HA（高可用）Cluster实现-创新互联

HA Cluster的目的，为了防止重要的服务器在提供服务时，出现不可抗力的因素，例如硬件故障

创新互联是一家以网络技术公司，为中小企业提供网站维护、成都网站设计、做网站、网站备案、服务器租用、域名注册、软件开发、小程序开发等企业互联网相关业务，是一家有着丰富的互联网运营推广经验的科技公司，有着多年的网站建站经验，致力于帮助中小企业在互联网让打出自已的品牌和口碑，让企业在互联网上打开一个面向全国乃至全球的业务窗口：建站欢迎联系：18980820575

自然灾害，断电，软件bug，操作系统bug等，导致提供服务的主机出现宕机，死机，从而影响正

常业务，因此采用高可用的方案，实现持续性稳定的提供主机服务的方案称为HA

A=可用

MTBF=平均无故障时间

MTTR=平均修复时长

换算公式

A=MTBF/(MTBF+MTTR)

注意提升A值的方法：

1.分子足够大，但是不划算，成本太高

2.降低分母，即降低平均修复时长，成本较低效果出色

如何降低平均修复时长？

建立备用服务器，实现Failover功能。

原理：在主服务器发生故障时，快速转移IP地址（floating ip），以及快速转移应用程序至备用服务器

需要相关软件转移IP（转移IP即实现ip地址从新配置），转移服务（即在备用主机上启动相关服务应用程序）

总体来说HA 即为在主服务器宕机时，实现备用服务器的快速切换

关键点：IP地址转移，数据共享

HA Cluster （ip，nginx）ip与nginx进程通常称为HA资源

备用服务器使用“心跳”检测，向主服务器发起udp报文（UDP报文不需要三次握手等),

根据主服务器的响应来判断主服务器是否正常工作

关键点：响应时间，资源争用共享存储

假如链接主从服务器的网线连接中断，则会导致，ip争夺，而且最主要的是存储的混乱（同一文件

一边在增加，另一边再删除）

会导致源数据的损坏，损害很大，解决方法，使用爆头设备（STONITH），在补刀，将未彻底断电的还在运行设备断电

当处理完主服务器的故障后，还要将服务器上线，即为Failback

Failover<------->Failback

HA Cluster实现方案

1、vrrp协议的实现

keepalived

2、ais（available Interface standard）：可用接口标准，完备HA集群

RHCS（cmam）

heartbeat

corosync

Keepalived：

vrrp协议：Virtual Redundant Protocol

术语:

虚拟路由：virtual router

虚拟路由器标识：VRID（0-255）

物理路由：

master：主设备

back：备用设备

priority：优先级

VIP：virtual Ip

VMAC:Virtual MAC (00-00-5e-00-01-VRID)

GraciousARP(免费arp)

通告：心跳，优先级等；周期性；

抢占式，非抢占式；

安全工作：

  认证：

无认证

简单字符认证

MD5

工作模式

主/备：单虚拟路由器；

主/主：主/备（虚拟路由器1），备/主（虚拟路径器2）

特点：

vrrp协议的软件实现，原生设计的目的为了高可用ipvs服务：

vrrp协议完成地址流动；

为vip地址所在的节点生成ipvs规则（在配置文件中预先定义）；

为ipvs集群的各RS做健康状态检测；

基于脚本调用接口通过执行脚本完成脚本中定义的功能，进而影响集群事务；

组件：

核心组件：

vrrp stack

ipvs wrapper

checkers

控制组件：配置文件分析器

IO复用器

内存管理组件

HA Cluster的配置前提：

（1）各节点时间必须同步

（2）确保iptables及selinux不会成为阻碍；

（3）各节点之间可通过主机名互相通信(对KA并非必须)；

建议使用/etc/hosts文件实现；

（4）各节点之间的root用户可以基于密钥认证的ssh服务完成互相通信（并非必须）

keepalived安装配置：

CentOS 6.4+随base仓库提供；

1、同步时间

配置chronyd服务器172.18。200.100

yum安装chrony，并启动服务

[root@localhost ~]# service chronyd start

Starting chronyd:                      [  OK  ]

使用ntpdate命令，同步172.18.10.10以及172.18.10.11的时间

[root@localhost ~]# ntpdate 172.18.200.100

2、清空iptables和selinux

iptables -F

setenforce 0

3、配置hosts文件（非必须）

4、安装keepalived

[root@localhost ~]# yum install keepalived

[root@localhost ~]# cd /etc/keepalived/

[root@localhost keepalived]# ls

keepalived.conf

[root@localhost keepalived]# cp keepalived.conf keepalived.conf.bak

[root@localhost keepalived]# ls

keepalived.conf  keepalived.conf.bak

[root@localhost keepalived]# vim keepalived.conf

主配置文件：/etc/keepalived/keepalived.conf

配置文件组成部分及相关选项解释

TOP HIERACHY

GLOBAL CONFIGURATION

Global definitions

Static routes/addresses

VRRPD CONFIGURATION

VRRP synchronization group(s)：vrrp同步组；

VRRP instance(s)：每个vrrp instance即一个vrrp路由器；

LVS CONFIGURATION

Virtual server group(s)

