Keepalived安装及配置实现虚拟IP与Namenode HA中Active地址端口的代理

一、背景

　　由于NameNode做了HA，因此需要使用Keepalived将NameNode的8020/9820端口代理出去，比如代理一个虚拟ip+端口

　　使用说明：如果在192.168.66.21、192.168.66.22节点安装了Keepalived，并创建了虚拟IP192.168.66.130和虚拟端口29870、28020，则在192.168.66.21、192.168.66.22节点上，若standby节点为192.168.66.22，则此192.168.66.22主机将无法telnet通192.168.66.130:29870、192.168.66.130:28020

二、下载

　　下载keepalived-2.2.4.tar.gz，并上传至服务器

　　https://www.keepalived.org/download.html

三、安装

1、解压

tar -zxvf keepalived-2.2.4.tar.gz -C /opt/module/

2、安装依赖包

yum install -y gcc openssl-devel popt-devel

3、编译安装

cd /opt/module/keepalived-2.2.4/
./configure --prefix=/usr/local/keepalived
make && make install

4、复制配置

cp /opt/module/keepalived-2.2.4/keepalived/etc/init.d/keepalived /etc/init.d/
mkdir -p /etc/keepalived
cp /usr/local/keepalived/etc/keepalived/keepalived.conf /etc/keepalived/
cp /opt/module/keepalived-2.2.4/keepalived/etc/sysconfig/keepalived /etc/sysconfig/
cp /usr/local/keepalived/sbin/keepalived /usr/bin/

5、主节点操作

（1）检测active节点

vim /etc/keepalived/chkHealth.sh

#!/bin/bash
status=`curl -HGET http://192.168.66.21:9870/jmx?qry=Hadoop:service=NameNode,name=NameNodeStatus`
active="active"
result=$(echo $status | grep "${active}")

if [ "$result" != "" ];then
echo "result=$result"
exit 0
else
exit 1
fi

　　最后赋予权限

chmod +x /etc/keepalived/chkHealth.sh

（2）keepalived.conf配置

vim /etc/keepalived/keepalived.conf

! Configuration File for keepalived

global_defs {
   notification_email {   }
   router_id HA_NETDISK
   script_user root
}

vrrp_script chk_ha_service {
   script "/etc/keepalived/chkHealth.sh"
   interval 3
}

vrrp_instance VI_1 {
    state MASTER
    interface ens33
    virtual_router_id 79
    priority 200
    advert_int 1
    authentication {
        auth_type PASS
        auth_pass 1111
    }
    virtual_ipaddress {
        192.168.66.130
    }
    track_script {
       chk_ha_service
    }
}

virtual_server 192.168.66.130 29820 {
    delay_loop 6
    lb_algo rr
    lb_kind NAT
    persistence_timeout 50
    persistence_timeout 1
    protocol TCP

    real_server 192.168.66.21 9820 {
         weight 1
         TCP_CHECK {
         connect_timeout 1
         nb_get_retry 3
         delay_before_retry 3
         connect_port 9820
         }
    }
}

virtual_server 192.168.66.130 29870 {
    delay_loop 6
    lb_algo rr
    lb_kind NAT
    persistence_timeout 50
    persistence_timeout 1
    protocol TCP

    real_server 192.168.66.21 9870 {
         weight 1
         TCP_CHECK {
         connect_timeout 1
         nb_get_retry 3
         delay_before_retry 3
         connect_port 9870
         }
    }
}

Keepalived.conf

　　说明：大致的思路是创建了一个192.168.66.130的虚拟IP，当建立TCP连接之后，访问虚拟IP的29820端口时，Keepalived会先调用chkHealth.sh脚本进行判断，若主节点192.168.66.21的NameNode处于active状态，则将请求转发至192.168.66.21的9820端口。同理，访问虚拟IP的29870端口时，Keepalived会先调用chkHealth.sh脚本进行判断，若主节点192.168.66.21的NameNode处于active状态，则将请求转发至192.168.66.21的9870端口

　　值得注意的是interface ens33中的ens33表示网卡名称，可以使用ifconfig查看你自己的网卡名称

　　完成之后赋予权限：

chmod 644 /etc/keepalived/keepalived.conf

（3）启停

service keepalived start
service keepalived stop
service keepalived status
service keepalived restart
ps -ef | grep keepalived

（4）日志路径说明：/var/log/messages

6、从节点操作

（1）检测active节点

vim /etc/keepalived/chkHealth.sh

#!/bin/bash
status=`curl -HGET http://192.168.66.22:9870/jmx?qry=Hadoop:service=NameNode,name=NameNodeStatus`
active="active"
result=$(echo $status | grep "${active}")

if [ "$result" != "" ];then
echo "result=$result"
exit 0
else
exit 1
fi

　　最后赋予权限

chmod +x /etc/keepalived/chkHealth.sh

（2）keepalived.conf配置

vim /etc/keepalived/keepalived.conf

! Configuration File for keepalived

global_defs {
   notification_email {   }
   router_id HA_NETDISK
   script_user root
}

vrrp_script chk_ha_service {
    script "/etc/keepalived/chkHealth.sh"
    interval 3
    weight -10
    fall 2
    rise 1
}

vrrp_instance VI_1 {
    state BACKUP
    interface ens33
    virtual_router_id 79
    priority 195
    advert_int 1
    authentication {
        auth_type PASS
        auth_pass 1111
    }
    virtual_ipaddress {
        192.168.66.130
    }
    track_script {
       chk_ha_service
    }
}

virtual_server 192.168.66.130 29820 {
    delay_loop 6
    lb_algo rr
    lb_kind NAT
    persistence_timeout 50
    persistence_timeout 1
    protocol TCP

    real_server 192.168.66.22 9820 {
         weight 1
         TCP_CHECK {
         connect_timeout 1
         nb_get_retry 3
         delay_before_retry 3
         connect_port 9820
         }
    }
}

virtual_server 192.168.66.130 29870 {
    delay_loop 6
    lb_algo rr
    lb_kind NAT
    persistence_timeout 50
    persistence_timeout 1
    protocol TCP

    real_server 192.168.66.22 9870 {
         weight 1
         TCP_CHECK {
         connect_timeout 1
         nb_get_retry 3
         delay_before_retry 3
         connect_port 9870
         }
    }
}

View Code

　　说明1：大致的思路是创建了一个192.168.66.130的虚拟IP，当建立TCP连接之后，访问虚拟IP的29820端口时，Keepalived会先调用chkHealth.sh脚本进行判断，若主节点192.168.66.22的NameNode处于active状态，则将请求转发至192.168.66.22的9820端口。同理，访问虚拟IP的29870端口时，Keepalived会先调用chkHealth.sh脚本进行判断，若主节点192.168.66.22的NameNode处于active状态，则将请求转发至192.168.66.22的9870端口

　　说明2：interface指网卡名称，可以使用ifconfig查看网卡名称

　　完成之后赋予权限：

chmod 644 /etc/keepalived/keepalived.conf

（3）启停

service keepalived start
service keepalived stop
service keepalived status
service keepalived restart
ps -ef | grep keepalived
#开机自启
systemctl enable keepalived.service

（4）日志路径说明：/var/log/messages

7、测试

（1）将keepalived启动起来

（2）访问虚拟IP+29870端口，发现处于激活状态的是slave3节点

　　http://192.168.66.130:29870/