Linux-Networking Digest #264, Volume #11 Mon, 24 May 99 13:13:47 EDT
Contents:
Re: Well got further with it. ("Curt")
Re: Kingston KNE100TX Ethernet cards and Linux (Rod Smith)
Ethernet wiring problem ("James in SC")
DNS setup Q. (Martin Ng)
Re: Kernel 2.2 TCP incompatible with certain Unices? (Joe Flasch)
Networking the internet ("v4cal")
Problem routing from cable modem (was "Can't communicate with 2nd NIC") ("Steve
Snyder")
Re: IP Masquerade/Routing (Athol Marshall)
Hiding a class C between two real class C's. Can it be done? ("John Antypas")
Diald problem (Wayne Chan)
Re: Cable modem connection under linux ("Alltek Energy Systems, Inc.")
Re: Linux: ICMP Redirect, IP Source Routing unterdruecken (Gert Doering)
Re: Logging downloads from PPP Server (Duncan Simpson)
Re: Appllications server (Henrik Carlqvist)
----------------------------------------------------------------------------
Reply-To: "Curt" <[EMAIL PROTECTED]>
From: "Curt" <[EMAIL PROTECTED]>
Crossposted-To:
comp.dcom.modems.cable,linux.redhat.announce,linux.redhat.digest,linux.redhat.install,linux.redhat.list,linux.redhat.misc,linux.redhat.rpm,linux.samba
Subject: Re: Well got further with it.
Date: Mon, 24 May 1999 08:54:59 -0500
If you can't ping something by name, your /etc/resolv.conf is probably
set wrong. It should have nameserver IP, where IP is the IP of a
DNS server, usually supplied by your ISP.
Matt Goebel <[EMAIL PROTECTED]> wrote in message
news:7i9rnb$eou$[EMAIL PROTECTED]...
> I pulled the NIC card from my Windows 98 machine that was connected to the
> Internet and put it into my Linux Box and reloaded 6.0. The card was auto
> detected and the install went well. When I booted up for the first time I
> was able to ping IP numbers but not urls, Netscape didn't work either.
I'll
> use the pump update tomorrow and try again. At least I'm one step ahead
of
> where I was. Now here is another problem(s): even though I was able to
get
> one NIC to work I'm not getting my others too. I had planned on
connecting
> the cable modem to a older 10 base card (Linksys Etherlink Combo Card II)
> since the cable modem doesn't even go that fast and save the 10/100 cards
> for connection to my hub for internet access via IP Masq and for file
> sharing. Linux seems to accept the 10 base card (a NE2000 PCI clone) but
> doesn't auto detect it. Linux won't let me use the card I was going to
use
> for connecting to the hub (ever though it is a slightly different version
of
> the same card I've hooked the cable modem up to.) So I'm going to use the
> two cards I managed to get working in the linux box and put the others in
my
> Windows machine since they are not as picky. Getting the two cards to
work
> together maybe a problem. Now for the static ip card do I need to enter a
> domain name and etc for it to work on my LAN? Or do I just need the IP
> (192.168.0.1 in this case)??
>
>
------------------------------
From: [EMAIL PROTECTED] (Rod Smith)
Crossposted-To: comp.os.linux.hardware
Subject: Re: Kingston KNE100TX Ethernet cards and Linux
Date: 24 May 1999 14:33:15 GMT
Reply-To: [EMAIL PROTECTED]
[Posted and mailed]
In article <7i8ees$6fn$[EMAIL PROTECTED]>,
[EMAIL PROTECTED] (John Winters) writes:
> Somebody asked the other day about Ethernet cards but I'm afraid I've
> lost the message. A word of warning about Kingston KNE100TX cards.
> I've been using these for some time and they were great under Linux.
> However, the latest pair I've received have had one of those nasty
> hidden upgrades, and it seems they no longer work at either 10 or 100
> Mbit.
