Linux-Networking Digest #223, Volume #12 Sat, 14 Aug 99 13:13:33 EDT
Contents:
Re: Odd Linux Modem Error (CRC) (Abdullah Ramazanoglu)
Re: DNS & IP-Masq. (Monte Phillips)
Re: Linux security update - newbie questions (hac)
Arg... All eggs in one basket! ("Stephanie R. Gesler")
Re: $15 card or $98 card? (Rod Smith)
Re: SCO 3.2 tape restore on linux (Paul Sherwin)
Re: Arg... All eggs in one basket! (Dave Edick)
Re: Automatically dial the internet from my firewall/router? (Bjoern Swierczek)
Re: name resolver problems (Seamus J Carr)
Re: ethernet problems and AMD79C974 (Mike Ching)
Samba newbie "where to start?" (Clemens Nijhof)
Re: Windows client can't Active FTP through firewall. (Mars)
ip masqing wowes... please help ("--[CAiN]--")
----------------------------------------------------------------------------
From: Abdullah Ramazanoglu <[EMAIL PROTECTED]>
Subject: Re: Odd Linux Modem Error (CRC)
Date: Sat, 14 Aug 1999 17:17:56 +0300
Bob Hart wrote:
>
> I've been racking my brain over this for a week
> now and would appreciate any insight.
>
> I've been able to successfully get every
> modem working in linux that I've tried, incluing
> ISA and PCI modems.
>
> Except for this one...
>
> I'm using RH 6.0.
>
> It's a PCI internal V.90 modem for a Transmonde
> Vibrant LS laptop. Transmonde's web site and
> technical support assures me (strongly) that the modem
> is not a winmodem. The modem itself is
> manufactured by Foresson. I use minicom
> and can dial out fine, connect at around 44kbps
> and can telnet adequately. Every so often
> the telnet session slows down dramatically, however.
>
> Further, when I try to use zmodem transfer in
> minicom, I get about 10k of the file, and then
> get error messages about "BAD CRC: Retry". A
> few seconds later, a few more kbyte are transferred,
> and the same error message. This happens whether
> I use zmodem or xmodem. Ultimately, I end up
> with transfer rates of about 0.5kbyte/second, or
> about 1/8 what they should be for the connection.
>
> I've used the same modem init string that windows
> uses for the modem. When the connection
> is made, I get V.42 BIS and LAPM connection messages,
> so it appears the hardware settings for the modem are
> ok. However, the CRC error would suggest it is
> not.
>
> I should point out that the ftp and zmodem
> transfer rates in windows 98 are fine (4kbyte/second
> on binary files).
>
> These problems also impact greatly Linux ppp transfers,
> which work fine otherwise. Linux ftp transfer
> rates through ppp typically average the same--
> about 0.5kbyte/second.
>
> setserial autodetects the port as
> uart 16550A spd_vhi
>
> I'm at a loss to explain the errors. Anyone have
> any ideas?
Could it be because of uart overrun / underrun that is not detected
(somehow) by linux?
May be xon/xoff characters get messed up with actual data?
May be crtscts is not working as it should for this particular modem?
Some modems use DTR instead of RTS. (see pppd man page, cdtrcts option)
Good luck
>
> Thanks,
>
> -Bob Hart
> [EMAIL PROTECTED]
--
Abdullah Ramazanoglu [ aramazanoglu AT demirbank DOT com DOT tr ]
------------------------------
From: [EMAIL PROTECTED] (Monte Phillips)
Subject: Re: DNS & IP-Masq.
Date: Sat, 14 Aug 1999 13:35:29 GMT
Most common problem I see on this group regarding masq/firewall, is
that they try to configure everything at once.
DO NOT DO THIS!
Once you are certain you have installed all the necessary progs, then
set up your masqing/firewall to work properly on ONE client or subnet,
then move to the second. If you have two up and working then you can
use them as the template too set up all the rest at a single go. If
you then have problems, you can bet your butt its a typo in one if the
entries.
If you are on a large net, then isolate a single unit to experiment on
until you get the 'bugs' worked out, then move it onto the main net.
Billy <[EMAIL PROTECTED]> wrote:
>I posted a question to this newgroup last week about a problem I had
>with my IP-masq clients trying to resolve hostnames forever when I had
>no net connection.
