Linux-Networking Digest #341, Volume #10 Mon, 1 Mar 99 08:13:39 EST
Contents:
Re: How to telnet as root (Richard Steiner)
Re: Ethernet for IBM ps/2 386 machine ("Moshe Bar")
Ethernet Problems (Paul Lambert)
Re: ipautofw ("Alain Coetmeur")
Re: NFS Export Questions (JK Dingwall)
a free software dynamic Firewall gateway for MICO & alike ? ("Alain Coetmeur")
Init Id * respawn error (Wazza)
Fetchmail 4.6.4 and SMTP ("laujon")
Configuration Conflicts between Dialup Adapter and Ethernet Card ("Vasilios E.
Tourloupis")
named on local network (Tim Herzog)
Re: pppd LCP problem with kernel 2.2 (Andreas Grabner)
Re: How to telnet as root (Paul)
Re: pppd LCP problem with kernel 2.2 (James Carlson)
Re: Desktop Hub VS. Rackmount Hub (Chris Comley)
----------------------------------------------------------------------------
From: [EMAIL PROTECTED] (Richard Steiner)
Subject: Re: How to telnet as root
Reply-To: [EMAIL PROTECTED]
Date: Mon, 01 Mar 1999 09:45:45 GMT
Here in comp.os.linux.networking, Paul <[EMAIL PROTECTED]>
spake unto us, saying:
>Yep i can do that, but the path doesn't get setup as root's it seems.
If you do an su -, it most certainly should. Note that this is not the
same as just su by itself.
--
-Rich Steiner >>>---> [EMAIL PROTECTED] >>>---> Bloomington, MN
OS/2 + Linux (Slackware+RedHat+SuSE) + FreeBSD + Solaris + BeOS +
WinNT4 + Win95 + PC/GEOS + MacOS + Executor = PC Hobbyist Heaven!
The Theorem Theorem: If If, Then Then
------------------------------
From: "Moshe Bar" <[EMAIL PROTECTED]>
Subject: Re: Ethernet for IBM ps/2 386 machine
Date: Mon, 1 Mar 1999 12:54:27 +0200
This machine uses a MicroChannel and you cannot use other cards.
Try one of the parallel port-based ethernet adapters.
Moshe Bar
Eldir Tomassen wrote in message <[EMAIL PROTECTED]>...
>Who knows what kind of ethernet card would fit in an
>IBM ps/2 386 PC.
>I was just trying to put an MCA card in it but that doesn't really
>fit...
>This machine has no ISA, PCI, VLB or any other common bus type.
>Of course the card has to be supported by linux (suse6.0)
>
>
>
>Eldir
>[EMAIL PROTECTED]
------------------------------
From: Paul Lambert <[EMAIL PROTECTED]>
Subject: Ethernet Problems
Date: Mon, 1 Mar 1999 11:03:19 +0000
Reply-To: [EMAIL PROTECTED]
Hi,
I am relatively new to linux, but have configured a couple of machines in the past.
We have a RedHat 5.1 installation, and in the past I have used it with success.
However, I now have a new PC. It is a PII 350, with a bx(100mhz) chipset. For some
reason, I am now unable to use any ethernet controller. PCI ethernet controllers
(Dlink with a 21140 - tulip and a 3com905) are detected but unfortunately, once I have
installed the OS, they just don't work. I have also tried an ISA NE200 (non plug and
prey) with no success (however I am not sure about the integrity of the card:-( )
Windows NT4 will run with the Dlink card and so I can assume that the hardware isn't
faulty.
Can anyone suggest what I need to do - do I need to upgrade the OS?
Are there any diagnostics that I can run that will give anyone a better idea of the
problem?
Please respond to [EMAIL PROTECTED]
Cheers,
Paul.
------------------------------
From: "Alain Coetmeur" <[EMAIL PROTECTED]>
Subject: Re: ipautofw
Date: Mon, 1 Mar 1999 09:46:30 +0100
Johannes Plachy a �crit dans le message ...
