keepalived双主模式高可用工作于NAT模型下的ipvs

一、Keepalived简介

ipvs提供高可用性，能够生成ipvs规则；也可以使用脚本来帮助高可用其他应用程序，没有heartbeart重量级；主要用在高可用ipvs，及一些反向代理应用程序

。

 

1.拓扑图

   说明：1.两个代理服务器通过VIP向外提供数据

              2.两个代理服务器都可以代理后端的服务器

              3.为测试方便，后端服务器至提供静态页面

2.ip规划

 

3.安装配置haproxy

   关于haproxy的详细配置，    

1.  
2. [root@node1 ~]# cat /etc/haproxy/haproxy.cfg
3. #---------------------------------------------------------------------
4. # Example configuration for a possible web application.  See the
5. # full configuration options online.
6. #
7. #   http://haproxy.1wt.eu/download/1.4/doc/configuration.txt
8. #
9. #---------------------------------------------------------------------
10. #---------------------------------------------------------------------
11. # Global settings
12. #---------------------------------------------------------------------
13. global
14. # to have these messages end up in /var/log/haproxy.log you will
15. # need to:
16. #
17. # 1) configure syslog to accept network log events.  This is done
18. #    by adding the '-r' option to the SYSLOGD_OPTIONS in
19. #    /etc/sysconfig/syslog
20. #
21. # 2) configure local2 events to go to the /var/log/haproxy.log
22. #   file. A line like the following can be added to
23. #   /etc/sysconfig/syslog
24. #
25. #    local2.*                       /var/log/haproxy.log
26. #
27. log         127.0.0.1 local2
28. chroot      /var/lib/haproxy
29. pidfile     /var/run/haproxy.pid
30. maxconn     4000
31. user        haproxy
32. group       haproxy
33. daemon
34. # turn on stats unix socket
35. stats socket /var/lib/haproxy/stats
36. #---------------------------------------------------------------------
37. # common defaults that all the 'listen' and 'backend' sections will
38. # use if not designated in their block
39. #---------------------------------------------------------------------
40. defaults
41. mode                    http
42. log                     global
43. option                  httplog
44. option                  dontlognull
45. option http-server-close
46. option forwardfor       except 127.0.0.0/8
47. option                  redispatch
48. retries                 3
49. timeout http-request    10s
50. timeout queue           1m
51. timeout connect         10s
52. timeout client          1m
53. timeout server          1m
54. timeout http-keep-alive 10s
55. timeout check           10s
56. maxconn                 3000
57. #---------------------------------------------------------------------
58. # main frontend which proxys to the backends
59. #---------------------------------------------------------------------
60. frontend  main *:80
61. default_backend static
62. #---------------------------------------------------------------------
63. # static backend for serving up images, stylesheets and such
64. #---------------------------------------------------------------------
65. backend static
66. balance     roundrobin
67. server      node2 192.168.1.202:80 check maxconn 2000
68. server      node3 192.168.1.203:80 check maxconn 2000
69. #---------------------------------------------------------------------
70. # round robin balancing between the various backends
71. #---------------------------------------------------------------------

   说明：两个节点的HAproxy的配置文件应该保持一样

4.测试haproxy的配置

5.安装配置keepalived

       ①、安装keepalived

           keepalived的安装可以通过yum源来安装，也可以通过编译源码来安装，本处通过yum源赖安装

       ②、配置keepalived主节点

1. [root@node1 ~]# cat /etc/keepalived/keepalived.conf
2. ! Configuration File for keepalived
3. global_defs {  //全局参数
4. notification_email { //邮件
5. [e-mail] //收件人地址
6. }
7. notification_email_from [e-mail] //发件人地址
8. smtp_server 127.0.0.1 //邮件服务器的地址
9. smtp_connect_timeout 30 //间隔时间
10. router_id LVS_DEVEL  //邮件服务器的组的id
11. }
12. vrrp_script chk_haproxy {
13. script "killall -0 haproxy" //检查haproxy是否在线
14. interval 1 //检查间隔时间
15. weight -5 //如果检查失败，则权重-5
16. }
17. vrrp_instance VI_1 {   //定义第一个集群
18. state MASTER      //初始状态为主节点，从节点应该为BACKUP
19. interface eth0    //配置ip的端口
20. virtual_router_id 51 //本组集群的id号，主从节点必须一样
21. priority 100 //主节点的优先级，备用节点的优先级必须低于主节点
22. advert_int 1 //心跳检查间隔时间
23. authentication {
24. auth_type PASS   //通信为明文密码通信
25. auth_pass 1111 //通信的密码，主从节点必须一样
26. }
27. virtual_ipaddress {
28. 192.168.1.99 //定义一个VIP
29. }
30. track_script {   //调用上面的命令
31. chk_haproxy
32. }
33. }

