Linux-Networking Digest #89, Volume #11 Sat, 8 May 99 23:13:40 EDT
Contents:
Re: mgetty blocks dialout (Bill Unruh)
Re: Redhat 6.0... the good, the bad, and the ugly (Bill Unruh)
Re: Reliable (!) nic for 2.2 kernel? (bryan)
Re: Reliable (!) nic for 2.2 kernel? (bryan)
Re: Reliable (!) nic for 2.2 kernel? (bryan)
Re: Reliable (!) nic for 2.2 kernel? (bryan)
Re: Problem with 100mbit Ethernet (bryan)
Re: IP Masquerade (Wisquatuk)
~~~~ WEBBOOK by SIMSON ~~~~ ("@T")
Telnet and external networks (Chris Moseng)
Upgrading kernel from 2.0 -> 2.2 broke PPP (Marcin Romaszewicz)
Re: RoadRunner problems ($teve)
Need help : AT1700 card no longer works with kernel 2.2 (Sven Liessem)
Re: Xircom CBEM56 ? (jeff)
Re: Reliable (!) nic for 2.2 kernel? (Don Baccus)
can't get IP from Cisco 675 ADSL router (DHCP) ("J.Maxwell")
----------------------------------------------------------------------------
From: [EMAIL PROTECTED] (Bill Unruh)
Subject: Re: mgetty blocks dialout
Date: 9 May 1999 01:09:52 GMT
In <[EMAIL PROTECTED]> Darren McClelland <[EMAIL PROTECTED]> writes:
>I am setting up a Linux box with dialin remote access and and a dial out
>PPP connection. The idea is to dial in to the box, hang up and then have
>it dial me back to a PPP server. This to provide some security and so
>that the long distance charges live at the remote box. I can dial in
>just fine and get a terminal and login, and I can dial out and establish
>a PPP connection, _but_ not both. When mgetty is running the dialout
>chat script never seems to talk to the modem and just hangs. Here's the
Run pppd with the lock option. (ie put the line
lock
into /etc/ppp/options)
mgetty sits there looking at the serial line. If it sees activity, it
checks to see if there is a lock file. If there is, it sits there and
checks every 20 sec or so to see if the lock file is still there. When
it disappears, it reinitialises the modem and keeps listening. for a
dialin.
Note that mgetty takes a little while to reset after it has answered teh
call. Thus if you pppd tries to phone out too soon after mgetty had
handled a call it will catch mgetty initialising the modem, and mgetty
will not get out of the way. Actually it puts up a lock file when this
happens, and pppd then just dies with a lock file found error.
------------------------------
From: [EMAIL PROTECTED] (Bill Unruh)
Crossposted-To: comp.os.linux.hardware,comp.os.linux.setup
Subject: Re: Redhat 6.0... the good, the bad, and the ugly
Date: 9 May 1999 01:13:01 GMT
In <[EMAIL PROTECTED]> James Stafford <[EMAIL PROTECTED]> writes:
>I was just at Fry's today and they had it there for$79.00 !!! That's more than I
>paid for Winblows, more worth it... but still!!!
And from cheapbytes.com, itis $1.99 plus shipping. However no manual, no
installation support.
------------------------------
From: bryan <[EMAIL PROTECTED]>
Subject: Re: Reliable (!) nic for 2.2 kernel?
Crossposted-To: comp.os.linux.development.system
Date: Sun, 09 May 1999 01:30:36 GMT
minor detail: I need a good PCI card, not ISA. 10/100 would be great,
but at least a clean 10 over PCI is minimal.
In comp.os.linux.networking Spiros Ioannou <[EMAIL PROTECTED]> wrote:
: In comp.os.linux.development.system bryan <[EMAIL PROTECTED]> wrote:
: : my tulip card is totally unreliable. I can bring it down with an ftp
: : has anyone done any load testing on the 2.2 kernel and found a
: : RELIABLE nic card they could recommend? one that stays up under close
: : to full load on a local 10/100 lan?
: With the 3c509 I get 8.5 Mbytes/sec with wget from 100Mbit switched ethernet
: lan. It has never crashed, the machine is a web/ftp server and completely
: reliable.
: (With the same card I get 3.5MB/sec max with W98) ;-)
: -Spiros
--
Bryan
------------------------------
From: bryan <[EMAIL PROTECTED]>
Subject: Re: Reliable (!) nic for 2.2 kernel?
