Hi all

Cant seem to get an FTP server working...

heres the setup

Cable Modem > Belkin Router > PC......FTP Client....(WSFTP Server)

Ok so I port forwarded port 20 and 21 in the Belkin Router setup as per portforwarding and Belkins own tech page...

The address of the PC is 192.168.2.10, if I enter this as an FTP address I can reach it and login OK

If I enter the outward facing IP address (the one the router is assigned by the cable modem) I cant connect,

It seems that even though the router is told to forward port 20-21 to IP 192.168.2.10 on the internal network, it wont, and connections are refused.

I have double checked the setup a few times, but it still wont let me FTP to the IP address assigned by the cable modem to the router.....

Any ideas, yes the router has the latest firmware, and I know if there is no suggestions then its a case of remove the router, I cant do that yet its in use by the family, and see if it connects, that is the next step, but i'm decorating at the moment so it will have to wait.

Even if I prove the router to be at fault, by connecting straight to the cable modem, what can I do with the router, will Belkin take it back or will I have to buy a new one...

Regards

wierd! I've done this many times and it does really work as you intend.
the only thing I can think of is (on my netgear router) that you can
DEFINE the forward without it being ENABLED. do you have an enable/disable
checkbox on the Belkin?

other thought: REBOOT the router

Hi Jobeard

Yep enable is ticked (did think that when it didint work) also the belkin has a list of predefined programs that you can choose from and FTP is in the list, that one didnt work.

Reboot, yep did that, last night as azureus caused it to lock up, still the same.

Looks like i'm going to have to take the router out of the equation and see if it works, if so wonder what Belkin will have to say...

Regards

Some routers do not support loopback connections - you cannot connect to your own public IP.

my my! My Netgear does, thank God!
Why that would be implemented is strange to me, in that it takes extra firmware
code to look for that special case. Sending packets from the router to the ISP gateway takes no extra code and that gateway only performs normal routing to
give it back to you -- the routing need not be performed in the router itself -- ever.

oh well, live and learn (I love tech

Nope, doesn't work that way.

When you are sending packets to an IP on your own subnet, they do not get sent to the gateway. Furthermore, TCP/IP stacks are optimised to not even send out packets that are supposed to loop back. The packet is just forwarded inside the TCP/IP stack to the incoming queue. This is standard and Works (TM) with any internet device.

Now, my guess is that the thing that breaks with loopback, is the NAT functionality.

I thought up a reason in my head once, but since I am no router expert, it is more of an hypothesis that I'm not going to disclose here.

If it were in the TCP stack, then it would be independent of the device and
inherent in ALL Tcp software. As we have different experiences, it clearly is not in all Tcp Stacks

As I said, the problem is not in TCP/IP, but NAT (sitting on top of TCP/IP)

Anyway, some vigorous googling gave me this pretty explanation:
http://nick.recoil.org/articles/2006/08/13/loopback-nat-with-pf

Basically, the NAT engine has to do some extra work to make loopback work.

Gota love good feedback and learning every day thanks