Linux-Misc Digest #727, Volume #24 Tue, 6 Jun 00 09:13:05 EDT
Contents:
Frage (Klaus Hirsch)
Re: Many questions and much dissatisfaction ("Ron Sinclair")
Re: Installing Netscape 6 (muzh)
News (Thomas Heil \\ WIS)
crontab: evironment not evaluated (peter pilsl)
Re: Installing Netscape 6 (=?iso-8859-1?Q?Beno=EEt?= Smith)
Re: How to set safe screen more than 7 minutes (David Efflandt)
Re: FTP in GNOME file manager? [gmc 4.5.50, Helix GNOME 1.2.1] (ray)
DELL's Linux price is HIGHER than Win98 ([EMAIL PROTECTED])
Re: partition lost (Svend Olaf Mikkelsen)
Re: partition lost (Svend Olaf Mikkelsen)
'mc' crashes after using FTP (Mihaly Gyulai)
How to create device on diskette? (Sverre Torjussen)
Re: how to start VNC during boot? (David Efflandt)
Re: partition lost (Svend Olaf Mikkelsen)
Seaching for Matlab for linux ... somebody has it.. (Jean-Philippe Desbiems)
WANTED Epson Printer Driver (=?iso-8859-1?Q?Beno=EEt?= Smith)
Re: samba troubles (Geoff Sullivan)
Re: Mandrake 7 setup question (David Efflandt)
loading module on red-hat 6.2 (Guy-Armand Kamendje)
Hideous X Fonts (Neurocrat)
Re: Where can i get XBF-neomagic-glibc-1.0.0-1.i386.rpm ? (Stephen Cornell)
Re: Serious fragmentation under Linux (Floyd Davidson)
----------------------------------------------------------------------------
From: Klaus Hirsch <[EMAIL PROTECTED]>
Subject: Frage
Date: Tue, 06 Jun 2000 13:16:36 +0200
Hallo
Oops false group
------------------------------
From: "Ron Sinclair" <[EMAIL PROTECTED]>
Crossposted-To:
alt.os.linux,alt.os.linux.mandrake,comp.graphics.rendering.renderman,comp.os.linux.setup
Subject: Re: Many questions and much dissatisfaction
Date: Tue, 6 Jun 2000 20:31:37 +0900
> This can happen when your BIOS detects a PnP >soundcard before the OS gets
to see it. Turning >off PnP detection in your BIOS might solve the
> problem.
I use Mandrake 7.0 and also had initial problems with some of my ISA cards.
The OS wouldn't detect my modem or soundcard until I turned off PnP in the
BIOS.
I'm assuming that PnP would be less of a problem with PCI cards, no?
------------------------------
From: muzh <[EMAIL PROTECTED]>
Subject: Re: Installing Netscape 6
Date: Tue, 06 Jun 2000 23:36:03 +1200
"David .." wrote:
>
> Jan Houtsma wrote:
> >
> [snip]
> > Isn't that this crappy mozilla monster with all the debug stuff slowing it
> > down tremendously?
>
> Could be but I don't use it so can't really say.
>
This "Crappy Mozilla Monster" is still only half the size of
Netscape-4.73
Sure the debug stuff slows it down tremendously. Think how fast it will
be without the debug stuff!
I urge as many people as possible to use it, and make bug reports, so
that there will be no use for all the debug stuff!
--
Never trust a man in a suit --
cll
------------------------------
From: Thomas Heil \\ WIS <[EMAIL PROTECTED]>
Subject: News
Date: Tue, 06 Jun 2000 13:52:50 +0200
Is this the right Newsgroup for Inn questions ??
Thomas Heil
E-mail: [EMAIL PROTECTED]
------------------------------
From: peter pilsl <[EMAIL PROTECTED]>
Subject: crontab: evironment not evaluated
Date: Tue, 06 Jun 2000 11:57:26 GMT
On many linux-server I use the following standard-crontab to get the mail:
MAILTO=""
WEEK=`date +%Y-week%W`
CRONLOG=$HOME/.log/cron.$WEEK
*/10 * * * * fetchmail >>$CRONLOG 2>&1
*/30 * * * * date +%Y-%m-%d-%H-%M-%S >>$CRONLOG
but now on a machine this doesnt work. The variables WEEK and CRONLOG are
not interpreted. when I enter
*/10 * * * * fetchmail >>/home/user/.log/cron.test
all is running fine, but I dont want to write a different tab for each
user and I want to have it all ordered by week.
