Linux-Networking Digest #710, Volume #9 Wed, 30 Dec 98 00:13:47 EST
Contents:
Re: Please Help..... (Joel Sloan)
Re: fetchmail/sendmail problem (Cliff Rayman)
Re: Problem setting up homenetwork with 3c509b. ([EMAIL PROTECTED])
Re: Virtual domain question. ("Dr. S. Robert Senay")
Re: Successful connection to Earthlink? (tyler)
Re: firewall tool for linux? (Holger Marzen)
Re: HEEEEEEEEEEEEEEEEELLLLLLLLLLLLLLLLP !!!!! Ethernet Card Trouble (Glenn Valenta)
Re: Help with automating fetchmail (Cliff Rayman)
RedHat 5.1 PPP/Shell account server problems (WB)
Re: PPP Options Setting for PAP Dialin Server? (Josh Gentry)
Re: dynamic dns system for free similiar to ml.org ("Dr. S. Robert Senay")
IPMasq works great...now SQUID - Just wanna cache on Linux (Kool Breeze)
Re: NIS and nis.domain (Mike Ireton)
Re: Insmod doesn't like Module NE2000. (Mike Ireton)
----------------------------------------------------------------------------
From: Joel Sloan <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.questions,comp.os.linux.setup,comp.os.linux.advocacy
Subject: Re: Please Help.....
Date: Tue, 29 Dec 1998 22:33:13 GMT
Glenn Valenta wrote:
> I'm not sure about RH5.2, but you may need to recompile the kernal with
> ipmasquerading and firewalling installed...
No need, Red Hat 5.x comes that way "out of the box".
js
------------------------------
From: Cliff Rayman <[EMAIL PROTECTED]>
Subject: Re: fetchmail/sendmail problem
Date: Tue, 29 Dec 1998 14:38:12 -0800
not sure what version of linux you are using, but I have never heard of
/bin/mail.local. My Redhat 5.2 version of Linux uses /usr/bin/procmail which
is pretty standard. In the file /etc/sendmail.cf should be a line something
like:
Mprocmail, P=/usr/bin/procmail, F=DFMSPhnu9, S=11/31, R=21/31,
T=DNS/RFC822
/X-Unix, A=procmail -Y -m $h $f $u
this is the program and necessary parameters to drop mail for local users. If
you have the /usr/bin/procmail program on your system, and a line like above
in your sendmail.cf then you have a sendmail configuration problem which is
causing the mail addressing to not be properly resolved.
Darryl Tang wrote:
> I am trying to configure fetchmail to download my mail from my isp via
> modem and put it into my mailbox. As I understand it, after fetchmail
> downloads the mail, it invokes sendmail to send it to my local mailbox.
> I'm at the point where fetchmail can pull the mail, but sendmail is having
> trouble doing its job.
>
> I get the following error in my /var/log/syslog:
>
> SYSERR(bin): Cannot exec /bin/mail.local: No such file or directory
>
> Am I missing this file or is sendmail looking for the wrong file? BTW,
> while testing this out, fetchmail downloaded two messages, the system
> error above was written twice to the syslog, and the mail appears to have
> been lost.
>
> Thanks for any help.
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: Problem setting up homenetwork with 3c509b.
Date: Tue, 29 Dec 1998 22:03:54 +0100
>> The computers are unable to PING each other.
I had a similar problem using the 3c509(B). I had everything configured,
but pinging didn't work. The problem was that the 3c509 has different
transceivers (AUI, 10base2, 10baseT). By default, the 10baseT transceiver
was enabled; you'll probably need the 10base2 (coax) type. I had to use
the program `3c509x.exe' (available from ftp.3com.com) to firstly disable
PnP and secondly reset the default.
You might also like to have a look at:
http://cesdis.gsfc.nasa.gov/linux/drivers/3c509.html
Good luck, Robert
--
Robert Vollmert [EMAIL PROTECTED]
------------------------------
From: "Dr. S. Robert Senay" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.misc
Subject: Re: Virtual domain question.
Date: Tue, 29 Dec 1998 17:52:50 -0500
mcv wrote:
> Hi,
> If I set-up my Linux PC with the virtual domain blahblah.blah, and set-up
> the DNS service on my PC, can other internet users connect to blahblah.blah
> or do I need to register this domain so it can be accessed?
> Cheers,
> Mark.
Okay folks here is a bit of wisdom that seems to be slow making it to market,
but so incredibly useful I'm amazed it isn't spreading like a forest fire in
dry woods... DNS can be a tricky beast yes, but impossible no... The first
smartest thing you can do is get your LOCAL domain working correctly and make
certain that your connection to the net if not static is at least reliable and
dynamic WITHOUT intervention... I.E. if the pipe goes down for any reason it
will come back up on its own... That said and done, there is really no need
to give out your IP address to have other people reach your lan, and you
should have some level of protection between you and the outside... So,
routing internally will dealt with by DNS, next you need an external
connection that SHOULD be on the otherside of some form of firewall,
potentially this could be as simple as two DIFFERENT adapters in you routing
machine, one for external connections on one IP and one internall with a
DIFFERENT IP. In this fashion you can provide yourself with some protection
see the Firewall HOWTO... Then once you have a connection tested and working
with an IP that you can see on the internet, stop in at www.tzo.com and
sign-up for service there... A virtual domain (yourname.tzo.com) is available
for something inexpensive (around 20 dollars US when I signed up) on an annual
basis, what they offer is not just an address but DNS server service on the
net that keeps things updated for you no matter how often your connection goes
up and down... You run a small client on one of your machines that reports
your NET IP address anytime your connection goes up or down and keeps the
tables fresh... In this fashion your server is ALWAYS available... Let the
folks at TZO know a happy customer sent you... Once you have this nailed, it
doesn't much matter what kind of connection you are using, or who your service
provider is... It's the ULTIMATE shoestring SOHO...
Scott...
------------------------------
From: tyler <[EMAIL PROTECTED]>
Subject: Re: Successful connection to Earthlink?
Date: Tue, 29 Dec 1998 14:48:38 -0800
J Scott Berg wrote:
> In article <[EMAIL PROTECTED]>, Tyler Dinh <[EMAIL PROTECTED]> wrote:
>
> > Does anyone have a successful connection to Earthlink? I was able to
> >connect, and received an IP, but immediately Earthlink dropped my
> >connection. Any suggestion?
>
> I connect to Earthlink, and while there are a lot of hassles to get
> things right, most of the problems I've had are fairly subtle.
>
> You're sure you received the IP? Once you get to that point, you
> should at least be able to do something. Does /sbin/ifconfig list
> ppp0 at any point? Does /var/log/messages (or maybe some other
> syslogd-created file) actually state that you have an IP address (it
> should)?
Yes, it did. In fact, after my password was accepted, it said:
Connected ...
IP:xxx.xxx.xxx.xxx
mtu:1500
>
> Do you have an ip-up script (or anything other than config files in
> /etc/ppp)? If so, get rid of it for now.
No, I have deleted all those scripts, except for : ppp-on, ppp-off, options,
chap- and pap-secrets, and ppp-on-dialer
> Make sure that you start up ppp with as little as possible running.
> The idea is to make sure that nothing is trying to send anything down
> your connection.
>
> What version of pppd are you running? Does it match what is compiled
> into your kernel?
>
I installed RH5., the kernel is 2.0.35 and pppd is 2.3.5
> Post your /etc/ppp/options file.
>
I used the options file posted here few days earlier. Basically it includes
the following:
crtstcs
defaultroute
modem
lock
mtu 1024
mru 1024
Nonetheless, I have tried using linuxconf, but I still got disconnected.
According to the /var/log/messages
NO CARRIER was received after Earthlink gave me an IP.
Thanks for your help.
Tyler
>
------------------------------
From: [EMAIL PROTECTED] (Holger Marzen)
Crossposted-To: comp.security.firewalls
Subject: Re: firewall tool for linux?
