Linux-Networking Digest #527, Volume #12 Thu, 9 Sep 99 12:13:33 EDT
Contents:
Re: Browsers and Linux (Norman Levin)
Re: Linux dynamic IP addr. (bill davidsen)
Re: Ftp Server (midknite)
setting kernel variables at boot time (Bernhard =?iso-8859-1?Q?K=FCpper?=)
Re: linux box vs switched hub (Greg Leblanc)
traffic conting ("roman")
Re: xdm while NCD to RedHat6.0 - no response ("Randall J. Parr")
NFS problem between AIX and SuSE ("Antony Mak")
2 NICS on the same machine (gendro)
ISP database package needed ("Frank Sinatra")
Re: Redirection Of Port to Internal Network ([EMAIL PROTECTED])
Re: icmp reply problem on home network ([EMAIL PROTECTED])
Re: xdm while NCD to RedHat6.0 - no response ("FEEB")
Re: list of network cards and drivers - where? (midknite)
Re: ipchains ICMP forging? (bill davidsen)
Re: IPChains and FTP (bill davidsen)
----------------------------------------------------------------------------
Date: Thu, 09 Sep 1999 07:21:03 -0400
From: Norman Levin <[EMAIL PROTECTED]>
Reply-To: [EMAIL PROTECTED]
Crossposted-To: alt.os.linux,comp.os.linux.setup
Subject: Re: Browsers and Linux
trader wrote:
>
> Norman Levin wrote:
> >
> > I wonder where the 'custom' of answering after the
> > question comes from? I've just be going thru some of
> > my offline usergroups, and I've gone through a dozen
> > appends that start with ">" and the same original
> > question ... and I have to scroll down to see
> > new stuff. If I'm really interested in the
> > original append (and I can't remember if from
> > the subject line), I can do that.
>
> i dunno, but if you answer my question before i ask it, i'd find that a
> bit ostentatious.
*** Ostentations? Don't think that's the word. But, maybe interspersed
answers is the best way.
i've been reading newsgroups since 1992, and i know
> the "custom" goes back at least until then.
>
> trader
>
> --
> The views expressed above are not necessarily those of anyone else on
> the planet.
> Email address spam protected; remove underwear to reply.
--
Norman Levin
vm/dynAmIX inc.
------------------------------
From: [EMAIL PROTECTED] (bill davidsen)
Subject: Re: Linux dynamic IP addr.
Date: 9 Sep 1999 14:49:05 GMT
In article <[EMAIL PROTECTED]>,
Michael Kelly <[EMAIL PROTECTED]> wrote:
| If you need to find out programmatically I believe most people
| use an ip-up script to call ipconfig to get the ip address assigned.
But... the ip-up script is called with the address as one of its
arguments, why would it use ifconfig? See the man page.
--
bill davidsen <[EMAIL PROTECTED]> CTO, TMR Associates, Inc
"So let it be written, so let it be dumb." Pharaoh Dufus the last...
------------------------------
From: midknite <[EMAIL PROTECTED]>
Subject: Re: Ftp Server
Date: Thu, 09 Sep 1999 09:39:33 -0400
Rivi wrote:
>
> I'm really new on linux and i really know very little...
>
> Could you tell me how to st up the ftp server on a slackware version?
> It is installed but i have no idea of how to have it charged and running
the documentation is found in the man pages. typing `man ftpd` will
show the man page for the ftp "daemon" (aka server). It should tell
you everything you need to know.
--
brian kowolowski
gpg key / infos http://www.cryogen.com/midknite/gpg.html
gpg print: F6B6 076D 4BFC CD14 7C14 1A2F 61DA BDE5 7A88 D6C3
------------------------------
From: Bernhard =?iso-8859-1?Q?K=FCpper?= <[EMAIL PROTECTED]>
Subject: setting kernel variables at boot time
Date: Thu, 09 Sep 1999 17:12:33 +0200
What is the offical, blessed way to set the kernel variables
in /proc/sys/net/ipv4/* at boot time? Of course I can make
a script that does all "echo 1> ..."'s I need but I wonder
if there isn't a smarter way?
