LVS理论篇
1、Client 向目标VIP 发出请求，Director（负载均衡器）接收。此时IP 包头及数据帧信息为：
2、Director 根据负载均衡算法选择RealServer_1，不修改也不封装IP 报文，而是将数据帧的MAC 地址改为RealServer_1 的MAC 地址，然后在局域网上发送。IP 包头及数据帧头信息如下：
3、RealServer_1 收到这个帧，解封装后发现目标IP 与本机匹配（RealServer 事先绑定了VIP，必须的！）于是处理这个报文，随后重新封装报文，发送到局域网。此时IP 包头及数据帧头信息如下：
4、Client 将收到回复报文。Client 认为得到正常的服务，而不会知道是哪台服务器处理的（注意，如果跨网段，那么报文通过路由器经由Internet 返回给用户）
LVS-DR 中的ARP 问题
在LVS-DR 负载均衡群集中，负载均衡器与节点服务器都要配置相同的VIP 地址，在局
域网中具有相同的IP 地址，势必会造成各种服务器ARP 通信紊乱。当一个ARP 广播发送到
LVS-DR 群集时，因为负载均衡器和节点服务器都是连接到相同的网络上的，它们都会接收
到ARP 广播，这个时候，应该只有前段的负载均衡器进行响应，其他节点服务器不应该响
应ARP 广播。
1、对节点服务器进行处理，使其不影响针对VIP 的ARP 请求。
使用虚拟接口lo:0 承载VIP 地址
设置内核参数arp_ignore=1，系统只响应目的IP 为本地IP 的ARP 请求
2、RealServer 返回报文（源IP 是VIP）经由路由器转发，在重新封装报文时，需要现货区路由器的MAC 地址，发送ARP 请求保重的源IP 地址，而不使用发送接口（如eth0）的IP地址
3、路由器收到ARP 请求后，将跟新ARP 表项，原有的VIP 对应的Director 的MAC 地址将会被跟新为VIP 对应的RealServer 的MAC 地址。
此时新来的报文请求，路由器根据ARP 表项，会将该报文转发给RealServer，从而导致
Director 的VIP 失效！
15
云计算
3、路由器收到ARP 请求后，将跟新ARP 表项，原有的VIP 对应的Director 的MAC 地址将会被跟新为VIP 对应的RealServer 的MAC 地址。此时新来的报文请求，路由器根据ARP 表项，会将该报文转发给RealServer，从而导致Director 的VIP 失效！
解决方法：
对节点服务器进行处理，设置内核参数arp_announce=2，系统不使用IP 包的源地址来
设置ARP 请求的源地址，而选择发送接口的IP 地址。解决ARP 的两个问题的设置方法：修改/etc/sysctl.conf 文件
net.ipv4.conf.lo.arp_ignore = 1
net.ipv4.conf.lo.arp_announce = 2
net.ipv4.conf.all.arp_ignore = 1
net.ipv4.conf.all.arp_announce = 2
LVS+DR实战篇：
实验环境：
VIP： 192.168.18.41
BL： 192.168.18.31
web1: 192.168.18.32
web2: 192.168.18.33
nfs: 192.168.18.34
NFS共享存储：
[[email protected] ~]# mount /dev/cdrom /media/cdrom/
mount: block device /dev/sr0 is write-protected, mounting read-only
[[email protected] ~]# yum -y install nfs-utils rpcbind
[[email protected] ~]# rpm -q nfs-utils rpcbind
[[email protected] ~]# mkdir /www
[[email protected] ~]# vi /etc/exports
/www 192.168.18.0/24(ro,sync,no_root_squash)
[[email protected] ~]# service rpcbind start
正在启动 rpcbind： [确定]
[[email protected] ~]# service nfs start
启动 NFS 服务： [确定]
启动 NFS mountd： [确定]
启动 NFS 守护进程： [确定]
正在启动 RPC idmapd： [确定]
[[email protected] ~]# showmount -e 192.168.18.34
Export list for 192.168.18.34:
/www 192.168.18.0/24
[[email protected] ~]# chkconfig rpcbind on
[[email protected] ~]# chkconfig nfs on
[[email protected] ~]# echo "<h1>ce shi ye</h1>">/www/index.html
web服务器配置:
web1
[[email protected] ~]# rpm -q httpd
httpd-2.2.15-29.el6.centos.x86_64
[[email protected] ~]# vi /etc/httpd/conf/httpd.conf
[[email protected] ~]# yum -y install nfs-utils
[[email protected] ~]# service httpd start
正在启动 httpd： [确定]
[[email protected] ~]# mount 192.168.18.34:/www /var/www/html/
[[email protected] ~]# df -h
Filesystem Size Used Avail Use% Mounted on
/dev/mapper/vg_doo-lv_root 18G 3.9G 13G 24% /
tmpfs 383M 0 383M 0% /dev/shm
/dev/sda1 485M 35M 426M 8% /boot
192.168.18.34:/www 18G 1.3G 16G 8% /var/www/html
[[email protected] ~]# vi /etc/fstab
