首页 > 代码库 > Linux 高可用(HA)集群之keepalived详解

Linux 高可用(HA)集群之keepalived详解

http://freeloda.blog.51cto.com/2033581/1280962

大纲

一、前言

二、Keepalived 详解

三、环境准备

四、LVS+Keepalived 实现高可用的前端负载均衡器

 

一、前言

       这篇文章是前几篇文章的总结,我们先简单的总结一下我们前面讲解的内容,前面我们讲解了,LVS(负载均衡器)、Heartbeat、Corosync、Pacemaker、Web高可用集群、MySQL高可用集群、DRDB、iscsi、gfs2、cLVM等,唯一没有讲解的就是LVS可用,也就是前端高可用,我们这一篇博文主要讲解内容。在说这个之前我们得和大家讨论一个问题,也是好多博友问的问题。Heartbeat、Corosync、Keepalived这三个集群组件我们到底选哪个好,首先我想说明的是,Heartbeat、Corosync是属于同一类型,Keepalived与Heartbeat、Corosync,根本不是同一类型的。Keepalived使用的vrrp协议方式,虚拟路由冗余协议 (Virtual Router Redundancy Protocol,简称VRRP);Heartbeat或Corosync是基于主机或网络服务的高可用方式;简单的说就是,Keepalived的目的是模拟路由器的高可用,Heartbeat或Corosync的目的是实现Service的高可用。所以一般Keepalived是实现前端高可用,常用的前端高可用的组合有,就是我们常见的LVS+Keepalived、Nginx+Keepalived、HAproxy+Keepalived。而Heartbeat或Corosync是实现服务的高可用,常见的组合有Heartbeat v3(Corosync)+Pacemaker+NFS+Httpd 实现Web服务器的高可用、Heartbeat v3(Corosync)+Pacemaker+NFS+MySQL 实现MySQL服务器的高可用。总结一下,Keepalived中实现轻量级的高可用,一般用于前端高可用,且不需要共享存储,一般常用于两个节点的高可用。而Heartbeat(或Corosync)一般用于服务的高可用,且需要共享存储,一般用于多节点的高可用。这个问题我们说明白了,又有博友会问了,那heartbaet与corosync我们又应该选择哪个好啊,我想说我们一般用corosync,因为corosync的运行机制更优于heartbeat,就连从heartbeat分离出来的pacemaker都说在以后的开发当中更倾向于corosync,所以现在corosync+pacemaker是最佳组合。但说实话我对于软件没有任何倾向性,所以我把所有的集群软件都和大家说了一下,我认为不管什么软件,只要它能存活下来都有它的特点和应用领域,只有把特定的软件放在特定的位置才能发挥最大的作用,那首先我们得对这个软件有所有了解。学习一种软件的最好方法,就是去查官方文档。好了说了那么多希望大家有所收获,下面我们来说一说keepalived。

二、Keepalived 详解

1.Keepalived 定义

       Keepalived 是一个基于VRRP协议来实现的LVS服务高可用方案,可以利用其来避免单点故障。一个LVS服务会有2台服务器运行Keepalived,一台为主服务器(MASTER),一台为备份服务器(BACKUP),但是对外表现为一个虚拟IP,主服务器会发送特定的消息给备份服务器,当备份服务器收不到这个消息的时候,即主服务器宕机的时候, 备份服务器就会接管虚拟IP,继续提供服务,从而保证了高可用性。Keepalived是VRRP的完美实现,因此在介绍keepalived之前,先介绍一下VRRP的原理。

2.VRRP 协议简介

在现实的网络环境中,两台需要通信的主机大多数情况下并没有直接的物理连接。对于这样的情况,它们之间路由怎样选择?主机如何选定到达目的主机的下一跳路由,这个问题通常的解决方法有二种:

  • 在主机上使用动态路由协议(RIP、OSPF等)

  • 在主机上配置静态路由

很明显,在主机上配置动态路由是非常不切实际的,因为管理、维护成本以及是否支持等诸多问题。配置静态路由就变得十分流行,但路由器(或者说默认网关default gateway)却经常成为单点故障。VRRP的目的就是为了解决静态路由单点故障问题,VRRP通过一竞选(election)协议来动态的将路由任务交给LAN中虚拟路由器中的某台VRRP路由器。

3.VRRP 工作机制

       在一个VRRP虚拟路由器中,有多台物理的VRRP路由器,但是这多台的物理的机器并不能同时工作,而是由一台称为MASTER的负责路由工作,其它的都是BACKUP,MASTER并非一成不变,VRRP让每个VRRP路由器参与竞选,最终获胜的就是MASTER。MASTER拥有一些特权,比如,拥有虚拟路由器的IP地址,我们的主机就是用这个IP地址作为静态路由的。拥有特权的MASTER要负责转发发送给网关地址的包和响应ARP请求。

       VRRP通过竞选协议来实现虚拟路由器的功能,所有的协议报文都是通过IP多播(multicast)包(多播地址224.0.0.18)形式发送的。虚拟路由器由VRID(范围0-255)和一组IP地址组成,对外表现为一个周知的MAC地址。所以,在一个虚拟路由 器中,不管谁是MASTER,对外都是相同的MAC和IP(称之为VIP)。客户端主机并不需要因为MASTER的改变而修改自己的路由配置,对客户端来说,这种主从的切换是透明的。

       在一个虚拟路由器中,只有作为MASTER的VRRP路由器会一直发送VRRP通告信息(VRRPAdvertisement message),BACKUP不会抢占MASTER,除非它的优先级(priority)更高。当MASTER不可用时(BACKUP收不到通告信息), 多台BACKUP中优先级最高的这台会被抢占为MASTER。这种抢占是非常快速的(<1s),以保证服务的连续性。由于安全性考虑,VRRP包使用了加密协议进行加密。

4.VRRP 工作流程

(1).初始化:    
路由器启动时,如果路由器的优先级是255(最高优先级,路由器拥有路由器地址),要发送VRRP通告信息,并发送广播ARP信息通告路由器IP地址对应的MAC地址为路由虚拟MAC,设置通告信息定时器准备定时发送VRRP通告信息,转为MASTER状态;否则进入BACKUP状态,设置定时器检查定时检查是否收到MASTER的通告信息。

(2).Master

  • 设置定时通告定时器;

  • 用VRRP虚拟MAC地址响应路由器IP地址的ARP请求;

  • 转发目的MAC是VRRP虚拟MAC的数据包;

  • 如果是虚拟路由器IP的拥有者,将接受目的地址是虚拟路由器IP的数据包,否则丢弃;

  • 当收到shutdown的事件时删除定时通告定时器,发送优先权级为0的通告包,转初始化状态;

  • 如果定时通告定时器超时时,发送VRRP通告信息;

