Dual Nics Windows Server 2003...

[Deviationer]

Need to pick the collective brains here on something.

Windows Server 2003 SP2 x64
(Does nothing other than FTP, not using IIS)

NIC 1 = Local Lan, connected to switch
IP=10.1.0.X
Subnet=255.255.255.0
Gateway=10.1.0.X
DNS=10.1.0.X

NIC 2 = Internet, no switch/router, direct line from fiber connection
IP=66.XXX.XXX.XXX
Subnet=255.255.255.0
Gateway=66.XXX.XXX.XXX
DNS=216.XXX.XXX.XXX

Does that look right?

 

[criccio]

Can you even use a class C subnet with a class A IP?

 

[Grentz]

Should be no problem on the LAN side. The subnet just determines how many subnets you can see within the IP range. I dont know on the outside IP though.

EX. 255.255.255.0 can see any 10.1.0.X, 255.255.0.0 can see any 10.1.X.X, and so on.

 

[Protoform-X]

Shouldn't your subnet on the WAN side be 255.255.255.255? You are only using one IP, so I don't see why you'd need a /24. I've never set a connection like this, but that just seems curious to me!

 

[Xipher]

the ONLY time you would ever use a /32 address is for loopbacks.

also, DON'T use multiple default gateways in Windows, drop the default gateway on the LAN side.

 

[ShockValue]

You're not on the same subnet. You're looking at the subnet mask, but in reality the range of IPs is determined by the address and the subnet combined.

On the LAN interface you're in the network range of:
10.1.0.0 through 10.1.0.255 (.1 through .254 being valid assignable addresses)

On the WAN interface you're in the network range of:
66.xxx.xxx.0 through 66.xxx.xxx.255 (.1 through .254 being valid assignable addresses)


The whole point of subnetting is to break up large groups (Like a Class A 10.x address) into smaller networks, so yes a 255.255.255.0 mask makes perfect sense in this case.

Xipher is right, remove the LAN gateway.