So why does cron on this machine not evaluate the variables ???
thanks,
peter
--
pilsl@
goldfisch.at.at
------------------------------
From: =?iso-8859-1?Q?Beno=EEt?= Smith <[EMAIL PROTECTED]>
Subject: Re: Installing Netscape 6
Date: Tue, 06 Jun 2000 14:07:46 +0200
muzh wrote:
> This "Crappy Mozilla Monster" is still only half the size of
> Netscape-4.73
Perhaps, but remember it's not a full version (there's no news, for instance...)
--
Beno�t Smith
Just a Rhyme Without a Reason
------------------------------
From: [EMAIL PROTECTED] (David Efflandt)
Subject: Re: How to set safe screen more than 7 minutes
Date: 6 Jun 2000 12:13:30 GMT
Reply-To: [EMAIL PROTECTED]
On Tue, 6 Jun 2000, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote:
>I forget where to set screen shut down.
>Please reply by e-mail.
See 'man setterm'. You can set it in /etc/rc.d/rc.local.
--
David Efflandt [EMAIL PROTECTED] http://www.de-srv.com/
http://www.autox.chicago.il.us/ http://www.berniesfloral.net/
http://hammer.prohosting.com/~cgi-wiz/ http://cgi-help.virtualave.net/
------------------------------
From: ray <[EMAIL PROTECTED]>
Subject: Re: FTP in GNOME file manager? [gmc 4.5.50, Helix GNOME 1.2.1]
Date: Tue, 06 Jun 2000 12:17:57 GMT
"D. D. Brierton" wrote:
> On Mon, 05 Jun 2000 21:30:46 GMT, ray <[EMAIL PROTECTED]>
> wrote:
>
> > > ftp.redhat.com was just an example. What I meant was that I can't connect to any
> > > FTP sites at all using gmc. In fact, as far as I can tell, it doesn't even try.
> >
> > Yes, I suspect this: I have used this feature a lot, and while reading your
>post I
> > ftp'd in GMC to sunsite and gnome, but NOT to redhat. Maybe this is what is
>happening.
> > That site is nearly always user limited, and you can't get in. GMC has no way of
> > notifying you of that. It APPEARS to have "hung up" or "not be trying", etc. when
>in
> > reality it has gotten a notice that there are "too many users" etc. Maybe the
>coders
> > could give us something in the command line of GMC that says that, preventing the
> > confusion, perhaps. Bear in mind that all this is supposition, perhaps others could
> > help us with this. --
>
> The problem is definitely with gmc and not with the site. If I open a
> terminal and ftp to ftp.cogsci.ed.ac.uk I have no problems whatsoever.
> If I then immediately try in gmc nothing happens. I hit return and the
> Location Bar instantly switches back to the address of the local
> directory gmc is open at.
>
> I have scoured the documentation and the preferences in gmc to no avail.
> This also happens whether I am logged in as root or as an ordinary user.
> I even created a new test account an let my new version of GNOME (Helix
> 1.2.1) create totally fresh default configuration, and ftp from gmc
> doesn't work in that account either.
>
> One thing that occurs to me: mcserv is not running. But I thought that
> mcserv was so users of other machines could browse *my* filesystem. Does
> mcserv have to be running for gmc to be able to ftp to *other* machines?
> (Other VFSs work by the way: I can click on tar and rpm files and their
> contents open as if they were directories. Its just ftp that won't
> work.)
> =====================================================================
> D. D. Brierton, Centre for Cognitive Science, University of Edinburgh
> [EMAIL PROTECTED] http://www.cogsci.ed.ac.uk/~ddb
> =====================================================================
Hmm, how interesting, I thought that also. Well, only take a few minutes to try
that,
let us know, the thing is simply infallible here, will ftp anywhere, and I just love
that
feature. Actually, I have done nothing in terms of any setup or other tweaking, when
GMC
comes up, it just does it. If that's not the culprit, only thing left I can think of is
some corruption occurred either at install time, or after, try -e the rpm and -i it
again.
LET US KNOW!!!!