Virtual server(s)：ipvs集群的vs和rs；

global_defs {   ###全局定义

  notification_email {

   acassen@firewall.loc

   failover@firewall.loc  ####定义出现问题后发送邮箱的地址

   sysadmin@firewall.loc

  }

  notification_email_from Alexandre.Cassen@firewall.loc  ##从哪里发过来

  smtp_server 192.168.200.1  ###邮件服务器地址

  smtp_connect_timeout 30#####超时时间

  router_id LVS_DEVEL###路由器IP

  vrrp_mcast_group4  224.0.100.5###ipv4多播地址

}

vrrp_instance VI_1 {  ##vrrp配置段

  state MASTER###表示是主还是从这里显示主，另一个则为从

  interface eth0###表明工作从哪个网卡发出 “多波心跳信息”

  virtual_router_id 51###虚拟路由ID

  priority 100###主的优先级

  advert_int 1    ##通告时间间隔

  authentication###认证

    auth_type PASS####认证类型：简单密钥认证

    auth_pass 1111#####认证密码：最多不能超过8位

  }

  virtual_ipaddress {##虚拟IP地址配在哪个网卡上

    192.168.200.16/24 dev eth0  ##定义配置在哪个网卡的别名上

    192.168.200.17

    192.168.200.18

  }

}

track_interface {  ##配置要监控的网络接口，一旦接口出现故障，则转为FAULT状态；即接口跟踪

eth0

eth2

...

}

nopreempt：定义工作模式为非抢占模式；

preempt_delay 300：抢占式模式下，节点上线后触发新选举操作的延迟时长；

5、修改配置文件

[root@localhost keepalived]# vim keepalived.conf

global_defs {

  notification_email {

    root@localhost

  }

  notification_email_from keepalived@localhost

  smtp_server 127.0.0.1

  smtp_connect_timeout 30

  router_id node1

  vrrp_mcast_group4  224.0.100.50

}

vrrp_instance myroute {

  state MASTER

  interface eth2

  virtual_router_id 50

  priority 100

  advert_int 1

  authentication {

    auth_type PASS

    auth_pass 123456

  }

  virtual_ipaddress {

    172.18.50.50/16 dev eth2

  }

}

6、将配置文件发送到另一台机器10上

[root@localhost keepalived]# scp keepalived.conf 172.18.10.10:/etc/keepalived/

修改配置文件

[root@localhost keepalived]# vim keepalived.conf

global_defs {

  notification_email {

    root@localhost

  }

  notification_email_from keepalived@localhost

  smtp_server 127.0.0.1

  smtp_connect_timeout 30

  router_id node2

  vrrp_mcast_group4  224.0.100.50

}

vrrp_instance myroute {

  state BACKUP

  interface eth2

  virtual_router_id 50

  priority 98

  advert_int 1

  authentication {

    auth_type PASS

    auth_pass 123456

  }

  virtual_ipaddress {

    172.18.50.50/16 dev eth2

  }

}

7、启动服务

启动备用服务器11

[root@localhost ~]# service keepalived start

查看地址

[root@localhost ~]# ip a

1: lo: mtu 65536 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: eth2: mtu 1500 qdisc pfifo_fast state UP qlen 1000

  link/ether 00:0c:29:07:27:ff brd ff:ff:ff:ff:ff:ff

  inet 172.18.10.10/16 brd 172.18.255.255 scope global eth2

  inet 172.18.50.50/16 scope global secondary eth2

  inet6 fe80::20c:29ff:fe07:27ff/64 scope link

    valid_lft forever preferred_lft forever

发现地址已经添加，这是若开启主服务器，由于没有设置抢断延迟，则会立刻抢断

8、启动主服务器

[root@localhost keepalived]# service keepalived start

Starting keepalived:                    [  OK  ]

[root@localhost keepalived]# ip a

1: lo: mtu 65536 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: eth2: mtu 1500 qdisc pfifo_fast state UP qlen 1000

  link/ether 00:0c:29:99:76:84 brd ff:ff:ff:ff:ff:ff

  inet 172.18.10.11/16 brd 172.18.255.255 scope global eth2

  inet 172.18.50.50/16 scope global secondary eth2

  inet6 fe80::20c:29ff:fe99:7684/64 scope link

    valid_lft forever preferred_lft forever

发现地址已经添加

而从服务器11上

[root@localhost ~]# ip a

1: lo: mtu 65536 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: eth2: mtu 1500 qdisc pfifo_fast state UP qlen 1000

