Keepalived+nginx实现双主高可用负载均衡

Nginx+keepalived高可用有两种配置方案：
1、Nginx+keepalived 主从配置
这种方案，使用一个vip地址，前端使用2台机器，一台做主，一台做备，但同时只有一台机器工作，另一台备份机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，所以本次不予采用。
2、Nginx+keepalived 双主配置

这种方案，使用两个vip地址，前端使用2台机器，互为主备，同时有两台机器工作，当其中一台机器出现故障，两台机器的请求转移到一台机器负担，非常适合于当前架构环境，故本次采用此方案对网站进行高可用架构。

下面我们以第二种配置演示实现双主高可用负载均衡。

第一组

VIP1：192.168.36.100

主nginx：192.168.36.99

备nginx：192.168.36.86

第二组

VIP2：192.168.36.100

主nginx：192.168.36.86

备nginx：192.168.36.99

web服务：

web1：192.168.36.215

web2：192.168.36.80

一、nginx配置（安装省略）

worker_processes  1;events {    worker_connections  1024;}http {    include       mime.types;    default_type  application/octet-stream;    sendfile        on;    upstream web1 {        server 192.168.36.215:80 max_fails=3 fail_timeout=3s;    }    upstream web2 {        server 192.168.36.80:80 max_fails=3 fail_timeout=3s;    }    server {            listen         88;            server_name  localhost;            location / {                root   html;                index  index.html index.htm;                proxy_pass http://web1;                proxy_set_header X-Real-IP $remote_addr;                proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;            }   }   server {            listen         99;            server_name  localhost;            location / {                root   html;                index  index.html index.htm;                proxy_pass http://web2;                proxy_set_header X-Real-IP $remote_addr;                proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;            }   }}

配置并启动192.168.36.99和192.168.36.86的nginx。检查访问是否正常.

web1

http://192.168.36.86:88/

http://192.168.36.99:88/

web2

http://192.168.36.86:99/

http://192.168.36.99:99/

二、安装、配置keepalived（安装省略）

1)192.168.36.86配置

vi /etc/keepalived/keepalived.conf

! Configuration File for keepalivedglobal_defs {   notification_email {     [e-mail]     #[e-mail]   }   notification_email_from [e-mail]   smtp_server smtp.163.com   smtp_connect_timeout 30   router_id LVS_DEVEL}vrrp_script chk_nginx {        script "/etc/keepalived/chk_nginx.sh"        interval 2        weight 2}vrrp_instance VI_1 {    state BACKUP    interface eth0    virtual_router_id 100    priority 50    advert_int 1    authentication {        auth_type PASS        auth_pass 1111    }    track_script {        chk_nginx    }    virtual_ipaddress {        192.168.36.100    }}vrrp_instance VI_2 {    state MASTER    interface eth0    virtual_router_id 200    priority 100    advert_int 1    authentication {        auth_type PASS        auth_pass 1111    }    track_script {        chk_nginx    }       virtual_ipaddress {        192.168.36.200    }}

2)192.168.36.99配置

vi /etc/keepalived/keepalived.conf

! Configuration File for keepalivedglobal_defs {   notification_email {     [e-mail]     #[e-mail]   }   notification_email_from [e-mail]   smtp_server smtp.163.com   smtp_connect_timeout 30   router_id LVS_DEVEL}vrrp_script chk_nginx {        script "/etc/keepalived/chk_nginx.sh"        interval 2        weight 2}vrrp_instance VI_1 {    state MASTER    interface eth0    virtual_router_id 100    priority 100    advert_int 1    authentication {        auth_type PASS        auth_pass 1111    }    track_script {        chk_nginx    }    virtual_ipaddress {        192.168.36.100    }}vrrp_instance VI_2 {    state BACKUP    interface eth0    virtual_router_id 200    priority 50    advert_int 1    authentication {        auth_type PASS        auth_pass 1111    }    track_script {        chk_nginx    }       virtual_ipaddress {        192.168.36.200    }}

