Linux-Networking Digest #850, Volume #9 Mon, 11 Jan 99 16:13:47 EST
Contents:
How to set up a DNS ?? (Ashutosh Sharma)
Re: How to setup routing? ("Jeff Volckaert")
Re: Proxy + firewall ("David H.")
Programming packet burst and long file names ([EMAIL PROTECTED])
Programming packet burst and long file names ([EMAIL PROTECTED])
Re: Newbie modem problem ("David H.")
Re: pppd with Rockwell HCF PCI Internal modem ("David H.")
Re: su won't let me shutdown/ifconfig (Sean McEwan)
what to filter in a packet filter ("Jan Stifter")
Re: ADSL (Chris)
Re: Q: DHCP & static IP coexist on one card possible? (Chris)
Re: Multiple NIC Cards.. HELP!! <solution> ("jay")
DMZ setup (Erwann CORVELLEC)
Re: SIOCADDRT: Invalid Argument ("Hans Verbeek")
Re: Apache keep alive cron (Juergen Heinzl)
Re: tcp wrappers - help with troubleshooting? (Juergen Heinzl)
Re: Pentium with CD -> 486 without ([EMAIL PROTECTED])
Re: sendmail queuing not delivering (Jayasuthan [VorHacker])
Public_html ("R. Brett Juergens")
Re: pppd error message ("Tim Underwood")
[Q] RH 5.1 with 3c905B TX NIC card (Chang Cheng Chao)
NFS from Win95 to Linux (Steve Ledford)
----------------------------------------------------------------------------
From: Ashutosh Sharma <[EMAIL PROTECTED]>
Subject: How to set up a DNS ??
Date: Tue, 12 Jan 1999 00:03:30 +0530
How can one set up a DNS on a linux box??
Please help.
Regards
Ashutosh
------------------------------
From: "Jeff Volckaert" <[EMAIL PROTECTED]>
Subject: Re: How to setup routing?
Date: Mon, 11 Jan 1999 10:42:31 -0500
Where is your default route? Try "route add default ppp0" if you want your
Net traffic to go out the dial up. Switch ppp0 for eth0 if you want Net
traffic to go out the firewall.
You firewall is probably blocking incoming Net stuff for your own good. You
should be able to telnet to the ppp0 address though.
Can you telnet to eth1?
Jeff Volckaert
Conrad Hagemans wrote in message <77b6m6$9b5$[EMAIL PROTECTED]>...
>I have a Red Hat (5.1) Linux box setup that is hooked up to the internet
>with a PPP connection to our ISP.
>I have a network card installed to hook the box onto out local network. And
>i added a second networkcard to
>connect the machine to other servers in front of the firewall. The box has
>an official IP address.
>
>Our company has an official c-class domain. 195.240.233.0
>
>Our internal network is using an in-official network range (but is behind
>the firewall) 120.120.0.0
>
>The PPP connection has ip address 195.240.233.1
>The internal network card ip address 120.120.18.111
>The card for use in front of the firewall IP address 195.240.233.10
>
>I connected the in front the fireall card to a hub to link the box to an
>extra server with IP address 195.240.233.20
>
>This server cannot be reached from the internet. Why???????
>
>Here are some printouts from the routing table and the rc.local with the
>ipfwadm statements:
>
>Kernel IP routing table
>Destination Gateway Genmask Flags MSS Window irtt Iface
>154.9.48.67 0.0.0.0 255.255.255.255 UH 1500 0 0 ppp0
>195.240.233.0 0.0.0.0 255.255.255.0 U 1500 0 0
eth0
>120.120.0.0 0.0.0.0 255.255.0.0 U 1500 0 0
>eth1
>127.0.0.0 0.0.0.0 255.0.0.0 U
3584
>0 0 lo
>0.0.0.0 154.9.48.67 0.0.0.0 UG 1500 0
>0 ppp0
>
>154.9.48.67 is the fixed IP address from the ISP side of the PPP
connection.
>eth0 is the NIC in front of the firewall
>eth1 is the NIC to the internal network
>
>The rc.local file (fragement)
>ipfwadm -F -p deny
>ipfwadm -F -a m -S 120.120.0.0/255.255.0.0 -D 0.0.0.0/0
>ipfwadm -F -a m -S 0.0.0.0/0 -D 195.240.233.0/255.255.255.0
>ipfwadm -F -a m -S 195.240.233.0/255.255.255.0 -D 0.0.0.0/0
>
>Can anyone help me to get this solved.
