LVS+Keepalived高可用负载均衡集群架构

3r2e3123

实验环境：
LVS和keepalived介绍：
1.keepalived的设计目标是构建高可用的LVS负载均衡群集，可以调用ipvsadm工具来创建虚拟服务器，管理服务器池，而不仅仅用来做双机热备。
使用keepalived构建LVS群集更加简便易用，主要优势体现在：
对LVS负载调度器实现热备切换，提高可用性；
对服务器池中的节点进行健康检查，自动移除失效节点，恢复后再重新加入。

2.在基于LVS+Keepalived实现的LVS群集结构中，至少包括两台热备的负载调度器，两台以上的节点服务器，本例将以DR模式的LVS群集为基础，增加一台从负载调度器，使用Keepalived来实现主、从调度器的热备，从而构建建有负载均衡、高可用两种能力的LVS网站群集平台。

3.通过LVS+Keepalived构建的LVS集群，LVS负载均衡用户请求到后端tomcat服务器，Keepalived的作用是检测web服务器的状态，如果有一台web服务器死机，或工作出现故障，Keepalived将检测到，并将有故障的web服务器从系统中剔除，当web服务器工作正常后Keepalived自动将web服务器加入到服务器群中，这些工作全部自动完成，不需要人工干涉，需要人工做的只是修复故障的web服务器。

Keepalived简介
什么是Keepalived：keepalived可以实现服务的高可用或热备，用来防止单点故障的问题；而Keepalived的核心VRRP协议，VRRP协议主要实现了在路由器或三层交换机处的冗余；Keepalived就是使用VRRP协议来实现高可用的；下面一起来看一下Keepalived的原理图：


Keepalived启动后会有三个进程：
    父进程：内存管理，子进程管理
    子进程：VRRP子进程
    子进程：healthchecker子进程
由上图可知：
    两个子进程都被系统WatchDog看管，两个子进程各自复杂自己的事，healthchecker子进程复杂检查各自服务器的健康程度，例如HTTP，LVS等，如果healthchecker子进程检查到MASTER上服务不可用了，就会通知本机上的VRRP子进程，让它删除通告，并且去掉虚拟IP，转换为BACKUP状态；

实验架构图：

1

实验过程：

HA高可用集群构建前提：
1.proxy和proxy2节点时间必须同步；
建议使用ntp协议进行；


2、节点之间必须要通过主机名互相通信；
建议使用hosts文件；

通信中使用的名字必须与其节点为上“uname -n”命令展示出的名字保持一致；

1 2 3 4 5 6 7 8

[iyunv@proxy ~]# cat /etc/hosts                         127.0.0.1   localhost localhost.localdomain localhost4 localhost4.localdomain4 ::1         localhost localhost.localdomain localhost6 localhost6.localdomain6 172.16.0.1  server.magelinux.com server 172.16.31.52 proxy.stu31.com proxy 172.16.31.53 proxy2.stu31.com proxy2 172.16.31.50 tom1.stu31.com tom1 172.16.31.51 tom2.stu31.com tom2

3、节点之间彼此root用户能基于ssh密钥方式进行通信；

1 2 3 4 5 6

节点proxy： # ssh-keygen -t rsa -P "" # ssh-copy-id -i .ssh/id_rsa.pub proxy2 节点proxy2： # ssh-keygen -t rsa -P "" # ssh-copy-id -i .ssh/id_rsa.pub proxy

测试ssh无密钥通信：

1 2 3

[iyunv@proxy ~]# date ; ssh proxy2 date Sat Jan 17 15:29:26 CST 2015 Sat Jan 17 15:29:26 CST 2015