>
> My working cards identify themselves as:
>
> DEC DC21140 (Rev 34)
>
> and the new non-functional ones are:
>
> DEC DC21141 (Rev 65)
>
> If anyone's got any pointers on how to get these going I'd be very
> grateful, but in the meantime I would avoid Kingston's Ethernet cards.
My original KNE100TX used a DEC 21140 (I don't know the board revision
number), but it failed and Kingston replaced it under warranty with one
that uses a 21143 (again, I don't know the board revision number
offhand). That board doesn't work with the drivers that come with any
Linux kernel I've tried to date (through 2.2.3), but it *DOES* work with
more recent drivers from the Linux Tulip driver website,
http://cesdis.gsfc.nasa.gov/linux/drivers/tulip-devel.html (I'm currently
using 0.90q, but I gather more recent drivers are available).
As a general comment, I'll say that not many current new "Tulip" boards
work with the drivers that come with the Linux kernel. This is because
the original DEC Tulip chips are no longer available, so these board
manufacturers have switched to other chipsets (the Kingston you got must
have been old, or perhaps they had a stockpile of original DEC chips or
reached some agreement with Intel to get a few more after Intel acquired
the rights to the Tulip chip). The Tulip drivers available from the Linux
Tulip driver site have been upgraded to work with most of these new
chipsets, but these drivers haven't yet worked their way into regular
kernel distributions. Fortunately, it's just a matter of replacing one
file in the source tree and recompiling to get the new drivers working.
So, I wouldn't advise avoiding Kingston's cards specifically. Since most
of the 10/100 PCI Ethernet boards these days are based on Tulip clones,
this is a VERY common problem, not just one with Kingston products.
--
Rod Smith
[EMAIL PROTECTED]
http://www.channel1.com/users/rodsmith
NOTE: Remove the "uce" word from my address to mail me
------------------------------
From: "James in SC" <[EMAIL PROTECTED]>
Subject: Ethernet wiring problem
Date: Mon, 24 May 1999 11:47:19 -0400
Hello,
I have 4 PCs connected to a 10baseT hub at home. The cat 5 wire currently
runs straight from the hub to each PC with no problem. When I tried to
wire up a wall jack straight to the hub (no patch panel) and then run a
patch cord from the wall to the PC it didn't work. I don't want to buy a
patch panel if I can avoid the expense.
All my cat 5 wires are T568B:
1-orange stripe
2-orange
3-green stripe
4-blue
5-blue stripe
6-green
7-brown stripe
8-brown
My wall jack is labeled T568A/B. I tried wiring by color code, I tried
wiring by what I thought made sense, and I guessed at a few. Now that I'm
writing this out, I'm thinking (it hurts a little) that maybe putting the
orange wire pair to the green jack pair and vice versa would work. What's
your opinion? Or have I made no sense at all? : )
Thanks,
James
[EMAIL PROTECTED]
------------------------------
From: Martin Ng <[EMAIL PROTECTED]>
Subject: DNS setup Q.
Date: Tue, 25 May 1999 00:02:09 +0800
Hi all,
I have successfully setup DNS on my LAN. I can ping
pc1.domain.com,
pc2.domain.com,etc.
However, I can't ping pc1 or pc2 ( without domain.com ).
Does anyone know the cause of the problem?
Martin
------------------------------
Date: Mon, 24 May 1999 11:03:12 -0500
From: Joe Flasch <[EMAIL PROTECTED]>
Subject: Re: Kernel 2.2 TCP incompatible with certain Unices?
Are you running 2.2.9 on a 6.0 build or on a 5.2 build?
What weve seen is anything on a 6.0 build is having
the problem that you have seen. I thing is a function of
box speed more than certain sites. The problem starts when
a IP frame gets lost. Then a long ack (8 sec +) retransmission loop
starts. This goes on and on then some sites give up and
reset the session. It may be that not all sites give up.
On a 5.2 build with 2.2.8 there are no problems like this.
So we think it is a 6.0 build problem and not a kernal problem.
If you are running on a 5.2 build please email me.
Joe
[EMAIL PROTECTED]
[EMAIL PROTECTED] wrote:
> I would be grateful to hear from any TCP/IP gurus regarding this
> puzzling problem.
