I'm running Red Hat 6.2 on an IBM server with 4 Ethernet interfaces on the same subnet. Each interface has been given its own IP address and all the interfaces have been assigned the same gateway and netmask. The driver being used is eepro100.
I have heard that the Linux O/S has the ability to detect problems on an Ethernet interface and send an ARP to the network assigning the troubled interface's IP address to the MAC address of a functional interface. I have witnessed this from the network view point, seeing the same MAC address assigned to 2 or more IP addresses of the server.
I'm currently experiencing slow response on varying interfaces. It may be one interface, then the server is rebooted and the slow response switches to two other interfaces. My network folks are pointing at the server because they are seeing the 'IPs sharing one mac address' situation I described above. Yet I have two other servers on different networks configured in the same manner showing this behavior and functioning fine.
I have looked and looked for documentation explaining this but have not found anything.
Can anyone give more detail on this behavior?
Thanks for your help!
Bill Hahnel
