Physical computer +oracle VM virtualbox+centos7+nat nic repeat reset Apdater (down/up) phenomenon possible cause

These days, when running CentOS on the VM box, there is a sudden abnormal regular network card, normal use of 1, 2 minutes and turn off the reset and normal repetition of the phenomenon,

Looking at this phenomenon, recall several actions that may be relevant before a phenomenon occurs:

1 Upgrading the CentOS kernel with Yum Update: tested and kernel-independent

2 suspected of large volume of communication data inside the CentOS system, causing the driver to be overwhelmed? Shut down all non-systematic services, or so;

3 suspect drive instability, switch to another network card or not;

4 Suspected physical system firewall blocked, shut down antivirus, firewall or not;

5/var/log/message a line to see, or there is no useful information hint, the most is to get the network card status changes and time-point information;

6 Run to the CentOS forum using BJ's English asked the next, there is a person said to be routed link-flap protection action caused, but I put all the Wifi/lan network disconnected, the phenomenon is still, just frequency li is not high, I think it should be a software detected network disconnection, delay after the transmission occurs ;

7 by reloading the VM box phenomenon is still in;

8 NAT to the Global NAT network interface form and bridging all solve this phenomenon, so I think the problem should be in the NAT network address translation process,

Although the approximate range has been found, the use of other network connections can solve this problem, but I have so far, if you must use this NAT mode, still does not solve the following problems:

1 Assume that there is a route protection mechanism like router flap, and there is something to send this signal, then who is sent, not clear, do not know how to troubleshoot;

2 Assuming this source is found, but can not stop it, then, if the CentOS setup ignore this signal, do not go to reset network card?

3 Assuming this signal is present, then, when the signal is received, the Reset action performer is the CentOS network card related to the operation (such as network manager), or VM box?

The welcome reply that knows.

English is very poor, laughed at, can solve the problem just fine. See the link, there are some test data at the time

https://www.centos.org/forums/viewtopic.php?f=50&t=51112

Physical computer +oracle VM virtualbox+centos7+nat nic repeat reset Apdater (down/up) phenomenon possible cause