在案例一介绍HAProxy高可用负载均衡集群架构中，虽然通过Keepalived实现了HAProxy的高可用，但是严重浪费了服务器资源，因为在一主一备的Keepalived环境中，只有主节点处于工作状态，而备用节点则一直处于空闲等待状态，仅当主节点出现问题时备用节点才能开始工作。对于并发量比大的Web应用系统来说，主节点可能会非常繁忙，而备用节点则十分空闲，这种服务器资源分布不均的问题，也是在做应用架构设计时必须要考虑的问题。对于一主一备资源不均衡的问题，可以通过双主互备的方式进行负载分流，下面就详细讲述双主互备的高可用集群系统是如何实现的。
1.系统架构图与实现原理
为了能充分利用服务器资源并将负载进行分流，可以在一主一备的基础上构建双主互备的高可用HAProxy负载均衡集群系统。双主互备的集群架构如图：
在这个架构中，要实现的功能是：通过haproxy1服务器将www.zb.com的访问请求发送到webapp1和webapp2两台主机上，要实现www.zb.com的负载均衡；通过haproxy2将img.zb.com的访问请求发送到webimg1和webimg2两台主机上，要实现img.zb.com的负载均衡；同时，如果haproxy1或haproxy2任何一台服务器出现故障，都会将用户访问请求发送到另一台健康的负载均衡节点，进而继续保持两个网站的负载均衡。
操作系统：
CentOS release 6.7
地址规划：
主机名 | 物理IP地址 | 虚拟IP地址 | 集群角色 |
haproxy1 | 10.0.0.35 | 10.0.0.40 | 主：www.zb.com |
备：img.zb.com | |||
haproxy2 | 10.0.0.36 | 10.0.0.50 | 主：img.zb.com |
备：www.zb.com | |||
webapp1 | 10.0.0.150 | 无 | Backend Server |
webapp2 | 10.0.0.151 | 无 | Backend Server |
webimg1 | 10.0.0.152 | 无 | Backend Server |
webimg2 | 10.0.0.8 | 无 | Backend Server |
主要：为了保证haproxy1和haproxy2服务器资源得到充分利用，这里对访问进行了分流操作，需要将www.zb.com的域名解析到10.0.0.40这个IP上，将img.zb.com域名解析到10.0.0.50这个IP上。
2.安装并配置HAProxy集群系统
在主机名为haproxy1和haproxy2的节点依次安装HAProxy并配置，配置好的haproxy.cfg文件，内容如下：
global
# to have these messages end up in /var/log/haproxy.log you will
# need to:
#
# 1) configure syslog to accept network log events. This is done
# by adding the '-r' option to the SYSLOGD_OPTIONS in
# /etc/sysconfig/syslog
#
# 2) configure local2 events to go to the /var/log/haproxy.log
# file. A line like the following can be added to
# /etc/sysconfig/syslog
#
# local2.* /var/log/haproxy.log
#
log 127.0.0.1 local2
pidfile /var/run/haproxy.pid
maxconn 4096
user haproxy
group haproxy
daemon
nbproc 1
# turn on stats unix socket
#---------------------------------------------------------------------
# common defaults that all the 'listen' and 'backend' sections will
# use if not designated in their block
#---------------------------------------------------------------------
defaults
mode http
retries 3
timeout connect 5s
timeout client 30s
timeout server 30s
timeout check 2s
listen admin_stats
bind 0.0.0.0:19088
mode http
log 127.0.0.1 local0 err
stats refresh 30s
stats uri /haproxy-status
stats realm welcome login\ Haproxy
stats auth admin:admin
stats hide-version
stats admin if TRUE
#---------------------------------------------------------------------
# main frontend which proxys to the backends
#---------------------------------------------------------------------
frontend www
bind *:80
mode http
option httplog
option forwardfor
log global
acl host_www hdr_dom(host) -i www.zb.com
acl host_img hdr_dom(host) -i img.zb.com
use_backend server_www if host_www
use_backend server_img if host_img
#---------------------------------------------------------------------
# static backend for serving up images, stylesheets and such
#---------------------------------------------------------------------
backend server_www
mode http
option redispatch
option abortonclose
balance roundrobin
option httpchk GET /index.html
server web01 10.0.0.150:80 weight 6 check inter 2000 rise 2 fall 3
server web02 10.0.0.151:80 weight 6 check inter 2000 rise 2 fall 3
backend server_img
mode http
option redispatch
option abortonclose
balance roundrobin
option httpchk GET /index.html
