LVS/NAT双主 + keepalived负载均衡实现
一、keepalived简介keepalived是分布式部署解决系统高可用的软件，结合lvs(LinuxVirtual Server)使用，解决单机宕机的问题。
keepalived是一个基于VRRP协议来实现IPVS的高可用的解决方案。对于LVS负载均衡来说，如果前端的调度器direct发生故障，则后端的realserver是无法接受请求并响应的。因此，保证前端direct的高可用性是非常关键的，否则后端的服务器是无法进行服务的。而我们的keepalived就可以用来解决单点故障（如LVS的前端direct故障）问题。keepalived的主要工作原理是：运行keepalived的两台服务器，其中一台为MASTER，另一台为BACKUP，正常情况下，所有的数据转换功能和ARP请求响应都是由MASTER完成的，一旦MASTER发生故障，则BACKUP会马上接管MASTER的工作，这种切换时非常迅速的。
二、测试环境
下面拿4台虚拟机进行环境测试,实验环境为centos6.6 x86_64,在lvs NAT模式环境下只有前端两台keepalived调度机有公网ip，后端真实机只有内网ip，具体用途和ip如下
服务器类型
公网ip
内网ip
LVS VIP1
192.168.214.70
192.168.211.254
LVS VIP2
192.168.214.71
192.168.211.253
Keepalived host1
192.168.214.76
192.168.211.76
Keepalived host2
192.168.214.77
192.168.211.77
Realserver A
192.168.211.79
Realserver B
192.168.211.83
拓扑图如下
http://s1.运维网.com/images/20180519/1526695242652921.jpg
三、软件安装
1、安装lvs所需包ipvsadm
yum install -y ipvsadm
ln -s /usr/src/kernels/`uname -r`/usr/src/linux
lsmod |grep ip_vs
#注意Centos 6.X安装lvs，使用1.26版本。并且需要先安装yuminstall libnl* popt* -y
执行ipvsadm(modprobe ip_vs)把ip_vs模块加载到内核
# ipvsadm -L -n
IP Virtual Server version 1.2.1 (size=4096)
Prot LocalAddress:Port Scheduler Flags
-> RemoteAddress:Port Forward Weight ActiveConn InActConn
#IP Virtual Server version 1.2.1 ---- ip_vs内核模块版本
2、安装keepalived
yum install -y keepalived
chkconfig keepalived on
注：在centos7系列系统中开机自动启动使用systemctl enable keepalived
四、开启路由转发
在lvs/NAT模式下前端两台调度机192.168.211.76和192.168.211.77上需要开启路由转发功能
vi /etc/sysctl.conf
修改net.ipv4.ip_forward = 0为net.ipv4.ip_forward= 1
保存退出后，使用systcl –p命令让其生效
五、keepalived配置
先看下211.76 keepalived配置文件
根据拓扑图得知211.76对应的是vip1 192.168.214.70的主，vip2192.168.214.71的备。
具体详细的参数说明见上篇LVS/DR + keepalived负载均衡实现
# cat /etc/keepalived/keepalived.conf
! Configuration File for keepalived
global_defs {
notification_email {
charles@test.com
}
notification_email_from reportlog@test.com
smtp_server mail.test.com
smtp_connect_timeout 30
router_id LVS_DEVEL
}
vrrp_sync_group VG1 { #lvs vip1组
group {
VI_1
VI_GATEWAY1
}
}
vrrp_sync_group VG2 { #lvs vip2组
group {
VI_2
VI_GATEWAY2
}
}
vrrp_instance VI_GATEWAY1 { #lvs vip1的配置
state MASTER
interface eth1
lvs_sync_daemon_inteface eth1
virtual_router_id 51
priority 150
advert_int 1
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
192.168.211.254
}
}
vrrp_instance VI_1 {
state MASTER
interface eth0
lvs_sync_daemon_inteface eth0
virtual_router_id 52
priority 150
advert_int 1
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
192.168.214.70
}
}
vrrp_instance VI_GATEWAY2 {#lvs vip2的配置
state BACKUP
interface eth1
lvs_sync_daemon_inteface eth1
virtual_router_id 53
priority 100
advert_int 1
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
192.168.211.253
}
}
vrrp_instance VI_2 {
state BACKUP
interface eth0
lvs_sync_daemon_inteface eth0
virtual_router_id 54
priority 100
advert_int 1
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
192.168.214.71
}
}
virtual_server 192.168.214.70 80 { #vip1
delay_loop 6
lb_algo rr
lb_kind NAT
#nat_mask 255.255.255.0
persistence_timeout 10
protocol TCP
real_server 192.168.211.79 80 { #vip1对应的后端真实机
weight 100
TCP_CHECK {
connect_timeout 3
connect_port 80
nb_get_retry 3
delay_before_retry 3
}
}
}
virtual_server 192.168.214.71 8080 { #vip2
delay_loop 6
lb_algo rr
lb_kind NAT
#nat_mask 255.255.255.0
persistence_timeout 10
protocol TCP
real_server 192.168.211.83 8080 { #vip2对应的后端真实机
weight 100
TCP_CHECK {
connect_timeout 3
connect_port 8080