Crossposted-To: comp.os.linux.development.system
Date: Sun, 09 May 1999 01:32:56 GMT
In comp.os.linux.networking Linus Torvalds <[EMAIL PROTECTED]> wrote:
: In article <0EXY2.11020$[EMAIL PROTECTED]>,
: bryan <[EMAIL PROTECTED]> wrote:
: >my tulip card is totally unreliable. I can bring it down with an ftp
: >xfer (local lan) at 10 or 100, in a minute or less. network hangs and
: >will NOT be reset by software.
: I would recommend the intel eepro100, although I also wonder whether you
: just have a flaky tulip card, because tulip would have been my second
: suggestion. There's a lot of different tulip-based cards out there..
I have 3 varieties:
- the original bay networks DEC card
- a card by a noname with Exact Same DEC chip on it
- a 'lite-on' pnic tulip clone
all 3 work identically. they work well for light traffic but under
heavy load (my ping flood, for example; or even a local rolled-cable
ftp at 100/full) the tulips lock up and need a system reboot. note
that the console (X, etc) is still alive and well; its just that the
net card goes out in the weeds and never comes back.)
--
Bryan
------------------------------
From: bryan <[EMAIL PROTECTED]>
Subject: Re: Reliable (!) nic for 2.2 kernel?
Crossposted-To: comp.os.linux.development.system
Date: Sun, 09 May 1999 01:35:15 GMT
In comp.os.linux.networking Preston F. Crow <[EMAIL PROTECTED]> wrote:
: I use a tulip card with no problems at all. However, I've noted that
: there are tons of different tulip cards, each with subtle differences.
: My particular card is sensitive to the version of the driver. With
: the latest 2.0.37-pre?? kernels, the included driver works fine. With
: the 2.2.x kernels, I have to install a different one from:
: http://www.bmen.tulane.edu/~siekas/driver.html
that's where I get the latest one from, myself. the kernel (shipped
with 2.2.x) is old and neither that nor the one listed above cure the
'lost in the weeds' problem.
: Interestingly, 2.0.37-pre?? has a newer version of tulip.c than 2.2.6
: (I haven't check 2.2.7). I would suggest trying several versions of
: the driver before giving up on the card, especially if it works under
: 2.0 but not 2.2. You might even be able to use the 2.0 driver under
: 2.2--I'm not sure if there are any complications in doing so.
if I wasn't running SMP, I'd go back to 2.0.36. but I get so much
more out of my sytem at 2.2 with smp (dual celerons) that I'd rather
find a stable card under this version than lose some smp efficiency
just for the sake of saving $50 on a nic card ;-)
--
Bryan
------------------------------
From: bryan <[EMAIL PROTECTED]>
Subject: Re: Reliable (!) nic for 2.2 kernel?
Crossposted-To: comp.os.linux.development.system
Date: Sun, 09 May 1999 01:39:01 GMT
In comp.os.linux.networking Leslie Mikesell <[EMAIL PROTECTED]> wrote:
: In article <0EXY2.11020$[EMAIL PROTECTED]>,
: bryan <[EMAIL PROTECTED]> wrote:
: >my tulip card is totally unreliable. I can bring it down with an ftp
: >xfer (local lan) at 10 or 100, in a minute or less. network hangs and
: >will NOT be reset by software.
: >
: >with a T1 download, it can hang the network in a few hours. this sucks ;-(
: >
: >has anyone done any load testing on the 2.2 kernel and found a
: >RELIABLE nic card they could recommend? one that stays up under close
: >to full load on a local 10/100 lan?
: >
: >(btw, I now have several dec tulip cards for sale. I refuse to use
: >them in 2.2 kernel boxes I have here - sigh.)
: Hmmm, mine (an early Linksys with a real Dec chip) seems to be working:
: cat /proc/net/dev
: Inter-| Receive | Transmit
: face |bytes packets errs drop fifo frame compressed multicast|bytes packets
:errs drop fifo colls carrier compressed
: lo:585407792 2204419 0 0 0 0 0 0 585407792 2204419
: 0 0 0 0 0 0
: eth0:3304241413 76566475 0 0 0 0 0 0 3709840852
:42453440 13 0 2 0 11 0
: I have several of these, some Intel 10/100 pros, and some of the 3Coms that
: Dell builds into their motherboards. I have never had any problems
: with the Intels. The only problem with the others have been occasional
: mis-negotiations of 10/100 or half/full duplex with certain hubs or
: switches.
