Linux-Networking Digest #220, Volume #12 Sat, 14 Aug 99 05:13:38 EDT
Contents:
sendmail to slow ! ("ArTec - Vincent MAURY")
Re: script problems (Roger)
Re: PPP negotiation failure (Vilmos Soti)
Re: FTP - I don't want an anonymous account (Roger)
Re: IMAP clients for linux (William McBrine)
Re: Displaying X on an IP Masq'd machine (David Crooke)
kppp/pppd not working with /dev/modem (Neoklis)
Re: tulip switches to AUI media when network down (David Hayes)
Re: Modem Binding ("Tony Platt")
Re: route problem: Can't add gateway address ("mark")
Re: What type of cable should I used? ("mark")
Linux network problem (randomly dies)... can anyone help? ("Jay Kubicky")
Re: Firewall on same subnet? ("mark")
Re: DHCP + 3C509B NIC ?? ("mark")
Re: How can I find IP addr when using dhcp and adsl? ("mark")
----------------------------------------------------------------------------
From: "ArTec - Vincent MAURY" <[EMAIL PROTECTED]>
Subject: sendmail to slow !
Date: 14 Aug 1999 06:29:21 GMT
I use madrake 5.3 (linux kernel 2.0.36).
When I want to send an email with sendmail (I need to because a cgi-script
uses it), it takes it more than 30 secs to do it !
If I use mail as client, it's normal (much faster)
I had a same config with RedHat 5.2 and sendmail worked normaly
Does anyone know the solution ?
------------------------------
From: Roger <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.questions
Subject: Re: script problems
Date: Sat, 14 Aug 1999 02:02:57 -0500
"Stuart R. Fuller" wrote:
> Hmm, in this script, include the line:
>
> echo $PATH
Took your advice and specified explicit paths - no help... :(
> and see if the directory that contains ifconfig (/sbin)is listed. If not,
> then explicitly specify the path (/sbin/ifconfig). Also, many of the other
> commands indicated are in the /usr/* directory, which may or may not be
> mounted at the time.
Hmmm ok... I believe that everything is mounted at the same time...
This is the last item called in the startup script and because I have a
small HD(that and its my first install at linux) I just slapped
everything as / so when the root dir is mounted everything should be
avaiable at that time..
Any other ideas on how I could get my machine to email me a bootup???
------------------------------
From: Vilmos Soti <[EMAIL PROTECTED]>
Subject: Re: PPP negotiation failure
Date: Sat, 14 Aug 1999 07:00:11 GMT
ORRIN wrote:
>
> The major problem is that I can't seen to negotiate a PPP
> session to my ISP who is running BSDI BSD/OS 3.0. I can sign-on
>
> Aug 9 15:16:49 notms pppd[410]: pppd 2.3.5 started by orrin, uid 500
> Aug 9 15:16:49 notms pppd[410]: Using interface ppp0
> Aug 9 15:16:49 notms pppd[410]: Connect: ppp0 <--> /dev/modem
> Aug 9 15:16:49 notms pppd[410]: sent [LCP ConfReq id=0x1 <magic 0x330b6914> <pcomp>
><accomp>]
> Aug 9 15:16:50 notms pppd[410]: rcvd [LCP ConfReq id=0x2 <mru 1500> <asyncmap
>0xa0000> <magic 0x616403e8> <pcomp> <accomp>]
> Aug 9 15:16:50 notms pppd[410]: sent [LCP ConfAck id=0x2 <mru 1500> <asyncmap
>0xa0000> <magic 0x616403e8> <pcomp> <accomp>]
> Aug 9 15:16:52 notms pppd[410]: sent [LCP ConfReq id=0x1 <magic 0x330b6914> <pcomp>
><accomp>]
> Aug 9 15:17:19 notms pppd[410]: LCP: timeout sending Config-Requests
> Aug 9 15:17:19 notms pppd[410]: Connection terminated.
> Aug 9 15:17:19 notms pppd[410]: Hangup (SIGHUP)
> Aug 9 15:17:19 notms pppd[410]: Exit.
>
>
> The hardware is the same as used to make a connection in WFWG with
> the Trumpet dialer (thats how I get to the internet).
