Nginx+keepalived双机热备（主主模式）

之前已经介绍了Nginx+Keepalived双机热备的主从模式，今天在此基础上说下主主模式的配置。

由之前的配置信息可知：
master机器（master-node）：103.110.98.14/192.168.1.14      VIP1:103.110.98.20
slave机器（slave-node）：103.110.98.24/192.168.1.24       VIP2:103.110.98.21

主主模式需要两个负载均衡的VIP，
之前设置了VIP（103.110.98.20）
所以还需要设置另一个VIP（103.110.98.21）

修改keepalived的配置

1）master负载机上的keepalived配置：（注意，这里是双主配置，MASTER-BACKUP和BACKUP-MASTER;如果是多主，比如三主，就是MATER-BACKUP-BACKUP、BACKUP-MASTER-BACKUP和BACKUP-BACKUP-MASTER）
注意：
配置中的虚拟路由标识virtual_router_id在MASTER和BACKUP处配置不能一样（但在主从模式下配置是一样的）

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66

[root@master-node ~] # vim /etc/keepalived/keepalived.conf ! Configuration File  for  keepalived         global_defs { notification_email {     ops@wangshibo.cn   tech@wangshibo.cn }     notification_email_from ops@wangshibo.cn  smtp_server 127.0.0.1      smtp_connect_timeout 30    router_id master-node     }     vrrp_script chk_http_port {           script  "/opt/chk_nginx.sh"        interval 2                        weight -5                         fall 2                         rise 1                    }     vrrp_instance VI_1 {      state MASTER         interface em1               mcast_src_ip 103.110.98.14      virtual_router_id 51             priority 101                      advert_int 1                      authentication {                      auth_type PASS                   auth_pass 1111                }   track_script {                         chk_http_port                    }   virtual_ipaddress {              103.110.98.20 }   notify_master  "/etc/keepalived/clean_arp.sh 103.110.98.20" }   vrrp_instance VI_2 {                 state BACKUP                interface em1                 mcast_src_ip 103.110.98.24       virtual_router_id 52            priority 99                    advert_int 1                    authentication {                     auth_type PASS                  auth_pass 1111               }    track_script {                         chk_http_port                 } virtual_ipaddress {             103.110.98.21      } notify_master  "/etc/keepalived/clean_arp.sh 103.10.86.21" }

[root@master-node ~]# vim /etc/keepalived/clean_arp.sh         //更新vip的arp记录到网关（注意脚本中的网卡别填错了，要跟vip所在网卡一致）
#!/bin/sh
VIP=$1
GATEWAY=103.110.98.1                                                         //负载均衡器的公网网关地址
/sbin/arping -I em1 -c 5 -s $VIP $GATEWAY &>/dev/null
[root@master-node ~]# chmod 755 /etc/keepalived/clean_arp.sh

2）slave负载机上的keepalived配置：

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66

[root@slave-node ~] # vim /etc/keepalived/keepalived.conf ! Configuration File  for  keepalived         global_defs { notification_email {     ops@wangshibo.cn   tech@wangshibo.cn }     notification_email_from ops@wangshibo.cn  smtp_server 127.0.0.1      smtp_connect_timeout 30    router_id slave-node     }     vrrp_script chk_http_port {           script  "/opt/chk_nginx.sh"        interval 2                        weight -5                         fall 2                         rise 1                    }     vrrp_instance VI_1 {      state BACKUP         interface em1               mcast_src_ip 103.110.98.14      virtual_router_id 51             priority 99                      advert_int 1                      authentication {                      auth_type PASS                   auth_pass 1111                }   track_script {                         chk_http_port                    }   virtual_ipaddress {              103.110.98.20 }   notify_master  "/etc/keepalived/clean_arp.sh 103.110.98.20" }   vrrp_instance VI_2 {                 state MASTER                interface em1                 mcast_src_ip 103.110.98.24      virtual_router_id 52            priority 101                    advert_int 1                    authentication {                     auth_type PASS                  auth_pass 1111               }    track_script {                         chk_http_port                 } virtual_ipaddress {             103.110.98.21      } notify_master  "/etc/keepalived/clean_arp.sh 21" }

[root@slave-node ~]# vim /etc/keepalived/clean_arp.sh 
#!/bin/sh
VIP=$1
GATEWAY=103.110.98.1 
/sbin/arping -I em1 -c 5 -s $VIP $GATEWAY &>/dev/null
[root@slave-node ~]# chmod 755 /etc/keepalived/clean_arp.sh

重启master和slave负载机的keepalive（保证两台机器的ngixn和keepalived服务都启动）
[root@master-node ~]# /etc/init.d/keepalived restart
[root@slave-node ~]# /etc/init.d/keepalived restart

将nginx中配置的域名解析到这两个VIP地址上：
103.110.98.20 dev.wangshibo.com
103.110.98.21 dev.wangshibo.com

浏览器访问是正常的（如果master或slave有一台宕机，或其中一个VIP故障，只要另一台是正常的就行）

 

关闭两台负载机其中一台的keepalived服务，那么它的VIP就会自动漂移到另一台机器上。
关闭两台机器的nginx，会自动重启（前提是keepalived服务要启动）！对网站域名的访问丝毫不受影响。

[root@master-node ~]# pkill -9 nginx
root 32365 9775 0 19:04 pts/0 00:00:00 grep --color=auto nginx
[root@master-node ~]# ps -ef|grep nginx
root 32367 9775 0 19:04 pts/0 00:00:00 grep --color=auto nginx
[root@master-node ~]# ps -ef|grep nginx
root 32369 32368 0 19:04 ? 00:00:00 /bin/bash /opt/chk_nginx.sh
root 32374 1 0 19:04 ? 00:00:00 nginx: master process /usr/local/nginx/sbin/nginx
www 32376 32374 1 19:04 ? 00:00:00 nginx: worker process
www 32377 32374 1 19:04 ? 00:00:00 nginx: worker process
www 32378 32374 1 19:04 ? 00:00:00 nginx: worker process
www 32379 32374 1 19:04 ? 00:00:00 nginx: worker process
www 32380 32374 1 19:04 ? 00:00:00 nginx: worker process
www 32381 32374 1 19:04 ? 00:00:00 nginx: worker process
www 32382 32374 1 19:04 ? 00:00:00 nginx: worker process
www 32383 32374 1 19:04 ? 00:00:00 nginx: worker process
www 32384 32374 0 19:04 ? 00:00:00 nginx: cache manager process
www 32385 32374 0 19:04 ? 00:00:00 nginx: cache loader process
root 32387 9775 0 19:04 pts/0 00:00:00 grep --color=auto nginx

***************当你发现自己的才华撑不起野心时，就请安静下来学习吧***************

分类:  LB+HA, Nginx

本文转自散尽浮华博客园博客，原文链接：http://www.cnblogs.com/kevingrace/p/6146031.html ，如需转载请自行联系原作者