keepalived诞生的目的是为了给ipvs提供高可用性的.
keppalived 服务一般会启动两个进程，一个是vrrp服务和后端服务通信的，一个是checker服务，检测后端real Server健康状况.
邮件服务器：
rhel5:sendmail
rhel6:postfix
keepalived最新版本1.3.5,keepalived配置文件共三部分组成.
global_defs { #全局配置
notification_email { #收件人
[email protected]
}
notification_email_from [email protected] 发件人
smtp_server 127.0.01 #发件服务器
smtp_connect_timeout 30 #超时时间
router_id nginx_slave #路由表示，自定义
}
vrrp_script chk_port { #脚本检测名称chk_port
script "/etc/keepalived/keepalived.jk2.sh" #脚本的路径
interval 2 #每隔2秒检测一次
weight -2 #一旦失败，权重减2
}
VRRP状态机，初始化(initialize)时,大家都是backup状态，通过选举产生master.收到startup且优先级是255时，直接定义为master，收到startup且优先级小雨255时，直接定义为backup.
vrrp_instance VI_1 { #定义虚拟路由和虚拟ip的.VI_1为名称.
state MASTER
interface eth0
virtual_router_id 51 #虚拟路由id，一般不大于255
priority 100 #初始优先级100,值越大优先级越高.
advert_int 1
authentication {
auth_type PASS #认证机制，明文认证
auth_pass 1111 #密码
}
virtual_ipaddress { #虚拟vip地址
192.168.30.129
}
track_script { #虚拟路由跟踪脚本.
chk_port
}
}
其他脚本定义使用
vrrp_script chk_file {
script "[[-f /etc/keepalived/down]] && exit 1 || exit 0"
interval 1
weight -2
}
实例：
系统Centos 6.5
2个node节点：
VIp:192.168.30.131
real server:192.168.30.129
real server:192.168.30.130
#两台real server 操作：
yum install nginx keepalived -y
#这两台real server 先配置好Nginx，做static server.
[[email protected] keepalived]# cat /etc/nginx/conf.d/admin.conf
#
# The default server
#
server {
listen 80 default_server;
server_name _;
# Load configuration files for the default server block.
include /etc/nginx/default.d/*.conf;
location / {
root /data/www/;
index.htm index index.html index.php;
}
}
以示区别/data/www/index.html 亦两台real server 静态页面取ip地址最后1位.
#配置keepalived：
Real Server 192.168.30.129 的keepalived配置文件：
[[email protected] keepalived]# cat keepalived.conf! Configuration File for keepalivedglobal_defs { notification_email {[email protected] #告警通知 } notification_email_from [email protected] smtp_server 127.0.0.1 smtp_connect_timeout 30 router_id nginx_slave#名字可以随便起，标示。}vrrp_script chk_port {#检测脚本 script "/etc/keepalived/keepalived.jk2.sh" interval 2#每个2秒运行一次 weight -2#失败，本机keepalived优先级减2}vrrp_instance VI_1 { state BACKUP#初始化此节点为backup interface eth0#网卡eth0 virtual_router_id 51#虚拟路由id priority 100#优先级，两台优先级可以是一样的，也可以一个高一个低. advert_int 1 authentication { auth_type PASS auth_pass 1111 } virtual_ipaddress { 192.168.30.131 }track_script {chk_port}}
Real Server 192.168.30.130的配置文件：
[[email protected] keepalived]# cat keepalived.conf ! Configuration File for keepalivedglobal_defs { notification_email {[email protected] } notification_email_from [email protected] smtp_server 127.0.0.1 smtp_connect_timeout 30 router_id nginx_master标示这是nginx master}vrrp_script chk_port { script "/etc/keepalived/keepalived.jk2.sh" interval 3 weight -2}vrrp_instance VI_1 { state MASTER #初始化状态为master,两台real server都可以初始化为BACKUP状态,让它们之间自己选举. interface eth0 virtual_router_id 51 priority 101#优先级高于从节点 advert_int 1 authentication { auth_type PASS auth_pass 1111 } virtual_ipaddress { 192.168.30.131 }track_script {chk_port}}
检测脚本路径和内容，赋予脚本可执行权限chmod a+x ...：
[[email protected] keepalived]# pwd/etc/keepalived[[email protected] keepalived]# cat keepalived.jk2.sh #!/bin/bashps -C nginxif [[ $? -eq 0 ]];then exit 0else /etc/init.d/nginx restart > /dev/null sleep 3 ps -C nginx if [[ $? -eq 0 ]];then exit 0 else exit 1 fifi
#此脚本主要判断本地nginx服务如果down 尝试启动1此，还是down就认为本节点下线，vip自动飘值bakcup节点.
#两台real server 启动keepalived服务：
# /etc/init.d/keepalived start
查看keepalived 的log：
[[email protected] conf.d]# tail -f /var/log/messagesAug 4 14:56:09 haproxy Keepalived[51515]: Starting VRRP child process, pid=51518Aug 4 14:56:09 haproxy Keepalived_vrrp[51518]: Netlink reflector reports IP 192.168.30.130 addedAug 4 14:56:09 haproxy Keepalived_healthcheckers[51517]: Netlink reflector reports IP 192.168.30.130 addedAug 4 14:56:09 haproxy Keepalived_healthcheckers[51517]: Netlink reflector reports IP fe80::20c:29ff:feca:1ae addedAug 4 14:56:09 haproxy Keepalived_healthcheckers[51517]: Registering Kernel netlink reflectorAug 4 14:56:09 haproxy Keepalived_healthcheckers[51517]: Registering Kernel netlink command channelAug 4 14:56:09 haproxy Keepalived_vrrp[51518]: Netlink reflector reports IP fe80::20c:29ff:feca:1ae addedAug 4 14:56:09 haproxy Keepalived_vrrp[51518]: Registering Kernel netlink reflectorAug 4 14:56:09 haproxy Keepalived_vrrp[51518]: Registering Kernel netlink command channelAug 4 14:56:09 haproxy Keepalived_vrrp[51518]: Registering gratuitous ARP shared channel
#查看master的vip地址,ifconfig 看不到，用ip a或者：ip addr show
[[email protected] conf.d]# ip a1: 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 forever2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000 link/ether 00:0c:29:ca:01:ae brd ff:ff:ff:ff:ff:ff inet 192.168.30.130/24 brd 192.168.30.255 scope global eth0 inet 192.168.30.131/32 scope global eth0 inet6 fe80::20c:29ff:feca:1ae/64 scope link valid_lft forever preferred_lft forever
#测试：打开浏览器访问http://192.168.30.131/ ，其中一台nginx 启动失败即可看到演示效果.