>
>
>
------------------------------
From: "David H." <[EMAIL PROTECTED]>
Subject: Re: Proxy + firewall
Date: Mon, 11 Jan 1999 13:07:44 -0600
I'd like info on that too. I'm going to try to get a network with the
TCI@home network, and need a proxy server --> Proxy server connection (tci
proxies everything). The sysadmin does not want the internet connection to
the hub, just to the one computer.
============= Posted via Newsfeeds.Com, Uncensored Usenet News ============
http://www.newsfeeds.com/ The Largest Usenet Servers in the World!
============= Over 66,000 Groups, Plus a Dedicated Binaries Server ============
------------------------------
From: [EMAIL PROTECTED]
Crossposted-To: comp.os.netware.connectivity
Subject: Programming packet burst and long file names
Date: Mon, 11 Jan 1999 19:12:13 GMT
Hallo!
I've written a NetWare client for a non-PC, non-mac platform sometimes known
as Acorn RiscOS, but have run into two problems:
1. When I try to scan a directory or open a file with a long file name
the server always returns an error. I am setting the NameSpace byte(s)
to 0x04 for OS2 name space.
Is there some magic handshaking required when I start my connection?
I've been looking at the Linux NCPFS stuff, and it does a few things
that I don't but these aren't commented and seem to be related to packet
signing.
2. How is packet bursting implemented? I know they have type 0x7777 in the
IPX header followed by a 36-byte header, but apart from a few fields in
this header that obviously indicate position and length inside the block,
I don't know how to set up a read from a file with this.
Are Novell people allowed to comment on this???
Regards,
Frank Vorstenbosch
See http://www.falstaff.demon.co.uk/nwclient.html for the software.
========================================================================
Frank A. Vorstenbosch <SPAM_ACCEPT="NONE"> Mobile: +44-976-430 569
Wimbledon, London SW19 Home: +44-181-544 1865
[EMAIL PROTECTED] Office: +44-181-636 3391
============= Posted via Deja News, The Discussion Network ============
http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own
------------------------------
From: [EMAIL PROTECTED]
Crossposted-To: comp.os.netware.connectivity
Subject: Programming packet burst and long file names
Date: Mon, 11 Jan 1999 19:12:40 GMT
Hallo!
I've written a NetWare client for a non-PC, non-mac platform sometimes known
as Acorn RiscOS, but have run into two problems:
1. When I try to scan a directory or open a file with a long file name
the server always returns an error. I am setting the NameSpace byte(s)
to 0x04 for OS2 name space.
Is there some magic handshaking required when I start my connection?
I've been looking at the Linux NCPFS stuff, and it does a few things
that I don't but these aren't commented and seem to be related to packet
signing.
2. How is packet bursting implemented? I know they have type 0x7777 in the
IPX header followed by a 36-byte header, but apart from a few fields in
this header that obviously indicate position and length inside the block,
I don't know how to set up a read from a file with this.
Are Novell people allowed to comment on this???
Regards,
Frank Vorstenbosch
See http://www.falstaff.demon.co.uk/nwclient.html for the software.
========================================================================
Frank A. Vorstenbosch <SPAM_ACCEPT="NONE"> Mobile: +44-976-430 569
Wimbledon, London SW19 Home: +44-181-544 1865
[EMAIL PROTECTED] Office: +44-181-636 3391
============= Posted via Deja News, The Discussion Network ============
http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own
------------------------------
From: "David H." <[EMAIL PROTECTED]>
Subject: Re: Newbie modem problem
Date: Mon, 11 Jan 1999 13:11:29 -0600
Winmodem. You're screwed.
============= Posted via Newsfeeds.Com, Uncensored Usenet News ============
http://www.newsfeeds.com/ The Largest Usenet Servers in the World!
============= Over 66,000 Groups, Plus a Dedicated Binaries Server ============
------------------------------
From: "David H." <[EMAIL PROTECTED]>
Subject: Re: pppd with Rockwell HCF PCI Internal modem
Date: Mon, 11 Jan 1999 13:14:07 -0600
It is a winmodem. Need I say more. The only way to get this to work (i have
the same modem) is to buy a new one, that does not say winmodem on it.
These modems are usually more expensive, but are the only way it will work.
============= Posted via Newsfeeds.Com, Uncensored Usenet News ============
http://www.newsfeeds.com/ The Largest Usenet Servers in the World!
============= Over 66,000 Groups, Plus a Dedicated Binaries Server ============
------------------------------
From: Sean McEwan <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux,linux.redhat.misc
Subject: Re: su won't let me shutdown/ifconfig
Date: Mon, 11 Jan 1999 13:28:38 -0600
When you "su root", look at your $PATH with echo. Shutdown is in the /bin or
/sbin directory, I think. When you "su root", you get superuser powers, but
your path doesn't change from the previous login. Once you find the directory
shutdown is in, you can type "./shutdown -h 0".