>It was suggested that I set up a caching DNS server to try and fix it,
>which I've now done. Alas, I'm getting the same problem still.
>Is there any way I can set named/bind (v8) to timeout more quickly if my
>PPP connection is unavailable?
>I had also thought on adding/removing a firewall rule for DNS when the
>connections is made and broken, but this seems a fairly inelegant way of
>doing it...
------------------------------
From: hac <[EMAIL PROTECTED]>
Crossposted-To:
athome.users-unix,comp.os.linux,comp.os.linux.misc,comp.os.linux.questions,comp.os.linux.redhat,comp.os.linux.setup
Subject: Re: Linux security update - newbie questions
Date: Sat, 14 Aug 1999 14:00:28 GMT
help wrote:
>
> From www.linux-mandrake.com :
>
> >SECURITY UPDATE: kernel 2.2.9
> >2.2.x kernel are all affected by a networking security bug. Please do not
> use your machine in a network >environment until you update with our new
> kernel.
> >Command line for upgrading:
>
> >rpm -ivh kernel-2.2.9-27mdk.i586.rpm
>
> >Important note: please don't forget to modify /etc/lilo.conf after
> upgrading, and re-run /sbin/lilo.
>
> Complete Newbie Questions:
> What do I modify in /etc/lilo.conf?
> Re-run /sbin/lilo = rebooting computer?
>
> Thanks
If your new kernel image has a different filename, than that filename
needs to be in /etc/lilo.conf.
Type "/sbin/lilo" as root, and lilo will rebuild the boot sector to
point to the physical location of your new kernel image. Then you can
reboot.
I advise making a boot floppy, just in case.
--
Howard Christeller Irvine, CA [EMAIL PROTECTED]
------------------------------
From: "Stephanie R. Gesler" <[EMAIL PROTECTED]>
Subject: Arg... All eggs in one basket!
Date: Sat, 14 Aug 1999 11:03:51 -0400
Wondering If anyone can help...
Here's my configuration:
I run a main DNS server on Redhat 5.2. On this same server I run apache as
a web server, sendmail and Pop services.
I want to move some of this stuff.
I have setup two other boxes. One is running authentication. Both of these
are RedHat 6.0.
I have setup another box as secondary dns and would like to use it for my
web server to get it off of my main server. I have about 15 Virtual Domains
on my Main DNS server. I can point one to another box, but the only way I
can figure to do it is assign the address of the linux box in my named file
for that domain. I have to assign all my virtual domains addresses on my DNS
server, how do I then point those addresses to the other linux box??? Do I
have to assign them to my ethernet interface on that box?? I can't when that
box isn't routing the domain...
Can anyone help?
------------------------------
Reply-To: [EMAIL PROTECTED]
From: [EMAIL PROTECTED] (Rod Smith)
Subject: Re: $15 card or $98 card?
Date: Sat, 14 Aug 1999 15:19:29 GMT
[Posted and mailed]
In article <[EMAIL PROTECTED]>,
Lindoze 2000 <[EMAIL PROTECTED]> writes:
> yes folks, I just went and got a $35USD tulip card by bay networks
> called net gear 310. I've had some no name and a d-link 530tx. the no
> name brand never worked and I could not get support for them. the d-link
> card worked fine, but was not detected during setup. the tulip card was
> detected at setup and is the first time I've ever installed linux over a
> network! I have an ftp server with RedHat, SuSE, and Slackware linux on
> it.
> I could install the SuSE over ftp, but I could not get the Redhat to
> install over ftp.
It's not clear precisely which OSes did and did not work with which cards.
The D-Link DFE-530TX uses a VIA Rhine chipset, but Red Hat apparently
forgot to include support for it in their network install floppy for RH
6.0. (Note there's also an older DE-530TX that used the Tulip chipset.)
There's an updated Red Hat 6.0 install CD that should work with that board
to install RH via FTP.
As to the Netgear board, if you bought it recently, it's probably got a
PNIC Tulip clone chipset. Reports vary, but most people have problems
getting these working with the stock drivers included with most
distributions. There are updated drivers available on
http://cesdis.gsfc.nasa.gov/linux/drivers/tulip-devel.html that work well
with it, though. Unfortunately, this won't do you much good for
installation unless you want to hack the install floppy to include the
updated driver.
