Linux-Networking Digest #620, Volume #10 Wed, 24 Mar 99 19:13:31 EST
Contents:
Re: IBM Token Ring card under RH5.2 (Alex Lang)
Re: Bought modem to work in LINUX! (bill davidsen)
Re: supra modem (Miguel Cruz)
Re: Any Masquerading Guru's out there? (Don Baccus)
Re: using linux box to console to cisco ("William R. Mattil")
Re: Linux - NFS mounting a UCX export drive. (Stephen Ashley)
Re: Can Linux Do This ??? ("The Lone Scribe")
Re: multi-link PPP? (bill davidsen)
ipchains, masq, question ("Scott MacDonald")
Re: Problem with Mail ("The Lone Scribe")
Re: Answering Software ("The Lone Scribe")
Solution: Multiple Subnets, WINS, and Problems Logging onto NT Domains (Cory R. King)
Re: Linuxconf and passwd rules (Rick Miller)
Re: Linux smaller than Windows NT ? (bill davidsen)
Remote printing to HP ([EMAIL PROTECTED])
Re: Yet another pppd/modem problem..... (Mike)
Network unreachable or server down. ("Georgios")
Re: ipchains, ipforwarding, kernel 2.2.3ac1 and ftp ("Dan")
----------------------------------------------------------------------------
From: Alex Lang <[EMAIL PROTECTED]>
Subject: Re: IBM Token Ring card under RH5.2
Date: Wed, 24 Mar 1999 23:14:58 GMT
PCI Token Ring cards are not yet supported in Linux. I believe that there is work
underway, but not complete.
David Aubespin wrote:
> Hi all,
>
> I am trying to set up a network access under my brand new RH5.2 but it
> doesn't work at all. I am using a IBM Token Ring Card (PCI), so I should have
> an interface tr0 using the module ibmtr (from what I've understood). I've
> configured everything via netcgf (static IP address, etc) + kernelcfg (for the
> ibmtr module), but when I try to activate the tr0 interface, I get a 'Delaying
> tr0 initialization'. Besides, I get a ibmtr: module not found at boot time...
> (although I compiled a new kernel..).
> If anyone can help, that would be GREAT!!!
> Thanks.
> david
------------------------------
From: [EMAIL PROTECTED] (bill davidsen)
Subject: Re: Bought modem to work in LINUX!
Date: 24 Mar 1999 23:15:34 GMT
In article <q0_J2.5835$[EMAIL PROTECTED]>,
Bob D <[EMAIL PROTECTED]> wrote:
| I also bought this modem to upgrade from a 28.8 on my Lan gateway machine.
| However, I can't get the ppp script to work. It dies waiting for ATZ to
| return an 'OK'. I can manually dial out using minicom but the modem response
| to any AT commands come back at the rate of 1 character every 6 seconds!
| Unfortunately there is no help for the Linux user on the 3Com website or
| knowledge base. Anyone using this modem successfully with a ppp dial-up
| script please help!
You have two problems here. One is that you probably haven't used isapnp
and setserial to get this modem configured in the right irq and io
places, or you didn't disable your mobo serial port, if you're using
com1/com2 addresses. That accounts for the slow part.
The lack of OK is because ATZ returns to the last user config saved, and
I bet you either set it wrong or didn't set it at all. It is *not* a
generic "figure out what I need and set it" command.
For USR modems, before you try any fancy init, use AT&F1 and nothing
else. Ignore the people who give you big long complex things, this is
the hardware flow control setting, and if you use crtscts in your ppp
script you should be 99% of the way there.
Other settings which might be of use involve &C and &D, possible &K, but
do *not* play unless something doesn't work. The &D3 setting is useful
for modems used for dial-in, it does a hard reset and probably isn't
needed or may be default by now. Start with the factory setting, the USR
folks got it right (and I have precious little good to say about them).
