Files
oldlinux-files/docs/mail-archive/linux-admin/Volume2/digest149
2024-02-19 00:23:35 -05:00

597 lines
21 KiB
Plaintext

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: Thu, 6 Oct 94 04:13:23 EDT
Subject: Linux-Admin Digest #149
Linux-Admin Digest #149, Volume #2 Thu, 6 Oct 94 04:13:23 EDT
Contents:
Re: booting in single user mode? (CVL staff member Nate Sammons)
Has anybody gotten bootpd to work? (Thomas Russell Hoover)
Re: Recommendation: Partitioning Linux (Baba Buehler)
Re: Security hole - has noone noticed so far? (Stephen David Wray)
gcc help
RLOGIN security - more info! (Gregory Trubetskoy)
Re: Zmodem errors at 38400 w/16C550 (Jim M. Kam)
Secure File System Wanted ... (Thomas Russell Hoover)
Re: SCSI vs IDE (Drew Eckhardt)
Re: <Q> Can Linux Mount a Mac Floppy (Mark J. Dulcey)
SCSI vs IDE
Anyone have the Acct package working with 1.1.51 kernel (Don Carroll)
Re: [Q] HP JetDirect Support on Linux??? (Baba Buehler)
Re: booting in single user mode? (Claus-Dieter Bredl)
Re: XFree86-3.1 - Whoopee! (Alex R. Moon)
Re: XFree86-3.1 - Whoopee! (Rene COUGNENC)
Re: Linux as KingGod NFS Server to DOS Slaves (James F. Morris)
----------------------------------------------------------------------------
From: nate@matisse.VIS.ColoState.Edu (CVL staff member Nate Sammons)
Subject: Re: booting in single user mode?
Date: 5 Oct 1994 14:07:06 GMT
Ross boswell says in email:
>
>If you boot with LILO:
>Hold the Alt key as LILO is loading. LILO will then prompt for
>parameters. Give the name of your boot option, followed by "single"
>as in:
>Linux single
>and it will boot in single-user mode.
>
>If you boot with loadlin:
>Give loadlin the name of your kernel image followed by "single"
>as in:
>LOADLIN VMLINUZ.IMG SINGLE
>
>Good luck -- Ross
>--
>| Ross Boswell | Email : drb@chmeds.ac.nz |
>| Department of Pathology | FAX : +64 3 364 0525 |
>| Christchurch School of Medicine | Phone : +64 3 364 0590 |
>| NEW ZEALAND | Post : PO Box 4345, Christchurch |
>
Which works.
thanks,
-nate
--
Nate Sammons <nate@vis.colostate.edu>
System Administrator - CSU Computer Visualization Laboratory
------------------------------
From: thoover@infi.net (Thomas Russell Hoover)
Subject: Has anybody gotten bootpd to work?
Date: 5 Oct 1994 15:33:03 GMT
I'm trying to setup a boot server for about 5 pcs. I'm not having any luck.
I've read the HOWTO and Network Admin Guide - zilch. Can anybody point me in
the right direction?
Thanks,
TOm Hoover
------------------------------
From: baba@ph-meter.beckman.uiuc.edu (Baba Buehler)
Subject: Re: Recommendation: Partitioning Linux
Date: 6 Oct 94 03:26:57 GMT
Reply-To: Baba Z Buehler <baba@uiuc.edu>
spencer@montego.umcc.umich.edu (Spencer PriceNash) writes:
>>Currently I'm planning a layout something like this:
>>
>>Root: 35meg
>>Swap: 16meg
>>Usr: (the rest of the disk)
>That's a teensy tiny root partition.
yeah, but all the binaries and such go in /usr under that setup, not in root
like you have it. this is the df from my install, based on Slackware 2.0.1,
with most options installed (and lots added by me):
Filesystem 1024-blocks Used Available Capacity Mounted on
/dev/hdb1 10213 4468 5218 46% /
/dev/hdb3 178598 156374 13001 92% /usr
/dev/hdb5 19974 14086 4857 74% /var
/dev/hdb6 15095 20 14296 0% /tmp
/dev/hdb7 10213 1250 8436 13% /home
/dev/hdb8 233229 130497 90688 59% /usr/local
i find that its good to keep thing as seperated as possible, namely its easier
to upgrade (just don't touch /usr/local and keep a copy of /etc), and if one
filesystem gets corrupted, it doesnt ditch your whole install.
