一 原理简介
1.HAProxy
HAProxy提供高可用性、负载均衡以及基于TCP和HTTP应用的代理，支持虚拟主机，它是免费、快速并且可靠的一种解决方案。HAProxy特别适用于那些负载特大的web站点，这些站点通常又需要会话保持或七层处理。HAProxy运行在时下的硬件上，完全可以支持数以万计的并发连接。并且它的运行模式使得它可以很简单安全的整合进当前的架构中， 同时可以保护web服务器不被暴露到网络上。
2.Keepalived
Keepalived 是一个基于VRRP协议来实现的LVS服务高可用方案，可以利用其来避免单点故障。一个LVS服务会有2台服务器运行Keepalived，一台为主服务器（MASTER），一台为备份服务器（BACKUP），但是对外表现为一个虚拟IP，主服务器会发送特定的消息给备份服务器，当备份服务器收不到这个消息的时候，即主服务器宕机的时候，备份服务器就会接管虚拟IP，继续提供服务，从而保证了高可用性。Keepalived是VRRP的完美实现。
3.vrrp协议
在现实的网络环境中，两台需要通信的主机大多数情况下并没有直接的物理连接。对于这样的情况，它们之间路由怎样选择？主机如何选定到达目的主机的下一跳路由，这个问题通常的解决方法有两种：
在主机上使用动态路由协议(RIP、OSPF等)
在主机上配置静态路由
很明显，在主机上配置动态路由是非常不切实际的，因为管理、维护成本以及是否支持等诸多问题。配置静态路由就变得十分流行，但路由器(或者说默认网关default gateway)却经常成为单点故障。VRRP的目的就是为了解决静态路由单点故障问题，VRRP通过一竞选(election)协议来动态的将路由任务交给LAN中虚拟路由器中的某台VRRP路由器。
二 运行环境
1.系统
CentOS Linux release 7.2.1511 (Core)
2.应用软件
haproxy-1.5.14-3.el7.x86_64
keepalived-1.2.13-7.el7.x86_64
httpd-2.4.6-45.el7.centos.x86_64
php-5.4.16-42.el7.x86_64
mariadb-server-5.5.52-1.el7.x86_64
mariadb.x86_64
wordpress-4.3.1-zh_CN.zip
nfs-utils-1.3.0-0.33.el7.x86_64
rpcbind-0.2.0-38.el7.x86_64
3.IP配置
负载均衡器
DIP1:172.18.67.13
DIP2:172.18.67.14
后端Real Server
RIP1:172.18.67.11
RIP2:172.18.67.12
数据库服务器
MIP:172.18.67.1
客户端IP
IP:172.18.67.3
VIP:172,.18.67.33
三 架构拓扑及应用软件安装
1.拓扑图
2.部署应用软件
在IP为172.18.67.13与172.18.67.14的服务器上安装部署haproxy、keepalived
[root@inode2 ~]# yum install haproxy keepalived -y [root@inode3 ~]# yum install haproxy keepalived -y
在IP为172.18.67.11与172.18.67.12的服务器上安装部署httpd、php
[root@inode4 ~]# yum install httpd php -y [root@inode5 ~]# yum install httpd php -y
在IP为172.18.67.1的服务器上部署mariadb、mariadb-server、php-mysql
[root@inode6 ~]# yum install mariadb mariadb-server php-mysql -y
另外由于数据库采用文件共享的方式，所以还要安装nfs，分别在Real Server端和数据库服务器端安装nfs的应用软件nfs-utils、rpcbind
[root@inode4 ~]# yum install nfs-utils rpcbind -y [root@inode5 ~]# yum install nfs-utils rpcbind -y [root@inode6 ~]# yum install nfs-utils rpcbind -y
四 配置
1.Real Server配置
在这里我们将IP为172.18.67.11和172.18.67.12的服务器80端口设置为动态资源站，将IP为172.18.67.11和172.18.67.12的8080端口服务器模拟另两台服务器设置成静态资源站：将wordpress应用分别解压至/var/www/html/下，并修改该目录的属主和属组
[root@inode4 ~]# unzip wordpress-4.3.1-zh_CN.zip -C /var/www/html/ [root@inode4 ~]# chown -R apache:apache /var/www/html/wordpress [root@inode5 ~]# unzip wordpress-4.3.1-zh_CN.zip -C /var/www/html/ [root@inode5 ~]# chown -R apache:apache /var/www/html/wordpress
修改后端Server的httpd配置文件将网站目录从默认的/var/www/html修改为/var/www/html/wordpress。
2.nfs配置
数据库端修改配置文件
[root@inode6 ~]# vim /etc/exports /data/ 172.18.67.11(rw,async) /data/ 172.18.67.12(rw,async)
修改保存完毕后启动Real Server和数据库服务器的nfs应用
[root@inode4 ~]# systemctl start rpcbind [root@inode4 ~]# systemctl start nfs [root@inode5 ~]# systemctl start rpcbind [root@inode5 ~]# systemctl start nfs [root@inode6 ~]# systemctl start rpcbind [root@inode6 ~]# systemctl start nfs
