Keepalived实践

keepalived集群高可用部署参考

需求描述：某服务A（后面都用A来表示该服务）需要高可用服务需求。当主服务故障时，需要切换到备服务上。目前为一主一备，后续为一主多备

需求提炼：

部署keepalived服务进行健康检查、故障切换
多备情况下，根据vrrp选举特性，备机权重值、state均设置为一致
编写脚本对进行服务进行定时检查，并动态调整keepalived服务启停。
考虑到频繁启停可能影响正常服务，设置keepalived集群节点均为非抢占模式

安装keepalived服务

下载keepalived源码安装包

wget https://www.keepalived.org/software/keepalived-2.0.10.tar.gz

解压安装

tar zxvf keepalived-2.0.10.tar.gz
cd keepalived-2.0.10
make && make install
#make的时候可能会遇到依赖问题，需要先解决依赖包问题再进行编译
#检查是否安装成功
keepalived --version

修改配置

! Configuration File for keepalived
 global_defs {
    notification_email {
      acassen@firewall.loc
      failover@firewall.loc
      sysadmin@firewall.loc
    }
    notification_email_from Alexandre.Cassen@firewall.loc
    smtp_server 192.168.200.1
    smtp_connect_timeout 30
    #路由id，需要保证集群内id唯一
    router_id LVS_DEVEL_2
 }
 vrrp_instance VI_1 {
     #初始状态
     state BACKUP
     #vip关联接口
     interface eth0
     #虚拟路由局域网id，需要保证为局域网内唯一
     virtual_router_id 131
     #选举权重值
     priority 100
     #非抢占模式
     nopreempt
     #心跳周期
     advert_int 1
     #集群鉴权信息
     authentication {
         auth_type PASS
         auth_pass 1993
     }
     #vip配置，可以设置多个
     virtual_ipaddress {
         10.25.74.124
     }
 }

启停与状态检查

#启停
systemctl start|stop|restart keepalived
#状态检查：1.查看ip地址
ip addr | grep eth0
#状态检查：2.查看日志
tail -n 200 /var/log/message | grep Keepalived

检查脚本

把下面脚本分别部署到keepalived各节点上

vim /etc/keepalived/check_port.sh
#!/bin/bash
#shell script for check service alive
#author: joe
#um:xianweijian323

while true;do
    count_9999=`netstat -lnpt | grep 9999| wc -l`
    count_kpa=`ps -ef | grep keepalived| grep -v grep | wc -l`
    if [ $count_9999 -gt 0 -a $count_kpa -eq 0 ];then
        systemctl start keepalived
        echo "`date +'%Y-%m-%d %H:%M:%S'` [WARN] 9999’s service is alive,but Keepalived service is dead.startup Keepalived serivice"
    elif [ $count_9999 -eq 0 -a $count_kpa -gt 0 ];then
        systemctl stop keepalived
        echo "`date +'%Y-%m-%d %H:%M:%S'` [WARN] 9999’s service is dead,but Keepalived service is active.shutdown Keepalived serivice"
#    else
#        echo "`date +'%Y-%m-%d %H:%M:%S'` [INFO] all serivce is actived.Nothing to do"
    fi
    sleep 1
done

nohup bash check_port.sh > ./check_port.log 2>&1 &