Linux-Misc Digest #902, Volume #20 Sat, 3 Jul 99 03:13:10 EDT
Contents:
Re: Contribution request (was: Idea for new *nix site: yes or no?) (Scott Lanning)
Re: Video timings (Dave Hammond)
Re: Commercial Applications (Christopher Browne)
Re: Kernel 2.3.9 (Frederic L. W. Meunier)
Re: Where can I get free Linux CD? (Christopher Browne)
Re: Intel could nip dual-Celeron move in bud (Christopher Browne)
Re: Tell me I'm wrong... (Christopher Browne)
Re: AOL Instant Messanger for UNIX (brian)
Re: Linux Presentation Software (Allin Cottrell)
Re: Filesystem for SCO OSR and Redhat Linux ("Scot Harkins on MSN")
Re: Documentation issues. (Christopher Browne)
Re: AOL Instant Messanger for UNIX (Adrian Hands)
Re: Modem speed (Bill Unruh)
Re: Visual programming language for linux? (Ken Williams)
moving an entire installation to a different partition ("Ryan T. Rhea")
Re: Fragile file system (Christopher Wong)
Re: 10GB disk and LILO - I tried EVERYTHING! (Cameron L. Spitzer)
----------------------------------------------------------------------------
From: [EMAIL PROTECTED] (Scott Lanning)
Subject: Re: Contribution request (was: Idea for new *nix site: yes or no?)
Date: 3 Jul 1999 04:08:30 GMT
Kris ([EMAIL PROTECTED]) wrote:
: So, if anyone has any home-made [Perl|Tcl/Tk/bash/csh/expect]
: scripts,
Here's a couple small one's, some of which I've posted before..
#!/usr/local/bin/tcsh -f
#
# Usage : d.tcsh file[s]
# Description : Removes file[s] specified on the command line, and
# tells how many bytes were removed.
# Language : tcsh script
# Author : Scott Lanning, November 1998
##################################################################
foreach file ($*) #command line glob
set SIZE=`ls -l $file | awk '{print $5}'`
@ TOTAL += $SIZE # @ is for doing operations
if ( -e $file ) then #remove if exists
echo "removing $file ($SIZE bytes)"
rm $file
endif
end
echo "removed $TOTAL bytes total"
#!/usr/local/bin/perl
# description: lists files with optional specified permissions
# defaults to current directory, read/write/execute
# usage: lsmod <mod> <dir>
# where <mod> is like octal version of chmod
# author: Scott Lanning, 5/6/99
################################################################
use File::stat;
($perms = @ARGV ? shift : 0777) &&
($dirname = @ARGV ? shift : ".");
$perms = oct($perms);
opendir(DIR, $dirname) || die "can't open $dirname: $!";
while (defined($file = readdir(DIR))) {
next if $file =~ /^\.\.?$/;
$stat = stat("$dirname/$file");
printf("%s\n",$file) if ($perms == (0777 & $stat->mode));
}
closedir(DIR);
#!/usr/local/bin/expect -f
#automated telnet
##########################
#change these
set USERNAME username
set PASSWORD password
set HOST your.host.com
spawn telnet $HOST
#may have to change these
expect ogin:
exp_send "$USERNAME\r"
expect assword:
exp_send "$PASSWORD\r"
interact
exit
#!/usr/local/bin/perl -n
# utod.pl -- convert Unix to DOS format
# Unix format: end lines with \n (\015)
# DOS format: end lines with \r\n (\012\015)
#
# If -r (reverse) option on command line, converts
# DOS to Unix.
#
# usage: utod [-r] filename.in > filename.out
#############################################################
use Getopt::Std;
getopts("r");
$opt_r ? s/\r//
: s/\n/\r\n/;
print;
--
Scott Lanning: [EMAIL PROTECTED], http://physics.bu.edu/~slanning
"I'm going to have fun telling you about this absurdity, because I
find it delightful." --Richard Feynman
------------------------------
From: Dave Hammond <[EMAIL PROTECTED]>
Crossposted-To: linux.help
Subject: Re: Video timings
Date: Thu, 01 Jul 1999 22:36:40 -0400
Dave Hammond wrote:
>
> I'm trying to configure XFree86 for my monitor (CTX VL700),
> [...]