--
Ray R. Jones
The Computer Shop
[EMAIL PROTECTED]
HTTP://gordo.penguinpowered.com
Linux 2.2.15 #2 Sat Jun 3 12:01:16 EDT 2000 i586
------------------------------
From: [EMAIL PROTECTED]
Subject: DELL's Linux price is HIGHER than Win98
Date: Tue, 06 Jun 2000 12:13:43 GMT
Can anyone answer this one? Why is the price for a Dell running Linux
more that the price for a Win98 box. It should be cheaper considering
it is a FREE OS!!! I checked the Dell web site and configured a "Dell
Dimension XPS T" and for Win98 the price was $1,658 but for RedHat Linux
6.1 (The older Version) $1,737. I have listed the Dell options below:
Date: 6/5/00 @ 8:00AM
Dell Dimension XPS T PIII Mini Tower: PIII @ 700MHz [220-2135]
Memory: 128MB 100MHz SDRAM [311-8410]
Keyboard: QuietKey Keyboard [310-7002]
Monitor: Dell Ultrascan P780 17" [320-4501]
Video Card: 32MB NVIDA geFORCE 4X AGP Graphics Card [320-0131]
Hard Drive: 20.4 GB Ultra ATA Hard Drive [340-2409]
Operating System: Win98 [310-8921] or Linux [420-2250]
Mouse: MS IntelliMouse [310-0050]
Network Card: 3COM 3C905C-TXM 10/100 Remote Wake Up [430-3280]
Modem: No Modem [313-3607]
Optical Drives: 48X Max / 20X Min CDROM [313-3922]
Sound Card: SoundBlaster Live! Value Digital [313-7869]
Speakers: Harmon/Kardon Speakers [313-3925]
Bundled Software: No MS Office [412-1397]
Iomega Zip Drives: No Zip Drives [460-8320]
Norton Antivirus: NAV 2000 [412-5620] ONLY ON THE MS SYSTEMS FOR FREE!!
Service: 1 Yr. Next Business Day On-Site P&L, Yrs 2&3, BSC [900-1590]
Sent via Deja.com http://www.deja.com/
Before you buy.
------------------------------
From: [EMAIL PROTECTED] (Svend Olaf Mikkelsen)
Subject: Re: partition lost
Date: Tue, 06 Jun 2000 12:25:18 GMT
"MAP" <[EMAIL PROTECTED]> skrev:
>Hi Svend,
>
>your program doesn't work for me...It only detect 1 of 4 partition (the
>primary), and it fails saying :
>
>Overflow in line No line number in module FINDPART at address 1634:410C
>
>the result fp.txt is the attached
>
>
>Any help will be appreciated
>
>Marc
>Findpart, version 3.6.
>Copyright Svend Olaf Mikkelsen, 2000.
>Searches for partitions type 01, 04, 06, 07, 0B, 0C, 0E, 83,
>logical partitions type 82, plus Fdisk F6 and Lilo sectors.
>Information based on bootsectors is marked B (ID guessed).
>If the disk is larger than supported by BIOS, the supported
>part of the disk is examined. Disks are numbered from 1.
>OS: DOS 7.10
>Disk: 2 Cylinders: 299043 Heads: 1 Sectors: 60 MB: 8761
>-PCyl N ID -----Rel -----Num ---MB -Start CHS- --End CHS-- BS CHS
> 204 1 83 60 377880 184 205 0 1 6502* 0 60 NB OK
> 204 1 83 60 377880 184 205 0 1 6502* 0 60 NB OK
It seems as something is wrong with the BIOS setting. 299043
cylinders, 1 head and 60 sectors is something that I deninately did
not see before.
Try this (assuming the disk is an IDE disk):
Set the disk to none in BIOS. Then boot do DOS and do if the disk is
primary slave:
findpart ps fp-a.txt
If the disk is not primary slave, then use in stead of ps:
pm for primary master
sm for secondary master
ss for secondary slave
Note that depending on which BIOS settings was used when the
partitions were created, it might be nessesary with further searches.
--
Svend Olaf
------------------------------
From: [EMAIL PROTECTED] (Svend Olaf Mikkelsen)
Subject: Re: partition lost
Date: Tue, 06 Jun 2000 12:27:43 GMT
"MAP" <[EMAIL PROTECTED]> skrev:
>Hi Svend,
>
>your program doesn't work for me...It only detect 1 of 4 partition (the
>primary), and it fails saying :
>
>Overflow in line No line number in module FINDPART at address 1634:410C
>
>the result fp.txt is the attached
>
>
>Any help will be appreciated
>
>Marc
>Findpart, version 3.6.
>Copyright Svend Olaf Mikkelsen, 2000.
>Searches for partitions type 01, 04, 06, 07, 0B, 0C, 0E, 83,
>logical partitions type 82, plus Fdisk F6 and Lilo sectors.
>Information based on bootsectors is marked B (ID guessed).
>If the disk is larger than supported by BIOS, the supported
>part of the disk is examined. Disks are numbered from 1.
>OS: DOS 7.10
>Disk: 2 Cylinders: 299043 Heads: 1 Sectors: 60 MB: 8761
>-PCyl N ID -----Rel -----Num ---MB -Start CHS- --End CHS-- BS CHS
> 204 1 83 60 377880 184 205 0 1 6502* 0 60 NB OK
> 204 1 83 60 377880 184 205 0 1 6502* 0 60 NB OK
It seems as something is wrong with the BIOS setting. 299043
cylinders, 1 head and 60 sectors is something that I deninately did
not see before.