  link/ether 00:0c:29:07:27:ff brd ff:ff:ff:ff:ff:ff

  inet 172.18.10.10/16 brd 172.18.255.255 scope global eth2

  inet6 fe80::20c:29ff:fe07:27ff/64 scope link

    valid_lft forever preferred_lft forever

IP地址已经删除

9、使用tcpdump抓包工具查看主从服务器的相应心跳测试

[root@localhost keepalived]# tcpdump -i eth2 host 224.0.100.50  ###在主服务器端抓包

tcpdump: verbose output suppressed, use -v or -vv for full protocol decode

listening on eth2, link-type EN10MB (Ethernet), capture size 65535 bytes

16:39:33.357307 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

16:39:34.358905 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

16:39:35.360605 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

16:39:36.362301 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

16:39:37.363904 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

16:39:38.365658 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

16:39:39.367266 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

16:39:40.368921 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

16:39:41.370599 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

[root@localhost ~]#  tcpdump -i  eth2  -nn host 224.0.100.50  ###在从服务器端抓包

tcpdump: verbose output suppressed, use -v or -vv for full protocol decode

listening on eth2, link-type EN10MB (Ethernet), capture size 65535 bytes

16:39:40.367044 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

16:39:41.368741 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

16:39:42.370289 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

16:39:43.371983 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

16:39:44.373750 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

16:39:45.375413 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

16:39:46.377092 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

16:39:47.378760 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

分析说明。实现简单的vrrp

即从服务器每隔一秒向主服务器发送1个通报报文。探测主服务器是否存活，实现具体实施软件keepalived

###############################################################################################################################

双主模型

1、我们在172.18.10.11上配置了主服务器配置，双主服务可在配置文件下面继续添如下内容，配置如下

[root@localhost keepalived]# vim keepalived.conf

vrrp_instance myroute2 {

  state BACKUP

  interface eth2

  virtual_router_id 51

  priority 98

  advert_int 1

  authentication {

    auth_type PASS

    auth_pass 123457

  }

  virtual_ipaddress {

    172.18.51.51/16 dev eth2

  }

}

2、将内容服务至粘贴至172.18.10.10服务器的keepalived.conf配置文件中,然后需要在state和priority上进行相应修改

vrrp_instance myroute2 {

  state MASTER

  interface eth2

  virtual_router_id 51

  priority 100

  advert_int 1

  authentication {

    auth_type PASS

    auth_pass 123457

  }

  virtual_ipaddress {

    172.18.51.51/16 dev eth2

  }

}

保存并退出，实现双主模型的设置

3、从启服务并测试

service keepalived restart

Stopping keepalived:                    [  OK  ]

Starting keepalived:                    [  OK  ]

使用tcpdump抓包，结果如下

172.18.10.11端

[root@localhost keepalived]#  tcpdump -i  eth2  -nn host 224.0.100.50

tcpdump: verbose output suppressed, use -v or -vv for full protocol decode

listening on eth2, link-type EN10MB (Ethernet), capture size 65535 bytes

00:50:20.150330 IP 172.18.10.10 > 224.0.100.50: VRRPv2, Advertisement, vrid 51, prio 100, authtype simple, intvl 1s, length 20

00:50:20.521639 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

00:50:21.151175 IP 172.18.10.10 > 224.0.100.50: VRRPv2, Advertisement, vrid 51, prio 100, authtype simple, intvl 1s, length 20

00:50:21.522539 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

00:50:22.152517 IP 172.18.10.10 > 224.0.100.50: VRRPv2, Advertisement, vrid 51, prio 100, authtype simple, intvl 1s, length 20

00:50:22.523232 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

00:50:23.154334 IP 172.18.10.10 > 224.0.100.50: VRRPv2, Advertisement, vrid 51, prio 100, authtype simple, intvl 1s, length 20

00:50:23.524046 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

172.18.10.10端

[root@localhost keepalived]# tcpdump -i eth2 host 224.0.100.50

tcpdump: verbose output suppressed, use -v or -vv for full protocol decode

listening on eth2, link-type EN10MB (Ethernet), capture size 65535 bytes

00:54:01.436075 IP 172.18.10.10 > 224.0.100.50: VRRPv2, Advertisement, vrid 51, prio 100, authtype simple, intvl 1s, length 20

00:54:01.437266 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

00:54:02.437295 IP 172.18.10.10 > 224.0.100.50: VRRPv2, Advertisement, vrid 51, prio 100, authtype simple, intvl 1s, length 20

00:54:02.438831 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

00:54:03.438695 IP 172.18.10.10 > 224.0.100.50: VRRPv2, Advertisement, vrid 51, prio 100, authtype simple, intvl 1s, length 20

00:54:03.439205 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

分析每次都会收到两次信息，一次发送，一次接收

使用iptable设置规则，拒绝172.18.10.11向224.0.100.50发送通知报文

[root@localhost keepalived]# iptables -A OUTPUT -s 172.18.10.11 -d 224.0.100.50 -j REJECT

在172.18.10.10端使用tcpdump抓包

[root@localhost keepalived]#  tcpdump -i  eth2  -nn host 224.0.100.50

tcpdump: verbose output suppressed, use -v or -vv for full protocol decode

listening on eth2, link-type EN10MB (Ethernet), capture size 65535 bytes

00:50:20.150330 IP 172.18.10.10 > 224.0.100.50: VRRPv2, Advertisement, vrid 51, prio 100, authtype simple, intvl 1s, length 20

