Linux-Networking Digest #863, Volume #9          Wed, 13 Jan 99 00:13:35 EST

Contents:
  Re: /etc/services (Villy Kruse)
  Re: Routing using RedHat 5.2 ("Jaspreet Singh")
  Can anyone suggest a good ATM card for Linux? (Ben Greear)
  Re: PLEASE HELP !!! PLEASE HELP !!! (Clueless Newbie)
  Re: Winmodem (Chris Fields)
  Re: ppp-2.3.5 compile problem (Clifford Kite)
  Re: Help with DNS Virtual Hosting (Alex Kamantauskas)
  Re: /etc/services (Stef)
  Re: /etc/services (Stef)
  Re: tunneling over masquerading (Vincent Zweije)
  Re: Netgear FA310TX, new tulip.c, still doesn't work (David Kaczynski)
  anyone use UUnet for ISDN w/ Unix? (Dave Klingler)
  Re: BellAtlantic.net and PPP 2.3.5 ("Hoyt")
  Re: HELP: Setting up a DIAL-IN PPP SERVER on my Linux box?? (Bill Unruh)
  Re: UUCP over TCP - no echo? (bill davidsen)

----------------------------------------------------------------------------

From: [EMAIL PROTECTED] (Villy Kruse)
Subject: Re: /etc/services
Date: 12 Jan 1999 18:13:32 +0100

In article <77fjj8$qid$[EMAIL PROTECTED]>,
 <[EMAIL PROTECTED]> wrote:
>Hey all,
>quick question...
>
>I'm playing around with the port addresses of various services...
>And I can't get HTTP to be anything but 80.
>Even if I change it to something like 7080, it only exists on 80!?!
>
>I'm sure I'm missing something stupid. But I kill -HUP inetd and I even tried
>rebooting to make sure. The file /etc/services still retains my changes, is
>there somewhere else I need to play around?
>


Not all server daemons get the port number from /retc/services.  For example
the apache http server define the server port in httpd.conf as 80 and that
is where it mey be re-defined to some other value.  



Villy

------------------------------

From: "Jaspreet Singh" <[EMAIL PROTECTED]>
Subject: Re: Routing using RedHat 5.2
Date: 12 Jan 1999 17:13:20 GMT



> I think you need to enable it at runtime as well.
> 
> Try something like:
> 
>  echo 1 > /proc/sys/net/ipv4/ip_forward
> 
> This should go/ should be in your /etc/rc.d/rc2.d/S10network file.
> 
> Ben
> 

Or Try something like:

edit file /etc/sysconfig/network (should exist in 5.2) and set
FORWARD_IPV4=true



------------------------------

From: Ben Greear <[EMAIL PROTECTED]>
Subject: Can anyone suggest a good ATM card for Linux?
Date: Tue, 12 Jan 1999 19:58:09 -0700

Looking for an OC3 NIC for Linux (PC).  I will be needing to hack the
ATM driver in all likely hood, so it is necessary that
it is open-source.

Can anyone suggest one?

(Please CC [EMAIL PROTECTED] so I can read it from work.)

Thanks,

-- 
Ben Greear ([EMAIL PROTECTED])  http://www.primenet.com/~greear 
Author of ScryMUD:  mud.primenet.com 4444
http://www.primenet.com/~greear/ScryMUD/scry.html

------------------------------

From: Clueless Newbie <[EMAIL PROTECTED]>
Crossposted-To: 
linux.redhat.install,comp.os.linux.questions,comp.os.linux.hardware,comp.os.linux.setup
Subject: Re: PLEASE HELP !!! PLEASE HELP !!!
Date: Tue, 12 Jan 1999 12:02:25 -0600

Clueless Newbie wrote:

>
> I know this is kind of a dumb question, but I am viewing the full header and
> I can't seem to extract anything quite that precise.
> is there some trick to decoding what I listed in the header to make it more
> meaningful?
>
> Thanks in advance for anylight that can be shed on this for me.

I think that I've answered my own question.

It does raise another question though? How do I keep my current IP address from
being included in the header?
I don't really mind if someone knows where to send complaints to, but it does
bother me that anyone has access to
the IP address that I'm currently on.

If it makes a difference on how to hide my IP  I am using  Redhat 5.2 and using
the Netscape News reader.

Thanks for taking the time to answer this, and I apologize if this posting is a
little off topic.


