Linux-Networking Digest #964, Volume #9 Fri, 22 Jan 99 00:13:31 EST
Contents:
Scheduled PPP? (Bruce Barnett)
Re: Why Does Linux Networking Suck So Badly ? ("Mark Vandersteen")
Botting diskless SPARCstation off redhat5.2 unsuccessfull...please help
("p.wojcieszak")
Re: Cable Modem problems (Buster_Bear)
Re: IP Masq & Civ 2 MPE (Dan Kegel)
Re: 3Com 3c509b, not working under RH 5.2 (Dan Kegel)
Re: Connect without hub (Wolfgang Kleinhappl)
Re: Extraction of IP address ("Rob Lauer")
TCP/IP help (Ziae)
Re: diald help. (Vincent Zweije)
Re: Multiple 3c509 ("Kyle Bowerman")
FTP files not visible (Ron Thompson)
Re: RedHat Linux 5.1 & Ethernet 3c509 NIC ("Kyle Bowerman")
Re: Transmit time out (David Kirkpatrick)
----------------------------------------------------------------------------
From: Bruce Barnett <[EMAIL PROTECTED]>
Crossposted-To: linux.dev.ppp
Subject: Scheduled PPP?
Date: 22 Jan 1999 02:50:10 GMT
I have a PPP link set up to disconnect when idle for 10 minutes.
However, it immediately reconnects. How is this decision made?
I want to scheule a connection ABOUT once an hour, that disconnects when
idle. The PPP docs suggest DIP, while anothe document suggests diald.
The PPP people don't want to talk about either, so I have to figure
out how to integrate the packages. It seems a shame, because PPP
ALMOST does what I want it to do.
The setup scripts are convoluted. Can someone give me a pointer on how
to accomplish the task? Or how the scripts are organized?
The README file does a very poor job of explaining the dozens of script
files used.
If I could put a DELAY between the shutdone and the startup, I'd
solve most of my problem. But this isn't documented as far as I can
tell.
Can anyone help?
------------------------------
From: "Mark Vandersteen" <[EMAIL PROTECTED]>
Subject: Re: Why Does Linux Networking Suck So Badly ?
Date: Fri, 22 Jan 1999 13:31:42 +1030
Your answers is plain and simple "DONT USE IT", go back to windoze, stop
rubbishing something you know nothing about.
If you take time to actually learn how to do something insead of using
something without reading the instructions like 99% of people in todays
world then you might actually benefit from it.
Keep your rubbish in the windows newsgroup thanks :)
>> From: [EMAIL PROTECTED] (j)
>> The sheer number of messages posted here - usually half going
>> unanswered - tells the tale ... Linux networking sucks.
>> Now you can take your Winders, add in file-sharing in just a
>> couple of minutes, bring up your Winders workstations and
>> locate and link in within just another few minutes. Easy.
>> Quick. Relatively painless. Not the Linux way ...
>> Now you may try to defend Linux by saying that it has more
>> "flexibility" - but quite frankly we could do with a lot
>> less "flexibility" and a lot more simplicity. I'll bet
>> that 95% of users just want to tie into a local NT server
>> or use Linux AS a file/print server for Winders PCs.
>> I think that "flexibility" is just an excuse - a euphamism
>> for "a bunch of really unsophisticated little utilities
>> written with no real plan in mind by people with a fetish
>> for cryptic poorly-documented command-line parameters and
>> well-hidden config files". As is, Linux is a user-UNfriendly
>> mirror of MSDOS - and that goes double for "connectivity" issues.
>> Want cheap PC networking & connectivity ? One "L" word comes
>> to mind and it's NOT "Linux" - try "LanTastic" instead.
>> Now Linux IS relatively compact and fast and DOES get past
>> some of the old buggaboos inherent to DOS-derived systems,
>> but Linux just isn't "ripe" yet. It needs DUN boxes and
>> "Network" panels which HELP you with things and show all
>> the pertinent options and conflicts right there.