192.168.18.34:/www /var/www/html nfs defaults,_netdev 1 2
（备份，第二检测）
web2:（同web1）
3、LVS-NAT 部署
ipvsadm 工具参数说明：
-A 添加虚拟服务器
-D 删除虚拟服务器
-C 删除所有配置条目
-E 修改虚拟服务器
-L 或-l，列表查看
-n 不做解析，以数字形式显示
-c 输出当前IPVS 连接
-a 添加真实服务器
-d 删除某个节点
-t 指定VIP 地址及TCP 端口
-s 指定负载调度算法，rr|wrr|lc|wlc|lblc|lblcr|dh|sh|sed|nq，默认wlc
-m NAT 群集模式
-g DR 模式
-i TUN 模式
-w 设置权重（权重为0 时表示暂停节点）
--help 帮助
BL：
[[email protected] ~]# modprobe ip_vs//加载ip_vs模块
[[email protected] ~]# yum -y install ipvsadm
[[email protected] ~]# service ipvsadm stop
ipvsadm: Clearing the current IPVS table: [确定]
ipvsadm: Unloading modules: [确定]
[[email protected] ~]# ipvsadm -C
[[email protected] ~]# vi /opt/vip.sh
#!/bin/bash
# VIP
VIP="192.168.18.41"
/sbin/ifconfig eth1:vip $VIP broadcast $VIP netmask 255.255.255.255
/sbin/route add -host $VIP dev eth1:vip
~
[[email protected] ~]# chmod +x /opt/vip.sh
[[email protected] ~]# /opt/vip.sh
[[email protected] ~]# route -n
Kernel IP routing table
Destination Gateway Genmask Flags Metric Ref Use Iface
192.168.18.41 0.0.0.0 255.255.255.255 UH 0 0 0 eth1
192.168.18.0 0.0.0.0 255.255.255.0 U 0 0 0 eth1
169.254.0.0 0.0.0.0 255.255.0.0 U 1002 0 0 eth1
0.0.0.0 192.168.18.2 0.0.0.0 UG 0 0 0 eth1
[[email protected] ~]# ipvsadm -A -t 192.168.18.41:80 -s rr
[[email protected] ~]# ipvsadm -a -t 192.168.18.41:80 -r 192.168.18.32:80 -g
[[email protected] ~]# ipvsadm -a -t 192.168.18.41:80 -r 192.168.18.33:80 -g
[[email protected] ~]# ipvsadm -ln
IP Virtual Server version 1.2.1 (size=4096)
Prot LocalAddress:Port Scheduler Flags
-> RemoteAddress:Port Forward Weight ActiveConn InActConn
TCP 192.168.18.41:80 rr
-> 192.168.18.32:80 Route 1 0 0
-> 192.168.18.33:80 Route 1 0 0
[[email protected] ~]# ipvsadm-save >/etc/sysconfig/ipvsadm
对web1：
[[email protected] ~]# vi /opt/lvs-dr
#!/bin/bash
# lvs-dr
VIP="192.168.18.41"
/sbin/ifconfig lo:vip $VIP broadcast $VIP netmask 255.255.255.255
/sbin/route add -host $VIP dev lo:vip
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
[[email protected] ~]# chmod +x /opt/lvs-dr
[[email protected] ~]# /opt/lvs-dr
[[email protected] ~]# 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
inet 192.168.2.41/32 brd 192.168.2.41 scope global lo:vip
[[email protected] ~]# scp /opt/lvs-dr 192.168.2.4:/opt
对web2也一样操作
对BL 可以看到分配的SYNC请求：
[[email protected] ~]# ipvsadm -lnc
IPVS connection entries
pro expire state source virtual destination
TCP 14:59 ESTABLISHED 192.168.18.140:53385 192.168.18.41:80 192.168.18.33:80
TCP 01:24 FIN_WAIT 192.168.18.140:53380 192.168.18.41:80 192.168.18.32:80
TCP 01:58 FIN_WAIT 192.168.18.140:53387 192.168.18.41:80 192.168.18.33:80
TCP 01:24 FIN_WAIT 192.168.18.140:53379 192.168.18.41:80 192.168.18.33:80
TCP 01:25 FIN_WAIT 192.168.18.140:53382 192.168.18.41:80 192.168.18.32:80
加上keepalived：
对BL：
实验环境：
VIP： 192.168.18.41