--
bill davidsen <[EMAIL PROTECTED]> CTO, TMR Associates, Inc
What I find astonishing is not that my cat has started to sing, but that
he has taken up country-western. This morning he sang `Momma, don't let
your kittens grow up to be barn cats' in the shower, followed by a
pretty decent yodeling version of `Roundup time in Texas when the catnip
is in bloom.'
------------------------------
From: [EMAIL PROTECTED] (Miguel Cruz)
Crossposted-To: comp.dcom.modems.cable
Subject: Re: supra modem
Date: 13 Mar 1999 01:05:21 GMT
In article <[EMAIL PROTECTED]>, <[EMAIL PROTECTED]> wrote:
> I have supra PCI modem ,when i enter enter the following
> how I can kown that my modem in not winmodem
Bad news: Almost all PCI modems are winmodems and thus not compatible with
Linux.
I've set follow-ups to comp.os.linux.networking, where they can help you
further.
miguel
------------------------------
Crossposted-To: comp.os.linux.setup,comp.os.linux.questions
Subject: Re: Any Masquerading Guru's out there?
From: [EMAIL PROTECTED] (Don Baccus)
Date: 12 Mar 1999 17:11:24 PST
In article <[EMAIL PROTECTED]>, GSM <[EMAIL PROTECTED]> wrote:
>Problem: How does one setup IPchains to allow filtering and setting up
>of class C reserved names with IPchains ie I want to masquarade the
>clients on the internal lan ? I have a static IP for linux box/hub and a
>couple of clients on a small baseT LAN.
Read all about it, starting from the ip-masquerading resource
page:
http://www.tor.shaw.wave.ca/~ambrose/index.html
It includes information on configuring your linux firewall,
as well as Win95 etc clients (which is no different than
you would for any other network with a gateway at a fixed
IP). I just set the primary and secondary DNS servers on
my windows boxes to be those of my ISP, and put names of
local boxes in the c:\windows\hosts file which is identical
to the /etc/hosts file on my linux box. Set the gateway
on the clients to 192.168.1.1 and you're off and running,
once you get masquerading working with ip chains (which I
haven't looked into as I'm running 2.0.36 which uses
ipfwadm, however I'm sure the principal's the same).
--
- Don Baccus, Portland OR <[EMAIL PROTECTED]>
Nature photos, on-line guides, at http://donb.photo.net
------------------------------
From: "William R. Mattil" <[EMAIL PROTECTED]>
Subject: Re: using linux box to console to cisco
Date: Wed, 24 Mar 1999 12:37:52 -0600
Albert Mintz wrote:
> Hi all,
> I need help seeting up A serial port to console to A cisco router.I
> looked at the minicom setup options and don't really understand how to
> make it work as A null connection.If anyone is using linux As A
> console i'd really appreciate if you'd share with me how to set it up.
>
> Thanks
Assuming that the cable is correct and applied to the proper port:
cu -l /dev/ttySX (x=[0-4]) if no getty is running this can be replaced
by:
cu -l /dev/cauX (x=[0-4] )
man cu will provide all of the gory details
Regards
Bill
--
William R. Mattil | Fred Astaire wasn't so great.
[EMAIL PROTECTED] | Ginger had to do it all backwards
(972) 399-4106 | and... in high heels.
------------------------------
From: Stephen Ashley <[EMAIL PROTECTED]>
Crossposted-To: vmsnet.networks.tcp-ip.ucx,vmsnet.networks.tcp-ip.misc
Subject: Re: Linux - NFS mounting a UCX export drive.
Date: Wed, 24 Mar 1999 23:07:46 +0000
Yep, I have several file systems mounted (nfs) from a Digital Unix host.
I have tried to umount those file systems and only mount the UCX VMS nfs
fs I was after. This has mabe no differance, the same message is output.
;>I have been trying to mount a VMS drive on my Linux work station. UCX
is
;>installed :-
;> DIGITAL TCP/IP Services for OpenVMS VAX Version V4.2 - ECO 1
;> on a VAXstation 4000-96 running OpenVMS V6.2
;>I get the following error in the linux system: -
;>mount: wrong fs type, bad option, bad superblock on yakka:/dkdata,
;> or too many mounted file systems
;>I have had no problems mounting NFS exported file systems from a
Digital
;>UNIX host. The DUNIX host also has no problems mounting the VMS file
;>system.