"A.G." wrote:
> I have read in many a post that one shouldn't logon as root.
>
> Alright, I have set up an account for myself to log on. But how do I
> shutdown the system in the end of the session? "su" doesn't help - I get
> "command not found" message when I try to enter shutdown or ifconfig for
> example.
>
> I have to log on as root at the end of each session to only shutdown. This
> is at the very least inconvenient :). I realize that it's probably my
> ignorance, that explains my not knowing a way out of this.
>
> Please advise,
>
> A.G.
------------------------------
From: "Jan Stifter" <j.stifter@[no-spam]usa.net>
Subject: what to filter in a packet filter
Date: Mon, 11 Jan 1999 16:55:03 +0100
hello out there,
I am setting up a firewall with ipchains.
question: is there an example of a *secure* configuration for a small sized
company with www, dns, mail, news, ping, traceroute ?
thank you all
Jan Stifter j.stifter@[no-spam].usa.net
------------------------------
From: [EMAIL PROTECTED] (Chris)
Subject: Re: ADSL
Date: Mon, 11 Jan 1999 19:31:44 GMT
On Fri, 08 Jan 1999 00:38:08 -0500, Dan Falk <[EMAIL PROTECTED]>
wrote in comp.os.linux.networking:
> Interestingly, the login is
>web-based! If this is the case with your provider ... look for a login
>script!! The whole web-based login is stupid.
Unfortunately, there is still no standard way of informing a DNS about a
DHCP lease. The web login pages tend to act as front-ends that launch a
CGI that "manually" updates the DNS tables.
My provider uses a secure web page for the login, so I also had to install
SSLeay and OpenSSL in order to implement a perl login script.
The whole issue is rather automated now-- I used the "-c" option of dhcpc
to run a post-lease script each time an address is obtained from the DHCP
server. That script takes care of any required changes that result from
leasing a different address, and then calls the perl login script.
------------------------------
From: [EMAIL PROTECTED] (Chris)
Subject: Re: Q: DHCP & static IP coexist on one card possible?
Date: Mon, 11 Jan 1999 19:42:09 GMT
[posted and mailed]
On 7 Jan 1999 16:54:48 -0500, [EMAIL PROTECTED] (Andrew
Duchowski) wrote in comp.os.linux.networking:
>Hi there! I normally have the laptop (running RH 5.1) plugged into the
>wall ethernet connection at my desk, for which I have a static IP address.
>I can start eth0 ok, works ok (save for the SIOCADDRT invalid argument error
>which I can't get rid off yet the network still functions). Occasionally I
>have to take the laptop and plug it into a mobile DHCP port at another
>location.
Assuming you are shutting down and powering up your laptop between
connections, you could configure the system to ask you what run-level you
want to use (by not putting a default in /etc/inittab), and then use
run-level 2 for the static-IP connection and run-level 3 for the DHCP
connection.
------------------------------
From: "jay" <[EMAIL PROTECTED]>
Crossposted-To:
alt.linux,alt.os.linux,comp.os.linux.hardware,comp.os.linux.help,comp.os.linux.misc,linux.redhat.install,linux.redhat.misc
Subject: Re: Multiple NIC Cards.. HELP!! <solution>
Date: Mon, 11 Jan 1999 19:43:31 GMT
Sorry, guys. you all are only part right- the OS already sees the two
identical cards, but it's delaying the one because the kernel needs to be
told to init the other - to do this add the following statement to your
lilo.conf in etc/
append="ether=0,0,eth1"
that's it, dude. e-mail me @ [EMAIL PROTECTED] if you have more problems- i
fought this for a couple months, but once you get it, it works great.
------------------------------
From: Erwann CORVELLEC <[EMAIL PROTECTED]>
Subject: DMZ setup
Date: Mon, 11 Jan 1999 16:53:39 -0100
I would like a network looking like this:
Linux Public WWW Server
| /-- Local Windows PC1
Internet (CISCO) === Linux FireWall --- HUB --- Local Windows PC2
\-- Local Network Printer
- Where can i install a DNS server for the local network ?
(on the FW or on a PC on the local network ???)
- I would like my PCs to be user independent. Can a PC server running under Linux let
the users log on any PC running Windows as a Windows NT solution would ? I mean, can a
PC under Win9x act as a NC ?
- What network standard would you suggest between the WWW server and the FW ? 10base2
BNC ? 100baseT ?
Thank you for your help !
