keepalived双主模型高可用+lvs
拓扑图：http://luoshixin.blog.运维网.com/attachment/201409/22/2817266_1411427483jYO0.png
环境准备：
Centos6.5x86_64 关闭防火墙和Selinux
node5: eth0:192.168.1.190/24 VIP：192.168.1.121/24
node1：eth1:192.168.1.191/24 VIP：192.168.1.122/24
node2：RIP：eth0: 192.168.19.2/24
node3：RIP：eth0: 192.168.19.3/24 所有节点网关/DNS都为：192.168.1.1
每个服务器的hosts文件
# cat /etc/hosts
127.0.0.1 localhost localhost.localdomain localhost4 localhost4.localdomain4
::1 localhost localhost.localdomain localhost6 localhost6.localdomain6
192.168.19.190 node5.luo.com node5
192.168.19.191 node1.luo.com node1
192.168.19.192 node2.luo.com node2
192.168.19.193 node3.luo.com node3
时间同步（每个节点都需要做）
为方便实验使用笔记本作为时间服务器，参考http://blog.csdn.net/irlover/article/details/7314530
ntpdate 192.168.1.105
如我想每隔10分钟和win7同步时间可以使用crontab -e 编辑crontab工作内容如下：
*/10 * * * * /usr/sbin/ntpdate 192.168.1.105 &>/dev/null&&/sbin/hwclock-w
配置后端real server ：node2与node3
node2与node3都执行下面的配置
echo 1 > /proc/sys/net/ipv4/conf/all/arp_ignore
echo 1 > /proc/sys/net/ipv4/conf/eth0/arp_ignore
echo 2 > /proc/sys/net/ipv4/conf/all/arp_announce
echo 2 > /proc/sys/net/ipv4/conf/eth0/arp_ignore
ifconfig lo:0 192.168.1.121 netmask 255.255.255.255 broadcast 192.168.1.121
route add -host 192.168.1.121 dev lo:0
ifconfig lo:1 192.168.1.122 netmask 255.255.255.255 broadcast 192.168.1.122
route add -host 192.168.1.122 dev lo:1
分别为node2与node3安装httpd并写好测试页面
# yum install httpd
node2测试页面
# echo "node2.luo.com" > /var/www/html/index.html
# service httpd restart
# curl http://192.168.1.192
node3测试页面
#echo "node3.luo.com" > /var/www/html/index.html
# service httpd restart
# curl http://192.168.1.193
二、配置前端的Director
node5与node1安装配置基本相同
在两个节点上都安装keepalived
# yum install -y keepalived ipvsadm
node5的配置与node1的不同之处在与两个实例中的角色刚好相反，node5为MASTER时，node1为BACKUP，其相应的优先级也有所改变，其他配置相同。
node5配置文件
! 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 localhost
smtp_connect_timeout 30
router_id LVS_DEVEL
}
vrrp_instance VI_1 {
state MASTER
interface eth0
virtual_router_id 61
priority 100
advert_int 1
authentication {
auth_type PASS
auth_pass 123456
}
virtual_ipaddress {
192.168.1.121
}
}
vrrp_instance VI_2 {
state BACKUP
interface eth0
virtual_router_id 71
priority 99
advert_int 1
authentication {
auth_type PASS
auth_pass 123456
}
virtual_ipaddress {
192.168.1.122
}
}
virtual_server 192.168.1.121 80 {
delay_loop 6
lb_algo rr
lb_kind DR
nat_mask 255.255.255.0
protocol TCP
real_server 192.168.1.192 80 {
weight 1
HTTP_GET {
url {
path /
status_code 200
}
connect_timeout 2
nb_get_retry 3
delay_before_retry 1
}
}
real_server 192.168.1.193 80 {
weight 1
HTTP_GET {
url {
path /
status_code 200
}
connect_timeout 2
nb_get_retry 3
delay_before_retry 1
}
}
}
virtual_server 192.168.1.122 80 {
delay_loop 6
lb_algo rr
lb_kind DR
nat_mask 255.255.255.0
protocol TCP
real_server 192.168.1.192 80 {
weight 1
HTTP_GET {
url {
path /
status_code 200
}
connect_timeout 2
nb_get_retry 3
delay_before_retry 1
}
}
real_server 192.168.1.193 80 {
weight 1
HTTP_GET {
url {
path /
status_code 200
}
connect_timeout 2
nb_get_retry 3
delay_before_retry 1
}
}
}
node1配置文件
! 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 localhost
smtp_connect_timeout 30
router_id LVS_DEVEL
}
vrrp_instance VI_1 {
state BACKUP
interface eth1
virtual_router_id 61
priority 99
advert_int 1
authentication {
auth_type PASS
auth_pass 123456
}
virtual_ipaddress {
192.168.1.121
}
}
vrrp_instance VI_2 {
state MASTER
interface eth1
virtual_router_id 71
priority 100
advert_int 1
authentication {
auth_type PASS
auth_pass 123456
}
virtual_ipaddress {
192.168.1.122
}
}
virtual_server 192.168.1.121 80 {
delay_loop 6
lb_algo rr
lb_kind DR
nat_mask 255.255.255.0
protocol TCP
real_server 192.168.1.192 80 {
weight 1
HTTP_GET {
url {
path /
status_code 200
}
connect_timeout 2
nb_get_retry 3
delay_before_retry 1
}
}
real_server 192.168.1.193 80 {
weight 1
HTTP_GET {
url {
path /
status_code 200
}
connect_timeout 2
nb_get_retry 3
delay_before_retry 1
}
}
}
virtual_server 192.168.1.122 80 {
delay_loop 6
lb_algo rr
lb_kind DR
nat_mask 255.255.255.0
protocol TCP
real_server 192.168.1.192 80 {
weight 1
HTTP_GET {
url {
path /
status_code 200
}
connect_timeout 2
nb_get_retry 3
delay_before_retry 1
}
}
real_server 192.168.1.193 80 {
weight 1
HTTP_GET {
url {
path /
status_code 200
}
connect_timeout 2
nb_get_retry 3
delay_before_retry 1
}
}
}
测试
先启动node5节点的keepalived服务，可以看到两个VIP地址都配置在了node5的eth0网卡上。
ipvs规则的启动效果为，可以看到两组规则也都启动在了node5上。
http://luoshixin.blog.运维网.com/attachment/201409/22/2817266_1411427483i2Ko.png
启动node1的keepalived服务，查看相关信息
http://luoshixin.blog.运维网.com/attachment/201409/22/2817266_1411427483AHRZ.png
模拟node5节点服务器故障，此处直接停止node5的keepalived服务，查看node1信息都无误，两个VIP都到node1上面了
http://luoshixin.blog.运维网.com/attachment/201409/22/2817266_1411427483gCN4.png
至此，实验完毕，如有错误之处，请多多指教。
页:
[1]