  • 收到VRRP通告信息时,如果优先权为0,发送VRRP通告信息;否则判断数据的优先级是否高于本机,或相等而且实际IP地址大于本地实际IP,设置定时通告定时器,复位主机超时定时器,转BACKUP状态;否则的话,丢弃该通告包;

(3).Backup

  • 设置主机超时定时器;

  • 不能响应针对虚拟路由器IP的ARP请求信息;

  • 丢弃所有目的MAC地址是虚拟路由器MAC地址的数据包;

  • 不接受目的是虚拟路由器IP的所有数据包;

  • 当收到shutdown的事件时删除主机超时定时器,转初始化状态;

  • 主机超时定时器超时的时候,发送VRRP通告信息,广播ARP地址信息,转MASTER状态;

  • 收到VRRP通告信息时,如果优先权为0,表示进入MASTER选举;否则判断数据的优先级是否高于本机,如果高的话承认MASTER有效,复位主机超时定时器;否则的话,丢弃该通告包;

5.ARP查询处理

       当内部主机通过ARP查询虚拟路由器IP地址对应的MAC地址时,MASTER路由器回复的MAC地址为虚拟的VRRP的MAC地址,而不是实际网卡的 MAC地址,这样在路由器切换时让内网机器觉察不到;而在路由器重新启动时,不能主动发送本机网卡的实际MAC地址。如果虚拟路由器开启的ARP代理 (proxy_arp)功能,代理的ARP回应也回应VRRP虚拟MAC地址;好了VRRP的简单讲解就到这里,我们下来讲解一下Keepalived的案例。

三、环境准备

1.操作系统

  • CentOS 6.4 X86_64

2.软件版本

  • ipvsadm.x86_64 0:1.25-10.el6                  

  • keepalived.x86_64 0:1.2.7-3.el6      

  • httpd-2.2.15-29.el6.centos.x86_64

3.实验拓扑

k00

4.时间同步

node1:

1
[root@node1 ~]# ntpdate 202.120.2.101

node2:

1
[root@node2 ~]# ntpdate 202.120.2.101

master:

1
[root@master ~]# ntpdate 202.120.2.101

slave:

1
[root@slave ~]# ntpdate 202.120.2.101

5.主机名互相解析

node1:

1
2
3
4
5
[root@node1 ~]# cat /etc/hosts 
127.0.0.1   localhost localhost.localdomain localhost4 localhost4.localdomain4  
::1         localhost localhost.localdomain localhost6 localhost6.localdomain6  
192.168.18.201    node1.test.com    node1  
192.168.18.202    node2.test.com    node2

node2:

1
2
3
4
5
[root@node2 ~]# cat /etc/hosts 
127.0.0.1   localhost localhost.localdomain localhost4 localhost4.localdomain4  
::1         localhost localhost.localdomain localhost6 localhost6.localdomain6  
192.168.18.201    node1.test.com    node1  
192.168.18.202    node2.test.com    node2

6.安装yum源

node1:

1
2
[root@node1 ~]# rpm -ivh http://download.fedoraproject.org/pub/epel/6/x86_64/epel-release-6-8.noarch.rpm
[root@node1 ~]# rpm -ivh http://elrepo.org/elrepo-release-6-5.el6.elrepo.noarch.rpm

node2:

1
2
[root@node2 ~]# rpm -ivh http://download.fedoraproject.org/pub/epel/6/x86_64/epel-release-6-8.noarch.rpm
[root@node2 ~]# rpm -ivh http://elrepo.org/elrepo-release-6-5.el6.elrepo.noarch.rpm

master:

1
2
[root@master ~]# rpm -ivh http://download.fedoraproject.org/pub/epel/6/x86_64/epel-release-6-8.noarch.rpm
[root@master ~]# rpm -ivh http://elrepo.org/elrepo-release-6-5.el6.elrepo.noarch.rpm

slave:

1
2
[root@slave ~]# rpm -ivh http://download.fedoraproject.org/pub/epel/6/x86_64/epel-release-6-8.noarch.rpm
[root@slave ~]# rpm -ivh http://elrepo.org/elrepo-release-6-5.el6.elrepo.noarch.rpm

四、LVS+Keepalived 实现高可用的前端负载均衡器

node1:

1.安装httpd

1
[root@node1 ~]# yum install -y httpd

2.配置httpd

1
2
[root@node1 ~]# vim /var/www/html/index.html
<h1>RS1.test.com</h1>

3.启动httpd

1
[root@node1 ~]# service httpd start

4.测试

k1

5.设置开机自启动

1
2
3
[root@node1 ~]# chkconfig httpd on  
[root@node1 ~]# chkconfig httpd --list   
httpd              0:关闭    1:关闭    2:启用    3:启用    4:启用    5:启用    6:关闭

6.配置node1

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
[root@node1 ~]# mkdir src  
[root@node1 ~]# cd src/   
[root@node1 src]# vim realserver.sh
#!/bin/bash  
#   
# Script to start LVS DR real server.   
# description: LVS DR real server   
#   
.  /etc/rc.d/init.d/functions
VIP=192.168.18.200 #修改你的VIP  
host=`/bin/hostname`
case "$1" in  
start)   
       # Start LVS-DR real server on this machine.   
        /sbin/ifconfig lo down   
        /sbin/ifconfig lo up   
        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
        /sbin/ifconfig lo:0 $VIP broadcast $VIP netmask 255.255.255.255 up  
        /sbin/route add -host $VIP dev lo:0
;;  
stop)
        # Stop LVS-DR real server loopback device(s).  
        /sbin/ifconfig lo:0 down   
        echo 0 > /proc/sys/net/ipv4/conf/lo/arp_ignore   
        echo 0 > /proc/sys/net/ipv4/conf/lo/arp_announce   
        echo 0 > /proc/sys/net/ipv4/conf/all/arp_ignore   
        echo 0 > /proc/sys/net/ipv4/conf/all/arp_announce
;;  
status)
        # Status of LVS-DR real server.  
        islothere=`/sbin/ifconfig lo:0 | grep $VIP`   
        isrothere=`netstat -rn | grep "lo:0" grep $VIP`   
        if [ ! "$islothere" -o ! "isrothere" ];then   
            # Either the route or the lo:0 device   
            # not found.   
            echo "LVS-DR real server Stopped."   
        else   
            echo "LVS-DR real server Running."   
        fi   
;;   
*)   
            # Invalid entry.   
            echo "$0: Usage: $0 {start|status|stop}"   
            exit 1   
;;   
esac   
[root@node1 src]# chmod +x realserver.sh  
[root@node1 src]# ll   
总用量 4   
-rwxr-xr-x 1 root root 1485 8月  22 10:18 realserver.sh
[root@node1 src]# ./realserver.sh start