>> Someone could make a LOT of money if they could put together
>> a turn-key Linux-based file/print server package which could
>> autoconfigure most things and be really easy and helpful on
>> the rest. Something you take out of the box and 20 minutes
>> later you have a good working server. You can do that with
>> Winders, but not with Linux unless you apparently have put
>> in a tremendous number of obsessive hours.
>> Now you thought I was just gonna flame ... nope. I've been
>> relatively polite and realistic here. There are MANY issues
>> to address, MANY refinements to be made before Linux becomes
>> a really good option for administrators of small/medium networks.
>> As much as I dislike Winders, it's a positive relief to boot
>> back into it after fooling around for hours in vain with Linux.
>> Preferring Winders to ANYTHING ... IMHO that's a pretty severe
>> indictment. Linux has been badly over-sold. If it weren't so
>> cheap it would be a rip-off ... a hobby OS masquerading as
>> the "real deal".
------------------------------
From: "p.wojcieszak" <[EMAIL PROTECTED]>
Subject: Botting diskless SPARCstation off redhat5.2 unsuccessfull...please help
Date: Thu, 21 Jan 1999 14:39:30 +1100
PROBLEM:
A diskless SPARCstation ELC will not boot of redhat 5.2.....
PRECISE DEFINITETION:
The tftp protocol works ok and sparc station downloads the bootimage of
linux bootparamd
without any problems...ARP, domainame calls etc are all successfull....
Eventually the Sparc Station mounts the directory off the linux server !
It then attempts to download a kernel file(486kb large)......
NFS READ TIMEOUT ERROR occurs...
Upon checking out the network traffic.....It seems that the linux box
sends the packets too
quickly for the poor sparc station...and the overrun
occurs.............Upon linux sending 5 packets
the sparc requests a retransmission of the entire block.
My feeling is that Linux NFS BLOCK_SIZE is by default set to 8192
bytes....BUT for some reason
upon sending that block to the client, the NFS_SERVER chops the block
size into 1024 b blocks!
.....Thus when the NFS server sends the 1024 byte blocks , it expects
each in turn to be acknowledged......It seems that the sparc diskless
client is not expecting 1024b blocks....It expects
the whole 8192 block....!!!! And thus does not reconstruct the block
correctly and requests a timeout.
HELP:
Am I on the right track?
What can I do....how can I force linux NFS to produce 8K block
sizes....?
btw...please remember that this problem is specifically related to
diskless sparc stations....
If you can help , it would be most appreciated....I am sick to death of
hacking and recompiling
the NFS code! This is a really bad problem!
HARDWARE INFO:
server:Linux Dell 3500 series , redhat 5.2 server, 256M ram , single
P2-350,Ami megaraid (raid 5 conf)3-scsii 9.5 gig drives.... PCI ETHERNET
intel express pro 100mbits
client: ELC sparc station, 10 Mbit connection
client-server path: 100/10 bay networks hub/switch.
thanks
paul. [EMAIL PROTECTED]
------------------------------
From: Buster_Bear <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.help,linux.redhat.install
Subject: Re: Cable Modem problems
Date: Fri, 22 Jan 1999 03:07:23 GMT
Hi
I have that same cable modem.. the 3Com 3c900 thernet card... and motorola
Cybersurfer cable modem. Hmm I have not ahd any problems with it though. I had
to download a dhcpcd program though to get it to work however so it catched my
IP which changes every once in a while. I am not sure if this will help you
but I used it and my cable modem is working. The PPP and Slip,, I am not
sure... but I don't think I need it so I would comment it is out in the
.config file for the kernel.
Dunno if any of this helps.. but I thought I should try to put my two cents
in.
Robert Montgomery wrote:
> Sorry, but I have no answers, but just to let you know, I'm in a similar
> position, so if anyone DOES know, they will probably be helping many
> of us...
>
> I've got a 3Com 3c900 ethernet card (which is tier-1 supported from RedHat
> 5.2)
> which runs to a Motorola CyberSurfer cable modem through the @Home network.
>
> The Network Configurator never seems to really "activate" the interface.
> I can try and activate the "eth0" interface and it appears to be active,
> until I
> close and re-open the Network Configurator again, only to find out it isnt
> active...