;>Any idea's.
Oswald Knoppers wrote:
> Stephen Ashley wrote:
>
> > mount: wrong fs type, bad option, bad superblock on yakka:/dkdata,
> > or too many mounted file systems
>
> Does your kernel have nfs support (either built in or as a module)?
>
> Regards,
>
> Oswald
------------------------------
From: "The Lone Scribe" <[EMAIL PROTECTED]>
Subject: Re: Can Linux Do This ???
Date: Wed, 24 Mar 1999 11:37:55 -0800
peter wrote in message <[EMAIL PROTECTED]>...
>1) Does linux support 56k v90 modems ???
Yes. But to take advantage of v.90, your ISP has to have v.90 modems
installed at their POP(s).
>2) Now, linux also supports multiple modems, does my IP have to
>support it as well???
Yes, your ISP will need to sell you a 'shotgun' type of account (if
available) so that you can use two phone lines/modem as a pair, not as two
separate accounts. I've never tried this with Linux because it was more cost
effective to get a xDSL account instead. Shotgun accounts were twice the
cost of a standard dial-up ($40/mo.), and needed a second phone line
($12/mo.), bringing the total cost to a bit higher than DSL ($49/mo.) at a
fraction of the speed of DSL, and I saved even more because I eliminated my
voice phone line (you can use both DSL and voice at the same time on the
same line). But perhaps someone else has done this and will post a solution
for you if you can't get affordable DSL in your area.
------------------------------
From: [EMAIL PROTECTED] (bill davidsen)
Subject: Re: multi-link PPP?
Date: 24 Mar 1999 23:25:38 GMT
In article <[EMAIL PROTECTED]>,
Christopher A. Gaul <[EMAIL PROTECTED]> wrote:
| Is there support for Multi-link PPP in the built in Linux PPP driver or
| via add on software?
I believe it's all in add-ons.
| Here's what I want to do. I want to put a Linux box on our corporate
| network with four modems in it. Then I want to enable two sys-admins to
| dial-up to two modems each and aggregate the bandwidth into one 67.2Kbps
| link (i.e. 33.6 x2) Also I want the dial-up connection to be assigned
| one IP that is one of the fixed IP's assigned to our Ethernet LAN so
| that the dial-up's appear to be on the LAN (which is on the Internet).
| This way the dial-up boxes can also be reached from the Internet as they
| will have legitimate Internet IP's. The dial-up boxes do not have to be
| Windows boxes, in fact I prefer they were Linux boxes.
There are three ways to attack this. You can get the multilink stuff,
learn it, install it, play with it, and have something which has a
chance of working with Windows if you must.
Then there are the better ways;-)
You can use equal cost routing on both ends and a bit of fancy routing
to get the machines to do load balancing. I got this up once as proof of
concept, didn't look too hard but I never did a script or anything.
You can get the eql package and use that. It works well, I used it a lot
for a while, but when ISPs started doing 56k I kind of got out of it,
since it needed a good bit of setup and used 2/3 of the incoming modem
pool. Also, I only used it in the middle of the night to avoid tying up
too many modems on the ISP during peak usage.
Either of the later two ways seems easier to setup. I looked at
multilink info, but since I have no connections which use it, it buys me
nothing.
--
bill davidsen <[EMAIL PROTECTED]> CTO, TMR Associates, Inc
What I find astonishing is not that my cat has started to sing, but that
he has taken up country-western. This morning he sang `Momma, don't let
your kittens grow up to be barn cats' in the shower, followed by a
pretty decent yodeling version of `Roundup time in Texas when the catnip
is in bloom.'
