Linux-Networking Digest #241, Volume #10 Thu, 18 Feb 99 13:13:19 EST
Contents:
linux networking questions (Allen Jantzen)
newbe, dial up PPP, busy. SuSE 6.0, KDE ("SteveC")
DHCP <-> DNS ("Oren Ben-Kiki")
activate software loopback interface ("Rudy Tuypens")
Re: Acer ALN ethernet problem (James)
Re: PPP problems under 2.2.1 (Clifford Kite)
Re: Can NT with NTFS coexist with RedHat Linux (Eric)
Re: Getting to the internet with my linux box ("Rocky Dean")
Re: Machine name themes - what do you use? (Malcolm Dunnett)
Re: MS Explorer 4.0 for Unix (Rowan Volvo)
Re: Linux on a MICROSOFT LAN (Yan Seiner)
Re: Changing IP address (Suresh Randawa)
Re: Poor man's failover? (Yan Seiner)
----------------------------------------------------------------------------
From: [EMAIL PROTECTED] (Allen Jantzen)
Subject: linux networking questions
Date: Wed, 17 Feb 1999 15:33:58 GMT
Well, sleep deprivation has started to kick in, so I am turning to
this NG for help. sorry....long message
I am a database admin/programmer w/ fair amount of HP-UX Unix
experience. I am attempting to connect a linux machine w/ 2 PCI PnP
NIC's to a cable modem. One 10baseT cable will goto the cable modem,
the other to a hub (w/ a PC attached). So proxies are next. Right
now, I am just trying to get the linux box working on the internet.
details: pentium 133, 32 megs, RH 5.2, Linux Kernel 2.0.36, IDE HD
Any internet access just hangs....it doesn't crash Linux (I can CTRL-C
out).....it just sits there.
A few questions:
==================================================================================
I am using 2 PCI PnP NIC's, both NE2000-compatible EtherLink II (I
believe). Here's the /bin/dmesg details:
ne2k-pci.c:v0.99L 2/7/98 D. Becker/P. Gortmaker
http://cesdis.gsfc.nasa.gov/linux/drivers/ne2k-pci.html
ne2k-pci.c: PCI NE2000 clone 'Winbond 89C940' at I/O 0xff40, IRQ 10.
eth0: PCI NE2000 found at 0xff40, IRQ 10, 00:20:78:16:22:9B.
ne2k-pci.c: PCI NE2000 clone 'Winbond 89C940' at I/O 0xff20, IRQ 9.
eth1: PCI NE2000 found at 0xff20, IRQ 9, 00:20:78:16:1F:10.
So linux is probing/finding/configuring both NIC's just fine. (Which
is interesting because the How-To's say it will *stop* probing after
the first NIC. It found both NIC's the very *first* time I booted,
w/o having to doctor /etc/lilo.conf or /etc/conf.modules. Is this new
in RH 5.2 or 2.0.36?)
If linux probes/finds/configures both NICS, should lo, eth0, and eth1
be "UP" and activated (ie /sbin/ifconfig) after boot? Only my eth0
and lo are "UP". Only /sbin/ifconfig -a shows eth1. Why would the
device eth1 not be activated if it has been successfully configured?
If I try a /sbin/ifconfig eth1 up, the machine just sits there
confused until I CTRL-C out.
======================================================================================
My dmesg also has the following:
ide: i82371 PIIX (Triton) on PCI bus 0 function 57
ide: BM-DMA base register is invalid (0x0000, PnP BIOS problem)
ide: setting BM-DMA base register to 0xe800
ide: BM-DMA base register is invalid (0x0000, PnP BIOS problem)
hda: QUANTUM FIREBALL EL5.1A, 4892MB w/418kB Cache, CHS=623/255/63
hdb: 685A, ATAPI CDROM drive
ide0 at 0x1f0-0x1f7,0x3f6 on irq 14
So there is some sort of PnP BIOS problem w/ the IDE HD??? The HD is
working great...so should this be a concern?
The only PnP/BIOS problem I've seen w/ ne2k-pci.c driver is when the
NIC is ISA. Mine are PCI.
=======================================================================================
After the linux box has been up for a few minutes, these start to
appear in dmesg:
ARP: arp called for own IP address
ARP: arp called for own IP address
ARP: arp called for own IP address
What the heck does this mean?
=========================================================================================
Even though I can't communicate on the internet thru eth0,
/sbin/ifconfig eth0 has values for RX and TX that keep going steadily
up. So the NIC is seeing something out there, and the something is
answering back??
===========================================================================================
I used as basis for my config another (working) linux box w/ cable
modem on same network as mine. So I think my hostname, DNS,
broadcast, mask and that stuff is OK.
Does anyone have any ideas?
Any help appreciated.
