inet 192.168.202.151  netmask 255.255.255.0  broadcast 192.168.202.255
        ether 00:0c:29:2e:28:f2  txqueuelen 1000  (Ethernet)

ens37: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500
        inet 192.168.202.134  netmask 255.255.255.0  broadcast 192.168.202.255
        inet6 fe80::707c:946e:3252:cf7f  prefixlen 64  scopeid 0x20<link>
        ether 00:0c:29:2e:28:fc  txqueuelen 1000  (Ethernet)
        RX packets 108  bytes 11608 (11.3 KiB)
        RX errors 0  dropped 0  overruns 0  frame 0
        TX packets 24  bytes 3900 (3.8 KiB)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

lo: flags=73<UP,LOOPBACK,RUNNING>  mtu 65536
        inet 127.0.0.1  netmask 255.0.0.0
        inet6 ::1  prefixlen 128  scopeid 0x10<host>
        loop  txqueuelen 1  (Local Loopback)
        RX packets 10  bytes 876 (876.0 B)
        RX errors 0  dropped 0  overruns 0  frame 0
        TX packets 10  bytes 876 (876.0 B)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

[[email protected] ~]# 
```
- 先去windows 上用浏览器访问下
![mark](http://oqxf7c508.bkt.clouddn.com/blog/20171108/232640129.png?imageslim)
- 这台机器的默认配置文件
```
[[email protected] ~]# cat /usr/local/nginx/conf/vhost/
aaa.com.conf   ld.conf        proxy.conf     ssl.conf       test.com.conf  
[[email protected] ~]# cat /usr/local/nginx/conf/vhost/aaa.com.conf
server
{
    listen 80 default_server; 
    server_name aaa.com;
    index index.html index.htm index.php;
    root /data/wwwroot/default;
location ~ \.php$
    {
        include fastcgi_params;
        fastcgi_pass unix:/tmp/aming.sock;
        fastcgi_index index.php;
        fastcgi_param SCRIPT_FILENAME /data/wwwroot/default$fastcgi_script_name;
    }

}

[[email protected] ~]# 

[[email protected] ~]# vim !$
vim /data/wwwroot/default/index.html

master master. This is the default site.
~                                                                                           
                                                                                          
~                                                                                           
:wq        
```
-  ![mark](http://oqxf7c508.bkt.clouddn.com/blog/20171108/233113698.png?imageslim)
-  关于backup ，从服务器  上面，它的nginx 是yum安装的
```
[[email protected] ~]# vim /usr/share/nginx/html/index.html
[[email protected] ~]# cat !$
cat /usr/share/nginx/html/index.html
backup backup.
[[email protected] ~]# 
```
![mark](http://oqxf7c508.bkt.clouddn.com/blog/20171108/233606639.png?imageslim)
![mark](http://oqxf7c508.bkt.clouddn.com/blog/20171109/000815047.png?imageslim)
- 查看下nginx服务,这个是我们手动起来的，可以停掉试下
```
[[email protected] ~]# ps aux |grep nginx
root       2561  0.0  0.2 122792  2084 ?        Ss   23:16   0:00 nginx: master process /usr/sbin/nginx
nginx      2562  0.0  0.3 123224  3576 ?        S    23:16   0:00 nginx: worker process
root       6985  0.0  0.0 112680   976 pts/0    S+   23:43   0:00 grep --color=auto nginx
[[email protected] ~]# 