一.配置安装LVS+keepalived
1.在Master与Backup服务器上分别安装Ipvsadm、Keepalived软件包、这里使用的是yum安装方式

1	# yum install -y keepalived ipvsadm

注释：这里安装Ipvsadm工具主要是为了查看lvs规则使用，不安装ipvsadm工具也能成功配置规则，但不方便；

2.修改keepalived配置文件
keepalived的文件路径/etc/keepalived/keepalived.conf

修改Master节点proxy的主配置文件

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

[iyunv@proxy ~]# vim /etc/keepalived/keepalived.conf ! Configuration File for keepalived global_defs {    notification_email {      #设置报警通知邮件地址，可以设置多个     root@localhost    }    notification_email_from admin@stu31.com  #设置邮件的发送地址    smtp_server 127.0.0.1        #设置smtp server的地址,该地址必须是存在的    smtp_connect_timeout 30      #设置连接smtp server的超时时间    router_id LVS_DEVEL          #运行Keepalived服务器的标识，发邮件时显示在邮件标题中的信息 } vrrp_instance VI_1 {            #定义VRRP实例,实例名自定义     state MASTER                #指定Keepalived的角色，MASTER为主服务器，BACKUP为备用服务器     interface eth0              #指定HA监测的接口     virtual_router_id 51        #虚拟路由标识，这个标识是一个数字(1-255)，在一个VRRP实例中主备服务器ID必须一样     priority 100                #优先级，数字越大优先级越高，在一个实例中主服务器优先级要高于备服务器     advert_int 1                #设置主备之间同步检查的时间间隔单位秒     authentication {            #设置验证类型和密码         auth_type PASS          #验证类型有两种{PASS|HA}         auth_pass oracle        #设置验证密码，在一个实例中主备密码保持一样     }     virtual_ipaddress {         #定义虚拟IP地址,可以有多个，每行一个     172.16.31.188     } } virtual_server 172.16.31.188 {    #设置虚拟服务器，需要指定虚拟IP与服务端口,用空格分隔     delay_loop 6                     #设置健康状态检查时间，单位为秒     lb_algo rr                       #设置负载高度算法，rr为轮询     lb_kind DR                       #设置LVS实现负载均衡的机制，可以为{NAT|TUN|DR}三种     nat_mask 255.255.0.0             #设置掩码     persistence_timeout 50           #会话保持时间，单位为秒;这个选项对于动态网页是非常有用的，为集群系统中session共享提供了一个很好的解决方案     protocol TCP                     #指定转发协议类型可以设置{TCP|UDP}两种     real_server 172.16.31.50 80 {    #后端服务器节点,需要指定Real_server的IP与端口,用空格分隔         weight 1                     #配置服务节点的权重,数字越大，权重越高         HTTP_GET {                   #设置检测Realserver的方式为http协议             url {                        path /           status_code 200       #设定返回状态码为200表示Realserver是存活的             }             connect_timeout 3        #设置响应超时时间             nb_get_retry 3           #设置超时重试次数             delay_before_retry 3     #设置超时后重试间隔         }     }     real_server 172.16.31.51 80 {         weight 1         HTTP_GET {             url {               path /           status_code 200             }             connect_timeout 3             nb_get_retry 3             delay_before_retry 3         }     } }

3.将Master服务器proxy节点上的主配置文件拷贝到Backup服务器proxy2节点稍作修改

1 2 3 4 5

# scp /etc/keepalived/keepalived.conf proxy2:/etc/keepalived/ ######修改如下两项 [iyunv@proxy2 ~]# vim /etc/keepalived/keepalived.conf state BACKUP priority 99

4.启动两台服务器上的Keepalived服务并设置为开机自启动

1 2 3 4 5

# chkconfig --add keepalived ; ssh proxy2 "chkconfig --add keepalived" # chkconfig  keepalived on ; ssh proxy2 "chkconfig keepalived on"            # chkconfig --list keepalived; ssh proxy2 "chkconfig --list keepalived" keepalived      0:off   1:off   2:on    3:on    4:on    5:on    6:off keepalived      0:off   1:off   2:on    3:on    4:on    5:on    6:off

启动服务：

1	# service keepalived start; ssh proxy2 "service keepalived start"

5.观察各节点的服务启动后的日志：
MASTER节点的日志；

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

[iyunv@proxy keepalived]# tail -f /var/log/messages Jan 17 15:57:42 proxy Keepalived[15019]: Starting Keepalived v1.2.13 (10/15,2014) Jan 17 15:57:42 proxy Keepalived[15021]: Starting Healthcheck child process, pid=15024 Jan 17 15:57:42 proxy Keepalived[15021]: Starting VRRP child process, pid=15025 Jan 17 15:57:42 proxy Keepalived_vrrp[15025]: Netlink reflector reports IP 172.16.31.52 added Jan 17 15:57:42 proxy Keepalived_vrrp[15025]: Netlink reflector reports IP fe80::a00:27ff:fe3b:2360 added Jan 17 15:57:42 proxy Keepalived_vrrp[15025]: Registering Kernel netlink reflector Jan 17 15:57:42 proxy Keepalived_vrrp[15025]: Registering Kernel netlink command channel Jan 17 15:57:42 proxy Keepalived_vrrp[15025]: Registering gratuitous ARP shared channel Jan 17 15:57:42 proxy Keepalived_healthcheckers[15024]: Netlink reflector reports IP 172.16.31.52 added Jan 17 15:57:42 proxy Keepalived_healthcheckers[15024]: Netlink reflector reports IP fe80::a00:27ff:fe3b:2360 added Jan 17 15:57:42 proxy Keepalived_healthcheckers[15024]: Registering Kernel netlink reflector Jan 17 15:57:42 proxy Keepalived_healthcheckers[15024]: Registering Kernel netlink command channel Jan 17 15:57:42 proxy Keepalived_vrrp[15025]: Opening file '/etc/keepalived/keepalived.conf'. Jan 17 15:57:42 proxy Keepalived_vrrp[15025]: Configuration is using : 63202 Bytes Jan 17 15:57:42 proxy Keepalived_vrrp[15025]: Using LinkWatch kernel netlink reflector... Jan 17 15:57:42 proxy Keepalived_healthcheckers[15024]: Opening file '/etc/keepalived/keepalived.conf'. Jan 17 15:57:42 proxy Keepalived_healthcheckers[15024]: Configuration is using : 16989 Bytes Jan 17 15:57:42 proxy Keepalived_vrrp[15025]: VRRP sockpool: [ifindex(2), proto(112), unicast(0), fd(10,11)] Jan 17 15:57:42 proxy Keepalived_healthcheckers[15024]: Using LinkWatch kernel netlink reflector... Jan 17 15:57:42 proxy Keepalived_healthcheckers[15024]: Activating healthchecker for service [172.16.31.50]:8080 Jan 17 15:57:42 proxy Keepalived_healthcheckers[15024]: Activating healthchecker for service [172.16.31.51]:8080 Jan 17 15:57:43 proxy Keepalived_vrrp[15025]: VRRP_Instance(VI_1) Transition to MASTER STATE Jan 17 15:57:44 proxy Keepalived_vrrp[15025]: VRRP_Instance(VI_1) Entering MASTER STATE Jan 17 15:57:44 proxy Keepalived_vrrp[15025]: VRRP_Instance(VI_1) setting protocol VIPs. Jan 17 15:57:44 proxy Keepalived_healthcheckers[15024]: Netlink reflector reports IP 172.16.31.188 added Jan 17 15:57:44 proxy Keepalived_vrrp[15025]: VRRP_Instance(VI_1) Sending gratuitous ARPs on eth0 for 172.16.31.188