------------------------------
From: "Hans Verbeek" <[EMAIL PROTECTED]>
Subject: Re: SIOCADDRT: Invalid Argument
Date: Mon, 11 Jan 1999 17:15:25 +0100
I have no suggestions. Only a remark: I see this message since I upgraded to
RedHat 5.2.
I guess there is a new driver in kernel 2.0.36. As I work with both NE2000
and 3COM I guess
it is not at the direct driver layer. Any ideas?
Greetings,
[Hans Verbeek]
Art Asbury wrote in message <[EMAIL PROTECTED]>...
>On Fri, 08 Jan 1999 21:50:34 -0500, "Brian D. Cook"
><[EMAIL PROTECTED]> wrote:
>
>>I have a dual 133 linux box. It's got two nic's in it, both are
>>3c905b-tx's. When I boot, I get this error, SIOCADDRT: Invalid
>>Argument.
>>Any Suggestions would be appreciated.
>Yes any suggestions. I have been chasing this for weeks.
>
>>Brian D. Cook
>>Web Space Administrator
>>Greer Toyota-Subaru
>>Wappingers Falls, NY
>
------------------------------
From: [EMAIL PROTECTED] (Juergen Heinzl)
Subject: Re: Apache keep alive cron
Date: Mon, 11 Jan 1999 20:26:01 GMT
In article <[EMAIL PROTECTED]>, Yaron wrote:
>I need a script that would check the apache server every 15 min and see
>if it is killed. If so it should revive it. The host is not running by
>root but by another user
>Can anyone help
use pidof but there are at least two versions floating around and with the
one here (sysvinit) ...
[/path/to/]pidof -s httpd > /dev/null 2>&1
if [ $? -ne 0 ]
then
restart
fi
...
You could evaluate the pid file apache creates too and call kill -0 pid,
yes 0 ... is that pid there, but if the uptime is long enough the pid
might be reused by another process and you would have to use ps + a
filter or evaluate /proc/<pid>/cmdline or status, whatever suits you.
Bye, Juergen
--
\ Real name : J�rgen Heinzl \ no flames /
\ EMail Private : [EMAIL PROTECTED] \ send money instead /
\ Phone Private : +44 181-332 0750 \ /
------------------------------
From: [EMAIL PROTECTED] (Juergen Heinzl)
Subject: Re: tcp wrappers - help with troubleshooting?
Date: Mon, 11 Jan 1999 20:26:00 GMT
In article <77dclv$tuv$[EMAIL PROTECTED]>, [EMAIL PROTECTED] wrote:
>what is spawn? Where does it come from? I hate using things I don't know,
>especially for something related to security.
[...]
>In article <[EMAIL PROTECTED]>,
> [EMAIL PROTECTED] wrote:
>
>> Use spawn in there. For example something along these lines.
>>
>> in.fingerd : ALL : spawn (/usr/sbin/traceroute %h | \
>> mail -s "Finger Trace Results" root) &
>>
... will run traceroute and mail the result too root via a subshell. One
disadvantage I see here is ...
while :
do
finger ...
done
... on the other side, but I guess it is not *that* much of a risk.
Cheers,
Juergen
--
\ Real name : J�rgen Heinzl \ no flames /
\ EMail Private : [EMAIL PROTECTED] \ send money instead /
\ Phone Private : +44 181-332 0750 \ /
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: Pentium with CD -> 486 without
Date: 11 Jan 99 09:53:43 MDT
I am currently doing this my self. I have tried SMB installs, FTP installs,
and am now doing an NFS install. I am using RedHat 5.1, but I am keeping W95
on the pentium for the family. So, as a result I have installed Linux onto a
Jaz disk on the Pentium. From here I am doing the NFS install to the 486. I'd
have this done now, but I didn't size all the partitions right on the Jaz
install and thus can't edit any files to facilitate the NFS install as the
partition containing /etc is full. I plan to redo everything tonight, so I
will report agian tomarrow.
Eric Winsor
[EMAIL PROTECTED]
In article <[EMAIL PROTECTED]>, Barry O'Neill
<[EMAIL PROTECTED]> writes:
> Mikhail Bovineck wrote:
>>
>> Hi - I have Linux installed on a hardrive on a pentium box which
>> originated on a CD ( RedHat Distribution). I also have a 486
>> with no CD which I would like to make a Linux box as well. In other
>> words, I would like to access the pentium box's CDrom from the 486.
>
> I've recently done this. To my shame, I opted to physically mode the
> CDROM to the 486 for the installation rather than learn something by
> using the existing network connection.
>
> That said, the installation (RedHat) worked without incident and the
> CDROM is now back in the Pentium, while the 486 is now web serving, news
> hosting and mail handling.