>
> Any help will be greatly appreciated since the SuSE installation
> help has told me they have no idea what is wrong.
>
Hi,
I had exactly the same problem, and my ISP also used BSDI. I finally
tracked down the problem that the first frame's starting character, a
"~", coming from the ISP, was missing. Try enable complete logging (So
you can see what's coming from the wire) and decode :-) it. I couldn't
find a solution and was forced to use a different package.
One way to check out if the opening "~" character is missing is to logon
to your ISP through minicom or seyon and login to PPP. See if the first
printable character in the ppp garbage (full with "}") is a tilde. If
not, then you have the same problem I had and I don't know any solution.
I think the problem is on the BSDI side. My ISP's tech support was a
nice person and logged into his account through HyperTerminal and sent
me the output. The frame opening tilde was missing there, too.
Here it seems the Linux TCP/IP code ignores the incomplete frame (and
causes headaches) but the Windows one (I used WIn95, long gone from my
machine...) "recreates" the missing info.
If you make a search on dejanews.com about this topic you can read about
it. It happened around late April or May.
Sorry for the not-so-encouraging info. Vilmos
--
Looking for a job in British Columbia.
Have you recompiled your kernel today?
Linus? Jesus Christ Superstar
Bill? Clueless in Seattle
------------------------------
From: Roger <[EMAIL PROTECTED]>
Subject: Re: FTP - I don't want an anonymous account
Date: Sat, 14 Aug 1999 02:08:24 -0500
I'm running wu-ftpd 2.5 and this is how I disabled anonymous ftp
access. In you /etc/ftpusers it has a listing of people that CAN NOT
log in via ftp. Just added ftp and anonymous to the list.
To test this add yourself to this list and see if your denied...
[root@peoria43 bb]# cat /etc/ftpusers
root
bin
daemon
adm
lp
sync
shutdown
halt
mail
news
uucp
operator
games
nobody
ftp
anonymous
------------------------------
From: William McBrine <[EMAIL PROTECTED]>
Subject: Re: IMAP clients for linux
Crossposted-To: comp.os.linux.x,comp.os.linux.development.apps,comp.windows.x.kde
Date: Sat, 14 Aug 1999 07:19:09 GMT
In comp.os.linux.networking Kaz Kylheku <[EMAIL PROTECTED]> wrote:
: Why do you think than an IMAP client also has to have an interface for
: browsing mail?
: The fetchmail program may be an alternative for you. It will pull mail
: from an IMAP account and deliver it to a local mailbox.
That would defeat a lot of the advantages of IMAP, wouldn't it? With an
IMAP client, like Pine, you pull down only the headers when you open the
INBOX; you pull down the bodies as you read, and even then, if there are
attachments, they're left on the server unless explicitly selected. I
haven't tried fetchmail, but I assume it must grab full messages only. No?
--
William McBrine | http://www.clark.net/~wmcbrine/
[EMAIL PROTECTED] | ./\./\./\./\./\./\./\./\./\./\.
------------------------------
From: David Crooke <[EMAIL PROTECTED]>
Reply-To: [EMAIL PROTECTED]
Crossposted-To: comp.os.linux.x
Subject: Re: Displaying X on an IP Masq'd machine
Date: Sat, 14 Aug 1999 05:44:26 GMT
Eric deRiel wrote:
>
> "Sunil P. Khatri" <[EMAIL PROTECTED]> writes:
>
> > I have successfully installed IP Masquerading on my home linux box.
> > The configuration is as shown below:
> > _______ _______
> > 192.168.0.3-------| | | |
> > | | | Local |
> > 192.168.0.2-------| HUB |------------------| GW |---------
> > local machines | | 192.168.0.1 | | w.x.y.z
> > 192.168.0.4-------|_______| |_______|
>
> I have a similar configuration, but would like to be able to direct
> traffic to any one of the internal machines, selected at client run
> time. Currently my GW runs an ipmasqadm rule similar to:
>
> ipmasqadm autofw -A -r tcp -d 6000 7000 -h 192.168.0.4
>
> Adding another, similar rule (-d 16000 17000 -h 192.168.0.3) is easy
> enough, but is it possible to make an X client run using a
> non-standard server control port?