> Can anyone help with the correct dot clock and video timings?
Thanks to Howard Mann <[EMAIL PROTECTED]> and his website
http://www.newbielinux.com for pointing the way to the video
timings calculator at
http://www.hut.fi/Misc/Electronics/faq/vga2rgb/calc.html
--
DaveH
------------------------------
From: [EMAIL PROTECTED] (Christopher Browne)
Subject: Re: Commercial Applications
Reply-To: [EMAIL PROTECTED]
Date: Sat, 03 Jul 1999 04:27:51 GMT
On Fri, 02 Jul 1999 14:29:41 GMT, Toffer <[EMAIL PROTECTED]> wrote:
>Do any of you folk know where I can get hold of a good list of all
>commercial applications that currently exist in linux. In particular I
>am interested in what large organisations are currently up to.
<http://www.ntlug.org/~cbbrowne/linuxcommercial.html> has a
(moderately outdated) list of vendors; on the order of 200 of 'em...
>Who is developing serious heavy-weight stuff for linux? I mean: things
>like oracle, tuxedo .. .. etc .. ..
Presumably Oracle Corporation is developing Oracle, and BEA Systems is
developing Tuxedo; if others were, that would probably cause
problems...
--
The shortest distance between two puns is a straight line.
[EMAIL PROTECTED] <http://www.ntlug.org/~cbbrowne/rdbms.html>
------------------------------
From: Frederic L. W. Meunier <[EMAIL PROTECTED]>
Subject: Re: Kernel 2.3.9
Date: 3 Jul 1999 04:11:05 GMT
Well can anybody make the kernel 2.3.9 sucessfully with Vfat support? I find
that I can't make the kernel if I select vfat and fat support~
,,,
Again, and for the last time I say that:
If you want to use devel series, just do it.
If you don't know what to do with a broken kernel, use the stable one.
Use 2.2.10.
------------------------------
From: [EMAIL PROTECTED] (Christopher Browne)
Crossposted-To: alt.os.linux,comp.os.linux.setup
Subject: Re: Where can I get free Linux CD?
Reply-To: [EMAIL PROTECTED]
Date: Sat, 03 Jul 1999 04:28:09 GMT
On Thu, 01 Jul 1999 14:14:26 -0400, James Knott <[EMAIL PROTECTED]> wrote:
>In article <[EMAIL PROTECTED]>, [EMAIL PROTECTED] wrote:
>>Go to www.linuxmall.com. They have CDs for $1.89 plus shipping
>
>Or do what I did. Download the ISO image, and burn a CD.
Only $1 for the CD-R, but $300 for the burner...
--
"Barf, what is all this prissy pedantry? Groups, modules, rings, ufds,
patent-office algebra. Barf!" -- R. William Gosper
[EMAIL PROTECTED] <http://www.hex.net/~cbbrowne/hardware.html>
------------------------------
From: [EMAIL PROTECTED] (Christopher Browne)
Crossposted-To: comp.os.linux.hardware
Subject: Re: Intel could nip dual-Celeron move in bud
Reply-To: [EMAIL PROTECTED]
Date: Sat, 03 Jul 1999 04:28:48 GMT
On Fri, 2 Jul 1999 20:27:40 GMT, Chris Mauritz <[EMAIL PROTECTED]> wrote:
>In comp.os.linux.hardware Philip Brown <[EMAIL PROTECTED]> wrote:
>> On Tue, 29 Jun 1999 08:51:17 -0400, [EMAIL PROTECTED] wrote:
>>>That will just shift the mind share away from them and onto AMD. If Intel
>>>doesn't want to sell two celerons then I'm sure AMD would be happy to sell
>>>two K7s. I just hope the K7s drop below $200 in the next few months.
