A network: network card 5 configured to 10.2.13.1, to the end client IP for 10.2.100.1 and 10.2.100.2, main, usually run
B network: the network card be 10.2.113.1 3, to the end client IP 10.2.100.1 and 10.2.100.2, channel, at ordinary times is not running, countless according to
Phenomenon is currently, 10.2.100.1 10.2.13.1 and client is running, if B network fiber cut off, so A netcom (normal,
If B network normal optical fiber, then restart the device, after A short time, no longer communications, crawl Ethernet message display all of the server sends A message, B all the client sends the message
Grab a message command for tcpdump -i eth4 - w ethe - s0, tcpdump -i for eth2 - w ethc - s0
Can consult, Linux such configuration? This is what is going wrong? Operating system or application issue?
The images to grab the following
Ethernet packets cut figure is as follows:
Nic 5
Nic 3
CodePudding user response:
This phenomenon should occur in two different network segment address, network adapter configuration is put on the same floor,13 and 113 if it is a separate routing, so it should belong to 113 segments should not appear on the second floor outside package sent to paragraph 13,
CodePudding user response:
The simplest is to solve the problem in routing to separate two segment really, don't disorderly plug cable between the 2 paragraph,To the server on a single machine to solve the problem will be very hard, there is no need,
CodePudding user response:
Two routing equally, loose hair, moreover Linux the default a in, only from port a back, to port a go out, come back from port b, don't accept, either do policy routing, either close the Linux routing detection,