>hi,
>i am new to linux and glad I made it so far...
>
>I am using redhad 2.0.36
>and ipfwadm let my local machines all connect to the internet that works
>fine !
>
>Now I want to map incoming requests from several ports directly to a given
>machine on my intranet.
>
>I tried IPAUTOFW but always get the message 'setsockopt' not available.
>What do I have to do else ? ( I did NOT recompile my kernel to make
>masquerading work !)
don't forget to do an echo 1>/proc/somewhere-i-dont-remember-about
so that the formwarding is activated, given it is already
compiled in the kernel, or more often accessible as a module.
look in dejanews in this group for your setsockopt error message...
the answsers are given.
------------------------------
Crossposted-To: comp.os.linux.setup
From: JK Dingwall <[EMAIL PROTECTED]>
Subject: Re: NFS Export Questions
Date: Mon, 1 Mar 1999 10:28:45 GMT
Fulton B. Gonzalez wrote:
>
> I'd like to find out how to configure /etc/exports so that two different
> users (my wife and myself) on two different linux machines have home
> directories which are readable/writable on the other's machine. Any
> help will be appreciated!
>
> --
> Fulton B. Gonzalez
> Department of Mathematics
> Tufts University
> Medford, MA 02155 USA
>
> Email: [EMAIL PROTECTED]
I think there are two possible solutions for this:
1) Set one machine up as an NFS server for /home, however this will mean
that this machine will always have to be on. This is how I do do it
with
my two boxes at home. The entry on the servers machine would then be:
/home ghost(rw)
This means that the directory /home on the server will be available to
the machine ghost as a read/write export.
>From the machine ghost you can then do (as root)
mount -t nfs spook:/home /home
ie mount the NFS volume on machine spook specified as /home in exports
on the
mount point /home on the local machine.
2) If you want to run the two machines independently then they will both
have to be NFS servers. You will each be able to access your own files,
and only each others when their machine is on.
the exports on your machine would then look like:
/home/husband ghost(rw)
and on your wife's
/home/wife spook(rw)
You can then mount each others home directory if their machine is on.
For further details on the exports file try man exports. Also when this
file is modified you will have to send a SIGHUP to rpc.mountd and
rpc.nfsd
for the changes to take effect. Further details on these man mountd and
man nfsd.
Hope this info is of help,
James
--
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| James
Dingwall |
+
[EMAIL PROTECTED] +
|
[EMAIL PROTECTED] |
+
http://www.fen.bris.ac.uk/students/jd5477/ +
| Maintainer:
http://www.nottingham.ac.uk/~lqzweb/ |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
------------------------------
From: "Alain Coetmeur" <[EMAIL PROTECTED]>
Crossposted-To: comp.object.corba
Subject: a free software dynamic Firewall gateway for MICO & alike ?
Date: Mon, 1 Mar 1999 11:28:52 +0100
I've looked for a gatway that would allow
CORBA clients to contact IIOP Services
through a blocking firewall.
Orbix Wonderwall and a similar system from Imprise do
the job on their supported platforms.
but I'll be working on Linux.
with Linux masquerading I thought
it would be easy, but talking about this
in various groups, I realized that
this is far more complex that plain
IP forwarding.
I'm looking for works around these ideas,
and maybe hitch the shoulder of some developer.
here are the state of my reflexions about
the difficulties.
when a client want to contact an internal service
it can only contact an IP port on the firewall,
and it is quite easy to relay it to the hidden service
with IP forwaring line in linux, or through a TIS-like TCP gateway.
but IIOP IP port are dynamically allocated and
one must hack the implementation repository so that
the IP forwarding is configured.
moreover the implementation repository on the firewall
must register all relay port as if they where local.
but in fact this cannot work very far since IOR that are exchanged
between services (name services, and user services)
refer to the internal network address.
one solution is to hack the internal network ORB so that
they think thay are located on the gateway, and set
the IP forwarding transparently. thus every internal client,
external client, and even the service itself think
the service is located on the gateway.
this is only possible if you hack the ORB
or maybe the system and socket library
so that bind(), listen(), getsocketaddress()
simulate this... anyway this is feasible on
Linux, providing a kind of "cluster"-like feature.
anyway the more realistic solution would
be to develop a IIOP gateway based on DII/DSI
relaying, and a virtual implementation repository.
depending on the address domain (inside, outside)
of the client, the gateway would convert the IOR
from "insider" to "insider" or reverse.