BL1:193.168.18.30（从）
BL2： 192.168.18.31(主)
web1: 192.168.18.32
web2: 192.168.18.33
nfs: 192.168.18.34 (raid+LVM)
[[email protected] ~]# yum -y install keepalived
[[email protected] ~]# cd /etc/keepalived/
[[email protected] keepalived]# cp keepalived.conf keepalived.conf.bak
[[email protected] keepalived]# vi keepalived.conf
! Configuration File for keepalived
global_defs {
notification_email {
[email protected]//自己的警报邮箱
}
notification_email_from [email protected]//发送邮箱
smtp_server ping.com.cn//邮件服务器的SMTP地址
smtp_connect_timeout 30//smtp服务器超时时间
router_id LVS_DEVEL_BLM//邮件标题识别，可乱写
}
vrrp_instance VI_1 {
state MASTER
interface eth1
virtual_router_id 51//同一个虚拟VIP的路由标记，同一个keepalived下主备必须一样
priority 100//优先级，数字越大优先级越高，1-254
advert_int 2//MASTER与SLAVE之间的同步检查时间
authentication {
auth_type PASS//验证类型，有PASS和AH两种
auth_pass 1111//同一个keepalived下必须密码一样
}
virtual_ipaddress {
192.168.18.41//VIP，可以设置多个
}
}
virtual_server 192.168.18.41 80 {//设置VIP
delay_loop 2//健康检查时间
lb_algo rr//轮询算法
lb_kind DR//负载均衡机制NAT ,DR,TUN
! nat_mask 255.255.255.0//非NAT，要注释，下同
! persistence_timeout 300//存留超时时间
protocol TCP
real_server 192.168.18.32 80 {
weight 1//权重值
TCP_CHECK {//readserver的状态检测部分
connect_timeout 10//10秒无响应超时
nb_get_retry 3//重试次数
delay_before_retry 3//两个重试时间间隔为3秒
connect_port 80//检测连接端口
}
}
real_server 192.168.18.33 80 {
weight 1
TCP_CHECK {
connect_timeout 10
nb_get_retry 3
delay_before_retry 3
connect_port 80
}
}
}
[[email protected] keepalived]# service keepalived start
正在启动 keepalived： [确定]
[[email protected] keepalived]# ipvsadm -ln
IP Virtual Server version 1.2.1 (size=4096)
Prot LocalAddress:Port Scheduler Flags
-> RemoteAddress:Port Forward Weight ActiveConn InActConn
TCP 192.168.18.41:80 rr
-> 192.168.18.32:80 Route 1 0 0
-> 192.168.18.33:80 Route 1 0 0
对从BL2：192.168.18.30
[[email protected] ~]# modprobe ip_vs
[[email protected] ~]# yum -y install keepalived ipvsadm
[[email protected] ~]# cd /etc/keepalived/
[[email protected] keepalived]# cp keepalived.conf keepalived.conf.bak
[[email protected] keepalived]# vi keepalived.conf.bak
! Configuration File for keepalived
global_defs {
notification_email {
[email protected]
}
notification_email_from [email protected]
smtp_server ping.com.cn
smtp_connect_timeout 30
router_id LVS_DEVEL_BLM
}
vrrp_instance VI_1 {
state BACKUP
interface eth1
virtual_router_id 51
priority 99
advert_int 2
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
192.168.18.41
}
}
virtual_server 192.168.18.41 80 {
delay_loop 2
lb_algo rr
lb_kind DR
! nat_mask 255.255.255.0
! persistence_timeout 300
protocol TCP
real_server 192.168.18.32 80 {
weight 1
TCP_CHECK {
connect_timeout 10
nb_get_retry 3
delay_before_retry 3
connect_port 80
}
}
real_server 192.168.18.33 80 {
weight 1
TCP_CHECK {
connect_timeout 10
nb_get_retry 3
delay_before_retry 3