创建数据存放目录及修改权
[root@inode6 ~]# mkdir /data [root@inode6 ~]# chown -R mysql:mysql /data
修改数据库配置文件数据存储目录
[root@inode6 ~]# vim /etc/my.cnf datadir=/data/
启动数据库
[root@inode6 ~]# systemctl start mariadb-server
将数据库目录挂载至web动态资源服务器
[root@inode4 ~]# mount -t nfs 172.18.67.1:/data/ /mnt [root@inode5 ~]# mount -t nfs 172.18.67.1:/data/ /mnt
3.keepalived配置
MASTER
! Configuration File for keepalived
global_defs {
notification_email {
root@localhost
}
notification_email_from keepalived@localhost
smtp_server 127.0.0.1
smtp_connect_timeout 30
router_id inode2
vrrp_macst_group4 224.0.67.67
}
vrrp_instance http {
state MASTER
interface eno16777736
virtual_router_id 67
priority 100
advert_int 1
authentication {
auth_type PASS
auth_pass lKZvQVv9
}
virtual_ipaddress {
172.18.67.33/16 dev eno16777736
}
}
notify_master "/etc/keepalived/notify.sh master"
notify_backup "/etc/keepalived/notify.sh backup"
notify_fault "/etc/keepalived/notify.sh fault"
}
BACKUP
! Configuration File for keepalived
global_defs {
notification_email {
root@localhost
}
notification_email_from keepalived@localhost
smtp_server 127.0.0.1
smtp_connect_timeout 30
router_id inode2
vrrp_macst_group4 224.0.67.67
}
vrrp_instance http {
state BACKUP
interface eno16777736
virtual_router_id 67
priority 98
advert_int 1
authentication {
auth_type PASS
auth_pass lKZvQVv9
}
virtual_ipaddress {
172.18.67.33/16 dev eno16777736
}
}
notify_master "/etc/keepalived/notify.sh master"
notify_backup "/etc/keepalived/notify.sh backup"
notify_fault "/etc/keepalived/notify.sh fault"
}
通知脚本
#!/bin/bash
#
contact='root@localhost'
notify() {
mailsubject="$(hostname) to be $1, vip floating"
mailbody="$(date +'%F %T'): vrrp transition, $(hostname) changed to be $1"
echo "$mailbody" | mail -s "$mailsubject" $contact
}
case $1 in
master)
notify master
;;
backup)
notify backup
;;
fault)
notify fault
;;
*)
echo "Usage: $(basename $0) {master|backup|fault}"
exit 1
;;
esac
4.haproxy配置
两节点的配置内容是一样的，如下：
[root@inode2 haproxy]# vim haproxy.cfg global log 127.0.0.1 local2 chroot /var/lib/haproxy pidfile /var/run/haproxy.pid maxconn 4000 user haproxy group haproxy daemon stats socket /var/lib/haproxy/stats defaults mode http log global option httplog option dontlognull option http-server-close option forwardfor except 127.0.0.0/8 option redispatch retries 3 timeout http-request 10s timeout queue 1m timeout connect 10s timeout client 1m timeout server 1m timeout http-keep-alive 10s timeout check 10s maxconn 3000 frontend web *:80 mode http maxconn 2000 acl url_static path_beg -i /static /images /javascript /stylesheets acl url_static path_end -i .jpg .gif .png .css .js .html .txt .htm use_backend staticsrvs if url_static default_backend appsrvs backend staticsrvs balance roundrobin server stcsrvs1 172.18.67.11:8080 check server stcsrvs2 172.18.67.12:8080 check backend appsrvs balance roundrobin server wp1 172.18.67.11:80 check server wp2 172.18.67.12:80 check listen stats bind :10086 stats enable stats uri /admin?stats stats auth admin:admin stats admin if TRUE