Ever wondered what the ":0.0" was for at the end of the DISPLAY setting?
This will do what you want, but it's not "non-standard" at all - the
port number is 6000 plus the server number given, so e.g. a DISPLAY
setting of "w.x.y.z:3.0" means screen 0 on the server on port 6003 on
w.x.y.z
I don't know if ":10000.0" would work though, stick to the 6000 range
>
> Or is there some other, better way of doing such a thing?
No, this is the right way. Get the gateway to forward each port from the
range 6001 to 60nn to port 6000 on one of the clients, or launch the
clients' display servers on higher numbered ports.
>
> Any suggestions, or even hints as to where an answer might be found,
> are greatly appreciated.
You're welcome :-)
--
David Crooke, Austin TX, USA. +1 (512) 656 6102
"Open source software - with no walls and fences, who needs Windows
and Gates?"
------------------------------
From: Neoklis <[EMAIL PROTECTED]>
Subject: kppp/pppd not working with /dev/modem
Date: Sat, 14 Aug 1999 05:03:10 GMT
Hi!
I have suse 6.0 installed on a PC and tried to connect on the net using
kppp, but if I specify /dev/modem pppd dies when started. If I specify
/dev/ttyS1, all is well. I tried a fresh symlink between /dev/ttyS1 and
/dev/modem but the fault persists. Also I cannot connect from the
terminal using ppp-up, I just get +++ATZ and a wait for the 1-minute
timeout .
Any ideas please? TIA!
--
Regards
Neoklis
My (ex-Acorn) RiscOS Homepage: http://www.arcsite.de/hp/neoklis
Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.
------------------------------
From: [EMAIL PROTECTED] (David Hayes)
Crossposted-To: comp.os.linux.hardware,comp.os.linux.setup,linux.redhat.axp
Subject: Re: tulip switches to AUI media when network down
Date: 14 Aug 1999 07:34:33 GMT
On Thu, 12 Aug 1999 22:19:17 GMT, [EMAIL PROTECTED]
<[EMAIL PROTECTED]> wrote:
>i've got a couple of machines with Digital DC21040 Tulip cards
>that've got 10baseT, AUI and BNC connectors. whenever we
>experience a power failure, these machines' NICs switch to
>AUI media.
>
>short of replacing the cards, is there any way to stop them
>from switching media? ifconfig doesn't seem to have any
>options.
ifconfig doesn't, but the tulip.o module does. Here's a section from
the source code:
/* The possible media types that can be set in options[] are: */
static const char * const medianame[] = {
"10baseT", "10base2", "AUI", "100baseTx",
"10baseT-FD", "100baseTx-FD", "100baseT4", "100baseFx",
"100baseFx-FD", "MII 10baseT", "MII 10baseT-FD", "MII",
"10baseT(forced)", "MII 100baseTx", "MII 100baseTx-FD", "MII
100baseT4",};
So you can load it with "modprobe tulip options=12" (10baseT-forced)
might do it for you.
--
David Hayes
[EMAIL PROTECTED]
------------------------------
From: "Tony Platt" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.hardware,redhat.networking.general
Subject: Re: Modem Binding
Date: Sat, 14 Aug 1999 17:45:30 +1000
EQL
Dood wrote in message ...
>ANyone knows how to do modem binding in Linux?
>System runs an adsl connection, Mandrake 2.2.9 kernel.
>I intend to bind 2 3com 56 modems together as a backup connection.
>Any ideeaz?
>
>DOOD
>
>------------------ Posted via CNET Linux Help ------------------
> http://www.searchlinux.com
------------------------------
From: "mark" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.misc,comp.os.linux.setup
Subject: Re: route problem: Can't add gateway address
Date: Sat, 14 Aug 1999 02:43:47 -0500
Your problem may very likely be this simple:
instead of just 'route', run 'route -n', which prevents the translation of
numeric addresses to names. I have a box that exhibits the same symptoms
you describe if I just run 'route' (hangs forever, until ^C), but shows
everything perfectly with 'route -n'.
Andrey Smirnov <[EMAIL PROTECTED]> wrote in message
news:7ovna3$ku8$[EMAIL PROTECTED]...
> Your ifconfig statement for eth0 looks like etho (with letter 'o'). Make
> sure it's eth0 (with '0' - zero!).
