docker部署keepalived

1、拉取镜像

docker pull osixia/keepalived:2.0.20

2、启动命令

192.168.1.1，192.168.1.2代表的是宿主机的IP，192.168.1.3代表的是虚拟IP。

1.master

docker run --cap-add=NET_ADMIN --cap-add=NET_BROADCAST --cap-add=NET_RAW --net=host -d --name keepalived \
    -v /etc/localtime:/etc/localtime \
    -e KEEPALIVED_INTERFACE='enp3s0' \
    -e KEEPALIVED_PASSWORD='123456' \
    -e KEEPALIVED_STATE='MASTER' \
    -e KEEPALIVED_ROUTER_ID='51' \
    -e KEEPALIVED_UNICAST_PEERS="#PYTHON2BASH:['192.168.1.1','192.168.1.2']" \
    -e KEEPALIVED_VIRTUAL_IPS='192.168.1.3' \
    osixia/keepalived:2.0.20

2.Slave

docker run --cap-add=NET_ADMIN --cap-add=NET_BROADCAST --cap-add=NET_RAW --net=host -d --name keepalived \
    -v /etc/localtime:/etc/localtime \
    -e KEEPALIVED_INTERFACE='enp3s0' \
    -e KEEPALIVED_PASSWORD='123456' \
    -e KEEPALIVED_STATE='BACKUP' \
    -e KEEPALIVED_ROUTER_ID='51' \
    -e KEEPALIVED_UNICAST_PEERS="#PYTHON2BASH:['192.168.1.1','192.168.1.2']" \
    -e KEEPALIVED_VIRTUAL_IPS='192.168.1.3' \
    osixia/keepalived:2.0.20	

3、注意

因为是docker部署的keepalived，所以keepalived里面所写的nginx检测脚本是运行不了的，因为不能从容器去操作宿主机，无法实现nginx的重启与keepalived的docker stop操作。所以采取crontab的定时任务策略去检测，实现VIP飘逸。

4、定时检测脚本

#!/bin/bash

while true
do
    if [ -z "$(lsof -i :80)" ]; then
        /usr/local/nginx/sbin/nginx
    fi

    sleep 5

    if [ -z "$(lsof -i :80)" ]; then
        docker stop keepalived
    fi

    sleep 5

    if [ -z "$(lsof -i :80)" ]; then
        /usr/local/nginx/sbin/nginx
    fi

    sleep 5

    if [ -z "$(lsof -i :80)" ]; then
        docker stop keepalived
    fi

    sleep 5

    if [ -z "$(lsof -i :80)" ]; then
        /usr/local/nginx/sbin/nginx
    fi

    sleep 5

    if [ -z "$(lsof -i :80)" ]; then
        docker stop keepalived
    fi
done

5、定时任务

crontab -e  进入定时任务编写页面

* * * * * check_nginx.sh

表示，每分钟执行一次检测脚本，crontab最小执行单位为分钟，所以check_nginx.sh中每五秒检测一次，写了三次，相当于每分支去检测6次nginx，第1.3.5次检测的是nginx，如果端口不存在则重启nginx，第2.4.6次是重启完了nginx再去检测nginx，nginx还没启动，则docker stop keepalived，实现虚拟IP转移。