LVS+keepalived+Nginx实现负载均衡时前后端故障全自动化处理
最近帮朋友公司配了一下LVS,实现全自动化处理故障问题,基本环境如下VIP : 192.168.101.100
LVS1: 192.168.101.103
LVS2: 192.168.101.104
RS1: 192.168.101.105
RS2: 192.168.101.108
后端还有一些服务器IP就不一一写出来了
首先各个服务器上配置好各种需要的环境,所需要依赖包,然后在master 和 backup上装好keepalived 和 LVS,直接yum安装 .不启动LVS,启动keepalived就好,下面直接说配置 全部在keepalived上
# cat /etc/keepalived/keepalived.conf
! Configuration File for keepalived
global_defs {
router_id LVS_DEVEL
}
vrrp_sync_group VGM { ###定义一个vrrp组
group {
VI_1
}
notify_master "/data/shell/arp.sh"
notify_backup "/data/shell/arp.sh"
}
vrrp_script chk_nginx { #### 定义脚本
script "/data/shell/check_nginx.sh"
interval 1
weight -2
}
vrrp_script chk_nginx2 { #### 定义脚本
script "/data/shell/check_nginx2.sh"
interval 1
weight -2
}
vrrp_instance VI_1 {
state MASTER ###backup上配置相同,只需要改下相关权重等
interface eth1
virtual_router_id 51
priority 101
advert_int 1
authentication {
auth_type PASS
auth_pass 1111
}
track_script { #### 执行脚本检测
chk_nginx
chk_nginx2
}
virtual_ipaddress {
192.168.101.100
}
}
virtual_server 192.168.101.100 80 {
delay_loop 2
lb_algo wrr
lb_kind DR
persistence_timeout 60
protocol TCP
real_server 192.168.101.105 80 {
weight 3
HTTP_GET {
url {
path /test.html
digest 9a8acfd3c07a45099bcde1a97a9fb7c5
}
connect_timeout 5
nb_get_retry 3
delay_before_retry 3
connect_port 80
}
}
real_server 192.168.101.108 80 {
weight 3
HTTP_GET {
url {
path /test2.html
digest 9649358e69985102d0da5ff240e1f2a4
}
connect_timeout 5
nb_get_retry 3
delay_before_retry 3
connect_port 80
}
}
}
# cat /etc/keepalived/keepalived.conf
! Configuration File for keepalived
global_defs {
router_id LVS_DEVEL
}
vrrp_sync_group VGM { ###定义一个vrrp组
group {
VI_1
}
notify_master "/data/shell/arp.sh"
notify_backup "/data/shell/arp.sh"
}
vrrp_script chk_nginx { #### 定义脚本
script "/data/shell/check_nginx.sh"
interval 1
weight -2
}
vrrp_script chk_nginx2 { #### 定义脚本
script "/data/shell/check_nginx2.sh"
interval 1
weight -2
}
vrrp_instance VI_1 {
state MASTER ###backup上配置相同,只需要改下相关权重等
interface eth1
virtual_router_id 51
priority 101
advert_int 1
authentication {
auth_type PASS
auth_pass 1111
}
track_script { #### 执行脚本检测
chk_nginx
chk_nginx2
}
virtual_ipaddress {
192.168.101.100
}
}
virtual_server 192.168.101.100 80 {
delay_loop 2
lb_algo wrr
lb_kind DR
persistence_timeout 60
protocol TCP
real_server 192.168.101.105 80 {
weight 3
HTTP_GET {
url {
path /test.html
digest 9a8acfd3c07a45099bcde1a97a9fb7c5
}
connect_timeout 5
nb_get_retry 3
delay_before_retry 3
connect_port 80
}
}
real_server 192.168.101.108 80 {
weight 3
HTTP_GET {
url {
path /test2.html
digest 9649358e69985102d0da5ff240e1f2a4
}
connect_timeout 5
nb_get_retry 3
delay_before_retry 3
connect_port 80
}
}
}
如果你前端开启了防火墙或者selinux,请关闭防火墙测试或者selinux测试,否则你会出现两个master,哈哈,两个vip地址同时出现,或者起不来服务,如果不关闭iptables,请加入以下策略
###允许80端口对外提供服务
-A INPUT -m state --state NEW -m tcp -p tcp --dport 80 -j ACCEPT
###DR模式，当用户请求LVS-DR VIP时，只有DR响应客户端的ARP广播包，允许vrrp虚拟路由器冗余协议-A INPUT -d 224.0.0.0/8 -j ACCEPT
-A INPUT -p vrrp -j ACCEPT
###允许80端口对外提供服务
-A INPUT -m state --state NEW -m tcp -p tcp --dport 80 -j ACCEPT
###DR模式，当用户请求LVS-DR VIP时，只有DR响应客户端的ARP广播包，允许vrrp虚拟路由器冗余协议-A INPUT -d 224.0.0.0/8 -j ACCEPT
-A INPUT -p vrrp -j ACCEPT
下面是几个在master 和backup上执行的高效脚本,检测后端nginx的
# cat arp.sh
#!/bin/bash
VIP=192.168.101.100
GATEWAY=192.168.101.1
/sbin/arping -I eth1 -c 5 -s $VIP $GATEWAY &>/dev/null
########################################################
# cat check_nginx.sh
!/bin/bash
url="http://192.168.101.105" ####realserver web地址检测
status=$(/usr/bin/curl -s --head "$url" | awk '/HTTP/ {print $2}')
if [ "$status" != "200" ]; then
/data/shell/arp.sh start
fi
# cat check_nginx2.sh
!/bin/bash
url="http://192.168.101.108" ####realserver web地址检测
status=$(/usr/bin/curl -s --head "$url" | awk '/HTTP/ {print $2}')
if [ "$status" != "200" ]; then
/data/shell/arp.sh start
fi
# cat arp.sh
#!/bin/bash
VIP=192.168.101.100
GATEWAY=192.168.101.1
/sbin/arping -I eth1 -c 5 -s $VIP $GATEWAY &>/dev/null
########################################################
# cat check_nginx.sh
!/bin/bash
url="http://192.168.101.105" ####realserver web地址检测
status=$(/usr/bin/curl -s --head "$url" | awk '/HTTP/ {print $2}')
if [ "$status" != "200" ]; then
/data/shell/arp.sh start
fi
# cat check_nginx2.sh
!/bin/bash
url="http://192.168.101.108" ####realserver web地址检测
status=$(/usr/bin/curl -s --head "$url" | awk '/HTTP/ {print $2}')
if [ "$status" != "200" ]; then
/data/shell/arp.sh start
fi
arp缓存清理shell{arp.sh}
#!/bin/bash
VIP=192.168.101.100
GATEWAY=192.168.101.1
/sbin/arping -I eth1 -c 5 -s 192.168.101.100 192.168.101.1 &>/dev/null
#!/bin/bash
VIP=192.168.101.100
GATEWAY=192.168.101.1
/sbin/arping -I eth1 -c 5 -s 192.168.101.100 192.168.101.1 &>/dev/null
这个配置会实现 无论前端调度或者后端nginx服务任意一台中断都会在最短的时间响应到正常的请求,经测试,lvs前端两台故障几乎是0秒切换,后端nginx的web服务故障也就是几秒钟的自动切换,如果是DNS解析,请做好解析定时缓存清理
页:
[1]