[[email protected] ~]# systemctl stop nginx
[[email protected] ~]# !ps
ps aux |grep nginx
root       7155  0.0  0.2 122792  2084 ?        Ss   23:44   0:00 nginx: master process /usr/sbin/nginx
nginx      7156  0.0  0.3 123224  3128 ?        S    23:44   0:00 nginx: worker process
root       7161  0.0  0.0 112680   980 pts/0    S+   23:44   0:00 grep --color=auto nginx
[[email protected] ~]# 

```
- 因为开启了keepalived 服务，所以即使停掉nginx服务，它也会自动加载起来的








# 18.5 用keepalived配置高可用集群（下）
- 先确定好两台机器上nginx差异，比如可以通过curl -I 来查看nginx版本
-  测试1：关闭master上的nginx服务
我们之前都已经测试过了，关闭还是会继续运行，因为加了那个脚本 开启了keepalived服务
-  测试2：在master上增加iptabls规则 
-  iptables -I OUTPUT -p vrrp -j DROP
-  测试3：关闭master上的keepalived服务
-  测试4：启动master上的keepalived服务
 
-  这个vip 是在主上 master 上
```
[[email protected] ~]# ip add
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN qlen 1
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
       valid_lft forever preferred_lft forever
    inet6 ::1/128 scope host 
       valid_lft forever preferred_lft forever
2: ens33: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
    link/ether 00:0c:29:2e:28:f2 brd ff:ff:ff:ff:ff:ff
    inet 192.168.202.131/24 brd 192.168.202.255 scope global ens33
       valid_lft forever preferred_lft forever
    inet 192.168.202.100/32 scope global ens33
       valid_lft forever preferred_lft forever
    inet 192.168.202.151/24 brd 192.168.202.255 scope global secondary ens33:0
       valid_lft forever preferred_lft forever
    inet6 fe80::ecdd:28b7:612b:cb7/64 scope link 
       valid_lft forever preferred_lft forever
3: ens37: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
    link/ether 00:0c:29:2e:28:fc brd ff:ff:ff:ff:ff:ff
    inet 192.168.202.134/24 brd 192.168.202.255 scope global dynamic ens37
       valid_lft 1369sec preferred_lft 1369sec
    inet6 fe80::707c:946e:3252:cf7f/64 scope link 
       valid_lft forever preferred_lft forever
[[email protected] ~]# 

```
- backup上是没有的
```
[[email protected] ~]# ip add
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN qlen 1
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
       valid_lft forever preferred_lft forever
    inet6 ::1/128 scope host 
       valid_lft forever preferred_lft forever
2: ens33: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
    link/ether 00:0c:29:58:33:e6 brd ff:ff:ff:ff:ff:ff
    inet 192.168.202.132/24 brd 192.168.202.255 scope global ens33
       valid_lft forever preferred_lft forever
    inet 192.168.202.152/24 brd 192.168.202.255 scope global secondary ens33:0
       valid_lft forever preferred_lft forever
    inet6 fe80::4500:6d42:8612:4e53/64 scope link 
       valid_lft forever preferred_lft forever
    inet6 fe80::ecdd:28b7:612b:cb7/64 scope link tentative dadfailed 
       valid_lft forever preferred_lft forever
[[email protected] ~]# 
```
- 咱们在主上加一个防火墙规则
- 把主上的vrrp进去的包给封掉 iptables -I OUTPUT -p vrrp -j DROP
```
[[email protected] ~]# iptables -I OUTPUT -p vrrp -j DROP
[[email protected] ~]# 

[[email protected] ~]# iptables -nvL
Chain INPUT (policy ACCEPT 109 packets, 5868 bytes)
 pkts bytes target     prot opt in     out     source               destination         

Chain FORWARD (policy ACCEPT 0 packets, 0 bytes)
 pkts bytes target     prot opt in     out     source               destination         

