Nginx+Keepalived双机热备（主主模式）
msg 2017-10-28
由于网站的访问需求不断加大，负载越来越高。现需要在Web前端放置Nginx负载均衡,同时结合Keepalived对前端Nginx实现HA高可用。
1、nginx进程基于Master+Slave(worker模式)多进程模型，自身具有非常稳定的子进程管理功能。在Master进程分配模式下，Master进程永远不进行业务处理，只是进行任务分发，从而达到Master进程的存活高可靠性，Slave(worker)进程所有的业务信号都由主进程发出，Slave(worker)进程所有的超时任务都会被Master中止，属于非阻塞式任务模型。
2、Keepalived是Linux下面实现VRRP备份路由的高可靠性运行件。基于Keepalived设计的服务模式能够真正做到主服务器和备份服务器故障时IP瞬间无缝交接。二者结合，可以构架出比较稳定的软件LB方案。
双机高可用方法目前分为两种：
1）双机主从模式：即前端使用两台服务器，一台主服务器和一台热备服务器，正常情况下，主服务器绑定一个公网虚拟IP，提供负载均衡服务，热备服务器处于空闲状态；当主服务器发生故障时，热备服务器接管主服务器的公网虚拟IP，提供负载均衡服务；但是热备服务器在主机器不出现故障的时候，永远处于浪费状态，对于服务器不多的网站，该方案不经济实惠。
2）双机主主模式：这种模式的效果很强大，即前端使用两台负载均衡服务器，互为主备，且都处于活动状态（这样达到不浪费服务器），同时各自绑定一个公网虚拟IP，提供负载均衡服务；当其中一台发生故障时，另一台接管发生故障服务器的公网虚拟IP（这时由非故障机器一台负担所有的请求）。这种方案，经济实惠，非常适合于当前架构环境。
一、环境介绍：
操作系统：
[root@CentOS-4 ~]# cat /etc/RedHat-release
CentOS release 6.9 (Final)
服务器对应关系：
KA1：192.168.5.129 centos-1
KA2：192.168.5.128 centos-4
Vip1：192.168.5.200 129master/128backup
VIP2：192.168.5.210 128master/129backup
Web1：192.168.5.131 centos-2
Web2：192.168.5.132 centos-3
二、环境安装：
安装依赖：
（在KA1和KA2机器上执行以下步骤）
[root@centos-4 ~]# yum -y install gcc pcre-devel zlib-devel openssl-devel
[root@centos-4~]# cd /usr/local/src/
[root@centos-4 src]# wget http://nginx.org/download/nginx-1.9.7.tar.gz
安装nginx
[root@centos-4 src]# tar -zvxfnginx-1.9.7.tar.gz
[root@centos-4 src]# cd nginx-1.9.7
[root@centos-4 nginx-1.9.7]#./configure --prefix=/usr/local/nginx --user=nginx --group=nginx--with-http_ssl_module --with-http_flv_module --with-http_stub_status_module--with-http_gzip_static_module --with-pcre
[root@centos-4 nginx-1.9.7]# make &&make install
[root@centos-1 ~]# yum install -ykeepalived
(在web1服务器和web2服务器上安装nginx)
[root@centos-2~]# yum -y install gcc pcre-devel zlib-devel openssl-devel
[root@centos-2~]# cd /usr/local/src/
[root@centos-2 src]# wget http://nginx.org/download/nginx-1.9.7.tar.gz
安装nginx
[root@centos-2 src]# tar -zvxfnginx-1.9.7.tar.gz
[root@centos-2 src]# cd nginx-1.9.7
[root@centos-2 nginx-1.9.7]# ./configure --prefix=/usr/local/nginx --user=nginx --group=nginx--with-http_ssl_module --with-http_flv_module --with-http_stub_status_module--with-http_gzip_static_module --with-pcre
[root@centos-2 nginx-1.9.7]# make &&make install
三、配置服务：
（所以服务器上配置）
[root@centos-1 ~]# cat/etc/sysconfig/selinux
SELINUX=disabled
[root@centos-1 ~]# getenforce
Disabled
[root@centos-1 ~]# service iptables stop
1、配置keepalived：
（KA1上操作）
[root@centos-1 ~]#cat /etc/keepalived/keepalived.conf
! Configuration Filefor keepalived
global_defs {
notification_email {
[email protected]
#[email protected]
#[email protected]
}
router_id LVS_DEVEL
}
vrrp_scriptchk_http_port {
script "/opt/check_nginx.sh"
interval 2
weight -5
fall 2
rise 1
}
vrrp_instance VI_1{
state MASTER
interface eth0
virtual_router_id 51
priority 100
advert_int 1
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
192.168.5.200
}
}
vrrp_instance VI_2{
state BACKUP
interface eth0
virtual_router_id 50
priority 90
advert_int 1
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
192.168.5.210
}
track_script {
chk_http_port
}
}
（KA2上操作）
[root@centos-2 ~]#cat /etc/keepalived/keepalived.conf
! Configuration Filefor keepalived
global_defs {
notification_email {
[email protected]
#[email protected]
#[email protected]
}
router_id LVS_DEVEL
}
vrrp_scriptchk_http_port {
script "/opt/check_nginx.sh"
interval 2
weight -5
fall 2
rise 1
}
vrrp_instance VI_1{
state BACKUP
interface eth0
virtual_router_id 51
priority 90
advert_int 1