00:50:20.521639 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

00:50:21.151175 IP 172.18.10.10 > 224.0.100.50: VRRPv2, Advertisement, vrid 51, prio 100, authtype simple, intvl 1s, length 20

00:50:21.522539 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

00:50:22.152517 IP 172.18.10.10 > 224.0.100.50: VRRPv2, Advertisement, vrid 51, prio 100, authtype simple, intvl 1s, length 20

00:50:22.523232 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

分析发现发送两条通告，因为172.18.10.11不通告，便认为172.18.10.11挂掉了，因此抢断，让自己变为主机。即别人不通告则认为对方挂掉了

可以使用ip a l 查看相应的ip地址获取：

[root@localhost keepalived]# ip a l

1: lo: mtu 65536 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: eth2: mtu 1500 qdisc pfifo_fast state UP qlen 1000

  link/ether 00:0c:29:07:27:ff brd ff:ff:ff:ff:ff:ff

  inet 172.18.10.10/16 brd 172.18.255.255 scope global eth2

  inet 172.18.51.51/16 scope global secondary eth2

  inet 172.18.50.50/16 scope global secondary eth2

  inet6 fe80::20c:29ff:fe07:27ff/64 scope link

    valid_lft forever preferred_lft forever

再次在172.18.10.11服务器上，清空iptables规则

[root@localhost keepalived]# iptables -F

再回到172.18.10.10服务器上使用ip a l 查询

[root@localhost keepalived]# ip a l

1: lo: mtu 65536 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: eth2: mtu 1500 qdisc pfifo_fast state UP qlen 1000

  link/ether 00:0c:29:07:27:ff brd ff:ff:ff:ff:ff:ff

  inet 172.18.10.10/16 brd 172.18.255.255 scope global eth2

  inet 172.18.51.51/16 scope global secondary eth2

  inet6 fe80::20c:29ff:fe07:27ff/64 scope link

    valid_lft forever preferred_lft forever

发现地址已经立马被夺回，是因为工作在抢占模式下。没有设置preempt_delay 300抢占延迟时间，

结论：实现双主模型实验

##################################################################################################################

如何实现自定义通知脚本

一、在172.18.10.11服务器上添加脚本，实现自动发邮件

1.编写邮件脚本

vim notify.sh

#!/bin/bash

#

contact='root@localhost'

notify() {

    mailsubject="vrrp： $(hostname) to be $1"

    mailbody="$(hostname) to be $1,vrrp transition, $(date)."

    echo "$mailbody" | mail -s "$mailsubject" $contact

}

case $1 in

master)

    notify master ;;

backup)

    notify backup ;;

fault)

    notify fault ;;

*)

    echo "Usage: $(basename $0 ) master|backup|fault"

    exit 1

    ;;

esac

2、测试脚本

语法检测

[root@localhost keepalived]# bash -n notify.sh

运行脚本测试

[root@localhost keepalived]# bash -x notify.sh master

+ contact=root@localhost

+ case $1 in

+ notify master

++ hostname

+ mailsubject='localhost.localdomain to be master'

++ hostname

++ date

+ mailbody='localhost.localdomain to be master,vrrp transition, Mon May 15 01:36:33 CST 2017.'

+ echo 'localhost.localdomain to be master,vrrp transition, Mon May 15 01:36:33 CST 2017.'

+ mail -s mailsubject root@localhost

[root@localhost keepalived]# vim notify.sh

You have mail in /var/spool/mail/root

3、查看收到的邮件

[root@localhost keepalived]# mail

Heirloom Mail version 12.4 7/29/08.  Type ? for help.

"/var/spool/mail/root": 1 message 1 new

>N  1 root          Mon May 15 01:36  18/696  "mailsubject"

&

Message  1:

From root@localhost.localdomain  Mon May 15 01:36:34 2017

Return-Path:

X-Original-To: root@localhost

Delivered-To: root@localhost.localdomain

Date: Mon, 15 May 2017 01:36:33 +0800

To: root@localhost.localdomain

Subject: mailsubject

User-Agent: Heirloom mailx 12.4 7/29/08

Content-Type: text/plain; charset=us-ascii

From: root@localhost.localdomain (root)

Status: R

localhost.localdomain to be master,vrrp transition, Mon May 15 01:36:33 CST 2017.