7.查看配置

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
[root@node1 src]# ifconfig  
eth0      Link encap:Ethernet  HWaddr 00:0C:29:11:92:E4
          inet addr:192.168.18.201  Bcast:192.168.18.255  Mask:255.255.255.0   
          inet6 addr: fe80::20c:29ff:fe11:92e4/64 Scope:Link   
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1   
          RX packets:115061 errors:0 dropped:0 overruns:0 frame:0   
          TX packets:14979 errors:0 dropped:0 overruns:0 carrier:0   
          collisions:0 txqueuelen:1000   
          RX bytes:43448483 (41.4 MiB)  TX bytes:1224926 (1.1 MiB)
lo        Link encap:Local Loopback
          inet addr:127.0.0.1  Mask:255.0.0.0   
          inet6 addr: ::1/128 Scope:Host   
          UP LOOPBACK RUNNING  MTU:16436  Metric:1   
          RX packets:2 errors:0 dropped:0 overruns:0 frame:0   
          TX packets:2 errors:0 dropped:0 overruns:0 carrier:0   
          collisions:0 txqueuelen:0   
          RX bytes:168 (168.0 b)  TX bytes:168 (168.0 b)
lo:0      Link encap:Local Loopback
          inet addr:192.168.18.200  Mask:255.255.255.255   
          UP LOOPBACK RUNNING  MTU:16436  Metric:1
[root@node1 src]# route -n  
Kernel IP routing table   
Destination     Gateway         Genmask         Flags Metric Ref    Use Iface   
192.168.18.200  0.0.0.0         255.255.255.255 UH    0      0        0 lo   
192.168.18.0    0.0.0.0         255.255.255.0   U     0      0        0 eth0   
169.254.0.0     0.0.0.0         255.255.0.0     U     1002   0        0 eth0   
0.0.0.0         192.168.18.254  0.0.0.0         UG    0      0        0 eth0
[root@node1 src]# cat /proc/sys/net/ipv4/conf/lo/arp_ignore  
1   
[root@node1 src]# cat /proc/sys/net/ipv4/conf/lo/arp_announce   
2   
[root@node1 src]# cat /proc/sys/net/ipv4/conf/all/arp_ignore   
1   
[root@node1 src]# cat /proc/sys/net/ipv4/conf/all/arp_announce
2

好了,node1到这里基本配置完成,下面我们来配置node2。

node2:

1.安装httpd

1
[root@node2 ~]# yum install -y httpd

2.配置httpd

1
2
[root@node2 ~]# vim /var/www/html/index.html
<h1>RS2.test.com</h1>

3.启动httpd

1
[root@node2 ~]# service httpd start

4.测试

k2

5.设置开机自启动

1
2
3
[root@node2 ~]# chkconfig httpd on  
[root@node2 ~]# chkconfig httpd --list   
httpd              0:关闭    1:关闭    2:启用    3:启用    4:启用    5:启用    6:关闭

6.配置node2

 

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
[root@node2 ~]# mkdir src  
[root@node2 ~]# cd src/   
[root@node2 src]# vim realserver.sh
#!/bin/bash  
#   
# Script to start LVS DR real server.   
# description: LVS DR real server   
#   
.  /etc/rc.d/init.d/functions
VIP=192.168.18.200  
host=`/bin/hostname`
case "$1" in  
start)   
       # Start LVS-DR real server on this machine.   
        /sbin/ifconfig lo down   
        /sbin/ifconfig lo up   
        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
        /sbin/ifconfig lo:0 $VIP broadcast $VIP netmask 255.255.255.255 up  
        /sbin/route add -host $VIP dev lo:0
;;  
stop)
        # Stop LVS-DR real server loopback device(s).  
        /sbin/ifconfig lo:0 down   
        echo 0 > /proc/sys/net/ipv4/conf/lo/arp_ignore   
        echo 0 > /proc/sys/net/ipv4/conf/lo/arp_announce   
        echo 0 > /proc/sys/net/ipv4/conf/all/arp_ignore   
        echo 0 > /proc/sys/net/ipv4/conf/all/arp_announce
;;  
status)
        # Status of LVS-DR real server.  
        islothere=`/sbin/ifconfig lo:0 | grep $VIP`   
        isrothere=`netstat -rn | grep "lo:0" grep $VIP`   
        if [ ! "$islothere" -o ! "isrothere" ];then   
            # Either the route or the lo:0 device   
            # not found.   
            echo "LVS-DR real server Stopped."   
        else   
            echo "LVS-DR real server Running."   
        fi   
;;   
*)   
            # Invalid entry.   
            echo "$0: Usage: $0 {start|status|stop}"   
            exit 1   
;;   
esac
[root@node2 src]# chmod +x realserver.sh  
[root@node2 src]# ./realserver.sh start

7.查看配置

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
[root@node2 src]# ifconfig  
eth0      Link encap:Ethernet  HWaddr 00:0C:29:B8:DF:6A
          inet addr:192.168.18.202  Bcast:192.168.18.255  Mask:255.255.255.0   
          inet6 addr: fe80::20c:29ff:feb8:df6a/64 Scope:Link   
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1   
          RX packets:110545 errors:0 dropped:0 overruns:0 frame:0   
          TX packets:10461 errors:0 dropped:0 overruns:0 carrier:0   
          collisions:0 txqueuelen:1000   
          RX bytes:32853789 (31.3 MiB)  TX bytes:889109 (868.2 KiB)
lo        Link encap:Local Loopback
          inet addr:127.0.0.1  Mask:255.0.0.0   
          inet6 addr: ::1/128 Scope:Host   
          UP LOOPBACK RUNNING  MTU:16436  Metric:1   
          RX packets:6 errors:0 dropped:0 overruns:0 frame:0   
          TX packets:6 errors:0 dropped:0 overruns:0 carrier:0   
          collisions:0 txqueuelen:0   
          RX bytes:616 (616.0 b)  TX bytes:616 (616.0 b)
lo:0      Link encap:Local Loopback
          inet addr:192.168.18.200  Mask:255.255.255.255   
          UP LOOPBACK RUNNING  MTU:16436  Metric:1
[root@node2 src]# route -n  
Kernel IP routing table   
Destination     Gateway         Genmask         Flags Metric Ref    Use Iface   
192.168.18.200  0.0.0.0         255.255.255.255 UH    0      0        0 lo   
192.168.18.0    0.0.0.0         255.255.255.0   U     0      0        0 eth0   
169.254.0.0     0.0.0.0         255.255.0.0     U     1002   0        0 eth0   
0.0.0.0         192.168.18.254  0.0.0.0         UG    0      0        0 eth0
[root@node2 src]# cat /proc/sys/net/ipv4/conf/lo/arp_ignore  
1   
[root@node2 src]# cat /proc/sys/net/ipv4/conf/lo/arp_announce   
2   
[root@node2 src]# cat /proc/sys/net/ipv4/conf/all/arp_ignore   
1   
[root@node2 src]# cat /proc/sys/net/ipv4/conf/all/arp_announce   
2