>>>
>
>> huh? Is AMD finally making multiprocessor capable cpus? Last I checked,
>> AMD chips didn't do that.
>
>The K7 (Altheon) supports SMP.
I *thought* that the K6 did too, but that mattered not at all, and
doesn't presently matter with the Altheon, as the critical component
is not merely the CPU, but the *MOTHERBOARD.*
There are no SMP motherboards for AMD chips (whatever the generation),
ergo no SMP systems based on AMD chips.
It doesn't appear that there are, in fact, *any* motherboards
available yet for the Altheon CPU, which means there are few systems
for sale thus far.
--
"I'd say the probability of Windows containing a backdoor is about the
same a spreadsheet containing a flight simulator."
--Phil Hunt <[EMAIL PROTECTED]>
[EMAIL PROTECTED] <http://www.hex.net/~cbbrowne/hardware.html>
------------------------------
From: [EMAIL PROTECTED] (Christopher Browne)
Crossposted-To: alt.linux
Subject: Re: Tell me I'm wrong...
Reply-To: [EMAIL PROTECTED]
Date: Sat, 03 Jul 1999 04:28:40 GMT
On Thu, 01 Jul 1999 17:39:40 GMT, HellNo <[EMAIL PROTECTED]> wrote:
>I downloaded a copy of SYSLINUX (linux on a floppy) and ran it of my
>floppy drive...
>
>After loging in as root (*no password*) I had complete access to /root
>after mounting my RH6.0 partition! It seems to have read/write/xecute
>access to all dirs. I even managed to reboot the machine
>using '/mnt/linux/usr/sbin/shutdown -r now' rather than using the
>reboot file that comes with SYSLINUX. Has anyone tried this? Someone
>please tell me I got it all wrong.
>
>Any input is more than welcome.
You want someone to tell you: "I got it all wrong."
OK. "I got it all wrong."
What happened here is that you have discovered the first principle of
computer security:
In order to construct a secure computer system, you must first ensure
that would-be intruders do not have physical access to your computer
facilities.
If someone can boot your computer system from a separate boot disk,
then your system is vulnerable to whatever they may wish to do with
it.
The *only* protection against this is to keep data on your hard drive
encrypted, using key information that is *not* maintained on the
computer system.
Any system that does not have tamper-resistant hardware and/or
encrypted filesystems will be vulnerable to this "attack."
Note that the flip side of this is that if you happen to *forget* the
root password on your system, you have discovered the "back door" that
may be used to fix things up.
--
"People are more vocally opposed to fur than leather because it's easier
to harass rich women than motorcycle gangs." [bumper sticker]
[EMAIL PROTECTED] <http://www.hex.net/~cbbrowne/security.html>
------------------------------
From: brian <[EMAIL PROTECTED]>
Subject: Re: AOL Instant Messanger for UNIX
Date: 3 Jul 1999 04:30:47 GMT
is there an RPM of any aim linux client? i have tried like 3 kinds and
have had install problems with them all. i run redhat6.0 and sometimes
mandrake
thanks a lot...
brian
Justin The Cynical wrote:
>
> On 10 Mar 1999 23:54:03 GMT, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote:
> ->In his obvious haste, eric <[EMAIL PROTECTED]> babbled thusly:
> ->: Hi,
> ->
> ->: I was wondering if AOL Instant messanger for unix will work with
> ->: Linux, more specific, Slackware 3.6?
> ->
> ->Is it available as source code, so it can be compiled for Linux?
>
> Yes and no. :-) The AIM client for Linux/*NIX is written in
> TCL/TK 8.0, which means it's a script of sorts. The client is viewable
with
> less, as it is just a plain text file/script. What this means is that
no, you
> don't have to compile anything, and yes, the download is an executable.
I am
> using it here on my home box, and it works nicely.
>
> Interesting, no? :-)
>
> --
> "NT disk, meet Mr. Microwave."