Allen Jantzen
------------------------------
From: "SteveC" <[EMAIL PROTECTED]>
Subject: newbe, dial up PPP, busy. SuSE 6.0, KDE
Date: Thu, 18 Feb 1999 00:19:03 -0000
Trying Kppp with all settings put in all I can get out is the modem is busy.
Trying all the howtos & mini howtos I tried to get minicom up. It
initializes the modem but then when trying to dial just sits there and times
out. Neither method actually makes the modem make a noise. Kernel compiled
ppp enabled.
I've tried everything I can & followed howtos. Only explanation I can think
of is that I have an IRQ conflict. This possible? Only way to change IRQs is
withing Win98 (ugh) I assume. Cards in machine:
BT848 capture card
SB16, on motherboard.
Modem, internal K65 com2 (been messing with the IRQ within windows, re
setting up in Linux and trying again)
Help! Would it help to post and config files, what can I do?
Thanks for replies,
SteveC [EMAIL PROTECTED]
http://www.fractalus.com/fracsaver ICQ#14047829
"In the time it takes you to read this sentence, eighty-six letters could
have been processed by your brain."
------------------------------
From: "Oren Ben-Kiki" <[EMAIL PROTECTED]>
Subject: DHCP <-> DNS
Date: Thu, 18 Feb 1999 17:35:19 +0200
We are running both DHCPD and a caching-only name server on the same machine
(RedHat 5.1). We'd really like the DNS server to also report the machines
which got their addresses via the DHCP daemon.
Has anyone ever written some magic which allows DHCPD to pass its leases to
the DNS server?
Thanks,
Oren Ben-Kiki
------------------------------
From: "Rudy Tuypens" <[EMAIL PROTECTED]>
Subject: activate software loopback interface
Date: Thu, 18 Feb 1999 01:12:27 +0100
ifconfig lo up <return>
or
ifconfig lo 127.0.0.1<return>
------------------------------
From: James <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.hardware
Subject: Re: Acer ALN ethernet problem
Date: Wed, 17 Feb 1999 20:43:16 -0400
Use ne2000 for your NIC type and you should get results. It's built into
the kernel.
Hope this helps.
ben domingue wrote:
> I can't seem to get any drivers (or support) for an Acer ALN-101
> Series pnp 10 mbps isa ethernet adapter. the disk comes with crazy
> sco unix drivers and lame installation programs but they're pretty
> much useless as far as i can tell. i'm running linux 2.0.36. any
> feedback or information would be appreciated.
------------------------------
From: [EMAIL PROTECTED] (Clifford Kite)
Subject: Re: PPP problems under 2.2.1
Date: 16 Feb 1999 22:46:49 -0600
Allan Olesen ([EMAIL PROTECTED]) wrote:
: [EMAIL PROTECTED] (Clifford Kite) wrote:
: >Well, I knew about asyncmap already. :) But now I do have a new symptom
: >which is indicative of an ACCM (Asyncronous Control Character Map)
: >problem.
: I too uses a Zyxel TA (OmniNet+). I experienced exactly the same
: behaviour until I put "acyncmap 0" in my /etc/ppp/options. I know of a
: few other Zyxel owners who have had the same problem and solved it
: this way.
: I really do not entirely understand the use of asyncmap [1], but
: perhaps there is a default setting which changed from version 2.2.0 to
: 2.3.5?
You seem to have hit on the answer. If the asyncmap is not given as an
option, then the default should be ffffffff. The default in ppp-2.2.0
must be 0000000, implied by this remark from the ppp-2.2.0f/pppd/lcp.c
file:
/*
* If the asyncmap hasn't been negotiated, we really should
* set the receive asyncmap to ffffffff, but we set it to 0
* for backwards contemptibility.
*/
The next statement sets the asyncmap to 00000000. The corresponding
statement in ppp-2.3.5/pppd/lcp.c sets it to ffffffff, which is the
correct default. Mystery solved.
: [1]: In which part of the communication process do you map the control
: characters? I thought that the purpose of hardware handshaking was
: eliminating control characters?
I *think* the purpose of the hardware flow control is to allow sending
and receiving 8 bits without using the software xonxoff handshake. This
makes for faster data exchange and eases the cpu burden.
The asyncmap defines which of the control characters 0-1f are to
be escaped when they are sent to pppd under hardware flow control.
The XON/XOFF (^S/^Q) start/stop characters and the telnet control-]
that drops telnet into the command mode are two that some PPP clients
might ask to have escaped.
--
Clifford Kite <[EMAIL PROTECTED]> Not a guru. (tm)
/* Microsoft is a great marketing organization.