------------------------------

From: Chris Fields <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.questions
Subject: Re: Winmodem
Date: Tue, 12 Jan 1999 11:37:40 -0600
Reply-To: "cjfields @jove.acs.unt.edu" <spamblocker>

I had a US Robotics Winmodem when I used Windows 95.  The problem with
winmodems is that they lack a chip used for data compression; the
(Windows-only, proprietary)software provided with the modem provides the
compression instead.  As you would probably guess by now, any software
that runs only on Windows will NOT run on LINUX unless you're running
Wine (and I'm not even sure about that one).  Even if the compression
software ran on WINE, why would you want to go through something like
that, when many of the 56K modems are less than $100?  Your best bet is
investing in a modem that is NOT a Winmodem (make sure the package
specifically states that it can run on other operating systems (if it
can run on a Mac or DOS, it can most likely work with your system as
well); especially look for ones that contain a Flash BIOS chip for
upgrading the modem's internal BIOS).  I changed over to a ZOOM 2919
Dual-Mode FaxModem (internal, NOT external), which all three operating
systems I have (LINUX, 95, NT) will detect, with some tinkering of the
jumper settings for the COM ports.

Bounty Hunter wrote:
> 
> On Thu, 29 Oct 1998 11:24:00 -0500, Shenzhi Yang <[EMAIL PROTECTED]>
> wrote:
> 
> >  I got a Winmodem in my computer. So, is there anyway to make in
> >working in Linux? If not, any recommendation for a modem that can work
> >in Linux?
> >  Thanks for any response.
> 
> I am looking right now at a package called isapnptools.
> It says that it will work with the newer win-type modems.
> Be advised though, some hacking is required! This package will help
> you get started though.
> 
> Steve

-- 
C. J. Fields
Graduate Student, Dept. of Biological Sciences
The University of North Texas
Denton, TX 

email : [EMAIL PROTECTED]
\\\\\\\\\\\\\\\\\\\\oooooooooooooooo////////////////////
"Giving money and power to government is like giving
 whiskey and car keys to teenage boys"
                                -P. J. O'Rourke
"Join the military.  Travel to exotic places, meet 
exciting people, then kill them"
                                -Anonymous
////////////////////oooooooooooooooo\\\\\\\\\\\\\\\\\\\\

------------------------------

From: [EMAIL PROTECTED] (Clifford Kite)
Subject: Re: ppp-2.3.5 compile problem
Date: 12 Jan 1999 11:52:37 -0600

Carl Koeppl ([EMAIL PROTECTED]) wrote:

:   I just ran across this problem. You need to copy ppp.c from the ppp
: source to .../drivers/net. Then recompile.

I could be wrong but I think this won't work since these error messages
don't fit with the suggested solution.


--
Clifford Kite <[EMAIL PROTECTED]>                       Not a guru. (tm)



------------------------------

From: [EMAIL PROTECTED] (Alex Kamantauskas)
Subject: Re: Help with DNS Virtual Hosting
Date: 12 Jan 1999 17:00:40 GMT

On Sat, 09 Jan 1999 13:13:24 -0500, Stephen <[EMAIL PROTECTED]> wrote:

> I have read books and learned all about how DNS works but I can't find a
> good example on Virtual Hosting with one IP address.
> I'm assuming the files "named.boot" and "named.conf" are the files I
> need to focus on.

I think its a web configuration issue.  All that DNS can do is assign the
hosts to an IP address.  It won't be able to determing port addressing or
anything like that.

> I have over 10 virtual hosts on one IP address.

Are they all on the same domain name, or separate domain names?

If all on the same domain name, then you just need to create one zone file
(for the domain name), with the following host records:

        foo.com.        IN A            10.0.0.1
        host1           IN CNAME        foo.com.
        host2           IN CNAME        foo.com.
        host3           IN CNAME        foo.com.

That will map host1.foo.com to point to 10.0.0.1, host2.foo.com to
10.0.0.1, etc.

If they are all separate domain names, then create a zone file for each
one and have an A record pointing to the IP address.

