Linux-Networking Digest #448, Volume #10 Wed, 10 Mar 99 14:13:42 EST
Contents:
Re: Total collision on RedHat machine (Rod Smith)
Linux fax server on linux/Win network ("Bruno Prior")
Help RealTek8019 or NE2000 Compatible ("Daniel Cederhammar")
Re: 2 Linux machines 2 nics and a hub and something is wrong? (Alexander Atkin)
Re: Modem setup problems (Clifford Kite)
apache final slash (ike icold)
Re: eth0: Rx frame corrupted (Stu Lewallen)
Re: newbie Puzzled over Samba connection ("Leopold Toetsch")
Re: Linux as a router to replace school NT4 box? (Vincent Fox)
Re: Machine name themes - what do you use? ("Stone Tiger")
NIS netgroup and mountd problems on RedHat 5.2 (Georg Schwarz)
Traffic shaper on a subinterafce (Kostas Zourkas)
Re: Linux+Win95 networking problem (Michael John Erskine)
Printing to Laserjet fails only on large jobs ([EMAIL PROTECTED])
Re: Can you tell the speed of your dialup connection?? ("Phantom")
----------------------------------------------------------------------------
From: [EMAIL PROTECTED] (Rod Smith)
Subject: Re: Total collision on RedHat machine
Date: 10 Mar 1999 17:05:29 GMT
Reply-To: [EMAIL PROTECTED]
[Posted and mailed]
In article <[EMAIL PROTECTED]>,
Jakob Kruse <[EMAIL PROTECTED]> writes:
> Problem is now solved.
>
> Solution: Don't use WDC network cards. Use SMC Ultra instead, and the
> problem disappears!
>
>> I have a very serious problem with a Linux box: It is unable to transmit
>> even a single IP packet. ifconfig shows all packets fail due to
>> collision - ALL of them!
I had a similar problem yesterday with a Kingston KNE100TX I got as a
factory replacement for a board of the same model that went south. The
original board has a DEC Tulip 21140 chip, while the new one has an Intel
Tulip 21143 chip (an aside: has Intel bought the Tulip design, or are
they a second source for these chips, or what?). Anyhow, the 21143 is
newer than the 21140, and the Tulip drivers in Linux 2.2.1 caused
problems. The de4x5 drivers were better, but a bit flakey. I got an
updated tulip.c file (version 0.90) and recompiled my kernel with that,
and it's been fine since then.
Now, I've no idea if the WDC cards you tried used the 21143. If so, that
might be the problem. If not, it could be that an equivalent driver
update would fix the problem. Of course, this is probably moot for you
unless you want to return hardware, but somebody else might find it
helpful, and it might at least clear up a mystery for you....
--
Rod Smith
[EMAIL PROTECTED]
http://www.channel1.com/users/rodsmith
NOTE: Remove the "uce" word from my address to mail me
------------------------------
From: "Bruno Prior" <[EMAIL PROTECTED]>
Subject: Linux fax server on linux/Win network
Date: Wed, 10 Mar 1999 17:18:50 -0000
Can I set up a fax server on a Linux server to act as a server for Windows
machines as well as linux machines. I can't find any information on the
HylaFAX web site to tell me one way or the other. Seems like it ought to be
possible in principle, but might be fun setting up a friendly Windows
front-end.
Cheers,
Bruno Prior [EMAIL PROTECTED]
------------------------------
From: "Daniel Cederhammar" <[EMAIL PROTECTED]>
Subject: Help RealTek8019 or NE2000 Compatible
Date: Wed, 10 Mar 1999 18:39:18 +0100
I can�t get my network card to work under Debian
Can someone help me whit this?
/Ceder
------------------------------
From: Alexander Atkin <[EMAIL PROTECTED]>
Subject: Re: 2 Linux machines 2 nics and a hub and something is wrong?
Date: Wed, 10 Mar 1999 10:29:52 +0000
I have a Linux router running Slackware 3.5 with a PCI Plug & Play NE2000 network
card and it works exceptionally well. It found the card and had it running twice
as fast as an ISA one which I had to load a module for instead.