好了,到这里node2也基本配置完成。下面我们来配置master与slave。

masterg与slave:

1.安装keepalived与ipvsadm

1
2
[root@master ~]# yum install -y keepalived ipvsadm
[root@slave ~]# yum install -y keepalived ipvsadm

2.修改配置文件

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
[root@master ~]# cat /etc/keepalived/keepalived.conf  
! Configuration File for keepalived
global_defs {  
   notification_email {   
15251076067@163.com #配置管理员邮箱   
   }   
   notification_email_from root #配置发件人   
   smtp_server 127.0.0.1 #配置邮件服务器   
   smtp_connect_timeout 30   
   router_id LVS_DEVEL   
}
vrrp_instance VI_1 {  
    state MASTER #配置模式   
    interface eth0   
    virtual_router_id 51   
    priority 101 #配置优先级   
    advert_int 1   
    authentication {   
        auth_type PASS   
        auth_pass 1111   
    }   
    virtual_ipaddress {   
        192.168.18.200 #配置虚拟IP地址   
    }   
}
virtual_server 192.168.18.200 80 {  
    delay_loop 6   
    lb_algo rr   
    lb_kind DR   
    nat_mask 255.255.255.0   
    #persistence_timeout 50   
    protocol TCP
    real_server 192.168.18.201 80 { #配置realaserver  
        weight 1   
        HTTP_GET { #监控配置   
            url {   
              path /   
          status_code 200   
            }   
            connect_timeout 2   
            nb_get_retry 3   
            delay_before_retry 1   
        }   
    }   
    real_server 192.168.18.202 80 {   
        weight 1   
        HTTP_GET {   
            url {   
              path /   
              status_code 200   
            }   
            connect_timeout 2   
            nb_get_retry 3   
            delay_before_retry 1   
        }   
    }   
}

3.将配置文件同步到slave

1
[root@master ~]# scp /etc/keepalived/keepalived.conf root@192.168.18.207:/etc/keepalived/

4.简单修改一下slave配置文件

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
[root@slave keepalived]# cat keepalived.conf  
! Configuration File for keepalived
global_defs {  
   notification_email {   
     15251076067@163.com   
   }   
   notification_email_from root   
   smtp_server 127.0.0.1   
   smtp_connect_timeout 30   
   router_id LVS_DEVEL   
}
vrrp_instance VI_1 {  
    state BACKUP #修改为BACKUP   
    interface eth0   
    virtual_router_id 51   
    priority 100 #修改优先级   
    advert_int 1   
    authentication {   
        auth_type PASS   
        auth_pass 1111   
    }   
    virtual_ipaddress {   
        192.168.18.200   
    }   
}
virtual_server 192.168.18.200 80 {  
    delay_loop 6   
    lb_algo rr   
    lb_kind DR   
    nat_mask 255.255.255.0   
    #persistence_timeout 50   
    protocol TCP
    real_server 192.168.18.201 80 {  
        weight 1   
        HTTP_GET {   
            url {   
              path /   
          status_code 200   
            }   
            connect_timeout 2   
            nb_get_retry 3   
            delay_before_retry 1   
        }   
    }   
    real_server 192.168.18.202 80 {   
        weight 1   
        HTTP_GET {   
            url {   
              path /   
              status_code 200   
            }   
            connect_timeout 2   
            nb_get_retry 3   
            delay_before_retry 1   
        }   
    }   
}

5.启动master与slave的keepalived服务

1
2
3
4
[root@master ~]# service keepalived start  
正在启动 keepalived:                                      [确定]
[root@slave ~]# service keepalived start  
正在启动 keepalived:                                      [确定]

6.查看一下LVS状态

1
2
3
4
5
6
7
[root@master ~]# ipvsadm -L -n  
IP Virtual Server version 1.2.1 (size=4096)   
Prot LocalAddress:Port Scheduler Flags   
  -> RemoteAddress:Port           Forward Weight ActiveConn InActConn   
TCP  192.168.18.200:80 rr   
  -> 192.168.18.201:80            Route   1      0          0       
  -> 192.168.18.202:80            Route   1      0          0

7.测试

k3

k4

8.模拟故障

(1).停止一下node1

1
2
[root@node1 src]# service httpd stop  
停止 httpd:                                               [确定]

(2).查看一下的lvs

1
2
3
4
5
6
[root@master ~]# ipvsadm -L -n  
IP Virtual Server version 1.2.1 (size=4096)   
Prot LocalAddress:Port Scheduler Flags   
  -> RemoteAddress:Port           Forward Weight ActiveConn InActConn   
TCP  192.168.18.200:80 rr   
  -> 192.168.18.202:80            Route   1      0          0

(3).测试一下

k4

(4).查看一下邮件

k5

(5).重新启动一下node1

1
2
[root@node1 src]# service httpd start  
正在启动 httpd:                                           [确定]

(6).再查看一下lvs状态

1
2
3
4
5
6
7
[root@master ~]# ipvsadm -L -n  
IP Virtual Server version 1.2.1 (size=4096)   
Prot LocalAddress:Port Scheduler Flags   
  -> RemoteAddress:Port           Forward Weight ActiveConn InActConn   
TCP  192.168.18.200:80 rr   
  -> 192.168.18.201:80            Route   1      0          0       
  -> 192.168.18.202:80            Route   1      0          0

(7).再查看一下邮件

k6

(8).关闭master上keepalived

1
2
3
4
5
6
[root@master ~]# service keepalived stop  
停止 keepalived:                                          [确定]   
[root@master ~]# ipvsadm -L -n   
IP Virtual Server version 1.2.1 (size=4096)   
Prot LocalAddress:Port Scheduler Flags   
  -> RemoteAddress:Port           Forward Weight ActiveConn InActConn

(9).查看一下slave状态

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
[root@slave ~]# ip addr show  
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: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000   
    link/ether 00:0c:29:f9:e6:26 brd ff:ff:ff:ff:ff:ff   
    inet 192.168.18.207/24 brd 192.168.18.255 scope global eth0   
    inet 192.168.18.200/32 scope global eth0   
    inet6 fe80::20c:29ff:fef9:e626/64 scope link   
       valid_lft forever preferred_lft forever   
[root@slave ~]# ipvsadm -L -n   
IP Virtual Server version 1.2.1 (size=4096)   
Prot LocalAddress:Port Scheduler Flags   
  -> RemoteAddress:Port           Forward Weight ActiveConn InActConn   
TCP  192.168.18.200:80 rr   
  -> 192.168.18.201:80            Route   1      0          0       
  -> 192.168.18.202:80            Route   1      0          0

(10).再次测试一下

k3

注,大家可以看到,经过上面的演示我们现在LVS的高可用即前端负载均衡的高可用,同时实现对后端realserver监控,也实现后端realserver宕机时会给管理员发送邮件。但还有几个问题我们还没有解决,问题如下:

  • 所有realserver都down机,怎么处理?是不是用户就没法打开,还是提供一下维护页面。

  • 怎么完成维护模式keepalived切换?

