We are running W2K server as the site RRAS server. It is setup to use a static pool of addresses. Dial in clients are configured to use the IP from this pool and to use this as the default gateway. when client PC's dial in they are given the correct IP from the static pool, but they have an incorrect subnet mask. Even though the subnet is wrong, sometimes they can still connect and ping other network resources. I have checked the setting on both the client and server. I can't see anything incorrect. I can't seem to discover what is wrong, and how to correct it. The RRAS server IP and subnet settings are correct. This has only started after we changed from using NT4.0 as the RRAS server. Any ideas? Thanks in advance

Yes. I do have an idea. What is probably happening may be expected behavior. The RFC (rules basically) that define IP address allocation across a PPP (dial in) connection only allow IP address info to be send across the connection. Let me take a guess here. The IP address is handed out from the pool but the subnet mask is a classful mask. In other words, if they obtain a class A address, they get a class A mask, regardless if you defined a subnetted mask? Is this right? I don't know how familiar you are with classes etc but if the first number,(octet) in the IP address is between 1 - 126, they will get a mask of 255.0.0.0. If the first octet is between 128-191 they will get a mask of 255.255.0.0, and if the first octet is between 192-223, they get a mask of 255.255.255.0. On rare occassions on W2k machines, the mask may be set to 255.255.255.255. Also, you may see the default gateway set to the same IP as the client computer. This is expected behavior. Am I close on this? \

Glen, Yes, you are correct - we have a class B subnet. The subnet assigned to RRAS clients is 255.255.0.0 Clients are configured to use default gateway on network. This would make me think that there is another issue preventing dial-in clients from sometimes accessing resources. Thanks for your explanation. Cheers, Peter.