I think it can be possible, but compute intensive,
since we should unmarshall each IIOP message
to find the IOR and convert them.
anyway the advantage from pure IP formarding
would be the only request that get through the
firewall are converted, and that local request
don't involve the firewall.
moreover one can control the object that are
allowed to be relayed depending on many factors,
including the content of the request, since
we must unmarshall it anyway.
all of this would be quite portable, except maybe
for the IP forwarding, which may anyway be
a separate module using IP forwading
or TCP relay.
this looks like an exciting project...
the kind Linux folks are fond of ?
------------------------------
From: Wazza <[EMAIL PROTECTED]>
Crossposted-To:
alt.os.linux,alt.uu.comp.os.linux.questions,aus.computers.linux,linux.redhat.misc
Subject: Init Id * respawn error
Date: Mon, 01 Mar 1999 22:09:08 +1100
G'day
I have not had much experience with Linux and and trying to set a box up
to masquerade to my other machines but have run accross this
intermittent error just before login.
INIT: Id "1" respawning too fast: disabled for 5 minutes
INIT: Id "2" respawning too fast: disabled for 5 minutes
INIT: Id "3" respawning too fast: disabled for 5 minutes
INIT: Id "4" respawning too fast: disabled for 5 minutes
INIT: Id "5" respawning too fast: disabled for 5 minutes
INIT: Id "6" respawning too fast: disabled for 5 minutes
Can't open "/var/lock/news/shlock447", Read-only file system
cat: /var/lock/news/LOCK.innwatch: No such file or directory
innwatch: [440] locked by []
The Init Id error's will then loop endlessly every 5 minutes.
is this to do with a call in the /etc/inittab file ?
1:12345:respawn:/sbin/mingetty tty1
2:2345:respawn:/sbin/mingetty tty2
3:2345:respawn:/sbin/mingetty tty3
4:2345:respawn:/sbin/mingetty tty4
5:2345:respawn:/sbin/mingetty tty5
6:2345:respawn:/sbin/mingetty tty6
Can anyone help and explain this problem to me??
the error is intermittent.
Thanks for any help
Warwick
------------------------------
From: "laujon" <[EMAIL PROTECTED]>
Subject: Fetchmail 4.6.4 and SMTP
Date: Mon, 01 Mar 1999 11:26:40 GMT
Hi,
I have installed fetchmail 4.6.4 on linux 2.0.36 (debian).
It polls POP3 server (Netscape Mail server 2.0). In syslog I get these
messages. I can telnet on port 25. I had to install smail to get that,
because else it said "connection refused".
========== start
Mar 1 10:09:46 pcmarfle in.smtpd[8121]: connect from localhost
Mar 1 10:09:46 pcmarfle fetchmail[812]: reading message 1 of 4 (39350
octets) (log message incomplete)
Mar 1 10:09:46 pcmarfle fetchmail[812]: SMTP listener doesn't like
recipient address `marfle@localhost'
Mar 1 10:09:46 pcmarfle fetchmail[812]: can't even send to postmaster!
Mar 1 10:09:46 pcmarfle fetchmail[812]: SMTP transaction error while
fetching from cmlag.fgov.be
I don't know what to do, please help.
Regards.
--
Posted via Talkway - http://www.talkway.com
Surf Usenet at home, on the road, and by email -- always at Talkway.
