keepalived

时间：2016-08-22 23:50:04 阅读：357 评论：0 收藏：0 [点我收藏+]

标签：keepalived 高可用

keepalived：

vrrp协议：Virtual Redundant Routing Protocol 虚拟冗余路由协议（网关高可用）

术语：

虚拟路由器（由一个或更多个物理设备组成）、

VRID（同一组的虚拟路由器的唯一标识）、

master路由器（承担转发报文）、backup路由器（预备军）、

VIP（虚拟ip地址，真正提供服务的ip，且在多个节点上面流转）、

VMAC（00-00-5e-00-01-VRID，vRID有效范围0-255）（虚拟mac地址，虚拟ip在哪则虚拟mac就在哪）

但其实实际上是当vip更换了主机之后，自己发起一个免费的ARP通告，并自问自答，让别人知道江山易主了

优先级（每个节点的优先级不同可以判定哪个能够指点江山）

抢占式（只要优先级高，我就抢）、非抢占式（等着主节点挂了，才能接受，暂时不管优先级）

vrrp工作流程：至少两个，一主一辅，当平时的时候主的上面跑着ip，发送免费arp让别人知道我这个节点是主动，提供服务；master发送周期性的arrp报文，说明自己的心跳，从在一旁观察这主的心跳信息；当主的服务器出现故障，从的不能接受到心跳信息之后，根据优先级选举，就把vip配置在自己的主机上面，接受服务，对外响应；这个地址是流动的，在这些节点之间进行流转，谁有这个ip，谁就可以响应服务；

vrrp主要是完成地址转移，保障服务始终可以对外响应

vrrp的认证：无认证，签名认证和md5认证三种

工作模式：

主/备（这样从的只有当主的挂了之后才能上台，有点儿浪费资源）

主/主：配置多个virtual router；（主要是不想让某个服务器始终闲着）

这里的双主，彼此为主，彼此为辅，配置两个vip，各自响应，让服务器同时响应请求

第一台

[root@zou ~]# hostnamectl set-hostname one.zou.com

[root@one ~]# vim /etc/hosts

172.16.1.3 one.zou.com one

172.16.1.5 two.zou.com two

第二台

[root@zou ~]# hostnamectl set-hostname two.zou.com

[root@one ~]# vim /etc/hosts

172.16.1.3 one.zou.com one

172.16.1.5 two.zou.com two

之后可以达到通过主机名通信：

[root@one ~]# ping two.zou.com

PING two.zou.com (172.16.1.5) 56(84) bytes of data.

64 bytes from two.zou.com (172.16.1.5): icmp_seq=1 ttl=64 time=1.72 ms

64 bytes from two.zou.com (172.16.1.5): icmp_seq=2 ttl=64 time=0.694 ms

同步时间（两台服务器都要同步时间，保持时间基本相同）

可以自己组建时间服务器

yum install -y chrony

vim /etc/chrony.conf

allow 172.16/16 允许哪些主机同步时间

server 172.16.1.1 iburst 时间服务器同步的时间来源

systemctl start chronyd.service

systemctl enable chronyd.service

chronyc -h 172.16.1.1 同步时间（指向这个搭建的时间服务器）

也可以在网上同步时间

[root@one ~]# ntpdate 172.16.0.1

安装keepalived

[root@one ~]# yum install keepalived -y

另一台也要安装

[root@two ~]# yum install keepalived -y

配置keepalived

主：

[root@one ~]# cp /etc/keepalived/keepalived.conf{,.bak}

[root@one ~]# vim /etc/keepalived/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 one

}

vrrp_instance VI_1 {

state MASTER

interface eno16777736

virtual_router_id 101

priority 100

advert_int 1

authentication {

auth_type PASS

auth_pass 123.com

}

virtual_ipaddress {

172.16.1.11

}

从：

[root@two ~]# vim /etc/keepalived/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 one

}

vrrp_instance VI_1 {

state BACKUP

interface eno16777736

virtual_router_id 101

priority 90

advert_int 1

authentication {

auth_type PASS

auth_pass 123.com

}

virtual_ipaddress {

172.16.1.11

}

[root@two ~]# systemctl start keepalived.service

[root@two ~]# systemctl status keepalived.service

● keepalived.service - LVS and VRRP High Availability Monitor

Active: active (running) since Mon 2016-08-01 07:13:32 CST; 20s ago

......