  • 如何在keepalived故障时,发送警告邮件给指定的管理员?

9.所有realserver都down机,怎么处理?

问题:在集群中如果所有real server全部宕机了,客户端访问时就会出现错误页面,这样是很不友好的,我们得提供一个维护页面来提醒用户,服务器正在维护,什么时间可以访问等,下面我们就来解决一下这个问题。解决方案有两种,一种是提供一台备用的real server当所有的服务器宕机时,提供维护页面,但这样做有点浪费服务器。另一种就是在负载均衡器上提供维护页面,这样是比较靠谱的,也比较常用。下面我们就来具体操作一下。

(1).master与slave安装上httpd

1
2
[root@master ~]# yum install -y httpd
[root@slave ~]# yum install -y httpd

(2).配置维护页面

1
2
3
4
[root@master ~]# vim /var/www/html/index.html 
Website is currently under maintenance, please come back later!
 [root@slave ~]# vim /var/www/html/index.html
Website is currently under maintenance, please come back later!

(3).启动httpd服务并测试

1
2
3
4
[root@master ~]# service httpd start 
正在启动 httpd:                                           [确定]
[root@slave ~]# service httpd start 
正在启动 httpd:                                           [确定]

k10

k11

(4).修改配置文件

master:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
[root@master ~]# cat /etc/keepalived/keepalived.conf  
! Configuration File for keepalived
global_defs { 
   notification_email {  
     15251076067@163.com  
   }  
   notification_email_from root  
   smtp_server 127.0.0.1  
   smtp_connect_timeout 30  
   router_id LVS_DEVEL  
}
vrrp_instance VI_1 { 
    state MASTER  
    interface eth0  
    virtual_router_id 51  
    priority 101  
    advert_int 1  
    authentication {  
        auth_type PASS  
        auth_pass 1111  
    }  
    virtual_ipaddress {  
        192.168.18.200  
    }  
}
virtual_server 192.168.18.200 80 { 
    delay_loop 6  
    lb_algo rr  
    lb_kind DR  
    nat_mask 255.255.255.0  
    protocol TCP
    real_server 192.168.18.201 80 { 
        weight 1  
        HTTP_GET {  
            url {  
              path /  
          status_code 200  
            }  
            connect_timeout 2  
            nb_get_retry 3  
            delay_before_retry 1  
        }  
    }  
    real_server 192.168.18.202 80 {  
        weight 1  
        HTTP_GET {  
            url {  
              path /  
              status_code 200  
            }  
            connect_timeout 2  
            nb_get_retry 3  
            delay_before_retry 1  
        }  
    }  
   sorry_server 127.0.0.1 80   #增加一行sorry_server   
}

slave:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
[root@slave ~]# cat /etc/keepalived/keepalived.conf  
! Configuration File for keepalived
global_defs { 
   notification_email {  
     15251076067@163.com  
   }  
   notification_email_from root  
   smtp_server 127.0.0.1  
   smtp_connect_timeout 30  
   router_id LVS_DEVEL  
}
vrrp_instance VI_1 { 
    state BACKUP  
    interface eth0  
    virtual_router_id 51  
    priority 100  
    advert_int 1  
    authentication {  
        auth_type PASS  
        auth_pass 1111  
    }  
    virtual_ipaddress {  
        192.168.18.200  
    }  
}
virtual_server 192.168.18.200 80 { 
    delay_loop 6  
    lb_algo rr  
    lb_kind DR  
    nat_mask 255.255.255.0  
    protocol TCP
    real_server 192.168.18.201 80 { 
        weight 1  
        HTTP_GET {  
            url {  
              path /  
          status_code 200  
            }  
            connect_timeout 2  
            nb_get_retry 3  
            delay_before_retry 1  
        }  
    }  
    real_server 192.168.18.202 80 {  
        weight 1  
        HTTP_GET {  
            url {  
              path /  
              status_code 200  
            }  
            connect_timeout 2  
            nb_get_retry 3  
            delay_before_retry 1  
        }  
    }  
    sorry_server 127.0.0.1 80 #增加一行sorry_server   
}

(5).关闭所有的real server并重新启动一下master与slave的keepalived

1
2
3
4
5
6
7
8
9
10
[root@node1 ~]# service httpd stop 
停止 httpd:                                               [确定]
[root@node2 ~]# service httpd stop 
停止 httpd:                                               [确定]
[root@master ~]# service keepalived restart 
停止 keepalived:                                          [确定]  
正在启动 keepalived:                                      [确定]
[root@slave ~]# service keepalived restart 
停止 keepalived:                                          [确定]  
正在启动 keepalived:                                      [确定]

(6).查看一下lvs

1
2
3
4
5
6
[root@master ~]# ipvsadm -L -n 
IP Virtual Server version 1.2.1 (size=4096)  
Prot LocalAddress:Port Scheduler Flags  
  -> RemoteAddress:Port           Forward Weight ActiveConn InActConn  
TCP  192.168.18.200:80 rr  
  -> 127.0.0.1:80                 Local   1      0          0

(7).测试

k12

注,sorry_server测试成功,下面我们继续。

10.怎么完成维护模式keepalived切换?

问题:我们一般进行主从切换测试时都是关闭keepalived或关闭网卡接口,有没有一种方法能实现在不关闭keepalived下或网卡接口来实现维护呢?方法肯定是有的,在keepalived新版本中,支持脚本vrrp_srcipt,具体如何使用大家可以man keepalived.conf查看。下面我们来演示一下具体怎么实现。

(1).定义脚本

1
2
3
4
5
6
7
vrrp_srcipt chk_schedown { 
   script "[ -e /etc/keepalived/down ] && exit 1 || exit 0"  
   interval 1 #监控间隔  
   weight -5 #减小优先级  
   fall 2 #监控失败次数  
   rise 1 #监控成功次数  
}

