Linux-Networking Digest #272, Volume #10         Sun, 21 Feb 99 14:13:41 EST

Contents:
  ppp dial-up connection/masquerade ("Marius Kaizerman")
  modem lock - more info (jack wallen)
  Re: pppd REJ's prot c029, should NAK it (Clifford Kite)
  Re: PPP failure on Slackware 3.4, Kernel 2.0.30 (Clifford Kite)
  Re: Apache proxy vs. Squid ("Carl R. Friend")
  Re: modem lock - more info (Henry)
  need help for pop3 and linux with win98 clients (Tobias Walkowiak)
  Re: Whoops! Subject _should_ have been multiple ISP connections (Andrzej Filip)
  Re: haow much user on a 100MBit Card ? (Eric Lee Green)
  Re: Linux and Microsoft Proxy Server 2.0 ("Carl R. Friend")
  Re: Multilink PPP in Linux with 2 x V90 = 105,333 bps? (Raphael Mankin)
  Kerneld ("Tim Champagne")
  Re: I give up with Linux and Win8 network (Eric Larson)
  Re: Telnet proxy (Juergen Heinzl)
  Re: Personal Netware & Linux (Mike Ching)
  Re: Driver for Madge Smart 16/4 AT Plus Ringnode? ([EMAIL PROTECTED])
  Re: Can I setup a standalone Apache Server?? ("JSH")
  Full duplex connection problem (Joewono Setiabudi)
  Re: network is unreachable ("Jeff")

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

From: "Marius Kaizerman" <[EMAIL PROTECTED]>
Subject: ppp dial-up connection/masquerade
Date: Sat, 20 Feb 1999 15:40:31 +0200

I'm using ip masquerading under linux rh5.0, accepting ppp connections
using 3 modems (kernel version is 2.0.36).

This is an error I received when a user logged in:

Feb 20 11:22:16 tty kernel: ip_masq_new(proto=TCP): no free ports.
Feb 20 11:23:33 tty kernel: ip_masq_new(proto=TCP): no free ports.
Feb 20 11:23:38 tty kernel: ip_masq_new(proto=TCP): no free ports.

Any idea what does it means and how to fix it ?

Thanks,
    Marius.






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

From: jack wallen <[EMAIL PROTECTED]>
Subject: modem lock - more info
Date: Sat, 20 Feb 1999 05:51:01 -0500

okay, i realized that in my original post it was lacking a lot of
information so here goes:

i'm running red hat 5.2.  i use afterstep and i'm having problems with,
every so often, my modem will lock.  it happens on disconnect and
there's no way i've found (other than powering down the machine) to
break the lock.

i have an internal rockwell 56k modem.   a compaq presario intel 200
mghz with 48 mb ram.  

i'm not sure what other info anyone would need. 

i've tried the following to kill this problem:  'echo ATHO > /dev/modem'
'killall -9 pppd', 'ip-down' (as root of course) but to no avail.

if anyone has any suggestions i'd certainly appreciate it!  you can mail
me direct.

thank you.
-- 
jack wallen,jr ICQ:20503940    \||/  LINUX -anything else...404!
                                00    
-against the run of the mill     >   -life in two dimensions
 swimming against the stream    -     is a mass production scheme

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

From: [EMAIL PROTECTED] (Clifford Kite)
Subject: Re: pppd REJ's prot c029, should NAK it
Date: 21 Feb 1999 11:07:25 -0600

Thanks for the update on the problem.

Looking back at your original post I *think* what happened was that
the ISP didn't like the IP address pppd asked for and Acked it to get
IPCP to the open state in order to close it down - and the connection.
I'd guess that this is "legal", but it's not likely to be obvious to
most of us when it's first encountered.

It's also something I'm not likely to forget either.  Thanks again.

--
Clifford Kite <[EMAIL PROTECTED]>                       Not a guru. (tm)
/* Better is the enemy of good enough. */

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

From: [EMAIL PROTECTED] (Clifford Kite)
Subject: Re: PPP failure on Slackware 3.4, Kernel 2.0.30
Date: 21 Feb 1999 11:34:03 -0600

Chris Small ([EMAIL PROTECTED]) wrote:

: I'm basically using the following parameters at the moment:

: lock
: defaultroute
: noipdefault
: modem
: /dev/ttyS3
: 19200   # Dropped down to a lower speed for testing.
: crtscts
: debug
: passive
: asyncmap 0 # also tried 0xa0000
: name "xxxxx"
: connect :chat -v -f /etc/ppp/cnetscript"
: ipcp-accept-local
: ipcp-accept-remote
: 0.0.0.0:<ip>.<ip>.<ip>.<ip>
: netmask 255.255.255.0

If you haven't already tried it, you might drop the last two pppd options
and see what happens.  You are trying to tell the ISP what address it
should use and you are assuming a netmask that may not be correct.

