Trying to replace my router with a Debian machine - but I can't understand my existing setup
Hello all,
This is only tangentially Debian related, but I can't really think of anywhere
better to ask, so maybe somebody can help.
I have an ADSL connection set up as follows: D-Link DSL-300G+ modem connected
to the ADSL line - well it calls itself a modem, but it connects to the rest of
our network via ethernet and appears to have both an external and an internal IP
address, which sounds rather more like a router to me, but more on that later. I
don't know what OS it runs; it has telnet access but the environment is
unfamiliar to me, so I've only used the web-based config. Its internal address
is 192.168.0.1.
Connected to this (via ethernet) is a device calling itself "RouteFinder SOHO
VPN Gateway". Once again I don't know what OS this is running, and I've not
found any telnet/ssh access to it. This has an internal IP address 192.168.2.1,
and is connected to the main switch for our intranet (where all machines have IP
addresses in the 192.168.2.0/24 subnet). I'd like to replace this router with a
PC running Lenny.
My problem lies in that I can't understand how the connection between the router
and the modem is actually working. The beginning of my confusion is that these
machines both claim to have the same external IP address. So, the setup looks
like this:
-------------------- -------------------- --------------
--- | Modem | | Router | | Network |
ADSL|---|$extip : 192.168.0.1|---|$extip : 192.168.2.1|---|192.168.2.0/24|
--- -------------------- -------------------- --------------
The modem does have a setting for "RFC 1483 Bridge", but it's disabled. Instead,
it's just set to the standard PPPoA.
The router claims to be getting its IP address, default gateway, and DNS via
DHCP, and it can be set up to do port forwarding, etc. I didn't set the network
up, and it was only when I physically followed along all the cabling that I
realised that the modem even *existed* - since the router appears to be
publically adressable I'd assumed that it included its own modem.
Reasoning that the modem must be somehow forwading the DHCP connection, I had a
short attempt (I don't want to cause too much downtime through experimentation)
to replace the router with the machine I've set up to be its replacement. When
that started up, it got the IP address 192.168.0.2 via DHCP, DNS server
192.168.0.1 (the modem doesn't actually provide DNS though) and no default
gateway. It did occur to me - after I'd replaced the original router - that
possibly I'd made some Shorewall configuration mistake, but it's definitely
configured to allow connections from '$FW' to 'net' - ie. all outgoing
connections from that machine, so I can't really see how there would be a
problem.
With the original setup, traceroute to $extip goes straight there, without any
hops, and traceroute to google.com looks something like this:
Tracing route to google.com [209.85.171.100]
over a maximum of 30 hops:
1 <1 ms <1 ms <1 ms 192.168.2.1
2 520 ms 555 ms 596 ms 62.241.160.185
...
Numerous internet addresses
...
16 172 ms 174 ms 179 ms cg-in-f100.google.com [209.85.171.100]
No mention in there of 192.168.0.1, or $extip.
So in a nutshell, when I'm using the dedicated router, the modem appears to be
completely transparent, but when I'm trying to use the computer replacement, it
acts like a normal router which has been misconfigured.
Does anyone have any idea how the existing setup actually works, or how I can
investigate further? Is there some common setup here which I'm completely
unaware of? Would it be worth temporarily allowing all connections from 'net' to
'$FW' in my Shorewall config and seeing what happens?
I'm quite stuck here so anyone with experience of this would be very welcome.
Thanks,
Nye
Reply to: