Linux-Networking Digest #861, Volume #9 Tue, 12 Jan 99 18:13:47 EST
Contents:
INN-2.1 man pages and more INN stuff (Allen Ahoffman)
Re: PPP/Online with Cwix.com ISP (John Feldman)
ppp0 vs ttyyS0 vs cau0 (John Feldman)
routing problems + aliasing ("Lovro Vre�")
Re: Remotely killing X-server (mike burrell)
Re: /usr over nfs? (Geoff Short)
Only use host name alias as root? (Jeffery Cann)
Re: ftp configuration problem ? (Luca Colombi)
Re: e-mail system setup (Duncan Simpson)
Re: Remote X-terminal (Charles Reindorf US/EE1 60/1/44 #44278)
Re: ISA network card in PCI motherboard's (Scott W. Petersen)
----------------------------------------------------------------------------
From: Allen Ahoffman <[EMAIL PROTECTED]>
Subject: INN-2.1 man pages and more INN stuff
Date: 12 Jan 1999 20:55:39 GMT
I installed the inn-2.1.tar.gz file taken from isc.org recently, it
compiled fine, but when make install ran I didn't get the man pages I
thought.
I tried this using Debian LInux 2.0.
Also, I have created my cycbuff files for inn 2.1 using
dd if=/dev/zero of=filname bs=1k count=1000000
I created cycbuff entries for each buff
I created metacycbuff entries for 2 metabuffs splitting the cycbuffs in
half.
I got to the point where inncheck didn't complain, and then ran
makehistory -i
it tells me there is a problem initializing the storage manager,
any thoughts.
I have chowned everything to news;News
chmoded files appropriately to 664
made srue /usr/local/news/bin was in the path
pointed patharticles to /home/buffs
set the storageapi to yes and tried also true
help!
ISC sells "support" contracts for $21,000 per year.
Heck if I pay that much I'll pay someone to write it myself.
--
=======================================================================
| Announce communications Inc. | voice: 301-731-5786 |
| 5004 West Lanham Dr. | email: [EMAIL PROTECTED] |
| Hyattsville, MD 20784 | http: www.announce.com |
======================================================================
------------------------------
From: John Feldman <[EMAIL PROTECTED]>
Crossposted-To: alt.os.linux,comp.os.linux.misc,comp.os.linux.setup,comp.os.linux.x
Subject: Re: PPP/Online with Cwix.com ISP
Date: Tue, 12 Jan 1999 16:07:27 -0500
You are absolutley right. MCI2000, now CWIX, does use CHAP authentication.
At least they did when I was a customer 6 months ago. I wasted an entire day
wrestling with ppp when I had that account. I finally gave up and decided to
call MCI and see what type of authentication they were using,. The first tech
that I spoke to didn't know what PAP or CHAP was, the second one knew they
were using CHAP. Both gave me a big hassle about not supporting Linux.
The thing was, I wasn't askin for any freagin Linux support! ..just some basic
info about how their RAS servers were setup. Guess they don't teach that
stuff in tech support school.
Anyway, edit your chap-secrets file and you'll be good to go. If I remember
correctly, I too was using Ezppp at the time and it worked almost instantly
after I made the corrections.
Good Luck
"Daniel P. Fraga" wrote:
> Draco wrote:
>
> > do probably most ISPs. When I try to connect through EzPPP in Linux, I
> > can't get on. It never asks for a password or login name. I have tried
> > using the terminal window and it still will not ask for that info. I was
> > wondering if anyone knows anything I should do, or should I just get a
> > new ISP??
>
> Maybe your ISP uses PAP or CHAP authentication. Please check my
> page and see if it helps.
>
> --
> http://members.xoom.com/ilovelinux/
------------------------------
From: John Feldman <[EMAIL PROTECTED]>
Subject: ppp0 vs ttyyS0 vs cau0
Date: Tue, 12 Jan 1999 16:39:30 -0500
I've been fighting with a ppp connection (me and every other user on
this board it seems) under RedHat 5.2 for the last few days. I connect
OK, but then the connection is terribly slow.
Someone suggested that there was a resourcse conflict and they were
absolutlely right. The Motorola 28.8 internal modem was set to use irq4
and so was serial port 1 with is attached to my mouse.
