> I've not seen it mentioned on this list, so its worth pointing out a
> that there is a 3 page article in LinuxUser (Issue 3 - September
> 2000), basically covering Open Source in education, comparing the UK
> to Mexico, and covering a few things that will be familiar to many
> here (Powys servers, Becta, Roger Whittaker at Suse...)
This is now available online as a PDF - see
for the articles index, and
for the article itself.
[ - Opinions expressed are personal and may not be shared by VData - ]
[ Nigel Metheringham Nigel.Metheringham(a)VData.co.uk ]
[ Phone: +44 1423 850000 Fax +44 1423 858866 ]
I am prototyping a squid box at home before installing it at school. I
made 1 change to squid.conf to allow access, used wvdial and then pointed
my client's browser to 172.19.50.2:3128.
Well it sort of works but I initially get errors along the lines of
"Unable to determine IP address from host name for www.etc....."
"The dnsserver returned- Name server unavailable"
If I browse on the server with Netscape I eventually access the
Internet/cache on the client.
I have seen references in this group to "named" which is not running on my
box. When is "named" needed and do I need any other changes for resolving
to work reliably.
When all this works well I will need to use it with a router which
accesses the LEA proxy server. This has an authentication process based on
IGear software. Anyone any experience with this and knowledge about how I
can get squid to talk to it.
I have my linuxbox dialing into NTL and running squid on a small network.
ATM Im using wvdial, manualy, but wnat to have dial on demand.
I know diald is the thing to use... but dont have a clue how to set it up.
Can anyone help.
Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com.
Share information about yourself, create your own public profile at
First post from me so hello folks! I have just installed Suse Linux 7.0 and
as a first time Linux user I seem to have fallen at the first hurdle. The
system was running OK, booting to the command line and I could run KDE with
startx. I then decided to (this is where I start talking gibberish because I
am now at work, my PC is at home and I forgot to bring in my notes) edit
id:3:initdefault to id:5:initdefault which should let me boot directly into
X. The system now hangs on bootup saying "No more processes at this
runlevel". What's happening?
The Gordon Schools
From: jlamb [SMTP:firstname.lastname@example.org]
Sent: Friday, October 06, 2000 9:12 AM
Subject: [suse-linux-uk-schools] Run levels
>First post from me so hello folks! I have just installed Suse Linux 7.0 and
>as a first time Linux user I seem to have fallen at the first hurdle. The
>system was running OK, booting to the command line and I could run KDE with
>startx. I then decided to (this is where I start talking gibberish because I
>am now at work, my PC is at home and I forgot to bring in my notes) edit
>id:3:initdefault to id:5:initdefault which should let me boot directly into
>X. The system now hangs on bootup saying "No more processes at this
>runlevel". What's happening?
I'm only guessing here, as I don't have much info about your setup, nor
that of Suse as I've never used that distribution.
I'd guess that you don't have anything in /etc/inittab set to run on
runlevel 5. On top of that, it looks like you haven't got any services
defined in /etc/rc.d/rc5.d (or the problem in your /etc/inittab is
causing these services to be omitted).
The "No more processes..." message is the one that you should normally
get when you are shutting down the PC. It means that there are no more
jobs running that were started when you booted/swapped to this runlevel.
Obviously this is happening immediately because none are getting
There is a way of telling LiLo to boot in single user mode (runlevel 1 I
think) to enable you to get back in and edit back to runlevel 3 to allow
you to fix the problem. Unfortunately, I can't remember what it is.
Hope this help,
>The Gordon Schools
Work: http://www.ringways.co.uk mailto:email@example.com
REVCOM: http://www.revcom.org.uk mailto:firstname.lastname@example.org
Murphy's Laws: (327) Build a system that even a fool can use, and only a fool will use it.
I would like to announce the Linux for Schools Project which I have been
working on. It is an open source software project for Linux, supported by
St John's School in Northwood, London.
There are downloads in RPM format for:
- Creating and removing users en-masse
- Automatically indexing users' home pages
- Providing simple online help in telnet sessions.
All the packages are available under the terms of the GPL. I would like to
invite everyone to evaluate the packages and let me know your feedback.
The address is http://www.lynxafrica.demon.co.uk/lfsp/.
FREE Personalized Email at Mail.com
Sign up at http://www.mail.com/?sr=signup
Now I am hoping that the Squid users out there can help me. I (well my boss !) would like to be able to set up Squid in such a way that Sixthform users can have free access to the Internet but lower school can only have access to sites that we specify.
