Centos7.3部署LVS负载均衡

LVS是Linux Virtual Server的简写，意即Linux虚拟服务器，是一个虚拟的服务器集群系统。本项目在1998年5月由章文嵩博士成立，是中国国内最早出现的自由软件项目之一。LVS 简单工作原理：用户请求LVS VIP，LVS根据转发方式和算法，将请求转发后端服务器，后端服务器接受到请求，返回给用户。对用户来说，看不到WEB后端具体的应用。LVS转发方式有三种，分别是NAT、DR、TUN模式，常用算法：RR、LC、WRR、WLC模式等（RR为轮询模式，LC为最少连接模式）LVS NAT原理：用户请求LVS到达director,director将请求的报文的目标地址改成后端的realserver地址，同时将报文的目标端口也改成后端选定的realserver相应端口，最后将报文发送到realserver，realserver将数据返给director，director再把数据发送给用户。（两次请求都经过director，所以访问大的话，director会成为瓶颈）LVS DR原理：用户请求LVS到达director,director将请求的报文的目标MAC地址改成后端的realserver MAC地址，目标IP为VIP（不变），源IP为用户IP地址（保持不变），然后Director将报文发送到realserver，realserver检测到目标为自己本地IP，如果在同一个网段，然后将请求直接返给用户。如果用户跟realserver不在一个网段，则通过网关返回用户。（此种转发效率最高）LVS TUN原理：跟LVS DR类似，也是改变封装MAC地址，多了一层隧道加密。实施环境复杂，比LVS DR模式效率略低。

---

实验环境：两台服务器部署，IP地址为：192.168.10.30、192.168.10.31 VIP：192.168.10.32
[root@server ~]# cat /etc/redhat-release
CentOS Linux release 7.3.1611 (Core)

[root@lvs ~]# yum install ipvsadm keepalived -y
[root@lvs ~]# ipvsadm -version //查看IPVS版本信息
[root@lvs ~]# ipvsadm -Ln //查看LVS转发列表
[root@lvs ~]# ipvsadm -Ln --timeout //查看LVS的超时时间
[root@lvs ~]# ipvsadm --set 5 10 10 //修改LVS对转发过的请求超时时间

---

LVS存在分设与合设，分设就是服务器单独安装LVS，合设就是跟业务网元一起安装在相同服务器上，合设时需要注意25%的几率会出现调度跟业务流出现死循环，需要在iptables里面对入流做标记过滤。
1. 分设主备配置示例：
[root@lvs ~]# vim /etc/keepalived/keepalived.conf //主用

! Configuration File for keepalived

global_defs {
router_id LVS_Master
}

vrrp_instance VI_WEB {
state MASTER
interface bond0
virtual_router_id 110
priority 200
advert_int 3
authentication {
auth_type PASS
auth_pass 1212
}
virtual_ipaddress {
192.168.10.32
}
}

virtual_server 192.168.10.32 80 {
delay_loop 6
lb_algo lc
lb_kind DR
persistence_timeout 60
protocol TCP
real_server 192.168.10.23 80 {
weight 100
TCP_CHECK {
connect_timeout 3
nb_get_retry 3
delay_before_retry 3
}
}
real_server 192.168.10.24 80 {
weight 100
TCP_CHECK {
connect_timeout 3
nb_get_retry 3
delay_before_retry 3
}
}
real_server 192.168.10.25 80 {
weight 100
TCP_CHECK {
connect_timeout 3
nb_get_retry 3
delay_before_retry 3
}
}
real_server 192.168.10.26 80 {
weight 100
TCP_CHECK {
connect_timeout 3
nb_get_retry 3
delay_before_retry 3
}
}
}

[root@lvs ~]# vim /etc/keepalived/keepalived.conf //备用

! Configuration File for keepalived

global_defs {
router_id LVS_Backup
}

vrrp_instance VI_WEB {
state BACKUP
interface bond0
virtual_router_id 110
priority 120
advert_int 3
authentication {
auth_type PASS
auth_pass 1212
}
virtual_ipaddress {
192.168.10.32
}
}

virtual_server 192.168.10.32 80 {
delay_loop 6
lb_algo lc
lb_kind DR
persistence_timeout 60
protocol TCP
real_server 192.168.10.23 80 {
weight 100
TCP_CHECK {
connect_timeout 3
nb_get_retry 3
delay_before_retry 3
}
}
real_server 192.168.10.24 80 {
weight 100
TCP_CHECK {
connect_timeout 3
nb_get_retry 3
delay_before_retry 3
}
}
real_server 192.168.10.25 80 {
weight 100
TCP_CHECK {
connect_timeout 3
nb_get_retry 3
delay_before_retry 3
}
}
real_server 192.168.10.26 80 {
weight 100
TCP_CHECK {
connect_timeout 3
nb_get_retry 3
delay_before_retry 3
}
}
}

2. 合设主备配置示例：

[root@lvs~]# vim /etc/sysconfig/iptables //主用（写备用LVS的MAC）

*mangle
:INPUT ACCEPT [0:0]
:FORWARD ACCEPT [0:0]
:OUTPUT ACCEPT [0:0]
-A PREROUTING -d 192.168.10.32/32 -p tcp -m tcp --dport 80 -m mac ! --mac-source 0c:c4:7a:82:dc:ce -j MARK --set-xmark 0x3
-A INPUT -d 224.0.0.18 -i bond0 -j ACCEPT
COMMIT