五 启动服务并测试
1.启动haproxy和keepalived
[root@inode2 ~]# systemctl restart haproxy [root@inode2 ~]# systemctl restart keepalived [root@inode3 ~]# systemctl restart haproxy [root@inode3 ~]# systemctl restart keepalived
2.测试
inode2：
[root@inode2 ~]# systemctl status -l keepalived ● keepalived.service - LVS and VRRP High Availability Monitor Loaded: loaded (/usr/lib/systemd/system/keepalived.service; disabled; vendor preset: disabled) Active: active (running) since Wed 2017-05-17 23:49:45 CST; 6s ago Process: 28940 ExecStart=/usr/sbin/keepalived $KEEPALIVED_OPTIONS (code=exited, status=0/SUCCESS) Main PID: 28941 (keepalived) CGroup: /system.slice/keepalived.service ├─28941 /usr/sbin/keepalived -D ├─28942 /usr/sbin/keepalived -D └─28943 /usr/sbin/keepalived -D May 17 23:49:45 inode2 Keepalived_vrrp[28943]: Registering gratuitous ARP shared channel May 17 23:49:45 inode2 Keepalived_vrrp[28943]: Opening file '/etc/keepalived/keepalived.conf'. May 17 23:49:45 inode2 Keepalived_vrrp[28943]: Configuration is using : 63025 Bytes May 17 23:49:45 inode2 Keepalived_vrrp[28943]: Using LinkWatch kernel netlink reflector... May 17 23:49:45 inode2 Keepalived_vrrp[28943]: VRRP sockpool: [ifindex(2), proto(112), unicast(0), fd(10,11)] May 17 23:49:46 inode2 Keepalived_vrrp[28943]: VRRP_Instance(http) Transition to MASTER STATE May 17 23:49:47 inode2 Keepalived_vrrp[28943]: VRRP_Instance(http) Entering MASTER STATE May 17 23:49:47 inode2 Keepalived_vrrp[28943]: VRRP_Instance(http) setting protocol VIPs. May 17 23:49:47 inode2 Keepalived_healthcheckers[28942]: Netlink reflector reports IP 172.18.67.33 added May 17 23:49:47 inode2 Keepalived_vrrp[28943]: VRRP_Instance(http) Sending gratuitous ARPs on eno16777736 for 172.18.67.33
我们看到inode2节点进入了MASTER模式，再查看inode3的状态
[root@inode3 ~]# systemctl start keepalived [root@inode3 ~]# systemctl status -l keepalived ● keepalived.service - LVS and VRRP High Availability Monitor Loaded: loaded (/usr/lib/systemd/system/keepalived.service; disabled; vendor preset: disabled) Active: active (running) since Wed 2017-05-17 23:51:08 CST; 5s ago Process: 42610 ExecStart=/usr/sbin/keepalived $KEEPALIVED_OPTIONS (code=exited, status=0/SUCCESS) Main PID: 42611 (keepalived) CGroup: /system.slice/keepalived.service ├─42611 /usr/sbin/keepalived -D ├─42612 /usr/sbin/keepalived -D └─42613 /usr/sbin/keepalived -D May 17 23:51:08 inode3 Keepalived_vrrp[42613]: Netlink reflector reports IP fe80::20c:29ff:fe78:24c3 added May 17 23:51:08 inode3 Keepalived_vrrp[42613]: Registering Kernel netlink reflector May 17 