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
192.168.5.200
}
}
vrrp_instance VI_2{
state MASTER
interface eth0
virtual_router_id 50
priority 100
advert_int 1
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
192.168.5.210
}
track_script {
chk_http_port
}
}
编写一个监控nginx的脚本：
需要注意的是，要判断本机nginx是否正常，如果发现nginx不正常，重启之后，等待三秒在校验，任然失败则不尝试，关闭keepalived，发送邮件，其他主机此时接管VIP；
[root@centos-4~]# cat /opt/check_nginx.sh
#!/bin/bash
check=$(ps-C nginx --no-heading|wc -l)
IP=`ipadd | grep eth0 | awk 'NR==2{print $2}'| awk -F '/' '{print $1}'`
if ["${check}" = "0" ]; then
/usr/local/nginx/sbin/nginx
sleep 2
counter=$(ps -C nginx --no-heading|wc -l)
if [ "${check}" = "0"]; then
/etc/init.d/keepalived stop
echo "check $IP nginx is down"| mail -s "check keepalived nginx" *********@qq.com
fi
fi
(KA1一样的监控脚本)
2、在两台前端服务器上启动keepalived服务，对于192.168.5.200的vip centos-1是master/192.168.5.210的vip centos-1是backup。
[root@centos-1 ~]#service keepalived start
[root@centos-4 ~]# service keepalived start
查看日志文件：
[root@centos-1 ~]# cat /var/log/messages
Oct 19 22:00:22 centos-1 Keepalived_vrrp[46184]: VRRP_Instance(VI_2)Sending gratuitous ARPs on eth0 for 192.168.5.210
Oct 19 22:00:22 centos-1 Keepalived_healthcheckers[46183]: Netlinkreflector reports IP 192.168.5.210 added
Oct 19 22:00:24 centos-1 Keepalived_vrrp[46184]: VRRP_Instance(VI_1)Sending gratuitous ARPs on eth0 for 192.168.5.200
Oct 19 22:00:27 centos-1 Keepalived_vrrp[46184]: VRRP_Instance(VI_2)Sending gratuitous ARPs on eth0 for 192.168.5.210
（因为KA1先启动keepalived服务所以两个vip都会在KA1上，但第二台keepaliver服务起来后vip2就会被KA2抢占回来。）
[root@centos-4 ~]# cat /var/log/messages
Oct 19 22:01:38 centos-4 Keepalived_healthcheckers[15009]: Netlinkreflector reports IP 192.168.5.210 added
Oct 19 22:01:38 centos-4 avahi-daemon[1513]: Registering new addressrecord for 192.168.5.210 on eth0.IPv4.
Oct 19 22:01:38 centos-4 Keepalived_vrrp[15010]: VRRP_Instance(VI_2)Sending gratuitous ARPs on eth0 for 192.168.5.210
Oct 19 22:01:43 centos-4 Keepalived_vrrp[15010]: VRRP_Instance(VI_2)Sending gratuitous ARPs on eth0 for 192.168.5.210
查看ip addr：
[root@centos-1 keepalived]# ip add
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_faststate UP qlen 1000
link/ether00:0c:29:0d:f3:5d brd ff:ff:ff:ff:ff:ff
inet 192.168.5.129/24 brd192.168.5.255 scope global eth0
inet 192.168.5.200/32scope global eth0
[root@centos-4 keepalived]#ip addr
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdiscpfifo_fast state UP qlen 1000
link/ether00:50:56:3a:84:30 brd ff:ff:ff:ff:ff:ff
inet 192.168.5.128/24 brd192.168.5.255 scope global eth0
inet 192.168.5.210/32 scope global eth0
3、配置nginx的反向代理
（在web1和web2服务器上配置两个web服务（可以http或者nginx）用来测试使用，这里就不一一演示了。）
[root@centos-2 ~]# curl localhost
2
[root@centos-3 ~]# curl localhost
3
（在两台前端服务器上配置）
[root@centos-1 ~]# vim/usr/local/nginx/conf/nginx.conf
……
……
……
upstreambackend {
ip_hash;
server 192.168.5.131:80 max_fails=2fail_timeout=30s;
server 192.168.5.132:80 max_fails=2fail_timeout=30s;
#ip_hash: 每个请求按访问ip的hash结果分配，这样每个访客固定访问一个后端服务器，可以解决session的问题。
#max_fails=2 为允许失败的次数，默认值为1
#fail_timeout=30s 当max_fails次失败后，暂停将请求分发到该后端服务器的时间
}
proxy_temp_path /usr/local/nginx/cache/tmp 1 2;
proxy_cache_path /usr/local/nginx/cache levels=1:2 keys_zone=cache1:100m inactive=1dmax_size=10g;
……
……
……
server {
listen 80;
server_name localhost;
#charset koi8-r;
#access_log logs/host.access.log main;
location / {
proxy_pass http://backend;
proxy_set_header Host $host;
proxy_set_header X-Real-IP $remote_addr;
proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
