keepalived容灾方案，实现nginx负载均衡主从架构（1）

一：环境准备：4台nginx服务器，两台用yum安装，两台使用源码安装

第一步：使用yum安装nginx服务器，在浏览器输入ip，可以显示以下内容，这步比较简单，安装好修改/usr/share/nginx/html/index.html就可以显示了：



第二步：使用源码部署nginx,操作步骤如下所示

~:$ yum -y install pcre-devel zlib-devel

~:$ yum -y groupinstall "Development Tools"

~:$ wget http://nginx.org/download/nginx-1.16.1.tar.gz -O /opt/nginx-1.16.1.tar.gz

~:$ tar xf /opt/nginx-1.16.1.tar.gz -C /opt/

~:$ useradd -s /sbin/nologin nginx

~:$ cd /opt/nginx-1.16.1/

/opt/nginx-1.16.1:$ ./configure --prefix=/usr/local/nginx --user=nginx --group=nginx

/opt/nginx-1.16.1:$ make && make install

$ vim /usr/bin/nginxctl -- 设置nginx控制启动命令 #以下是制作一个控制命令的脚本

#!/usr/bin/env bash

        #
        # nginx service control
        case $1 in
        stop)
        /usr/local/nginx/sbin/nginx -s quit
        ;; start)
         /usr/local/nginx/sbin/nginx
         ;; reload)
        /usr/local/nginx/sbin/nginx -s reload
         ; ; *)
        echo "please input (start|stop|reload)"
         ;;
        esac
       #接下来便是准备nginx主配置文件，用于配置负载均衡加反向代理，有3种配置方案，可以进行随意搭配
        user	nginx;

worker_processes auto;

error_log logs/error.log;

events {

use epoll;#使用异步方式

worker_connections 1024;#最大并发数

}

http {

include mime.types;

default_type application/octet-stream;

log_format main '$remote_addr' - '$remote_user [$time_local] "$request" '

'$status $body_bytes_sent "$http_referer" '

'"$http_user_agent" "$http_x_forwarded_for"';

access_log logs/loadbalance_access.log main;

sendfile on;

keepalive_timeout 65;

轮询方式:按照时间顺序依次分配请求到后端

upstream myweb01 {

server 192.168.161.21:80;

server 192.168.161.22:80;

server 192.168.161.23:80;

}

ip_hash会话保持:针对解决session会话问题

upstream myweb02 {

server 192.168.161.21:80;

server 192.168.161.22:80;

server 192.168.161.23:80;

ip_hash;

}

weight权重:用于解决后端服务器性能不均情况;权重值越大,被访问到的几率就越大;

upstream myweb03 {

server 192.168.161.21:80 weight=6;

server 192.168.161.22:80 weight=3;

server 192.168.161.23:80 backup;

}

server {

listen 80;

server_name www.bavdu.com;

location / {
  #配置反向代理至upstream地址池,由upstream分配请求至地址池中的某台机器;
  proxy_pass	http://myweb;
  #设置反向代理可以获取到用户的真实IP地址;
  proxy_set_header	Host	$host;
  proxy_set_header	X-Real-IP	$remote_addr;
  proxy_set_header	X-Forwarded-For	$proxy_add_x_forwarded_for;
  #允许客户端请求的最大单文件字节数;
  client_max_body_size	10m;
  #缓冲区代理缓冲用户端请求的最大字节数;
  client_body_buffer_size	128K;
  #和后端服务器连接超时时间;
  proxy_connect_timeout	90;
  #接收后端服务器响应超时时间;
  proxy_read_timeout	90;
  #设置保存用户头部信息的缓冲区大小;
  proxy_buffer_size	4k;
  #高负荷下缓冲区大小;
  proxy_busy_buffers_size	64k;
}

}

}

负载均衡加反向代理实现好之后，接下来就是安装keepalived，主从的配置也不一样

nginx-loadbalance-master:设置vip

~:$ ifconfig ens33:1 192.168.161.100 broadcast 192.168.161.255 netmask 255.255.255.0 up

~:$ route add -host 192.168.161.100 dev ens33:1

主机安装keepalived服务

~:$ yum -y install openssl-devel popt-devel libnl-devel kernel-devel

~:$ wget https://www.keepalived.org/software/keepalived-2.0.18.tar.gz

~:$ tar xf keepalived-2.0.18.tar.gz

~:$ cd keepalived-2.0.18

keepalived-2.0.18:$ ./configure --prefix=/usr/local/keepalived

keepalived-2.0.18:$ make && make install

配置keepalived服务

~:$ mkdir -p /etc/keepalived

~:$ ln -s /usr/local/keepalived/etc/keepalived/keepalived.conf /etc/keepalived/#从机不需要此操作

~:$ systemctl start keepalived && systemctl status keepalived

~:$ cp /usr/local/keepalived/etc/keepalived/{keepalived.conf,keepalived.conf.old}

~:$ vim /etc/keepalived/keepalived.conf

! Configuration File for keepalived

global_defs {

router_id loadbalance01 # 设定本机的唯一标识,一般为主机名

}

vrrp_script check_nginx_status {

script /etc/keepalived/check_nginx_status.sh # 指定脚本位置

interval 5 # 执行脚本的间隔

weight -20 # 设置脚本执行优先级

}

前提nginx要启动才会有虚拟ip出现，当nginx停止后，便会执行脚本，停止keepalived,便会进行ip漂移

vrrp_instance VI_1 {

state MASTER #标示状态为MASTER 备份机为BACKUP

interface ens33 #设置实例绑定的网卡

virtual_router_id 51 #virtual_router_id必须相同

priority 100 #MASTER权重要高于BACKUP 比如BACKUP为99

advert_int 1 #MASTER与BACKUP负载均衡器之间同步检查的时间间隔,单位是秒

authentication { #主从服务器验证方式

auth_type PASS

auth_pass 123456

}

virtual_ipaddress { #设置vip

192.168.161.100

}

track_script { #执行检查脚本,当检测到nginx宕掉后,把keepalived也杀死,从而进行故障转移

check_nginx_status

}

}

这个脚本是检测nginx是否运行，如果没有运行就执行杀死keepalived的进程，一旦进程杀死，虚拟ip便会漂移到从机上，

!/usr/bin/env bash

file: /etc/keepalived/check_nginx_status.sh

usage: monitor nginx status;

nginx_number=$(ps -C nginx | grep -v PID | wc -l)

if [ ${nginx_number} -eq 0 ];then

yum -y install psmisc &

killall keepalived &

fi

从机节点的设置

从节点配置

! Configuration File for keepalived

global_defs {

router_id loadbalance02

}

vrrp_script check_nginx_status {

script /etc/keepalived/check_nginx_status.sh

interval 5

weight -20

}

vrrp_instance VI_1 {

state BACKUP

interface ens33

virtual_router_id 51 #必须和主设置一样

priority 99 #必须比主机的小，优先级低

advert_int 1 #1s检测一次

authentication {

auth_type PASS

auth_pass 123456

}

virtual_ipaddress {

192.168.161.100 #一个虚拟ip，但是要在同网段上

}

track_script {

check_nginx_status

}

}