一、keepalived简介
keepalived是分布式部署解决系统高可用的软件，结合lvs(LinuxVirtual Server)使用，解决单机宕机的问题。
keepalived是一个基于VRRP协议来实现IPVS的高可用的解决方案。对于LVS负载均衡来说，如果前端的调度器direct发生故障，则后端的realserver是无法接受请求并响应的。因此，保证前端direct的高可用性是非常关键的，否则后端的服务器是无法进行服务的。而我们的keepalived就可以用来解决单点故障（如LVS的前端direct故障）问题。keepalived的主要工作原理是：运行keepalived的两台服务器，其中一台为MASTER，另一台为BACKUP，正常情况下，所有的数据转换功能和ARP请求响应都是由MASTER完成的，一旦MASTER发生故障，则BACKUP会马上接管MASTER的工作，这种切换时非常迅速的。
二、测试环境
下面拿4台虚拟机进行环境测试,实验环境为centos6.6 x86_64,具体用途和ip如下
|
服务器类型 |
IP地址 |
|
Lvs VIP |
192.168.214.89 |
|
Keepalived Master |
192.168.214.85 |
|
Keepalived Backup |
192.168.214.86 |
|
Realserver A |
192.168.214.87 |
|
Realserver B |
192.168.214.88 |
三、软件安装
1、安装lvs所需包ipvsadm
yum install -y ipvsadm
ln -s /usr/src/kernels/`uname -r` /usr/src/linux
lsmod |grep ip_vs
#注意Centos 6.X安装lvs，使用1.26版本。并且需要先安装yuminstall libnl* popt* -y
执行ipvsadm(modprobe ip_vs)把ip_vs模块加载到内核
[[email protected] ~]# ipvsadm -L -n
IP Virtual Server version 1.2.1 (size=4096)
Prot LocalAddress:Port Scheduler Flags
-> RemoteAddress:Port Forward Weight ActiveConn InActConn
#IP Virtual Server version 1.2.1 ---- ip_vs内核模块版本
2、安装keepalived
yum install -y keepalived
chkconfig keepalived on
注：在centos7系列系统中开机自动启动使用systemctl enable keepalived
四、keepalived配置
先看下214.85keepalived主机上的配置
[[email protected] ~]# cat /etc/keepalived/keepalived.conf
! Configuration File for keepalived
global_defs {
#配置报警邮箱
notification_email {
[email protected]
}
notification_email_from [email protected]
smtp_server mail.test.com
smtp_connect_timeout 30
router_id LVS_DEVEL
}
vrrp_sync_group VG1 {
group {
VI_1
}
}
vrrp_instance VI_1 {
state MASTER #指定keepalived的角色，MASTER表示此主机是主服务器，BACKUP表示此主机是备用服务器
interface eth0 #指定HA监测网络的接口
lvs_sync_daemon_inteface eth0
virtual_router_id 55
#虚拟路由标识，这个标识是一个数字，同一个vrrp实例使用唯一的标识。即同一vrrp_instance下，MASTER和BACKUP必须是一致的
priority 100 #定义优先级，数字越大，优先级越高，在同一个vrrp_instance下，MASTER的优先级必须大于BACKUP的优先级
advert_int 1 #设定MASTER与BACKUP负载均衡器之间同步检查的时间间隔，单位是秒
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress { #设置虚拟IP地址
192.168.214.89
}
}
virtual_server 192.168.214.89 80 {
delay_loop 6 #(每隔6秒查询realserver状态)
lb_algo rr #(lvs 算法)
lb_kind DR #(使用lvs的DR模式)
#nat_mask 255.255.255.0
persistence_timeout 10 #(同一IP的连接10秒内被分配到同一台realserver)
protocol TCP #(用TCP协议检查realserver状态)
real_server 192.168.214.87 80 {
weight 100 #(权重)
TCP_CHECK {
connect_timeout 3 #(3秒无响应超时)
connect_port 80
nb_get_retry 3
delay_before_retry 3
}
}
real_server 192.168.214.88 80 {
weight 100
TCP_CHECK {
connect_timeout 3
connect_port 80
nb_get_retry 3
delay_before_retry 3
}
}
}
再看下214.86keepalived备机上的配置
! Configuration File for keepalived
global_defs {
notification_email {
[email protected]
}
notification_email_from [email protected]
smtp_server mail.test.com
smtp_connect_timeout 30
router_id LVS_DEVEL
}
vrrp_sync_group VG1 {
group {
VI_1
}
}
vrrp_instance VI_1 {
state BACKUP
interface eth0
