can't traceroute to gateway.
John
expect to see my gateway early in the output when using traceroute to
determine where my connection is going. Why would my gateway not
appear?
I have a small internal network, connecting to the Internet through a
linux box running RH 7.2, connected to a cable modem (Cox). I'm using
dhcpc to get my IP address, gateway and DNS server information.
Masquerading using very simple iptables works, most of the time. While
trying to diagnose my trouble, I made some trial pings from the linux
box, then decided to run traceroute. Fascinating.
I still have a lot to learn about my /etc/dhcpc/dhcpcd-eth0.info file,
but some of it makes sense. My gateway is listed as 68.4.72.1, which I
find in my routing table. When I ping 68.4.72.1, I get very quick
response (usec), but here's what traceroute returns:
1 * * 10.73.0.1 (10.73.0.1) 15.560 ms
after which it quits. From what Cox says, 10.73.0.1 is my cable modem's
IP addr.
I also receive a primary and secondary nameserver as part of my lease.
I can ping both nameservers most of the time (not always, but I'm not
sure that's >my< problem). However, if I try to traceroute to them
(68.6.16.30, for example), I get output like this:
1 10.73.0.1 (10.73.0.1) 18.723 ms 13.836 ms 23.210 ms
2 68.4.8.a (68.4.8.a) 10.883 ms 10.925 ms 46.535 ms
3 68.4.15.b (68.4.15.b) 16.407 ms 25.242 ms 40.481 ms
4 rsmtdsrc01-gew0304.rd.oc.cox.net (68.4.15.c) 12.518 ms 24.641 ms
13.637 ms
5 rsmtbbrc01-pos0101.rd.oc.cox.net (68.1.0.d) 12.837 ms 17.754 ms
19.991 ms
6 fed1bbrc01-pos0102.rd.sd.cox.net (68.1.0.e) 21.948 ms 15.075 ms
17.734 ms
7 fed1dsrc01-pos0901.rd.sd.cox.net (68.1.0.f) 16.736 ms * 21.008 ms
This raises a couple of questions for me. Firstly, I don't seem to be
getting anywhere near the DNS address, and secondly, if my gateway is
supposed to be 68.4.72.1, why don't I see it?
Jon
purpose, don't you think?
-Jon
John <w...@cox.net> wrote in message news:3CA01052...@cox.net...
Villy Kruse
Jon <j...@unixgroup.com> wrote:
>Why do you assume your gateway has only 1 IP? That would kinda defeat the
>purpose, don't you think?
>
An ISDN router is often set up with only one IP number and the ppp
link is then a numberless point-to-point link. The internal routing
table in the router then uses the physical link or perhaps the ISDN
telephone number as the next hop address rather than the IP number.
Villy
Jon
configs. Yes, it is possible to configure a router with only 1 IP. I don't
think this is what most cable folks are doing (I don't know how you got the
idea to use ISDN as an analogy in this context), particularly since
presumably a router with only 1 IP would answer with it's only IP in a
traceroute, and not be confusing the poor lad. In fact, I'm not sure why
anyone uses unnumbered interfaces on point-to-point circuits, except that
maybe they are uncomfortable with IP routing (and the stupid vendors support
such unconscienable behaviour)- it's not like there's a shortage of RFC1918
space out there, and even if you don't want to break pmutd, you are only
burning 4 IPs per connection, in any case...
-Jon.
Villy Kruse <v...@pharmnl.ohout.pharmapartners.nl> wrote in message
news:slrnaa095...@pharmnl.ohout.pharmapartners.nl...
John
Jon wrote:
> Why do you assume your gateway has only 1 IP? That would kinda defeat the
> purpose, don't you think?
>
> -Jon
>
I'm still pretty much in the dark when it comes to 'real' networks, and I
sorta figured someone would tell me that Cox is using some kinda NAT trick to
make their hardware pretend to be a few physical servers, when in fact there's
a lot more behind the addresses.
What bothers me about the gateway is that if I try to traceroute to it, the
only hop I get is to my cable modem. I'm not even too disturbed by the fact
that acting gateway (68.4.8.85) is not even in the network I was told I was in
(68.4.72.0), and does not appear in my routing table. I don't know how they
handle that, but if it works, ok.
But I started the traceroute stuff trying to figure out why my connection
intermittantly stops working - so far tech support has told me every time that
no one else is reporting problems. And if I stop and restart my network
services, the problem clears up. I think I'm misconfigured in a mostly benign
way, but I don't have enough information yet even to ask an answerable
question about that. As far as busted chops go, I expected to be sent back to
the traceroute man page, which may contain the answer, but not in a form I
recognize.
Karl Heyes
>> Why do you assume your gateway has only 1 IP? That would kinda defeat the
>> purpose, don't you think?
