LVS+Keepalived构建高可用负载均衡(测试篇)

一、 启动lvs高可用集群服务

首先，启动每个real server节点的服务：
[root@localhost ~]# /etc/init.d/lvsrs start
start lvs of realserver
然后，分别在主备director server启动keepalived服务：
[root@dr1 ~]#/etc/init.d/keepalived start
[root@dr1 ~]#/ ipvsadm -l
ip virtual server version 1.2.1 (size=4096)
prot localaddress:port scheduler flags
-> remoteaddress:port forward weight activeconn inactconn
tcp bogon:http rr
-> real-server1:http route 1 1 0
-> real-server2:http route 1 1 0
此时查看keepalived服务的系统日志信息如下:
[root@localhost ~]# tail -f /var/log/messages
feb 28 10:01:56 localhost keepalived: starting keepalived v1.1.19 (02/27,2011)
feb 28 10:01:56 localhost keepalived_healthcheckers: netlink reflector reports ip 192.168.12.25 added
feb 28 10:01:56 localhost keepalived_healthcheckers: opening file '/etc/keepalived/keepalived.conf'.
feb 28 10:01:56 localhost keepalived_healthcheckers: configuration is using : 12063 bytes
feb 28 10:01:56 localhost keepalived: starting healthcheck child process, pid=4623
feb 28 10:01:56 localhost keepalived_vrrp: netlink reflector reports ip 192.168.12.25 added
feb 28 10:01:56 localhost keepalived: starting vrrp child process, pid=4624
feb 28 10:01:56 localhost keepalived_healthcheckers: activating healtchecker for service [192.168.12.246:80]
feb 28 10:01:56 localhost keepalived_vrrp: opening file '/etc/keepalived/keepalived.conf'.
feb 28 10:01:56 localhost keepalived_healthcheckers: activating healtchecker for service [192.168.12.237:80]
feb 28 10:01:57 localhost keepalived_vrrp: vrrp_instance(vi_1) transition to master state
feb 28 10:01:58 localhost keepalived_vrrp: vrrp_instance(vi_1) entering master state
feb 28 10:01:58 localhost keepalived_vrrp: vrrp_instance(vi_1) setting protocol vips.
feb 28 10:01:58 localhost keepalived_healthcheckers: netlink reflector reports ip 192.168.12.135 added
feb 28 10:01:58 localhost avahi-daemon[2778]: registering new address record for 192.168.12.135 on eth0.

二、 高可用性功能测试

高可用性是通过lvs的两个director server完成的，为了模拟故障，我们先将主director server上面的keepalived服务停止，然后观察备用director server上keepalived的运行日志，信息如下：
feb 28 10:08:52 lvs-backup keepalived_vrrp: vrrp_instance(vi_1) transition to master state
feb 28 10:08:54 lvs-backup keepalived_vrrp: vrrp_instance(vi_1) entering master state
feb 28 10:08:54 lvs-backup keepalived_vrrp: vrrp_instance(vi_1) setting protocol vips.
feb 28 10:08:54 lvs-backup keepalived_vrrp: vrrp_instance(vi_1) sending gratuitous arps on eth0 for 192.168.12.135
feb 28 10:08:54 lvs-backup keepalived_vrrp: netlink reflector reports ip 192.168.12.135 added
feb 28 10:08:54 lvs-backup keepalived_healthcheckers: netlink reflector reports ip 192.168.12.135 added
feb 28 10:08:54 lvs-backup avahi-daemon[3349]: registering new address record for 192.168.12.135 on eth0.
feb 28 10:08:59 lvs-backup keepalived_vrrp: vrrp_instance(vi_1) sending gratuitous arps on eth0 for 192.168.12.135
从日志中可以看出，主机出现故障后，备机立刻检测到，此时备机变为master角色，并且接管了主机的虚拟ip资源，最后将虚拟ip绑定在eth0设备上。
接着，重新启动主director server上的keepalived服务，继续观察备用director server的日志状态：
备用director server的日志状态：
feb 28 10:12:11 lvs-backup keepalived_vrrp: vrrp_instance(vi_1) received higher prio advert
feb 28 10:12:11 lvs-backup keepalived_vrrp: vrrp_instance(vi_1) entering backup state
feb 28 10:12:11 lvs-backup keepalived_vrrp: vrrp_instance(vi_1) removing protocol vips.
feb 28 10:12:11 lvs-backup keepalived_vrrp: netlink reflector reports ip 192.168.12.135 removed
feb 28 10:12:11 lvs-backup keepalived_healthcheckers: netlink reflector reports ip 192.168.12.135 removed
feb 28 10:12:11 lvs-backup avahi-daemon[3349]: withdrawing address record for 192.168.12.135 on eth0.
从日志可知，备机在检测到主机重新恢复正常后，重新返回backup角色，并且释放了虚拟ip资源。

三、 负载均衡测试

这里假定两个real server节点配置www服务的网页文件根目录均为/webdata/www目录，然后分别执行如下操作：
在real server1 执行：
echo "this is real server1" /webdata/www/index.html
在real server2 执行：
echo "this is real server2" /webdata/www/index.html
接着打开浏览器，访问http://192.168.12.135这个地址，然后不断刷新此页面，如果能分别看到“this is real server1”和“this is real server2”就表明lvs已经在进行负载均衡了。

四、 故障切换测试

故障切换是测试当某个节点出现故障后，keepalived监控模块是否能及时发现，然后屏蔽故障节点，同时将服务转移到正常节点来执行。
这里我们将real server 1节点服务停掉，假定这个节点出现故障，然后查看主、备机日志信息，相关日志如下：
feb 28 10:14:12 localhost keepalived_healthcheckers: tcp connection to [192.168.12.246:80] failed !!!
feb 28 10:14:12 localhost keepalived_healthcheckers: removing service [192.168.12.246:80] from vs [192.168.12.135:80]
feb 28 10:14:12 localhost keepalived_healthcheckers: remote smtp server [192.168.12.1:25] connected.
feb 28 10:14:12 localhost keepalived_healthcheckers: smtp alert successfully sent.
通过日志可以看出，keepalived监控模块检测到192.168.12.246这台主机出现故障后，将此节点从集群系统中剔除掉了。
此时访问http://192.168.12.135这个地址，应该只能看到“this is real server2”了，这是因为节点1出现故障，而keepalived监控模块将节点1从集群系统中剔除了。
下面重新启动real server 1节点的服务，可以看到keepalived日志信息如下：
feb 28 10:15:48 localhost keepalived_healthcheckers: tcp connection to [192.168.12.246:80] success.
feb 28 10:15:48 localhost keepalived_healthcheckers: adding service [192.168.12.246:80] to vs [192.168.12.135:80]
feb 28 10:15:48 localhost keepalived_healthcheckers: remote smtp server [192.168.12.1:25] connected.
feb 28 10:15:48 localhost keepalived_healthcheckers: smtp alert successfully sent.
从日志可知，keepalived监控模块检测到192.168.12.246这台主机恢复正常后，又将此节点加入了集群系统中。
此时再次访问http://192.168.12.135这个地址，然后不断刷新此页面，应该又能分别看到“this is real server1”和“this is real server2”页面了，这说明在real server 1节点恢复正常后，keepalived监控模块将此节点加入了集群系统中。

本文出自 “技术成就梦想” 博客