> David Parsons in comp.os.linux.advocacy (e-mail addy deleted for spam
reasons)
>
> Justin The Cynical - [EMAIL PROTECTED]
================== Posted via SearchLinux ==================
http://www.searchlinux.com
------------------------------
From: Allin Cottrell <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.development.apps
Subject: Re: Linux Presentation Software
Date: Sat, 03 Jul 1999 00:30:06 -0400
Kenvyn Davies wrote:
> I'm currently working on a web site (www.ivojo.com) for my business. I have
> a hint's and tips page which covers presentation software for Windows and I
> would very much like to extend this to cover Linux software...
>
> If anyone has any hints, tips etc. or knows of a resource which may be
> willing to share theirs then please email me of reply to this post.
http://www.ecn.wfu.edu/~cottrell/xdvi
--
Allin Cottrell
Department of Economics
Wake Forest University, NC
------------------------------
From: "Scot Harkins on MSN" <[EMAIL PROTECTED]>
Subject: Re: Filesystem for SCO OSR and Redhat Linux
Date: Fri, 2 Jul 1999 21:49:22 -0700
Crossposted-To: comp.unix.sco.misc
Good. This makes it easy. No.
Your only alternative is to run the data on a second machine and mount the
info from NFS for both OS'.
Unless someone out there has ported a driver for a Linux fs to OSR5.
Scot
--
Scot Harkins (KA5KDU) | Systems Administrator, Thurman Ind, Bothell, WA
North Bend, WA | [EMAIL PROTECTED]
[EMAIL PROTECTED] | SCA: Ld. Scot MacFin, Barony of Madrone, An Tir
[EMAIL PROTECTED]/msn.com | URL <http://www.bigfoot.com/~scoth>
<[EMAIL PROTECTED]> wrote in message news:7lfvpp$pap$[EMAIL PROTECTED]...
> In comp.unix.sco.misc [EMAIL PROTECTED] wrote:
> : I am in the process of moving from SCO OSR 5.0.5 to RedHat 6.0.
> : My machine dual boots between the two OS'es. My user's files
> : are in a separate partition which is currently SCO OSR's HTFS format.
> : RedHat 6.0 (at least as I configured it) does not "know" HTFS.
> : I would like to copy the entire contents of that file system
> : into another file system which both RedHat and SCO can access.
> : The file system MUST support soft links.
> : Does such a file system format exist? The only file systems I
> : found common to the two are the generic UNIX S51K file system and
> : DOS. Neither supports soft links.
>
> Perhaps I need to clarify because I have gotten multiple response
suggesting
> I create a tar file on a DOS partition, then mount the dos partition in
> Linux and untar. I am NOT interested in a one time copy from an SCO
filesystem
> to a Linux filesystem. I am interested in copying to a file system with
> softlinks and 255 char file names so I can continue to use both OSes
> over the next 6 months as I slowly migrate, each OS access the files
> in the common file system.
>
> --
> Arch
> +---------------------------------------------------+
> | Dr. J. Archer Harris Dept of Computer Science |
> | [EMAIL PROTECTED] James Madison University |
> | (540) 568 - 2774 Harrisonburg, VA 22807 |
> +---------------------------------------------------+
------------------------------
From: [EMAIL PROTECTED] (Christopher Browne)
Crossposted-To: gnu.misc.discuss,comp.unix.bsd.misc,comp.text.sgml,comp.text.xml
Subject: Re: Documentation issues.
Reply-To: [EMAIL PROTECTED]
Date: Sat, 03 Jul 1999 04:29:41 GMT
On Thu, 01 Jul 1999 08:06:44 GMT, Ketil Z Malde <[EMAIL PROTECTED]> wrote:
>[EMAIL PROTECTED] (Christopher B. Browne) writes:
>>> So, you're saying that in order to use MS Word, you need to know the
>>> user interface, the file format, the Win32 API, and HPCL or
>>> Postscript?