BACKUP节点的日志：

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

[iyunv@proxy2 keepalived]# tail -f /var/log/messages Jan 17 15:57:42 proxy2 Keepalived[15005]: Starting Keepalived v1.2.13 (10/15,2014) Jan 17 15:57:42 proxy2 Keepalived[15007]: Starting Healthcheck child process, pid=15009 Jan 17 15:57:42 proxy2 Keepalived[15007]: Starting VRRP child process, pid=15010 Jan 17 15:57:42 proxy2 Keepalived_vrrp[15010]: Netlink reflector reports IP 172.16.31.53 added Jan 17 15:57:42 proxy2 Keepalived_healthcheckers[15009]: Netlink reflector reports IP 172.16.31.53 added Jan 17 15:57:42 proxy2 Keepalived_vrrp[15010]: Netlink reflector reports IP fe80::a00:27ff:fe6e:bd28 added Jan 17 15:57:42 proxy2 Keepalived_vrrp[15010]: Registering Kernel netlink reflector Jan 17 15:57:42 proxy2 Keepalived_vrrp[15010]: Registering Kernel netlink command channel Jan 17 15:57:42 proxy2 Keepalived_vrrp[15010]: Registering gratuitous ARP shared channel Jan 17 15:57:42 proxy2 Keepalived_healthcheckers[15009]: Netlink reflector reports IP fe80::a00:27ff:fe6e:bd28 added Jan 17 15:57:42 proxy2 Keepalived_healthcheckers[15009]: Registering Kernel netlink reflector Jan 17 15:57:42 proxy2 Keepalived_healthcheckers[15009]: Registering Kernel netlink command channel Jan 17 15:57:42 proxy2 Keepalived_vrrp[15010]: Opening file '/etc/keepalived/keepalived.conf'. Jan 17 15:57:42 proxy2 Keepalived_healthcheckers[15009]: Opening file '/etc/keepalived/keepalived.conf'. Jan 17 15:57:42 proxy2 Keepalived_healthcheckers[15009]: Configuration is using : 16991 Bytes Jan 17 15:57:42 proxy2 Keepalived_vrrp[15010]: Configuration is using : 63204 Bytes Jan 17 15:57:42 proxy2 Keepalived_vrrp[15010]: Using LinkWatch kernel netlink reflector... Jan 17 15:57:42 proxy2 Keepalived_vrrp[15010]: VRRP_Instance(VI_1) Entering BACKUP STATE Jan 17 15:57:42 proxy2 Keepalived_vrrp[15010]: VRRP sockpool: [ifindex(2), proto(112), unicast(0), fd(10,11)] Jan 17 15:57:42 proxy2 Keepalived_healthcheckers[15009]: Using LinkWatch kernel netlink reflector... Jan 17 15:57:42 proxy2 Keepalived_healthcheckers[15009]: Activating healthchecker for service [172.16.31.50]:8080 Jan 17 15:57:42 proxy2 Keepalived_healthcheckers[15009]: Activating healthchecker for service [172.16.31.51]:8080

6.开启Master节点与Backup节点服务器的路由转发功能

1	#sed -i 's/net.ipv4.ip_forward = 0/net.ipv4.ip_forward = 1/g' /etc/sysctl.conf

二.安装配置后端的tomcat服务器节点
1.配置后端tomcat服务器节点；
具体配置参考我的前篇博客，都是连续性的实验；
配置完成后，访问测试如图：

tom1节点：


tom2节点：


实现了负载均衡访问；

2.为两台Realserver提供Sysv格式的脚本来自动修改内核参数与虚拟IP并运行脚本

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

#vim rs.sh .  /etc/rc.d/init.d/functions VIP=172.16.31.188 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

注释：脚本中的VIP定义的是虚拟IP地址

3.脚本两个节点都存放一份；然后执行此脚本：

1	#sh rs.sh start

4.查看设置后的状态：

1 2	# sh rs.sh status LVS-DR real server Running.

查看节点的VIP地址设置完成与否：
tom1节点；

1 2 3 4 5 6 7

[iyunv@tom1 ~]# ip addr show lo 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 172.16.31.188/32 brd 172.16.31.188 scope global lo:0     inet6 ::1/128 scope host        valid_lft forever preferred_lft forever

tom2节点：

1 2 3 4 5 6 7

[iyunv@tom2 ~]# ip addr show lo 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 172.16.31.188/32 brd 172.16.31.188 scope global lo:0     inet6 ::1/128 scope host        valid_lft forever preferred_lft forever

5.查看MASTER节点proxy的ipvs规则：

1 2 3 4 5 6 7

[iyunv@proxy 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  172.16.31.188:0 rr persistent 50   -> 172.16.31.50:8080            Route   1      0          0            -> 172.16.31.51:8080            Route   1      0          0

从规则中可以看出虚拟IP与Port及调度算法为rr；其中有两个Realserver

三.访问测试及故障模拟ces
1.访问测试：



由上可知，使用的是rr调度算法，在访问测试时可能需要多访问几次或换个浏览器来测试访问

2.模拟Master服务器出现故障，将Master主机上的Keepalived服务停止，查看Backup服务器是否接管所有服务

1 2	[iyunv@proxy keepalived]# service keepalived stop    Stopping keepalived:                                       [  OK  ]