23:51:08 inode3 Keepalived_healthcheckers[42612]: Using LinkWatch kernel netlink reflector... May 17 23:51:08 inode3 Keepalived_vrrp[42613]: Registering Kernel netlink command channel May 17 23:51:08 inode3 Keepalived_vrrp[42613]: Registering gratuitous ARP shared channel May 17 23:51:08 inode3 Keepalived_vrrp[42613]: Opening file '/etc/keepalived/keepalived.conf'. May 17 23:51:08 inode3 Keepalived_vrrp[42613]: Configuration is using : 63023 Bytes May 17 23:51:08 inode3 Keepalived_vrrp[42613]: Using LinkWatch kernel netlink reflector... May 17 23:51:08 inode3 Keepalived_vrrp[42613]: VRRP_Instance(http) Entering BACKUP STATE May 17 23:51:08 inode3 Keepalived_vrrp[42613]: VRRP sockpool: [ifindex(2), proto(112), unicast(0), fd(10,11)]
不难看出inode3节点处于BACKUP状态，此时我们将inode2的keepalived服务停掉
[root@inode2 ~]# systemctl stop keepalived [root@inode3 ~]# systemctl status -l keepalived ● keepalived.service - LVS and VRRP High Availability Monitor Loaded: loaded (/usr/lib/systemd/system/keepalived.service; disabled; vendor preset: disabled) Active: active (running) since Wed 2017-05-17 23:51:08 CST; 1min 2s ago Process: 42610 ExecStart=/usr/sbin/keepalived $KEEPALIVED_OPTIONS (code=exited, status=0/SUCCESS) Main PID: 42611 (keepalived) CGroup: /system.slice/keepalived.service ├─42611 /usr/sbin/keepalived -D ├─42612 /usr/sbin/keepalived -D └─42613 /usr/sbin/keepalived -D May 17 23:51:08 inode3 Keepalived_vrrp[42613]: Opening file '/etc/keepalived/keepalived.conf'. May 17 23:51:08 inode3 Keepalived_vrrp[42613]: Configuration is using : 63023 Bytes May 17 23:51:08 inode3 Keepalived_vrrp[42613]: Using LinkWatch kernel netlink reflector... May 17 23:51:08 inode3 Keepalived_vrrp[42613]: VRRP_Instance(http) Entering BACKUP STATE May 17 23:51:08 inode3 Keepalived_vrrp[42613]: VRRP sockpool: [ifindex(2), proto(112), unicast(0), fd(10,11)] May 17 23:52:07 inode3 Keepalived_vrrp[42613]: VRRP_Instance(http) Transition to MASTER STATE May 17 23:52:08 inode3 Keepalived_vrrp[42613]: VRRP_Instance(http) Entering MASTER STATE May 17 23:52:08 inode3 Keepalived_vrrp[42613]: VRRP_Instance(http) setting protocol VIPs. May 17 23:52:08 inode3 Keepalived_healthcheckers[42612]: Netlink reflector reports IP 172.18.67.33 added May 17 23:52:08 inode3 Keepalived_vrrp[42613]: VRRP_Instance(http) Sending gratuitous ARPs on eno16777736 for 172.18.67.33
我们发现inode3节点进入了MASTER状态，因此体现出了高可用的特性
接下来我们测试haproxy的特性，在haproxy的配置文件里有下面这一段
listen stats bind :10086 stats enable stats uri /admin?stats stats auth admin:admin stats admin if TRUE
这段配置可以让我们在浏览器中查看和修改haproxy统计接口启用相关的参数，在浏览器中输入http://172.18.67.33:10086/admin?stats，就会出现下面这种状态，输入账号和密码，就进入了haproxy相关参数配置页面。
在下图中可以看出负载均衡的两台web服务器一台负责动态资源解析，另一台负责静态资源。动态资源的数据存放于后端的nfs服务器上。
接下来我们在浏览器中访问http://172.18.67.33就可以安装wordpress了。至此一个简单的高可用负载均衡服务搭建完毕。