10/100 speed always works. full/half is broken - almost everywhere.
I spent 2 yrs managing a large campus network (lots of cisco and ctron
hi-end hubs) and the duplex thing was always messing with us. nasty
stuff - if in doubt, set the hub or the host to auto but THE OTHER to
a fixed state. it makes the negotiation go a bit more reliably.
: I've never seen a problem where the card works correctly
: for a while and then fails.
its totally repeatable. I wonder if its my SMP that is throwing a
monkey wrench into the works? is anyone happy with their tulip in
2.2.7 AND smp??
: If you have this with more than one card
: I would suspect some other problem.
its DEFINITELY not the card.
: Try connecting to a different hub
: or if you are using a switch lock it in full duplex, 100M for your
: port.
its a rolled cable. ie, NO hub at all - just a crossover cable.
nothing to fail ;-)
: You might also have cable problems that only show up under
: certain conditions.
not in this case. the cable is brand new and only about 6 feet long.
and it works when I switch this card out to a different card.
: 100M equipment can be more sensitive to this
: than the 10M versions, even when running at 10M. Be sure everything
: is cat 5.
its cat5 - but even 6 feet of cat3 would work fine for 100/full.
--
Bryan
------------------------------
From: bryan <[EMAIL PROTECTED]>
Subject: Re: Problem with 100mbit Ethernet
Date: Sun, 09 May 1999 01:40:56 GMT
can you force to 100/full on BOTH ends? disable auto-negotiation (if
possible) and try again.
Moe Koenig <[EMAIL PROTECTED]> wrote:
: i have a problem with my 100mbit ethernet here..
: i connected two machines (linux/win98) using a twisted pair crossover
: cable, both are equipped with 100mbit ethernet cards..
: it all runs pretty fine, the only problem is that i keep getting
: overrun errors on the windows machine (reported by the linux box,
: wouldn't know how to find out on the winbox itself) and transfer rate
: never gets any higher than 2,2mb/s.
: also there is quite an amount of receive errors (linux box).
: this is what ifconfig reports (after some data went thru the cable):
: eth0 Link encap:10Mbps Ethernet HWaddr 00:00:E8:6F:DF:5E
: UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
: RX packets:1058918 errors:12617 dropped:0 overruns:0
: TX packets:0 errors:0 dropped:0 overruns:9857067
: Interrupt:11 Base address:0xfc00
: i already tried switching the ethernet card inside the win-box to
: half-duplex and other modes but it didn't help.
: does anyone have any idea how i could fix that?
: (all services work as well as ftp transfers, its just annoying to
: know that its not properly configured and that its not as fast as
: it could be)
: this all happens on a 2.2.6 linux using donald becker's rtl8139.c
: ethernet-driver.
: not knowing if its relevant i'd like to add that i do get errors
: and overruns from my isdn-dialup too:
: ippp0 Link encap:Point-Point Protocol
: UP POINTOPOINT RUNNING NOARP MTU:1500 Metric:1
: RX packets:16644722 errors:16569 dropped:0 overruns:0
: TX packets:0 errors:0 dropped:0 overruns:916770
: probably its just "normal", do you get same output from your dialups?
: my linux box runs on "only" a 100mhz pentium, could that be the
: reason?
: any help is highly appreciated!
: --
: Moritz Koenig [email:[EMAIL PROTECTED] fax:089-666-1718-659]
: [ Wir programmieren Ihren Erfolg! @ http://www.holoplex.de ]
--
Bryan
------------------------------
From: Wisquatuk <[EMAIL PROTECTED]>
Subject: Re: IP Masquerade
Date: 3 May 1999 18:07:28 GMT
Tero Niemi <[EMAIL PROTECTED]> wrote:
> Wisquatuk wrote:
>
>> All data would get sent back to 123.123.123.123, which would then
>> easily think, 'Oh, this data is a part of the connection that
>> 192.168.1.2 made to the outside world. I'll send it back to
>> 192.168.1.2 then.'
>
> Does this work for ping's and other 'straight' internet
> connections too?