Aug 01 07:13:36 two.zou.com Keepalived_vrrp[128056]: VRRP_Instance(VI_1) Transition to MASTER STATE

上条说先去找主节点

Aug 01 07:13:37 two.zou.com Keepalived_vrrp[128056]: VRRP_Instance(VI_1) Entering MASTER STATE

由于主节点还没有开启，所以虽然它属于从，主的不在，它就变成了主

Aug 01 07:13:37 two.zou.com Keepalived_vrrp[128056]: VRRP_Instance(VI_1) setting protocol VIPs.

准备协议VIPS

Aug 01 07:13:37 two.zou.com Keepalived_vrrp[128056]: VRRP_Instance(VI_1) Sending gratuitous ARPs on eno16777736 for 172.16.1.11 设置好了vip

Aug 01 07:13:37 two.zou.com Keepalived_healthcheckers[128055]: Netlink reflector reports IP 172.16.1.11 added 免费通告，自己有vip，为主节点

Aug 01 07:13:42 two.zou.com Keepalived_vrrp[128056]: VRRP_Instance(VI_1) Sending gratuitous ARPs on eno16777736 for 172.16.1.11 免费通告

[root@two ~]# ip addr show eno16777736

inet 172.16.1.5/16 brd 172.16.255.255 scope global eno16777736

valid_lft forever preferred_lft forever

inet 172.16.1.11/32 scope global eno16777736

valid_lft forever preferred_lft forever

好了这个时候该把，172.16.1.3上的主节点启动了，由于节点是争用的，主节点优先级高，所以很开就会把主节点抢过去

[root@one ~]# systemctl start keepalived.service

You have new mail in /var/spool/mail/root 主节点发生变化，就要发送邮件了

之后通过状态监测keepalived.service，可以看到

one.zou.com Keepalived_vrrp[2385]: VRRP_Instance(VI_1) Sending gratuitous ARPs on eno16777736 for 172.16.1.11 已经把VIP给争到了，它就是主的了，相对而言1.5那边就变成了从节点

two.zou.com Keepalived_healthcheckers[128055]: Netlink reflector reports IP 172.16.1.11 removed

主主模型（互为主从）

使用环境：

（1）两个主都配置在一个服务器上面，剩下的两个从配置一个上面；

在有内外网的时候，且同时提供对内和对外的服务，为了保障服务不中断，因此配置两个虚拟段，互为主从，保障在keepalived服务器即使在一台有问题的情况下可以运转

（这里这个配置和上面一样，就是多加个虚拟路由而已，所以就不写操作步骤了）

（2）互为主从，配置两个虚拟路由，一个上面为主，同时住上面还有一个虚拟路由是另一个的从节点

在网络配置完毕，同时都安装了keepalived服务之后，修改配置文件：

172.16.1.3上面相关操作与配置

172.16.1.5上面相关操作与配置

@one ~]# vim /etc/keepalived/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 one

}

vrrp_instance VI_1 {

state MASTER

interface eno16777736

virtual_router_id 101

priority 100

advert_int 1

authentication {

auth_type PASS

auth_pass 123.com

}

virtual_ipaddress {

172.16.1.11

}

vrrp_instance VI_2 {

state BACKUP

interface eno16777736

virtual_router_id 111

priority 90

advert_int 1

authentication {

auth_type PASS

auth_pass 456.com

}

virtual_ipaddress {

172.16.1.22

}

@two ~]# vim/etc/keepalived/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 one