Michael Wilson wrote:
> On Sun, 24 Jan 1999, Darren Ford wrote:
>
> >Folks, I am about to go crazy here with problems. I have two machines
> >both running Red Hat 5.0. Both have NICs and are recognized during the
> >boot process. I reinstalled Red Hat on both machines to get a fresh
> >start. During the install I chose to set up networking for each
> >machine.
> >I want machine one(M1) to be able to connect to a cable modem later on
> >so I want it as a gateway.
> >
> >M1(linux1.blank.net) M2(linux2.blank.net)
> >IP = 192.168.9.1 IP = 192.168.9.2
> >netmask = 255.255.255.0 netmask = 255.255.255.0
> >gateway = 192.168.9.1 gateway = 192.168.9.1
> >domain = blank.net domain = blank.net
> >
> >These machines are both connected to hub. I cannot ping one machine
> >from the other but they can ping themselves and their dummy interface.
> >
> >If I want M1 to be a gateway later on, setting the gateway IP to
> >192.168.9.1 is correct right?
> >
> >Why can't these guys ping each other? I tried adding each of them to
> >the others routing table but that didn't work or I didn't do it right.
> >When I do a 'route -n' it shows that the entry is there.
> >
> >Can somebody tell me what I'm doing wrong?
>
> I had this problem a while ago whilst experimenting with IP Masquerading using
> a Win98 box into a Linux Box connected to the net, the solution was quite
> simple, The NIC that I had in the Linux Box was a clone NE2000 PCI card which
> in all fairness initialised on bootup but failed to establish any networking
> protocols whilst trying to ping the other machine and vice versa. I would go
> to the lengths if you are using a PCI NIC that maybe a change to ISA may solve
> the problem. As far as your setup goes with linking two machines to the
> Internet via one box, I would suggest a look at the doc entitiled
> IP-Masquerade, the setup facility in it worked for me. If you haven't got it,
> I can forward it to you no bother.
>
> Regards, Michael
------------------------------
From: [EMAIL PROTECTED] (Clifford Kite)
Subject: Re: Modem setup problems
Date: 10 Mar 1999 10:59:12 -0600
Somnath Banerjee ([EMAIL PROTECTED]) wrote:
: Then I pressed activate from the control panel menu.
: The activate button is greyed but I do not see any action. The modem does
: not even start dialling
: So I weent to /var/log/messages. I am attaching the tail of this file here:
: ============================= Beginning of the message file
: ===========================
: Mar 9 22:53:39 localhost pppd[678]: pppd 2.3.5 started by root, uid 0
: Mar 9 22:53:39 localhost pppd[678]: tcgetattr: Input/output error(5)
tcgetattr almost always means that there is another process besides
pppd that is contending for the modem IRQ. Some examples are (1) an
ethernet card that used the same IRQ as the modem, (2) a misconfigured
ppp script that launched pppd twice, and (3) a mingetty that somehow
had been configured for call-in. If yours is different, I would be
interested in adding to this list. :)
--
Clifford Kite <[EMAIL PROTECTED]> Not a guru. (tm)
/* Governments should be changed like diapers - often and for the
* same reason. */
------------------------------
From: ike icold <[EMAIL PROTECTED]>
Subject: apache final slash
Date: 10 Mar 1999 14:31:38 GMT
My apache server works fine except....
when I request a directory name. the server returns the index.html file
only if there is a slash at the end of the request. Otherwise it just
waits and waits and waits.
I am running slackware 96 which I got with a book from slackware. I think
it is linux 2.0 and apache 1.1 but I am not sure.
I am assuming that there is a configuration fix for this. Something like
FileNamesWithoutExtentionsAreDirectories on
but I cannot find it in the conf files.
================== Posted via SearchLinux ==================
http://www.searchlinux.com
------------------------------
From: Stu Lewallen <[EMAIL PROTECTED]>
Crossposted-To:
alt.os.linux,comp.os.hardware,comp.os.linux.misc,comp.os.linux.setup,alt.linux.slakware,alt.os.linux.slackware
Subject: Re: eth0: Rx frame corrupted
Date: Wed, 10 Mar 1999 18:14:59 GMT
I would start by getting the 2.0.36 source and recompiling.