3.观察节点切换后的系统日志：
MASTER节点的日志：

1 2 3 4 5 6 7 8

[iyunv@proxy keepalived]# tail -f /var/log/messages Jan 17 16:20:40 proxy Keepalived[15021]: Stopping Keepalived v1.2.13 (10/15,2014) Jan 17 16:20:40 proxy Keepalived_vrrp[15025]: VRRP_Instance(VI_1) sending 0 priority Jan 17 16:20:40 proxy Keepalived_vrrp[15025]: VRRP_Instance(VI_1) removing protocol VIPs. Jan 17 16:20:40 proxy Keepalived_healthcheckers[15024]: Netlink reflector reports IP 172.16.31.188 removed Jan 17 16:20:40 proxy Keepalived_healthcheckers[15024]: Removing service [172.16.31.50]:8080 from VS [172.16.31.188]:0 Jan 17 16:20:40 proxy Keepalived_healthcheckers[15024]: Removing service [172.16.31.51]:8080 from VS [172.16.31.188]:0 Jan 17 16:20:40 proxy kernel: IPVS: __ip_vs_del_service: enter

BACKUP节点的日志：

1 2 3 4 5 6 7

[iyunv@proxy2 keepalived]# tail -f /var/log/messages Jan 17 16:20:41 proxy2 Keepalived_vrrp[15010]: VRRP_Instance(VI_1) Transition to MASTER STATE Jan 17 16:20:42 proxy2 Keepalived_vrrp[15010]: VRRP_Instance(VI_1) Entering MASTER STATE Jan 17 16:20:42 proxy2 Keepalived_vrrp[15010]: VRRP_Instance(VI_1) setting protocol VIPs. Jan 17 16:20:42 proxy2 Keepalived_healthcheckers[15009]: Netlink reflector reports IP 172.16.31.188 added Jan 17 16:20:42 proxy2 Keepalived_vrrp[15010]: VRRP_Instance(VI_1) Sending gratuitous ARPs on eth0 for 172.16.31.188 Jan 17 16:20:47 proxy2 Keepalived_vrrp[15010]: VRRP_Instance(VI_1) Sending gratuitous ARPs on eth0 for 172.16.31.188

4.查看各节点的VIP地址状态及IPVS规则：
原MASTER节点：

1 2 3 4 5 6 7 8 9 10 11

[iyunv@proxy keepalived]# ip addr show 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: eth0: mtu 1500 qdisc pfifo_fast state UP qlen 1000     link/ether 08:00:27:3b:23:60 brd ff:ff:ff:ff:ff:ff     inet 172.16.31.52/16 brd 172.16.255.255 scope global eth0     inet6 fe80::a00:27ff:fe3b:2360/64 scope link        valid_lft forever preferred_lft forever

1 2 3 4

[iyunv@proxy keepalived]# ipvsadm -L -n IP Virtual Server version 1.2.1 (size=4096) Prot LocalAddress:Port Scheduler Flags   -> RemoteAddress:Port           Forward Weight ActiveConn InActConn

由上可见Master服务器上已删除虚拟IP与LVS规则

BACKUP节点：

1 2 3 4 5 6 7 8 9 10 11 12

[iyunv@proxy2 keepalived]# ip addr show 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: eth0: mtu 1500 qdisc pfifo_fast state UP qlen 1000     link/ether 08:00:27:6e:bd:28 brd ff:ff:ff:ff:ff:ff     inet 172.16.31.53/16 brd 172.16.255.255 scope global eth0     inet 172.16.31.188/32 scope global eth0     inet6 fe80::a00:27ff:fe6e:bd28/64 scope link        valid_lft forever preferred_lft forever

由上可见，虚拟IP地址已成功在Backup服务器启动

1 2 3 4 5 6 7

[iyunv@proxy2 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  172.16.31.188:0 rr persistent 50   -> 172.16.31.50:8080            Route   1      0          0            -> 172.16.31.51:8080            Route   1      0          0