>
> Good luck!
>
> Youjip Won <[EMAIL PROTECTED]> wrote in message
> news:[EMAIL PROTECTED]...
> > Dear Linux guru,
> > I am fan of linux, but never ever able to get the network setup in
> > one shot. This time was not an exception. After several trials and
> > errors, I am finally writing this message in a hope that I could get
> > some help.
> >
> > To make long story short, I cannot add gateway to routing table.
> >
> > Here's what had happened. I can ifconfig and route add 'lo' and 'etho'.
> > When I do "route add default gw 166.104.88.1", it does not complain.
> > BUT, if I type 'route' to verify the gw in routing table, it goes to the
> > infinite loop. It only shows the first two lines in routing table(lo,
> > eth0) and goes into infinite loop. I eagerly hope that someone have had
> > similar problems and provide me some clue.
> > Please refer to the attached script.
> > Thanks a mil!!!!
> >
> > Youjip
> >
> > Script started on Thu Aug 12 21:40:56 1999
> > [root@sobaek network-scripts]# ifconfig
> > [root@sobaek network-scripts]# route
> > Kernel IP routing table
> > Destination Gateway Genmask Flags Metric Ref Use
> > Iface
> > [root@sobaek network-scripts]# ifconfig lo 127.0.0.1
> > [root@sobaek network-scripts]# route add -net 127.0.0.0
> > [root@sobaek network-scripts]# ifconfig etho 166.104.88.108
> > [root@sobaek network-scripts]# route add -net 166.104.88.0
> > [root@sobaek network-scripts]# route
> > Kernel IP routing table
> > Destination Gateway Genmask Flags Metric Ref Use
> > Iface
> > 127.0.0.0 * 255.255.255.0 U 0 0 0
> > lo
> > 166.104.88.0 * 255.255.255.0 U 0 0 0
> > eth0
> > [root@sobaek network-scripts]# route add gw 166.104.88.1
> > [root@sobaek network-scripts]# route
> > Kernel IP routing table
> > Destination Gateway Genmask Flags Metric Ref Use
> > Iface
> > 127.0.0.0 * 255.255.255.0 U 0 0 0
> > lo
> > 166.104.88.0 * 255.255.255.0 U 0 0 0
> > eth0
> > ===> Goes to infinite loop
> >
> >
>
>
>
------------------------------
From: "mark" <[EMAIL PROTECTED]>
Subject: Re: What type of cable should I used?
Date: Sat, 14 Aug 1999 02:58:36 -0500
If the cable you are using truly is 'straight through', then you're fine.
Ethernet talks on 1,2,3, and 6. With most cables, if you hold the ends
together side-by-side with the tabs down, the wires are orange-white,
orange, green-white, blue, blue-white, green, brown-white, brown. Both ends
should look the same. If one end has the greens and oranges flipped, then
you have a crossover cable, which probably isn't going to work for you.
Twisted pair doesn't refer to a pinout, it's refering to the twists in the
wires as you run the length of the cable, which are supposed to help reduce
interference.
Chris <[EMAIL PROTECTED]> wrote in message
news:7p15tt$gb3$[EMAIL PROTECTED]...
> Currently, I am using 3Com's 3c905B 'Cyclone' NIC connected to an ADSL
> modem. Another ATM card (from Virata - the VL1025) is also connected to
the
> ADSL modem. The ADSL modem is Alcatel 1000. It has 2 sockets for the RJ-45
> type cable. One socket is labelled '10-BaseT' and the other 'ATMF-25'. The
> 3c905B NIC is connected to the socket labelled '10-BaseT' and the VL 1025
> ATM card is connected to the 'ATMF-25' socket at the ADSL modem.
>
> The RJ-45 cable I am using, I think, is the normal 'pin-to-pin' type
cable.
> I've heard people mentioning a 'twisted-pair' RJ-45 cable. They said it is
> cross-linked to some other pins at each end of the cable. Is it true?
>
> I used 2 cables to connect respectively both my NIC and ATM card to the
ADSL
> modem. Am I using the correct cable for my 3c905B NIC to the ADSL modem
> because my ATM card is working fine except for my 10-BaseT connection to
the
> ADSL modem. Can surf the net and everything else using the ATM connection
> except for the 10-BaseT one.