3）两台机器都需要有nginx进程检查脚本，当检测不到进程的时候重启nginx，无法重启就关闭keepalived，以便让vip切换到备机。

vi /etc/keepalived/chk_nginx.sh

#!/bin/sh#description: # 如果启动失败，则停止keepalivedstatus=$( ps -C nginx --no-heading| wc -l)if [ "${status}" = "0" ]; then        /usr/local/nginx/sbin/nginx        status2=$( ps -C nginx --no-heading| wc -l)        if [ "${status2}" = "0" ]; then                service keepalived stop        fifi

设置执行权限：chmod +x /etc/keepalived/chk_nginx.sh

三、启动、测试

分别启动nginx和keepalived：

 /usr/local/nginx/sbin/nginx

service keepalived start

访问：

web1

http://192.168.36.100:88/

http://192.168.36.200:88/

web2

http://192.168.36.100:99/

http://192.168.36.200:99/

可以正常访问。

我们可以试一下，关闭一个keepalived。关闭192.168.36.86上的keepalived

service keepalived stop

查看日志：

tail -f /var/log/messages 

192.168.36.86日志：

Feb  7 00:39:05 slave-d Keepalived[5738]: Stopping Keepalived v1.2.15 (02/07,2015)Feb  7 00:39:05 slave-d Keepalived_vrrp[5741]: VRRP_Instance(VI_2) sending 0 priorityFeb  7 00:39:05 slave-d Keepalived_vrrp[5741]: VRRP_Instance(VI_2) removing protocol VIPs.Feb  7 00:39:05 slave-d Keepalived_healthcheckers[5740]: Netlink reflector reports IP 192.168.36.200 removedFeb  7 00:39:05 slave-d avahi-daemon[1823]: Withdrawing address record for 192.168.36.200 on eth0.

192.168.36.99日志：

Feb  7 00:39:11 slave-c Keepalived_vrrp[25103]: VRRP_Instance(VI_2) Transition to MASTER STATEFeb  7 00:39:12 slave-c Keepalived_vrrp[25103]: VRRP_Instance(VI_2) Entering MASTER STATEFeb  7 00:39:12 slave-c Keepalived_vrrp[25103]: VRRP_Instance(VI_2) setting protocol VIPs.Feb  7 00:39:12 slave-c Keepalived_healthcheckers[25102]: Netlink reflector reports IP 192.168.36.200 addedFeb  7 00:39:12 slave-c avahi-daemon[1832]: Registering new address record for 192.168.36.200 on eth0.IPv4.Feb  7 00:39:12 slave-c Keepalived_vrrp[25103]: VRRP_Instance(VI_2) Sending gratuitous ARPs on eth0 for 192.168.36.200Feb  7 00:39:17 slave-c Keepalived_vrrp[25103]: VRRP_Instance(VI_2) Sending gratuitous ARPs on eth0 for 192.168.36.200