--
Clifford Kite <[EMAIL PROTECTED]>                       Not a guru. (tm)
/* Editing with vi is a lot better than using a huge swiss army knife. */

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

From: "Carl R. Friend" <[EMAIL PROTECTED]>
Subject: Re: Apache proxy vs. Squid
Date: Sat, 20 Feb 1999 10:24:24 -0500

[EMAIL PROTECTED] wrote:
> 
> What do you folks recommend [for proxies]...?

   Personally, I like Squid. But then again, I don't have a lot of
experience with Apache.

  Squid does do some really slick, stuff, though which I require at
work (like ICP - allowing multiple caches to interact with one
another).

   If you don't require that multiple caches be aware of each others'
contents, and are running Apache anyway, Apache may be a better fit.

-- 
 ______________________________________________________________________
|                                                |                     |
| Carl Richard Friend (UNIX Sysadmin)            | West Boylston       |
| Minicomputer Collector / Enthusiast            | Massachusetts, USA  |
| mailto:[EMAIL PROTECTED]                |                     |
| http://www.ultranet.com/~crfriend/museum       | ICBM: N42:22 W71:47 |
|________________________________________________|_____________________|

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

From: Henry <[EMAIL PROTECTED]>
Subject: Re: modem lock - more info
Date: Sat, 20 Feb 1999 10:37:38 -0500

I have a similar problem but with a Diamond SupraExpress.  Every once in
a while the dialing will timeout and then there is no more talking to
the modem until the machine is powered off and on.  

I don't know if this is related or just similar.  I was assuming that
the modem firmware/hardware was in a "stuck" state (i.e. a bug in the
modem) and therefore I had no hope of getting its attention again,
without a power cycle.

Henry

jack wallen wrote:
> 
> okay, i realized that in my original post it was lacking a lot of
> information so here goes:
> 
> i'm running red hat 5.2.  i use afterstep and i'm having problems with,
> every so often, my modem will lock.  it happens on disconnect and
> there's no way i've found (other than powering down the machine) to
> break the lock.

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

From: Tobias Walkowiak <[EMAIL PROTECTED]>
Subject: need help for pop3 and linux with win98 clients
Date: 20 Feb 1999 16:38:12 +0100


hi.

my configuration:
* linux [suse 6.0] server running samba
* windoze98 clients
* a ppp account and some pop3 accounts

now i connect to the ppp account at Provider1 with the linux comp and get
the mail from the pop3 accounts at Provider2. fetchmail does the job
and distributes the mail correctly to the appropriate users on the linux box.

questions:
1. can i read the mail at the windoze clients even when the mail clients
   only let the mail server be a pop3 server? there are only the settings
   for an smtp server as *outgoing* mail server.
2. what are the correct header settings when i send mail using the different
   remote pop3 accounts? is the sender then [EMAIL PROTECTED] [s. a.]
   or must it be [EMAIL PROTECTED]?
3. will the whole thing then work with 'sendmail -q' sending outgoing, queued
   mail to Provider1?
4. how can i configure ppp/isdn such that at a connection request from
   a windoze client [esp. using the www] the connection will be established
   automatically?

answering any of these questions will be very appreciatied :)

thanx in advance
walko
-- 
-- [EMAIL PROTECTED]       http://www.cadlab.tu-berlin.de/~walko/ --
-- pgp key available via finger [EMAIL PROTECTED]          --
-- pgp key fingerprint = 91 96 22 DA 5E 5B 6B 98  97 59 0C BA D2 06 EF 50 --
--

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

From: Andrzej Filip <[EMAIL PROTECTED]>
Subject: Re: Whoops! Subject _should_ have been multiple ISP connections
Date: Sat, 20 Feb 1999 16:40:30 +0100

Evan Wolenzik wrote:

> Ok, I'm _sure_ this one's been gone over before, but it never pertained to
> me, so I ignored it. Please forgive me for rehashing this topic:
>
> Let's say I have a LAN with 3 boxes, A, B and C.  A is currently the
> default gateway to the Internet, with a 24-7 PPP connection to a provider.
>
> Now, let's say I want greater bandwidth, and I also want some redundancy,
> so rather than upgrading my connection with my existing provider, I want
> to have a second account with a different provider. Can I have this second
> account / connection on machine _C_, and have it also be a gateway? How
> will packets know that they can use either A or C to access the Internet?
>
> Or, would the second connection have to be on A also?