nb_get_retry 3
delay_before_retry 3
}
}
} 再看下211.77 keepalived配置文件
根据拓扑图得知211.77对应的是vip1 192.168.214.70的备，vip2192.168.214.71的主。
# cat /etc/keepalived/keepalived.conf
! Configuration File for keepalived
global_defs {
notification_email {
charles@test.com
}
notification_email_from reportlog@test.com
#smtp_server 192.168.200.1
smtp_connect_timeout 30
router_id LVS_DEVEL
}
vrrp_sync_group VG1 {
group {
VI_1
VI_GATEWAY1
}
}
vrrp_sync_group VG2 {
group {
VI_2
VI_GATEWAY2
}
}
vrrp_instance VI_GATEWAY1 {
state BACKUP
interface eth1
lvs_sync_daemon_inteface eth1
virtual_router_id51
priority 100
advert_int 1
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
192.168.211.254
}
}
vrrp_instance VI_1 {
state BACKUP
interface eth0
lvs_sync_daemon_inteface eth0
virtual_router_id 52
priority 100
advert_int 1
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
192.168.214.70
}
}
vrrp_instance VI_GATEWAY2 {
state MASTER
interfaceeth1
lvs_sync_daemon_inteface eth1
virtual_router_id 53
priority 150
advert_int 1
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
192.168.211.253
}
}
vrrp_instance VI_2 {
state MASTER
interface eth0
lvs_sync_daemon_inteface eth0
virtual_router_id 54
priority 150
advert_int 1
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
192.168.214.71
}
}
virtual_server 192.168.214.70 80 {
delay_loop 6
lb_algo rr
lb_kind NAT
#nat_mask 255.255.255.0
persistence_timeout 10
protocol TCP
real_server 192.168.211.79 80 {
weight 100
TCP_CHECK {
connect_timeout 3
connect_port 80
nb_get_retry 3
delay_before_retry 3
}
}
}
virtual_server 192.168.214.71 8080 {
delay_loop 6
lb_algo rr
lb_kind NAT
#nat_mask 255.255.255.0
persistence_timeout 10
protocol TCP
real_server 192.168.211.83 8080 {
weight 100
TCP_CHECK {
connect_timeout 3
connect_port 8080
nb_get_retry 3
delay_before_retry 3
}
}
} 六、后端真实机配置
后端web应用机器192.168.211.79需要配置默认网关为vip1内网ip地址192.168.211.254，web应用机器192.168.211.83需要配置默认网关为vip2内网ip地址192.168.211.253
注意，线上环境如果更改默认网关为vip内网地址后，可能会造成连接不上服务器，需要提前添加相关静态路由。
七、启动keepalived服务及查看相关信息
在211.76和211.77上分别启动keepalived服务
在211.76主机上查看信息
通过ip addr可以看到vip1 地址已经绑定在eth0和eth1网口上
# ip addr
1: lo:mtu65536 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: eth0: mtu 1500 qdisc pfifo_fast state UP qlen1000
link/ether 52:54:00:55:b2:d4 brd ff:ff:ff:ff:ff:ff
inet 192.168.214.76/24 brd 192.168.214.255 scope global eth0
inet 192.168.214.70/32 scope global eth0
inet6 fe80::5054:ff:fe55:b2d4/64 scope link
valid_lft forever preferred_lft forever
3: eth1: mtu 1500 qdisc pfifo_fast state UP qlen1000
link/ether 52:54:00:85:11:95 brd ff:ff:ff:ff:ff:ff
inet 192.168.211.76/24 brd 192.168.211.255 scope global eth1
inet192.168.211.254/32 scope global eth1
inet6 fe80::5054:ff:fe85:1195/64 scope link
valid_lft forever preferred_lft forever
在211.76上查看日志信息，看到已成功进入keepalived vip1组的主机模式，vip2组的备机模式。
http://s1.运维网.com/images/20180519/1526695311394449.jpg
在211.77主机上查看信息
通过ip addr可以看到vip2 地址已经绑定在eth0和eth1网口上
# ip addr
1: lo:mtu65536 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: eth0: mtu 1500 qdisc pfifo_fast state UP qlen1000
link/ether 52:54:00:1b:a2:11 brd ff:ff:ff:ff:ff:ff
inet 192.168.214.77/24 brd 192.168.214.255 scope global eth0
inet 192.168.214.71/32 scope global eth0
inet6 fe80::5054:ff:fe1b:a211/64 scope link
valid_lft forever preferred_lft forever
3: eth1: mtu 1500 qdisc pfifo_fast state UP qlen1000
link/ether 52:54:00:64:47:7d brd ff:ff:ff:ff:ff:ff