>
> Reason why I used the 3c905B NIC to connect to the ADSL modem is because
> there is currently no drivers or ATM support for my Virata ATM card.
>
> ifconfig reported the IRQ, base address and eveything else correctly. Only
> weird thing is that the TX queue value is always rising (meaning that
> something is sent) BUT RX queue value is always ZERO (meaning that nothing
> is received). Route settings and DNS and named services all running and
> correct.
>
> Only thing is that when I ping to anything other than my own static IP (as
> given by my ISP), the ping program hangs there as if there is no response
or
> whatsoever.
>
> Do I need to use the 'twisted pair' instead of the current one I have?
> Heard that my standard cable should be connected via a hub then from the
hub
> to the modem. Btw, in my scenario, is my ADSL modem the router or some hub
> device or what role is it? What RJ-45 type cable (exact pls) do I need? Or
> am I already using the correct one (which I doubt so) and I am missing out
> something. Check all those settings (in Linux) and have been playing about
> with the netcfg for a month already yet I still cannot get the connection
> up.
>
> Can anyone give me their precious views/solutions? Really am desperate.
> Thanks.
>
>
------------------------------
From: "Jay Kubicky" <[EMAIL PROTECTED]>
Subject: Linux network problem (randomly dies)... can anyone help?
Date: Sat, 14 Aug 1999 01:07:20 -0700
I've been having problem with my Linux box networking for the past couple of
months (sys is P-200/64MB/RedHat 5.2/Bay Networks 10/100 adapter/static IP
address/connected to cable-modem through hub - two other Win-PCs on subnet
using DHCP to get their IP addresses - yes I can get multiple IP addresses,
and these PC's have been working fine all along). Basically, after some
random length of time, all of the networking 'dies' on the box - I stop
receiving all incoming requests (telnet, ftp, http, etc.) and I can't send
anything out (can't even ping to a numeric IP address (i.e.
100.100.100.100)).
Once I'm in this 'crashed' state, if I type /sbin/route, it displays the
entry for my local subnet, for the loopback interface, but then hangs when
it gets to my gateway entry (I think because it can't do the name lookup to
print this line). If I type /sbin/route -n, I get basically the same thing
I got when everything was working (except the values in the 'Use' column are
a bit different). The only way I've found to get out of this state, short
of rebooting, is to go to /etc/sysconfig/network-scripts and run 'ifdown
eth0' followed by 'ifup eth0'. Of course these scripts do quite a bit, so
it's not clear to me what part of them is responsbile for the recovery.
Something else interesting is that if I'm running sniffit (a packet
sniffer), it continues to operate during these 'crashes'.
Oh - one other thing. When I reboot the system, the networking doesn't
always 'work' (i.e. can't get in from an outside machine). However, if I
log in and then do anything that accesses the external net (i.e. ping or
whatever), then everything is fine and I can get in from outside (well,
until it crashes).
This behavior has only been happening within the past couple of months.
Prior to that, everything was fine (I never observed anything like this
before). I'm pretty sure everything is configured correctly (like I said,
it worked for several months with the identical config files). I sent a
request to my ISP and they couldn't think of anything that had changed
(although I'm still suspicious - I know I've had at least one DHCP problem
that was related to a configuration problem on their end). Anyone have any
ideas?
Thanks in advance,
Jay
------------------------------
From: "mark" <[EMAIL PROTECTED]>
Subject: Re: Firewall on same subnet?
Date: Sat, 14 Aug 1999 02:52:24 -0500
Why do you want both sides of your firewall to appear to be on the same
subnet? Your DSL router's ethernet port has, from your picture, an
interface at x.x.86.161. Your firewall's 'dirty' side will be x.x.86.162.
That's all fine and good. Take your second NIC, throw any address you
please on it, and do IP masq'ing. The outside world will see all requests
as originating from your firewall. You can use ipchains to configure access
to boxes behind the firewall based on protocol or port, but you won't even
need to do that unless you have something really complex planned for the
internal LAN. If all you're wanting to do is give 2, 3, or 20 systems in
your house the ability to use your 'net connection, and all you are allowed
is a single IP address, then IP masq'ing is perfect for you...