Date: 29 Dec 1998 08:48:27 GMT
(Please note that there is an idiot who posts his crap with my
email-address. He uses other person's email addresses as well. If he
didn't fake his postings perfectly then his real email-address might be
[EMAIL PROTECTED] I sent a complaint to [EMAIL PROTECTED])
In article <768kpo$o11$[EMAIL PROTECTED]>,
Paul B. Brown <[EMAIL PROTECTED]> wrote:
>>I use it and I like it. It can create dynamic rules, send alert mails
>>and perform actions.
>
>My only question is stability. Has SINUS been around long enough and
>have a large enough installed user base to really be thoroughly tested?
>When dealing with firewalls, I prefer to be conservative. The older
>ipfwadm certainly has. IP Chains is coming up to speed as well.
There are some people who report stability problems. But I have none. As
far as I noticed only RedHat based systems have such problems. Maybe
they have some distributor-specific kernel modifications.
I use a libc5-based Linux-2.0.35 and 2.0.36 and have no problems.
Performance and stability are excellent. SINUS firewall can check tcp
sequence numbers. I am not sure if ipchains can do that.
------------------------------
From: Glenn Valenta <[EMAIL PROTECTED]>
Reply-To: [EMAIL PROTECTED]
Crossposted-To:
linux.redhat.install,comp.os.linux.questions,comp.os.linux.hardware,comp.os.linux.setup
Subject: Re: HEEEEEEEEEEEEEEEEELLLLLLLLLLLLLLLLP !!!!! Ethernet Card Trouble
Date: Tue, 29 Dec 1998 22:58:09 GMT
T1 wrote:
>
> So how do I get it working? The RH manual, which I borrowed, won't
> tell me that. All it says is that if your ethernet card is PCI, you
> will have to change your PCI bios to get it working correctly
>
> On Tue, 29 Dec 1998 06:38:09 -0700, mk <[EMAIL PROTECTED]>
> wrote:
>
> >T1 wrote:
> >>
> >> I am trying to install Redhat Linux 5.2 onto my system via FTP since I
> >> am sick and tired of buying Linux by the time a new version always
> >> comes out. Besides, I don't have anymore hard drive space although I
> >> formated a hard drive for RH linux. I am using an SMC PCI 10 Ethernet
> >> card and I have a direct connection via Cox@home. I downloaded both
> >> the supp and boot images and placed them to two floppies, booted via
> >> my floppy, chose FTP as installation option, inserted supp disk, and
> >> then I had no problem choosing NE2000 PCI. However, when I got to the
> >> autoprobing option, I got that annoying and irritating error message
> >> saying "SIOCSIFFLAGS: Resource Temporarily Unavailable." Anybody got a
> >> solution to this error message?
> >
> >
> >
> >The SMC PCI card that @home distributes is I believe a DEC chip. It uses
> >the tulip.o driver.
> >
Re-compile the kernel with the tulip driver only and not as a module. After
installing the kernel and log back in, type "dmesg | less" and scroll around
and confirm that Linux sees the network card. The line should look something
like this:
tulip.c:v0.89H 5/23/98 [EMAIL PROTECTED]
eth0: Lite-On 82c168 PNIC at 0xdd00, 00 a0 cc 30 bf b0, IRQ 5.
--
x-no-archive: yes
--
Glenn Valenta Engineering @ http://www.coloradostudios.com
[EMAIL PROTECTED] http://ouray.cudenver.edu/~gavalent/
[EMAIL PROTECTED] Personal mail
[EMAIL PROTECTED] Work mail
------------------------------
From: Cliff Rayman <[EMAIL PROTECTED]>
Subject: Re: Help with automating fetchmail
Date: Tue, 29 Dec 1998 15:05:24 -0800
fetchmail will grab the mail off of the target system, then run it through your
local sendmail program. The mail will go where all of the regular user mail
goes now, so people will find all of their mail in their mailboxes. Only the
To: address and the mail headers will indicate that this mail was popped from
another server.
Jonas wrote:
> Thanks! That just what i needed.
> This must have been the fastes answers ever. I had the answer 'bout ten
> second after I posted the question, incredible.