> I was noticing a message on the Console after every "activate" that said
> something like "delaying activation of eth0", but now even that has
> stopped...
>
> [EMAIL PROTECTED]
>
> Paul Tiseo wrote:
>
> > Well,
> >
> > All is good in linux land for me. All I have left (which is
> > apparently the source of much pain to newbies) is sound and networking.
> > So, here's a few questions for the more experienced:
> >
> > 1. Do I need to include PPP and SLIP in my kernel if I have
> > the one-way Mediaone cable modem whenever I get around
> > to a kernel recompile?
> >
> > 2. When I create an interface entry in the Network Configurator
> > in RH5.2, it never seems to activate. Why?
> >
> > 3. I looked into the DHCP HOWTO as it seems to be rather
> > important for cable modems. (Got that out of the cable modem
> > mini-HOWTO) The first thing one must do replace rc.inet1 in
> > /etc/rc.d with a new script. Problem? I don't seem to have an
> > rc.inet1 to replace! Does it have to do with an incomplete
> > networking installation?
> >
> > So many questions, so little time...
> >
> > -----------------------
> > Paul Tiseo
> > [EMAIL PROTECTED]
------------------------------
From: Dan Kegel <[EMAIL PROTECTED]>
Subject: Re: IP Masq & Civ 2 MPE
Date: Wed, 20 Jan 1999 20:03:31 -0800
Like I said, the upcoming Civ 3 from Activision won't
require you to do any port forwarding. It'll work
through NATs without any special configuration.
- Dan
Mark Vandersteen schrieb:
>
> You actually need to redirect the packets that come in from the outside
> world to you PC behind the firewall.
> Have a look at http://www.ox.compsoc.org.uk/~steve/portforwarding.html
> you will need to know which port the CIV2 games uses and redirect any
> incoming packets to the machine which runs the game.
>
> Mark Vandersteen [EMAIL PROTECTED]
> Powered by Linux ! Obsessed with women !
> The thoughts here are not necessariy mine !
>
> Dan Kegel wrote in message <[EMAIL PROTECTED]>...
> >Ray Clouse schrieb:
> >> I'm running into problems when trying to host a Civilization 2 game. I
> >> can join a game just fine, it's the hosting that's the problem. When
> >> people try to join, the get a "connection refused" from me, and, of
> >> course, they never show up in the "waiting for players" dialog. Here's
> >> my setup:
> >>
> >> P120, RH 5.1, 2.0.36, eth0->cable modem, eth1->10baseT hub
> >> K6-200, dual boot RH 5.2/Windoze95
> >>
> >> I think that using ipfwadm might fix the prob, but the docs I find on it
> >> are sparse. I'm hesitant about using it, cause then I might be open for
> >> other nastiness. Will that work, or do I need to wait for an
> >> ip_masq_civ2mpe.o module to be written?
> >
> >No need for a masq module. The game has to be made a little more
> >NAT-aware, I think. See
> >http://www.alumni.caltech.edu/~dank/peer-nat.html
> >Civilization [3]: Call To Power will definitely not suffer from
> >this problem.
> >- Dan
> >
> >--
> >Speaking only for myself, not for my employer
--
Speaking only for myself, not for my employer
------------------------------
From: Dan Kegel <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.hardware
Subject: Re: 3Com 3c509b, not working under RH 5.2
Date: Wed, 20 Jan 1999 20:06:20 -0800
Melvin Mathew Meadlin II schrieb:
> I have a 3com 3c509b ethernet card which works fine under Windows 95,
> but I can't get it to work under Red Hat 5.2. Does anyone have any
> suggestions????
Is that ISA or PCI? I have a PCI 3c509b, and it works fine with
Red Hat 5.2.