lvs_sync_daemon_inteface eth0
virtual_router_id 55
priority 90
advert_int 1
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
192.168.214.89
}
}
virtual_server 192.168.214.89 80 {
delay_loop 6
lb_algo rr
lb_kind DR
#nat_mask 255.255.255.0
persistence_timeout 10
protocol TCP
real_server 192.168.214.87 80 {
weight 100
TCP_CHECK {
connect_timeout 3
connect_port 80
nb_get_retry 3
delay_before_retry 3
}
}
real_server 192.168.214.88 80 {
weight 100
TCP_CHECK {
connect_timeout 3
connect_port 80
nb_get_retry 3
delay_before_retry 3
}
}
}
五、后端realserver操作
DR模式需要在后端真实机上运行以下脚本
#!/bin/bash
# description: Config realserver lo
#Written by :Charles
VIP1=192.168.214.89
. /etc/rc.d/init.d/functions
case “$1” in
start)
ifconfig lo:0 $VIP1 netmask 255.255.255.255 broadcast $VIP1
/sbin/route add –host $VIP1 dev lo:0
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
sysctl –p >/dev/null 2>&1
echo “RealServer Start OK”
;;
stop)
ifconfig lo:0 down
route del $VIP1 >/dev/null 2>&1
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
echo “RealServer Stoped”
;;
*)
echo “Usage: $0 {start|stop}”
exit 1
esac
exit 0
#执行realserver.sh start开启，stop关闭
#脚本设置成755权限，并放入rc.local下让其开机启动运行
六、启动keepalived服务及查看相关信息
在214.85和214.86上分别启动keepalived服务
在214.85keepalived主机上查看信息
通过ip addr可以看到vip 地址已经绑定在eth0网口上
[[email protected] ~]# ip addr
1: lo: <LOOPBACK,UP,LOWER_UP> mtu65536 qdisc noqueue state UNKNOWN
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
2: eth0:<BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen1000
link/ether 00:0c:29:85:7a:67 brd ff:ff:ff:ff:ff:ff
inet 192.168.214.85/24 brd 192.168.214.255 scope global eth0
inet 192.168.214.89/32 scope global eth0
inet6 fe80::20c:29ff:fe85:7a67/64 scope link
valid_lft forever preferred_lft forever
在214.85上查看日志信息，看到已成功进入keepalived主机模式
[[email protected] ~]# tail -f /var/log/messages
May 4 14:12:34 test85 Keepalived_vrrp[7977]: VRRP_Instance(VI_1) Entering MASTER STATE
May 4 14:12:34 test85 Keepalived_vrrp[7977]: VRRP_Instance(VI_1) settingprotocol VIPs.
May 4 14:12:34 test85 Keepalived_healthcheckers[7975]: Netlink reflector reports IP 192.168.214.89 added
May 4 14:12:34 test85 Keepalived_vrrp[7977]: VRRP_Instance(VI_1) Sendinggratuitous ARPs on eth0 for 192.168.214.89
May 4 14:12:34 test85 Keepalived_vrrp[7977]: VRRP_Group(VG1) Syncinginstances to MASTER state
May 4 14:12:36 test85 ntpd[1148]: Listen normally on 7 eth0 192.168.214.89UDP 123
May 4 14:12:36 test85 ntpd[1148]: peers refreshed
May 4 14:12:39 test85 Keepalived_vrrp[7977]: VRRP_Instance(VI_1) Sendinggratuitous ARPs on eth0 for 192.168.214.89
May 4 14:12:40 test85 root[7924] 192.168.5.80 53823 192.168.214.85 22:#1525414360
May 4 14:12:40 test85 root[7924] 192.168.5.80 53823 192.168.214.85 22: ipaddr
在214.86上查看日志信息，看到已成功进入keepalived备机模式
May 4 14:12:37 web86 Keepalived_vrrp[31009]: Using LinkWatch kernel netlinkreflector...