}

vrrp_instance VI_1 {

state BACKUP

interface eno16777736

virtual_router_id 101

priority 90

advert_int 1

authentication {

auth_type PASS

auth_pass 123.com

}

virtual_ipaddress {

172.16.1.11

}

vrrp_instance VI_2 {

state MASTER

interface eno16777736

virtual_router_id 111

priority 100

advert_int 1

authentication {

auth_type PASS

auth_pass 456.com

}

virtual_ipaddress {

172.16.1.22

}

之后两台服务器启动服务，两台服务器上分配的ip

[root@one ~]# ip addr show eno16777736

inet 172.16.1.3/16 brd 172.16.255.255 scope global eno16777736

inet 172.16.1.11/32 scope global eno16777736

[root@two ~]# ip addr show eno16777736

inet 172.16.1.5/16 brd 172.16.255.255 scope global eno16777736

inet 172.16.1.22/32 scope global eno16777736

根据拓扑图完成后续操作

首先定义好后端的两台wab服务器，同时给他们设置好限制通告和vip设置

[root@luffy www]# echo 1 > /proc/sys/net/ipv4/conf/all/arp_ignore

[root@luffy www]# echo 1 > /proc/sys/net/ipv4/conf/lo/arp_ignore

[root@luffy www]# echo 2 > /proc/sys/net/ipv4/conf/lo/arp_announce

[root@luffy www]# echo 2 > /proc/sys/net/ipv4/conf/all/arp_announce

[root@luffy www]# ip addr add 172.16.1.44/32 broadcast 172.16.1.44 dev lo label lo:0

让每次首先经由vip转发（此命令可以不执行）

[root@luffy www]# ip route add 172.16.1.44 dev lo:0

准备第二台的web服务器

[root@zou ~]# echo 1 > /proc/sys/net/ipv4/conf/all/arp_ignore

[root@zou ~]# echo 1 > /proc/sys/net/ipv4/conf/lo/arp_ignore

[root@zou ~]# echo 2 > /proc/sys/net/ipv4/conf/lo/arp_announce

[root@zou ~]# echo 2 > /proc/sys/net/ipv4/conf/all/arp_announce

[root@zou ~]# ifconfig lo:0 172.16.1.44 255.255.255.255 broadcast 172.16.1.44 up

[root@zou ~]# route add -host 172.16.1.44 dev lo:0

在keepalive上（两台）都准备好sorry web，当然也要安装web服务（简单配置即可）

[root@one ~]# yum install httpd -y

[root@one ~]# systemctl start httpd.service

[root@one ~]# vim /var/www/html/index.html

[root@two ~]# yum install httpd -y 或者安装nginx也可以

[root@two ~]# systemctl start httpd.service

[root@two ~]# vim /var/www/html/index.html

生产环境这个sorry网页是一样的，但是里测试故意把这个服务器的sorry网页写的不一样，看看是否东能够正常工作

好了可以准备配置keepalived了，这里还需要说明一下，这个时候不用安装ipvsadm，因为keepalived会自动向内核中找ipvs的系统调用去完成dr模型的调度，完成负载，同时两台keepalived完成高可用设置，

不过，为了我们在实验环节能够更好的查看（仅为查看使用）可以安装上ipvsadm软件包

[root@one etc]# yum install -y ipvsadm

[root@two ~]# yum install -y ipvsadm

[root@one 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 one 另一个为two

vrrp_mcast_group4 224.0.101.11

}

vrrp_instance VI_1 {

state MASTER 另一台主机为BACKUP

interface eno16777736

virtual_router_id 22

priority 100 另一个为90

advert_int 1

authentication {

auth_type PASS

auth_pass 123.com

}

virtual_ipaddress {

172.16.1.44/32 dev eno16777736 brd 172.16.1.44 label eno16777736:0

}

virtual_server 172.16.1.44 80 {

delay_loop 3

lb_algo rr

lb_kind DR

nat_mask 255.255.0.0

#persistence_timeout 50

protocol TCP

sorry_server 127.0.0.1 80

real_server 172.16.1.1 80 {

weight 1

HTTP_GET {

url {

path /index.html

#digest 640205b7b0fc66c1ea91c463fac6334d

status_code 200

}

#url {

# path /testurl2/test.jsp

#digest 640205b7b0fc66c1ea91c463fac6334d

connect_timeout 3

nb_get_retry 3

delay_before_retry 3

}

real_server 172.16.1.2 80 {

weight 1

HTTP_GET {

url {

path /index.html

status_code 200

}

connect_timeout 3

nb_get_retry 3

delay_before_retry 3

}

之后两台keepalived主机重启服务，查看状态

[root@two keepalived]# systemctl restart keepalived.service

[root@two keepalived]# systemctl status keepalived.service

[root@one keepalived]# ip a show eno16777736

inet 172.16.1.3/16 brd 172.16.255.255 scope global eno16777736

inet 172.16.1.44/32 brd 172.16.1.44 scope global eno16777736:0

当然我们一开始安装ipvsadm就是为了这个时候查看状态的，不然安装了也就用不着了

[root@one keepalived]# ipvsadm -ln

IP Virtual Server version 1.2.1 (size=4096)