Jerry Craker <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> As far as I know, that is not possible. The problem arises with routing.
> You cannot have the same subnet/subnet mask on 2 separate cards on a
> machine. The only real way to do this with the same "subnet" is to fix
the
> subnet mask such that each side of your firewall "looks" like a different
> network. Do the IPs on the LAN need to remain exactly what they are, or
> can you change them, such that if you changed the subnet mask, you could
> have 2 separate lans. BTW, I am not sure what the subnet mask itself
would
> have to be.
>
> --Jerry Craker --
> Kyle Page wrote:
>
> > I'm not sure if this is even possible (thus the post) but...
> > I'm trying to setup a Linux Firewall (using Red Hat v5.2) between my DSL
> > router and my local LAN. It goes something like this:
> >
> > ------ ----------- ---------- -----
> > | WAN | -- | My router | ------- | Firewall | ------------ | LAN |
> > ------ ----------- ---------- -----
> > x.x.86.161 x.x.86.162 x.x.86.163
> >
> > Subnet: 255.255.255.240
> > Network: x.x.86.160
> >
> > As you can see from the above, I'm trying to place the Firewall on the
> > same subnet as the network for which I'd like to protect. That is,
> > between my router and LAN. It seems like this would be possible
> > somehow??? The problem I'm having now is that when I activate both
> > cards on the firewall, only the last card to be activated is reachable..
> > that is ping'able from the firewall machine. For example, if eth0 and
> > eth1 are configured as x.x.86.162 and x.x.86.163 respectively, then I
> > can only ping x.x.86.163 (assuming it is activated last). If I activate
> > eth0 last, then only eth0 is reachable. However, if I configure them so
> > they are on separate networks, let's say eth0 = x.x.85.122, and eth1 =
> > x.x.86.163 and activate them both, they are both reachable. It is only
> > when both NICs are configured to be on the same network that I have this
> > problem. Therefore, I have ruled out any hardware problems. Would this
> > "routing" issue be handled by "ipfwadm"? I've looked many places for a
> > solution and I'm starting to lose gumption (sited "Zen and the Art of
> > Motorcycle Maintenance") :) My DSL router does have Firewall software
> > but I'd rather not spend the extra cash.. and I'd rather do it myself.
> >
> > Cheers,
> > Kyle
>
------------------------------
From: "mark" <[EMAIL PROTECTED]>
Subject: Re: DHCP + 3C509B NIC ??
Date: Sat, 14 Aug 1999 03:02:14 -0500
pump, RedHat's proprietary DHCP client, is horrible. Absolutely flaky. Do
yourself a huge favor and download the ISC DHCP distribution from
ftp.isc.org/isc/dhcp. If that isn't all you need to fix your problems, I'll
be very surprised. I haven't heard anything good at all about pump... it
really wasn't ready for 'prime time'.
Craig Mankiewicz <[EMAIL PROTECTED]> wrote in message
news:7ovpbm$mu6$[EMAIL PROTECTED]...
> I have the exact configuration you do and I am having the exact same
> problem. If you figure it out, please post your results. Thanks.
>
>
> Lance C Dial <[EMAIL PROTECTED]> wrote in message
> news:[EMAIL PROTECTED]...
> > On Fri, 6 Aug 1999 02:32:54 -0500, "Stew Rappaport"
> > <[EMAIL PROTECTED]> wrote:
> >
> > >Hello- I have installed two 3C509B cards in my RedHat 6.0 system. I
> have
> > >configured them with the DOS utility to both have PnP Disabled and they
> have
> > >separate IRQ's. Using the control-panel applet in Gnome I configured
> eth0
> > >to use DHCP and activate at boot. I configured eth1 with a 192.168.x.x
> IP
> > >address.
> > > My problem is then when I connect eth0 to my cable modem it says on
> boot
> > >that:
> > >
> > >bringing up device eth0
> > >FAILED: unable to determine IP information. FAILED.
> > >
> > >After this if I run ifconfig I get only information for lo and eth1.
> > >
> > >I KNOW my ISP uses DHCP, and I have heard of some who use it
successfully
> > >with this ISP. What is my problem. I have read the DHCP mini-HOWTO
> without
> > >much luck. Am I being to vague, missing something obvious, or just
> stupid.