So I jumpered the modem to IRQ3 and then, using "setserial", changed the
/dev/modem file to IRQ3 and the standard I/O for COM2 (0x2f8 I think).
The problem now is that the modem still seems to be responding to IRQ4.
Under control panel, I went into the modem configuration tool and
selected cau1. I was thinking that this was the right setting because
I've configured the modem for COM2.
Under the network configuration tool, I set the modem to "ttyS1".
(/modem and cau1 are not available options).
Now when I try to initiate a connection (using kppp), the modem only
responds to cau0 which I know is set to use IRQ4!
So my question here is, what is the relationship, in this scenario,
between ttyS1, cau1 and /dev/modem? If I use "setserial" to check their
rescources, they're all are using the same settings. If this were an
external modem, I think I'd have it working by now. The fact that it's
internal and sharing resources with an unused COM2 port is confusing the
heck out of me.
My other question is, how can this modem be responding to IRQ4 when It's
physically jumpered to IRQ3?
Please set this Linux virgin straight.
Thanks.
------------------------------
From: "Lovro Vre�" <[EMAIL PROTECTED]>
Subject: routing problems + aliasing
Date: Tue, 12 Jan 1999 13:21:10 GMT
Hi,
I have the following situation:
2 NIC
first NIC has address 172.16.1.18/255.255.0.0 and also have alias
193.77.168.220/255.255.255.128
Second NIC has address 193.77.168.18/255.255.255.128 and also have an alias
193.77.168.20/255.255.255.128
Here is my routing table:
Destination Gateway Genmask Flags MSS Window Use
Iface
193.77.168.20 * 255.255.255.255 UH 1500 0 0
eth1:0
193.77.168.0 * 255.255.255.128 U 1500 0 26 eth1
193.77.168.128 * 255.255.255.128 U 1500 0 1
eth0:0
172.16.0.0 * 255.255.0.0 U 1500 0 640 eth0
127.0.0.0 * 255.0.0.0 U 3584 0 19 lo
default 193.77.168.19 * UG 1500 0 5862 eth1
I have another machine connected to first NIC with address
193.77.168.169/255.255.255.128 and another to second NIC with address
193.77.168.23/255.255.255.128.
>From linux machine I can ping all mentioned addresses (193.77.168.18,
193.77.168.20, 193.77.168.23, 193.77.168.169 and 193.77.168.220).
>From machine 193.77.168.169 I can ping addresses 193.77.168.18,
193.77.168.20 and 193.77.168.220 (all Linux addresses), but I cannot ping
address 193.77.168.23.
The same situation is from station 193.77.168.23. I can ping all Linux
addresses, but I cannot ping address 193.77.168.169.
What should I do so machine 193.77.168.23 and 193.77.168.169. I must not use
IP masquerading, I just want that my Linux machine act as a router. IP
forwarding/gatewaying is included in Linux kernel.
Thank you, very much.
Regards,
Lovro
------------------------------
Crossposted-To: comp.os.linux,linux.redhat.misc
Subject: Re: Remotely killing X-server
From: [EMAIL PROTECTED] (mike burrell)
Date: Sun, 10 Jan 1999 19:06:01 GMT
In article <77asa2$nil$[EMAIL PROTECTED]>, A.G. <[EMAIL PROTECTED]> wrote:
*snip*
>I did "kill -s $DIFFERENT_NUMBERS 644", and most of them didn't do anythig,
>but I think that "kill -s 9 644" turned 644 into a "zombie", and so it did
>with all other processes, belonging to X. But the screen on the hung machine
>didn't change.
IINM, zombie processes will eventually be killed off by the kernel by
killing their parents. It's probably best (and fastest) to just kill their
parents yourself. To find their parenst, do a 'ps fax' and track down where
this zombie X is. Starting X (say, through 'startx') usually launches a
bunch of different scripts and stuff to load X, so the actual X server will
probably be nested pretty deeply. Anyway you may see something like this:
1556 1 S 0:00 -bash
1802 1 S 0:00 \_ sh /usr/X11R6/bin/startx -- -bpp 16
1803 1 S 0:00 \_ xinit /home/mikpos/.xinitrc -- -bpp 16
1804 ? Z 0:00 \_ /usr/X11R6/bin/Xwrapper :0 -bpp 16
1806 1 S 0:00 \_ wmaker
1813 1 S 0:00 \_ rxvt -transparent -bg black -fg white +sb
1814 p0 S 0:00 \_ bash
1816 p0 R 0:00 \_ ps fax
this is a good example of how deeply things can get nested in X. You'll
notice the 'Z' for the Xwrapper, which would mean that it's a zombie
process...being a zombie process, it cannot be killed via a SIGKILL ('kill
-SIGKILL', or 'kill -9'), so you have to kill its parent (xinit in this
case). So try 'kill -SIGKILL 1803'...if that doesn't work, 'kill -SIGKILL
1802', or 'kill -SIGKILL 1556'. Sometimes killing the controlling terminal
(by killing the shell) is the only way to get rid of it.