* It _has_ to be */
------------------------------
From: Eric <[EMAIL PROTECTED]>
Crossposted-To:
comp.os.linux,comp.os.linux.admin,comp.os.linux.misc,comp.os.linux.questions
Subject: Re: Can NT with NTFS coexist with RedHat Linux
Date: Wed, 17 Feb 1999 16:03:50 -0700
Tat M. Leung wrote:
> I have a computer currently running NT server using NTFS. I like to add a
> second harddisk to the computer and install RedHat Linux on this drive. The
> second harddisk will be used for Linux exclusively. I understand that if NT
> is using FAT, there would not be any problem. Since my system is using
> NTFS, can I install RedHat Linux on this system?
>
> Thank you for any advise.
>
> Tat
For starters, you should not be running a pure NTFS install. This is an
absolute recipe for disaster. It is virtually impossible to recover (it can be
done but it really sucks) a NTFS system if you are booting from a NTFS
partition and it becomes corrupt. For safety you should have a small dos
partition (<50 mb's is fine), then create a NTFS partition on the rest of the
drive. In my case I have 2 drives, the first drive has dos and linux, the
second is NTFS. I have the NT loader files installed in the dos partition on
the first drive, and NT installed on the second. As far as the rest of your
question, install Linux as normal, set Lilo up to dual boot dos and Linux.
since the NT loader files are in the dos partition, as soon as you choose dos
you will get the NT dual boot menu and you can choose dos or NT. This is
extremely stable, easy to recover if you have problems, and is a hell of a lot
more stable than 3rd party crap like partition magic.
Eric
------------------------------
From: "Rocky Dean" <[EMAIL PROTECTED]>
Subject: Re: Getting to the internet with my linux box
Date: Thu, 18 Feb 1999 12:28:07 -0500
You can use Netscape in X, it uses proxy. also, if you want an IRC client,
the only one i've seen that works with socks proxy is sirc. but there is no
client software for linux to use msproxy.
Scott Baker wrote in message <[EMAIL PROTECTED]>...
>I have a linux box that is hooked up to an internal network. The
>other machines on the network are running NT. We have one machine
>that is responsible for calling our internet service provider, and it
>is through the proxy that all other NT machines access the internet.
>
>The question is, how do I get the Linux box to communicate with the
>proxy and see the internet?
>
>Thanks
>Scott
------------------------------
From: [EMAIL PROTECTED] (Malcolm Dunnett)
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: 17 Feb 99 16:27:40 -0700
In article <[EMAIL PROTECTED]>,
Dennis Clarke <[EMAIL PROTECTED]> writes:
>
> I'm giggling. I spent the afternoon at a company that is having
> stability problems with their DEC Alpha based NT boxes. I asked them
> what the uptime was like and they kinda shrugged and said about six to
> seven hours at a time. No kidding. I suggested scrapping them and
> putting in Solaris based Sun boxes and they looked at each other and
> then said - "We can't administer unix".
>
Why would you suggest they scrap their hardware when they can already
run multiple flavours of Unix ( "OSF1/Digital Unix/Tru64 Unix" or linux )
or ( even better ) VMS on the existing hardware?
> I said "What could be worse, a set of servers that crash three times a
> day - OR - a server that you look at once every week and then
> call in someone to do administration for you - remotely. Meanwhile your
> users and management are happy cause your email system hasn't stopped
> working. You have uptime measured in months?"
>
> You can guess the answer. People are learning - slowly - but they are
> learning.
>
Yep, now if some other folks could learn Solaris isn't the only
alternative to NT :-) <-- look, I'm smiling, keep the flames on low please!
--
=============================================================================
Malcolm Dunnett Malaspina University-College Email: [EMAIL PROTECTED]
Computer Services Nanaimo, B.C. CANADA V9R 5S5 Tel: (250)755-8738
"OpenVMS is today what Microsoft wants Windows NT v8.0 to be!"
posted on www.openvms.digital.com on or about Sep 22,1998
Quashed on Sep 23,1998
------------------------------
Crossposted-To:
alt.os.linux,alt.linux,linux.redhat,linux.redhat.misc,comp.windows.x.kde
From: [EMAIL PROTECTED] (Rowan Volvo)
Subject: Re: MS Explorer 4.0 for Unix
Date: Thu, 18 Feb 1999 16:42:16 GMT
In article <7a03so$[EMAIL PROTECTED]>,
Alexander Viro <[EMAIL PROTECTED]> wrote:
>People who think that there are only two browsers have no business doing web
>development. Period.
Like the USPO site that runs a script "if IE..... else if netscape... "
and gives a blank screen in red baron.
>People who never heard of validator.w3.org *definitely*
>should not be allowed to put a page. And those who think that HTML is OK
>if two browsers manage to parse it and show something have no fscking idea
>of HTML. Sheeesh...
so enlighten us. Being a linux user means you know the greatest bugaboo
seems to be undocumented or esoteric details like this.
>
>--
>"You're one of those condescending Unix computer users!"