connect_port 80
}
}
}
[[email protected] keepalived]# ip a//没有VIP，就对了
[[email protected] keepalived]# ipvsadm -ln
IP Virtual Server version 1.2.1 (size=4096)
Prot LocalAddress:Port Scheduler Flags
-> RemoteAddress:Port Forward Weight ActiveConn InActConn
TCP 192.168.18.41:80 rr
-> 192.168.18.32:80 Route 1 0 0
-> 192.168.18.33:80 Route 1 0 0
对Web1：
[[email protected] ~]# service httpd stop
停止 httpd： [确定]
对BL：
[[email protected] keepalived]# ipvsadm -ln
IP Virtual Server version 1.2.1 (size=4096)
Prot LocalAddress:Port Scheduler Flags
-> RemoteAddress:Port Forward Weight ActiveConn InActConn
TCP 192.168.18.41:80 rr
-> 192.168.18.33:80 Route 1 0 0
对web1：
[[email protected] ~]# service httpd start
正在启动 httpd： [确定]
对BL：
[[email protected] keepalived]# ipvsadm -ln
IP Virtual Server version 1.2.1 (size=4096)
Prot LocalAddress:Port Scheduler Flags
-> RemoteAddress:Port Forward Weight ActiveConn InActConn
TCP 192.168.18.41:80 rr
-> 192.168.18.32:80 Route 1 0 0
-> 192.168.18.33:80 Route 1 0 0
高可用测试：
对主BL：
[[email protected] keepalived]# 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: eth1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
link/ether 00:0c:29:9d:cf:f3 brd ff:ff:ff:ff:ff:ff
inet 192.168.18.31/24 brd 192.168.18.255 scope global eth1
inet 192.168.18.41/32 scope global eth1
inet6 fe80::20c:29ff:fe9d:cff3/64 scope link
valid_lft forever preferred_lft forever
[[email protected] keepalived]# service keepalived stop
停止 keepalived： [确定]
对BL从：
[[email protected] keepalived]# 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: eth1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
link/ether 00:0c:29:34:d3:f8 brd ff:ff:ff:ff:ff:ff
inet 192.168.18.30/24 brd 192.168.18.255 scope global eth1
inet 192.168.18.41/32 scope global eth1
inet6 fe80::20c:29ff:fe34:d3f8/64 scope link
valid_lft forever preferred_lft forever
[[email protected] keepalived]# ipvsadm -l -n
IP Virtual Server version 1.2.1 (size=4096)
Prot LocalAddress:Port Scheduler Flags
-> RemoteAddress:Port Forward Weight ActiveConn InActConn
TCP 192.168.18.41:80 rr
-> 192.168.18.32:80 Route 1 0 0
-> 192.168.18.33:80 Route 1 0 0
NFS+raid5+LVM 192.168.18.34
对NFS：
[[email protected] ~]# fdisk -l|grep dev
Disk /dev/sda: 21.5 GB, 21474836480 bytes
/dev/sda1 * 1 64 512000 83 Linux
/dev/sda2 64 2611 20458496 8e Linux LVM
Disk /dev/sdb: 2147 MB, 2147483648 bytes
Disk /dev/sdc: 2147 MB, 2147483648 bytes
Disk /dev/sdd: 2147 MB, 2147483648 bytes
Disk /dev/sde: 2147 MB, 2147483648 bytes
Disk /dev/mapper/vg_doo2-lv_root: 18.9 GB, 18865979392 bytes
Disk /dev/mapper/vg_doo2-lv_swap: 2080 MB, 2080374784 bytes
[[email protected] ~]# yum -y install parted
[[email protected] ~]# parted /dev/sdb
GNU Parted 2.1
使用 /dev/sdc
Welcome to GNU Parted! Type ‘help‘ to view a list of commands.