>Again, I think your choice of root partition size is by far too small.
like i said, its generally a bad idea to keep everything in root.
--
%>- Baba Z Buehler
%>- Beckman Institute Systems Services, Urbana Illinois
%>- WWW: http://www.beckman.uiuc.edu/groups/biss/people/baba/
%>- PGP Public Key available via WWW & public key servers
------------------------------
From: swra01@cs.aukuni.ac.nz (Stephen David Wray)
Crossposted-To: comp.mail.smail
Subject: Re: Security hole - has noone noticed so far?
Date: 06 Oct 1994 03:23:51 GMT
> /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?
------------------------------
From: s010dls@alpha.wright.edu ()
Subject: gcc help
Date: Wed, 5 Oct 1994 01:13:24 GMT
I want add a path to the list of paths automatically searched for
include files. I would also like to disable one of the warning
messages. How do I do this?
Thanks
------------------------------
From: grisha@cais.cais.com (Gregory Trubetskoy)
Subject: RLOGIN security - more info!
Date: 5 Oct 1994 18:22:34 GMT
This is regarding my earlier post about the possibilty
of making an .rhosts file in the bin directory.
This wasn't my idea - a friend of mine told me that he
can crack my linux box in three minutes. I gave him an account and
said "Go for it!". Three minutes later he became bin.
Unfortunately, I cannot find out how it was done...
Now, here is the .bash_history file. Can someone guess what
is in the xxx file? This user did not have the write rights to
/bin (as I thought earlier) , so there is some kind of a trick
in the xxx script...
The user's name has been replaced with "username".
The resulting .rhosts looked like this (I wonder what the zeros mean):
localhost username 0
b 0 0
The following is the TRUE .bash_history file with MY comments preceeded by #
less < /etc/inetd.conf
less < /etc/hosts.deny
finger root
cd /root
ls -la
cd /
ls -la
cd /var/tmp
ls -la
ftp
ls -la
vi
chmod u+x xxx
ls -la
ls -la
xxx #this is the misterious script
xxx /root/.rhosts username localhost
rlogin
rlogin -l root localhost # this didn't work!
finger bin
finger bin
ls -lad /etc
xxx /bin/.rhosts username localhost
rlogin -l bin localhost # but this works
rm xxx # i wish there was a way to undelete in linux
fg
rpcinfo -p
# the rest is irrelevant...
Any comments?
--
================================================================
Gregory Trubetskoy grisha@cais.com
================================================================
------------------------------
From: jimk@jimk.sys.hou.compaq.com (Jim M. Kam)
Subject: Re: Zmodem errors at 38400 w/16C550
Date: Mon, 3 Oct 1994 04:25:20 GMT
Dennis Heltzel (dheltzel@crl.com) wrote:
: I just installed a pair of internal BOCA V.FAST 28,800 bps modems into 2
: Linux systems. I have the line speed set to 38400 on both ends (stty
: confirms speed) and Zmodem works great with text files (3800 bps).
: However, when I try to transfer a compressed file, I get lots of errors
: and the transfer aborts. I know that the modem can't actually handle data
: that is already compressed at that high a speed, but shouldn't flow
: control prevent over running the chip's FIFO ? If I reduce the speed to
: 19,200, the compressed files transfer fine, but less than half the speed
: (1800 bps). I'd be really happy if the compressed files would transfer at
: a higher speed. Anyone have any ideas ?
: BTW: I have the same problem when using DOS/Telix on the receiving end.
: Dennis
:
I bet what's happening is that you're using xon/xoff instead of hardware
handshaking. Easy enough to fix. Do the following
setserial /dev/cua0 spd_vhi
(or whatever serial port you're using) This sets speed to 115Kbps. I know.
You have trouble getting it to work at 38.4Kbps. Bear with me.
stty crtscts -ixon -ixoff < /dev/cua0
This turns hardware handshaking on and xon/xoff off.