&

4、将脚本发送至172.18.10.10端

[root@localhost keepalived]# scp notify.sh 172.18.10.10:/etc/keepalived/

root@172.18.10.10's password:

notify.sh                                    100%  367   0.4KB/s  00:00

5、调用脚本

[root@localhost keepalived]# vim keepalived.conf

在172.18.10.11上的vrrp_instance myrouter1下面添加如下内容,注意是放在vrrp_instance myrouter1上下文中调用

    notify_master "/etc/keepalived/notify.sh master"

    notify_backup "/etc/keepalived/notify.sh backup"

    notify_fault "/etc/keepalived/notify.sh fault"

在172.18.10.10上的vrrp_instance myrouter2下面添加如下内容

    notify_master "/etc/keepalived/notify.sh master"

    notify_backup "/etc/keepalived/notify.sh backup"

    notify_fault "/etc/keepalived/notify.sh fault"

6，为了实现测试效果，将之前定义的双主模型删除，并停止服务（在10.10和10.11上做同样的操作）

：.,$d  表示从当前行都最后一行全部删除

[root@localhost keepalived]# service keepalived stop

Stopping keepalived:                    [  OK  ]

7、给之前编写的脚本加上执行权限

[root@localhost keepalived]# chmod +x  notify.sh

[root@localhost keepalived]# ll

total 8

-rw-r--r-- 1 root root 658 May 15 02:01 keepalived.conf

-rwxr-xr-x 1 root root 367 May 15 01:41 notify.sh

8、启动服务

在172.18.10.11端

[root@localhost keepalived]# service keepalived start

Starting keepalived:                    [  OK  ]

[root@localhost keepalived]# ip a l

1: lo: mtu 65536 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: eth2: mtu 1500 qdisc pfifo_fast state UP qlen 1000

  link/ether 00:0c:29:99:76:84 brd ff:ff:ff:ff:ff:ff

  inet 172.18.10.11/16 brd 172.18.255.255 scope global eth2

  inet 172.18.50.50/16 scope global secondary eth2

  inet6 fe80::20c:29ff:fe99:7684/64 scope link

    valid_lft forever preferred_lft forever

[root@localhost keepalived]# mail

Heirloom Mail version 12.4 7/29/08.  Type ? for help.

"/var/spool/mail/root": 3 messages 2 unread

  1 root          Mon May 15 01:36  19/707  "mailsubject"

>U  2 root          Mon May 15 11:03  19/735  "vrrp: localhost.localdomain to be master"

&

9、启动172.18.10.10端的keepalived，并且再次到172.18.10.11端查看邮件

[root@localhost ~]# mail

Heirloom Mail version 12.4 7/29/08.  Type ? for help.

"/var/spool/mail/root": 7 messages 5 new 7 unread

 U  1 root          Mon May 15 11:09  19/735  "vrrp: localhost.localdomain to be backup"

 U  2 root          Mon May 15 11:11  19/735  "vrrp: localhost.localdomain to be backup"

>N  3 root          Mon May 15 11:11  18/725  "vrrp: localhost.localdomain to be master"

 N  4 root          Mon May 15 11:11  18/725  "vrrp: localhost.localdomain to be backup"

 N  5 root          Mon May 15 11:26  18/725  "vrrp: localhost.localdomain to be backup"

 N  6 root          Mon May 15 11:26  18/725  "vrrp: localhost.localdomain to be master"

 N  7 root          Mon May 15 11:26  18/725  "vrrp: localhost.localdomain to be backup"

&

结论：通知脚本功能实现

######################################################################################################

如何实现 keepalived 高可用LVS （重点）

实验准备：4台虚拟主机

其中172.18.10.10和172.18.10.11做为VS端分别为VS2和VS1

172.18.200.100和172.18.249.57做为RS分别为RS1和RS2

首先分别再RS1和RS2端安装httpd

1、进行如下配置

[root@localhost ~]# cat /var/www/html/index.html

RS1:172.18.200.100

[root@localhost ~]# cat /var/www/html/index.html

RS2:172.18.249.57

2、编写VIP配置脚本

vim setparam.sh

#!/bin/bash

#

vip='172.18.50.50'

netmask='255.255.255.255'

iface='lo:0'

case $1 in

start)

    echo 1 > /pro/sys/net/ipv4/conf/all/arp_ignore

    echo 1 > /pro/sys/net/ipv4/conf/lo/arp_ignore

    echo 2 > /pro/sys/net/ipv4/conf/all/arp_ignore

    echo 2 > /pro/sys/net/ipv4/conf/lo/arp_ignore

    ifconfig $iface $vip netmask $netmask broadcast $vip up

    route add -host $vip dev $iface

    ;;

stop)

    ifconfig $iface down

    echo 0 > /pro/sys/net/ipv4/conf/all/arp_ignore

    echo 0 > /pro/sys/net/ipv4/conf/lo/arp_ignore

    echo 0 > /pro/sys/net/ipv4/conf/all/arp_ignore

    echo 0 > /pro/sys/net/ipv4/conf/lo/arp_ignore

    ;;

esac

3、测试脚本

[root@localhost ~]# bash -n setparam.sh

[root@localhost ~]# bash -x setparam.sh start

+ vip=172.18.50.50

+ netmask=255.255.255.255

+ iface=lo:0

+ case $1 in

+ echo 1

setparam.sh: line 9: /pro/sys/net/ipv4/conf/all/arp_ignore: No such file or directory

+ echo 1

setparam.sh: line 10: /pro/sys/net/ipv4/conf/lo/arp_ignore: No such file or directory