(2).执行脚本

1
2
3
track_script {
   chk_schedown #执行chk_schedown脚本
}

(3).修改配置文件

master:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
[root@master ~]# cat /etc/keepalived/keepalived.conf  
! Configuration File for keepalived
global_defs { 
   notification_email {  
     15251076067@163.com  
   }  
   notification_email_from root  
   smtp_server 127.0.0.1  
   smtp_connect_timeout 30  
   router_id LVS_DEVEL  
}
vrrp_script chk_schedown { #定义vrrp执行脚本  
   script "[ -e /etc/keepalived/down ] && exit 1 || exit 0" #查看是否有down文件,有就进入维护模式  
   interval 1 #监控间隔时间  
   weight -5 #降低优先级  
   fall 2 #失败次数  
   rise 1 #成功数次  
}
vrrp_instance VI_1 { 
    state MASTER  
    interface eth0  
    virtual_router_id 51  
    priority 101  
    advert_int 1  
    authentication {  
        auth_type PASS  
        auth_pass 1111  
    }  
    virtual_ipaddress {  
        192.168.18.200  
    }  
    track_script { #执行脚本  
        chk_schedown  
    }  
}
virtual_server 192.168.18.200 80 { 
    delay_loop 6  
    lb_algo rr  
    lb_kind DR  
    nat_mask 255.255.255.0  
    protocol TCP
    real_server 192.168.18.201 80 { 
        weight 1  
        HTTP_GET {  
            url {  
              path /  
          status_code 200  
            }  
            connect_timeout 2  
            nb_get_retry 3  
            delay_before_retry 1  
        }  
    }  
    real_server 192.168.18.202 80 {  
        weight 1  
        HTTP_GET {  
            url {  
              path /  
              status_code 200  
            }  
            connect_timeout 2  
            nb_get_retry 3  
            delay_before_retry 1  
        }  
    }  
   sorry_server 127.0.0.1 80  
}

slave:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
[root@slave ~]# cat /etc/keepalived/keepalived.conf  
! Configuration File for keepalived
global_defs { 
   notification_email {  
     15251076067@163.com  
   }  
   notification_email_from root  
   smtp_server 127.0.0.1  
   smtp_connect_timeout 30  
   router_id LVS_DEVEL  
}
vrrp_script chk_schedown { 
   script "[ -e /etc/keepalived/down ] && exit 1 || exit 0"  
   interval 1  
   weight -5  
   fall 2  
   rise 1  
}
vrrp_instance VI_1 { 
    state BACKUP  
    interface eth0  
    virtual_router_id 51  
    priority 100  
    advert_int 1  
    authentication {  
        auth_type PASS  
        auth_pass 1111  
    }  
    virtual_ipaddress {  
        192.168.18.200  
    }  
    track_script {  
    chk_schedown  
    }
}
virtual_server 192.168.18.200 80 { 
    delay_loop 6  
    lb_algo rr  
    lb_kind DR  
    nat_mask 255.255.255.0  
    protocol TCP
    real_server 192.168.18.201 80 { 
        weight 1  
        HTTP_GET {  
            url {  
              path /  
          status_code 200  
            }  
            connect_timeout 2  
            nb_get_retry 3  
            delay_before_retry 1  
        }  
    }  
    real_server 192.168.18.202 80 {  
        weight 1  
        HTTP_GET {  
            url {  
              path /  
              status_code 200  
            }  
            connect_timeout 2  
            nb_get_retry 3  
            delay_before_retry 1  
        }  
    }  
    sorry_server 127.0.0.1 80

(4).测试

master:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
[root@master keepalived]# touch down #新建一下down文件
[root@master keepalived]# ll 
总用量 4  
-rw-r--r-- 1 root root    0 8月  22 13:39 down  
-rw-r--r-- 1 root root 1317 8月  22 13:35 keepalived.conf
[root@master keepalived]# tail -f /var/log/messages #查看一下日志 
Aug 22 13:43:52 master Keepalived_vrrp[12003]: VRRP_Instance(VI_1) Entering MASTER STATE  
Aug 22 13:43:52 master Keepalived_vrrp[12003]: VRRP_Instance(VI_1) setting protocol VIPs.  
Aug 22 13:43:52 master Keepalived_vrrp[12003]: VRRP_Instance(VI_1) Sending gratuitous ARPs on eth0 for 192.168.18.200  
Aug 22 13:43:52 master Keepalived_vrrp[12003]: VRRP_Instance(VI_1) Received higher prio advert  
Aug 22 13:43:52 master Keepalived_vrrp[12003]: VRRP_Instance(VI_1) Entering BACKUP STATE  
Aug 22 13:43:52 master Keepalived_vrrp[12003]: VRRP_Instance(VI_1) removing protocol VIPs.  
Aug 22 13:43:52 master Keepalived_healthcheckers[12002]: Netlink reflector reports IP 192.168.18.200 added  
Aug 22 13:43:52 master Keepalived_healthcheckers[12002]: Netlink reflector reports IP 192.168.18.200 removed  
Aug 22 13:43:52 master Keepalived_healthcheckers[12002]: SMTP alert successfully sent.  
Aug 22 13:43:52 master Keepalived_healthcheckers[12002]: SMTP alert successfully sent.  
^C  
[root@master keepalived]# ip add show #查看VIP  
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: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000  
    link/ether 00:0c:29:4b:a1:85 brd ff:ff:ff:ff:ff:ff  
    inet 192.168.18.208/24 brd 192.168.18.255 scope global eth0  
    inet6 fe80::20c:29ff:fe4b:a185/64 scope link   
       valid_lft forever preferred_lft forever

slave:

1
2
3
4
5
6
7
8
9
10
11
12
[root@slave ~]# ip addr show #查看一下VIP已转移到slave上 
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: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000  
    link/ether 00:0c:29:f9:e6:26 brd ff:ff:ff:ff:ff:ff  
    inet 192.168.18.207/24 brd 192.168.18.255 scope global eth0  
    inet 192.168.18.200/32 scope global eth0  
    inet6 fe80::20c:29ff:fef9:e626/64 scope link   
       valid_lft forever preferred_lft forever

好了,自写监测脚本,完成维护模式切换,到这里就演示成功,下面我们来解决最后一个问题,就是keepalived主从切换的邮件通告。

11.如何在keepalived故障时(或主备切换时),发送警告邮件给指定的管理员?

(1).keepalived通知脚本进阶示例

下面的脚本可以接受选项,其中

  • -s, --service SERVICE,...:指定服务脚本名称,当状态切换时可自动启动、重启或关闭此服务;

  • -a, --address VIP: 指定相关虚拟路由器的VIP地址;

  • -m, --mode {mm|mb}:指定虚拟路由的模型,mm表示主主,mb表示主备;它们表示相对于同一种服务而方,其VIP的工作类型;

  • -n, --notify {master|backup|fault}:指定通知的类型,即vrrp角色切换的目标角色;

  • -h, --help:获取脚本的使用帮助;