It can be on A BUT
If you have one connection on A with default route via ISP_A
and another one on B with default route via ISP_B and
A & B use ip_masqueraading THAN
you may configure any computer on your network to connect
to any IP address via ISP of its choice

e.g.

computer D has default route via A (ISP_A) and
only addresses from ISP_B address space are
accessed via B

and at the same moment

computer E has default route via B (ISP_B) and
only addresses from ISP_A address space are
accessed via A

If you are brave and determined you may use gated
daemons on A & B to make failures of one of the lines
almost invisible to computers on your net
(computers receive  routing tables from gateds
on A & B, gateds detect links failures)

P.S.
If both providers use squid (www cache)
you have even more possibilities to
explore. Squids  exchange data how "close"
www servers are so your squid will receive
www pages via squid of the ISP "closer"
to the page.
--
"Andrzej (Andrew) A. Filip" <[EMAIL PROTECTED]>
Warsaw, Poland
[EMAIL PROTECTED] (backup)



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

From: [EMAIL PROTECTED] (Eric Lee Green)
Subject: Re: haow much user on a 100MBit Card ?
Date: 21 Feb 1999 17:53:38 GMT
Reply-To: [EMAIL PROTECTED]

On Sat, 20 Feb 1999 20:54:26 +0100, Bernd Broermann <[EMAIL PROTECTED]> wrote:
>How much user can work on a 100 Mbit network interface on one Server.
>They only connect over telnet session in 
>a CSMA/CD 100 MBit Network (for example 192.168.1.0) .

The Linux 2.0.36 kernel is limited to 255 telnet sessions. Actually, fewer,
since a few PTY's are used internally. 

The 2.2 kernel eliminates that restriction.

I have seen 50 users telnet'ed into a Pentium 120 with 32mb of memory
and a 3gb IDE hard drive over a 10base-T card doing character-based
data entry into a database. I do not think that you'll have any
trouble doing 500 users with modern hardware, assuming that you give
enough memory so that the machine is not swapping all the time. (Note
that not all of those users were necessarily working, but they were
all logged in). Walnut Creek CD-ROM regularly has hundreds of users
logged into their FTP server simultaneously, and these users are all
doing FTP downloads -- much harder on network bandwidth than
interactive telnet sessions.

--
Eric Lee Green         [EMAIL PROTECTED]     http://www.linux-hw.com/~eric
  "Microsoft will compete ... by adding features" -- Ed Muth, Microsoft

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

From: "Carl R. Friend" <[EMAIL PROTECTED]>
Subject: Re: Linux and Microsoft Proxy Server 2.0
Date: Sat, 20 Feb 1999 10:40:04 -0500

Tony Melendez wrote:
> 
> Greetings,
>     I have been trying to get Netscape Communicator under Linux to
> use MS Proxy 2.0 to connect to the internet. However, I don't seem
> to be having much luck.

   This symptom exhibits itself with _all_ browsers save Internet
Exploder. M$ buggered with the HTTP standard and produced a bogus
header which is used in their proprietary authentication scheme. This
is to be expected - see the "Halloween Documents".

   The way around this is to disable authentication on the Proxy
server (it's on by default). This _should_ get rid of the bogus
header. (The proxy passes back the "access denied" page by default
with the "Authentication: NTLM" header. This can actually make some
browsers dump core!)

   Of course, the other way to solve the problem - permanently - is
to scrap M$ Proxy and run Squid or some other standards-compliant
proxy system. Generally, performance of the M$ proxy will be sub-
standard for all browsers other than IE (it has to do with the way
that MS implemented cache coherency).