>
> I have an SMP Pentium Pro box that runs kernel 2.2.9. While the other
> machines on the LAN (Windows and Linux) have no problems, this single
> machine cannot fetch data from certain sites -- not all sites, just a
> few. It seems to stall, even while the 2.0.36 machine next to it
> happily downloads the site.
>
> The two sites I've noticed this stall on are:
> http://www.werbach.com
> http://www.winfiles.com (sorry :)
>
> It's interesting to note that queso reports that both these sites run
> "BSDi or IRIX."
>
> I've booted the SMP box back to a 2.0.36 kernel (the stock RH 5.2 one)
> and it then works fine! Am I finding evidence that the 2.2 kernel (or
> at least, the 2.2 kernel as I've configured it) has a TCP
> incompatibility with a BSDi/IRIX TCP stack? For what it's worth, I've
> not only tried recompiling the kernel as minimally as possible
> (including non-SMP), I've also swapped NICs to see if it made a
> difference; nothing changed the stalling. I've also ruled out the
> firewall by deactivating it. I've been using GNU 'wget' in runlevel 3
> to do this testing.
>
> I'm happy to hear any and all suggestions! Thanks.
>
> Regards,
> Chris Powell
> --
> Christopher Powell Brick and Ivy Corporate Consulting
> [EMAIL PROTECTED] http://www.brickandivy.com/
>
> --== Sent via Deja.com http://www.deja.com/ ==--
> ---Share what you know. Learn what you don't.---
------------------------------
From: "v4cal" <[EMAIL PROTECTED]>
Subject: Networking the internet
Date: Mon, 24 May 1999 09:04:14 -0700
I am trying to network the net through all my computer that are pluged into
linux
the other computers are useing windows
i would like a step by step help on haw to do this
I want Linux to be the Internet server in my local network and i also want
to dail into linux and also get access to the net and my files
haw do i do this
Norbert
------------------------------
From: "Steve Snyder" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.hardware
Subject: Problem routing from cable modem (was "Can't communicate with 2nd NIC")
Date: Mon, 24 May 1999 11:10:54 -0400 (EST)
Reply-To: "Steve Snyder" <[EMAIL PROTECTED]>
Thanks to all who responded to my post "Can't communicate through 2nd NIC".
I actually *can* communicate through the 2nd NIC; it's the routing of
packets through it to the cable modem that's screwed-up.
In that original thread I passed along some bad information. I claimed
that the IP address of the cable modem (a 3Com USR CMX) was 24.4.162.173.
In fact, according to @Home's tech support:
10.71.5.44 = The IP address of the cable modem itself.
24.4.162.173 = The IP address assigned to me for use by my NIC.
Given those numbers, what, if anything, is wrong with this configuration:
# /sbin/ifconfig -a
===================
eth0 Link encap:Ethernet HWaddr 00:10:4B:9A:82:E5
inet addr:192.168.0.12 Bcast:192.168.0.255 Mask:255.255.255.0
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:1650 errors:0 dropped:0 overruns:0 frame:0
TX packets:1047 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:100
Interrupt:11 Base address:0xe400
eth1 Link encap:Ethernet HWaddr 00:60:97:C8:01:C8
inet addr:24.4.162.173 Bcast:24.4.162.255 Mask:255.255.255.0
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:1391 errors:0 dropped:0 overruns:0 frame:0
TX packets:1043 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:100
Interrupt:10 Base address:0xe800
# cat /etc/sysconfig/network
============================
NETWORKING=yes
FORWARD_IPV4=false
HOSTNAME="corona.snydernet.lan"
DOMAINNAME=snydernet.lan
#GATEWAY=
#GATEWAYDEV=
GATEWAY=24.4.162.173
GATEWAYDEV=eth1
Please bear in mind that I'm not trying to do name resolution at this
point, but just to ping static IP numbers outside my local subnet.
My goal now is simply to get a response when I ping the IP address of
@Home's name server.
Thank you.