Here are my thoughts. I can dictate the IP address and port number that a user has to connect to Squid. So what I thought was have Squid on either 2 IP addresses or 2 ports (8080 & 8181). Then set ACL lists accordingly. This is where I come unstuck. I can't quite see how to block all outgoing requests on port 8181 except for requests that are contained in a list.
I hope someone understands the question for a start let alone has an answer !
I suggested to my boss that he buys another machine, it would be dead easy with 2 ! He didn't like that much . . .
The Purbeck School
> building a diskless workstation based around a 2.2.13 kernel
> (Mandrake) and having problems with NFS once the machine has
> booted from the /tftpboot/vmlinuz.xterm, run from the NFS-mounted
> root and come up with a mingetty logon.
So - I assume that your root is mounted ok, and you are
now trying to mount the contents of fstab. Do any of the
additional mounts (from fstab) work - or do they all fail ?
I am a little confused that you say it fails after coming up
with the logon prompt - does this mean that they were
all originally mounted on startup and now failing ?
Is there something in the startup process that is stopping
or restarting the mountd/nfsd servers or something,
after the Xterm has already mounted the disks ?
I think you can turn on additional logging for mountd/nfsd
if you run them in a console - so that you can debug things
a little more (can't remember the details - sorry).
Don't know if this helps or not, but when I was experimenting
with diskless machines for our clustering project, I found
the beowulf howto very useful. This had info on diskless
machines, also a script that setup the diskless NFS root
images on the server, using copies and symlinks (note:
I had problems, as I setup my main servers partition
and the NFS root images on separate partitions
- so symlinks did not work - does not sound like you
have this problem, but thought I would mention it !)
This was for RH 5.2 (when I used it) - but was quite useful
in seeing what you actually need to create.
Don't know if this helps or not ;-)
(I assume that you have seen the nfs-root mini
howos, etc) ...
> Kevin Taylors XDM document is good stuff
Thank you ;-)
I was just re-reading my list and you seemed to know something about proxies?
I have Squid running at my school on a rerasonably powerful server. We are
using dialnet as the parent cache and I am pointing all of the school's
stations (Win95) to the proxy in order to control things a bit (port
8080--Dialnet does this but I can control by room with Squid). I now want to
use Apache to serve our web site on the same machine but it is proving
1. Do I need to have the Linux box running Samba for the clients to access the
2. Can they both exist on the same box or do I have to use just Apache?
3. What is the key configuration in Apache for this to happen. I have squid
running but when I start up Apache (using it on port 8000).....hang on...
I have just experimented and I can now reach the Apache message if I point the
host machines to port 8000 but I have then lost the proxy capability. Is it
possible to point Apache to go to Squid for the outside world? How is this
Thanks Christopher (apologies Simon, I didn't choose Dialnet, I only work with
I'll give that a go later on. I did have a read of a Linux book last night
and got a few ideas but the port idea was alluding me. I have the internet on
8080 on Squid (for the clients) so I'll shuffle it around. I do have to go
out through 80 to get to the ISP though, does this have to be on the
>===== Original Message From Christopher Dawkins <cchd(a)felsted.essex.sch.uk>
>> I have Squid running at my school on a rerasonably powerful server. We are
>> using dialnet as the parent cache and I am pointing all of the school's
>Simon works for RM, so may not be so interested in Dialnet customers!
>> 1. Do I need to have the Linux box running Samba for the clients to
>> access the web pages?
>No, not at all, but Samba is very useful anyway and you can.
>> 2. Can they both exist on the same box or do I have to use just Apache?
>Yes, Samba and Apache and Squid co-exist quite happily.
>> 3. What is the key configuration in Apache for this to happen. I have
>> squid running but when I start up Apache (using it on port 8000) ...
>Use Apache on port 80 for your intranet, Squid on 8080 for proxying, don't
>use 8000 at all.
>Set your intranet domain as "excluded domains" on your browsers, and
>point them all to Squid on 8080, they will then go to 80 on Apache for the
>intranet. Don't ask me how. If it doesn't work check very carefully your
>Squid and Apache configs. If Netscape, leave the "socks" entry blank.
>Christopher Dawkins, Felsted School, Dunmow, Essex CM6 3JG
>01371-820527 or 07798 636725 cchd(a)felsted.essex.sch.uk