-- 
 ______________________________________________________________________
|                                                |                     |
| Carl Richard Friend (UNIX Sysadmin)            | West Boylston       |
| Minicomputer Collector / Enthusiast            | Massachusetts, USA  |
| mailto:[EMAIL PROTECTED]                |                     |
| http://www.ultranet.com/~crfriend/museum       | ICBM: N42:22 W71:47 |
|________________________________________________|_____________________|

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

From: Raphael Mankin <[EMAIL PROTECTED]>
Crossposted-To: 
comp.os.linux.misc,comp.os.linux.questions,comp.os.linux.x,uk.comp.os.linux
Subject: Re: Multilink PPP in Linux with 2 x V90 = 105,333 bps?
Date: 21 Feb 1999 14:30:27 GMT

In comp.os.linux.misc Jason Clifford <[EMAIL PROTECTED]> wrote:
: On Fri, 19 Feb 1999, Doodle wrote:

:> I know Win95/98 has support for Multilink PPP to 'channel bond' two
:> V90 modems to gain something like 10k/sec connection speeds. Can I use
:> two V90s in Linux to get 10k/sec connection speeds?

: Yes. You need to build support for this into the kernel and read the
: documentation that accompanies the kernel with regard to this facility.

And before this you will also need to speak to your ISP to make
sure that they will support this.

-- 
==================================
                                        2 + 2 = 5 (for large values of 2)
Raphael Mankin
E-Mail: [EMAIL PROTECTED]
==================================

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

From: "Tim Champagne" <[EMAIL PROTECTED]>
Subject: Kerneld
Date: Sat, 20 Feb 1999 11:12:01 -0500

I am setting up a Linux box with Red Hat 5.2. I am setting it up similar to
an existing one. My question involves PPP and kerneld. On my old system I
have a file request-route which kicks off my ppp-on script whenever an
unknown address is recognized (this is done by kerneld I think). I have done
the same thing on the new box but kerneld does not seem to kick it off.
Whenever I try to ping an address like www.xxx.com I get an immediate
response indicating that address not found. When I do this on the old system
it waits until the ppp connection is established and then finally responds
to my ping request. Does anybody know anything about this?

Thanks,
-Tim



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

From: [EMAIL PROTECTED] (Eric Larson)
Subject: Re: I give up with Linux and Win8 network
Date: Sun, 21 Feb 1999 16:59:27 GMT
Reply-To: [EMAIL PROTECTED]

The best web resource for IP Masquerading BY FAR is:

http://rlz.ne.mediaone.net/linux/faq/index.html.

This includes a detailed FAQ and an online script generation tool that
will result in your being up on the net quickly. It's aimed at
cable/ASDL users so you should have no problems.

For dialup users like me you have to hack the scripts a bit, but we
are working on it....

On Sun, 21 Feb 1999 00:56:40 GMT, "Jose" <[EMAIL PROTECTED]> wrote:

>Okay, I have a cable modem, 2 nics in a llinux box, one in my Win98
>machine.. I've read the firewall How-Tos, I looked a probably a dozen sites,
>and I haven't been able fully understand how to get my win98 machine on the
>web routing thru my linux box. I'm done a fresh install like 7 times, I know
>I have ifwadm 2.3.0 compiled in the kernel, but I still haven't been able to
>get this to work.
>
>I have Linux unleashed 2nd and third edition, I have printed out the NET-3
>Howto, Firewalling and Proxy Howto and I'm still lost..
>
>It's funny becuase I can ping the nic from my win98 machine. but I can't get
>out.. I'm really starting to go back to NT for my network.. Can some one
>help me out, there's got to be like a IP masquerading for dummies or
>something cause damn this is really kick my but.. And the reason I want
>linux is for Hosting my web site and other issues..
>
>TIA
>Jose
>
>


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

From: [EMAIL PROTECTED] (Juergen Heinzl)
Subject: Re: Telnet proxy
Date: Sun, 21 Feb 1999 18:26:00 GMT

In article <[EMAIL PROTECTED]>, Roman Spitzbart wrote:
>For what??????? Telnet is interactive, what should the proxy cache?????

Try to telnet from 192.168.21.1 -> wherever ... no, is not going
to work since this is an address not to be routed to the outside world and
if it does than your routing is pants 8)

Can't masquerading do the job ? Cannot tell, I do not need it.

