From:     Digestifier <Linux-Admin-Request@senator-bedfellow.mit.edu>
To:       Linux-Admin@senator-bedfellow.mit.edu
Reply-To: Linux-Admin@senator-bedfellow.mit.edu
Date:     Mon, 17 Oct 94 15:14:20 EDT
Subject:  Linux-Admin Digest #209

Linux-Admin Digest #209, Volume #2               Mon, 17 Oct 94 15:14:20 EDT

Contents:
  Re: [Q] unerase? undelete? (Jerry Davis)
  Re: XFree86 and fonts (Al Sutton)
  Re: ftp sites for linux (Al Sutton)
  xdm and xconsole (Daniel M. Coleman)
  lilo and big drives (Chuck Mattern)
  Re: Security hole - has noone noticed so far? (Joerg Mertin)
  Re: Extreme delays telnetting into linux box (Alan Cox)
  Re: xdm and xconsole (Gerrit Nieuwenhuizen)
  Re: xdm and xconsole (Gerrit Nieuwenhuizen)
  Re: xdm and xconsole (Byron A Jeff)
  Re: shadow-332: -f bug present (Wolfgang Jung)
  Re: Booting different configurations ? (Werner Almesberger)
  Re: SCSI vs IDE (ralph@dci.rubicon.org)
  Re: lilo and big drives (Werner Almesberger)
  [** URGENT **] e2fsck/mount/partition error (Joerg Fries)

----------------------------------------------------------------------------

From: jfdecd@earth.execpc.com (Jerry Davis)
Subject: Re: [Q] unerase? undelete?
Date: 17 Oct 1994 14:12:01 GMT


: Shouldn't there be a way to write one, though?  I'm no kernel hacker, but I
: just have this intuition that if MeSsy-Doz can write a pretty effective
: undelete, linux should be able to too.

I remember once, seeing an alias for rm that pointed to a script that put 
all the rm'd files into a separate subdirectory.  When you accidentally 
rm'd a file (which has happened to ALL of us) that you didn't want 
deleted then you just went to this subdirectory to get it back again.

Then again I guess you could alias rm to 'mv \!^ ~/deleted and see if 
that works.  When you REALLY want to delete it just go into ~/deleted and 
do a /usr/bin/rm filename on it.  I don't have my C shell book with me so 
you'll have to make sure of the syntax.

Hope this helps in the FUTURE.

======================================================================
| Jerry & Ellen Davis  | If work were like my hobbies, then I would  |
|                      | want to WORK all the time!                  |
======================================================================




------------------------------

Subject: Re: XFree86 and fonts
From: sutton@dcs.kcl.ac.uk (Al Sutton)
Date: 17 Oct 94 12:36:40 GMT

Michael Esler (mike@myhost.subdomain.domain) wrote:
:    I just put XFree86 3.1 on my box (a Gateway P-5 90 w/ 32megs and
: kernel 1.1.51) and I can't seem to get X to run. I installed the entire
: distibution but I get the following errors:

:    failed to set default font path: '/usr/X11R6/lib/X11/fonts'

:    Fatal Server error:
:    could not open default font 'fixed'
:    because FontPath is invalid

:    The font path it cannot find *is* present. The distibution created it
: and I put in my Xconfig. I also know I do not have the font 'fixed'. It
: was not in any of the tar balls I downloaded. 
:    How can I resolve this problem. Thanks in advance.

: Mike
: mike@gauss.dorm.virginia.edu


I expect a few people are having trouble with this, it threw me for a while
as well, heres two steps that should cure your problem.

1. Check the directory exsists.

2. go to the directory and type mkfontdir

everything should now work fine.

Al.





------------------------------

Subject: Re: ftp sites for linux
From: sutton@dcs.kcl.ac.uk (Al Sutton)
Date: 17 Oct 94 12:45:36 GMT

Chris Sorge (crsorge@sgcpu1.sdrc.com) wrote:
: Hi all,

:      This is my first time posting in this group, and would really appreciate
: help you could give me as far as sites go to ftp linux.  All responses
: are greatly appreciated.  Thanks in advance.

: Chris


Try using src.doc.ic.ac.uk, in the packages/linux directory they have
mirrors of all of the major Linux sites.

Al.