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
#!/bin/bash 
# Author: freeloda   
# description: An example of notify script  
# Usage: notify.sh -m|--mode {mm|mb} -s|--service SERVICE1,... -a|--address VIP  -n|--notify {master|backup|falut} -h|--help
contact=‘1521076067@163.com‘
helpflag=0  
serviceflag=0  
modeflag=0  
addressflag=0  
notifyflag=0
Usage() { 
  echo "Usage: notify.sh [-m|--mode {mm|mb}] [-s|--service SERVICE1,...] <-a|--address VIP>  <-n|--notify {master|backup|falut}>"   
  echo "Usage: notify.sh -h|--help"  
}
ParseOptions() { 
  local I=1;  
  if [ $# -gt 0 ]; then  
    while [ $I -le $# ]; do  
      case $1 in  
      -s|--service)  
        [ $# -lt 2 ] && return 3  
         serviceflag=1  
         services=(`echo $2|awk -F"," ‘{for(i=1;i<=NF;i++) print $i}‘`)  
        shift 2 ;;  
      -h|--help)  
         helpflag=1  
        return 0  
        shift  
        ;;  
      -a|--address)  
        [ $# -lt 2 ] && return 3  
        addressflag=1  
        vip=$2  
        shift 2  
        ;;  
      -m|--mode)  
        [ $# -lt 2 ] && return 3  
        mode=$2  
        shift 2  
        ;;  
      -n|--notify)  
        [ $# -lt 2 ] && return 3  
        notifyflag=1  
        notify=$2  
        shift 2  
        ;;  
      *)  
        echo "Wrong options..."  
        Usage  
        return 7  
        ;;  
       esac  
    done  
    return 0  
  fi  
}
#workspace=$(dirname $0)
RestartService() { 
  if [ ${#@} -gt 0 ]; then  
    for in $@; do  
      if [ -x /etc/rc.d/init.d/$I ]; then  
        /etc/rc.d/init.d/$I restart  
      else  
        echo "$I is not a valid service..."  
      fi  
    done  
  fi  
}
StopService() { 
  if [ ${#@} -gt 0 ]; then  
    for in $@; do  
      if [ -x /etc/rc.d/init.d/$I ]; then  
        /etc/rc.d/init.d/$I stop  
      else  
        echo "$I is not a valid service..."  
      fi  
    done  
  fi  
}
Notify() {  
    mailsubject="`hostname` to be $1: $vip floating"  
    mailbody="`date ‘+%F %H:%M:%S‘`, vrrp transition, `hostname` changed to be $1."  
    echo $mailbody | mail -s "$mailsubject" $contact  
}
# Main Function  
ParseOptions $@  
[ $? -ne 0 ] && Usage && exit 5
[ $helpflag -eq 1 ] && Usage && exit 0
if [ $addressflag -ne 1 -o $notifyflag -ne 1 ]; then 
  Usage  
  exit 2  
fi
mode=${mode:-mb}
case $notify in 
‘master‘)  
  if [ $serviceflag -eq 1 ]; then  
      RestartService ${services[*]}  
  fi  
  Notify master  
  ;;  
‘backup‘)  
  if [ $serviceflag -eq 1 ]; then  
    if "$mode" == ‘mb‘ ]; then  
      StopService ${services[*]}  
    else  
      RestartService ${services[*]}  
    fi  
  fi  
  Notify backup  
  ;;  
‘fault‘)  
  Notify fault  
  ;;  
*)  
  Usage  
  exit 4  
  ;;  
esac

(2).在keepalived.conf配置文件中,其调用方法如下所示:

  • notify_master "/etc/keepalived/notify.sh -n master -a 192.168.18.200"  

  • notify_backup "/etc/keepalived/notify.sh -n backup -a 192.168.18.200"  

  • notify_fault "/etc/keepalived/notify.sh -n fault -a 192.168.18.200"  

(3).修改配置文件

master:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
[root@master keepalived]# cat keepalived.conf  
! Configuration File for keepalived
global_defs { 
   notification_email {  
     15251076067@163.com  
   }  
   notification_email_from root  
   smtp_server 127.0.0.1  
   smtp_connect_timeout 30  
   router_id LVS_DEVEL  
}
vrrp_script chk_schedown { 
   script "[ -e /etc/keepalived/down ] && exit 1 || exit 0"  
   interval 1  
   weight -5  
   fall 2  
   rise 1  
}
vrrp_instance VI_1 { 
    state MASTER  
    interface eth0  
    virtual_router_id 51  
    priority 101  
    advert_int 1  
    authentication {  
        auth_type PASS  
        auth_pass 1111  
    }  
    virtual_ipaddress {  
        192.168.18.200  
    }  
    track_script {  
        chk_schedown  
    }
    #增加以下三行 
    notify_master "/etc/keepalived/notify.sh -n master -a 192.168.18.200"  
    notify_backup "/etc/keepalived/notify.sh -n backup -a 192.168.18.200"
    notify_fault "/etc/keepalived/notify.sh -n fault -a 192.168.18.200"
}
virtual_server 192.168.18.200 80 { 
    delay_loop 6  
    lb_algo rr  
    lb_kind DR  
    nat_mask 255.255.255.0  
    protocol TCP
    real_server 192.168.18.201 80 { 
        weight 1  
        HTTP_GET {  
            url {  
              path /  
          status_code 200  
            }  
            connect_timeout 2  
            nb_get_retry 3  
            delay_before_retry 1  
        }  
    }  
    real_server 192.168.18.202 80 {  
        weight 1  
        HTTP_GET {  
            url {  
              path /  
              status_code 200  
            }  
            connect_timeout 2  
            nb_get_retry 3  
            delay_before_retry 1  
        }  
    }  
   sorry_server 127.0.0.1 80  
}

slave:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
[root@slave keepalived]# cat keepalived.conf  
! Configuration File for keepalived
global_defs { 
   notification_email {  
     15251076067@163.com  
   }  
   notification_email_from root  
   smtp_server 127.0.0.1  
   smtp_connect_timeout 30  
   router_id LVS_DEVEL  
}
vrrp_script chk_schedown { 
   script "[ -e /etc/keepalived/down ] && exit 1 || exit 0"  
   interval 1  
   weight -5  
   fall 2  
   rise 1  
}
vrrp_instance VI_1 { 
    state BACKUP  
    interface eth0  
    virtual_router_id 51  
    priority 100  
    advert_int 1  
    authentication {  
        auth_type PASS  
        auth_pass 1111  
    }  
    virtual_ipaddress {  
        192.168.18.200  
    }  
    track_script {  
    chk_schedown  
    }
    #增加以下三行 
    notify_master "/etc/keepalived/notify.sh -n master -a 192.168.18.200"
    notify_backup "/etc/keepalived/notify.sh -n backup -a 192.168.18.200"  
    notify_fault "/etc/keepalived/notify.sh -n fault -a 192.168.18.200"  
}
virtual_server 192.168.18.200 80 { 
    delay_loop 6  
    lb_algo rr  
    lb_kind DR  
    nat_mask 255.255.255.0  
    protocol TCP
    real_server 192.168.18.201 80 { 
        weight 1  
        HTTP_GET {  
            url {  
              path /  
          status_code 200  
            }  
            connect_timeout 2  
            nb_get_retry 3  
            delay_before_retry 1  
        }  
    }  
    real_server 192.168.18.202 80 {  
        weight 1  
        HTTP_GET {  
            url {  
              path /  
              status_code 200  
            }  
            connect_timeout 2  
            nb_get_retry 3  
            delay_before_retry 1  
        }  
    }  
    sorry_server 127.0.0.1 80  
}