May 4 14:12:37 web86 Keepalived_vrrp[31009]: VRRP_Instance(VI_1) Entering BACKUP STATE
May 4 14:12:37 web86 Keepalived_vrrp[31009]: VRRP sockpool: [ifindex(2),proto(112), unicast(0), fd(10,11)]
May 4 14:12:37 web86 Keepalived_healthcheckers[31007]: Opening file'/etc/keepalived/keepalived.conf'.
May 4 14:12:37 web86 Keepalived_healthcheckers[31007]: Configuration isusing : 14713 Bytes
May 4 14:12:37 web86 Keepalived_healthcheckers[31007]: Using LinkWatchkernel netlink reflector...
May 4 14:12:37 web86 Keepalived_healthcheckers[31007]: Activatinghealthchecker for service [192.168.214.87]:80
May 4 14:12:37 web86 Keepalived_healthcheckers[31007]: Activatinghealthchecker for service [192.168.214.88]:80
后端真实机启动脚本后，查看网卡信息，看到vip已成功绑定在回环口上。
[[email protected] ~]# ipaddr
1: lo: <LOOPBACK,UP,LOWER_UP> mtu65536 qdisc noqueue state UNKNOWN
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
inet 192.168.214.89/32 brd 192.168.214.89 scope global lo:0
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 qlen1000
link/ether 00:0c:29:38:31:ad brd ff:ff:ff:ff:ff:ff
inet 192.168.214.87/24 brd 192.168.214.255 scope global eth0
inet6 fe80::20c:29ff:fe38:31ad/64 scope link
valid_lft forever preferred_lft forever
通过ipvsadm –L –n查看相应lvs连接信息
[[email protected] ~]# ipvsadm -L -n
IP Virtual Server version 1.2.1 (size=4096)
Prot LocalAddress:Port Scheduler Flags
-> RemoteAddress:Port Forward Weight ActiveConn InActConn
TCP 192.168.214.89:80 rr persistent 10
-> 192.168.214.87:80 Route 100 2 2
-> 192.168.214.88:80 Route 100 0 0
七、keepalived测试
使用vip地址192.168.214.89访问后端192.168.214.87和192.168.214.88的页面
正常访问没问题后，我们来模拟lvs集群故障
首先把keepalived master主机214.85宕机，看备机能否接管过来，vip地址是否会漂移过来
在214.86上查看日志，发现备机成功切换成了主机状态，vip地址成功漂移了过来
May 4 14:35:34 web86 Keepalived_vrrp[31009]: VRRP_Instance(VI_1) Transition to MASTER STATE
May 4 14:35:34 web86 Keepalived_vrrp[31009]: VRRP_Group(VG1) Syncinginstances to MASTER state
May 4 14:35:35 web86 Keepalived_vrrp[31009]: VRRP_Instance(VI_1) EnteringMASTER STATE
May 4 14:35:35 web86 Keepalived_vrrp[31009]: VRRP_Instance(VI_1) settingprotocol VIPs.
May 4 14:35:35 web86 Keepalived_vrrp[31009]: VRRP_Instance(VI_1) Sendinggratuitous ARPs on eth0 for 192.168.214.89
May 4 14:35:35 web86 Keepalived_healthcheckers[31007]: Netlink reflectorreports IP 192.168.214.89 added
May 4 14:35:36 web86 ntpd[1230]: Listen normally on 7 eth0 192.168.214.89UDP 123
May 4 14:35:36 web86 ntpd[1230]: peers refreshed
May 4 14:35:40 web86 Keepalived_vrrp[31009]: VRRP_Instance(VI_1) Sendinggratuitous ARPs on eth0 for 192.168.214.89
然后，再把214.85主机恢复，由于214.85拥有较高的优先级，会从214.86抢回MASTER状态，相应的214.86会回归到原来的Backup状态
214.85日志记录，重新回到了MASTER状态
May 4 14:41:55 test85 Keepalived_vrrp[8066]: VRRP_Instance(VI_1) Transitionto MASTER STATE
May 4 14:41:55 test85 Keepalived_vrrp[8066]: VRRP_Instance(VI_1) Receivedlower prio advert, forcing new election
May 4 14:41:55 test85 Keepalived_vrrp[8066]: VRRP_Group(VG1) Syncinginstances to MASTER state
May 4 14:41:56 test85 Keepalived_vrrp[8066]: VRRP_Instance(VI_1) EnteringMASTER STATE
May 4 14:41:56 test85 Keepalived_vrrp[8066]: VRRP_Instance(VI_1) settingprotocol VIPs.