+ echo 2

setparam.sh: line 11: /pro/sys/net/ipv4/conf/all/arp_announce: No such file or directory

+ echo 2

setparam.sh: line 12: /pro/sys/net/ipv4/conf/lo/arp_announce: No such file or directory

+ ifconfig lo:0 172.18.50.50 netmask 255.255.255.255 broadcast 172.18.50.50 up

+ route add -host 172.18.50.50 dev lo:0

4、使用scp将脚本分发至RS2

[root@localhost ~]# scp setparam.sh 172.18.249.57:/root

root@172.18.249.57's password:

setparam.sh                                          100%  610   0.6KB/s  00:00

5、在RS2端执行脚本，并查看是否生成VIP

[root@localhost ~]# bash -x setparam.sh start

+ vip=172.18.50.50

+ netmask=255.255.255.255

+ iface=lo:0

+ case $1 in

+ echo 1

setparam.sh: line 9: /pro/sys/net/ipv4/conf/all/arp_ignore: No such file or directory

+ echo 1

setparam.sh: line 10: /pro/sys/net/ipv4/conf/lo/arp_ignore: No such file or directory

+ echo 2

setparam.sh: line 11: /pro/sys/net/ipv4/conf/all/arp_announce: No such file or directory

+ echo 2

setparam.sh: line 12: /pro/sys/net/ipv4/conf/lo/arp_announce: No such file or directory

+ ifconfig lo:0 172.18.50.50 netmask 255.255.255.255 broadcast 172.18.50.50 up

+ route add -host 172.18.50.50 dev lo:0

[root@localhost ~]# ip a

1: lo: mtu 65536 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 172.18.50.50/32 brd 172.18.50.50 scope global lo:0

  inet6 ::1/128 scope host

    valid_lft forever preferred_lft forever

2: eth0: mtu 1500 qdisc pfifo_fast state UP qlen 1000

  link/ether 00:0c:29:b2:ca:ea brd ff:ff:ff:ff:ff:ff

  inet 172.18.249.57/16 brd 172.18.255.255 scope global eth0

  inet6 fe80::20c:29ff:feb2:caea/64 scope link

    valid_lft forever preferred_lft forever

6、启动RS1和RS2的httpd服务，并查看端口，两端都要查看，这里只演示一端的

[root@localhost ~]# service httpd start

[root@localhost ~]# ss -tnl

State    Recv-Q Send-Q                Local Address:Port                 Peer Address:Port

LISTEN    0    128                       :::80                       :::*

LISTEN    0    128                       :::22                       :::*

LISTEN    0    128                        *:22                        *:*

LISTEN    0    100                       ::1:25                       :::*

LISTEN    0    100                    127.0.0.1:25

7、在两个前段节点生成ipvs规则

在VS2端

停止keepalived服务

配置sorry server页面

vim /var/www/html/index.html

Director2 sorry server2

启动httpd服务

[root@localhost ~]# service httpd start

在VS1端

首先停止keepalived服务

[root@localhost ~]#  service keepalived stop

vim /var/www/html/index.html

Director1

启动httpd服务

[root@localhost ~]# service httpd start

在VS1端编辑keepalived配置文件，添加如下内容：

virtual_server 172.18.50.50 80 {

  delay_loop 6

  lb_algo wrr

  lb_kind DR

  persistence_timeout 0

  protocol TCP

  real_server 172.18.10.11 80 {

    weight 1

    HTTP_GET {

  persistence_timeout 0

  protocol TCP

sorry_server 127.0.0.1 80

  real_server 172.18.10.11 80 {

    weight 1

    HTTP_GET {

      url {

       path /

        status_code 200

      }

      connect_timeout 3

      nb_get_retry 3

      delay_before_retry 3

    }

  }

  real_server 172.18.10.10 80 {

    weight 1

    HTTP_GET {

      url {

       path /

        status_code 200

      }

      connect_timeout 3

      nb_get_retry 3

      delay_before_retry 3

    }

  }

}

在VS2端，同样编辑keepalived.conf文件，添加如下内容

virtual_server 172.18.50.50 80 {

  delay_loop 6

  lb_algo wrr

  lb_kind DR

  persistence_timeout 0

  protocol TCP

sorry_server 127.0.0.1 80

  real_server 172.18.10.11 80 {

    weight 1

    HTTP_GET {

  persistence_timeout 0

  protocol TCP

  real_server 172.18.10.11 80 {

    weight 1

    HTTP_GET {

      url {

       path /

        status_code 200

      }

      connect_timeout 3

      nb_get_retry 3

      delay_before_retry 3

    }

  }

  real_server 172.18.10.10 80 {

    weight 1

    HTTP_GET {

      url {

       path /

        status_code 200

      }

      connect_timeout 3

      nb_get_retry 3

      delay_before_retry 3

    }

  }

}

在VS2上启动keepalived服务

[root@localhost ~]# service keepalived start

Starting keepalived:                    [  OK  ]

[root@localhost ~]# ipvsadm -ln

IP Virtual Server version 1.2.1 (size=4096)