inet 192.168.211.77/24 brd 192.168.211.255 scope global eth1
inet192.168.211.253/32 scope global eth1
inet6 fe80::5054:ff:fe64:477d/64 scope link
valid_lft forever preferred_lft forever
在211.77上查看日志信息，看到已成功进入keepalived vip1组的备机模式，vip2组的主机模式。
http://s1.运维网.com/images/20180519/1526695355835726.jpg
通过ipvsadm -L –n 查看相应lvs连接信息
211.76上查看
# ipvsadm -L -n
IP Virtual Server version 1.2.1 (size=4096)
Prot LocalAddress:Port Scheduler Flags
-> RemoteAddress:Port Forward Weight ActiveConn InActConn
TCP 192.168.214.70:80 rr persistent 10
-> 192.168.211.79:80 Masq 100 3 5
TCP 192.168.214.71:8080 rr persistent 10
-> 192.168.211.83:8080 Masq 100 0 0
211.77上查看
# ipvsadm -L -n
IP Virtual Server version 1.2.1 (size=4096)
Prot LocalAddress:Port Scheduler Flags
-> RemoteAddress:Port Forward Weight ActiveConn InActConn
TCP 192.168.214.70:80 rr persistent 10
-> 192.168.211.79:80 Masq 100 0 0
TCP 192.168.214.71:8080 rr persistent 10
-> 192.168.211.83:8080 Masq 100 3 4
八、keepalived测试
使用vip1地址192.168.214.70访问后端web192.168.211.79的页面
http://s1.运维网.com/images/20180519/1526695378523908.jpg
使用vip2地址192.168.214.71访问后端web192.168.211.83的页面
http://s1.运维网.com/images/20180519/1526695399660047.jpg
正常访问没问题后，我们来模拟lvs集群故障，把前端lvs调度机主机211.76宕机，看211.77能否把vip1的地址漂移过来
通过ip addr命令查看发现vip1的地址都已经漂移过来了
# ip addr
1: lo:mtu65536 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: eth0: mtu 1500 qdisc pfifo_fast state UP qlen1000
link/ether 52:54:00:1b:a2:11 brd ff:ff:ff:ff:ff:ff
inet 192.168.214.77/24 brd 192.168.214.255scope global eth0
inet 192.168.214.71/32 scope global eth0
inet 192.168.214.70/32scope global eth0
inet6 fe80::5054:ff:fe1b:a211/64 scope link
valid_lft forever preferred_lft forever
3: eth1: mtu 1500 qdisc pfifo_fast state UP qlen1000
link/ether 52:54:00:64:47:7d brd ff:ff:ff:ff:ff:ff
inet 192.168.211.77/24 brd 192.168.211.255 scope global eth1
inet 192.168.211.253/32 scope global eth1
inet192.168.211.254/32 scope global eth1
inet6 fe80::5054:ff:fe64:477d/64 scope link
valid_lft forever preferred_lft forever
再211.77上通过日志查看，看到vip1的地址已漂移了过来，变成了主机状态
May 15 14:38:26 localhostKeepalived_vrrp: VRRP_Instance(VI_1) Transitionto MASTER STATE
May 15 14:38:26 localhostKeepalived_vrrp: VRRP_Group(VG1) Syncing instances to MASTER state
May 15 14:38:26 localhostKeepalived_vrrp: VRRP_Instance(VI_GATEWAY1) Transitionto MASTER STATE
May 15 14:38:27 localhostKeepalived_vrrp: VRRP_Instance(VI_GATEWAY1)Entering MASTER STATE
May 15 14:38:27 localhostKeepalived_vrrp: VRRP_Instance(VI_GATEWAY1) setting protocol VIPs.
May 15 14:38:27 localhostKeepalived_vrrp: VRRP_Instance(VI_GATEWAY1) Sending gratuitous ARPs oneth1 for 192.168.211.254
May 15 14:38:27 localhostKeepalived_healthcheckers: Netlink reflector reports IP 192.168.211.254added
May 15 14:38:27 localhostKeepalived_vrrp: VRRP_Instance(VI_1) EnteringMASTER STATE
May 15 14:38:27 localhostKeepalived_vrrp: VRRP_Instance(VI_1) setting protocol VIPs.
May 15 14:38:27 localhostKeepalived_vrrp: VRRP_Instance(VI_1) Sending gratuitous ARPs on eth0 for192.168.214.70
May 15 14:38:27 localhostKeepalived_healthcheckers: Netlink reflector reports IP 192.168.214.70added
May 15 14:38:32 localhostKeepalived_vrrp: VRRP_Instance(VI_GATEWAY1) Sending gratuitous ARPs oneth1 for 192.168.211.254
May 15 14:38:32 localhostKeepalived_vrrp: VRRP_Instance(VI_1) Sending gratuitous ARPs on eth0 for192.168.214.70
如果想了解更多，请关注我们的公众号
公众号ID:opdevos
扫码关注
http://s1.运维网.com/images/20180519/1526695425563062.jpg
页:
[1]