Thanks in advance.
--
Bernhard K�pper |Tel.: (49)-(30)-20192310
AGr Quantenchemie |Fax.: (49)-(30)-20192302
Humboldt-Universit�t |Mail: [EMAIL PROTECTED]
J�gerstr. 10/11, D-10117 Berlin|WWW: http://www.chemie.hu-berlin.de
------------------------------
From: Greg Leblanc <[EMAIL PROTECTED]>
Subject: Re: linux box vs switched hub
Date: Thu, 09 Sep 1999 14:56:11 GMT
In article <[EMAIL PROTECTED]>,
[EMAIL PROTECTED] (David C.) wrote:
> Greg Leblanc <[EMAIL PROTECTED]> writes:
> > [EMAIL PROTECTED] (David C.) wrote:
> >> Greg Leblanc <[EMAIL PROTECTED]> writes:
> >>>
> >>> Well, the way that he phrased the question, yes. However, Linux
can
> >>> function as a switch (for some damn reason the howto is called
> >>> bridging, but they MEAN what I call switching).
> >>
> >> Switching is an ambiguous term.
> >>
> >> It sometimes refers to layer-2 switches, which is perform bridging,
> >> and sometimes to layer-3 switches, which perform routing. And
> >> sometimes it refers to other kinds of switches, which may perform
> >> bridging, routing, prioritization, filtering, firewalling, and
other
> >> services.
> >>
> >> I prefer to keep the terms unambiguous and say "bridging" and
> >> "routing" when I'm talking about a single service. I suspect the
> >> author of the howto feels the same way.
> >
> > I hadn't thought of it that way. That sort of throws my perceptions
> > off a bit, since they say that the routing code in the kernel
handles
> > the bridging. And don't layer 3 switches function a bit differenly
> > from routers? I know that they work on the protocol layer, but our
> > layer3 switches (different from our layer2 switches) act a bit
> > differently from our routers (a pair of Cisco 2500s). I can't put
it
> > into words very easily, since I don't have a reference book or the
> > devices available to me from home.
>
> I can't see how. If it's really switching on layer-3, then it must
> decrement TTL and rewrite the layer-2 header information as the
packets
> pass through. They may build their forwarding tables from static
> configuration, routing protocols (like RIP and OSPF), or through other
> means, but it's still routing.
Well, it doesn't decrement TTL on ip packets that pass through it, at
least according to NT tools. I know that it has the ability to build
forwarding tables based on IP address information, and not just MAC
addresses, which means that it can't be purely layer 2, right? I don't
think it's using RIP, because it examines the header and destination on
all of the packets that travel through it. It does this to allow
devices that were connected to one port on the switch, to move to
another port and still work without needing an address/subnet change.
>
> It is possible that the switch is really switching at layer-2, but is
> using layer-3 (or higher) information to perform filtering,
> prioritization, firewalling or other non-routing activities.
That seems like a posibility, but I'm not sure how to test out what the
device is actually doing, and 3Com's product specs aren't exactly clear
on this.
>
> If you know more about the difference between these two devices,
please
> let me know. I'm actually quite curious.
>
> -- David
>
--
It's pronounced "sexy" not "scuzzy"!
Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.
------------------------------
From: "roman" <[EMAIL PROTECTED]>
Subject: traffic conting
Date: Thu, 9 Sep 1999 17:23:46 +0200
I want to count the traffic over the line for each ip-address.
I have a redhat Linux 6.0 and 2 network cards in my computer.
Does anybody know which software can do this.
Thanx for any help.
Roman Berngruber
------------------------------
From: "Randall J. Parr" <[EMAIL PROTECTED]>
Reply-To: [EMAIL PROTECTED]
Crossposted-To: comp.os.linux.x
Subject: Re: xdm while NCD to RedHat6.0 - no response
Date: Thu, 09 Sep 1999 08:55:47 -0500
skip.............