--
Rod Smith
[EMAIL PROTECTED]
http://members.bellatlantic.net/~smithrod
Author of _Special Edition Using Corel WordPerfect 8 for Linux_, from Que
------------------------------
From: [EMAIL PROTECTED] (Paul Sherwin)
Crossposted-To: alt.os.linux,comp.security.firewalls,comp.os.linux.networking
Subject: Re: SCO 3.2 tape restore on linux
Date: Sat, 14 Aug 1999 15:36:14 GMT
On Fri, 13 Aug 1999 10:00:45 GMT, [EMAIL PROTECTED] (J Bade)
wrote:
>Hi,
>
>The Trick I want to transfere data backed up on a DAT tape using SCO
>3.2 (without any networking) and restore the data on Linux or Windows
>NT.
>Amount 350 MB of data.
>
>What is the best approach?
It depends what the tape was written with. Assuming it is tar, you
should be able to put the tape in a DAT drive on your Linux system and
read it with tar. There are often strange incompatabilities between
one DAT drive and another, though - these are nothing to do with
Linux. If the SCO system is running on PC hardware you can probably
take out the DAT drive and SCSI card, transfer it to the Linux system,
load the appropriate SCSI card kernel module, and read the tape. You
can't mount the SCO disks on your Linux system.
>
>Can NT be made to read tapes backed-up with SCO 3.2 and TAR?
>
No, not without commercial third party software.
Best regards and good luck, Paul
Paul Sherwin Consulting 22 Monmouth Road, Oxford OX1 4TD, UK
Phone +44 (0)1865 721438 http://www.telinco.co.uk/psherwin/index.htm
Mobile +44 (0)7931 578334 mailto:[EMAIL PROTECTED]
Pager +44 (0)7666 797228
------------------------------
From: [EMAIL PROTECTED] (Dave Edick)
Subject: Re: Arg... All eggs in one basket!
Reply-To: [EMAIL PROTECTED]
Date: Sat, 14 Aug 1999 15:39:26 GMT
Simple. The SOA records should point to the DNS server address. The entries
in the A and/or CNAME records should point to the web server address.
On Sat, 14 Aug 1999 11:03:51 -0400, Stephanie R. Gesler <[EMAIL PROTECTED]> wrote:
>Wondering If anyone can help...
>
>Here's my configuration:
>
>I run a main DNS server on Redhat 5.2. On this same server I run apache as
>a web server, sendmail and Pop services.
>
>I want to move some of this stuff.
>
>I have setup two other boxes. One is running authentication. Both of these
>are RedHat 6.0.
>
>I have setup another box as secondary dns and would like to use it for my
>web server to get it off of my main server. I have about 15 Virtual Domains
>on my Main DNS server. I can point one to another box, but the only way I
>can figure to do it is assign the address of the linux box in my named file
>for that domain. I have to assign all my virtual domains addresses on my DNS
>server, how do I then point those addresses to the other linux box??? Do I
>have to assign them to my ethernet interface on that box?? I can't when that
>box isn't routing the domain...
>
>Can anyone help?
>
>
------------------------------
From: [EMAIL PROTECTED] (Bjoern Swierczek)
Crossposted-To: linux.redhat.misc
Subject: Re: Automatically dial the internet from my firewall/router?
Reply-To: [EMAIL PROTECTED]
Date: Sat, 14 Aug 1999 17:28:33 +0200
On Sat, 14 Aug 1999 13:28:28 GMT, Jim Orfanakos <[EMAIL PROTECTED]> wrote:
> How can I set it up so if a computer on the LAN needs to connect to the
> internet, the firewall/router will automatically dial out...and disconnect
> after a certain tine of inactivity?
You want to use the dial-daemon ('diald').
Bjoern.
--
LinuxHaus Stuttgart fon +49 (711) 2851905 | Bj�rn Swierczek
Birkenweg 20 fax +49 (711) 5780692 | fon +49 (171) 6036857
D-70734 Fellbach <[EMAIL PROTECTED]> | <[EMAIL PROTECTED]>
------------------------------
From: [EMAIL PROTECTED] (Seamus J Carr)
Subject: Re: name resolver problems
Date: Sat, 14 Aug 1999 15:33:23 GMT
Reply-To: [EMAIL PROTECTED]
I've done that. The boot script seems intent on rewriting it. Part
of the boot init copies the resolv.conf to resolv.conf.save and the
resets the nameserver to 0.0.0.0. Can I edit the init script and take
this out without having to spend hours tracking the consequences?