Then configure your web server (if Apache 1.3.3+) like such for each
virtual host:

        NameVirtualHost ##.##.##.##
        <VirtualHost host.some_domain.com>
        ServerAdmin [EMAIL PROTECTED]
        DocumentRoot /www/docs/host.some_domain.com
        ServerName host.some_domain.com
        ErrorLog logs/host.some_domain.com-error_log
        TransferLog logs/host.some_domain.com-access_log
        </VirtualHost>


>  Anyone willing to help me out on this is a life saver. I just want one
> good example and I'm a happy man!
> 
> Thanks in advance.
> 
> email:[EMAIL PROTECTED]
> 


-- 
Alex Kamantauskas
Tugger Networks

------------------------------

From: Stef <[EMAIL PROTECTED]>
Subject: Re: /etc/services
Date: 12 Jan 1999 17:11:11 +0100

: I'm playing around with the port addresses of various services...
: And I can't get HTTP to be anything but 80.
: Even if I change it to something like 7080, it only exists on 80!?!

: I'm sure I'm missing something stupid. But I kill -HUP inetd and I even tried
: rebooting to make sure. The file /etc/services still retains my changes, is
: there somewhere else I need to play around?

You have to change the configuration file of your httpd (what are you
using?). For apache under Debian 2.0 this is
/etc/apache/httpd.conf
There you have to find the line saying
Port 80
change it to whatever you desire, and then do a
/etc/init.d/apache reload

Your mileage may vary.

Stef
-- 
WebMaster D-WERK
UNIX and Windows NT administration, SOS-ETH 
ETH Zurich
[EMAIL PROTECTED]        http://hoes.li

------------------------------

From: Stef <[EMAIL PROTECTED]>
Subject: Re: /etc/services
Date: 12 Jan 1999 17:15:06 +0100

: I'm sure I'm missing something stupid. But I kill -HUP inetd and I even tried
: rebooting to make sure. The file /etc/services still retains my changes, is
: there somewhere else I need to play around?

P.S. /etc/services only maps service names and their port numbers, so
you can do a
telnet host http
instead of
telnet host 80
and don't have to remeber all the portnumbers for each
service. /etc/services does however not controll the ports your daemons are
using.

Stef
-- 
WebMaster D-WERK
UNIX and Windows NT administration, SOS-ETH 
ETH Zurich
[EMAIL PROTECTED]        http://hoes.li

------------------------------

From: Vincent Zweije <[EMAIL PROTECTED]>
Subject: Re: tunneling over masquerading
Date: 12 Jan 1999 13:45:10 +0100

In article <77cqoh$m6c$[EMAIL PROTECTED]>, Eric Kluft <[EMAIL PROTECTED]>
wrote:

||  My provider gives me 1 ip-number. They don't have the abillity to route
||  more ip-numbers to me.

Nonsense, of course; they don't want to go through the trouble to get
it working, or you must pay.

||                         A friend of mine, however, is working at another
||  ISP. By tunneling (linux IPIP) I route 32 ip-numbers to me from that ISP. I
||  just route all computers on my local network over the tunnel to the
||  internet.  (/sbin/route add default gw w.x.y.z tunl0)
||  So far so good.
||  The problem is that the tunnel uses 8 extra hops so the connections get
||  slower. The solution to this problem is to masquerade the workstations and
||  to tunnel the servers.
||  I can't do this on 1 server (because i must choose to route the default
||  gateway over the tunnel or ethernet interface (can't use both)).
||  If I use 2 computers, 1 computer is connected to the internet and
||  masquerades all computers behind it. The second computer contains the
||  tunnel. The source address of this tunnel server should be rewritten by the
||  masquerading server.

Although incoming traffic *to* the servers must come over the tunnel (or
you won't be reachable), *outgoing* traffic doesn't have to.  That is,
unless your provider is filtering out traffic that doesn't have your
address as source address.

The only traffic that must go out through the tunnel is to addresses that
are local to your friend (and otherwise unroutable).  From your story,
I induce this is not the case.

Now you only have to determine which packets to masquerade.

I suspect that for TCP connections, it's automatic; any connection
initiated from within your network will be masqueraded.  Others coming
from outside through the tunnel are left alone.

A problem might be connectionless services (UDP? ICMP?), where the
"initiating" host is not necessarily a defined thing.  I think a solution
is to use *two* addresses on each server.  It must use one address
(the offcial one coming in through the tunnel) for traffic that must
not be masqueraded.  It must use another, unofficial/reserved, one for
traffic that must be masqueraded.  I believe that named, for instance,
can do this.