------------------------------
From: "Vasilios E. Tourloupis" <[EMAIL PROTECTED]>
Crossposted-To: aus.computers.linux,comp.os.linux.misc,comp.os.linux.hardware
Subject: Configuration Conflicts between Dialup Adapter and Ethernet Card
Date: Mon, 1 Mar 1999 22:41:46 +1100
Dear Linux Users,
I know this is the wrong newsgroup to post this question, however, there may
be some people
who have experienced or are experiencing the same problem as myself. The
following message
is the one I sent to the Windows 95 newsgroups:
Does anybody know how to resolve a configuration conflict that I am
currently experiencing with my dial-up adapter and my ethernet card.
In particular, I am experiencing a clash between the DNS Configuration
for the dial-up adapter and the Ethernet card. Because I am on a LAN,
I want to use my ISP's DNS for my dial-up adapter, and the my LAN's
DNS for my ethernet card. At the moment, I am getting the same
configuration for both - i.e. if I modify the configuration for one
device,
Windows 95 is happy to modify the configuration for the other device.
As you've probably guessed by now, the LAN that I am refering to is none
other than my
Linux box acting as the DNS server. I suppose I should describe the
symptoms of the problem
before jumping to any conclusion: basically, I am getting a VERY slow
response time from my
ISP when I dial up under Windows 95 - so slow that some of the time I get
disconnected!
Am I correct to assume that the above mentioned problem is the cause of
this?
Thanks in advance,
Regards,
Bill Tourloupis - [EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED] (Tim Herzog)
Subject: named on local network
Date: Mon, 01 Mar 1999 05:51:22 -0600
Bear with: I'm a new Linux user...
How do I go about configuring named to work on a local network if named
does not have Internet access? I have a local network of, say, three
machines:
192.168.6.1 tom.foo.com
192.168.6.2 dick.foo.com
192.168.6.3 harry.foo.com (Harry's the Linux box)
I want my Linux box to provide DNR services not only for itself, but for
the other two machines as well. These are the only three IP addresses
named need be aware of.
I added these three entries to /etc/hosts. That works great on the Linux
box (using say, Netscape), but if tom or dick query named on harry,
harry's named apparently ignores /etc/hosts and queries the Internet
instead. Eventually, the network times out.
Tim
[EMAIL PROTECTED]
--
Two Bits Worth
778 Hague Avenue
Saint Paul, MN 55104
Phone/Fax: 1 (612) 227-2920
[EMAIL PROTECTED]
------------------------------
From: Andreas Grabner <[EMAIL PROTECTED]>
Subject: Re: pppd LCP problem with kernel 2.2
Date: Mon, 01 Mar 1999 09:44:29 +0100
Clifford Kite wrote:
> : Feb 27 22:21:57 film chat[438]: CONNECT
> : Feb 27 22:21:57 film chat[438]: -- got it
> : Feb 27 22:21:57 film chat[438]: send (\d)
> : Feb 27 22:21:58 film chat[438]: expect (sername:)
> : Feb 27 22:21:58 film chat[438]: 115200^M
>
> Here is something strange. No ordinary modem can send or receive at
> 115200 bit/sec over the usual PPP connection except with help from modem
> compression and error correction. Connecting at that speed, specified by
> the pppd option, very likely means that no error correction or compression
> is configured in the modem itself.
>
> <snip>
>
> : an my options (i tryed a lot of other)
>
> : /dev/ttyS2
> : 115200
> : crtscts
> : modem
> : debug
> : defaultroute
> : asyncmap 0
> : (lcp-restart 2)
>
> --
> Clifford Kite <[EMAIL PROTECTED]> Not a guru. (tm)
> /* I gave up on politics when no matter who I voted for, I regretted it.
> * -- Pepper...and Salt, WSJ */
Tanks now it works dont ask me what was te problem maybe the UART chip
Andi
------------------------------
From: [EMAIL PROTECTED] (Paul)
Subject: Re: How to telnet as root
Date: Mon, 01 Mar 1999 12:17:32 GMT
OK ta.
[EMAIL PROTECTED] (Richard Steiner) wrote:
>Here in comp.os.linux.networking, Paul <[EMAIL PROTECTED]>
>spake unto us, saying:
>
>>Yep i can do that, but the path doesn't get setup as root's it seems.