QuestionExchange <[EMAIL PROTECTED]> wrote
>> I'm setting up a machine with dual boot capabilities.
>> (NT/Caldera 2.2) I'm tying into a hybrid WinNT/SCOUnix network
>> where the DHCP is handled by the NT side. I'm not getting a
>> clean address when I logon the Linux and I have to reenter the
>> name resolver address for the DNS every time I come back to
>> Linux. I have already updated the hosts table.(have to do it
>> whenever I lease an address). Can someone point out what I;m
>> missing? Sent via Deja.com http://www.deja.com/ Share what
>> you know. Learn what you don't.
>
>Hi, List your nameserver in /etc/resolv.conf and save it!
>"nameserver ip.address" is ok. HTH!
>
>--
> This answer is courtesy of QuestionExchange.com
>
>http://www.questionexchange.com/servlet1/showUsenetGuest?ans_id=2463&cus_id=USENET&qtn_id=1581
------------------------------
From: [EMAIL PROTECTED] (Mike Ching)
Crossposted-To: comp.os.linux.setup
Subject: Re: ethernet problems and AMD79C974
Reply-To: [EMAIL PROTECTED]
Date: Sat, 14 Aug 1999 15:52:40 GMT
I was only able to support my PCnet board by compiling the driver into the
kernel when using version 2.0.30 but was able to use a module when I got
version 2.2.5. I recommend getting the current kernel.
On Sat, 14 Aug 1999 21:55:39 +0800, Floyd <[EMAIL PROTECTED]> wrote:
>Hi,
>I meet such problem before. My Pc is Compaq Deskpro 5120. The problem
>resolved only after I recompiled the kernel. My kernel is 2.0.36. You can
>try to change the status from DHCP to Bootp. Well, I recompiled the kernel
>for 6+ times!! It always prompts me the kernel is too large(>512K).
>Hope will helps.
>
>floyd.
>
>Jonathan Bowman wrote in message <[EMAIL PROTECTED]>...
>> When activating my eth0 interface, the system responds, "eth0:
>>unknown interface." Using netcfg, I have added the eth0 interface,
>>specifying dhcp, and activate interface at boot time, but nothing else
>>(is this enough, since I am using dhcp?) I have also given netcfg the
>>nameserver addresses, and in "Routing" I have given the default gateway
>>address, and specified the device as "eth0." Is any of this wrong or
>>not enough?
>> The only problem I can see is perhaps with my ethernet card itself.
>>I have a Compaq Deskpro XL 575 with onboard ethernet (AMD79C974), which
>>uses the pcnet32 driver. When I select the driver to use (using
>>linuxconf), I do not see pcnet32 -- can I add this driver (which is in
>>my /lib/modules/net directory) without recompiling the kernel, or do I
>>not need this driver (I have tried ne2k-pci)? Any suggestions would be
>>helpful.
>>
>>e-mail is preferable,
>>Jonathan Bowman
>>
>
>
------------------------------
From: Clemens Nijhof <[EMAIL PROTECTED]>
Subject: Samba newbie "where to start?"
Date: Sat, 14 Aug 1999 11:02:39 -0300
Reply-To: [EMAIL PROTECTED]
Hello everyone,
I ordered a book about samba from H W Sams but, while this book is in
the mail I was wondering where to start? From our local usegroup I got
an e-mail to install sendmail,Bind, fetchmail, Samba and dhcpd.
Now what to do next? running RH v 6.0
-Clem
------------------------------
From: Mars <[EMAIL PROTECTED]>
Subject: Re: Windows client can't Active FTP through firewall.
Date: Sat, 14 Aug 1999 23:03:10 +0800
I have the same problem. After trying your suggestion, it still fails.
Is the order of loading module ip_masq_ftp and ipchains/ipmasqadm
matter?
I try "modprobe -c" and couldn't see ip_masq_ftp is loaded. How could I
know it is properly loaded?