------------------------------
From: "Scott MacDonald" <[EMAIL PROTECTED]>
Subject: ipchains, masq, question
Date: Wed, 24 Mar 1999 13:49:26 -0600
Hi,
If an ip address is being masqueraded behind a linux firewall, can the real
ip address still be seen by other types of hacking software like port
scanning software? Or will that ip only be seen from that subnetwork if it
is active on the internet or passing packets through the firewall?
Scott
------------------------------
From: "The Lone Scribe" <[EMAIL PROTECTED]>
Subject: Re: Problem with Mail
Date: Wed, 24 Mar 1999 11:49:16 -0800
Moses Ling wrote in message <[EMAIL PROTECTED]>...
> I was think maybe I can have my linux box pull all the mail from the
>POP and the reroute them to each clients. What I have in mind is to have
>some kind of "KeyWord" in the "Subject" field or some where in the
>message to tell who the mail belongs to under my local network.
Try procmail (type 'man procmail' for info). Of course, it depends on your
incoming mail having the keywords that you want to procmail scan for. If you
can train the folks who send you email to do that, that is. The keyword
doesn't necessarily have to be in the subject line, either. It can be
anywhere in the email, not just the header.
What you could do is set up an account ('postman' or something like that)
that runs fetchmail at regular intervals (or whenever your ppp connection
comes up), and write a .procmailrc for that account, which will then
redistribute the mail automatically. You can end .procmailrc with a
'catch-all' that will route all email that doesn't have any keywords to be
sent to your (or another) account for manual redistribution, and so that you
can tune .procmailrc even more.
------------------------------
From: "The Lone Scribe" <[EMAIL PROTECTED]>
Subject: Re: Answering Software
Date: Wed, 24 Mar 1999 11:53:57 -0800
Landen Stoker wrote in message <[EMAIL PROTECTED]>...
>Is there any software out there that can be used for voice mail with a
>modem? I currently run it on a NT box but would like to move it to my
>Linux one. I have looked around but have not been able to find any.
vgetty will do that, usually included in most Linux distributions. I was on
freshmeat the other night and saw that someone is writing a multiple
voicemailbox application, but it's still extremely beta.
------------------------------
From: [EMAIL PROTECTED] (Cory R. King)
Crossposted-To: linux.samba
Subject: Solution: Multiple Subnets, WINS, and Problems Logging onto NT Domains
Date: Wed, 24 Mar 1999 18:43:12 GMT
Greetings, I figured since this caused me so much trouble, I might as
well post the solution to save other people the grief.
Configuration:
2 subnets, routed /w a linux box. Linux samba running WINS, a Windows
NT PDC on one subnet, Win95/98/WfW clients on other subnet logging
onto Windows NT PDC.
Nitpicks:
Linux 2.0.1 and up
Samba 1.9 and up
WinNT 4.0sp4
Problem: Win95/98/WfW clients complain they cannot find a PDC to
validate their password.
Other sympoms: Windows NT complains about not being domain master
browser and a whole slew of other odd error messages all found in the
event log.
Resolution: Reboot the NT box!
Why?
This is my 99% acurate conclusion ('er theory):
When Clients try to find a PDC/BDC to authenticate with, they go to
their WINS server to find where the PDC is for the domain:
NTDOMAIN<1c>
When the NT box boots, it adds the above entry into the Wins database
on the linux box
In my case:
"MYDOMAIN#1c" 922557425 xxx.xxx.xxx.xxx e4R
where xxx.xxx.xxx.xxx is the IP address of my NT domain controller
Without this entry, your clients cannot find a place to log onto, and
you get the "cannot find a domain controller"
The other thing to make sure of is that PDC is also the domain master
browser.
NTDOMAIN<1b>
You should see this entry in your wins.dat file:
"MYDOMAIN1b" 922557425 xxx.xxx.xxx.xxx 64R
where xxx.xxx.xxx.xxx is, of course, the NT domain controller
Those entrys not there? You need to reboot your NT PDC to let it
write those entrys to the WINS database
Now my question; what are the MYDOMAIN<00> and MYDOMAIN<1e> entrys?
