keepalived健康检查问题

Question

使用tcpdump监听，发现drector不断向real server发送会话请求，连接就断开。我知道这个是健康检查，但为什么同样配置的另一个集群没有这个问题，下面是配置文件

global_defs {
   notification_email {
     root@dd.com
   }
   notification_email_from root@SS_lvs
   smtp_server 192.168.14.16
   smtp_connect_timeout 30
   router_id LVS_DEVEL
}

vrrp_instance LVS {
    state MASTER
    interface eth0
    virtual_router_id 52
    priority 100
    advert_int 2
    authentication {
        auth_type PASS
        auth_pass 1111
    }
    virtual_ipaddress {
        192.168.14.23
    }
}

virtual_server 192.168.14.23 5001
{
delay_loop 3
lb_algo wrr
lb_kind DR
persistence_timeout 120
protocol TCP
 
real_server 192.168.14.16 5001 {
        weight 1
        TCP_CHECK {
                connect_port 5001
                connect_timeout 10
                nb_get_retry 3
                delay_before_retry 3
        }
}
}

virtual_server 192.168.14.23 5002
{
delay_loop 3
lb_algo wrr
lb_kind DR
persistence_timeout 120
protocol TCP
 
real_server 192.168.14.16 5002 {
        weight 1
        TCP_CHECK {
                connect_port 5002
                connect_timeout 10
                nb_get_retry 3
                delay_before_retry 3
        }
}
}

virtual_server 192.168.14.23 5008
{
delay_loop 3
lb_algo wrr
lb_kind DR
persistence_timeout 120
protocol TCP
 
real_server 192.168.14.16 5008 {
        weight 1
        TCP_CHECK {
                connect_port 5008
                connect_timeout 10
                nb_get_retry 3
                delay_before_retry 3
        }
}
}

Answer 1

哎，自己来解答吧。有健康检查才是正确的，另一套集群的配置文件中TCP_CHECK {之间缺少空格，导致配置有问题，万恶的空格啊。