Chain OUTPUT (policy ACCEPT 32 packets, 2968 bytes)
 pkts bytes target     prot opt in     out     source               destination         
   74  2960 DROP       112  --  *      *       0.0.0.0/0            0.0.0.0/0           
[[email protected] ~]# 
```
- 看下日志
```
[[email protected] ~]# tail /var/log/messages
Nov  9 00:21:58 aming-01 Keepalived_vrrp[124429]: Sending gratuitous ARP on ens33 for 192.168.202.100
Nov  9 00:21:58 aming-01 Keepalived_vrrp[124429]: Sending gratuitous ARP on ens33 for 192.168.202.100
Nov  9 00:21:58 aming-01 Keepalived_vrrp[124429]: Sending gratuitous ARP on ens33 for 192.168.202.100
Nov  9 00:21:58 aming-01 Keepalived_vrrp[124429]: VRRP_Instance(VI_1) Received advert with lower priority 90, ours 100, forcing new election
Nov  9 00:21:58 aming-01 Keepalived_vrrp[124429]: Sending gratuitous ARP on ens33 for 192.168.202.100
Nov  9 00:21:58 aming-01 Keepalived_vrrp[124429]: VRRP_Instance(VI_1) Sending/queueing gratuitous ARPs on ens33 for 192.168.202.100
Nov  9 00:21:58 aming-01 Keepalived_vrrp[124429]: Sending gratuitous ARP on ens33 for 192.168.202.100
Nov  9 00:21:58 aming-01 Keepalived_vrrp[124429]: Sending gratuitous ARP on ens33 for 192.168.202.100
Nov  9 00:21:58 aming-01 Keepalived_vrrp[124429]: Sending gratuitous ARP on ens33 for 192.168.202.100
Nov  9 00:21:58 aming-01 Keepalived_vrrp[124429]: Sending gratuitous ARP on ens33 for 192.168.202.100
[[email protected] ~]# 
```
- 再看下从
```
[[email protected] ~]# less /var/log/messages




