***** Steve Snyder *****
------------------------------
From: Athol Marshall <[EMAIL PROTECTED]>
Subject: Re: IP Masquerade/Routing
Date: Mon, 24 May 1999 16:58:46 +0100
George,
thanks for your input. The problem is for ME to dial into a remote system
which has the address 128.100.0.0
This may be (and now I know is) actually on the net somewhere, so I shouldn't
route to it directly - it should go via the default route to my ISP.
The chances of a conflict are vanishingly small. I was just wondering if
I could translate a private IP address for my routing purposes into a
different IP address for the router at the other end. The reason being that
I will soon be connecting to various systems via ISDN/PPP and these networks
have invariably been set-up without reference to the valid private IP numbers.
Even if all the networks had used private address space, they might all have
chosen 192.168.0.0, for example, in which case I would still need some way of
differentiating them for routing purposes.
George Georgakis wrote:
> If I understand you correctly ...
------------------------------
Reply-To: "John Antypas" <[EMAIL PROTECTED]>
From: "John Antypas" <[EMAIL PROTECTED]>
Crossposted-To: comp.dcom.sys.cisco
Subject: Hiding a class C between two real class C's. Can it be done?
Date: Mon, 24 May 1999 16:21:26 GMT
Good morning all,
The subject says it all. I've got a network problem, I think our Cisco can
handle, but I'm not sure....
We are trying to build a redundant path to our ISP resources. In the old
days, we would had a NIC-assigned /24 and run appropriate routing protocols
between two ISPs. No problem here right?
Today, one cannot get their own /24, and getting ISPs to handle the routing
info is difficult, so we're going to have to hack this one.
What I'd LIKE to do is this, but I'm not sure it's possible...
Provider A provides a circuit with a /24 address space (real)
Provider B provides a separate circuit with a separate real /24 address
space
We have a Cisco on our end (2611, IOS 12)
A) Can we assign all our internal resources a non-routable space, say
10.1.1.* and have the Cisco translate, blindly, to one of the ISP's /24
spaces. True, it would be a manual switchover if that ISP failed, but at
least I wouldn't have to renumber every host.
B) Can the poor Cisco handle this for an entire /24 If not, what is
reasonable (20 hosts? 10 hosts?)
C) Is there a way to do automatic failover?
D) More important, are we solving the problem the wrong way?
John Antypas
[EMAIL PROTECTED]
------------------------------
From: Wayne Chan <[EMAIL PROTECTED]>
Subject: Diald problem
Date: Mon, 24 May 1999 21:59:09 +0800
Hi all,
I installed the diald rpm of version 0.14 in my RH5.2 machine. Diald is
running fine except the fire-up time in /var/log/messages. I am located
in HK (GMT+8:00) and the time show in /var/log/messages is my correct
local time. However, when diald fire up, the time recorded is always
8:00 behind. I have tried enabling and disabling the option "[ ]
Hardware clock set to GMT ". But it didn't help.
I am not sure whether my time configuration is wrong or I need to fix
some diald config. Could someone please help me out?
Here is the extraction of /var/log/messages:
May 24 20:49:26 linux2 named[244]: Ready to answer queries.
May 24 20:49:27 linux2 named[244]: sysquery: sendto([128.9.0.107].53):
Network is unreachable
May 24 20:49:28 linux2 kernel: CSLIP: code copyright 1989 Regents of the
University of California
May 24 20:49:28 linux2 kernel: SLIP: version
0.8.4-NET3.019-NEWTTY-MODULAR (dynamic channels, max=256).
May 24 20:49:28 linux2 kernel: SLIP linefill/keepalive option.
May 24 12:49:32 linux2 diald[287]: Running connect (pid =
295). <== It is always 8 hours behind!!!