Prot LocalAddress:Port Scheduler Flags

-> RemoteAddress:Port Forward Weight ActiveConn InActConn

TCP 172.16.1.44:80 rr

-> 172.16.1.1:80 Route 1 0 0

-> 172.16.1.2:80 Route 1 0 0

查看调度器是否能够轮训调度

[root@two keepalived]# for i in {1..6};do curl http://172.16.1.44;done

this is web 2

this is web 1

this is web 2

this is web 1

this is web 2

this is web 1

测试web服务器宕机，是否还可正常使用：

[root@zou ~]# /etc/init.d/httpd stop （这里停掉的是1.2）

之后我们用ipvsadm去查看

[root@one ~]# ipvsadm -ln

IP Virtual Server version 1.2.1 (size=4096)

Prot LocalAddress:Port Scheduler Flags

-> RemoteAddress:Port Forward Weight ActiveConn InActConn

TCP 172.16.1.44:80 rr

-> 172.16.1.1:80 Route 1 0 0

我们可以从结果看到现在只有一个服务在线

[root@two ~]# for i in {1..6};do curl http://172.16.1.44;done

this is web 1

测试，同时把web1和web2给停掉

[root@luffy ~]# nginx -s stop

继续查看和监测

[root@one ~]# ipvsadm

IP Virtual Server version 1.2.1 (size=4096)

Prot LocalAddress:Port Scheduler Flags

-> RemoteAddress:Port Forward Weight ActiveConn InActConn

TCP 172.16.1.44:http rr

-> localhost:http Route 1 0 0

可以看到sorryserver上台了

[root@one ~]# curl http://172.16.1.44

sorry!!! this is web 1.3

测试停掉keepalived主，节点

[root@one ~]# systemctl stop keepalived.service

[root@two ~]# curl http://172.16.1.44

sorry!this is web 1.5

基于状态的监测

[root@one ~]# vim /etc/keepalived/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 one

vrrp_mcast_group4 224.0.101.11

}

vrrp_script test_down {

script "[[ -f /etc/keepalived/down ]] && exit1 || exit 0"

interval 10

weight -5

}

vrrp_instance VI_1 {

state MASTER

interface eno16777736

virtual_router_id 22

priority 100

advert_int 1

authentication {

auth_type PASS

auth_pass 123.com

}

virtual_ipaddress {

172.16.1.66

}

track_script {

test_down

}

[root@one ~]# scp /etc/keepalived/keepalived.conf two:/etc/keepalived/keepalived.conf

[root@two ~]# vim /etc/keepalived/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 two

vrrp_mcast_group4 224.0.101.11

}

vrrp_script test_down {

script "[[ -f /etc/keepalived/down ]] && exit 1 || exit 0" 判断是否有down这个文件，如果有文件就返回失败值1，如果不存在就返回成功值0。

interval 10

weight -5 当失败的时候权重减少5

}

vrrp_instance VI_1 {

state BACKUP

interface eno16777736

virtual_router_id 22

priority 98

advert_int 1

authentication {

auth_type PASS

auth_pass 123.com

}

virtual_ipaddress {

172.16.1.66

}

track_script {

test_down

}

之后两个服务器重启keepalived服务，这个时候应该是主节点，也就是有优先级为100的为主

这个时候在主几点上创建/etc/keepalived/down文件，那么其优先级就会降低3，变成97，而从的默认为98，则从节点上就会变成主节点了

[root@one ~]# ip addr show eno16777736 | grep 32

inet 172.16.1.66/32scope global eno16777736

[root@two ~]# ip addr show eno16777736 | grep 32

之后主节点创建down文件

[root@one ~]# touch /etc/keepalived/down

[root@one ~]# ip addr show eno16777736 | grep 32

[root@one keepalived]# systemctl status keepalived.service

VRRP_Instance(VI_1)Received higher prio advert

VRRP_Instance(VI_1)Entering BACKUP STATE

: VRRP_Instance(VI_1) removing protocol VIPs.

