This is for you switching gurus out there. I know its not good but I am trying to explain why and I cant come up with the correct words...
Two Unix Solaris boxes with 2 nics each. Nic1 for Prod and Nic 2 for BU on both boxes. On the same switch (3750) and the Prod subnet is x.x.1.0/24 and BU subnet is x.x.2.0/24 for both boxes. SolarisA has both nics set to the same mac adx (:::ABCD) as well as SolarisB(:::1234). (This is a default setting to this type of physical server ( I think its a Fujitsu chassis. And yes, I am forcing the unix team to change it to unique adxs but I just want assistance for a pro-active write up.)
Ive been asked to sniff these servers due to the fact that ftp transfers sometimes stop during its process and they want to be proactive about it and find out why. These ftp transfers are log transfers that are normally very large.
Since it is same switch and same subnet; they are only communicating via Layer 2 and not using any L3 type of communication. Sniffing is not coming up with anything inclusive (Im probably using it wrong due to the fact that the filters are based upon mysql ports and ip adxs.) But, I am trying to explain to the client that having the same mac address on both ports (prod and BU) can cause this problem. I realize that communication will be via L2 but if it goes to the BU port it should drop due to the wrong vlan while if it goes to the PROD port then communication is good. But, will there be any time where it is communicating via the PROD port and then switch to the BU port?
Please ask questions if I need to clarify.