Mars
Paskal van Lomm wrote:
>
> Did you try: modprobe ip_masq_ftp ??
>
> I've done this and that solved my similar problem.
>
> Regards,
> Paskal
------------------------------
From: "--[CAiN]--" <[EMAIL PROTECTED]>
Subject: ip masqing wowes... please help
Date: Sat, 14 Aug 1999 13:08:06 -0400
I'm running redhat 6.0 and i'm trying to get ip masquerading to work. I
have my linux box connected to the internet by means of a cable modem
and i have various windows machines trying to connect. In the linux
box: eth0 and eth1 are working fine, all machines can ping each other,
and the linux box can reach the internet, but no one else can. Here's a
copy of my rc.firewall file:
/sbin/depmod -a
/sbin/modprobe ip_masq_ftp
/sbin/modprobe ip_masq_raudio
/sbin/modprobe ip_masq_irc
/sbin/modprobe ip_masq_quake ports=26000,27000,27910,27960
/sbin/modprobe ip_masq_cuseeme
/sbin/modprobe ip_masq_vdolive
echo "1" > /proc/sys/net/ipv4/ip_forward
echo "1" > /proc/sys/net/ipv4/ip_dynaddr
/sbin/ipchains -M -S 7200 10 60
/sbin/ipchains -A input -j ACCEPT -i eth1 -s 0/0 67 -d 0/0 68 -p udp
/sbin/ipchains -P forward DENY
#TO ENABLE ALL MACHINES TO INTERNET
/sbin/ipchains -A forward -s 192.168.0.0/24 -j MASQ
#TO ENABLE ONLY SPECIFIC MACHINES TO INTERNET
#ipchains -A forward -s 192.168.0.10/32 -j MASQ
this stuff is all taken from mini HOWTO on ip masquerading. Although,
no computer but the linux box itself can reach the internet. I did a
'chmod 700 rc.firewall' and put /etc/rc.d/rc.firewall in the rc.local
file. But when i do a ./rc.firewall to check if it is executing ok i
get this output:
/sbin/depmod: invalid option --
Usage: depmod [-e -s -v] -a [FORCED_KERNEL_VER]
depmod [-e -s -v] MODULE_1.o MODULE_2.o ...
Create module-dependancy information for modprobe.
-a, --all visit all modules
-d, --debug run in debug mode
-e output unresolved symbols
-i ignore symbol versions
-m, --system-map <file> use the symbols in <file>
-s, --system-log use the system log for error
reporting
--help display this help and exit
-v, --verbose run in verbose mode
-V, --version output version information and
exit
can't locate module ip_masq_ftp
can't locate module ip_masq_raudio
can't locate module ip_masq_irc
/lib/modules/2.2.5-15/ipv4/ip_masq_quake.0: invalid parameter ports
can't locate module ip_masq_cuseeme
can't locate module ip_masq_vdolive
: No such file or directoryt/ipv4/ip_forward
: No such file or directoryt/ipv4/ip_dynaddr
' specifiedins: invalid timeout value '60
Try '/sbin/ipchains -h' or '/sbin/ipchains --help' for more information.
' specifiedins: unknown protocol'udp
Try '/sbin/ipchains -h' or '/sbin/ipchains --help' for more information.
' for -Pchains :Invalid policy 'DENY
Try '/sbin/ipchains -h' or '/sbin/ipchains --help' for more information.
'sbin/ipchains: Invalid targe name 'MASQ
Try '/sbin/ipchains -h' or '/sbin/ipchains --help' for more information.
: command not found
I know that i have everything installed correctly, i even recompiled the
kernel and made sure i had all of the necesarry things for ip
masquerading to work. With some of the commands, if typed in manually
in the root, will work, but not all. A 'depmod -a' in the root gives me:
/lib/modules/2.2.5-15/net/plip.o: unresolved symbol(s)
/lib/modules/2.2.5-15/ipv4/ip_masq_user.o: unresolved symbol(s)
/lib/modules/2.2.5-15/misc/nvram.o: unresolved symbol(s)
but things like '/sbin/modprobe ip_masq_ftp' will execute fine. ???
i'm very confused, any help is extremely appreciated.
------------------------------
** 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
******************************