Netlink reflectorreports IP 172.16.1.66 removed

[root@two ~]# ip addr showeno16777736 | grep 32

inet172.16.1.66/32 scope global eno16777736

如果把这个down文件给删了，原先的主节点一会就会成为主的

[root@one keepalived]# rm -rf down

[root@one keepalived]#systemctl status keepalived.service 这是回复节点的日志

:VRRP_Script(chk_down) succeeded

VRRP_Instance(VI_1)forcing a new MASTER election

VRRP_Instance(VI_1)Transition to MASTER STATE

VRRP_Instance(VI_1)Entering MASTER STATE

VRRP_Instance(VI_1)setting protocol VIPs.

VRRP_Instance(VI_1)Sending gratuitous ARPs on eno16777736 for 172.16.1.66

Netlink reflectorreports IP 172.16.1.66 added

VRRP_Instance(VI_1)Sending gratuitous ARPs on eno16777736 for 172.16.1.66

从上面这个实验，down文件的存在与否可以判断哪个是作为主节点；同理，我们可以判断一个服务时候存在，之后来调整keepalived的活动节点，从而完成高可用的。不如我们可以监测httpd或者nginx服务是否在线，从而来调整keepalived的活动节点，从而让httpd或者nginx服务达到高可用的目的

（这里补充一下，如果killall在centos7上没有的话，就安装psmisc软件包）

注：如果式样过程有不顺利的地方，可能是keepalived没有启动好，先用killall keepalived 将其关闭，之后启动再测试

[root@two keepalived]# killall -0 nginx 探测服务是否存在

[root@two keepalived]# echo $?

0 如果为0则说明服务正常响应

[root@two keepalived]# nginx -s stop

[root@two keepalived]# killall -0 nginx

nginx: no process found

[root@two keepalived]# echo $?

[root@one keepalived]# killall -0 httpd

[root@one keepalived]# echo $?

0 所以以后可以使用kill -0 来探测某服务是否存在了

下面来设置一下，keepalived.conf ,在里面加入

global_defs {

notification_email {

root@localhost

}

notification_email_from keepalived@localhost

smtp_server 127.0.0.1

smtp_connect_timeout 30

router_id one

vrrp_mcast_group4 224.0.101.11

}

vrrp_script chk_httpd {

script "killall -0 httpd && exit 0 || exit 1"

interval 1

weight -5

}

vrrp_instance VI_1 {

state MASTER

interface eno16777736

virtual_router_id 22

priority 100

advert_int 1

authentication {

auth_type PASS

auth_pass 123.com

}

virtual_ipaddress {

172.16.1.66

}

track_script {

chk_httpd

}

接下来设置另一个服务器上的，这是1.5上面的，我这上面跑的是nginx

vrrp_script chk_nginx {

script "killall -0 nginx && exit 0 || exit 1"

interval 1

weight -5

}

track_script {

test_down

chk_nginx

}

重启两个keepalived服务

测试，将1.3上面的httpd关闭，看起主节点是否发生变化：

[root@one keepalived]# systemctl stop httpd

[root@one keepalived]# systemctl status keepalived.service

Aug 01 08:22:18 one.zou.com Keepalived_vrrp[8805]: VRRP_Instance(VI_1) Entering MASTER STATE

Aug 01 08:22:18 one.zou.com Keepalived_vrrp[8805]: VRRP_Instance(VI_1) setting protocol VIPs.

Aug 01 08:22:18 one.zou.com Keepalived_vrrp[8805]: VRRP_Instance(VI_1) Sending gratuitous ARPs on eno16777736 for 172.16.1.66

Aug 01 08:22:18 one.zou.com Keepalived_healthcheckers[8804]: Netlink reflector reports IP 172.16.1.66 added

Aug 01 08:22:23 one.zou.com Keepalived_vrrp[8805]: VRRP_Instance(VI_1) Sending gratuitous ARPs on eno16777736 for 172.16.1.66

Aug 01 08:22:31 one.zou.com Keepalived_vrrp[8805]: VRRP_Script(chk_httpd) failed

Aug 01 08:22:32 one.zou.com Keepalived_vrrp[8805]: VRRP_Instance(VI_1) Received higher prio advert

Aug 01 08:22:32 one.zou.com Keepalived_vrrp[8805]: VRRP_Instance(VI_1) Entering BACKUP STATE

Aug 01 08:22:32 one.zou.com Keepalived_vrrp[8805]: VRRP_Instance(VI_1) removing protocol VIPs.