Prot LocalAddress:Port Scheduler Flags

 -> RemoteAddress:Port      Forward Weight ActiveConn InActConn

TCP  172.18.50.50:80 wrr

 -> 172.18.200.100:80       Route  1    0      0

 -> 172.18.249.57:80       Route  1    0      0

在客户端使用curl进行访问测试（配置完有一定延迟，稍等片刻在访问）

[root@localhost ~]# curl http://172.18.50.50

RS2:172.18.249.57

[root@localhost ~]# curl http://172.18.50.50

RS1:172.18.200.100

[root@localhost ~]# curl http://172.18.50.50

RS2:172.18.249.57

[root@localhost ~]# curl http://172.18.50.50

RS1:172.18.200.100

[root@localhost ~]# curl http://172.18.50.50

RS2:172.18.249.57

[root@localhost ~]# curl http://172.18.50.50

RS1:172.18.200.100

[root@localhost ~]# curl http://172.18.50.50

RS2:172.18.249.57

[root@localhost ~]# curl http://172.18.50.50

RS1:172.18.200.100

在172.18.200.100端停止httpd服务

[root@localhost ~]# service httpd stop

Stopping httpd:                       [  OK  ]

在VS2端使用ipvsadm观察

[root@localhost keepalived]# ipvsadm -ln

IP Virtual Server version 1.2.1 (size=4096)

Prot LocalAddress:Port Scheduler Flags

 -> RemoteAddress:Port      Forward Weight ActiveConn InActConn

TCP  172.18.50.50:80 wrr

 -> 172.18.249.57:80       Route  1    0      2

在172.18.200.100端停止httpd服务

[root@localhost ~]# service httpd start

在VS2端使用ipvsadm观察

[root@localhost keepalived]# ipvsadm -ln

IP Virtual Server version 1.2.1 (size=4096)

Prot LocalAddress:Port Scheduler Flags

 -> RemoteAddress:Port      Forward Weight ActiveConn InActConn

TCP  172.18.50.50:80 wrr

 -> 172.18.200.100:80       Route  1    0      0

 -> 172.18.249.57:80       Route  1    0      0

启动VS1上的keepalived服务，并且关闭VS2，客户端使用curl测试发现，仍然能够访问

[root@localhost keepalived]# curl http://172.18.50.50

RS2:172.18.249.57

[root@localhost keepalived]# curl http://172.18.50.50

RS1:172.18.200.100

[root@localhost keepalived]# curl http://172.18.50.50

RS2:172.18.249.57

[root@localhost keepalived]# curl http://172.18.50.50

RS1:172.18.200.100

更改配置文件，将之前删除的双主内容添加进去

VS1端

vrrp_instance myroute2 {

  state BACKUP

  interface eth2

  virtual_router_id 51

  priority 98

  advert_int 1

  authentication {

    auth_type PASS

    auth_pass 123457

  }

  virtual_ipaddress {

172.18.51.51/16 dev eth2

  }

}

VS2端

vrrp_instance myroute2 {

  state MASTER

  interface eth2

  virtual_router_id 51

  priority 98

  advert_int 1

  authentication {

    auth_type PASS

    auth_pass 123457

  }

  virtual_ipaddress {

172.18.51.51/16 dev eth2

  }

}

重启keepalived服务,相当重要，，，，，不重启不会有效果，这就是个坑

总结VS端

VS2端keepalived配置

! Configuration File for keepalived

global_defs {

  notification_email {

root@localhost

  }

  notification_email_from keepalived@localhost

  smtp_server 127.0.0.1

  smtp_connect_timeout 30

  router_id node2

  vrrp_mcast_group4  224.0.100.50

}

vrrp_instance myroute1 {

  state BACKUP

  interface eth2

  virtual_router_id 50

  priority 98

  advert_int 1

  authentication {

    auth_type PASS

    auth_pass 123456

  }

  virtual_ipaddress {

172.18.50.50/16 dev eth2

  }

notify_master "/etc/keepalived/notify.sh master"

notify_backup "/etc/keepalived/notify.sh backup"

notify_fault "/etc/keepalived/notify.sh fault"