Nov  7 22:01:01 aming-02 systemd: Started Session 4 of user root.
Nov  7 22:01:01 aming-02 systemd: Starting Session 4 of user root.
Nov  7 23:01:01 aming-02 systemd: Started Session 5 of user root.
Nov  7 23:01:01 aming-02 systemd: Starting Session 5 of user root.
Nov  7 23:49:27 aming-02 kernel: hrtimer: interrupt took 6003226 ns
Nov  8 00:01:01 aming-02 systemd: Started Session 6 of user root.
Nov  8 00:01:01 aming-02 systemd: Starting Session 6 of user root.
Nov  8 00:33:47 aming-02 rsyslogd: [origin software="rsyslogd" swVersion="7.4.7" x-pid="834" x-info="http://www.rsyslog.com"] exiting on signal 15.
Nov  8 21:55:29 aming-02 rsyslogd: [origin software="rsyslogd" swVersion="7.4.7" x-pid="827" x-info="http://www.rsyslog.com"] start
Nov  8 21:55:09 aming-02 kernel: Initializing cgroup subsys cpuset
Nov  8 21:55:09 aming-02 kernel: Initializing cgroup subsys cpu
Nov  8 21:55:09 aming-02 kernel: Initializing cgroup subsys cpuacct
Nov  8 21:55:09 aming-02 kernel: Linux version 3.10.0-514.el7.x86_64 ([email protected]) (gcc version 4.8.5 20150623 (Red Hat 4.8.5-11) (GCC) ) #1 SMP Tue Nov 22 16:42:41 UTC 2016
Nov  8 21:55:09 aming-02 kernel: Command line: BOOT_IMAGE=/vmlinuz-3.10.0-514.el7.x86_64 root=UUID=9a2a0d05-a7ab-4948-bbd2-77b2b7d5f565 ro crashkernel=auto rhgb quiet LANG=zh_CN.UTF-8
Nov  8 21:55:09 aming-02 kernel: Disabled fast string operations
Nov  8 21:55:09 aming-02 kernel: e820: Bios-provided physical RAM map:
Nov  8 21:55:09 aming-02 kernel: BIOS-e820: [mem 0x0000000000000000-0x000000000009ebff] usable
Nov  8 21:55:09 aming-02 kernel: BIOS-e820: [mem 0x000000000009ec00-0x000000000009ffff] reserved
Nov  8 21:55:09 aming-02 kernel: BIOS-e820: [mem 0x00000000000dc000-0x00000000000fffff] reserved
/var/log/messages
```
- ![mark](http://oqxf7c508.bkt.clouddn.com/blog/20171109/002641992.png?imageslim)
- 这个也照常可以访问，说明直接封这里的ip是不能直接达到目的的
- 先把这个iptables 规则恢复
```
[[email protected] ~]# iptables -F
[[email protected] ~]# 
```
- 那怎么样模仿现实中宕机的环境，
- 只能暂停keepalived服务试下
```
[[email protected] ~]# systemctl stop keepalived
[[email protected] ~]# ip add
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN qlen 1
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
       valid_lft forever preferred_lft forever
    inet6 ::1/128 scope host 
       valid_lft forever preferred_lft forever
2: ens33: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
    link/ether 00:0c:29:2e:28:f2 brd ff:ff:ff:ff:ff:ff
    inet 192.168.202.131/24 brd 192.168.202.255 scope global ens33
       valid_lft forever preferred_lft forever
    inet 192.168.202.151/24 brd 192.168.202.255 scope global secondary ens33:0
       valid_lft forever preferred_lft forever
    inet6 fe80::ecdd:28b7:612b:cb7/64 scope link 
       valid_lft forever preferred_lft forever
3: ens37: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
    link/ether 00:0c:29:2e:28:fc brd ff:ff:ff:ff:ff:ff
    inet 192.168.202.134/24 brd 192.168.202.255 scope global dynamic ens37
       valid_lft 1452sec preferred_lft 1452sec
    inet6 fe80::707c:946e:3252:cf7f/64 scope link 
       valid_lft forever preferred_lft forever
[[email protected] ~]# 
```
- 再去从服务器看看
```
[[email protected] ~]# ip add
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN qlen 1
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
       valid_lft forever preferred_lft forever
    inet6 ::1/128 scope host 
       valid_lft forever preferred_lft forever
2: ens33: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
    link/ether 00:0c:29:58:33:e6 brd ff:ff:ff:ff:ff:ff
    inet 192.168.202.132/24 brd 192.168.202.255 scope global ens33
       valid_lft forever preferred_lft forever
    inet 192.168.202.100/32 scope global ens33
       valid_lft forever preferred_lft forever
    inet 192.168.202.152/24 brd 192.168.202.255 scope global secondary ens33:0
       valid_lft forever preferred_lft forever
    inet6 fe80::4500:6d42:8612:4e53/64 scope link 
       valid_lft forever preferred_lft forever
    inet6 fe80::ecdd:28b7:612b:cb7/64 scope link tentative dadfailed 
       valid_lft forever preferred_lft forever
[[email protected] ~]# 