>
> If I have understood fetchmail correct (not likely)it by default it
> downloads mail to the current ~/mail (home directory) but how can I make the
> mail get in different directories for different users. I have at least one
> account at the internet POP-server for every user on the LAN so the easiest
> way would be if I could specify where it should store the mail for each
> fetchmail connection, or do I have to use procmail for that?
>
> TIA /Jonas
------------------------------
From: WB <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.help,comp.os.linux.setup
Subject: RedHat 5.1 PPP/Shell account server problems
Date: Wed, 30 Dec 1998 04:14:23 GMT
I hit a wall once every few months with some topic on Linux. After I
feel I've suffered
enough on my own I try to make an educated post. So far I'm 100% on
getting the
solutions I've needed. I think I may just be missing something stupid
here, please let
me know if you have any insights on this.
I am trying to set up Machine 1 to be a dialup PPP server. I -have- read
all the
HOWTOs (serial, PPP, etc). I have also installed and read the docs on
mgetty.
I -am- able to receive faxes properly via mgetty+sendfax! Thus I assume
my problem
MUST be configuration related that I can't get a PPP connection going.
I have two questions I need help on:
1. HOW does one set up just a generic "shell" dial-in account. The PPP
HOWTO
says "this is NOT included in this document, read the SERIAL HOWTO for
that
information". Unfortunately, that info is -not- in the SERIAL HOWTO that
I could see.
The main reason I'd like to set this up is to get SOME type of access.
Is it the default
for mgetty if I don't set autoppp or what? Is my login.config defined
correctly for PPP?
2. What do I need to change on the server from what I have now to get
this PPP
connection working? /var/log/messages on the client shows a connection,
but I never
get the "login" prompt, just garbage. I don't care if it's PAP, CHAP, or
nothing. I
just want to get the connection going! In the past I've only gotten
garbage AFTER
the login/pw connection (ie during the PPP negotiation). But here you
can see that
I'm getting garbage right after the initial connection.
(Excerpt from client log, full text below)
Dec 29 21:11:24 storm chat[331]: 300 NoEC^M
Dec 29 21:11:27 storm chat[331]:
~?~?~?~?~?~?~?~?~?~?^_^@aX,^V^KaX^P^N^G^Rh ^P^H^D^M
If you have some ideas, PLEASE reply and cc an email to me at
[EMAIL PROTECTED]
I will (obviously) read reply posts, but only have a news server
available during specific hours,
so a cc email would be appreciated! Specs are included below...
(Remove the spam-blocking character Z in my email address).
Thanks! -- Bill
Machine 1: P2 with (unknown?) ISA modem (came from inside Acer Acros
P100)
- Redhat 5.1 is installed, TCP/IP networking configured and working, and
PPP set up
for module load. Remember, mgetty+sendfax -does- receive faxes
properly..!
- PPP 2.2.0 (modules loaded by rc.local)
- IP forwarding is turned on
- mgetty+sendfax manually installed (fax receiving works fine)
# login.config ----- (played around with various combinations)
/AutoPPP/ - /usr/sbin/pppd login -pap -chap modem crtscts
# NOTE I'VE TRIED USING THE FILE OPTION TO POINT
# TO /ETC/PPP/OPTIONS.SRV, BUT DOESNT MATTER
# -----------------
# mgetty configuration file (played around with various combinations)
debug 5
fax-id 49 89 xxxxxxxx
port ttyS1
debug 5
speed 19200
switchbd 19200
#---------------------------
- /etc/ppp/options.srv file, which I've played around with using (and
not)
asyncmap 0
:192.168.1.5
debug
modem
crtscts
dns-addr a.b.c.d
dns-addr a1.b1.c1.d1
proxyarp
lock
================
- /var/log/log_mgttyS1 output
12/29 22:11:23 ##### failed A_FAIL dev=ttyS1, pid=554, caller='none',
conn='', name=''
=============================
- /var/log/messages output ----------------------
Dec 29 19:38:04 tornado kernel: CSLIP: code copyright 1989 Regents of
the University of California
Dec 29 19:38:04 tornado kernel: PPP: version 2.2.0 (dynamic channel
allocation)
Dec 29 19:38:04 tornado kernel: PPP Dynamic channel allocation code
copyright 1995 Caldera, Inc.