>
> Unfortunately, that did not work. But I upgraded to the newer version of
> the NCD software (3.3.2) and I am getting the CHOOSER screen with a menu of
> about 20 *NIX boxes I can connect to EXCEPT the one I want to connect to.
> If I force the address of the RedHat6.0 box in CHOOSER I get:
>
> asking host "the.RedHat.box" for XDMCP session
>
> and after a while
>
> XDMCP error - login host not responding
>
> Any hints, please?
Had the same problem. I got the fix from some other threads.
As installed, RedHat 6.0 uses gdm 1.0.0 instead of xdm. Actually, in
/etc/initab it calls /etc/X11/prefdm which is a link to gdm.
The gdm 1.0.0 docs imply that is supports XDMCP but it does not. It is my
understanding that gdm 2.0.0 beta does support XDMCP (supposedly more stable
than 1.0.0 even though a beta). I found gdm 2.0.0 beta source via a link from
the www.gnome.org site software map. I decided NOT to install this but rather
to wait for the gdm 2.0.0 final (and rpms) to be available.
The other fix is to change /etc/inittab to use /usr/X11R6/bin/xdm. I did this
and it started working fine from PC xservers, etc. It also provides a more
"standard" X login sequence (albiet a little more messy and limited).
In the long run, I plan to install gdm 2.0.0 so as to provide and use the gdm
login sequence with its extra features and options (ability to choose desktop,
etc.).
I, for one, wasted too much time tracking this down and finding a solution. I
really wish the people who decided to use gdm 1.0.0 (even though it didn't
support XDMCP as it's doc's stated) would have added some notes, comments, RH
FAQ's or something letting the world know. This kind of things is what feeds
the oppositions argument that Linux is harder to install and maintain.
R.Parr
Temporal Arts.
------------------------------
From: "Antony Mak" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.misc,comp.unix.aix
Subject: NFS problem between AIX and SuSE
Date: Thu, 9 Sep 1999 23:19:09 +0800
Hi all,
I recently setup a NFS server on my SuSE 6.0 box for backup purpose.
Everything was fine when I exported a FS between two SuSE Box. But when I
exported a FS to a AIX 4.3.2 Box, it take over one hour to copy a 5MB file
from the exported filesystem(SuSE) to local filesystem(AIX). It didn't
provide any logs or messages either in the linux bos nor AIX box. Can anyone
have any idea on this problem?
thanks
antony
------------------------------
From: gendro <[EMAIL PROTECTED]>
Subject: 2 NICS on the same machine
Date: Thu, 09 Sep 1999 14:37:42 GMT
I have some problems configuring a PC as firewall/router. I have a ADSL
internet connection which I want to pipe to my other PC (p200 running
Win98/Linux).
Facts:
On my router-to-be, I have RedHat 6.0 installed - no X - customized with
minimal apps to do what I want.
PC: 486-dx2-66 VLB - 400 Meg HD, 16 Ram
NICs : 2 DLink DE220t
My problems resides in the NIC detection and configuration.
Iniatial config of the cards have been done in accordance with the
supplier specs ie I have used the DOS software supplied with the driver
to configure both cards and make sure that they have been PNP disabled
and set to different I/O adress and IRQ.
Problem:
At first boot, eth0 was detected which was the IO300 IRQ10 NIC.
So I launched linuxconf to configure the second card (eth1) as IO 320
and IRQ5. Next, I logged off and rebooted the machine to ensure that the
new setting would take place.
At the second boot where ethx are loaded, eth0 successfully installed
but not eth1.
On top of that, eth0 has connected itself to IO 320 - IRQ which was
supposed to be IO 300 - IRQ !!!
How come ?
Is it possible that linux does not allow having two NIC of same
brand/model ?
OR
Is there any specific setting that has to be done in order to ensure
that the two card are treated seperately at installation time even
though they use the same driver ?
Thanks
Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.
------------------------------
From: "Frank Sinatra" <[EMAIL PROTECTED]>
Subject: ISP database package needed
Date: Thu, 9 Sep 1999 10:11:17 -0400
Reply-To: "Frank Sinatra" <[EMAIL PROTECTED]>
We are looking for a Linux based ISP database/billing package. Any leads or
tips will be appreciated.