[EMAIL PROTECTED] wrote:
>
> Hi
>
> I just installed Slackware 3.6. Everything seems to work fine, except
> the network communications. When I start something that involves an
> extensive transfer of data over the network, I get loads of following
> message:
>
> eth0: Rx frame 5675 corrupted, status dbc8, cmd ae1e, next 8a23, pbuf
> c702, len 6e3f
>
> after which the system hangs/crashes. The hex numbers are different
> each time the problem occurs.
>
> Some network comms seem to be working, though. E.g. I can open an ftp
> session to a ftp server, log on, cd to another directory. But when I
> actually get the file, the problem occurs. Same with nfs.
>
> I've changed base address and IRQ of the network card, without
> success.
>
> The PC is a Dell OptiPlex 466/Le with 24 MB RAM and an Intel
> EtherExpress 16 network card. The software is Linux Slackware 3.6,
> kernel 2.0.35 (net.i)
>
> Any clues anyone ? Any help would be greatly appreciated.
>
> Ivan
------------------------------
From: "Leopold Toetsch" <[EMAIL PROTECTED]>
Subject: Re: newbie Puzzled over Samba connection
Date: Wed, 10 Mar 1999 17:24:56 +0100
Ron S wrote in message <[EMAIL PROTECTED]>...
>I set up a redhat 5.2 with samba server to connect to a win98 system.
>Everything worked fine..until today...not sure what I goofed up.
>I can ping from the linux to the win98 using ip address or name of the
win98
>system. I can ping the linux system from win98 only using the ip address of
>the linux, cant ping it by name(I used to be able to ping it by name from
>the win98 system, the linux system can ping itself by ip address or name) ,
>it gives 'host unknown'
>==CUT
>the Lmhosts.sam file on the win98 machine looks ok.
lmhosts.sam is a _sample_ file.
On my system (NT) I configured
\winnt\system32\drivers\etc\hosts
>I can print from the linux system to the printer on my win98 system. It
>seems somewhere the ip address isn't converted to the name of the linux
>samba server
Also make sure that authentication is on boths sides either encrypted or
not.
leo
------------------------------
From: [EMAIL PROTECTED] (Vincent Fox)
Subject: Re: Linux as a router to replace school NT4 box?
Date: 10 Mar 1999 10:22:23 -0500
In <7c5loe$fib$[EMAIL PROTECTED]> "Andrew Dunlop" <[EMAIL PROTECTED]>
writes:
>Stefano has now emailed me the disk at it's time to play setup. In some
>respects I wish I had more fully considered the Linux option before I set up
>a MS based, email, proxy, and web server.
>The prospect of managing and maintaining a Linux box seemed a little
>daunting for a school, but I guess the reliability would have meant that it
>would not have been such a problem once it is all set up.
Here's my short perspective on the UNIX vs MS server issue:
UNIX came from a mainframe environment and is migrating down to
the desktop. It's original design framework came from that mainframe
mindset where you could not have 200 users on the system and have
user 201 bring down the whole system by running their accounting package.
It may not be friendly, but it's solid. The uptime on my main Sun
server is 172 days at present. This is a machine that is the
file server, web server, e-mail, NIS, DNS, *and* has 20-30 logins
at any time, all running anything from elm to ANSYS.
NT, etc. came from the desktop and is still struggling it's way
up into utility as a server OS. It's very nice as a desktop OS for
many people, let me not disparage it for that. However, in it's server
role it's stability and performance leave much to be desired.
Even MS tried a year or so ago to move all it's web servers over
to NT. They were crushed. They could not handle the load. They had
to move back to UNIX-based web servers to handle it.
Right now in any client/server setup I would lean toward a Sun
or Linux-based server, and NT clients. You can run Samba on the
server to handle the NT fileserver needs, etc.
--
"Who needs horror movies when we have Microsoft"?
-- Christine Comaford, PC Week, 27/9/95
------------------------------
From: "Stone Tiger" <[EMAIL PROTECTED]>
Crossposted-To:
vmsnet.networks.misc,microsoft.public.windowsnt.domain,comp.unix.solaris,comp.os.os2.networking.server,comp.os.ms-windows.nt.admin.networking,comp.infosystems.www.servers.unix,comp.protocols.tcp-ip.domains
Subject: Re: Machine name themes - what do you use?