>
>> In order to troubleshoot MS Word problems properly, YES
>
>But that wasn't the question, and I maintain that with a properly
>designed and implemented (SG|X)ML publishing system, writers would only
>have to know the semantic structure of the documents - and the phone
>number to call when something goes wrong.
I'm not solely worried about the "dumb user," for whom the solution
you describe may very well be enough.
I'm also concerned about the person whose phone number gets called.
In the case of MS Word, that person is likely to be:
a) Some LAN administrator type, who likely doesn't know how to
program, and has no internal knowledge of what's going on in the many
layers involved.
b) Someone at a MSFT "Help Line," charging higher rates than the
Psychic Friends Network, and yet (based on the reports I've seen :-))
having roughly equally equivalent ability to resolve such problems.
The *right* answer would be:
c) One of the Word developers, who really understands what's going on,
but this is Not Going To Happen.
The newsgroups here don't include any MSFT-related ones, so it is
appropriate to establish a Linux equivalent...
Supposing the "document processing" system involves a similar number
of layers of components, and the user got them in predigested binary
form (e.g. - RPM/tarball-of-binaries), there may be no way of the user
connecting to someone that really understands the system.
This is decidedly true for proprietary document systems like
ApplixWare, Star Office, and WordPerfect.
It may be *effectively* true if there's not a programming-capable
person on the other end of that phone line.
--
Editing is a rewording activity.
-- Alan Perlis
[And EMACS a rewording editor. Ed.]
[EMAIL PROTECTED] <http://www.hex.net/~cbbrowne/sgml.html>
------------------------------
From: Adrian Hands <[EMAIL PROTECTED]>
Subject: Re: AOL Instant Messanger for UNIX
Date: Sat, 03 Jul 1999 01:29:32 -0400
Reply-To: [EMAIL PROTECTED]
Oh, so that's what "aim" stands for!
I was just at the gnome site and noticed something called "gaim".
http://www.marko.net/gaim/
brian wrote:
>
> is there an RPM of any aim linux client? i have tried like 3 kinds and
> have had install problems with them all. i run redhat6.0 and sometimes
> mandrake
>
> thanks a lot...
> brian
>
> Justin The Cynical wrote:
> >
> > On 10 Mar 1999 23:54:03 GMT, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote:
> > ->In his obvious haste, eric <[EMAIL PROTECTED]> babbled thusly:
> > ->: Hi,
> > ->
> > ->: I was wondering if AOL Instant messanger for unix will work with
> > ->: Linux, more specific, Slackware 3.6?
> > ->
> > ->Is it available as source code, so it can be compiled for Linux?
> >
> > Yes and no. :-) The AIM client for Linux/*NIX is written in
> > TCL/TK 8.0, which means it's a script of sorts. The client is viewable
> with
> > less, as it is just a plain text file/script. What this means is that
> no, you
> > don't have to compile anything, and yes, the download is an executable.
> I am
> > using it here on my home box, and it works nicely.
> >
> > Interesting, no? :-)
> >
> > --
> > "NT disk, meet Mr. Microwave."
> > David Parsons in comp.os.linux.advocacy (e-mail addy deleted for spam
> reasons)
> >
> > Justin The Cynical - [EMAIL PROTECTED]
>
> ------------------ Posted via SearchLinux ------------------
> http://www.searchlinux.com
------------------------------
From: [EMAIL PROTECTED] (Bill Unruh)
Subject: Re: Modem speed
Date: 3 Jul 1999 05:35:31 GMT
In <[EMAIL PROTECTED]> "Robert J. Schweikert"
<[EMAIL PROTECTED]> writes:
>I did set the Baude rate to 56600 in the script that executes the ppp command.
56600 is an illegal speed. Your pppd will use the default 9600 bd.
The legal speed is 57600
(Youcould also try 115200 once you get it working.)
------------------------------
From: [EMAIL PROTECTED] (Ken Williams)
Subject: Re: Visual programming language for linux?