- Dan
--
Speaking only for myself, not for my employer
------------------------------
From: Wolfgang Kleinhappl <[EMAIL PROTECTED]>
Crossposted-To:
comp.dcom.lans.ethernet,comp.sys.sun.admin,comp.os.ms-windows.networking.win95
Subject: Re: Connect without hub
Date: Fri, 22 Jan 1999 04:37:23 +0100
hi:))
sure...it is possibel to connect 2 pcs directly via a crossovercable...
i have done it for myself and it works without problems...
bye:))
Jerry Mendes wrote:
> I may be wrong, but I believe 10BaseT and 10/100 cards need to see a "Link Beat"
> signal from a hub before they will work; therefore, a simple crossover cable
> won't work.
>
> It's so easy to buy an inexpensive hub, why not just do it?
>
> Paul E Larson wrote:
>
> > In article <77s6cr$6ap$[EMAIL PROTECTED]>, Mike Humski <[EMAIL PROTECTED]> wrote:
> > >
> > >I am planning to connect the following two machines with ethernet
> > >without using a hub.
> > >
> > >
> > >These machines are only two feet apart.
> > >I heard it is possible to directly connect two machines with
> > >ethernet cards if you use a crossover 10 baseT cable.
> > >
> > Find or make a crossover cable
> > (http://www.linksys.com/support/faqs/howbuild/wiring.htm)
> >
> > Paul
> >
> > Get rid of the blahs to email me :}
------------------------------
From: "Rob Lauer" <[EMAIL PROTECTED]>
Subject: Re: Extraction of IP address
Date: Wed, 20 Jan 1999 23:30:47 -0500
[EMAIL PROTECTED] wrote in message
<785q4k$78n$[EMAIL PROTECTED]>...
>Can somebody help me out? I'm trying to use a shell script that needs my
IP
>address (from the ppp0 interface), so that i can use a command such as:
>
> echo `date` : `myipaddress.sh` >>/var/log/temp.log
>
>Anyone know of any script or program to do this? Thanks!
>
> Luke Reeves
> [EMAIL PROTECTED]
> http://members.xoom.com/LukeReeves/
>
>-----------== Posted via Deja News, The Discussion Network ==----------
>http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own
It's ugly but it works...
ifconfig pp0 | grep "inet addr:" | cut -f 2 -d ":" | cut -f 1 -d " "
This should also be shell independent.
------------------------------
From: Ziae <[EMAIL PROTECTED]>
Subject: TCP/IP help
Date: 21 Jan 1999 03:55:13 GMT
I am a new user and am trying to connect Win95 and Linux PC's to connect
via Ethernet using TCP/IP. Can someone point me in the right direction
for getting some documentation? I can't even ping either machines!
Thanks!
[EMAIL PROTECTED]
------------------------------
From: Vincent Zweije <[EMAIL PROTECTED]>
Subject: Re: diald help.
Date: 20 Jan 1999 23:50:10 +0100
In article <[EMAIL PROTECTED]>, tony mollica
<[EMAIL PROTECTED]> wrote:
|| Hi. I have a working diald setup on a Debian2.0
|| box but there are a couple of messages in the log
|| that I can't seem to find a solution to. The items
|| are logged when the link comes down. These are
|| the lines with my comments added:
Please don't edit your logs.
|| : Closing down idle link ;no problem here
|| : Terminating on signal 2 ;can't find a reference
|| ;for this.
This is pppd shutting down because diald sent it a...
$ kill -l
1) SIGHUP 2) SIGINT 3) SIGQUIT 4) SIGILL
5) SIGTRAP 6) SIGABRT 7) SIGBUS 8) SIGFPE
9) SIGKILL 10) SIGUSR1 11) SIGSEGV 12) SIGUSR2
13) SIGPIPE 14) SIGALRM 15) SIGTERM 17) SIGCHLD
18) SIGCONT 19) SIGSTOP 20) SIGTSTP 21) SIGTTIN
22) SIGTTOU 23) SIGURG 24) SIGXCPU 25) SIGXFSZ
26) SIGVTALRM 27) SIGPROF 28) SIGWINCH 29) SIGIO
30) SIGPWR
$
... an INTerrupt signal. This is normal.
|| : Connection terminated ;no problem here either
|| : Failed ;what failed?
Well, maybe if you hadn't edited your logs, I could have said something
about this.
|| : disconnect script failed ;can't seem to get any
|| ;disconnect script to work.
|| ;Any suggestions?
