Subnet Mask "Slippage"

Hi,
Depending on how often your DHCP is setup to check on connections and how your effected PC reply to it (Network Card Traffic Config) it could show as if these PCs are slower than the rest is replying back where the DHCP will think they belong to another subnet.
Also, are the one effected similar in anything else, like on same hub, switch ??

Hope this helps.

George

Hey jb. Need more info.

What OSs are the client machines?

What scope options is the DHCP server set to push out to clients?
Is the DHCP server on a different subnet than the clients? If so, does it have a different subnet mask?

This is very odd behavior!

What IP address are you using? is it 169.x.x.x? The only thing I can think of (very far fetched) is that your lease is expiring and your clients can not find a DHCP server so they drop to APIPA mode and pick up a 169.x.x.x address.

I know...but I had to ask!

Another far fetched alternetive could be that you have a rouge DHCP sever (non-2000) handing out a bad scope.

As you probably know, the DHCP cliet requests an update halfway through the lease period. If the client is on the otherside of a routerand you are using DHCP Relay or have an RFC 1542 capable router, then you might look around that segment for a rogue DHCP server that would respond quicker then the authorized one.

Not an answer as such, but is there any relationship between the IP addresses of the machines that are slipping (like they are all in a small range of IPs ?), are they on a particular hub/switch. Are you blocking any possible DCHP requests getting outside of your subnet (thats unlikely to be the problem if you say that the leases are not expiring)
I further question, if you renew the lease using winipcfg or ipconfig, do the masks get back to full 24 bit C class or remain as an effective (or should that read ineffective) A class. ?

A.C

JB:

This is an off of the wall question. Has your system been through a lightning strike or power surge? Not your users, but your server (room). If so, check the NIC's. I had a similar problem, but more intense. My users were loosing ther complete IP addresses. I found that the backup power supply had been hit in my server room. Now, this did not affect all users. Only users that had Intel NIC's were affected. The 3Com users were fine. This screwed with me until I had a weekend alone, with a bottle of jack ;), and tore into the whole damn thing. Check your NICS.