------------------------------

From: dcoleman@mail.utexas.edu (Daniel M. Coleman)
Subject: xdm and xconsole
Date: 17 Oct 1994 00:11:31 GMT

How does one get xdm to automatically launch an xconsole?  I have various
important system messages written to console, and I would like to see them
both before logging in with xdm and after.  I noticed the config file for xdm
has some stuff about xconsole in it, but I haven't figured out how to get it
to load it automatically.

Thanks for any help,

Dan

-- 
Daniel Matthew Coleman             | Internet: dcoleman@mail.utexas.edu
===================================+   DECnet: UTXVMS::DCOLEMAN
The University of Texas at Austin  |      IRC: Shiner
Electrical/Computer Engineering    |           "Sure thing, Giant Beer!"

------------------------------

From: cmattern@ronin.mindspring.com (Chuck Mattern)
Crossposted-To: comp.os.linux.help
Subject: lilo and big drives
Date: 16 Oct 1994 22:11:48 -0400
Reply-To: cmattern@mindspring.com

I am trying to install lilo on a Maxtor P17-S (1.5 gig formatted) drive and 
every time I try I get an error indicating that the drive has more than 1024 
cylinders.  Is there anything I can do to work around this short of booting 
from a floppy?

Config:
Adaptec 1542b
Maxtor P17-S
        3 partitions:
        /dev/sda1: dos 75 megs
        /dev/sda2: Linux 1.4 gigs
        /dev/sda3: Linux swap 25 megs
AMI 486/25 MB
Pioneer DRM-600
16M RAM

I get the same error regardless of whether I try to write lilo to the MBR, the
root partition or to floppy.
-- 
======================================================================
|Chuck Mattern              | "Not failure, but low aim, is crime."  |
|cmattern@mindspring.com    | -James Russell Lowell-                 |
======================================================================

------------------------------

Reply-To: smurphy@stardust.bln.sub.org
From: smurphy@stardust.bln.sub.org (Joerg Mertin)
Subject: Re: Security hole - has noone noticed so far?
Date: 08 Oct 1994 11:05:05 +0000
Crossposted-To: comp.mail.smail

In comp.os.linux.admin Fred Condo (fred@lightside.com) wrote:
: In article <SWRA01.94Oct6162351@cs19.cs.aukuni.ac.nz>,
: swra01@cs.aukuni.ac.nz (Stephen David  Wray) wrote:

: > > /usr/lib/sendmail is a symbolic link to /usr/bin/smail.
: > > 
: > > try
: > > 
: > > /usr/lib/sendmail -d -D/etc/nologin noone@empty.space
: > > 
: > > as a normal user and have fun explaining it to your sysadmin. I was
: > > awed when I found out...
: > 
: > 
: > Umm... I have a recentish slackware distribution, and just tried this
: > out -- nothing seems to have happened.
: > 
: > It just sits there, doing nothing...
: > 
: > What is it supposed to do that is so bad?

: It sits there waiting for standard input to close. Press ^D. Then you will
: find that /etc/nologin has been created and can have been created by any
: user. The patch recently posted makes this file creation happen under the
: UID of the calling user rather than as root.
Not ojly this file has been created, but it has rights for everyone to
modify it. So just think about what's gonna happen if someone make:

smail or /usr/lib/sendmail -d -D/etc/passwd bla@bla
then smail -d -D/etc/shadow bla@bla

Then, all he has to do is to copy some prepared files to these
location, set the rights and here it goes, he's root...
Fine thing, no  ???
Fortunatly, it doesnot work on my system anymore :-) Someone posted a
patch to fix it and it worked :-)

--
Solong & Happy Hacking
************************************************************************
*   Joerg Mertin          :   smurphy@stardust.bln.sub.org (Home)      *  
* in Berlin Spandau at    :   jorgagif@w203zrz.zrz.tu-berlin.de        *
* Stardust's Linux System :   Data, Fax & Voice 49 30 3615569          *
************************************************************************
`Fatal Error: Found [MS-Windows] System -> Terminating Virus...'

------------------------------

From: iialan@iifeak.swan.ac.uk (Alan Cox)
Subject: Re: Extreme delays telnetting into linux box
Date: Mon, 17 Oct 1994 13:18:36 GMT

In article <37jjbp$fhf@library.erc.clarkson.edu> komarimf@craft.camp.clarkson.edu (Mark 'Enry' Komarinski) writes:
>We have a similar login problem, especially when connecting to a MUSH
>port.  The connection from a remote host can (sometimes) sit there forever.
>If, however, from the machine I connect to that port (telnet localhost 7567)
>the connection from remote becomes instantly connected.  We were at first
>thinking this is a problem with our code, as regular telnet appears to
>work okay(who knows where that lag comes from? :).  But these
>problems  may be related.  Running 1.1.49 on a Slackware setup.