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: Redirection Of Port to Internal Network
Date: Thu, 09 Sep 1999 15:00:19 GMT
Thanks that clears some things up. So all redirection is based on
external IP to internal machine? Doing it this way will all everyone
to be redirected? Can I restrict it to a single Internet IP? Would
that be accomplished via an IPChains Rule?
Also, The /usr/sbin/modprobe ip_masq_portfw gives me :
Can't Locate module ip_masq_portfw. I thought I had the kernel
configured already according to the Readme.
Thanks,
Shawn
In article <[EMAIL PROTECTED]>,
"Cowles, Steve" <[EMAIL PROTECTED]> wrote:
> EXT_IP is the "external" IP address of my Linux box. EXCH_IP is the
Exchange
> servers "internal" IP address.
>
> Steve Cowles
> SWCowles at gte dot net
>
> <[EMAIL PROTECTED]> wrote in message
news:7r5lev$b9l$[EMAIL PROTECTED]...
> > Thanks for the info,
> > I finally figured out I needed to get Ipmasqadm installed and
finally
> > found it and installed it.
> > The $EXT_IP is the IP of the outside NIC on the firewall correct?
Or is
> > this a remote internet site?
>
>
Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.
------------------------------
From: [EMAIL PROTECTED]
Crossposted-To: comp.os.linux.setup
Subject: Re: icmp reply problem on home network
Date: Thu, 09 Sep 1999 14:48:37 GMT
James Hunt <[EMAIL PROTECTED]> wrote:
> Machine 1 (desktop: 192.168.1.1) running RH6.0 (2.2.2-15 kernel) with
Please include the output of "ifconfig" and "route" on the RH6 machine.
Maybe a problem with the netmask? Send the details... -Tom
Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.
------------------------------
From: "FEEB" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.x
Subject: Re: xdm while NCD to RedHat6.0 - no response
Date: Thu, 09 Sep 1999 09:54:01 -0400 (EDT)
Reply-To: "FEEB" <[EMAIL PROTECTED]>
On Thu, 09 Sep 1999 07:59:50 -0400 (EDT), FEEB wrote:
>On Wed, 08 Sep 1999 15:25:39 -0400, Jeffery Browning wrote:
>
>>FEEB wrote:
>>>
>>> I have an NCD terminal that connects regularly to HP-UX box. I can
>connect
>>> with it to all kinds of other *NIX boxes like SUN and SGI without
>problems
>>> just by changing the NCD's configuration and using 'bt' monitor
command.
>>>
>>> Now I am trying to connect to RedHat6.0. I can connect to tftpd, read
>in
>>> all software and fonts (no error messages), then the terminal screen
>>> becomes shaded with a little cross in the middle (it moves with the
>mouse)
>>> and that's it. No xdm, no login screen. I checked /var/log - no
errors.
>>> No errors on NCD either. I tried to completely relax /etc/pam.d/xdm in
>>> case the problem had something to do with the PAM. I read NCD-X
>Terminal
>>> mini HOWTO back and forth and searched dejanews for any clues.
Nothing.
>>>
>>> Could please someone help me out here?
>>>
>>> Thanks
>>>
>>> P.S. When I hit SETUP on NCD, I can telnet to the host.
>>> X works fine on the console (including xdm).
>
>>comment out the following line in /etc/X11/xdm/Xservers
>>
>>:0 local /usr/X11R6/bin/X
>
>Unfortunately, that did not work. But I upgraded to the newer version of
>the NCD software (3.3.2) and I am getting the CHOOSER screen with a menu
of
>about 20 *NIX boxes I can connect to EXCEPT the one I want to connect to.
>If I force the address of the RedHat6.0 box in CHOOSER I get:
>
>asking host "the.RedHat.box" for XDMCP session
>
>and after a while
>
>XDMCP error - login host not responding
>
>Any hints, please?