Thanks, and I hope this helps somebody with the same problems I had!
Cory R. King
Azalea Software Inc.
http://www.azalea.com
------------------------------
From: Rick Miller <[EMAIL PROTECTED]>
Crossposted-To: linux.redhat.misc
Subject: Re: Linuxconf and passwd rules
Date: Wed, 24 Mar 1999 15:02:21 -0500
This is a multi-part message in MIME format.
==============536A6944C7E84B55DDCB8E53
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Well, yes, it accepts it, for root. My users are not able to change the
password unless they select something which Linux will accept which is usually
something way off the wall. I don't want it to do this. It seems as though
it is checking the password against a dictionary before assigning it. I do
not want it to do this.
Kevin Dempster wrote:
> Rick Miller wrote:
>
> > I have a RH 5.2 box running linuxconf. When I add a new user and
> > set their initial password, I typically get an error stating "BAD
> > PASSWORD: it is based on a dictionary word". My user's are also stating
> > that they are having problems changing their passwords.
> >
> > I ignore this message and the "bad"
> > password seems to be accepted OK.
==============536A6944C7E84B55DDCB8E53
Content-Type: text/x-vcard; charset=us-ascii;
name="vmiller.vcf"
Content-Transfer-Encoding: 7bit
Content-Description: Card for Rick Miller
Content-Disposition: attachment;
filename="vmiller.vcf"
begin:vcard
n:Miller;Vincent
x-mozilla-html:FALSE
org:The MITRE Corporation;R107 - End User Computing
adr:;;;;;;
version:2.1
email;internet:[EMAIL PROTECTED]
title:Help Desk Specialist
fn:Rick Miller
end:vcard
==============536A6944C7E84B55DDCB8E53==
------------------------------
From: [EMAIL PROTECTED] (bill davidsen)
Subject: Re: Linux smaller than Windows NT ?
Date: 24 Mar 1999 23:17:30 GMT
In article <[EMAIL PROTECTED]>,
=?iso-8859-1?Q?J=E9r=F4me?= Baldo <[EMAIL PROTECTED]> wrote:
| I recently transfered all file sharing services from a Windows NT
| machine (Cyrix 166+, 32 Mb RAM) onto a Linux server (a Pentium 200 MMX,
| with 96 Mb of RAM).
| And now, the users of those files complain about the speed of the new
| server.
| I am using samba 1.9.18 and a 2.0.36 kernel. Is there any way to make
| samba faster ? How can I change the priority for the samba daemon ?
| Will the speed increase if I put a new kernel and if I upgrade Samba ?
I have no idea about the kernel upgrade, I have been on more recent
kernels for a very long time, but the upgrade to 2.x SAMBA will do
wonders for speed and ease of admin.
--
bill davidsen <[EMAIL PROTECTED]> CTO, TMR Associates, Inc
What I find astonishing is not that my cat has started to sing, but that
he has taken up country-western. This morning he sang `Momma, don't let
your kittens grow up to be barn cats' in the shower, followed by a
pretty decent yodeling version of `Roundup time in Texas when the catnip
is in bloom.'
------------------------------
From: [EMAIL PROTECTED]
Subject: Remote printing to HP
Date: Wed, 24 Mar 1999 15:01:11 -0500
I am trying to send printouts from a PC running Linux
redhat 5.2 to a remote HP workstation running HP-UX 10.20.
The HP has a printer with the quename of lexxy.
The Linux hostname is: cyclone.
The HP hostname is: haze.
The printer quename is: lexxy.
The problem occurs when the HP workstation receives the
cf and df files from the PC. The file are copied to the
directory /usr/spool/lp/request/lexxy with file names of:
cfA029Aa20530
dfA029Aa20530
An error occurs on the HP and is reported in log file
with the message.