It's a combination of two problems.

1. Mush is buggy - it shouldn't use select() as a reliable indication that
an accept() will work - ie it should have a non blocking master socket for
the accept()

2. BSD happens to work so that select() is reliable in this case. 

I've fixed the current Linux kernel code to emulate BSD more so that these
numerous broken programs work happily.

Alan
-- 
  ..-----------,,----------------------------,,----------------------------,,
 // Alan Cox  //  iialan@www.linux.org.uk   //  GW4PTS@GB7SWN.#45.GBR.EU  //
 ``----------'`----------------------------'`----------------------------''

------------------------------

From: nieuwhzn@dxgsia.cern.ch (Gerrit Nieuwenhuizen)
Subject: Re: xdm and xconsole
Date: Mon, 17 Oct 1994 08:44:09 GMT

In <37sffj$1a2@geraldo.cc.utexas.edu> dcoleman@mail.utexas.edu (Daniel M. Coleman) writes:

>How does one get xdm to automatically launch an xconsole?  I have various
>important system messages written to console, and I would like to see them
>both before logging in with xdm and after.  I noticed the config file for xdm
>has some stuff about xconsole in it, but I haven't figured out how to get it
>to load it automatically.

>Thanks for any help,

>Dan

>-- 
>Daniel Matthew Coleman            | Internet: dcoleman@mail.utexas.edu
>-----------------------------------+   DECnet: UTXVMS::DCOLEMAN
>The University of Texas at Austin  |     IRC: Shiner
>Electrical/Computer Engineering           |           "Sure thing, Giant Beer!"


Try the file Xsetup_0 in /usr/lib/X11. With the Slackware 1.0.2 distr.
this contained a commemted out line for Xconsole.


                                Gerrit J. van Nieuwenhuizen
                                CERN-PPE-IO
                                Building 595, R-004
                                CH-1211  Geneve 23
                                Switzerland
                                nieuwhzn@dxgsib.cern.ch
                                  (or NIEUWHZN@VXWA80.CERN.CH)
                                tel.: +41 22 767 4740
                                fax : +41 22 782 4897



------------------------------

From: nieuwhzn@dxgsia.cern.ch (Gerrit Nieuwenhuizen)
Subject: Re: xdm and xconsole
Date: Mon, 17 Oct 1994 08:46:32 GMT

Ah, one thing I forgot to ask.
When I do a shutdown of a system running XDM then X dies,
the screen goes into normal ascii mode, but I don't get the 
system messages (like unmounting filesystems) anymore.
Why is that? I like to see that the system properly shuts down
and like to know when it is halted.


                                Gerrit J. van Nieuwenhuizen
                                CERN-PPE-IO
                                Building 595, R-004
                                CH-1211  Geneve 23
                                Switzerland
                                nieuwhzn@dxgsib.cern.ch
                                  (or NIEUWHZN@VXWA80.CERN.CH)
                                tel.: +41 22 767 4740
                                fax : +41 22 782 4897



------------------------------

From: byron@gemini.cc.gatech.edu (Byron A Jeff)
Subject: Re: xdm and xconsole
Date: 17 Oct 1994 12:59:59 GMT

In article <37sffj$1a2@geraldo.cc.utexas.edu>,
Daniel M. Coleman <dcoleman@mail.utexas.edu> wrote:
-How does one get xdm to automatically launch an xconsole?  I have various
-important system messages written to console, and I would like to see them
-both before logging in with xdm and after.  I noticed the config file for xdm
-has some stuff about xconsole in it, but I haven't figured out how to get it
-to load it automatically.

Make sure it's in your DisplayManager System Path.

BAJ
-- 
Another random extraction from the mental bit stream of...
Byron A. Jeff - PhD student operating in parallel - And Using Linux!
Georgia Tech, Atlanta GA 30332   Internet: byron@cc.gatech.edu

------------------------------

From: woju@keep.in-berlin.de (Wolfgang Jung)
Subject: Re: shadow-332: -f bug present
Date: 17 Oct 1994 01:57:36 +0100

Chris Haos (chris@ideal.com) wrote:
: Harald Milz (hm@ix.de) wrote:
: : I installed the shadow-3.3.2 suite a couple of minute ago and found
: : that the old -f bug is again in there. What's wrong?