proxy_cache cache1;
add_header X-Cache$upstream_cache_status;
proxy_cache_key $host$uri$is_args$args;
proxy_cache_valid 200 304 10m;
expires 30d;
# root /web;
index index.php index.html index.htm;
}
[root@centos-2 ~]# vim/usr/local/nginx/conf/nginx.conf
……
……
……
upstreambackend {
ip_hash;
server 192.168.5.131:80 max_fails=2fail_timeout=30s;
server 192.168.5.132:80 max_fails=2fail_timeout=30s;
}
proxy_temp_path /usr/local/nginx/cache/tmp 1 2;
proxy_cache_path /usr/local/nginx/cache levels=1:2 keys_zone=cache1:100m inactive=1dmax_size=10g;
……
……
……
server {
listen 80;
server_name localhost;
#charset koi8-r;
#access_log logs/host.access.log main;
location / {
proxy_pass http://backend;
proxy_set_header Host $host;
proxy_set_header X-Real-IP $remote_addr;
proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
proxy_cache cache1;
add_header X-Cache$upstream_cache_status;
proxy_cache_key $host$uri$is_args$args;
proxy_cache_valid 200 304 10m;
expires 30d;
# root /web;
index index.php index.html index.htm;
}
（两台KA1和KA2服务器重启nginx、keepalived服务）
[root@centos-1~]# /usr/local/nginx/sbin/nginx -t
nginx:the configuration file /usr/local/nginx/conf/nginx.conf syntax is ok
nginx:configuration file /usr/local/nginx/conf/nginx.conf test is successful ###检查配置文件没问题后再执行重启nginx。
[root@centos-1~]# /usr/local/nginx/sbin/nginx -s reload
[root@centos-4~]# /usr/local/nginx/sbin/nginx -t
nginx:the configuration file /usr/local/nginx/conf/nginx.conf syntax is ok
nginx:configuration file /usr/local/nginx/conf/nginx.conf test is successful
[root@centos-4~]# /usr/local/nginx/sbin/nginx -s reload
[root@centos-1~]# service keepalived restart
停止keepalived： [确定]
正在启动keepalived： [确定]
[root@centos-4~]# service keepalived restart
停止keepalived： [确定]
正在启动keepalived： [确定]
四、测试：
验证方法（保证从负载均衡器本机到后端真实服务器之间能正常通信）：
（1）、先测试完成后的效果访问vip1、vip2
[root@centos-1 ~]# curl 192.168.5.200
10.2
[root@centos-1 ~]# curl 192.168.5.210
10.3
（注意在KA1、KA2上做了缓存和ip_hash）
（2）、把KA1keepalived stop掉（模拟KA1主机的keepalived故障）
[root@centos-1 ~]# service keepalived stop
停止 keepalived：
[root@centos-1 ~]# ip addr
2: eth0:<BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen1000
link/ether 00:0c:29:0d:f3:5d brd ff:ff:ff:ff:ff:ff
inet 192.168.5.129/24 brd 192.168.5.255 scope global eth0
inet6 fe80::20c:29ff:fe0d:f35d/64 scope link
valid_lft forever preferred_lft forever
（KA1主机上查看ip addr已经没有vip了。）
在KA2主机上查看日志文件：
[root@centos-4 ~]# cat /var/log/messages
Oct 19 23:20:46 centos-4Keepalived_vrrp[15412]: VRRP_Instance(VI_1) Sending gratuitous ARPs on eth0 for192.168.5.200
Oct 19 23:20:46 centos-4avahi-daemon[1513]: Registering new address record for 192.168.5.200 oneth0.IPv4.
Oct 19 23:20:46 centos-4 Keepalived_healthcheckers[15411]:Netlink reflector reports IP 192.168.5.200 added
Oct 19 23:20:51 centos-4Keepalived_vrrp[15412]: VRRP_Instance(VI_1) Sending gratuitous ARPs on eth0 for192.168.5.200
（日志文件显示已经把vip：192.168.5.200接管了）
查看KA2主机的ip addr
[root@centos-4 ~]# ip addr
2: eth0:<BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen1000
link/ether 00:50:56:3a:84:30 brd ff:ff:ff:ff:ff:ff
inet 192.168.5.128/24 brd 192.168.5.255 scope global eth0
inet 192.168.5.210/32 scope global eth0
inet 192.168.5.200/32 scope global eth0
（可以看到已经有两个vip）
检查nginx服务是否被KA2接管且不中断
[root@centos-1 ~]# curl 192.168.5.210
10.3
[root@centos-1 ~]# curl 192.168.5.200
10.2
（可以看到服务还是进行的而且缓存还在。ip_hash的作用）
一些关于Keepalived相关教程集合：
相关推荐
环境说明：本环境由两台mysql 数据库和heartbeat 组成，一台的ip 为192.168.10.197，一台为192.168.10.198，对外提供服务的vip 为192.168.10.200