May 4 14:41:56 test85 Keepalived_vrrp[8066]: VRRP_Instance(VI_1) Sendinggratuitous ARPs on eth0 for 192.168.214.89
May 4 14:41:56 test85 Keepalived_healthcheckers[8064]: Netlink reflectorreports IP 192.168.214.89 added
May 4 14:41:58 test85 ntpd[1148]: Listen normally on 8 eth0 192.168.214.89UDP 123
May 4 14:41:58 test85 ntpd[1148]: peers refreshed
May 4 14:42:01 test85 Keepalived_vrrp[8066]: VRRP_Instance(VI_1) Sendinggratuitous ARPs on eth0 for 192.168.214.89
218.86日志记录,接收到了高优先级请求，从之前的MASTER状态变回了BACKUP状态
May 4 14:35:34 web86 Keepalived_vrrp[31009]: VRRP_Group(VG1) Syncinginstances to MASTER state
May 4 14:35:35 web86 Keepalived_vrrp[31009]: VRRP_Instance(VI_1) Entering MASTER STATE
May 4 14:35:35 web86 Keepalived_vrrp[31009]: VRRP_Instance(VI_1) settingprotocol VIPs.
May 4 14:35:35 web86 Keepalived_vrrp[31009]: VRRP_Instance(VI_1) Sendinggratuitous ARPs on eth0 for 192.168.214.89
May 4 14:35:35 web86 Keepalived_healthcheckers[31007]: Netlink reflectorreports IP 192.168.214.89 added
May 4 14:35:36 web86 ntpd[1230]: Listen normally on 7 eth0 192.168.214.89UDP 123
May 4 14:35:36 web86 ntpd[1230]: peers refreshed
May 4 14:35:40 web86 Keepalived_vrrp[31009]: VRRP_Instance(VI_1) Sendinggratuitous ARPs on eth0 for 192.168.214.89
May 4 14:36:41 web86 root[30963] 192.168.5.80 53824 192.168.214.86 22:#1525415801
May 4 14:36:41 web86 root[30963] 192.168.5.80 53824 192.168.214.86 22: ipaddr
May 4 14:41:55 web86 Keepalived_vrrp[31009]: VRRP_Instance(VI_1) Received higher prio advert
May 4 14:41:55 web86 Keepalived_vrrp[31009]: VRRP_Instance(VI_1) Entering BACKUP STATE
May 4 14:41:55 web86 Keepalived_vrrp[31009]: VRRP_Instance(VI_1) removingprotocol VIPs.
May 4 14:41:55 web86 Keepalived_vrrp[31009]: VRRP_Group(VG1) Syncinginstances to BACKUP state
May 4 14:41:55 web86 Keepalived_healthcheckers[31007]: Netlink reflectorreports IP 192.168.214.89 removed
May 4
14:41:56 web86 ntpd[1230]: Deleting interface #7
eth0,192.168.214.89#123, interface stats: received=0, sent=0,
dropped=0,active_time=380 secs
最后，再模拟后端真实机214.87服务宕掉，看是否vip只请求214.88
通过日志查看得知，keepalived集群探测到后端真实机214.87的80端口不通，把它从vip请求列表中移除了
May 414:48:00 test85 Keepalived_healthcheckers[8064]: TCP connection to[192.168.214.87]:80 failed !!!
May 4 14:48:00 test85Keepalived_healthcheckers[8064]: Removing service [192.168.214.87]:80 from VS[192.168.214.89]:80
当重新探测到后端真实机214.87服务恢复后，又把它加入了请求列表中
May 4 14:52:55 test85 Keepalived_healthcheckers[8064]: TCP connection to[192.168.214.87]:80 success.
May 4 14:52:55 test85 Keepalived_healthcheckers[8064]: Adding service[192.168.214.87]:80 to VS [192.168.214.89]:80
如果想了解更多，请关注我们的公众号
公众号ID:opdevos
扫码关注
原文地址：http://blog.51cto.com/5ydycm/2118107