>
Well, I'm not sure what you mean by 'straight' -- if you mean UDP
(connectionless), then I'm really not sure, and would tend to think
not. But I know that if you enable ICMP Masquerading in the kernel
config, yes, ping will work, since every echo-request packet should
have one and only one echo-reply packet, so it's simple for the
masquerading server to forward the right ones. I use masquerading
24/7, and it operates pretty much transparently, provided I have the
ip_masq_ftp module installed to masquerade FTP packets. (I upgraded
the kernel and made some mistakes installing the modules.. boy did I
notice quickly.) :)
--
- Wisquatuk (name[1..4]@netrover.com to email)
=====BEGIN GEEK CODE BLOCK=====
Version: 3.12
GCS/CC/M d-(--) a--- C++(+++)>$ UL++++>$ P+++ L+++ E>++ W-(+>++) N+++
!o>++ K- w--- O- M- !V PS++(+++) PE- Y+ PGP+++@ t+@ 5 X+++@ R+ tv b+
DI+@ D+ G>+++ e- h!(++) !r z
======END GEEK CODE BLOCK======
------------------------------
From: "@T" <[EMAIL PROTECTED]>
Subject: ~~~~ WEBBOOK by SIMSON ~~~~
Date: Sat, 8 May 1999 22:35:03 +0800
Hi,
Anyone knows where i can download webbook web-based address book by simson
???
Thanks in advance
------------------------------
From: Chris Moseng <[EMAIL PROTECTED]>
Subject: Telnet and external networks
Date: Sat, 08 May 1999 21:17:08 -0500
I've got a DSL connection to the world, and it operates well.
The short story is that all connections I can think to make outgoing
work fine (telnet, www, everything). This is even with a masqeraded
machine, miracle of all miracles.
The downside is that incoming telnet connections do not happen. I can
ping from the outside-in, traceroute, and finger, but can't telnet. When
I try from a VAX, I'm told "No route to host," with other clients, it
just times-out. This operation works fine on my internal network.
I haven't had a chance to set up apache, so I don't know what the status
of http is...
Is this a configuration problem on my end (I'm all set to Slackware
defaults, as far as telnet goes), or does my ISP maybe have to allow
such connections explicitly?
And, yes, I am talking about a real IP address, as far as I can figure,
and not about my masq'ed machines.
Thanks for any suggestions.
------------------------------
From: Marcin Romaszewicz <[EMAIL PROTECTED]>
Subject: Upgrading kernel from 2.0 -> 2.2 broke PPP
Date: 08 May 1999 19:32:24 PDT
Reply-To: [EMAIL PROTECTED]
Hi All,
I've upgraded my kernel from 2.0.36 to the latest 2.2, and I can no
longer use pppd (with a simple chat script) to establish connections.
The modem dials in, a PPP connection is established, but no ppp0 device
is created, nor do I get entries in the routing table.
I can manually add interface ppp0 and make it the default route, but
I would like to know what the underlying cause is. I figured that maybe
the values of some constants changed, so I recompiled pppd, and I
still see the problem (with all pppd's versions 2.2.0 and up).
Does anyone know if pppd has some dependencies which need to be
recompiled
to work with kernel 2.2.x?
Thanks in advance.
-- Marcin
------------------------------
From: [EMAIL PROTECTED] ($teve)
Subject: Re: RoadRunner problems
Date: Sun, 09 May 1999 02:39:11 GMT
if RIP is happening, i don't know enough about Linux to tell. it is
quite possible, since I selected 'everything' in the install . I do
know that IP forwarding is off.
The Linux box is the only device connected to the cable modem - no
hub, not other devices, no IP masq going on ... I do know that named,
httpd, etc are running.
$teve
On Sat, 8 May 1999 17:10:57 -0400, "Doug Pitek" <[EMAIL PROTECTED]>
wrote:
>do you have any rip packets transferring on servers for different routes..
>
>because that's all it sounds like...
>
>$ <[EMAIL PROTECTED]> wrote in message
>news:[EMAIL PROTECTED]...
>> Hi!
>>
>> I live in El Paso TX and use Time Warner RR. Been playing with Linux -
>> I
>> am a powerful and infulential (translate harried and stressed) LAN
>> supervisor (translate whipping boy) of the EPISD. I do have lots of
>> hardware and software to play with. We are using Linux ( Caldera
>> mostly) in production as DNS, Apache servers, etc. I also manage about
>> 1000 seats on a 80 NW 4.11 LAN/WAN.