You probably don't need a disconnect script at all, even though it's
specified somewhere in your diald or pppd configuration. Take it out.
|| : Exit. ;OK.
|| : Delaying 10 seconds before clear to dial.
||
|| I changed a parameter to make it 10 seconds between
|| redials, but why does diald end it's session with
|| this line, which makes you wait 10 seconds before
|| dialing a new connection?
There may be lingering (hung) connections which might produce traffic
for some while. You don't want these to bring the link back up. Then
again, maybe you might. It's just generally sensible to delay for a bit
between connections.
--
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: "Kyle Bowerman" <[EMAIL PROTECTED]>
Subject: Re: Multiple 3c509
Date: Thu, 21 Jan 1999 22:14:24 -0700
How did you get it to detect the second card?
Do you edit Lilo?
thanks for the Reply
Kyle
Sandy Culver wrote in message <[EMAIL PROTECTED]>...
>Greetings,
>
>For what it worth on 5.2 I am able to have two 3c509's at least tagged
>and noticed (not fully tested yet) and these are not compiled and are
>reachable as well by modprobe. I read somewhere that the lower
>numbered machine addressed ethernet card has to be called eth0.
>
>Regards,
>
>Sandy
>
>On Mon, 18 Jan 1999 20:19:19 -0700, "Kyle Bowerman"
><[EMAIL PROTECTED]> wrote:
>
>>I have two 3c509 nics running Slackware and can only find one.
>>I recomplied the kernel and specified the nic but it stil only find it if
I
>>decomment the driver in /etc/rc.d/rc.modules.
>>
>>I have tried the following combinations in lilo
>> 1. append="ether=0,0,eth0 ether=0,0,eth1"
>> and
>> 2. append="ether=10,0x300,eth0 ether=0,0,eth1"
>> and
>> 3. append="ether=10,0x300,eth0 ether=15,0x330,eth1"
>> and
>> 4. append="ether=0,0,eth0 ether=10,0x300,eth1"
>> and
>> 5. append="ether=10,0x300,eth0 ether=15,0x280,eth1"
>>
>>to no avail
>>
>>I read on of the post that said I had to put "alias eth1 3c509" in
>>/etc/moules.conf but I don't have this file and I added it and it did not
>>work either.
>>
>>An unusual thing happed when I took out the one that was discovered and
>>replace it with a 3c507. -- A had all linklights on the hub lite (this
>>didn't occur with both 3c509s) --and the transmit light on the 3c507 card
>>blinked and the transmit light of the 3c509 on the hub blinked.
>>
>>I have read the documentaion from http://cedis/....../multicard.html and
it
>>was not much help. It merely said that the pnp feature of the 3c509 made
>>addtion probling difficult.
>>
>>I also tried modprobe -ta net \* and it does not probe the 3c509 driver
>>
>>Is there a place that I cant bind this driver to eth1?
>>
>>Any suggestions?
>>
>>
>>
>>
>>
>
>============================================
>Mr.Sandy Culver fax: (978) 623-0082
>HR Consultant office: (978) 623-0942
> [EMAIL PROTECTED]
>============================================
------------------------------
From: Ron Thompson <[EMAIL PROTECTED]>
Subject: FTP files not visible
Date: Wed, 20 Jan 1999 20:18:39 -0800
Okay, from reading a lot of the other messages, it looks like I might
want to change to WU-FTPD (I'm not sure what I'm using right now), but
for now maybe someone can answer this question for me.
When I log into my ftp server as anonymous, I can't see any files. The
permissions are:
drwxr-xr-x 7 root root 1024 Aug 11 23:16 .
drwxr-xr-x 12 root root 1024 Jan 14 20:34 ..
-rw-r--r-- 1 root root 157 Jul 31 23:54 .welcome
-rw-r--r-- 1 root root 139 Aug 6 19:17 README
drw-r--r-- 2 root root 1024 Jul 17 1998 bin
drw-r--r-- 2 root root 1024 Aug 11 22:55 etc
drwxrwxr-- 2 root ftp 1024 Aug 6 19:12 incoming
drw-r--r-- 2 root root 1024 Jul 17 1998 lib
drwxr-xr-x 4 root ftp 1024 Jan 6 08:26 pub
I can cd to the directories, but ls, or ls -la doesn't see anything...