LVS的规则也已成功配置在Backup服务器上面


5.再次访问测试服务器是否正常提供服务



6.假如Master服务器修复好已重新上线，则虚拟IP地址与LVS规则会重新配置到Master服务器上而在Backup服务器上删除

1 2	[iyunv@proxy keepalived]# service keepalived start Starting keepalived:                                       [  OK  ]

观察启动日志：
MASTER节点的日志：

1 2 3 4 5 6 7 8

[iyunv@proxy keepalived]# tail -f /var/log/messages Jan 17 16:34:50 proxy Keepalived_vrrp[15366]: VRRP_Instance(VI_1) Transition to MASTER STATE Jan 17 16:34:50 proxy Keepalived_vrrp[15366]: VRRP_Instance(VI_1) Received lower prio advert, forcing new election Jan 17 16:34:51 proxy Keepalived_vrrp[15366]: VRRP_Instance(VI_1) Entering MASTER STATE Jan 17 16:34:51 proxy Keepalived_vrrp[15366]: VRRP_Instance(VI_1) setting protocol VIPs. Jan 17 16:34:51 proxy Keepalived_vrrp[15366]: VRRP_Instance(VI_1) Sending gratuitous ARPs on eth0 for 172.16.31.188 Jan 17 16:34:51 proxy Keepalived_healthcheckers[15365]: Netlink reflector reports IP 172.16.31.188 added Jan 17 16:34:56 proxy Keepalived_vrrp[15366]: VRRP_Instance(VI_1) Sending gratuitous ARPs on eth0 for 172.16.31.188

自动进入了MASTER状态；

BACKUP节点的日志：

1 2 3 4 5

[iyunv@proxy2 keepalived]# tail -f /var/log/messages Jan 17 16:34:50 proxy2 Keepalived_vrrp[15258]: VRRP_Instance(VI_1) Received higher prio advert Jan 17 16:34:50 proxy2 Keepalived_vrrp[15258]: VRRP_Instance(VI_1) Entering BACKUP STATE Jan 17 16:34:50 proxy2 Keepalived_vrrp[15258]: VRRP_Instance(VI_1) removing protocol VIPs. Jan 17 16:34:50 proxy2 Keepalived_healthcheckers[15257]: Netlink reflector reports IP 172.16.31.188 removed

自动进入BACKUP状态，移除VIP；


查看各节点的VIP 和IPVS规则：
MASTER节点的的VIP生成成功，IPVS规则生成：

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

[iyunv@proxy keepalived]# ip addr show 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: eth0: mtu 1500 qdisc pfifo_fast state UP qlen 1000     link/ether 08:00:27:3b:23:60 brd ff:ff:ff:ff:ff:ff     inet 172.16.31.52/16 brd 172.16.255.255 scope global eth0     inet 172.16.31.188/32 scope global eth0     inet6 fe80::a00:27ff:fe3b:2360/64 scope link        valid_lft forever preferred_lft forever [iyunv@proxy 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  172.16.31.188:0 rr persistent 50   -> 172.16.31.50:8080            Route   1      0          0            -> 172.16.31.51:8080            Route   1      0          0

BACKUP节点自动降级，VIP自动移除，但是ipvs规则还是存在的，这对我们是没有影响的，没有了IP地址只有规则也是不生效的

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

[iyunv@proxy2 keepalived]# ip addr show 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: eth0: mtu 1500 qdisc pfifo_fast state UP qlen 1000     link/ether 08:00:27:6e:bd:28 brd ff:ff:ff:ff:ff:ff     inet 172.16.31.53/16 brd 172.16.255.255 scope global eth0     inet6 fe80::a00:27ff:fe6e:bd28/64 scope link        valid_lft forever preferred_lft forever [iyunv@proxy2 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  172.16.31.188:0 rr persistent 50   -> 172.16.31.50:8080            Route   1      0          0            -> 172.16.31.51:8080            Route   1      0          0

7.如果后端Realserver出现故障，则LVS规则会清除相应Realserver的规则
我们将后端的tom1节点的tomcat服务停止：

1 2 3 4 5 6 7

[iyunv@tom1 ~]# catalina.sh stop Using CATALINA_BASE:   /usr/local/tomcat Using CATALINA_HOME:   /usr/local/tomcat Using CATALINA_TMPDIR: /usr/local/tomcat/temp Using JRE_HOME:        /usr/java/latest Using CLASSPATH:       /usr/local/tomcat/bin/bootstrap.jar:/usr/local/tomcat/bin/tomcat-juli.jar [iyunv@tom1 ~]#