: : -- 
: : Gordon's first law:
: :     If a research project is not worth doing, it is not worth doing
: : well.
: : -- 
: : Harald Milz (hm@ix.de)      WWW: http://www.ix.de/editors/hm.html
: : iX Multiuser Multitasking Magazine  phone +49 (511) 53 52-377
: : Helstorfer Str. 7, D-30625 Hannover fax   +49 (511) 53 52-378
: : Opinions stated herein are my own, not necessarily my employer's.

: Nothing is wrong.  Shadow simply hasnt been patched yet.  A quick hack is 
: to comment out the "case 'f'" in lmain.c.  This will disable -f.  I am 
: not sure what the correct patch to login was.  Does anyone have the other 
: login.c patch?  I will make up a new diff file if anyone does.  For now 
: you are best to just remove that "feature" of login to prevent 
: unauthorized access.

I once fixed this not disabling the -f feature...
Its a bug (not really a bug) which is introduced by a feature of
getopt. with checking optind in conjunktion with -f you can rip the
login from this REALLY bad security problem.
main raeson is: getting a program which calls login with -f and then 
using -froot as a Login name makes login think its preauthorized, if
you check optind you will see.

I am using a fixed version, and can upload it to some ftpsite..
(There are other Bugs with it..)

-- 
===============================================================================
| Gruss                 |ISO-8859-1 Mail: woju@keep.in-berlin.de              |
|       Wolfgang        |                 woju@keep.bln.sub.org               |
|       Jung            |                 wong@cs.tu-berlin.de                |
|                       |             wojuacac@w250zrz.zrz.tu-berlin.de       |
===============================================================================

------------------------------

From: almesber@nessie.cs.id.ethz.ch (Werner Almesberger)
Subject: Re: Booting different configurations ?
Date: 17 Oct 1994 14:28:08 GMT

In article <37m40h$17n@sun2.ruf.uni-freiburg> ps@kis.uni-freiburg.de writes:
> I suppose this to be done in one of the /etc/rc.d scripts (local or
> inet).

Yup, /etc/rc* and /etc/inittab are the easiest places.

> BUT, how do I pass some information to this scripts ? Using different
> kernel versions and using uname isn't what I would like to do.

Maybe the environment variable BOOT_IMAGE is what you want to use. 
(You can see it with
tr '\000' '\012' </proc/1/environ; echo
note that login normally removes this variable, so you won't see it
from an ordinary login shell.)

Alternatively, you could also create your own variable and set it with
MYCONFIG=foobar  on the boot command line or in /etc/lilo.conf with
APPEND.

- Werner
-- 
   _________________________________________________________________________
  / Werner Almesberger, sending this from    almesber@nessie.cs.id.ethz.ch /
 /______________________..._but_now_at_home_at________almesber@di.epfl.ch_/

------------------------------

From: ralph@dci.rubicon.org
Subject: Re: SCSI vs IDE
Date: 14 Oct 1994 05:09:09 GMT

In article <37h4a8$b3p@bmerhc5e.bnr.ca> mlord@bnr.ca (Mark Lord) writes:
>In article <DMW.94Oct12084235@prism1.prism1.com> dmw@prism1.prism1.com writes:
><
><      No, even on systems with one drive SCSI is a better choice than IDE
><unless you are running a single-tasking OS like DOS. It is definately
><possible to get transfer speeds out of IDE that are close to SCSI (but not
><Fast & Wide SCSI). But you are paying for the speed with CPU busy waiting.
>
>Nope.  Just plain false.  No measurable "busy waiting" is used for IDE drives
>under linux.  Read/Write operations are interrupt driven, not polled.
>Note that the effect of this is unnoticeable under most single-user situations,
>since writes are deferred by the buffer-cache, and reads usually imply that
>the user is sitting there waiting for the data anyhow.  With lots of processes
>running and performing I/O, the general (unproven) opinion is that SCSI really
>ought to perform better.  I agree.
>-- 
>mlord@bnr.ca   Mark Lord       BNR Ottawa,Canada       613-763-7482

