如何部署带状态检测的负载均衡架构(LVS+Keepalived)?
- By : Will
- Category : Load balancing
- Tags: Keepalived, Load balancing, LVS, 负载均衡
1 理论部分
基于lvs的负载均衡无法检测Real Server的状态,当Real Server的服务单点故障,Director照样把请求转发给故障的机器,造成请求无法得到正常的服务。
所以就需要引入一种检测机制来弥补缺陷,这就需要Keepalived来实现,他不但可以检测Real Server的状态,同时也可以检测Director的状态,达到Failover的目的。
2 实验部分
2.1 实验基础
LVS均衡负载基础:
http://cmdschool.blog.51cto.com/2420395/1702421
以上基础并增加一台副Director主机
2.2 主机信息
Dr1:
director ipaddress=10.168.0.89
vip ipaddress=10.168.0.91
hostname=dr1
Dr2:
director ipaddress=10.168.0.90
vip ipaddress=10.168.0.91
hostname=dr2
Rs1:
real ipaddress=10.168.0.94
vip-lo ipaddress=10.168.0.91
hostname=rs1
Rs2:
real ipaddress=10.168.0.95
vip-lo ipaddress=10.168.0.91
hostname=rs2
2.3 yum源安装
In Director
yum install -y keepalived
2.4 设置配置文件
2.4.1 step1
IN DR1
vim /etc/keepalived/keepalived.conf
加入如下配置:
vrrp_instance VI_1 { state MASTER #备用服务器上为 BACKUP interface eth0 virtual_router_id 51 priority 100 #备用服务器上为90 advert_int 1 authentication { auth_type PASS auth_pass 1111 } virtual_ipaddress { 10.168.0.91 } } virtual_server 10.168.0.91 80 { delay_loop 6 #(每隔10秒查询realserver状态) lb_algo wlc #(lvs 算法) lb_kind DR #(Direct Route) persistence_timeout 60 #(同一IP的连接60秒内被分配到同一台realserver) protocol TCP #(用TCP协议检查realserver状态) real_server 10.168.0.92 80 { weight 100 #(权重) TCP_CHECK { connect_timeout 10 #(10秒无响应超时) nb_get_retry 3 delay_before_retry 3 connect_port 80 } } real_server 10.168.0.93 80 { weight 100 TCP_CHECK { connect_timeout 10 nb_get_retry 3 delay_before_retry 3 connect_port 80 } } }
IN DR2
vim /etc/keepalived/keepalived.conf
加入如下配置:
vrrp_instance VI_1 { state BACKUP #主服务器上为 MASTER interface eth0 virtual_router_id 51 priority 90 #主服务器上为100 advert_int 1 authentication { auth_type PASS auth_pass 1111 } virtual_ipaddress { 10.168.0.91 } } virtual_server 10.168.0.91 80 { delay_loop 6 #(每隔10秒查询realserver状态) lb_algo wlc #(lvs 算法) lb_kind DR #(Direct Route) persistence_timeout 60 #(同一IP的连接60秒内被分配到同一台realserver) protocol TCP #(用TCP协议检查realserver状态) real_server 10.168.0.92 80 { weight 100 #(权重) TCP_CHECK { connect_timeout 10 #(10秒无响应超时) nb_get_retry 3 delay_before_retry 3 connect_port 80 } } real_server 10.168.0.93 80 { weight 100 TCP_CHECK { connect_timeout 10 nb_get_retry 3 delay_before_retry 3 connect_port 80 } } }
注意,DR1&DR2差异仅如下:
state MASTER <-> state BACKUP priority 100 <-> priority 90
2.4.2 step2
IN DR1&DR2
开启端口转发:
echo 1 > /proc/sys/net/ipv4/ip_forward
2.4.3 step3
IN RS1&RS2
启动脚本:
/usr/local/sbin/lvs_dr_rs.sh
2.4.4 step4
IN DR1&DR2
启动keeplive服务:
/etc/init.d/keepalived start
注意:
以下脚本不需要再像在LVS实验中那样执行“/usr/local/sbin/lvs_dr.sh”
没有评论