观察LVS节点的启动日志：

MASTER节点：
出现连接错误，发送了邮件：

1 2 3 4 5

[iyunv@proxy keepalived]# tail -f /var/log/messages Jan 17 16:40:39 proxy Keepalived_healthcheckers[15365]: Error connecting server [172.16.31.50]:8080. Jan 17 16:40:39 proxy Keepalived_healthcheckers[15365]: Removing service [172.16.31.50]:8080 from VS [172.16.31.188]:0 Jan 17 16:40:39 proxy Keepalived_healthcheckers[15365]: Remote SMTP server [127.0.0.1]:25 connected. Jan 17 16:40:39 proxy Keepalived_healthcheckers[15365]: SMTP alert successfully sent.

BACKUP节点：
出现连接错误，发送了邮件：

1 2 3 4 5

[iyunv@proxy2 keepalived]# tail -f /var/log/messages Jan 17 16:40:42 proxy2 Keepalived_healthcheckers[15257]: Error connecting server [172.16.31.50]:8080. Jan 17 16:40:42 proxy2 Keepalived_healthcheckers[15257]: Removing service [172.16.31.50]:8080 from VS [172.16.31.188]:0 Jan 17 16:40:42 proxy2 Keepalived_healthcheckers[15257]: Remote SMTP server [127.0.0.1]:25 connected. Jan 17 16:40:42 proxy2 Keepalived_healthcheckers[15257]: SMTP alert successfully sent.

在到MASTER节点查看ipvs规则：

1 2 3 4 5 6 7

[iyunv@proxy 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  172.16.31.188:0 rr persistent 50   -> 172.16.31.51:8080            Route   1      0          0          You have new mail in /var/spool/mail/root

规则已经将故障的后端tomcat服务器移除，并且我们还收到了邮件；查看邮件：

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

[iyunv@proxy keepalived]# mail Heirloom Mail version 12.4 7/29/08.  Type ? for help. "/var/spool/mail/root": 16 messages 12 new 14 unread 略..... N 12 admin@stu31.com       Sat Jan 17 15:44  17/594   "[LVS_DEVEL] Realserver [172.16.31.51]:80 - DOWN" N 13 admin@stu31.com       Sat Jan 17 15:44  17/594   "[LVS_DEVEL] Realserver [172.16.31.50]:80 - DOWN" N 14 admin@stu31.com       Sat Jan 17 15:50  17/594   "[LVS_DEVEL] Realserver [172.16.31.51]:80 - DOWN" N 15 admin@stu31.com       Sat Jan 17 15:50  17/594   "[LVS_DEVEL] Realserver [172.16.31.50]:80 - DOWN" N 16 admin@stu31.com       Sat Jan 17 16:40  17/596   "[LVS_DEVEL] Realserver [172.16.31.50]:8080 - DOWN" & 16 Message 16: From admin@stu31.com  Sat Jan 17 16:40:39 2015 Return-Path: <admin@stu31.com> X-Original-To: root@localhost Delivered-To: root@localhost.stu31.com Date: Sat, 17 Jan 2015 08:40:39 +0000 From: admin@stu31.com Subject: [LVS_DEVEL] Realserver [172.16.31.50]:8080 - DOWN X-Mailer: Keepalived To: root@localhost.stu31.com Status: R => CHECK failed on service : connection error <= & quit Held 16 messages in /var/spool/mail/root

注：我在这里还测试了优先级相同时，LVS的主备节点的选举情况，如果优先级都是100的话，我们停止MASTER节点的keepalived服务后，VIP和IPVS规则自动转移到了BACKUP节点，但是我们将MASTER节点的keepalived服务启动后，VIP还是在 BACKUP节点上，ipvs规则会在MASTER节点上生成；
所以我们在生产环境中一定要注意优先级的设置，主节点的优先级必须比备用节点的优先级高；

至此，LVS+Keepalived实现高可用负载均衡集群已全部完成