Dec 29 19:38:04 tornado kernel: PPP line discipline registered.
Dec 29 22:11:23 tornado mgetty[554]: failed A_FAIL dev=ttyS1, pid=554,
caller='none', conn='', name=''
=================================================
Machine 2: IBM ThinkPad laptop that is able to dial up and connect to
other machines.
(Has both RedHat 5.0 and Win95 loaded on it)
- PPPD 2.3.3
- Script used to connect to Machine 1--------
echo Dialing...
/usr/sbin/pppd connect '/usr/sbin/chat -v "" ATM1DT1234567 CONNECT ""
login: <USER> word: <PW>' \
/dev/ttyS1 -chap -pap debug noipdefault modem defaultroute
==============================================
- /var/log/messages on the client --------------
Dec 29 21:11:10 storm kernel: CSLIP: code copyright 1989 Regents of the
University of California
Dec 29 21:11:10 storm kernel: PPP: version 2.2.0 (dynamic channel
allocation)
Dec 29 21:11:10 storm kernel: PPP Dynamic channel allocation code
copyright 1995 Caldera, Inc.
Dec 29 21:11:10 storm kernel: PPP line discipline registered.
Dec 29 21:11:10 storm kernel: registered device ppp0
Dec 29 21:11:10 storm pppd[330]: pppd 2.3.3 started by root, uid 0
Dec 29 21:11:11 storm chat[331]: send (ATM1DT7629497^M)
Dec 29 21:11:11 storm chat[331]: expect (CONNECT)
Dec 29 21:11:24 storm chat[331]: ATM1DT1234567^M^M
Dec 29 21:11:24 storm chat[331]: CONNECT
Dec 29 21:11:24 storm chat[331]: -- got it
Dec 29 21:11:24 storm chat[331]: send (^M)
Dec 29 21:11:24 storm chat[331]: expect (login:)
Dec 29 21:11:24 storm chat[331]: 300 NoEC^M
Dec 29 21:11:27 storm chat[331]:
~?~?~?~?~?~?~?~?~?~?^_^@aX,^V^KaX^P^N^G^Rh ^P^H^D^M
Dec 29 21:11:27 storm chat[331]: NO CARRIER^M
Dec 29 21:12:09 storm chat[331]: alarm
Dec 29 21:12:09 storm chat[331]: Failed
Dec 29 21:12:09 storm pppd[330]: Connect script failed
Dec 29 21:12:10 storm pppd[330]: Exit.
==========================================================
------------------------------
From: [EMAIL PROTECTED] (Josh Gentry)
Subject: Re: PPP Options Setting for PAP Dialin Server?
Date: Tue, 29 Dec 1998 15:05:19 -0800
Charles,
Since you have enabled AutoPPP the alias is unneccessary. Also, since you
did create an account for
pppuser that started pppd as soon as the user successfully logged in, then
the alias is also
unneccessary.
The point of the alias was for if a user with a shell account connected (our
original setup) without
AutoPPP (which we never enabled). The machines would negotiate a SLIP
connection by default,
allowing login from a terminal screen. This means that pppd still had to be
started on both machines.
The alias allowed the user to start pppd on the host machine.
This is a functional but obviously clumsy method which we improved on by
creating the pppuser account,
and you improved on farther in your setup.
In short, no, you don't need the alias, but there originally was a reason
for it.
Josh
[EMAIL PROTECTED]
*** Posted from RemarQ - http://www.remarq.com - Discussions Start Here (tm) ***
------------------------------
From: "Dr. S. Robert Senay" <[EMAIL PROTECTED]>
Subject: Re: dynamic dns system for free similiar to ml.org
Date: Tue, 29 Dec 1998 18:07:55 -0500
ben wrote:
> hi folks, i've been running my linux box and the dyns system
> from ml.org so the machine can have a domain name. well,
> unfortunately, ml.org went down and so i am looking for another
> simliar service. they provided a good service while it lasted..
> anyone have suggestions for alternatives?