I'm not sure if this is entirely accurate, but it's interesting anyway. I
heard somewhere that the Linux kernel will kill off parents of zombies after
a while. However, if it finds a zombie process with no parents (the kernel
is it's only parent), then it comes to the conclusion that the only way to
kill the process is to kill the kernel, and you end up with a kernel panic,
and a dead machine. I've never heard of anyone running into one of these
orphaned zombies, but it's kind of interesting.
>I admit that I didn't really know what I was doing with all these -s
>switches, so can anyone please explain, what would have been a correct
>procedure for this?
Might I suggest, 'man kill', 'man killall', and 'man 7 signal'...for day to
day rogue applications, you can probably get away with just knowing SIGTERM,
SIGKILL (and maybe SIGHUP). Most of the signals aren't designed to be sent
to processes manually anyway (like SIGINT, SIGSEGV, SIGPIPE, etc.). One of
the most misleading things about the 'kill' and 'killall' applications is
that they're seldom used for actually 'killing' a process, but rather just
send signals to them.
Hope this is of some help.
--
m i k e b u r r e l l
[EMAIL PROTECTED]
http://mikpos.dyndns.org
------------------------------
From: [EMAIL PROTECTED] (Geoff Short)
Subject: Re: /usr over nfs?
Date: 12 Jan 1999 12:07:46 GMT
Markus Hauke ([EMAIL PROTECTED]) wrote:
:
: I�d like to install Linux on a Network of about 10 PC�s. One of them
: should be a nfs server with lots of disk space. My question is:
: Is it possible to install my Linux distribution with all the needed
: programs on the server and then do a "mount -t nfs server:/usr /usr" on
: all clients???
: On the clients should only run a minimal Linux system, so that they can
: boot on their own and then do mount the /usr tree from the server.
Yes, most unix systems seem to be designed to do this, and allow /usr
to be read-only if you like. Certainly Slackware assumes that /usr
may be nfs mounted.
As long as the system startup scripts mount /usr as soon as possible, you
shouldn't have any problems.
Geoff
--
============================================================================
Ever sit and watch ants? They're always busy with Geoff Short
something, never stop for a moment. I just [EMAIL PROTECTED]
can't identify with that kind of work ethic. http://kipper.york.ac.uk/~geoff
------------------------------
From: Jeffery Cann <[EMAIL PROTECTED]>
Crossposted-To: alt.os.slackware
Subject: Only use host name alias as root?
Date: Tue, 12 Jan 1999 06:29:54 -0700
Greetings.
I have built a simple 2 node ethernet LAN. I have 2 Linux machines
connected via x-over cable. No gateway and no local nameserver. I am
running Slackware 3.5 on each machine, kernel 2.0.34. I used the
'netconfig' program for configuration.
Network IP: 192.168.0.0
Jumanji: 192.168.0.1
Dijbuti: 192.168.0.2
Everything is working great. I can login in to either console and ping,
ftp, telnet the other machine. I have added the host name aliases to
the /etc/hosts file on each machine:
#
# hosts
#
# For loopbacking.
127.0.0.1 localhost
192.168.0.1 Jumanji.Kilamenjaro Jumanji
192.168.0.2 Dijbuti.Kilamenjaro Dijbuti
The only problem I have is that when logged on to Jumanji, I can only
'telnet (ping, ftp) Dijbuti' as root user. Any other user will report
'unknown host Dijbuti'. Note that as any user, I can directly ping,
ftp, telnet the IP address 192.168.0.2. The reveerse is not true: When
logged on to Dijbuti, I can telnet, ping, ftp Jumanji using both its
host name alias or IP address.