- -------------------------------------------------------
lexxy-29 doug lexxy Mar 23 15:08
/usr/sbin/lpsched: Unable to open and lock "request/lexxy/cfA029cyclone"
- -------------------------------------------------------
The data in cfA029Aa20530 file is:
- -------------------------------------------
Hcyclone
Pdoug
Ja.c
Ccyclone.sigmet.com
Ldoug
B
K1
O -oBSDJa.c -oBSDCcyclone.sigmet.com
FdfA029Aa20530
fdfA029Aa20530
UdfA029Aa20530
Na.c
- -------------------------------------------
It seems the HP is expecting the PC's hostname cyclone
to be the last 6 characters of the cf file,
instead of Aa20530.
Does anyone have any suggestions?
Thanks
Doug Paris
[EMAIL PROTECTED]
------------------------------
From: Mike <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.misc,comp.os.linux.setup
Subject: Re: Yet another pppd/modem problem.....
Date: Wed, 24 Mar 1999 15:31:45 -0800
Reply-To: [EMAIL PROTECTED]
Karl Soar wrote:
>
> Dear Gurus,
Heh. Just a lowly sysadmin responding...
> pppd started for ppp0 on /dev/cua2 at 9600
> pppd 2.2.3 started by root, uid 0
> tcgetaddr: Input/output error(5) ( <----- HMMMM....curious.....)
> Exit
I noticed below that your modem is on COM2, wouldn't this be /dev/cua1?
(COM1 = /dev/cua0, COM2 = /dev/cua1) Also, why 9600? Why not 19200?
> So then I tried playing with minicom, but strange things happen when it
> tries to initialise
> the modem...
>
> The Init string characters are echoed to the console v e r y s
> l o w l y indeed
> (maybe one every 10-20 secs) getting progressively slower until nothing
> appears to
> happening at all and I usually end up killing the process...
This is generally indicitive of an IRQ conflict. Do you have a built-in
COM port set to COM4? They tend to go to IRQ3 and conflict with COM2.
> Can anyone shed any light or point me in the direction of some
> enlightened soul
> before I tear any more hair out and admit defeat (2 weeks or
> investigation with little
> prior Linux experience)
> System Spec
> -----------------
> 8mb RAM, Quantum Pioneer 2.1GB IDE HDD,
> Adaptec 1542B on IRQ11 (base: 330),
> 'Old Faithful' Micropolis 135MB(!) SCSI HDD,
> Miro 10SD w/GENDAC (VLB),
> Mirosoft compatible serial mouse on COM1 (x03f8, IRQ4),
> USR Sportster 14.4 external modem on COM2 (x02f8, IRQ 3),
> SMC/WD8013 ethernet card (base 280, IRQ 5) (kernel module:eth0),
> Aztech Sound Galaxy NX Pro 16 (s/b compatible) (base: 220, IRQ7)
Hope this helps!
-Mike
------------------------------
From: "Georgios" <[EMAIL PROTECTED]>
Subject: Network unreachable or server down.
Date: Wed, 24 Mar 1999 20:52:06 +0100
I 've just set-up my RedHat 5.2 ppp but..I have an access problem to my ISP
After connecting to my ISP without any problem I get the "Network error
Unable to connect to server TCP error Network unreachable or the server is
down".
Does anybody has an idea??
------------------------------
From: "Dan" <[EMAIL PROTECTED]>
Subject: Re: ipchains, ipforwarding, kernel 2.2.3ac1 and ftp
Date: Wed, 24 Mar 1999 18:05:36 GMT
I'm sorry. The interface (-i) should be the one out to the internet. But
you should include the source address (-s) as your private network. So you
would have:
ipchains -A forward -i eth1 -s 192.168.1.0/24 -j MASQ
Dan wrote in message ...
>I think you want to forward the other way. The interface for forwarding
>should be the one for your private network.
>
>DR
>
>root wrote in message <7cpoe1$eui$[EMAIL PROTECTED]>...
>>
>>ipchains -A forward -i eth1 -j MASQ
>>
>>Eth1 being the card that is connected to the internet.
>>
>
>
>
------------------------------
** 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
******************************