Further development:
The chooser mentioned above is not coming from the RedHat box. It is
generated by the NCD itself. But, if I manually start xdm on the RedHat
box, I get the connection and I can log in. The problem seems to be, that
RedHat6.0 does not run xdm. It runs something called 'prefdm' (xdm is
linked to prefdm in /etc/X11) and gdmgreeter.
So that's the reason, why RedHat5.x boxes respond to me and RedHat6.0 does
not. Now, does anyone know, where I can get any info on the above
mentioned prefdm and gdmgreeter? There are no manual pages for any such
things.
Thanks
Frank Bures, [EMAIL PROTECTED] (use this address for replies)
http://frank.chem.utoronto.ca/electronics
Warning: Received flame-mail will be reposted on the UseNet in full
------------------------------
From: midknite <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.setup
Subject: Re: list of network cards and drivers - where?
Date: Thu, 09 Sep 1999 09:48:28 -0400
Bert Douglas wrote:
>
> Hi all,
>
> Where can I find a list of NICs and the associated drivers for Linux?
most of the drivers for NICs that I know of for linux come with the
kernel.
kernels come from the higher beings <grin> and are found at
http://www.kernel.org
if you're looking for a list of NICs the kernel supports, the docs
that come with the kernel are probably the best source. a perhaps
slightly outdated list can be found in the howtos. check
http://www.linux.org in the support/howto section for howtos that look
ethernet or networking related.
--
brian kowolowski
gpg key / infos http://www.cryogen.com/midknite/gpg.html
gpg print: F6B6 076D 4BFC CD14 7C14 1A2F 61DA BDE5 7A88 D6C3
------------------------------
From: [EMAIL PROTECTED] (bill davidsen)
Subject: Re: ipchains ICMP forging?
Date: 9 Sep 1999 15:13:41 GMT
In article <7r65tc$ojt$[EMAIL PROTECTED]>, <[EMAIL PROTECTED]> wrote:
| That would only solve 1/2 of the problem. I *do* want to send the ICMP
| port unreachable message, but I want the source IPaddr on that packet to
| be the destination IPaddr of the rejected packet. e.g.:
|
| Packet 1: [IP/TCP]
| src: a.b.c.d srcport: xxx
| dst: e.f.g.h dstport: 23
| [data]
|
| Response packet from firewall host: [IP/ICMP]
| src: e.f.g.h
| dst: a.b.c.d
| [data]
|
| See? No packet lists the address of the firewall. This way, no one
| knows that system exists, and no one attacks it. STO is a bad thing, but
| I don't need to give them a running start either.
|
| I don't see anything in the ipchains docs about re-writing or forging
| the ICMP source, but I'm hoping that this is either a hidden feature or
| something on the drawing board.
There are several things you can do, starting by a look at the REDIRECT
target for inspiration. If that doesn't give you what you want, go to
ftp://ftp.tux.org/people/alexey-kuznetsov/ip-routing
and look at route2 and iputils.
--
bill davidsen <[EMAIL PROTECTED]> CTO, TMR Associates, Inc
"So let it be written, so let it be dumb." Pharaoh Dufus the last...
------------------------------
From: [EMAIL PROTECTED] (bill davidsen)
Subject: Re: IPChains and FTP
Date: 9 Sep 1999 13:20:24 GMT
In article <7qll1m$a3t$[EMAIL PROTECTED]>,
Cedric Blancher <[EMAIL PROTECTED]> wrote:
| Rick Orwig <[EMAIL PROTECTED]> a �crit dans le message :
| [EMAIL PROTECTED]
|
| > ipchains -P input ACCEPT
| > ipchains -P output ACCEPT
| > ipchains -P forward ACCEPT
|
| That's a silly set if rules for a firewall... Default policy _must_ be
| DENY for forward !
In a secure setup setup it doesn't matter, you never reach it. Every
chain looks like:
- things we allow
- things we disallow silently
- DENY and log everything else
Then you bash the logs and examine them daily.
--
bill davidsen <[EMAIL PROTECTED]> CTO, TMR Associates, Inc
"So let it be written, so let it be dumb." Pharaoh Dufus the last...
------------------------------
** 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
******************************