>
> You would be better off (IMHO) with a couple of cheap network cards to
> link your two boxes, rather than having to go down the PLIP/Serial
> connection route.
>
> regards,
>
> Barry
> --
> Linux Redhat 5.2. BeOS R4.
> Who needs Micro$lop?
------------------------------
From: Jayasuthan [VorHacker] <[EMAIL PROTECTED]>
Subject: Re: sendmail queuing not delivering
Date: 11 Jan 99 19:57:02 GMT
You have send a big mail ! My server limiting me
you need to edit smart relay .......
: # ... define this only if sendmail cannot automatically determine your domain
: #Dj$w.Foo.COM
: CP.
: # "Smart" relay host (may be null)
: DS
Try edit /etc/syslog.conf
*.* /var/log/messages
*.* /dev/console
reboot and see what appear on this two place went sending email
tail -f /var/log/messages
: # deal with plussed users so aliases work nicely
: R$+ + * $#local $@ $&h $: $1
: R$+ + $* $#local $@ + $2 $: $1 + *
------------------------------
From: "R. Brett Juergens" <[EMAIL PROTECTED]>
Subject: Public_html
Date: Mon, 11 Jan 1999 13:57:08 -0600
I'm trying to allow users on our web server to have home pages. I've
placed UserDir public_html in srm.conf and
<Directory ~/public_html>
AllowOverride None
Options Indexes Includes FollowSymLinks
order allow,deny
allow from all
</Directory>
in the access.conf file. I've also added the directory
/home/someuser/public_html
and placed an index.html file there. The directory and file are set to
someuser.someuser.
I haven't made a .htaccess file in the directory. Do I need this file
and what do I place in it. Also, is there any other files that I should
make? Thanks in advance Brett
------------------------------
From: "Tim Underwood" <[EMAIL PROTECTED]>
Subject: Re: pppd error message
Date: Mon, 11 Jan 1999 15:46:02 GMT
I am also getting the message. I have tried rebuilding the kernel (the docs
are kind of fuzzy here - it appears that part of ppp is built into the
kernel).
I D/L'd 2.0.36 source (gave up trying to figure out how to rebuild the
kernel with the RH 5.2 Source CD). Installed the .rpm for ppp 2.3.5, ran
per README in /usr/src/ppp-2.3.5/README.
make dep - OK
make modules - errors in ppp.c (wrong # of args, etc.)
How do I rebuild the kernel to get the 2.3.5 ppp in? I must have demand
dialing for what I want to be able to do. Or should I just drop 10, punt,
and use diald instead?
Art Asbury wrote in message <[EMAIL PROTECTED]>...
>On Wed, 6 Jan 1999 08:23:31 -0800, "Darrin" <[EMAIL PROTECTED]> wrote:
>
>You me and several others. I have one machine with RH5.1 which
>I updated with the RPM 2.3.5 from RH and another with RH5.2 which
>includes 2.3.5 both give the same error.
>
>>I get an error message of:
>>
>>pppd: demand dialling is not supported by kernel driver version 2.2.0
>>
>>But my ppp version is 2.3.5-2.
>>
>>Can anyone tell me what kernel driver version this message is referring
to??
>>
>>I'm trying to get on demand dialing to work, any pointers that would lead
me
>>in that direction would be greatly appreciated.
>>
>>Thanks
>>
>>
>
------------------------------
From: Chang Cheng Chao <[EMAIL PROTECTED]>
Subject: [Q] RH 5.1 with 3c905B TX NIC card
Date: Mon, 11 Jan 1999 11:29:57 -0800
Has anyone made this card work with RH 5.1 before? During the
installation, it detected it and I had chose DHCP as the option. But it
still does not work. What should I do?
Chang
------------------------------
From: Steve Ledford <[EMAIL PROTECTED]>
Subject: NFS from Win95 to Linux
Date: Mon, 11 Jan 1999 13:15:03 -0600
Hello,
I have a bare bones 486 running RH 5.1 with a puny 200 meg hard drive
in it. I can't load the entire RH packages because of disk space
constraints for obvious reasons. I was wondering if anyone has
experience with Win95 on a 10BaseT LAN exporting via NFS or other
mechanism volumes so I can complete the installation of the kernel
source, X, etc for the Linux box to use. The Win95 maching has a 6.4G
hard drive that I could have a portion dedicated to the Linux machine
but the Win95 box needs to run Win95 all of the time. What
programs/setup is required on the Win95 box? Can this be accomplished
with Samba as I already have that running and if so how?
Thanks in advance.
------------------------------
** 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
******************************