AFAIK, the only difference between configuration on Jumanji and Dijbuti
is that Dijbuti has extra configuration parameters in my
/etc/resolv.conf to allow nameserver lookups when dialed out to my ISP
(via ppp).
Suggestions are appreciated. Thank you.
Jeffery Cann
--
"Who does not trust enough will not be trusted."
- Lao Tsu
------------------------------
From: Luca Colombi <[EMAIL PROTECTED]>
Crossposted-To: alt.os.linux,comp.os.linux.misc,uk.comp.ol.linux
Subject: Re: ftp configuration problem ?
Date: Tue, 12 Jan 1999 13:28:54 +0000
==============468FF85DD2C7DB23B55A1F97
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Try to edit your /etc/hosts file and enter the following.
192.168.0.1 name1'machine@yourdomain namemachine
192.168.0.2 name2'machine@yourdomain namemachine
192.168.0.3 name3'machine@yourdomain namemachine
[EMAIL PROTECTED] wrote:
> Hello,
> I am new to Linux. I have just installed, Red Hat 5.1.
> It is very stable !
> Now, I have created a little local network, 3 pcs:
>
> 1 192.168.0.1 Linux on it
> 2 192.168.0.2 W 95
> 3 192.168.0.3 W 95
>
> I have ftp, telnet and apache running. Apache works great. I have a problem
> with both ftp and telnet. If I type the command "ftp localhost", or ftp
> 192.168.0.1 from the Linux terminal itself, (samething with telnet) everything
> is fine, but if I do it from one of the W 95 machines, I instantly receive a
> message saying connected, but then it looks like "hang" for 1 or 2 minutes,
> until I see a welcome screen. If I try, WS_fgp95, for example, I never get
> connected.
> Does anybody have any idea of what is happening ?
>
> Thank you very much,
>
> Daniel Fernandez
> [EMAIL PROTECTED]
> Montevideo,
> Uruguay
>
> -----------== Posted via Deja News, The Discussion Network ==----------
> http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own
--
Luca Colombi
System Administrator
The Hub Communications Co. Ltd.
The Farmhouse
Syon Park
Middlesex
TW8 8JF
Tel: +44(0)181 560 9222 Fax: +44(0)181 560 9333
E-mail: [EMAIL PROTECTED] URL: http://www.thehub.co.uk
==============468FF85DD2C7DB23B55A1F97
Content-Type: text/html; charset=us-ascii
Content-Transfer-Encoding: 7bit
<HTML>
Try to edit your /etc/hosts file and enter the following.
<P>192.168.0.1 name1'machine@yourdomain namemachine
<BR>192.168.0.2 name2'machine@yourdomain namemachine
<BR>192.168.0.3 name3'machine@yourdomain namemachine
<BR>
<P>[EMAIL PROTECTED] wrote:
<BLOCKQUOTE TYPE=CITE>Hello,
<BR>I am new to Linux. I have just installed, Red Hat 5.1.
<BR>It is very stable !
<BR>Now, I have created a little local network, 3 pcs:
<P>1 192.168.0.1 Linux on it
<BR>2 192.168.0.2 W 95
<BR>3 192.168.0.3 W 95
<P>I have ftp, telnet and apache running. Apache works great. I have a
problem
<BR>with both ftp and telnet. If I type the command "ftp localhost", or
ftp
<BR>192.168.0.1 from the Linux terminal itself, (samething with telnet)
everything
<BR>is fine, but if I do it from one of the W 95 machines, I instantly
receive a
<BR>message saying connected, but then it looks like "hang" for 1 or 2
minutes,
<BR>until I see a welcome screen. If I try, WS_fgp95, for example, I never
get
<BR>connected.
<BR>Does anybody have any idea of what is happening ?
<P>Thank you very much,
<P>Daniel Fernandez
<BR>[EMAIL PROTECTED]
<BR>Montevideo,
<BR>Uruguay
<P>-----------== Posted via Deja News, The Discussion Network ==----------
<BR><A
HREF="http://www.dejanews.com/">http://www.dejanews.com/</A>
Search, Read, Discuss, or Start Your Own</BLOCKQUOTE>
<PRE>--
Luca Colombi
System Administrator
The Hub Communications Co. Ltd.