[[email protected] ~]# tail /var/log/messages
Nov  9 00:29:54 aming-02 Keepalived_vrrp[10474]: Sending gratuitous ARP on ens33 for 192.168.202.100
Nov  9 00:29:54 aming-02 Keepalived_vrrp[10474]: Sending gratuitous ARP on ens33 for 192.168.202.100
Nov  9 00:29:54 aming-02 Keepalived_vrrp[10474]: Sending gratuitous ARP on ens33 for 192.168.202.100
Nov  9 00:29:54 aming-02 Keepalived_vrrp[10474]: Sending gratuitous ARP on ens33 for 192.168.202.100
Nov  9 00:29:59 aming-02 Keepalived_vrrp[10474]: Sending gratuitous ARP on ens33 for 192.168.202.100
Nov  9 00:29:59 aming-02 Keepalived_vrrp[10474]: VRRP_Instance(VI_1) Sending/queueing gratuitous ARPs on ens33 for 192.168.202.100
Nov  9 00:29:59 aming-02 Keepalived_vrrp[10474]: Sending gratuitous ARP on ens33 for 192.168.202.100
Nov  9 00:29:59 aming-02 Keepalived_vrrp[10474]: Sending gratuitous ARP on ens33 for 192.168.202.100
Nov  9 00:29:59 aming-02 Keepalived_vrrp[10474]: Sending gratuitous ARP on ens33 for 192.168.202.100
Nov  9 00:29:59 aming-02 Keepalived_vrrp[10474]: Sending gratuitous ARP on ens33 for 192.168.202.100
[[email protected] ~]# 

```

- 访问日志可以看到，ip被增加了，现在来访问下web
- ![mark](http://oqxf7c508.bkt.clouddn.com/blog/20171109/003125547.png?imageslim)
- 可以看到发生了改变，说明vip 已经到了从机器上
- 那再把这个keepalived服务 开启
```
[[email protected] ~]# systemctl start keepalived
[[email protected] ~]# ip add
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN qlen 1
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
       valid_lft forever preferred_lft forever
    inet6 ::1/128 scope host 
       valid_lft forever preferred_lft forever
2: ens33: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
    link/ether 00:0c:29:2e:28:f2 brd ff:ff:ff:ff:ff:ff
    inet 192.168.202.131/24 brd 192.168.202.255 scope global ens33
       valid_lft forever preferred_lft forever
    inet 192.168.202.100/32 scope global ens33
       valid_lft forever preferred_lft forever
    inet 192.168.202.151/24 brd 192.168.202.255 scope global secondary ens33:0
       valid_lft forever preferred_lft forever
    inet6 fe80::ecdd:28b7:612b:cb7/64 scope link 
       valid_lft forever preferred_lft forever
3: ens37: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
    link/ether 00:0c:29:2e:28:fc brd ff:ff:ff:ff:ff:ff
    inet 192.168.202.134/24 brd 192.168.202.255 scope global dynamic ens37
       valid_lft 1276sec preferred_lft 1276sec
    inet6 fe80::707c:946e:3252:cf7f/64 scope link 
       valid_lft forever preferred_lft forever
[[email protected] ~]# 
```
- 可以看到马上就能变回来，再看下web访问
- ![mark](http://oqxf7c508.bkt.clouddn.com/blog/20171109/003318786.png?imageslim)
- 看下从上面的日志, 显示 removing protocol VIPs
```
[[email protected] ~]# tail /var/log/messages
Nov  9 00:29:59 aming-02 Keepalived_vrrp[10474]: VRRP_Instance(VI_1) Sending/queueing gratuitous ARPs on ens33 for 192.168.202.100
Nov  9 00:29:59 aming-02 Keepalived_vrrp[10474]: Sending gratuitous ARP on ens33 for 192.168.202.100
Nov  9 00:29:59 aming-02 Keepalived_vrrp[10474]: Sending gratuitous ARP on ens33 for 192.168.202.100
Nov  9 00:29:59 aming-02 Keepalived_vrrp[10474]: Sending gratuitous ARP on ens33 for 192.168.202.100
Nov  9 00:29:59 aming-02 Keepalived_vrrp[10474]: Sending gratuitous ARP on ens33 for 192.168.202.100
Nov  9 00:32:17 aming-02 chronyd[482]: Selected source 120.25.108.11
Nov  9 00:32:17 aming-02 chronyd[482]: System clock wrong by -1.225785 seconds, adjustment started
Nov  9 00:32:51 aming-02 Keepalived_vrrp[10474]: VRRP_Instance(VI_1) Received advert with higher priority 100, ours 90
Nov  9 00:32:51 aming-02 Keepalived_vrrp[10474]: VRRP_Instance(VI_1) Entering BACKUP STATE
Nov  9 00:32:51 aming-02 Keepalived_vrrp[10474]: VRRP_Instance(VI_1) removing protocol VIPs.
[[email protected] ~]# 
```
- 因为主机器宕机，从机器很快的就加入了vip地址，这个时候访问vip地址的时候，看到的页面是

- 从机器上的默认索引页，证明整个实验成功
 
- 这就是高可用的一个集群，通常会用俩台机器，有的企业为了保险，可能会用2-3台backup