Someone else here mentioned TZO and I thought I'd drop in my two cents
worth... I've been using TZO now for about six months and I have to say
if you go for nothing more than the twenty dollars a year subdomain
(yourdomain.tzo.com) its one hell of a bargain that CAN NOT be beat...
Set up a COMPLETE server with Mail, FTP, WEB and whatever else you want
and its good to go... I started with just the subdomain, but quickly
realized the POWER and functionality I could have at my finger tips and
now have what MUST BE the ultimate SOHO with a subdomain and a toplevel
domain all to myself with a dozen machines inside my firewall and BOTH a
COMPLETE NT4 Server AND a COMPLETE Redhat Linux 5.2 Server... I have a
router and firewall in between me and the net, and am running WEB, FTP,
MAIL and Database servers on BOTH boxes with fulltime connectivity and
access from the net... It's kinda COOL, best part of it is its all
based around a dynamic connection to the net with several v90 modems
ganged together... When they finally be faster pipes in here, I just
change the cable and everything else remains un touched since the second
the connection goes back up, the machine calls the TZO DNS and I'm live
again... No fuss, muss or bother...
------------------------------
From: [EMAIL PROTECTED] (Kool Breeze)
Subject: IPMasq works great...now SQUID - Just wanna cache on Linux
Date: Wed, 30 Dec 1998 04:21:43 GMT
Reply-To: [EMAIL PROTECTED]
Now that Ive been running IPMasq for about a year (via isdn) I wish to
cache web stuff on the linux box.
Is squid feasible since this is all I want to do?
Am I going to get better performance with only two computers that
actually do the browsing?
------------------------------
From: Mike Ireton <[EMAIL PROTECTED]>
Subject: Re: NIS and nis.domain
Date: Tue, 29 Dec 1998 23:08:38 +0000
Reply-To: [EMAIL PROTECTED]
elegall wrote:
> I try to install a client/server Nis on a local Network but I have few
> problems.
>
> At boot time(on server NIS) I have this message: domain not bound
>
It sounds to me as if you are not running ypbind. The programs you have
to be runing to make this fly are:
SERVER SIDE -
portmap
ypserv
ypbind
CLIENT SIDE -
portmap
ypbind
Check with both machines that these are in fact running (ps aux | grep
"portmap\|ypserv\|ypbind"), I bet you that one or more of them is missing.
good luck,
Mike-
------------------------------
From: Mike Ireton <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.hardware,comp.os.linux.setup
Subject: Re: Insmod doesn't like Module NE2000.
Date: Tue, 29 Dec 1998 23:11:15 +0000
Reply-To: [EMAIL PROTECTED]
>
Hey buddy,
You need to insmod 8390.o as well - This module, 8390, is the actual hardware
interface driver for ne2000's. Those 'unresolved symbols' are supplied by it and
so if you load this first and then ne, your problem will go away.
Mike
>
>
> > I have been trying to get my ISA D-Link DE-220PCT working under SuSE 5.3. I
> > have already used the utility on the installation disk to turn off PnP to
> > avoid problems there. There are no conflicts with other cards that I know
> > of... But when I do an insmod ne, I get six lines of errors after it:
> > /lib/modules/2.0.35/net/ne.o: unresolved symbol ei_open
> > /lib/modules/2.0.35/net/ne.o: unresolved symbol ethdev_init
> > /lib/modules/2.0.35/net/ne.o: unresolved symbol ei_debug
> > /lib/modules/2.0.35/net/ne.o: unresolved symbol ei_interrupt
> > /lib/modules/2.0.35/net/ne.o: unresolved symbol NS8390_init
> > /lib/modules/2.0.35/net/ne.o: unresolved symbol ei_closed
> >
> > This thing is really starting to tick me off. I have spent the last two
> > days working on this thing, to no avail. I have tried different settings,
> > recompiling, and removing other cards. Any and all help would be
> > appreciated.
> >
> > RC
>
> --
> ________________________________________________
> Sean Harper [EMAIL PROTECTED]
> Technician/Programmer Computer Gallery
> www.cgallery.com Redhat Linux Reseller
> ________________________________________________
--
Mike Ireton
Network Systems Manager
Broadlink Communications
------------------------------
** 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
******************************