Date: Wed, 10 Mar 1999 18:18:22 GMT
For new servers, we were using Disney names, particularly the Seven
Dwarves - Dopey, Sneezy, Grumpy, etc.
But out Senior VP of Operations is a HUGE Warner Bros cartoon fan - so, at
his behest, our last three new servers have been Marvin, Daffy, and
Sylvester. I'm just dreading the day he requests our Domain be renamed Bugs
Bunny.
------------------------------
From: [EMAIL PROTECTED] (Georg Schwarz)
Subject: NIS netgroup and mountd problems on RedHat 5.2
Date: 10 Mar 1999 14:40:27 GMT
On a newly installed RedHat 5.2 (Intel) Linux machine, which runs as a
NIS client, I don't succeed in exporting file systems using NIS
netgroup. When I directly specify in /etc/exports a host that should be
allowed to NFS mount a directory from the PC it works fine. However, if
I put that host in netgroups (on the NIS server) it does not work.
The situations is as follows:
- ypdomasiname works (on the PC == NIS client == NFS server)
- ypcat netgroup gives me the correct map (with the host that should be
NFS client included)
- the netgroup's name, with a leading @, is specified on the PC's
/etc/exports
- showmount -e from remote shows the directories to be exported with the
correct name of the netgroup (with a leading @)
- running mountd in debug mode, on NFS mount attemps one gets syslog
messages such as:
Mar 10 13:34:52 bofh mountd[19958]: flushed host access cache
Mar 10 13:34:52 bofh mountd[19958]: NFS mount of /usr attempted from
130.149.162.136
Mar 10 13:34:52 bofh mountd[19958]: check unknown clnt addr 130.149.162.136
Mar 10 13:34:52 bofh mountd[19958]: ^Iclient name is m21.physik.TU-Berlin.DE
Mar 10 13:34:52 bofh mountd[19958]: Unauthorized access by NFS client
130.149.162.136.
Mar 10 13:34:52 bofh mountd[19958]: Blocked attempt of 130.149.162.136
to mount /usr
....
is it possible that RedHat 5.2's mountd maybe does not include NIS
support?
What could I have done wrong, and what else could I check for?
Thanks.
--
Georg Schwarz ([EMAIL PROTECTED], [EMAIL PROTECTED], PGP 2.6ui)
Institut f�r Theoretische Physik +49 30 314-24254 FAX -21130 IRC kuroi
Technische Universit�t Berlin http://home.pages.de/~schwarz/
------------------------------
From: Kostas Zourkas <[EMAIL PROTECTED]>
Subject: Traffic shaper on a subinterafce
Date: 10 Mar 1999 15:31:38 GMT
Hi,
iam using linux 2.2.3 with the build in traffic shaper. Using two ethernet
interfaces (eth0, eth1), i have attached the shaper at eth1 with a speed of
32K and is working fine. Now i want to add another subinterface: eth1:0
With an ip address from another class C network (/24). When i try
to attach shaper1 to this newlly created subinterface, i get :
no such device, even though when i issued the command:
insmod -o shaper1 /lib/modules/2.2.3/net/shaper.0 , i didn't get any
errors.
Is there something wrong with the way iam trying to make it work?
i want to have two groups of peaple in side my network that will have their
own banwidth 32k ( eth1) and 16K (eth1:0). Iam using ipchains to forward
the traffic.
Thank you in advance for any help.
================== Posted via SearchLinux ==================
http://www.searchlinux.com
------------------------------
Date: Wed, 10 Mar 1999 18:05:38 +0000
From: Michael John Erskine <[EMAIL PROTECTED]>
Subject: Re: Linux+Win95 networking problem
> I just set a two-machine network between a Linux and a Win95 (OSR 2) boxes.
Running PPP hopefully.
> They are connected with a crossover twisted-pair cable. The connection
> seems to be working - I can ping each machine from the other one. I can
> also ping the machines themselves.
Which means that the PPP connection is working properly.