On the firewall, you can then masquerade traffic according to IP source
address.  Pass through the official addresses and masquerade the rest.
Remember that you don't have to tunnel outgoing traffic.

||  This last piece however is not possible. Tunneling uses RAW ip for it's
||  frames and you can't masquerade RAW ip.
||  I know this is the case for standard 2.0.35 kernels, but maybe there's a
||  solution for it. The solution as offered for microsoft PPTP does not work.
||  Does anyone know if ipchains has the abillity to masquerade raw ip?
||  Does anyone know if NAT has the abillity to masquerade raw ip?
||  Does anyone know a different solution?

I don't think raw IP *can* be masqueraded.  There must be some way to
associate returned packets with previously outgoing masqueraded packets,
in order to de-masquerade them, and in raw IP this way doesn't exist.

Some ideas if you insist on tunneling outgoing traffic:

A different solution might be to tunnel PPP over TCP (ssh).  Gives the
added benefit of a secure tunnel.

Another solution: instead of masquerading the tunnel as you tried,
*tunnel* the tunnel!  From a local server, tunnel to a reserved address
at your friend's.  On the firewall, again tunnel this reserved address
to your friend.

Hope this helps.                                                 Vincent.
-- 
Vincent Zweije <[EMAIL PROTECTED]>    | "If you're flamed in a group you
<http://www.xs4all.nl/~zweije/>      | don't read, does anybody get burnt?"
[Xhost should be taken out and shot] |            -- Paul Tomblin on a.s.r.

------------------------------

From: [EMAIL PROTECTED] (David Kaczynski)
Subject: Re: Netgear FA310TX, new tulip.c, still doesn't work
Date: Tue, 12 Jan 1999 18:31:22 GMT
Reply-To: /dev/null

On Tue, 12 Jan 1999 12:51:02 GMT, [EMAIL PROTECTED] (Bob Koss) wrote:

[snip]

>Can't PCI cards share an interrupt?

In theory, yes.

>What I'm getting on boot-up is, "device or resource busy."  Okay, busy
>doing what and how do I unbusy it?????

Like I said, check which devices are using what IRQs and if the nic
card is sharing an IRQ with something else, try giving it its own IRQ
if possible.  And if that corrects the problem, there you have it.


------------------------------

From: [EMAIL PROTECTED] (Dave Klingler)
Crossposted-To: comp.dcom.isdn
Subject: anyone use UUnet for ISDN w/ Unix?
Date: 12 Jan 1999 19:22:58 -0700

Hi everyone.  I've purchased an unlimited ISDN dialup account from UUNet,
but so far their tech support seems abysmal, i.e., they only do Windows
and they can't give me any technical information.  I'm a little surprised
to get this level of expertise from a great-grandaddy of the net like
UUnet.

Basically, I've got a RedHat 5.2 machine that I'm trying to make talk to
their POP through a 3com Impact IQ modem.  I think I've verified that the
modem is okay, but I still haven't managed to make the two ends carry on a
conversation.

I either need to find a new ISP or fill in the technical details myself.
Does anyone else use UUnet for ISDN with a relatively public-domain
version of PPP?  Does anyone know whether UUnet uses sync or async PPP?
Whether they allow non-chap-authenticated connections?

Thanks!
Dave Klingler

------------------------------

From: "Hoyt" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.setup,alt.uu.comp.os.linux.questions,alt.os.linux.setup
Subject: Re: BellAtlantic.net and PPP 2.3.5
Date: 12 Jan 1999 18:31:57 GMT


[EMAIL PROTECTED] wrote in message <[EMAIL PROTECTED]>...
>Frank Hale writes:
>> Anyway the whole connection process is wrapped up in 1 perl script which
>> can hardly be called complicated.
>
>They tell you that you must recompile your kernel, download and compile an
>obsolete version of pppd, and then type in and run a perl script.  A good
>way to scare off new users.

I contacted Worldnet and informed them that the version of ppp was not only
dated, but the link was broken (the file no longer exists in that location).
Their reply was on the order of "The guy who wrote that stuff isn't here any
more."


>
>> All you have to do is plug in your user data.
>
>You can do that with Debian (and, I believe, Red Hat) without their script.
>Just run pppconfig and answer the questions.  All AT&T needs to do is say
>that they use CHAP, explain how to figure out the username and password,
>and tell you the nameserver numbers.to this address.

I was able to use kppp in KDE (RH 5.2, KDE 1.0) to configure my dial-up
without any help from Worldnet (or messing with any ppp scripts). I actually
had more problems because the RH distribution I had did not have ppp
installed in the kernel by default and, as a newbie, it took me a while to
figure it out. Works fine now.