Date: Sat, 03 Jul 1999 06:12:07 GMT
>H.Eitjes <[EMAIL PROTECTED]> wrote in message
>news:[EMAIL PROTECTED]...
>> Does there exist a Visual language for linux. Like there is visual C or
>> visual basic for MS-windows?
http://www.cs.uni-potsdam.de/~smeier/kdevelop/
------------------------------
From: "Ryan T. Rhea" <[EMAIL PROTECTED]>
Subject: moving an entire installation to a different partition
Date: Fri, 02 Jul 1999 22:19:54 -0400
I have a linux server which is entirely installed on one 450 MB
partition. I have a need to move the entire install to a second larger
partition on the same drive. There is a separate linux swap partition
that will not change.
Has anyone tried this? I examined the boot logs, and I don't see any
references to partitions other than the obvious ones to the root file
system and the swap space - I will change that in fstab and lilo.conf.
Did I miss anything? I did see a reference to kswapd - I assume this is
a daemon for kernel swap space, I can find no info on it.
Also, can I just use the 'cp' command, with perhaps '-dpr' for the
options (no link Deference, Preserve ownerships and permissions,
Recursive) to copy all the files from one partition to the other? Or
will this miss some files such as cp itself or any other open files?
Thanks in advance,
Ryan T. Rhea
please send a courtesy copy of any replies to [EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED] (Christopher Wong)
Subject: Re: Fragile file system
Date: 2 Jul 1999 16:26:32 GMT
Reply-To: [EMAIL PROTECTED]
On 2 Jul 1999 02:53:17 GMT, Paul Thompson <[EMAIL PROTECTED]> wrote:
>Did you check your syslog for IDE errors? You mentioned that you
>did not think that it was a disk problem but not if you actually looked
>for any errors in the log.
>
>Not being able to erase files as root points to an education problem.
>You you try deleting by inode or other ways around undeletable files?
>The ext2 debug utility might have been the way.
>
>Too bad there was not output presented of your attempts to delete the
>files.
>
>What was the syntax of the network backup command you were running?
>I have run network backups for years; back when I used NFS on older
>kernels I would occasionally hang, but never a crash like what you
>described and not recently when using rsh.
>
>In article <[EMAIL PROTECTED]>,
> [EMAIL PROTECTED] (Christopher Wong) writes:
>> Not too long ago, my Mandrake 5.3 system (kernel 2.0.36 based on Red Hat
>> 5.2) crashed. While an OS crash is not too unusual for me -- having used
>> Windows since its 3.0 days -- I am surprised by the apparent fragility
>> of the Linux file system. A Dos or OS/2 system would typically dust
>> itself and bounce right back up with a simple chkdsk or two. This crash,
>> however, wreaked havoc all over my ext2fs partition.
OK, here are more details to satisfy your curiousity. In response to
your questions: no, there were no IDE-related errors. I have a sample
session below showing my attempts to delete wierd files the "normal"
way. The backup command involves the following command over in.rshd (I
do not completely understand our network backup scheme ... but I believe
NFS is involved) among others.
gtar -c --atime -f - --ignore-fail --sparse -l --totals \
./tmp/xxx.tdir .
More gory details follow, but my point is that there does not appear to
be a hardware failure, and yet the thing blew up spectacularly. This led
to a system state that a nonexpert would find difficult to recover
from. I suppose it is encouraging to hear that my experience is rare,
and I hope not to see it again.