> I'm still pretty much in the dark when it comes to 'real' networks, and I
> sorta figured someone would tell me that Cox is using some kinda NAT trick to
> make their hardware pretend to be a few physical servers, when in fact there's
> a lot more behind the addresses.
All the 10.x.x.x. addresses are NAT type so te gateway will also have a
10.x.x.x address.
> What bothers me about the gateway is that if I try to traceroute to it, the
> only hop I get is to my cable modem. I'm not even too disturbed by the fact
> that acting gateway (68.4.8.85) is not even in the network I was told I was in
> (68.4.72.0), and does not appear in my routing table. I don't know how they
> handle that, but if it works, ok.
I'm guessing what cox have told you is wrong and that the 10.73.0.1
address is the address of the gateway not the modem, then modem is
probably acting like a bridge maybe with a IP.
> But I started the traceroute stuff trying to figure out why my connection
> intermittantly stops working - so far tech support has told me every time that
> no one else is reporting problems. And if I stop and restart my network
> services, the problem clears up. I think I'm misconfigured in a mostly benign
> way, but I don't have enough information yet even to ask an answerable
> question about that. As far as busted chops go, I expected to be sent back to
> the traceroute man page, which may contain the answer, but not in a form I
> recognize.
Does traceroute to any internet site work ok, eg google. The usual
problem with these is MTU size, try reducing it with
ifconfig eth0 mtu 1000
karl.
John
Karl Heyes wrote:
> On Wed, 27 Mar 2002 04:05:37 +0000, John wrote:
>
> > What bothers me about the gateway is that if I try to traceroute to it, the
> > only hop I get is to my cable modem. I'm not even too disturbed by the fact
> > that acting gateway (68.4.8.85) is not even in the network I was told I was in
> > (68.4.72.0), and does not appear in my routing table. I don't know how they
> > handle that, but if it works, ok.
>
> I'm guessing what cox have told you is wrong and that the 10.73.0.1
> address is the address of the gateway not the modem, then modem is
> probably acting like a bridge maybe with a IP.
>
That's an interesting thought. It would make a lot of sense, since I get the
feeling they consider subscribers to be on their local network, anyway. It would
make me wonder if my IP address isn't subject to alteration when my packets leave
'cox.net', but I was acknowledged by address during an ftp session recently.
>
> > But I started the traceroute stuff trying to figure out why my connection
> > intermittantly stops working - so far tech support has told me every time that
> > no one else is reporting problems.
>
> Does traceroute to any internet site work ok, eg google.
I was too absorbed in the failures to mention that tracing to external sites works
normally.
> The usual problem with these is MTU size, try reducing it with
>
> ifconfig eth0 mtu 1000
I will try this right away, since the trouble is intermittant, it will take a little
time to decide whether it has helped.
Thanks.
Jon
read a lot more about networking, TCP/IP, and set up a baby test network so
you can get actual experiance playing around.
Karl is quite confused, and in some cases wrong. RFC1918 addresses
[10.x.x.x, etc] are not necessarily NAT'd IPs, in fact many cable providers
love to use those addresss for their routers that provide *public* transit,
with no address translation. Karl is getting warmer when he talks about
your cable modem - generally they bridge traffic, but still contain an
administrative IP, and most will route packets if you want to play enough
stupid ARP tricks, but generally bridge in the default config used by most
providers. Your modem is therefor likely *not* your gateway, and will
probably not appear in a traceroute.
You do have a route for every network in your routing table - you just don't
quite understand ranges and netmasks. That network called 0.0.0.0 with a
mask of 0.0.0.0 in your routing table? That in and of itself is an explicit
route for any destination not more specifically referenced in the routing
table, and it goes through your gateway - your gateway is answering with one
of it's other IPs - not really much rocket science there...
Traceroute is a lot like probability throey - much is contra-intuitive, but
if you think about it, it all makes sense.
-Jon
John <w...@cox.net> wrote in message news:3CA2B2B5...@cox.net...
Karl Heyes
> Karl is quite confused, and in some cases wrong. RFC1918 addresses
> [10.x.x.x, etc] are not necessarily NAT'd IPs, in fact many cable providers
> love to use those addresss for their routers that provide *public* transit,
> with no address translation. Karl is getting warmer when he talks about
> your cable modem - generally they bridge traffic, but still contain an
> administrative IP, and most will route packets if you want to play enough
> stupid ARP tricks, but generally bridge in the default config used by most
> providers. Your modem is therefor likely *not* your gateway, and will
> probably not appear in a traceroute.
I was relating the issue to the posters question and 10.x.x.x addresses
have to be NAT'd for internet access as it is reserved IP range, try
and stick with the plot! I wasn't going the mention the admin IP as
it bears no relation to the posters problem. Configuring the modem may
cause more problems and will violate his agreement with the provider.
karl.