>"Here's a nickel, kid. Get yourself a better computer" - Dilbert.
--
Never fly in an airplane that was designed by an optimist
------------------------------
From: Yan Seiner <[EMAIL PROTECTED]>
Subject: Re: Linux on a MICROSOFT LAN
Date: Thu, 18 Feb 1999 11:40:16 -0500
Reply-To: [EMAIL PROTECTED]
You should be using samba www.samba.org
Yan
patrick wrote:
>
> Greetings.
>
> I`m having big time difficulties trying to connect my linux workstation to
> the company LAN.
>
> How can I make my linux be on the LAN?
> My eth0 is working and I can ping others.
> Now I suppose all I have to do is to get access to the server, to the shared
> files ... and more.
>
> How can I do that?
> I have 3 linux books but the more I read the more I get confused.
> Help please.
>
> what should I use ? Samba or NFS ?
> any useful informations I can find on the net?
> Thanks in advance.
------------------------------
From: Suresh Randawa <[EMAIL PROTECTED]>
Subject: Re: Changing IP address
Date: Wed, 17 Feb 1999 19:44:49 -0500
"Shane S." wrote:
>
> Hi there,
>
> I was wondering if someone could help me out.
> Our Linux admin bailed out. We are changing our network IP address.
> I was wondering what are the files I need to edit to change the IP address,
> subnet mask, Gateway IP address, and DNS enteries?
>
> Many thanks
> Shane
/etc/resolv.conf for DNS.
/etc/sysconfig/network for stuff like gateways among others.
/etc/sysconfig/network-scripts/ has specific information on different
interfaces you can have.
I know this is exactly how it is on Red Hat systems, bit I don't know
about others. If you don't find all these files/directories, do a grep
for the IP #'s or names you are looking for in the /etc directory, or
find a file in /etc called network (or something like that).
--
Suresh Randawa
mailto:[EMAIL PROTECTED]
------------------------------
From: Yan Seiner <[EMAIL PROTECTED]>
Subject: Re: Poor man's failover?
Date: Thu, 18 Feb 1999 11:09:58 -0500
Reply-To: [EMAIL PROTECTED]
I run a similar setup with a linux box and an NT ws. I don't try to do
an automatic switchover, and my file dump takes place every night.
I also mount and umount the NT fs as I need it (the theory being that if
we get hacked and the data on the linux box is lost, at least the mirror
may be a little harder to get to....)
I have not figured out how to delete files from the mirror that are no
longer on the primary; disk space is not a problem (yet) and I don't
really trust autmated delete routines....
Let me know if you resolve that issue.
Yan
[EMAIL PROTECTED] wrote:
>
> I run a small hodge-podge of computers loosely referred to
> as a 'network.' We've got two servers, one in a dorm running
> the local LAN, the other across campus (on a T1-connected
> backbone) running our Internet services (web hosting, email).
> The two are connected with a pair of 33.6 modems.
>
> We're getting an xDSL line run into the dorm, and we're going
> to at the very least move the cross-campus server into the
> dorm. Since we'll have two relatively comparable machines,
> and reliability and up-time are going to be imperative, what
> I'm thinking of doing is this:
>
> +-------------------------------+ +-------------------------------+
> | EBHON SERVER | | SPRAWL SERVER |
> | eth0 eth1 | | eth1 eth0 |
> +-------------------------------+ +-------------------------------+
> # # # #
> # ################ #
> # #
> # ##############################################################
> # #
> # # +----------------+
> # # ##################| xDSL interface |
> # # # +----------------+
> +--------------+
> | HUB |
> +--------------|
>
> We're using a firewall, so all our local IP addresses are
> 'generic' Class-C (192.168.x.x). The configuration would
> be something like this:
>
> EBHON SPRAWL
> eth0 192.168.0.1 not assigned
> eth1 192.168.1.1 192.168.1.2
>
> Once a week or so, SPRAWL uses eth1 to mount whatever variable
> directories are on EBHON (nfs) and does a backup to the
> corresponding directories on its local filestructure. (EBHON
> also has a tape drive.)
>
> Additionally, SPRAWL would periodically ping EBHON using eth1.
> If for whatever reason it got no response, it would bring up
> its own eth0 as 192.168.0.1 and thereby pick up the network
> traffic destined for the primary server.
>
> I think I can do all of this with some fancy shell scripts
> and creative use of cron. It's cheap, and it just might work.
> But before I lose sleep and hair, is this a workable solution?
> I know the hardware duplication is a bit extreme, but we've
> had bad karma with hardware, and we've definitely got more
> than a surplus of hardware...
>
> Thanks all!
>
> -------------------------------------------------
> R. Christopher Harshman: [EMAIL PROTECTED]
>
> -----------== Posted via Deja News, The Discussion Network ==----------
> http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own
------------------------------
** 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
******************************