server webimg1 10.0.0.152:80 weight 6 check inter 2000 rise 2 fall 3
server webimg2 10.0.0.8:80 weight 6 check inter 2000 rise 2 fall 3在这个HAProxy配置中，通过ACL规则将www.zb.com站点转到webapp1、webapp2两个后端节点，将img.zb.com站点转到webimg1和webimg2两个后端服务节点，分别实现负载均衡。
最后将haproxy.conf文件分别复制到haproxy1和haproxy2两台服务器上，然后在两个负载均衡器上依次启动HAProxy服务。
3.安装并配置双主的Keepalived高可用系统
依次在主、备两个节点上安装Keepalived。keepalived.conf文件内容:
! Configuration File for keepalived
global_defs {
notification_email {
acassen@firewall.loc
failover@firewall.loc
sysadmin@firewall.loc
}
notification_email_from Alexandre.Cassen@firewall.loc
smtp_server 192.168.200.1
smtp_connect_timeout 30
router_id LVS_DEVEL
}
vrrp_script check_haproxy {
script "killall -0 haproxy"
interval 2
}
vrrp_instance HAProxy_HA {
state MASTER #在haproxy2主机上，此处为BACKUP
interface eth0
virtual_router_id 80 #在一个实例下，virtual_router_id是唯一的，因此在haproxy2上，virtual_router_id也为80
priority 100 #在haproxy2主机上，priority值为80
advert_int 2
authentication {
auth_type PASS
auth_pass aaaa
}
notify_master "/etc/keepalived/mail_notify.sh master"
notify_backup "/etc/keepalived/mail_notify.sh backup"
notify_fault "/etc/keepalived/mail_notify.sh fault"
track_script {
check_haproxy
}
virtual_ipaddress {
10.0.0.40/24 dev eth0
}
}
vrrp_instance HAProxy_HA2 {
state BACKUP #在haproxy2主机上，此处为MASTER
interface eth0
virtual_router_id 81 #在haproxy2主机上，此外virtual_router_id也必须为81
priority 80 #在haproxy2主机上，priority值为100
advert_int 2
authentication {
auth_type PASS
auth_pass aaaa
}
notify_master "/etc/keepalived/mail_notify.sh master"
notify_backup "/etc/keepalived/mail_notify.sh backup"
notify_fault "/etc/keepalived/mail_notify.sh fault"
track_script {
check_haproxy
}
virtual_ipaddress {
10.0.0.50/24 dev eth0
}
}priority值。
在完成所有配置修好后，依次在haproxy1和haproxy2两个节点启动Keepalived服务，并观察VIP地址是否正常加载到对应的节点上。
4.测试
在haproxy1和haproxy2节点依次启动HAProxy服务和Keepalived服务后，首先观察haproxy节点Keepalived的启动日志，信息如下：
Jul 25 10:06:34 data-1-1 Keepalived_vrrp[33167]: VRRP_Instance(HAProxy_HA) Transition to MASTER STATE
Jul 25 10:06:36 data-1-1 Keepalived_vrrp[33167]: VRRP_Instance(HAProxy_HA) Entering MASTER STATE
Jul 25 10:06:36 data-1-1 Keepalived_vrrp[33167]: VRRP_Instance(HAProxy_HA) setting protocol VIPs.
Jul 25 10:06:36 data-1-1 Keepalived_vrrp[33167]: VRRP_Instance(HAProxy_HA) Sending gratuitous ARPs on eth0 for 10.0.0.40
Jul 25 10:06:36 data-1-1 Keepalived_healthcheckers[33166]: Netlink reflector reports IP 10.0.0.40 added
Jul 25 10:06:53 data-1-1 Keepalived_vrrp[33167]: VRRP_Instance(HAProxy_HA2) Received higher prio advert
Jul 25 10:06:53 data-1-1 Keepalived_vrrp[33167]: VRRP_Instance(HAProxy_HA2) Entering BACKUP STATE
下面测试一下双主互备的故障切换功能，这里为了模拟故障，将haproxy1节点上HAProxy服务关闭，然后在haproxy1节点观察Keepalived的启动日志，信息如下：
Jul 25 14:42:09 data-1-1 Keepalived_vrrp[33167]: VRRP_Script(check_haproxy) failed
Jul 25 14:42:10 data-1-1 Keepalived_vrrp[33167]: VRRP_Instance(HAProxy_HA) Entering FAULT STATE
Jul 25 14:42:10 data-1-1 Keepalived_vrrp[33167]: VRRP_Instance(HAProxy_HA) removing protocol VIPs.
Jul 25 14:42:10 data-1-1 Keepalived_vrrp[33167]: VRRP_Instance(HAProxy_HA) Now in FAULT state
Jul 25 14:42:10 data-1-1 Keepalived_healthcheckers[33166]: Netlink reflector reports IP 10.0.0.40 removed
从切换过程看，keepalived运行完全正常。
转载于:https://blog.51cto.com/6284444/2149987