Chris
======================================================================
The following are my attempts to delete the files with wierd permissions
or modes. There are lots of these files littered all over my clobbered
ext2fs filesystem. Granted, there are ways to delete these things using
some low level fs debugging tool as you suggested. My question was more
on the lines of how someone accustomed to running "chkdsk" or "scandisk"
would deal with this sort of mess, especially if it is as extensive as
mine.
root@wchris /old/var/log: ls -l ju*
cr-sr-Sr-t 1 24941 29999 112, 117 May 11 2031 junkbuster
root@wchris /old/var/log: rmdir junkbuster
rmdir: junkbuster: Operation not permitted
root@wchris /old/var/log: rm junkbuster
rm: remove `junkbuster'? y
rm: cannot unlink `junkbuster': Operation not permitted
root@wchris /old/var/log: ls -l net*
br-sr-xr-- 1 28261 12140 82, 47 Nov 18 2031 netconf.log
-rw-r--r-- 1 root root 390 Apr 1 15:35 netconf.log.1
-rw-r--r-- 1 root root 0 Mar 1 04:02 netconf.log.2
-rw-r--r-- 1 root root 1809 Feb 24 16:47 netconf.log.3
-rw-r--r-- 1 root root 732 Feb 9 16:32 netconf.log.4
root@wchris /old/var/log: rm netconf.log
rm: remove `netconf.log'? y
rm: cannot unlink `netconf.log': Operation not permitted
I looked through the syslog as you suggested. There are no IDE-related
messages in /var/log/messages. While I no longer recall when Linux
actually blew up, there are two dates that might pertain to the
event. Perhaps they might prove more enlightening to you than to me. I
wrapped some lines to make them more readable. This is on a kernel
version 2.0.36.
May 18 05:46:30 wchris kernel: 991840
May 18 05:46:30 wchris kernel: EXT2-fs error (device 03:05):
ext2_readdir: directory #14606 contains a hole at offset 1818802176
May 18 05:46:30 wchris kernel: attempt to access beyond end of device
May 18 05:46:30 wchris kernel: 03:05: rw=0, want=795161139, limit=991840
May 18 05:46:30 wchris kernel: EXT2-fs error (device 03:05):
ext2_readdir: directory #14606 contains a hole at offset 1818803200
May 18 05:46:30 wchris kernel: attempt to access beyond end of device
(lots more of the above)
May 23 04:02:04 wchris PAM_pwdb[7069]: (su) session opened for user
nobody by (uid=99)
May 23 04:04:28 wchris PAM_pwdb[7069]: (su) session closed for user nobody
May 23 04:22:25 wchris kernel: general protection: fffc
May 23 04:22:25 wchris kernel: CPU: 0
May 23 04:22:25 wchris kernel: EIP: 0010:[general_protection+0/12]
May 23 04:22:25 wchris kernel: EFLAGS: 00010207
May 23 04:22:25 wchris kernel: eax: bfffe234 ebx: 08074dc8
ecx: 08083608 edx: 08080261
May 23 04:22:25 wchris kernel: esi: 0806f5c0 edi: bfffe548
ebp: bfffe158 esp: 02fa5178
May 23 04:22:25 wchris kernel: ds: 002b es: 002b fs: 002b gs: 002b
ss: 0018
May 23 04:22:25 wchris kernel: Corrupted stack page
May 23 04:22:25 wchris kernel: Process gawk (pid: 7139, process nr: 58,
stackpage=02fa9000)
May 23 04:22:25 wchris kernel: Stack: 0000fffc 0010ae28 00000010
00010207 0000fffc 0010ae28 00000010
00010207
May 23 04:22:25 wchris kernel: 0000fffc 0010ae28 00000010 00010207
0000fffc 0010ae28 00000010 00010207
May 23 04:22:25 wchris kernel: 0000fffc 0010ae28 00000010 00010207
0000fffc 0010ae28 00000010 00010207
May 23 04:22:25 wchris kernel: Call Trace: [general_protection+0/12]
[general_protection+0/12]
[general_protection+0/12]
[general_protection+0/12]
[general_protection+0/12]
[general_protection+0/12]
[general_protection+0/12]
May 23 04:22:25 wchris last message repeated 27 times
May 23 04:22:25 wchris kernel: [general_protection+0/12]
[general_protection+0/12]
May 23 04:22:25 wchris kernel: Code: 68 e4 b4 10 00 e9 ce fe ff ff 89 f6
68 5c b4 10 00 e9 c2 fe
May 23 04:22:25 wchris kernel: release: gawk kernel stack corruption. Aiee
------------------------------
From: [EMAIL PROTECTED] (Cameron L. Spitzer)
Subject: Re: 10GB disk and LILO - I tried EVERYTHING!