I think the poster may have been assuming a DMA controller with SCSI.
I agree that IDE can give SCSI a run for its money.
I noticed going from IDE to SCSI (both on ISA bus, and with a FD1680
controller) that lilo takes noticably longer to load the kernel than
it did with an IDE drive (IDE was a Quantum LPS120AT, SCSI is a Micropolis
660MB 5.25")
I believe this is due to the higher latencies of SCSI commands vs IDE, and
lilo is probably loading one sector at a time.
I expect to see the benefit of SCSI once I set up my 2nd SCSI disk for swap;
SCSI allows for more eficient use of multiple drives at the same time.
-Ralph

------------------------------

From: almesber@nessie.cs.id.ethz.ch (Werner Almesberger)
Crossposted-To: comp.os.linux.help
Subject: Re: lilo and big drives
Date: 17 Oct 1994 14:43:31 GMT

In article <37smh4$5rd@ronin.mindspring.com> cmattern@mindspring.com writes:
> I am trying to install lilo on a Maxtor P17-S (1.5 gig formatted) drive and 
> every time I try I get an error indicating that the drive has more than 1024 
> cylinders.  Is there anything I can do to work around this short of booting 
> from a floppy?

Yes, create two partitions for Linux instead of one. Some disk controllers
are able to map the geometry ("address translation") so that the whole
disk appears to have 1024 or less sectors, but I don't think the 1542B can
do this. (At least mine doesn't.)

> Maxtor P17-S
>       3 partitions:
>       /dev/sda1: dos 75 megs
>       /dev/sda2: Linux 1.4 gigs
>       /dev/sda3: Linux swap 25 megs

Either your kernel or one of the files in /boot is probably located beyond
the 1024th cylinder. If you make two partitions, one for /, the other for,
say, /home, where the first partition is entirely within the first 1024
cylinders, you're safe. Also, having two smaller partitions instead of one
big partition tends to simplify system maintenance.

> I get the same error regardless of whether I try to write lilo to the MBR,
> the root partition or to floppy.

Yes, that's because it's not the boot sector, but one of the other files
LILO tries to map.

- Werner
-- 
   _________________________________________________________________________
  / Werner Almesberger, sending this from    almesber@nessie.cs.id.ethz.ch /
 /______________________..._but_now_at_home_at________almesber@di.epfl.ch_/

------------------------------

From: fries@iti.informatik.th-darmstadt.de (Joerg Fries)
Crossposted-To: comp.os.linux.help,comp.os.linux.misc
Subject: [** URGENT **] e2fsck/mount/partition error
Date: 17 Oct 1994 15:03:34 GMT

This is an urgent call for help, because I have probs to repair my
var-partition! (Yes, there is all my configuration ... >>:o

My system broke down (HD hangs, X down) and I had to reboot. After
that, e2fsck finds an error in my partition /dev/hdb1 (using e2fsck
and fs version 0.5a):

"Block bitmap 0 for group 0 not in group"

All other partitions were mounted correctly. When I tried to fix the
problem with "e2fsck -v -r -b 8193 /dev/hdb1" and then mount the
partition, (other values like 16385 didn't worked either...)

"EXT2-fs error (device 3/65): ext2_check_descriptors:
 Block bitmap 0 for group 0 not in group (block 0)

 EXT2-fs: group descriptors corrupted!"

appears. When I make fdisk for this HD, verifying tells me

"Warning: bad start-of-data in partition 1
 271 unallocated sectors"

WHAT CAN I DO? Is there a new e2fsck which could repair that fault?
What kind of fault is that (fs,partition,HD)?

Any help via email would be greatly appreciated..

Thanx in advance, Joerg

PS. Please excuse the crosspostings but I'm in great trouble because I
can't proceed in my thesis.
--

==============================================================================
Joerg Fries
Department of Computer Science
Technical University of Darmstadt, Germany

email: fries@iti.informatik.th-darmstadt.de
==============================================================================

------------------------------


** 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: Linux-Admin-Request@NEWS-DIGESTS.MIT.EDU

You can send mail to the entire list (and comp.os.linux.admin) via:

    Internet: Linux-Admin@NEWS-DIGESTS.MIT.EDU

Linux may be obtained via one of these FTP sites:
    nic.funet.fi				pub/OS/Linux
    tsx-11.mit.edu				pub/linux
    sunsite.unc.edu				pub/Linux

End of Linux-Admin Digest
******************************