}

vrrp_instance myroute2 {

  state MASTER

  interface eth2

  virtual_router_id 51

  priority 98

  advert_int 1

  authentication {

    auth_type PASS

    auth_pass 123457

  }

  virtual_ipaddress {

172.18.51.51/16 dev eth2

  }

}

virtual_server 172.18.50.50 80 {

  delay_loop 6

  lb_algo wrr

  lb_kind DR

  persistence_timeout 0

  protocol TCP

  sorry_server 127.0.0.1 80

  real_server 172.18.200.100 80 {

    weight 1

    HTTP_GET {

      url {

       path /

        status_code 200

      }

      connect_timeout 3

      nb_get_retry 3

      delay_before_retry 3

    }

  }

  real_server 172.18.249.57 80 {

    weight 1

    HTTP_GET {

      url {

       path /

        status_code 200

      }

      connect_timeout 3

      nb_get_retry 3

      delay_before_retry 3

    }

  }

}

virtual_server 172.18.51.51 80 {

  delay_loop 6

  lb_algo wrr

  lb_kind DR

  persistence_timeout 0

  protocol TCP

  sorry_server 127.0.0.1 80

  real_server 172.18.200.100 80 {

    weight 1

    HTTP_GET {

      url {

       path /

        status_code 200

      }

      connect_timeout 3

      nb_get_retry 3

      delay_before_retry 3

    }

  }

  real_server 172.18.249.57 80 {

    weight 1

    HTTP_GET {

      url {

       path /

        status_code 200

      }

      connect_timeout 3

      nb_get_retry 3

      delay_before_retry 3

    }

  }

}

VS1端keepalived配置

! Configuration File for keepalived

global_defs {

  notification_email {

root@localhost

  }

  notification_email_from keepalived@localhost

  smtp_server 127.0.0.1

  smtp_connect_timeout 30

  router_id node1

  vrrp_mcast_group4  224.0.100.50

}

vrrp_instance myroute1 {

  state MASTER

  interface eth2

  virtual_router_id 50

  priority 100

  advert_int 1

  authentication {

    auth_type PASS

    auth_pass 123456

  }

  virtual_ipaddress {

172.18.50.50/16 dev eth2

  }

notify_master "/etc/keepalived/notify.sh master"

    notify_backup "/etc/keepalived/notify.sh backup"

    notify_fault "/etc/keepalived/notify.sh fault"

}

vrrp_instance myroute2 {

  state BACKUP

  interface eth2

  virtual_router_id 51

  priority 98

  advert_int 1

  authentication {

    auth_type PASS

    auth_pass 123457

  }

  virtual_ipaddress {

172.18.51.51/16 dev eth2

  }

}

virtual_server 172.18.50.50 80 {

  delay_loop 6

  lb_algo wrr

  lb_kind DR

  persistence_timeout 0

  protocol TCP

sorry_server 127.0.0.1 80

  real_server 172.18.200.100 80 {

    weight 1

    HTTP_GET {

      url {

       path /

status_code 200

      }

      connect_timeout 3

      nb_get_retry 3

      delay_before_retry 3

    }

  }

  real_server 172.18.249.57 80 {

    weight 1

    HTTP_GET {

      url {

       path /

status_code 200

      }

      connect_timeout 3

      nb_get_retry 3

      delay_before_retry 3

    }

  }

}

virtual_server 172.18.51.51 80 {

  delay_loop 6

  lb_algo wrr

  lb_kind DR

  persistence_timeout 0

  protocol TCP

sorry_server 127.0.0.1 80

  real_server 172.18.200.100 80 {

    weight 1

    HTTP_GET {

      url {

       path /

status_code 200

      }

      connect_timeout 3

      nb_get_retry 3

      delay_before_retry 3

    }

  }

  real_server 172.18.249.57 80 {

    weight 1

    HTTP_GET {

      url {

       path /

status_code 200

      }

      connect_timeout 3

      nb_get_retry 3

      delay_before_retry 3

    }

  }

}

VIP配置脚本 （由用户是双主模型因此VIP有两个）

#!/bin/bash

#

vip='172.18.50.50'

vip2='172.18.51.51'

netmask='255.255.255.255'

iface='lo:0'

iface2='lo:1'

case $1 in

start)

echo 1 > /pro/sys/net/ipv4/conf/all/arp_ignore

echo 1 > /pro/sys/net/ipv4/conf/lo/arp_ignore

echo 2 > /pro/sys/net/ipv4/conf/all/arp_announce

echo 2 > /pro/sys/net/ipv4/conf/lo/arp_announce

ifconfig $iface $vip netmask $netmask broadcast $vip up

ifconfig $iface2 $vip2 netmask $netmask broadcast $vip2 up

route add -host $vip dev $iface

;;

stop)

ifconfig $iface down

ifconfig $iface2 down

echo 0 > /pro/sys/net/ipv4/conf/all/arp_ignore

echo 0 > /pro/sys/net/ipv4/conf/lo/arp_ignore

echo 0 > /pro/sys/net/ipv4/conf/all/arp_announce

echo 0 > /pro/sys/net/ipv4/conf/lo/arp_announce

;;

esac

实验结论：实现keepalived 高可用lvs负载均衡

另外有需要云服务器可以了解下创新互联cdcxhl.cn，海内外云服务器15元起步，三天无理由+7*72小时售后在线，公司持有idc许可证，提供“云服务器、裸金属服务器、高防服务器、香港服务器、美国服务器、虚拟主机、免备案服务器”等云主机租用服务以及企业上云的综合解决方案，具有“安全稳定、简单易用、服务可用性高、性价比高”等特点与优势，专为企业上云打造定制，能够满足用户丰富、多元化的应用场景需求。

网站名称：HA（高可用）Cluster实现-创新互联
转载注明：http://cxhlcq.com/article/ehjso.html