CentOS6.4 LVS+keepalived高可用负载均衡服务配置

CentOS6.4 LVS+keepalived高可用负载均衡服务配置

转自

http://songxj.blog.51cto.com/620981/1171882

一：测试环境信息

服务器DELL-R720

虚拟化 KVM

虚拟机4台： ipvs01，ipvs02，web01，web02

[root@KVM01~]# virsh list

 Id    Name                           State

----------------------------------------------------

 1     SN-web01                    running

 2     SN-web02                    running

 3     SN-ipvs01                    running

 4     SN-ipvs02                    running

 

网络环境：

Ipvs01:192.168.40.90

Ipvs02:192.168.40.91

Web01:192.168.40.86

Web02:192.168.40.87

Vip:192.168.40.6

 

系统版本：

物理机CentOS6.4 64bit

虚拟机CentOS6.4 64bit

二：服务配置

软件包6.4系统已经自带ipvs1.25和keepalived1.2.7 测试用yum自行安装。

 

主LVS服务器ipvs01配置：

[root@ipvs01 ~]# yum list >yum.list

[root@ipvs01 ~]# cat yum.list|grep ipvs

ipvsadm.x86_64                         1.25-10.el6                       @base

[root@ipvs01 ~]# cat yum.list |grep keepalived

keepalived.x86_64                       1.2.7-3.el6                        @base

[root@ipvs01 ~]#yum install ipvsadm keepalived

[root@ipvs01 ~]#ipvsadm

[root@ipvs01 ~]#lsmod |grep ip_vs  查看模块是否加载成功

ip_vs_rr                1420  1

ip_vs_wrr               2179  0

ip_vs                 115643  5 ip_vs_rr,ip_vs_wrr

libcrc32c               1246  1 ip_vs

ipv6                  321422  11 ip_vs

[root@ipvs01 ~]#cd /etc/keepalived/

[root@ipvs01 ~]#cp -a keepalived.conf ./keepalived.bak

[root@ipvs01 ~]#vi keepalived.conf

 

 

! 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_MASTER

}

 

vrrp_instance VI_1 {

    state MASTER   #角色主MASTER，备服务器改为BACKUP

    interface eth0   #HA侦听接口

    virtual_router_id 51   #虚拟路由标记ID，同一组vrrp一致

   priority 100   #优先级自定义，MASTER高于BACKUP即可

    advert_int 1   #HA 侦听间隔：1秒

    authentication {   #认证形式

        auth_type PASS   #认证类型PASS：PASS/AH 2种可选

        auth_pass 1111   #认证密码，同一组vrrp密码一致

    }

    virtual_ipaddress {   #虚拟服务地址，可以多个，分多行

        192.168.40.6

    }

}

 

virtual_server 192.168.40.6 80 {   #虚拟服务地址和端口

    delay_loop 6   #运行情况检查，单位秒

    lb_algo rr   #负载调度算法，RR为轮询

    lb_kind DR   #LVS负载工作模式为DR，三大模式NAT,TUN,DR

    nat_mask 255.255.255.0   #网络掩码

    #persistence_timeout 50   #会话保持时间，50秒内分配同一节点，测试时候为了查看均衡效果可以先注释掉

    protocol TCP    #协议类型TCP/UDP

 

    real_server 192.168.40.86 80 {   #配置真实服务器节点1 的IP和端口

        weight 5            #权值大小，越大权值越高

        TCP_CHECK {           #realserver 状态检测时间，单位秒

            connect_timeout 3           #连接超时时间3秒

            nb_get_retry 3                  #重试次数:3次

            delay_before_retry 3       #重试间隔

            connect_port 80                #连接端口

            }

    }

real_server 192.168.40.87 80 {

        weight 5

        TCP_CHECK {

            connect_timeout 3

            nb_get_retry 3

            delay_before_retry 3

            connect_port 80

        }

    }

}

 

备服务器ipvs02上将主服务器配置文件复制过去，然后修改2处：

state BACKUP

priority 100

其他保持不变。

 

三：服务节点配置：

Web01，web02 ，http服务安装略

[root@web01 ~]# cat /var/www/html/index.html

<h1>WEB01/192.168.40.86</h1>

[root@web02 ~]# cat /var/www/html/index.html

<h1>WEB02/192.168.40.87</h1>

 

在web01和web02上分别运行脚本

ifconfig lo:0 192.168.40.6 netmask 255.255.255.0 up    

route add -host 192.168.40.6 dev lo:0       

echo "1" >/proc/sys/net/ipv4/conf/lo/arp_ignore       

echo "2" >/proc/sys/net/ipv4/conf/lo/arp_announce       

echo "1" >/proc/sys/net/ipv4/conf/all/arp_ignore       

echo "2" >/proc/sys/net/ipv4/conf/all/arp_announce

 

执行完毕后查看路由有一条虚拟IP的路由指向lo接口

[root@web01 ~]# route -n

Kernel IP routing table

Destination     Gateway         Genmask         Flags Metric Ref    Use Iface

192.168.40.6    0.0.0.0         255.255.255.255 UH    0      0        0 lo

 

四：测试LVS

将ipvs01，ipvs02 ，keepalived服务启动

将web01，web02， http服务启动

在主LVS服务器上查看lvs状态如下 说明配置生效

[root@ipvs01 ~]# ipvsadm -ln

IP Virtual Server version 1.2.1 (size=4096)

Prot LocalAddress:Port Scheduler Flags

  -> RemoteAddress:Port           Forward Weight ActiveConn InActConn

TCP  192.168.40.6:80 rr persistent 50

  -> 192.168.40.86:80             Route   5      0          0        

  -> 192.168.40.87:80             Route   5      0          0 

 

从ip a 命令来看 虚拟IP现在是在ipvs01上的，而从服务器上ip a没有虚拟IP

[root@ipvs01 ~]# ip a

1: lo: <LOOPBACK,UP,LOWER_UP> mtu 16436 qdisc noqueue state UNKNOWN

    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00

    inet 127.0.0.1/8 scope host lo

    inet6 ::1/128 scope host

       valid_lft forever preferred_lft forever

2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000

    link/ether 52:54:00:06:88:f4 brd ff:ff:ff:ff:ff:ff

    inet 192.168.40.90/24 brd 192.168.40.255 scope global eth0

    inet 192.168.40.6/32 scope global eth0

    inet6 fe80::5054:ff:fe06:88f4/64 scope link

       valid_lft forever preferred_lft forever

 

>测试负载均衡效果：

打开浏览器测试：

刷新一下

 

>测试一下HA效果

将ipvs01 keepalived服务器停掉，然后查看ipvs02 ip a状态

[root@ipvs01 ~]# /etc/init.d/keepalived stop

Stopping keepalived:                                       [  OK  ]

[root@SN349_ipvs02 ~]# ip a

1: lo: <LOOPBACK,UP,LOWER_UP> mtu 16436 qdisc noqueue state UNKNOWN

    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00

    inet 127.0.0.1/8 scope host lo

    inet6 ::1/128 scope host

       valid_lft forever preferred_lft forever

2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000

    link/ether 52:54:00:45:f8:78 brd ff:ff:ff:ff:ff:ff

    inet 192.168.40.91/24 brd 192.168.40.255 scope global eth0

    inet 192.168.40.6/32 scope global eth0

    inet6 fe80::5054:ff:fe45:f878/64 scope link

       valid_lft forever preferred_lft forever

 

此时ipvs02绑定了虚拟IP 顶替了ipvs01的服务。