   ③、配置keepalived从节点

1. [root@node4 ~]# cat /etc/keepalived/keepalived.conf
2. ! Configuration File for keepalived
3. global_defs {
4. notification_email {
5. [e-mail]
6. }
7. notification_email_from [e-mail]
8. smtp_server 127.0.0.1
9. smtp_connect_timeout 30
10. router_id LVS_DEVEL
11. }
12. vrrp_script chk_haproxy {
13. script "killall -0 haproxy"
14. interval 1
15. weight -5
16. }
17. vrrp_instance VI_1 {
18. state BACKUP
19. interface eth0
20. virtual_router_id 51
21. priority 99
22. advert_int 1
23. authentication {
24. auth_type PASS
25. auth_pass 1111
26. }
27. virtual_ipaddress {
28. 192.168.1.99
29. }
30. track_script {
31. chk_haproxy
32. }
33. }
34. 2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
35. link/ether 00:0c:29:56:78:cd brd ff:ff:ff:ff:ff:ff
36. inet 192.168.1.204/24 brd 192.168.1.255 scope global eth0
37. inet 192.168.1.99/32 scope global eth0
38. inet6 fe80::20c:29ff:fe56:78cd/64 scope link tentative dadfailed
39. valid_lft forever preferred_lft forever

      可以看到ip地址已经配置上去了

   我们可以看到网页访问正常

   下面，我们模拟服务器损坏，将node1上的haproxy关掉，看看ip地址是否会转移到node4上

1. [root@node4 ~]# ip add
2. 1: lo: <LOOPBACK,UP,LOWER_UP> mtu 16436 qdisc noqueue state UNKNOWN
3. link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
4. inet 127.0.0.1/8 scope host lo
5. inet6 ::1/128 scope host
6. valid_lft forever preferred_lft forever
7. 2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
8. link/ether 00:0c:29:56:78:cd brd ff:ff:ff:ff:ff:ff
9. inet 192.168.1.204/24 brd 192.168.1.255 scope global eth0
10. inet 192.168.1.99/32 scope global eth0
11. inet6 fe80::20c:29ff:fe56:78cd/64 scope link tentative dadfailed
12. valid_lft forever preferred_lft forever

   查看网页是否能够访问正常   

    OK！！可以看到我们的页面访问正常，这就可以实现当前段的一个反向代理服务器宕机或者后端的一个web服务宕机，服务都可以正常对外提供

     7.扩展

         我们还可以自定义通知机制

 

1. #!/bin/bash
2. # Author: zero<[e-mail]>
3. # description: An example of notify script
4. #
5. vip=192.168.1.99
6. contact='root@localhost'
7. notify() {
8. mailsubject="`hostname` to be $1: $vip floating"
9. mailbody="`date '+%F %H:%M:%S'`: vrrp transition, `hostname` changed to be $1"
10. echo $mailbody | mail -s "$mailsubject" $contact
11. }
12. case "$1" in
13. master)
14. notify master
15. exit 0
16. ;;
17. backup)
18. notify backup
19. exit 0
20. ;;
21. fault)
22. notify fault
23. exit 0
24. ;;
25. *)
26. echo 'Usage: `basename $0` {master|backup|fault}'
27. exit 1
28. ;;
29. esac

       在keepalived的配置文件中通过notify来调用，如下所示

1. notify_master "/etc/keepalived/notify.sh master"
2. notify_backup "/etc/keepalived/notify.sh backup"
3. notify_fault "/etc/keepalived/notify.sh fault"