Pre
双机热备是指两台机器都在运行，但并不是两台机器都同时在提供服务。
环境信息
节点 | 软件 | IP |
|---|---|---|
节点A | keepalived tomcat | 192.168.126.137 |
节点B | keepalived tomcat | 192.168.126.138 |
虚拟IP | - | 192.168.126.200 |
官网
https://keepalived.readthedocs.io/en/latest/#
原理
keepalived用于保证环境高可用，防止单点故障。
keepalived主要有三个模块，分别是core、check和VRRP。
- core核心模块，负责主进程的启动、维护以及全局配置文件的加载和解析
- check负责健康检查，包括常见的各种检查方式
- VRRP模块实现VRRP协议
VRRP（Virtual Router Redundancy Protocol） 即虚拟路由冗余协议，实现路由器高可用的协议，即将N台提供相同功能的路由器组成一个路由器组，这个组里面有一个master和多个backup，master上面有一个对外提供服务的vip（该路由器所在局域网内其他机器的默认路由为该vip），master会发组播，当backup收不到VRRP包时就认为master宕掉了，这时就需要根据VRRP的优先级来选举一个backup当master。这样的话就可以保证路由器的高可用了。
keepalived 安装
两个节点均需要安装， 在137 和 138 节点上安装 ，执行 yum install keepalived
137 安装 keepalived
代码语言：javascript复制[root@localhost ~]# yum install keepalived
Loaded plugins: fastestmirror
.......
.......
Complete!
[root@localhost ~]# 138 安装 keepalived
代码语言：javascript复制[root@localhost ~]# yum install keepalived
Loaded plugins: fastestmirror
.......
.......
Complete!
[root@localhost ~]# tomcat 也都装下哈 ，用于验证，可以yum安装
CentOS 7 yum安装与配置 Tomcat 写的挺不错，没有的话可以参考
配置keepalived的配置文件/etc/keepalived/keepalived.conf
关键参数解读
代码语言：javascript复制state 主服务器为MASTER，其他服务器均为BACKUP
interface 通信用的网卡，ip addr可以查看，以实际为准
virtual_router_id 主从服务器此ID必须一致
priority 优先级，主服务器必须大于其他服务器，数值越大优先级越高
authentication 主从服务器必须一致
virtual_ipaddress 虚拟IP地址，主从服务器必须一致，该VIP为客户端访问时使用的IP137
代码语言：javascript复制! Configuration File for keepalived
global_defs {
notification_email {
acassen@firewall.loc
failover@firewall.loc
sysadmin@firewall.loc
}
notification_email_from example@example.com
smtp_server mail.example.com
smtp_connect_timeout 30
router_id LVS_DEVEL
}
vrrp_script chk_tomcat {
script "/etc/keepalived/tomcat_check.sh"
interval 5
weight -5
fall 3
rise 2
}
vrrp_instance VI_1 {
state MASTER
interface ens33
mcast_src_ip 192.168.126.137
virtual_router_id 51
priority 101
advert_int 2
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
192.168.126.200
}
track_script {
chk_tomcat
}
}138
代码语言：javascript复制! Configuration File for keepalived
global_defs {
notification_email {
acassen@firewall.loc
failover@firewall.loc
sysadmin@firewall.loc
}
notification_email_from example@example.com
smtp_server mail.example.com
smtp_connect_timeout 30
router_id LVS_DEVEL
}
vrrp_script chk_tomcat {
script "/etc/keepalived/tomcat_check.sh"
interval 5
weight -5
fall 3
rise 2
}
vrrp_instance VI_1 {
state BACKUP
interface ens33
mcast_src_ip 192.168.126.138
virtual_router_id 51
priority 99
advert_int 2
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
192.168.126.200
}
track_script {
chk_tomcat
}
}与主节点的配置上的区别，需要注意
state 表示一个是主机一个是备机，mcast_src_ip 当前服务器的ip，priority 主机要比备机的要高
检测脚本
代码语言：javascript复制!/bin/bash
JAVA_PROCESS=`ps -C java --no-heading| wc -l`
if [ $JAVA_PROCESS -eq 0 ];then
echo "tomcat is stop"
sleep 2
if [ `ps -C java --no-heading| wc -l` -eq 0 ];then
systemctl stop keepalived.service
fi
fi比较简单，根据实际情况调整
检测是否有tomcat在运行，如果没有的话，等2秒，如果还没有就把keepalived给停止掉，这样vip就能从主机master切换到备机BACKUP上了。
日志
可以查看 /var/log/messages 日志 观察切换和选主的过程。
状态发生变更，触发脚本
https://keepalived.readthedocs.io/en/latest/configuration_synopsis.html
! Configuration File for keepalived
global_defs {
notification_email {
acassen@firewall.loc
failover@firewall.loc
sysadmin@firewall.loc
}
notification_email_from example@example.com
smtp_server mail.example.com
smtp_connect_timeout 30
router_id LVS_DEVEL
}
vrrp_script chk_tomcat {
script "/etc/keepalived/tomcat_check.sh"
interval 5
weight -5
fall 3
rise 2
}
vrrp_instance VI_1 {
state MASTER
interface ens33
mcast_src_ip 192.168.126.137
virtual_router_id 51
priority 101
advert_int 2
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
192.168.126.200
}
track_script {
chk_tomcat
}
notify_master /etc/keepalived/a.sh
notify_backup /etc/keepalived/b.sh 当节点从BACKUP 转为 MASTER的时候 ，执行 /etc/keepalived/a.sh
当节点从MASTER 转为 BACKUP的时候 ，执行 /etc/keepalived/b.sh
小结
MASTER和 BACKUP，预先分配好， 通过priority 来确定， 当MASTER宕机以后，BACKUP 没有收到 MASTER的 VRRP消息，BACKUP就认为 MASTER宕机了，这个时候就把自己选举为MASTER。
MASTER宕机，本节点的日志
代码语言：javascript复制Dec 30 01:40:01 localhost Keepalived[6784]: Stopping
Dec 30 01:40:01 localhost systemd: Stopping LVS and VRRP High Availability Monitor...