Cheers,
Juergen

-- 
\ Real name     : J�rgen Heinzl                 \       no flames      /
 \ EMail Private : [EMAIL PROTECTED] \ send money instead /
  \ Phone Private : +44 181-332 0750              \                  /

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

From: [EMAIL PROTECTED] (Mike Ching)
Subject: Re: Personal Netware & Linux
Reply-To: [EMAIL PROTECTED]
Date: Sat, 20 Feb 1999 16:17:35 GMT

On Fri, 19 Feb 1999 18:36:25 GMT, Holger Petersen <[EMAIL PROTECTED]> wrote:
>
>
>Is there any way to have a marriage between those two ?
>
>asking, Holger

Does the Caldera package have what you're looking for?

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

From: [EMAIL PROTECTED]
Crossposted-To: comp.os.linux.hardware
Subject: Re: Driver for Madge Smart 16/4 AT Plus Ringnode?
Date: 20 Feb 1999 16:19:02 GMT

In comp.os.linux.hardware Stephan Loescher <[EMAIL PROTECTED]> wrote:
> Does Linux support the "Madge Smart 16/4 AT Plus Ringnode"-token ring
> adapter?

I called Madge on the phone last summer and they had just developed a beta
Linux driver.  They uuencoded it and mailed it to me, and it worked fine.

Craig Ludington

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

From: "JSH" <[EMAIL PROTECTED]>
Subject: Re: Can I setup a standalone Apache Server??
Date: Sat, 20 Feb 1999 10:52:50 -0500

You do not need to register a domain name to run Apache on your computer.
You do not need to be connected to the Internet to run Apache, though if you
aren't connected to the Internet, the only way you can access the server is
from your own computer or from a computer on your LAN.

 If you are planning to use the server ONLY for yourself on the same
computer, do the following:

1.) Make sure your loopback interface (lo) is enabled by typing ifconfig. If
there is no "lo" listed, type ifconfig lo 127.0.0.1
2.) What domain names do you want to use? Place them in /etc/hosts as
aliases after 127.0.0.1. For example, if you wanted to use the domain name
"duckytown" (like I do on my computer), your entry would look like this:

127.0.0.1    localhost duckytown

Leave "localhost" as the first entry, though.

3.) Set up VirtualHost(s) in the /var/apache/etc/httpd.conf file. The
instructions are somewhere on the bottom. In my case, the ServerName of my
VirtualHost would be "duckytown."

This is the simplest way to do what you are asking... it gets more
complicated when you want your server to be accessible through a LAN or if
you are running DNS.

Dominik Slusarczyk
[EMAIL PROTECTED]

Qiang Wan wrote in message ...
>Hi,there:
> I am a beginner to Linux network setup and I tried to setup an Apache
>web server on my PC. Is that possible for me to run an standalone Apache
>as a server for Netscape in my PC without connecting to internet. Do I had
to
>get a IP address and domain name for my PC in order to run Apache. Is there
any way that I can
>emulate an "fake" IP address just for stand alone use within a PC. Or if I
had
>to get a IP address, how can I get a IP and what is the cost?
> Thanks in advance for your help!
>
>
>Qiang
>



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

Date: Sun, 21 Feb 1999 13:58:12 -0500
From: Joewono Setiabudi <[EMAIL PROTECTED]>
Crossposted-To: alt.os.linux
Subject: Full duplex connection problem

I have 2 computers,  A (windows) and B (linux) that I want to connect in
full duplex
with crossover cable of course.

Whenever I am trying to telnet from A to B, B responded so slowly. The
telnet prompt
appear after about 1/2-1 minutes. This also happen to the ftp service.
But, once I got connected to B, the connection is fast (normal).

With the same configuration (ethernet, IP , etc) this case never happen
when I connect A and B
through hub (half duplex).

Anyone know what is wrong with this ?
Is there any setting that I need to set ???

FYI both cards are the same, NetGear FA310Tx, linux is using tulip
driver.
Also, there were no collision when I check it.

Thanks



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

From: "Jeff" <[EMAIL PROTECTED]>
Subject: Re: network is unreachable
Date: Sun, 21 Feb 1999 14:00:44 -0500

Do a route -n
and an ifconifg
post the results..
sounds like you don;t have any route set up..
or the hardware is not running
try ping 127.0.0.1

Jeff
* wrote in message <7HVz2.2021$[EMAIL PROTECTED]>...
>Im trying to set up a network between my SUSE 6.0 box and a win 98 box.
When
>i�m trying to ping the win98 box it says: network is unreachable. I can�t
>ping my self either.
>
>Please help me .. im new to linux ...
>
>
>
>



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


** 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