Now run kermit or some such program and set the modem to use a fixed DTE
rate, use hardware flow control and use V.42bis compression. On my
USRobotics modem the commands are as follows :
AT &B1&H1&K3&W
Now try it. Dollars to donuts it'll work alot better. You can make the
settings permanent by adding those lines to your /etc/rc.d/rc.local file. It
should work in rc.serial, but for some reason that doesn't seem to do it on
my box.
Cheers,
jimbo
#include <std_disclaimer.h>
------------------------------
From: thoover@infi.net (Thomas Russell Hoover)
Subject: Secure File System Wanted ...
Date: 5 Oct 1994 01:16:05 GMT
Does anybody know where I can get a "secure" filesystem? I read about CFS
1.1 and it looks interesting - Does it port easily??
Thanks in advance,
Tom Hoover
------------------------------
From: drew@frisbee.cs.Colorado.EDU (Drew Eckhardt)
Subject: Re: SCSI vs IDE
Date: 5 Oct 1994 01:26:17 GMT
In article <36ss1f$b5v@newsflash.concordia.ca>,
Iain J. Bryson <iain@ece.concordia.ca> wrote:
>Hi. I am interested in hearing people advocating
>which is better, IDE or SCSI.
If you only want one small drive, it doesn't matter.
If you want multiple drives, large/fast drives, etc,
SCSI is the unquestionable winner.
>One big advantage
>for SCSI would be more disks and CD-ROMS not
>taking up a slot... But it that worth the
>extra cost of a (good?) controller?
That's something you'll have to decide for yourself. Depending on your
immediate and future requirements, the cost may be negligible (NCR boards run
about $70, PCI only, closeout Ultrastor 34F boards about $90), or in the
neighborhood of $200-$300.
>How about speed?
1. With a reasonable host adapter and driver software, you can sustain
the sum of the transfer rates of your drives rather than the
average as with IDE (assuming you haven't run into a bottleneck
somewhere else - an i486-66 seems to utilize about 10% of the
available CPU cycles per megabyte per second of transfer rate,
fast SCSI-II is only good to 8-10M/sec, ISA only works to about
5M/sec).
2. With a reasonable host adapter, driver, and drives, you can
sustain about 75% of head rate on reading (through
the filesystem), and close to 100% writing (again,
through the filesystem), with the above mentioned limits
applying.
With modern multi-gigabyte drives, I've seen this work
out to be to be 3+M/sec and 4+M/sec through the filesystem.
Smaller drives will be in the 1-2M/sec range.
3. CPU utilization on my system works out to be about 1/3
as much to run roughly similar SCSI drives on a busmastering
SCSI controller (NCR53c810) versus IDE drives.
--
Since our leaders won't respect The Constitution, the highest law of our
country, you can't expect them to obey lesser laws of any country.
Boycott the United States until this changes.
------------------------------
From: mdulcey@pryder.pn.com (Mark J. Dulcey)
Subject: Re: <Q> Can Linux Mount a Mac Floppy
Date: Wed, 05 Oct 94 22:46:56 GMT
In article <1994Oct5.153300@hammer.westboro-ma.peritus.com> cummings@hammer.westboro-ma.peritus.com writes:
>
>Is there even a project ongoing to support it? I thought that at least at some
>level Apple used some variable speed drives and varied the speed of the drive
>depending on which track was being written to. This allowed them to write
>more sectors onto some tracks than others resulting in more storage. The
>floppy drives found in most "IBM-compatible" PCs are fixed speed drives
>making it near impossible to read Apple formatted floppies (never mind
>write them).
400K and 800K Mac floppies used variable-speed recording, with more
sectors on the outer tracks. It's actually possible to read and write
such disks without a variable-speed disk drive -- you use a variable
speed controller instead. The Copy II Deluxe Option Board that Central
Point Software used to sell was capable of this trick, as was the
Spectre GCR add-on for the Atari ST.
1.44MB Mac floppies use IBM-style fixed-speed MFM recording, but still
use a Mac-style HFS directory structure, so you can't read them on a PC
without special software. PC-compatible 720K and 1.44MB floppies
written by a Mac use MFM recording and a standard DOS directory and FAT.