>
>
> Now here is the problem. On my Win95 machine I also have a couple of
> dial-up accounts configured (one to my ISP and one to my work LAN). When I
> try to do telnet or ftp from Win95 to Linux, I get a dial-up popup window
> asking me to connect to my ISP account.
Because it is possible to configure W95 to automatically dial out when you do
something that it interprets to mean you want to go to the net.
> When I cancel this, nothing
> happens.
I bet you could telnet to the W95 machine if you were running WinQVT server.
> It seems that neither telnet nor ftp can see the ethernet
> connection.
>
The connection is not made because W95 thought you wanted it cancelled. W95 is
brain dead in this regard and assumes you are like the rest of the herd... you
have only one connection to one net via dialup.
> Why is this happening and how can I fix it? I tried to remove the
> "TCP/IP -> Dial-up adapter" from the Network applet, but this did not change
> anything.
>
The W95 machine probably does not have anything set up to indicate where (which
interface) packets are to go for each network.
> Additional info. I have a small Win98 partition on my Linux box and when I
> boot it, my "network" works with no problems.
When you do that your are using NETBUI and NETBIOS. Set up Samba on your linux
machine and you will see the same thing,
> I can see both machines in
> the "network neigborhood" and I can mount local disk dirives and copy files
> between them.
>
Different protocol and name service.
> Any info will be appreciated. Thanks in advance,
>
> Andy
>
> -----------------------------------------
> Andy Jaworski
> [EMAIL PROTECTED]
>
> Opinions expressed herein are my own and may not represent those of my employer.
--
Michael Erskine, Simply Computers!, Urbanna, Virginia, 804-758-3798
[EMAIL PROTECTED]
============= Posted via Newsfeeds.Com, Uncensored Usenet News ============
http://www.newsfeeds.com/ The Largest Usenet Servers in the World!
============= Over 66,000 Groups, Plus a Dedicated Binaries Server ============
------------------------------
From: [EMAIL PROTECTED]
Subject: Printing to Laserjet fails only on large jobs
Date: Wed, 10 Mar 1999 19:03:50 GMT
I am having problems printing large jobs from a vanilla RedHat 5.2 box
to a Laserjet 4si (using its JetDirect MIO card) via ip. Hold on to
your seat, this isn't as straightforward as it sounds at first ...
A 400+ page report is generated vi Universe on HPUX 10.20. This report
is pretty much plain text, just very large. The report prints
correctly if I print from HPUX directly to the printer via IP (not
using JetDirect software). However, if I redirect the print job to a
queue on the RedHat box, the printer spits out 20-30 pages and gives
an error: "80 SERVICE 001F", then promptly dies. A power cycle is
needed to get the printer working again.
Printing from HPUX through the Linux queue works fine on all jobs
except the big one - the smaller ones are anywhere between 1-20 pages.
The Linux box has sufficient resources - it's ONLY used for print
spooling, has 128MB RAM and 4GB dedicated to /var/spool/lpd ...
The ONLY difference in printer definitions on the HPUX box is the
destination IP address and queue name (direct: printer_ip & raw;
linux: linux_ip & queuename). This anomaly shows up whether the linux
queue is using the laserjet filter or none at all. The printcap entry
on linux looks like:
pr0_manual:\
:sd=/var/spool/lpd/pr0_manual:\
:mx#0:\
:sh:\
:rm=pr0:\
:rp=raw:
(pr0 is listed in the /etc/hosts file)
The inability to print this report correctly is causing several people
to lose hair. It's being generated at one of the larger local
hospitals. I'd greatly appreciate any help you can offer me in this
matter. Thanks!
[EMAIL PROTECTED]
------------------------------
From: "Phantom" <[EMAIL PROTECTED]>
Subject: Re: Can you tell the speed of your dialup connection??
Date: Wed, 10 Mar 1999 15:42:02 -0000
you could look in /var/adm/messages, the chat script
normally writes stuff there including the connect string
from the modem which should include line speed
Eriksson wrote in message ...
>How?
>
>I'm using ordinary PPP.
>
>Thanx!
>
>/Martin
>
>
------------------------------
** 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
******************************
