Possible causes of duplicate reset apdater (down/up) on the physical computer + Oracle VM virtualbox + centos7 + NAT Nic; virtualboxcentos7

Possible causes of duplicate reset apdater (down/up) on the physical computer + Oracle VM virtualbox + centos7 + NAT Nic; virtualboxcentos7

In the past few days, when centos is running on the vm box, an abnormal and regular Nic suddenly occurs. Normally, the reset is disabled for 1 or 2 minutes and the reset is normal and repetitive,

Looking at this phenomenon, we can recall several operations that may be relevant before the phenomenon occurs:

1. Use yum update to upgrade centos kernel: tested and irrelevant to Kernel

2. Why can't the driver be unable to cope with the large amount of internal communication data in the centos system? Shut down all non-systematic services;

3. If the driver is unstable, switching to another NIC will not work;

4. It is suspected that the firewall of the physical system has stopped and the anti-virus service has been disabled. The firewall still does not work;

5. Check/var/log/message in one row. no useful information is displayed. At most, the information about the NIC status change and time point is obtained;

6. I ran to the centos forum and asked bj in English. One person said it was caused by the link-flap protection action of the route, but I disconnected all wifi/lan networks, I think the frequency of li is not high. I think it should be because some software detects that the network is disconnected, and the delay occurs only when the message is sent;

7. The vm box is still being reinstalled;

8. Converting NAT into the global NAT network interface form and bridge interface solves this problem. Therefore, I think the problem should be caused by the NAT network address translation process,

Although the scope is found, other network connection methods can solve this problem, but so far, if you must use this NAT method, the following problems are still not solved:

1 Suppose there is a router flap like a routing protection mechanism, and something is sending this signal, so it is unclear who sent it and how to troubleshoot it;

2. If the signal source is found but cannot be stopped, what if we ignore this signal in centos and do not reset the Nic?

3. If this signal exists, will the reset operator perform operations related to the NIC of centos (such as the network administrator) or vm box when the signal is received?

The one you know is welcome to reply.

The English language is very poor. I have a smile and it is good to solve the problem. See the link for some testing data.

Https://www.centos.org/forums/viewtopic.php? F = 50 & t = 51112