Date: 3 Jul 1999 06:24:12 GMT
In article <[EMAIL PROTECTED]>, Robert Heller wrote:
> Alessandro Magni <[EMAIL PROTECTED]>,
> In a message on Fri, 02 Jul 1999 15:27:22 +0200, wrote :
>
>AM> Sorry if you already read about it many times, but I tried almost
>AM> everything I saw on USENET about this topic, and failed...
>AM>
>AM> I got this WesternDigital 9GB disk.
>AM> Windows fdisk says it's only 8GB, but I know about the 1024 cylinder
>AM> problem, OK.
>AM> HOW can I put a boot partition inside the damned 1024cyl limit?
>AM> (By the way, diskdruid does not tell you where it creates its
>AM> partitions,
Disk druid is not up to this task. Complain to Red Hat.
Use cfdisk or fdisk.
>/dev/hda5 seems way to far up to be under the 1024 cylinder. What you
>want is something like:
>
>'primary partitions':
>/dev/hda1 2gig FAT16 or FAT32
>/dev/hda2 64meg Ext2 (Linux Native) / (root) Hopefully this will be under
> the 1024th cylinder
>/dev/hda3 64-128meg Swap
This would work too:
/dev/hda1 1.5 MB /u1 (for LILO and kernels)
/dev/hda2 4 GB C:
/dev/hda3 150 MB /
/dev/hda4 (extended)
/dev/hda5 120 MB swap
/dev/hda6 2.5 GB /u6 (for /usr)
/dev/hda7 2.5 GB /u7 (/home, /var)
Do your install in two stages. First stage, partition the
drive, exercise each partition with "badblocks -w", make
file systems, then create a /usr directory in /u6
and a symlink in / pointing to it. Example follows.
You're in the ramdisk shell, on the install floppy.
cfdisk # create partition table
mkdir /m3 /m6 /m1
badblocks -w /dev/hda1 1500
mke2fs /dev/hda1 1500
badblocks -w /dev/hda3 150000
mke2fs /dev/hda3 150000
badblocks -w /dev/hda6 2500000
mke2fs /dev/hda6 2500000
badblocks -w /dev/hda7 2500000
mke2fs /dev/hda7 2500000
badblocks -w /dev/hda5 128000
mkswap /dev/hda5 128000
mount -t ext2 /dev/hda6 /m6
mkdir /m6/usr
mount -t ext2 /dev/hda3 /m3
cd m3
ln -s u6/usr .
mount -t ext2 /dev/hda1 /m1
mkdir /m1/boot
ln -s u1/boot .
Reboot. Run the install. It will ask what to do with the existing
partitions. Activate the existing swap. Mount /dev/hda3 as /
and /dev/hda6 as /u6 (Not as /usr!)
Mound /dev/hda1 as /u1
Let the install program mount them. Go to a shell on the second
virtual console and verify that /usr is on /dev/hda6.
df # (prints a table of mounted partitions)
cd /target/usr # (or whereever Red Hat uses)
df .
Now let the install proceed.
This is how to install a small root and big usr in one pass,
without filling up any partitions and without creating the space
management problems you get when /usr is, itself, a mount point.
The badblocks -w operation isn't going to report any bad blocks
on a modern drive. What it's really doing is warming the drive up
to operating temperature before you write anything important on it,
and giving the drive electronics a chance to spare out any sectors that
have gone bad in shipping.
By using symlinks from root to the other traditional partitions,
your storage allocation remains flexible. You can move parts or all of
/usr to a second drive, for example.
It's too bad the distributions' install programs aren't smart enough
to do this stuff. People would have a lot less trouble.
Cameron
------------------------------
** 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.misc) 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-Misc Digest
******************************