Rumor has it that future Mac models will drop the variable speed disk
drives, so they won't be able to read old Mac floppies. They're also
going to drop the auto-eject mechanism. (Basically, they're planning
to use standard IBM-style floppy drives; because of the economies of
scale, those drives are significantly less expensive.)
If there is any Linux software to read Macintosh disks, I would expect
it to only be able to read the 1.44MB ones, which can be easily read
by a standard IBM-compatible floppy controller.
------------------------------
From: s010dls@alpha.wright.edu ()
Subject: SCSI vs IDE
Date: Wed, 5 Oct 1994 01:45:53 GMT
> Hi. I am interested in hearing people advocating
> which is better, IDE or SCSI. One big advantage
> for SCSI would be more disks and CD-ROMS not
> taking up a slot... But it that worth the
> extra cost of a (good?) controller? How about
> speed?
If all you're going to do is add another drive, stay IDE. There really
isn't any speed difference between them anymore. You'll hear people
scream that when they updated to SCSI, things got faster. But, they're
probably comparing an older IDE drive with a new VLB SCSI card. I
recently added a SCSI drive. I now have a 500MB IDE drive, and a 1GB
SCSI-2 drive. I have a SCSI-2 ISA controller, and enhanced IDE
controller. The SCSI drive has a faster spin rate, lower access time,
higher transfer rate (it's newer), but the IDE drive outperforms the
SCSI drive by about 150K/sec.
However, if you want to have more than 2 hard drives, or a CD-ROM, go
SCSI. Keep in mind, that most SCSI cards have BIOS on them that scan
the drive chain everytime you reboot (which is necessary for SCSI HDs).
It usually takes 5-30 seconds for the card to do this, and you're
sitting there waiting the whole time. If you're using DOS, you'll need
to load several device drivers to ASPI, int 13, CAM, etc support eating
up conventional memory.
I bought SCSI for the CD-ROM, and disabled the bios on the card. I
purchased a SCSI hd because I thought they were faster for multitasking.
This hasn't proved to be so.
One more thing, if you do SCSI and end up having an IDE drive and a SCSI
drive, don't forget the LEDs. You probably have one LED on the case for
a hard drive and it's connected to the IDE card. The SCSI card will
have a connector too. You shouldn't connect them together (it'll burn
up the LED in time). I had to drill a new hole in the case, buy a 3V LED
from radio shack, beg for a connector from a computer store, and install
a new LED myself.
------------------------------
From: don@ds9.us.dell.com (Don Carroll )
Subject: Anyone have the Acct package working with 1.1.51 kernel
Date: Tue, 4 Oct 1994 21:08:27
the subject says it all
------------------------------
From: baba@ph-meter.beckman.uiuc.edu (Baba Buehler)
Subject: Re: [Q] HP JetDirect Support on Linux???
Date: 6 Oct 94 03:50:36 GMT
Reply-To: Baba Z Buehler <baba@uiuc.edu>
schmittl@cc.memphis.edu (Larry Schmitt) writes:
>Hi All - We are considering placing our HP Laser Printers directly on the
>network using the HP Jet Direct interface. Has anyone been able to configure
>one of these printers in a Linux environment. The perferred method is to use a
>bootp server. Any help will be appreciated greatly.
the plp replacement printing system (on ftp.iona.ie) is a reverse-engineered
version of Berkeley's system, with many enhancements, including support of
printers with their own TCP/IP interfaces.
it compiles and runs under linux.
linux bootp/tftp should be sufficient to boot these printers. the only problem
you will run into is that HP only distributes binaries for its printer
utlities, so their "JetAdmin" config program will not work under Linux (they
only give out Sun & HP versions)... however with plp, you can still communicate
with the printers, so you'd just have to send the postscript commands yourself
to configure it remotely.
--
%>- Baba Z Buehler
%>- Beckman Institute Systems Services, Urbana Illinois
%>- WWW: http://www.beckman.uiuc.edu/groups/biss/people/baba/
%>- PGP Public Key available via WWW & public key servers
------------------------------
From: cdb@tph116.fkp.physik.th-darmstadt.de (Claus-Dieter Bredl)
Subject: Re: booting in single user mode?