8-/
The permissions are fine for ls, and it doesn't give me any errors or
anything, so I'm just totally stumped... if I remember correctly, one of
my past configurations of anonymous FTP (on a long since wiped hard
drive) worked fine... 8-/
Hmmm... stumped...
-Ron ([EMAIL PROTECTED])
------------------------------
From: "Kyle Bowerman" <[EMAIL PROTECTED]>
Subject: Re: RedHat Linux 5.1 & Ethernet 3c509 NIC
Date: Thu, 21 Jan 1999 22:22:52 -0700
Try ifconfig eth0
and modprobe -ta net /*
and check /proc/interrupts and /proc/ioports for IRQs and Base addresses
Kyle
Abdul wrote in message <[EMAIL PROTECTED]>...
>Hello,
>
>I have 2 computers running RedHat Linux 5.1. The Pentium II runs fine
>with an NE2000 compatible NIC. The 80486 with 3c509 NIC, however, does
>not initialize eth0 and ifconfig only shows loopback. Here are the
>specifics:
>
>At boot:
>(1) First error message:
>
> Finding module dependencies: conf:2:missing module argument
>
> /lib/modules/preferred/modules.dep does not exist
>
>I checked and the file modules.dep does exist in the directory
>specified. I'm not sure if this has anything to do with the NIC
>problem. I don't know what other types of problems might occur because
>of this. I'm not sure why it's having trouble with this file. It looks
>like an ascii file.
>
>(2) The second message:
>
> Delaying eth0 initialization
>
>When I look at eth0 after bootup, it is not active. I activate it and
>restart the network daemon. ifconfig shows only loopback. Is there a
>way I can initialize eth0 at the command line, activate it, restart the
>network daemon and get this to work? Should I load RedHat Linux 5.2?
>Should I get another NE2000 compatible NIC? (This NIC worked fine with
>Windows 95.). I have all of the ip address and host information in the
>databases. It doesn't like it when I specify the ioaddress and irq, so
>I took that out. Are there any other diagnostic utilities I can run?
>
>I appreciate your time and any suggestions or references to resources
>you can point me to.
>
>[EMAIL PROTECTED]
>
------------------------------
From: David Kirkpatrick <[EMAIL PROTECTED]>
Subject: Re: Transmit time out
Date: Wed, 20 Jan 1999 23:24:30 +0000
Reply-To: [EMAIL PROTECTED]
Daren,
There is not "exact", error as you have written it. There is
one
like it though that is proably from the eepro100.c driver. The
error
message is generated by a call to speedo_tx_timeout. The code is
gotten to in the call in speedo_start_xmit by not returning after
some checks so continues on to executing speedo_tx_timeout.
The check looks for the time to be greater than some fixed
time
and if greater falls through to the timeout. I think the coder
suspects that something is wrong in the queue layer so falls
through
to speedo_tx_timeout. In speedo_tx_timeout he prints the timeout
error then checks for an 0x0080 which is probably a more fatal
error. If its that error he sends DRVR_INIT to the board (or
lower
level layer on the board I think). If not then tries to restart
by
sending a CU_START probably to do a milder start/restart.
This is not very specific but I think the problem is comming
from somewhere on the board, cables or remote but below the
linux eepro100.c driver.
David Kirkpatrick
[EMAIL PROTECTED]
Daren Jacobs wrote:
>
> I have just set up my Linux 5.2 server and I am getting the following error
> message:
> "
> eth0: transmit timed out status 0050 command 0000
> eth0: Trying to restart the transmitter"
> Can anyone assist with this as I have never seen this error message before.
>
> [EMAIL PROTECTED]
--
[EMAIL PROTECTED]
[EMAIL PROTECTED]
[EMAIL PROTECTED]
------------------------------
** 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
******************************