(parted) mklabel
新的磁盘标签类型？ gpt
(parted) mkpart
分区名称？ []? a
文件系统类型？ [ext2]? ext3
起始点？ 1
结束点？ -1
(parted) p
Model: VMware, VMware Virtual S (scsi)
Disk /dev/sdc: 2147MB
Sector size (logical/physical): 512B/512B
Partition Table: gpt
Number Start End Size File system Name 标志
1 1049kB 2146MB 2145MB a
(parted) q
信息: You may need to update /etc/fstab.
同样的方法对 c d e....
[[email protected] ~]# yum -y install mdadm
[[email protected] ~]# mdadm -Cv /dev/md5 -a yes -n3 -x1 -l5 /dev/sd[b-e]1
mdadm: layout defaults to left-symmetric
mdadm: layout defaults to left-symmetric
mdadm: super1.x cannot open /dev/sdb1: Device or resource busy
mdadm: Cannot use /dev/sdb1: It is busy
mdadm: cannot open /dev/sdb1: Device or resource busy
[[email protected] ~]# mkfs.ext3 /dev/md5
[[email protected] ~]# mdadm -D -s >/etc/mdadm.conf
[[email protected] ~]# sed -i ‘1 s/$/ auto=yes/‘ /etc/mdadm.conf
LVM:
[[email protected] ~]# pvcreate /dev/md5
Physical volume "/dev/md5" successfully created
[[email protected] ~]# vgcreate vg0 /dev/md5
Volume group "vg0" successfully created
[[email protected] ~]# lvcreate -L 2G -n web vg0
Logical volume "web" created
[[email protected] ~]# mkdir /web
[[email protected] ~]# mkfs.ext4 /dev/vg0/web
[[email protected] ~]# mount /dev/vg0/web /web/
[[email protected] ~]# echo "doo">/web/index.html
[[email protected] ~]# vim /etc/fstab
……
/dev/vg0/web /web ext4 defaults 1 2
[[email protected] ~]# vim /etc/exports
/web 192.168.1.0/24(rw,sync,no_root_squash)
[[email protected] ~]# /etc/init.d/rpcbind start
[[email protected] ~]# /etc/init.d/nfs start
[[email protected] ~]# showmount -e 192.168.18.34
Export list for 192.168.18.34:
/web 192.168.18.0/24
对web1和web2挂载
[[email protected] ~]# vim /etc/fstab
……
192.168.18.34:/web /var/www/html nfs defaults,_netdev 1 2
[[email protected] ~]# yum -y install nfs-utils
[[email protected] ~]# mount 192.168.18.34:/web /var/www/html/
客户端测试：
多刷新几次
[[email protected] keepalived]# ipvsadm -L -n -c
IPVS connection entries
pro expire state source virtual destination
TCP 01:58 FIN_WAIT 192.168.18.140:53988 192.168.18.41:80 192.168.18.32:80
TCP 01:57 FIN_WAIT 192.168.18.140:53977 192.168.18.41:80 192.168.18.33:80
TCP 00:58 SYN_RECV 192.168.18.140:53998 192.168.18.41:80 192.168.18.33:80
TCP 01:57 FIN_WAIT 192.168.18.140:53981 192.168.18.41:80 192.168.18.32:80
TCP 01:57 FIN_WAIT 192.168.18.140:53974 192.168.18.41:80 192.168.18.32:80
TCP 01:57 FIN_WAIT 192.168.18.140:53976 192.168.18.41:80 192.168.18.32:80