>>
>> Been playing with Red Hat at home. I will be getting 6.0 next week.
>>
>> I found some old hardware, I am running RH 5.0 (2.0.31) on an HP
>> Netserver LC 5/133 with an Intel E100B nic. I applied your excellent
>> scripts ( I had to guess at our login server - I used the DNS/DHCP
>> address, so that may be part of the problem). All worked swimmingly,
>> and
>> I _do_ login - I know I am because if I misconfigure my password in
>> rrconf, I can't ping or resolve anything).
>>
>> The puzzling part is this - about 3 minutes after boot, most of the
>> world becomes unreachable (according to ping). The only thing I can
>> still ping is my default gateway. Prior to this, I can surf a bit with
>> Lynx, ping, etc. If I do a traceroute the "unreachable" thing happens
>> sooner. I never see anything other than * * * in the traces. Name
>> resolution is fine until it bombs.
>>
>> This has driven me crazy for about 3 days! If we can fix this, I bet
>> it
>> will help a lot of other people in El Paso.
>>
>> I also did the RH install with "everything" checked, so I am running
>> named, http, sendmail, etc. I wonder if these daemons are somehow
>> freaking out RR. I am too newbieish to know how to 'comment them out',
>> but I am learning fast. I did have a lot of Unix in college, but that
>> was on a vax and a PDP-11 20 years ago ....
>>
>> Any suggestions would be great!
>>
>> $teve Crye
>> [EMAIL PROTECTED]
>>
>
>
Please remove NOSPAM to reply via e-mail
------------------------------
From: Sven Liessem <[EMAIL PROTECTED]>
Subject: Need help : AT1700 card no longer works with kernel 2.2
Date: Sat, 08 May 1999 22:47:26 -0400
Hi,
I have an AT1700 card that works fine with kernel version 2.0.36.
After upgrading the kernel to 2.2.7 the card no longer works. It is
correctly
detected at boot time but immediately after that the driver seems to
turn
it off.
In the same manner the install disk (ftp method) for Redhat 6.0 is
unable to
initialize the card (it says it can't find it even if I specify the io
and irq). This
worked perfectly with Redhat 5.2
Is there anything I can do about that or missed I something important ?
Any help would be highly appreciated.
Sven
------------------------------
From: jeff <[EMAIL PROTECTED]>
Subject: Re: Xircom CBEM56 ?
Date: Sun, 09 May 1999 02:03:18 +0000
Dmitry Koshkin wrote:
> Hi, is it true that Xircom Ethernet Card + Modem is not supported by
> RedHat5.1/5.2/6.0
> in Laptop Dell Latitute.
>
> I can't believe in it.
>
> Dmitry
Well, the CEM56 10mbit LAN+modem works like a charm.
I'm not certain if the REM56 version works, though 32 bit support is
still in its early stages. Whether the Dell's have a problem with the
cards, I have NO idea.
------------------------------
Crossposted-To: comp.os.linux.development.system
Subject: Re: Reliable (!) nic for 2.2 kernel?
From: [EMAIL PROTECTED] (Don Baccus)
Date: 8 May 1999 19:54:34 PST
In article <0z5Z2.14665$[EMAIL PROTECTED]>,
bryan <[EMAIL PROTECTED]> wrote:
>minor detail: I need a good PCI card, not ISA. 10/100 would be great,
>but at least a clean 10 over PCI is minimal.
>: With the 3c509
?? 3c509b comes in PCI as well as ISA flavor
--
- Don Baccus, Portland OR <[EMAIL PROTECTED]>
Nature photos, on-line guides, at http://donb.photo.net
------------------------------
From: "J.Maxwell" <[EMAIL PROTECTED]>
Subject: can't get IP from Cisco 675 ADSL router (DHCP)
Date: Sat, 8 May 1999 18:56:57 -0500
I am having problems being assigned an IP address from my Cisco675 ADSL
router (which acts as a DHCP client assigning 192.168.x.x IP's to all the
computer on my mini home network.) Does anyone here use this ADSL router
with Linux?
I am using RedHat 5.2, 2.0.36
Thanks in advance,
-Jeff
------------------------------
** 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
******************************