>
>If you do an su -, it most certainly should. Note that this is not the
>same as just su by itself.
------------------------------
From: James Carlson <[EMAIL PROTECTED]>
Crossposted-To: comp.protocols.ppp
Subject: Re: pppd LCP problem with kernel 2.2
Date: 01 Mar 1999 07:25:16 -0500
[EMAIL PROTECTED] (Clifford Kite) writes:
> One thing to check is the setserial command in the /etc/rc.* boot-up
> files. The UART type specified must agree with the actual type of UART
> that the modem uses.
The logs seem to show that he can hear the peer correctly, but that
the peer cannot hear him during LCP. That could be a flow control
problem.
Since he's doing "&F" to his modem, I'd recommend trying "&K3" as
well. Some modems apparently have a factory default of no flow
control (!).
> Another thing is the configuration of the modem itself. This is likely
> to be your problem and there is another comment about it below.
[...]
> Here is something strange. No ordinary modem can send or receive at
> 115200 bit/sec over the usual PPP connection except with help from modem
> compression and error correction. Connecting at that speed, specified by
> the pppd option, very likely means that no error correction or compression
> is configured in the modem itself.
Uh, I don't think so. Modern modems can be told to report either DTE
or DCE speed at connection time. Many PC users (for reasons that I
don't quite understand) have their modems set to report the DTE rate
-- which happens to be fixed here to 115.2Kbps async. Often the other
interesting connect parameters (V.42 and so on) are left out when this
is done.
He'd probably never have gotten this far if no error correction was
being negotiated. V.34 and V.90 connections are pretty much unusable
without V.42. Even on perfectly clean digital lines in a lab test, I
would get a burst of errors every one or two seconds on a V.34bis
connection.
Now, it is possible that he's just going too fast either for that port
hardware or for that brand of modem and causing overflows.
--
James Carlson, Consulting S/W Engineer <[EMAIL PROTECTED]>
IronBridge Networks / 55 Hayden Avenue 71.246W Vox: +1 781 372 8132
Lexington MA 02421-7996 / USA 42.423N Fax: +1 781 372 8090
"PPP Design and Debugging" --- http://people.ne.mediaone.net/carlson/ppp
------------------------------
From: [EMAIL PROTECTED] (Chris Comley)
Crossposted-To: comp.dcom.lans.ethernet,comp.os.ms-windows.nt.admin.networking
Subject: Re: Desktop Hub VS. Rackmount Hub
Date: Mon, 01 Mar 1999 12:34:09 GMT
On Tue, 23 Feb 1999 13:27:10 -0800, "- AJS"
<[EMAIL PROTECTED]> wrote:
>
>Chris Comley wrote in message <[EMAIL PROTECTED]>...
>>Er, no.
>>
>>There are desktop hubs, un-managed rack-mount hubs, and managed
>>rack-mount hubs. Managed hubs are a LOT more expensive than either of
>>the unmanaged type, mainly I guess because they're a lot less used,
>>economies of scale and all that. We sell about 50 hubs a month, almost
>>all of them rack-mount, only one or two of them managed.
>
>
>Uh huh.
>One more thought for you... I use stackable hubs with 1 manageable hub in
>each stack. The rest of the stack are slaved by the management hub. (My
>brand of choice: Neteligent 3024/3124.) The Mgmt unit is more expensive, but
>overall reduces the cost of managed hubs by controlling and reporting for 5
>others.
>
>All depends on what you need. Define that first and then find equipment that
>suits your needs.
This is also a good point. But 95% of our customers do NOT use
"managed" hubs *at all*. It all depends. If you buy managed hubs, you
probably have a massive network spread over a large area and already
have OpenView or some similar management tool in use. Most of our
customers are "SME" sized, e.g. networks of up to about 200 nodes,
these generally have a single wiring closet, and it's a lot easier and
cheaper to fault-find by opening the closet door and looing at the
front panel lights on the hubs and switches! :-)
------------------------------
** 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
******************************