Date: 5 Oct 1994 06:33:49 GMT
Matt Beal (publius@eng.umd.edu) wrote:
: CVL staff member Nate Sammons (nate@seurat.VIS.ColoState.Edu) wrote:
: : Can you issue a "boot -s" type command to get into single user mode?
: : The problem is, my machine goes to init 6 (xdm) but X is not
: : configured properly, so it sits there and flicks in and out of sync
: : with the monitor, and I cannot do anything. I need to be able to go in
: : and change the initial init state of the machine, and the easiest
: : way I can think of (with other UN*Xs) is to boot it into single user mode.
: : Any suggestions? (not in the FAQ that I could find)
: To change the initial startup mode, edit /etc/inittab. Mine has:
[...]
You might have to boot Linux from floppy (e.g. an installation disk)
and mount the hard disk partition where /etc/inittab resides. Thus,
you'll have to edit (e.g.) /mnt/etc/inittab ;-)
If you use LILO, there should be a way to pass parameters on to the
Linux kernel, one of them for single user startup (I never needed it
so far, so RT(LILO)M yourself:)
CDB
------------------------------
From: moon@symphony.cc.purdue.edu (Alex R. Moon)
Subject: Re: XFree86-3.1 - Whoopee!
Date: 5 Oct 1994 06:55:01 GMT
In article <36qbmd$7ra@interport.net>,
Carlos Dominguez <carlos@interport.net> wrote:
>Michael_Nelson (nelson@seahunt.imat.com) wrote:
>
>: I think I'll wait awhile before attempting 3.1 again... :-(
>
>I think I'll await for it to become part of a future slackware
>distribution .
>
>I've worked too hard to get X up and running, and to futz with fvwm and
>its neat-o window sound event manager, to start from scratch all over again.
>
Don't wait because of that! I spent this past weekend customizing fvwm
"just right" (especially the neat-o window sound manager which I 'enhanced'
a bit), then v3.1 came out Monday, and I just installed it in the new directory
and am still running fvwm off the old directories with the old shared libs.
I'll recompile it to use X11R6 one of these days, but not today. :)
--Alex
moon@symphony.cc.purdue.edu
------------------------------
From: rene@renux.frmug.fr.net (Rene COUGNENC)
Subject: Re: XFree86-3.1 - Whoopee!
Date: 5 Oct 1994 01:01:18 GMT
Reply-To: cougnenc@hsc.fr.net (Rene COUGNENC)
Ce brave Dr. Raimund K. Ege ecrit:
> BTW: XFree86 3.1 and XFree 2.1 co-exist without trouble
> on my machine. Here is a snapshot of the ld situation:
Well, not on mine :-(
Of course, everything still works, but on my 8Mb RAM machine,
the old clients loading the old libs, the new ones using the
new incompatible libs, the system is slow, there is not enough
RAM.
Since I cant't spend months recompiling all the programs installed,
I'll keep XFree 2.1.1 and delete X11R6.
If, one day, I buy a new system (not planned before 2 or 3 years),
then it will run XFREE 3.1 or later; but for now I am happy with
XFree 2.1.1, so...
--
linux linux linux linux -[ cougnenc@renux.frmug.fr.net ]- linux linux linux
------------------------------
From: jfmorris@netcom.com (James F. Morris)
Subject: Re: Linux as KingGod NFS Server to DOS Slaves
Date: Wed, 5 Oct 1994 03:56:30 GMT
In article <Cwwwsv.Gzu@gcs.com>, Mark Bolzern <mark@gcs.com> wrote:
>In article <COSC176T.94Sep25013323@menudo.uh.edu>,
>
>Very happy with Century Software's TinyTerm Plus/NFS. Not too expensive
>and TCP/IP stack is from wollongong.... uses only 50K of my low ram...
How much is "not too expensive"? And do you have a telephone number for
Century Software? Where can I get more info?
Thanks!
--
/--------------------------------------------------\
| Jim Morris | Internet: jfmorris@netcom.com |
| | CompuServe: 73670,762 |
\--------------------------------------------------/
------------------------------
** 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
******************************