(4).增加脚本

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
[root@slave keepalived]# pwd 
/etc/keepalived  
[root@slave keepalived]# vim notify.sh
[root@slave keepalived]# cat notify.sh  
#!/bin/bash  
# Author: freeloda   
# description: An example of notify script  
# Usage: notify.sh -m|--mode {mm|mb} -s|--service SERVICE1,... -a|--address VIP  -n|--notify {master|backup|falut} -h|--help
contact=‘15251076067@163.com‘ 
helpflag=0  
serviceflag=0  
modeflag=0  
addressflag=0  
notifyflag=0
Usage() { 
  echo "Usage: notify.sh [-m|--mode {mm|mb}] [-s|--service SERVICE1,...] <-a|--address VIP>  <-n|--notify {master|backup|falut}>"   
  echo "Usage: notify.sh -h|--help"  
}
ParseOptions() { 
  local I=1;  
  if [ $# -gt 0 ]; then  
    while [ $I -le $# ]; do  
      case $1 in  
      -s|--service)  
        [ $# -lt 2 ] && return 3  
         serviceflag=1  
         services=(`echo $2|awk -F"," ‘{for(i=1;i<=NF;i++) print $i}‘`)  
        shift 2 ;;  
      -h|--help)  
         helpflag=1  
        return 0  
        shift  
        ;;  
      -a|--address)  
        [ $# -lt 2 ] && return 3  
        addressflag=1  
        vip=$2  
        shift 2  
        ;;  
      -m|--mode)  
        [ $# -lt 2 ] && return 3  
        mode=$2  
        shift 2  
        ;;  
      -n|--notify)  
        [ $# -lt 2 ] && return 3  
        notifyflag=1  
        notify=$2  
        shift 2  
        ;;  
      *)  
        echo "Wrong options..."  
        Usage  
        return 7  
        ;;  
       esac  
    done  
    return 0  
  fi  
}
#workspace=$(dirname $0)
RestartService() { 
  if [ ${#@} -gt 0 ]; then  
    for in $@; do  
      if [ -x /etc/rc.d/init.d/$I ]; then  
        /etc/rc.d/init.d/$I restart  
      else  
        echo "$I is not a valid service..."  
      fi  
    done  
  fi  
}
StopService() { 
  if [ ${#@} -gt 0 ]; then  
    for in $@; do  
      if [ -x /etc/rc.d/init.d/$I ]; then  
        /etc/rc.d/init.d/$I stop  
      else  
        echo "$I is not a valid service..."  
      fi  
    done  
  fi  
}
Notify() {  
    mailsubject="`hostname` to be $1: $vip floating"  
    mailbody="`date ‘+%F %H:%M:%S‘`, vrrp transition, `hostname` changed to be $1."  
    echo $mailbody | mail -s "$mailsubject" $contact  
}
# Main Function  
ParseOptions $@  
[ $? -ne 0 ] && Usage && exit 5
[ $helpflag -eq 1 ] && Usage && exit 0
if [ $addressflag -ne 1 -o $notifyflag -ne 1 ]; then 
  Usage  
  exit 2  
fi
mode=${mode:-mb}
case $notify in 
‘master‘)  
  if [ $serviceflag -eq 1 ]; then  
      RestartService ${services[*]}  
  fi  
  Notify master  
  ;;  
‘backup‘)  
  if [ $serviceflag -eq 1 ]; then  
    if "$mode" == ‘mb‘ ]; then  
      StopService ${services[*]}  
    else  
      RestartService ${services[*]}  
    fi  
  fi  
  Notify backup  
  ;;  
‘fault‘)  
  Notify fault  
  ;;  
*)  
  Usage  
  exit 4  
  ;;  
esac

(5).给脚本增加执行权限

1
[root@slave keepalived]# chmod +x notify.sh

(6).将master上脚本复制到slave上

1
[root@slave keepalived]# scp -p notify.sh root@192.168.18.207:/etc/keepalived/

(7).测试一下脚本

1
2
3
4
5
6
7
[root@slave keepalived]# ./notify.sh -h 
Usage: notify.sh [-m|--mode {mm|mb}] [-s|--service SERVICE1,...] <-a|--address VIP>  <-n|--notify {master|backup|falut}>  
Usage: notify.sh -h|--help  
[root@slave keepalived]# ./notify.sh --help  
Usage: notify.sh [-m|--mode {mm|mb}] [-s|--service SERVICE1,...] <-a|--address VIP>  <-n|--notify {master|backup|falut}>  
Usage: notify.sh -h|--help  
[root@slave keepalived]# ./notify.sh -m mb -a 1.1.1.1 -n master

(8).查看一下邮件

k13

注,大家可以看到成功收到邮件,测试成功。在模拟故障时先重启一下keepalived服务。

(9).模拟故障

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
[root@master keepalived]# ip addr show #查看一下VIP 
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: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000  
    link/ether 00:0c:29:4b:a1:85 brd ff:ff:ff:ff:ff:ff  
    inet 192.168.18.208/24 brd 192.168.18.255 scope global eth0  
    inet 192.168.18.200/32 scope global eth0  
    inet6 fe80::20c:29ff:fe4b:a185/64 scope link   
       valid_lft forever preferred_lft forever  
[root@master keepalived]# touch down #进入维护模式  
[root@master keepalived]# ll  
总用量 8  
-rw-r--r-- 1 root root    0 8月  22 14:39 down  
-rw-r--r-- 1 root root 1543 8月  22 14:04 keepalived.conf  
-rwxr-xr-x 1 root root 2516 8月  22 14:15 notify.sh  
[root@master keepalived]# ip addr show #再次查看VIP  
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: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000  
    link/ether 00:0c:29:4b:a1:85 brd ff:ff:ff:ff:ff:ff  
    inet 192.168.18.208/24 brd 192.168.18.255 scope global eth0  
    inet6 fe80::20c:29ff:fe4b:a185/64 scope link   
       valid_lft forever preferred_lft forever
[root@slave keepalived]# ip addr show #大家可以看到VIP成功移动到slave上 
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: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000  
    link/ether 00:0c:29:f9:e6:26 brd ff:ff:ff:ff:ff:ff  
    inet 192.168.18.207/24 brd 192.168.18.255 scope global eth0  
    inet 192.168.18.200/32 scope global eth0  
    inet6 fe80::20c:29ff:fef9:e626/64 scope link   
       valid_lft forever preferred_lft forever

(10).查看一下邮件

k14

注,大家可以看到,主备切换时,会发送邮件报警,好了到这里所有演示全部完成。希望大家有所收获^_^……

 

Linux 高可用(HA)集群之keepalived详解