The Farmhouse
Syon Park
Middlesex
TW8 8JF
Tel: +44(0)181 560 9222 Fax: +44(0)181
560 9333
E-mail: [EMAIL PROTECTED] URL: <A
HREF="http://www.thehub.co.uk">http://www.thehub.co.uk</A></PRE>
</HTML>
==============468FF85DD2C7DB23B55A1F97==
------------------------------
From: [EMAIL PROTECTED] (Duncan Simpson)
Subject: Re: e-mail system setup
Date: 11 Jan 1999 13:40:47 GMT
In <773h1b$gp4$[EMAIL PROTECTED]> "Chrispy" <[EMAIL PROTECTED]> writes:
>Hi,
> I work in a small company (only 5 of us) and we currently have an e-mail
>account with an ISP here in the UK that allows unlimited addresses. I would
>like to set up a Linux machine to retreive all the mail from this account
>and put it into individual mailboxes for retreival in our office. Similarly,
>anyone in the office sending mail should be able to do this through the
>Linux box using a client (such as Netscape Mail or Outlook Express) on their
>own PC.
>I just need some pointers as to where to start with a setup like this, any
>help would be appreciated.
You can do this with a little sendmail trickery and a few other tools:
1. Mark all outgoing mailers as "expensive"
2. Set the HoldExpensive option in sendmail.cf
Now all outgoing mail gets queued and only sent when you explicitly
run the queue (do this when you are connected). You may want to use
your ISPs mnail server as a snart mail routing host---it is almost
certaintly be willing to act as one for you. (Windows users need this
feature due to mail software which is incapable of MX lookup).
3. Regularly fetch your mail using fetchmail. IMAP is the protocol of
choice if you can use it.
Now for the clients:
4. Install appropiate servers on your liunx box. Sendmail is a fine MTA
for sending mail. POP and IMAP servers are easily avialable and
some come on almost all linux CDs. (They may even be running after
in the default configuration).
I use fetchmail, sendmail, procmail and exmh for my mail.
Duncan (-:
------------------------------
From: Charles Reindorf US/EE1 60/1/44 #44278 <[EMAIL PROTECTED]>
Subject: Re: Remote X-terminal
Date: 12 Jan 1999 14:12:17 +0100
Gilbert Jacolbia <[EMAIL PROTECTED]> writes:
> Hi to all,
>
>
> Is it possible to acces my Linux box (RedHat 5.2) at home from the
> office and run programs under KDE environment?.
>
> I'm using Sun Sparc with Open windows at work . I usually telnet to my
> box at home and work on something but working under X-environment (GUI)
> is much faster and easier for me. By the way my connection at home is
> via cable modem and my linux box is always on.
>
> Is there any way to do it and How?.
You may find that running your KDE window manager on your home machine
so that it displays on your work machine does not perform that well,
depending on the network bandwidth available between the two machines.
Something like dxpc (*) may help.
Even if you find the above configuration hopeless, you can keep your
OpenLook window manager running at work and just run the X client
applications on the home machine, displaying on the work
machine. Again, dxpc may help you overcome network overhead problems.
The biggest hassle I have experienced with this kind of setup would be
(in your case) having to telnet from work to home, set-up the correct
authentication records, setting your DISPLAY environment variable and
launching the application you want. I ended up writing an 'expect'
script which did this (and solved some other problems to do with my
particular setup). email me if interested.
Alternatively: Look into 'ssh' - I know little about it but I've read
somewhere that it can tunnel 'X' sessions or something like that.
Regards,
Charles.
(*) Here it is:
dxpc - a Differential X Protocol Compressor
home page: http://ccwf.cc.utexas.edu/~zvonler/dxpc/
------------------------------
From: [EMAIL PROTECTED] (Scott W. Petersen)
Subject: Re: ISA network card in PCI motherboard's
Date: Tue, 12 Jan 1999 22:38:49 GMT
On Mon, 11 Jan 1999 12:24:34 +1000, ftc <[EMAIL PROTECTED]> wrote:
>yo..
>
>somebody told me that having a ISA network card in a PCI motherboard can
>cause problems, like packet loss and stuff...
>
>I'm just using some cheap ne2000's
>
>does anybody know if this is true ?
>
>
I have yet to see a problem with that. We have a lot of customers
running isa in a pci board. Never heard a problem. PCI is better
though for speed.
>
------------------------------
** 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
******************************