It was helpful to know that Worldnet uses CHAP, but that's all the help the
Worldnet site gave me. BellAtlantic.net uses PAP.

I plan to leave Worldnet because BeOS does not support CHAP, only PAP and
Worldnet doesn't support PAP (apparently some of their POP's used to support
PAP, but not after the latest upgrades they made). I am considering
BellAtlantic but don't think much of their usenet support (limited groups,
articles only cached for a few days), but the have a 30-day money back deal,
so I may give it a shot. There is also a local ISP that looks good, but
again, their usenet support is weak - no money back deal there. Worldnet has
a rep for having great usenet support (lots of groups and the posts stay on
there a long time). I also get consistent 46.333 connects with them, little
down time and for about 10 months, no busy signals. I hate to leave them.

So, we shall see

If anyone can port ppp to BeOS, it would be great as Be has official plans
to _never_ support CHAP - they say that third-party developers can implement
it if they desire.

Hoyt



------------------------------

From: [EMAIL PROTECTED] (Bill Unruh)
Crossposted-To: 
alt.os.linux,comp.os.linux,comp.os.linux.help,comp.os.linux.admin,comp.os.linux.misc,comp.os.linux.questions,comp.os.linux.setup,alt.linux
Subject: Re: HELP: Setting up a DIAL-IN PPP SERVER on my Linux box??
Date: 12 Jan 1999 18:35:47 GMT

In <[EMAIL PROTECTED]> [EMAIL PROTECTED] writes:

>I know others have probably asked, but I'm not sure so I'll
>ask.....Does anyone know how to set up a Dial-In PPP Server on a Linux
>Machine so i can dial into it from afar and get tcp/ip routing to work
>with it??? Sort of like having a single user ISP going on, ya know? If
>anyone can help, thanks in advance, if not......ask someone who might
>know AND thanks in advance...or something

a) Get mgetty and install it and configure /etc/inittab to run mgetty on
the serial port.
S1:345:respawn:/sbin/mgetty  ttyS1  
(my modem is on Com2 or ttyS1)
b) Set up mgetty in the files /etc/mgetty+sendfax/*

c) Make SURE IP forwarding is turned on in your system ( On Redhat 5.x,
make sure that 
NETWORKING=yes
FORWARD_IPV4="yes"
in /etc/sysconfig/network

d) put the options in the /etc/ppp/options file
eg
nodetach
noauth
proxyarp
idle 1800


noauth means the outside user is to use standard login authentication,
not PAP or CHAP, proxyarp is so that other machines know that yours is
where to send packets to for the remote machine, idle 1800 is to
disconnect after 1800 sec of no activity.
(Do not run AutoPPP in mgetty if you are using noauth)
(If you want AutoPPP-- ie ppp starts up on receipt of a ppp negotiation
packet from the remote end immediately after a connection is made-- then
you must set up PAP or CHAP authentication)

------------------------------

From: [EMAIL PROTECTED] (bill davidsen)
Subject: Re: UUCP over TCP - no echo?
Date: 13 Jan 1999 04:55:00 GMT

In article <[EMAIL PROTECTED]>,
Edgar J. Zaragoza <[EMAIL PROTECTED]> wrote:

| I've set up UUCP over TCP using all the info I could find, and I am
| running into a strange problem. We've got a PC running Red Hat 5.1,
| 3com network board, network is uo and running. We can FTP, ping, rlogin
| - no problem. UUCP doesn't work. 
| 
|       According to the Debug file, we see and respond to the login prompt, we see and
| respond to the password prompt, but we still get
| login incorrect. Now before you say I typed the password in wrong,
| If I cu to the machine, and type in my login, NO ECHO. hit return,
| passwd prompt, type in passwd. waits.....login incorrect.
| 
|       Rlogin with same user name and passwd works fine.

If you are logging into a Taylor UUCP via TCP be aware that the password
comes not from the /etc/passwd or /etc/shadow file but the UUCp internal
password file. Sorry I can't give you a path off the top of my head, but
it is a separate file and has no relation to the rlogin password.

-- 
  bill davidsen <[EMAIL PROTECTED]>  CTO, TMR Associates, Inc
"Too soon we grow old, and too late we grow smart" -Arthur Godfrey


------------------------------


** 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
******************************

Reply via email to