May 24 20:49:32 linux2 connect: Initializing Modem
May 24 20:49:32 linux2 chat[297]: timeout set to 5 seconds
May 24 20:49:32 linux2 chat[297]: send (ATZ&C1&D2%C0^M)
May 24 20:49:32 linux2 chat[297]: timeout set to 45 seconds
------------------------------
From: "Alltek Energy Systems, Inc." <[EMAIL PROTECTED]>
Subject: Re: Cable modem connection under linux
Date: Mon, 24 May 1999 12:31:19 -0400
set your eth0 for dhcp.....
mine works that way.
Yves wrote in message ...
>Hii folks
>I am running red hat 5.2 on a pentium 200 machine.
>64 megs ram and also 10.2 gig hard drive . I am trying to figure out how to
>setup my cable modem to work with red hat . I tried
>a few things with tcp/ip with netconfig in xwindows but no luck.
>can anybody help me out ???/
>
>
>Bye :) YvesC
>
>
>
------------------------------
Crossposted-To:
de.comm.internet.routing,comp.os.linux.networking,de.comp.os.unix.networking
From: [EMAIL PROTECTED] (Gert Doering)
Subject: Re: Linux: ICMP Redirect, IP Source Routing unterdruecken
Date: Mon, 24 May 1999 16:30:17 GMT
[EMAIL PROTECTED] (Detlef Bosau) writes:
>Aber ein aktives Wegloeschen waere noch nuetzlicher.
Wenn ich mich richtig entsinne, hiess die Option "drop source routed
frames" (in der Kernel-Konfiguration).
gert
--
[EMAIL PROTECTED]
18 24 61 B 17 17 4
------------------------------
From: [EMAIL PROTECTED] (Duncan Simpson)
Subject: Re: Logging downloads from PPP Server
Date: 24 May 1999 16:49:43 GMT
In <7ib1ho$iop$[EMAIL PROTECTED]> [EMAIL PROTECTED] writes:
>Hi there. I was wondering if it's possible to log what people are
>downloading from a PPP server. The server is running RH 5.2. I've been
>looking around, and there doesn't seem to be anything that will do it.
>Just wanted to know if anyone has actually managed to do this?
Have you tried reading Apache's logs? I think RH puts them somewhere
like /var/log/httpd. Last time I looked they contained IP numbers and
requested URLs...which should translate inot the information you want
somehow.
--
Duncan (-:
"software industry, the: unique industry where selling substandard goods is
legal and you can charge extra for fixing the problems."
------------------------------
From: Henrik Carlqvist <[EMAIL PROTECTED]>
Subject: Re: Appllications server
Date: Sun, 23 May 1999 18:36:13 +0200
c4u wrote:
> Could anyone point to some HOWTO's or doc's concerning linux and
> aplications server.
>
> What do I want to do.
>
> Running (eg. Xwordperfect) on a few old 486 machines (486DX66 8 to 32
> mb, 200 mb hdd) but not locally but from an aplications server
> (pentium II 400, 256 MB, UW-SCSI 2x 9 GB).
This isn't very hard, X gives these possibilities. Let's say you are
sitting by a machine called my486.mydomain.com and want to use the
cpu-power of the server bigpii.mydomain.com. Then you should do the
following:
my486 login: username
Password: the_secret_password
bash$ startx
(Then you should start som kind of terminal window like xterm, in this
window you will type the rest)
bash$ xhost +bigpii
bigpii being added to access control list
bash$ telnet bigpii
bigpii login: username
Password: the_secret_password
bash$ export DISPLAY=my486:0
bash$ /usr/local/WordPerfect/wpbin/xwp &
Before you will be able to run startx you must configure X with a
program like xf86config or XF86Setup.
xhost +bigpii means that the server bigpii will be allowed to put
windows on your screen.
export DISPLAY=my486:0 tells bigpii that all X programs should send
their windows to the first display on my486.
Running X-windows takes some resources so you shouldn't have less than
16 MB in your 486 machines unless you really like swapping.
Hope this helps.
regards Henrik
--
spammer strikeback:
root@localhost [EMAIL PROTECTED] [EMAIL PROTECTED]
[EMAIL PROTECTED] [EMAIL PROTECTED] [EMAIL PROTECTED]
------------------------------
** 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
******************************