Aug 01 08:22:32 one.zou.com Keepalived_healthcheckers[8804]: Netlink reflector reports IP 172.16.1.66 removed

当然这个时候ip已经转移到1.5主机上面去了

这是时候原主机如果开启httpd服务的话：

[root@one keepalived]# systemctl start httpd

[root@one keepalived]# ip a show eno16777736 | grep 32

inet 172.16.1.66/32 scope global eno16777736

[root@one keepalived]# systemctl status keepalived.service

Aug 01 08:22:32 one.zou.com Keepalived_healthcheckers[8804]: Netlink reflector reports IP 172.16.1.66 removed

Aug 01 08:24:48 one.zou.com Keepalived_vrrp[8805]: VRRP_Script(chk_httpd) succeeded

Aug 01 08:24:48 one.zou.com Keepalived_vrrp[8805]: VRRP_Instance(VI_1) forcing a new MASTER election

Aug 01 08:24:49 one.zou.com Keepalived_vrrp[8805]: VRRP_Instance(VI_1) Transition to MASTER STATE

Aug 01 08:24:50 one.zou.com Keepalived_vrrp[8805]: VRRP_Instance(VI_1) Entering MASTER STATE

Aug 01 08:24:50 one.zou.com Keepalived_vrrp[8805]: VRRP_Instance(VI_1) setting protocol VIPs.

Aug 01 08:24:50 one.zou.com Keepalived_vrrp[8805]: VRRP_Instance(VI_1) Sending gratuitous ARPs on eno16777736 for 172.16.1.66

Aug 01 08:24:50 one.zou.com Keepalived_healthcheckers[8804]: Netlink reflector reports IP 172.16.1.66 added

Aug 01 08:24:55 one.zou.com Keepalived_vrrp[8805]: VRRP_Instance(VI_1) Sending gratuitous ARPs on eno16777736 for 172.16.1.66

从上面的实验来看，一旦主的上面的服务挂了之后，keepalived的优先级降低，这样会促成新的keepalived选举，从而，主节点发生变化，这样的话，从的上面的几点就能对外响应web服务了，从而达到了高可用web服务器；

那同样的道理，这个方案也适合其他服务，keepalived在响应的过程中消耗资源非常小，适合中等并发量的高可用服务

当然以上的后续步骤是在状态监测完毕之后，可以执行脚本，从而完成更多的自动化任务：

通知脚本示例：（每个节点都有一份儿）

~】# vim /etc/keepalived/notify.sh

#!/bin/bash

#指定接收脚本的用户root以及地址

contact=‘root@localhost‘

notify() {

mailsubject="$(hostname) to be $1, vip floating."

mailbody="$(date +‘%F %T‘): vrrp transition, $(hostname) changed to be $1"

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

}

case $1 in

master)

notify master

;;

backup)

notify backup

systemctlstart httpd.service

;;

fault)

notify fault

;;

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

exit 1

;;

esac

脚本调用方法示例：

在vrrp_instance VI_1 { 。。} 上下文里面写入一下三行，这里是在虚拟路由配置项里面写的，如果有多个虚拟路由配置段，这个就要写多个；且每台keepalived服务器也都要写上，这样就可以以邮件的方式知道节点之间的状态信息变化了。

~】# vim /etc/keepalived/keepalived.conf

vrrp_instance VI_2 {

......

virtual_ipaddress {

172.16.1.44

}

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

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

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

}

上面这个脚本的和在keepalived的相应配置所达到的效果：

一旦主节点上面的httpd服务宕机，就会使主节点的优先级下降，主节点变成back从节点；脚本定义好了一旦参数为backup，就会启动httpd服务，启动httpd服务之后，节点的优先级回复，这样又变成主节点；所以主节点永远是主的了，（当然这里所有的配置为抢占式）；如果把keepalived配置成为非抢占式，那主从节点就有可能轮流坐庄，但还是默认主节点会响应更多的服务，所以如果要使用这个脚本配置，配置成非抢占式

非抢占模式为 nopreempt 如

vrrp_instance VI_2 {

......

virtual_ipaddress {

172.16.1.44

}

nopreempt

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

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

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

}

本文出自 “北极的linux” 博客，请务必保留此出处http://941012521.blog.51cto.com/9253690/1841141