再开启192.168.36.86上的keepalived：

192.168.36.86日志：

Feb  7 00:40:42 slave-d Keepalived[6004]: Starting Keepalived v1.2.15 (02/07,2015)Feb  7 00:40:42 slave-d Keepalived[6005]: Starting Healthcheck child process, pid=6007Feb  7 00:40:42 slave-d Keepalived[6005]: Starting VRRP child process, pid=6008Feb  7 00:40:42 slave-d Keepalived_vrrp[6008]: Netlink reflector reports IP 192.168.36.86 addedFeb  7 00:40:42 slave-d Keepalived_vrrp[6008]: Netlink reflector reports IP fe80::20c:29ff:fe6a:66ff addedFeb  7 00:40:42 slave-d Keepalived_vrrp[6008]: Registering Kernel netlink reflectorFeb  7 00:40:42 slave-d Keepalived_vrrp[6008]: Registering Kernel netlink command channelFeb  7 00:40:42 slave-d Keepalived_vrrp[6008]: Registering gratuitous ARP shared channelFeb  7 00:40:42 slave-d Keepalived_vrrp[6008]: Opening file '/etc/keepalived/keepalived.conf'.Feb  7 00:40:42 slave-d Keepalived_healthcheckers[6007]: Netlink reflector reports IP 192.168.36.86 addedFeb  7 00:40:42 slave-d Keepalived_healthcheckers[6007]: Netlink reflector reports IP fe80::20c:29ff:fe6a:66ff addedFeb  7 00:40:42 slave-d Keepalived_healthcheckers[6007]: Registering Kernel netlink reflectorFeb  7 00:40:42 slave-d Keepalived_healthcheckers[6007]: Registering Kernel netlink command channelFeb  7 00:40:42 slave-d Keepalived_healthcheckers[6007]: Opening file '/etc/keepalived/keepalived.conf'.Feb  7 00:40:42 slave-d Keepalived_vrrp[6008]: Configuration is using : 44182 BytesFeb  7 00:40:42 slave-d Keepalived_vrrp[6008]: Using LinkWatch kernel netlink reflector...Feb  7 00:40:42 slave-d Keepalived_vrrp[6008]: VRRP_Instance(VI_1) Entering BACKUP STATEFeb  7 00:40:42 slave-d Keepalived_vrrp[6008]: VRRP sockpool: [ifindex(2), proto(112), unicast(0), fd(10,11)]Feb  7 00:40:42 slave-d Keepalived_healthcheckers[6007]: Configuration is using : 7257 BytesFeb  7 00:40:42 slave-d Keepalived_healthcheckers[6007]: Using LinkWatch kernel netlink reflector...Feb  7 00:40:42 slave-d Keepalived_vrrp[6008]: VRRP_Instance(VI_2) Transition to MASTER STATEFeb  7 00:40:42 slave-d Keepalived_vrrp[6008]: VRRP_Instance(VI_2) Received lower prio advert, forcing new electionFeb  7 00:40:43 slave-d Keepalived_vrrp[6008]: VRRP_Instance(VI_2) Entering MASTER STATEFeb  7 00:40:43 slave-d Keepalived_vrrp[6008]: VRRP_Instance(VI_2) setting protocol VIPs.Feb  7 00:40:43 slave-d avahi-daemon[1823]: Registering new address record for 192.168.36.200 on eth0.IPv4.Feb  7 00:40:43 slave-d Keepalived_vrrp[6008]: VRRP_Instance(VI_2) Sending gratuitous ARPs on eth0 for 192.168.36.200Feb  7 00:40:43 slave-d Keepalived_healthcheckers[6007]: Netlink reflector reports IP 192.168.36.200 addedFeb  7 00:40:48 slave-d Keepalived_vrrp[6008]: VRRP_Instance(VI_2) Sending gratuitous ARPs on eth0 for 192.168.36.200

192.168.36.99日志：

Feb  7 00:40:47 slave-c Keepalived_vrrp[25103]: VRRP_Instance(VI_2) Received higher prio advertFeb  7 00:40:47 slave-c Keepalived_vrrp[25103]: VRRP_Instance(VI_2) Entering BACKUP STATEFeb  7 00:40:47 slave-c Keepalived_vrrp[25103]: VRRP_Instance(VI_2) removing protocol VIPs.Feb  7 00:40:47 slave-c Keepalived_healthcheckers[25102]: Netlink reflector reports IP 192.168.36.200 removedFeb  7 00:40:47 slave-c avahi-daemon[1832]: Withdrawing address record for 192.168.36.200 on eth0.

关闭后依旧可以正常访问。

关闭nginx后，可以看到nginx立即有启动。

参考文章：

CentOS6.5 keepalived详解及实现Nginx服务的高可用性

用NginX+keepalived实现高可用的负载均衡

Keepalived+Nginx实现高可用和双主节点负载均衡

nginx+keepalived实现nginx双主高可用的负载均衡