> > >Why am I not getting a public IP address? Any help or suggestions
would
> be
> > >greatly appreciated. Thanks.
> > >Stew Rappaport
> > >
> > >
> >
> > I had the exact same problems. I found out that I needed to
> > power-cycle the cable modem and let it sit off for a few minutes. Then
> > I turned it back on and ran this command:
> >
> > /sbin/pump -h hostname -i eth1
> >
> > Well, for you it would be eth0 and the hostname would be your Linux
> > hostname, which should be the one they supplied you. If they did not
> > give you one, use whatever hostname you are currently using on the
> > Linux box.
> >
> > In order to make this stick after a reboot, you have to edit your ifup
> > file to include the /sbin/pump command just like above. I think the
> > ifup file is in /etc/sysconfig/network-scripts but it may be somewhere
> > else, I can't remember.
> >
> > I have an identical setup with the two 3C509Bs (PnP disabled) RH6.0
> > and I was getting the same errors. Good luck
> >
>
>
------------------------------
From: "mark" <[EMAIL PROTECTED]>
Subject: Re: How can I find IP addr when using dhcp and adsl?
Date: Sat, 14 Aug 1999 03:12:39 -0500
Most likely, the 10.0.0.2 interface is indeed your eth0, and your ADSL's
ethernet interface is 10.0.0.1. Is your ADSL modem a Cisco? And is your
connection PPP? If so, telnet to it and type "sho nat" at the prompt.
You'll see a list of entries with your 10.0.0.2 address, and a corresponding
'real world' address. That address is the ticket to putting up any server
you want on your local LAN. To set it up, you'll need to do 'enable' at the
prompt to get into exec mode. Then, for example, let's say you want a web
server there in your house...
set nat entry add 10.0.0.2 80 204.43.43.21 80 tcp
The 204 address is, of course, made up, but you get the idea. Now any
packet destined for port 80 that hits your 'real' address will be routed
internally to your web server at 10.0.0.2. Damned easy ; )
Works for whatever you want to put up. If you do FTP, remember that port 21
isn't enough, you also have to point port 20 at the same host.
Ok, ok... so I rambled : ]
mark
Stephen Satchell <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> [EMAIL PROTECTED] (Brian Smith) wrote in <[EMAIL PROTECTED]>:
>
> >I'm using DHCP on an ADSL connection. Using arp and ifconfig, I'm only
> >able to see the ip addr as 10.0.0.2. I've looked at the dhcpc-etho-info
> >file and it shows the same thing. Is there somewhere else that I can
> >look?
> >
> >Thanks,
> >Brian Smith
> >
>
> You could use tcpdump or an Ethernet packet sniffer to monitor the DHCP
> exchange. The thing is, ifconfig does give the address assigned to you
> as part of its output. (For example, ifconfig on my firewall system
> shows my address to be part of a PBI/NBI subnetwork.)
>
> Given that, I would suspect that your ISP gets around certain problems by
> using a private network for ADSL users and provides NAT services to map
> "real" addresses into the private Net-10 space. This would also prevent
> you from putting up a server without them knowing about it and you paying
> for the bandwidth to do so.
>
> "ifconfig" *is* the answer. If you have more than one interface, make
> sure that you are looking at all of them. In my case, I'm running two
> ethernet cards with three subnetworks, which makes life, er, interesting.
> Two of the subnetworks are in Net 10 (private and DMZ) while the third
> network (on the second interface) is in a real IP address range.
>
> You might want to reconsider your problem -- if you have NAT between you
> and the real world, it would be real hard for a script kiddie to get to
> your system... :)
------------------------------
** FOR YOUR REFERENCE **
The service address, to which questions about the list itself and requests
to be added to or deleted from it should be directed, is:
Internet: [EMAIL PROTECTED]
You can send mail to the entire list (and comp.os.linux.networking) via:
Internet: [EMAIL PROTECTED]
Linux may be obtained via one of these FTP sites:
ftp.funet.fi pub/Linux
tsx-11.mit.edu pub/linux
sunsite.unc.edu pub/Linux
End of Linux-Networking Digest
******************************