Dec 30 01:40:01 localhost Keepalived_vrrp[6786]: VRRP_Instance(VI_1) sent 0 priority
Dec 30 01:40:01 localhost Keepalived_vrrp[6786]: VRRP_Instance(VI_1) removing protocol VIPs.
Dec 30 01:40:01 localhost Keepalived_healthcheckers[6785]: Stopped
Dec 30 01:40:02 localhost Keepalived_vrrp[6786]: Stopped
Dec 30 01:40:02 localhost Keepalived[6784]: Stopped Keepalived v1.3.5 (03/19,2017), git commit v1.3.5-6-g6fa32f2
Dec 30 01:40:02 localhost systemd: Stopped LVS and VRRP High Availability Monitor.BACKUP节点的日志
代码语言：javascript复制Dec 30 01:40:05 localhost Keepalived_vrrp[9173]: VRRP_Instance(VI_1) Entering MASTER STATE
Dec 30 01:40:05 localhost Keepalived_vrrp[9173]: VRRP_Instance(VI_1) setting protocol VIPs.当MASTER恢复以后，MASTER发送VRRP消息给BACKUP，这个时候虽然BACKUP是MASTER，但是 比较Prority后，发现自己的Prority小，会进入BACKUP模式，这个时候，之前设置的MASTER就又变成了MASTER了 。
MASTER恢复，本节点的日志
代码语言：javascript复制Dec 30 01:42:09 localhost Keepalived[6915]: Starting Healthcheck child process, pid=6916
Dec 30 01:42:09 localhost Keepalived[6915]: Starting VRRP child process, pid=6917
Dec 30 01:42:09 localhost Keepalived_healthcheckers[6916]: Opening file '/etc/keepalived/keepalived.conf'.
Dec 30 01:42:09 localhost systemd: Started LVS and VRRP High Availability Monitor.
Dec 30 01:42:09 localhost Keepalived_vrrp[6917]: Registering Kernel netlink reflector
Dec 30 01:42:09 localhost Keepalived_vrrp[6917]: Registering Kernel netlink command channel
Dec 30 01:42:09 localhost Keepalived_vrrp[6917]: Registering gratuitous ARP shared channel
Dec 30 01:42:09 localhost Keepalived_vrrp[6917]: Opening file '/etc/keepalived/keepalived.conf'.
Dec 30 01:42:09 localhost Keepalived_vrrp[6917]: WARNING - default user 'keepalived_script' for script execution does not exist - please create.
Dec 30 01:42:09 localhost Keepalived_vrrp[6917]: SECURITY VIOLATION - scripts are being executed but script_security not enabled.
Dec 30 01:42:09 localhost Keepalived_vrrp[6917]: VRRP_Instance(VI_1) removing protocol VIPs.
Dec 30 01:42:09 localhost Keepalived_vrrp[6917]: Using LinkWatch kernel netlink reflector...
Dec 30 01:42:09 localhost Keepalived_vrrp[6917]: VRRP sockpool: [ifindex(2), proto(112), unicast(0), fd(10,11)]
Dec 30 01:42:09 localhost Keepalived_vrrp[6917]: VRRP_Script(chk_tomcat) succeeded
Dec 30 01:42:10 localhost Keepalived_vrrp[6917]: VRRP_Instance(VI_1) Transition to MASTER STATE
Dec 30 01:42:12 localhost Keepalived_vrrp[6917]: VRRP_Instance(VI_1) Entering MASTER STATE
Dec 30 01:42:12 localhost Keepalived_vrrp[6917]: VRRP_Instance(VI_1) setting protocol VIPs.
Dec 30 01:42:12 localhost Keepalived_vrrp[6917]: Sending gratuitous ARP on ens33 for 192.168.126.200BACKUP节点的日志
代码语言：javascript复制Dec 30 01:42:09 localhost Keepalived_vrrp[9173]: /etc/keepalived/tomcat_check.sh exited with status 127
Dec 30 01:42:12 localhost Keepalived_vrrp[9173]: VRRP_Instance(VI_1) Received advert with higher priority 101, ours 94
Dec 30 01:42:12 localhost Keepalived_vrrp[9173]: VRRP_Instance(VI_1) Entering BACKUP STATE
Dec 30 01:42:12 localhost Keepalived_vrrp[9173]: VRRP_Instance(VI_1) removing protocol VIPs.Questions
exited due to signal 15
/var/log/messages 日志提示：/etc/keepalived/check_nginx.sh exited due to signal 15
调整如下参数
代码语言：javascript复制 interval 5 #监控脚本的执行时间要大于advert_int
advert_int 2 #心跳检查间隔时间notify_master notify_backup未生效
检查 getenforce
getenforce 如果 开启的话，关闭 执行 setenforce 0