[root@lvs ~]# vim /etc/keepalived/keepalived.conf //主用

! Configuration File for keepalived

global_defs {
router_id LVS_Master
}

vrrp_instance VI_WEB {
state MASTER
interface bond0
virtual_router_id 110
priority 200
advert_int 3
authentication {
auth_type PASS
auth_pass 1212
}
virtual_ipaddress {
192.168.10.32
}
}

virtual_server fwmark 3 {
delay_loop 6
lb_algo lc
lb_kind DR
persistence_timeout 60
protocol TCP
real_server 192.168.10.23 80 {
weight 100
TCP_CHECK {
connect_timeout 3
nb_get_retry 3
delay_before_retry 3
}
}
real_server 192.168.10.24 80 {
weight 100
TCP_CHECK {
connect_timeout 3
nb_get_retry 3
delay_before_retry 3
}
}
real_server 192.168.10.25 80 {
weight 100
TCP_CHECK {
connect_timeout 3
nb_get_retry 3
delay_before_retry 3
}
}
real_server 192.168.10.26 80 {
weight 100
TCP_CHECK {
connect_timeout 3
nb_get_retry 3
delay_before_retry 3
}
}
}

[root@lvs~]# vim /etc/sysconfig/iptables //备用（写主用LVS的MAC）

*mangle
:INPUT ACCEPT [0:0]
:FORWARD ACCEPT [0:0]
:OUTPUT ACCEPT [0:0]
-A PREROUTING -d 192.168.10.32/32 -p tcp -m tcp --dport 80 -m mac ! --mac-source 0c:c4:7a:82:dc:cf -j MARK --set-xmark 0x4
-A INPUT -d 224.0.0.18 -i bond0 -j ACCEPT
COMMIT

[root@lvs ~]# vim /etc/keepalived/keepalived.conf

! Configuration File for keepalived

global_defs {
router_id LVS_Backup
}

vrrp_instance VI_WEB {
state BACKUP
interface bond0
virtual_router_id 110
priority 120
advert_int 3
authentication {
auth_type PASS
auth_pass 1212
}
virtual_ipaddress {
192.168.10.32
}
}

virtual_server fwmark 4 {
delay_loop 6
lb_algo lc
lb_kind DR
persistence_timeout 60
protocol TCP
real_server 192.168.10.23 80 {
weight 100
TCP_CHECK {
connect_timeout 3
nb_get_retry 3
delay_before_retry 3
}
}
real_server 192.168.10.24 80 {
weight 100
TCP_CHECK {
connect_timeout 3
nb_get_retry 3
delay_before_retry 3
}
}
real_server 192.168.10.25 80 {
weight 100
TCP_CHECK {
connect_timeout 3
nb_get_retry 3
delay_before_retry 3
}
}
real_server 192.168.10.26 80 {
weight 100
TCP_CHECK {
connect_timeout 3
nb_get_retry 3
delay_before_retry 3
}
}
}

[root@lvs ~]# systemctl enable keepalived && systemctl start keepalived
[root@lvs ~]# systemctl list-unit-files | grep keepalived
[root@lvs ~]# tail -f /var/log/messages

---

如上设置，LVS 主备配置完毕，接下来需要在realserver配置LVS VIP，为什么要在realserver绑定VIP呢？客户端访问director的VIP，director接收请求，将通过相应的算法将请求转发给相应的realserver。在转发的过程中，会修改请求包的目的mac地址，目的ip地址不变。Realserver接收请求，并直接响应客户端。这时便出现一个问题，director此时与realserver位于同一个网络中，当director直接将请求转发给realserver时，realserver检测到该请求包的目的ip是vip而并非自己，便会丢弃，而不会响应。为了解决这个问题，所以需要在所有Realserver上都配上VIP。为什么一定要配置在lo接口上呢？在realserver上的lo口配置VIP,这样限制了VIP不会在物理交换机上产生MAC地址表，从而避免IP冲突。
[root@realserver ~]# vim /usr/bin/realserver.sh

#! /bin/bash
#description: Config realserver lo and apply noarp
LVS_VIP=192.168.10.32
case "$1" in
start)
ifconfig lo:0 $LVS_VIP netmask 255.255.255.255 broadcast $LVS_VIP
/sbin/route add -host $LVS_VIP dev lo:0
echo "1" >/proc/sys/net/ipv4/conf/bond0/arp_ignore
echo "2" >/proc/sys/net/ipv4/conf/bond0/arp_announce
echo "1" >/proc/sys/net/ipv4/conf/all/arp_ignore
echo "2" >/proc/sys/net/ipv4/conf/all/arp_announce
sysctl -p >/dev/null 2>&1
echo "RealServer Start OK"
exit 0
;;
stop)
ifconfig lo:0 down
route del $EPG_VIP >/dev/null 2>&1
echo "0" >/proc/sys/net/ipv4/conf/bond0/arp_ignore
echo "0" >/proc/sys/net/ipv4/conf/bond0/arp_announce
echo "0" >/proc/sys/net/ipv4/conf/all/arp_ignore
echo "0" >/proc/sys/net/ipv4/conf/all/arp_announce
echo "RealServer Stoped"
exit 1
;;
status)

Status of LVS-DR real server.

    islothere=`/sbin/ifconfig lo:0 | grep $LVS_VIP`
    isrothere=`netstat -rn | grep "lo:0" | grep $LVS_VIP`
    if [ ! "$islothere" -o ! "isrothere" ];then
        # Either the route or the lo:0 device
        echo "LVS-DR real server Stopped."
    else
        echo "LVS-DR Running."
    fi

;;
*)

Invalid entry.

    echo "$0: Usage: $0 {start|status|stop}"
    exit 1

;;
esac

[root@realserver ~]# chmod +x /usr/bin/realserver.sh
[root@realserver ~]# sh /usr/bin/realserver.sh start

转载于:https://blog.51cto.com/263393281/2045649