We have been having a weird set of issues come up recently. We have several subnets that we are getting reports of IP address conflicts errors from end users. We are assigning addresses via DHCP. The addresses (hardware, and IP) that are being listed as being in conflict are usually not the same. I initialy thought this could be a dhcp server issue, so I changed one subnet with issues to a different DHCP server with a differant scope of addresses. No change the issue still exists. Then errors started occuring on a subnet in a remote location, running a totally differant setup for dhcp(The main site uses Microsoft’s DHCP Server propagating dhcp traffic across a Cisco 6509 Switch’s subnets with helper addresses setup, and the remote site with a single subnet with a Cisco 4006 switch, and running DHCP Server on the 1750 Cisco Router at that location) Finally I found one lead but am not sure what it means. We just got in a new Fluke Optiview Network analyser, and I put it on each subnet with issues. The Optiview shows that on each subnnet about 6 PC’s have the wrong subnet mask, It should be a class C subnet mask but is coming up as class A or B. I verified that these PC’s were not set with static IP addresses, and if you release andrenew the ip address manually the correct mask appears, but within a couple days the incorect mask reappears. The only thing in common with these computers seems to be that they are running Windows 98, they are all of several differant brands, and also are running several differant brands of Network cards (3Com, integrated Dell, and Aopen) My thoughts are that since the subnet masks are not right the dhcp servers cannot tell that the IP addresses they are assigning are in use and they reassign the same address. Any ideas? Am I even on the right track here?
Jim Matuska