Try this (assuming the disk is an IDE disk):
Set the disk to none in BIOS. Then boot do DOS and do if the disk is
primary slave:
findpart ps fp-a.txt
If the disk is not primary slave, then use in stead of ps:
pm for primary master
sm for secondary master
ss for secondary slave
Depending on which BIOS settings was used when the partitions were
created, it might be nessesary with further searches after this.
Note that some ISP's (including my primary) filter messages containing
attached files. It is better to insert the output files as text.
--
Svend Olaf
------------------------------
From: Mihaly Gyulai <[EMAIL PROTECTED]>
Subject: 'mc' crashes after using FTP
Date: Tue, 06 Jun 2000 12:19:12 GMT
I used Midnight Commander (mc) 4.1.36 for local FTP
without serious problems.
Now I upgraded to mc-4.5.50, and it crashes after
every FTP copy, if I copy _to_ the remote machine...
It copies only the first 8192 bytes and dies...
It can copy _from_ the remote machine with FTP, it's OK.
Shall I go back to an earlier version of MC??
What can cause the problem?
Q2. Anyone recommend another filemanager on console?
--
Mihaly Gyulai
mailto:[EMAIL PROTECTED]
http://www.freeyellow.com/members5/gyulai/
Sent via Deja.com http://www.deja.com/
Before you buy.
------------------------------
From: Sverre Torjussen <[EMAIL PROTECTED]>
Subject: How to create device on diskette?
Date: 06 Jun 2000 14:30:59 +0200
Hello,
I have a PC running RH 6.2(Zoot)/WIN NT 2k. I use a boot diskette to
load the Linux environment and would like to take a back up of the
boot diskette. But I have some problems trying to do this... The
device properties/attributes I would like to create on diskette are:
[root@wst_pc /dev]# ls -l hdb1 fd0
brw-rw---- 1 root floppy 2, 0 May 5 1998 fd0
brw-rw---- 1 root disk 3, 65 May 5 1998 hdb1
[root@wst_pc dev]# cd; umount /mnt/floppy/; mount /dev/fd0 /mnt/floppy; cd
/mnt/floppy
[root@wst_pc floppy]# cd dev/
[root@wst_pc dev]# mknod -m 660 hdb1 b 3 65
mknod: hdb1: Operation not permitted
[root@wst_pc dev]# mknod -m 600 fd0 b 2 0
mknod: fd0: Operation not permitted
I may successfully run the same mknod command if the device NAME
(hdb1/fd0) resides on any partition on the hard disk run. No write
protection of floppy (mounted rw).
Any help very much appreciated. For example, alternative methods or work arounds...
Cheers
//Sverre
Sverre Torjussen, seniorconsultant
WM-data Consulting AS
P.O. 1832 Myrene
N-4801 Arendal (Norway)
Phone office: +47 370 73706
Fax +47 370 73701
e-mail: [EMAIL PROTECTED]
Internet: www.wmdata.no
------------------------------
From: [EMAIL PROTECTED] (David Efflandt)
Crossposted-To: alt.os.linux.suse
Subject: Re: how to start VNC during boot?
Date: 6 Jun 2000 12:33:21 GMT
Reply-To: [EMAIL PROTECTED]
On Mon, 05 Jun 2000, Sascha Kicken <[EMAIL PROTECTED]> wrote:
>And if you want to use VNC for X on a remote PC you will still need a
>running X on the Server since VNC transfers the contents of the Screen. If
>no X-Server shall run on the Linux-server you'll still need a XServer for
>the 'client' like KeaX or Exceed (if you are using a Windows machine as
>client).
Apparently you are not familiar with VNC. It runs a background vncserver
(which is an X server on Unix/Linux systems) on the box you want to
access. There is also a Windows VncServer. On the remote (any OS), you
just need the vncviewer or any Java capable web browser to access the
server at the main box. You don't need KeaX or Exceed.
http://www.uk.research.att.com/vnc/index.html
--
David Efflandt [EMAIL PROTECTED] http://www.de-srv.com/
http://www.autox.chicago.il.us/ http://www.berniesfloral.net/
http://hammer.prohosting.com/~cgi-wiz/ http://cgi-help.virtualave.net/
------------------------------
From: [EMAIL PROTECTED] (Svend Olaf Mikkelsen)
Subject: Re: partition lost
Date: Tue, 06 Jun 2000 12:36:36 GMT
"MAP" <[EMAIL PROTECTED]> wrote:
>Hi Svend,
>
>your program doesn't work for me...It only detect 1 of 4 partition (the
>primary), and it fails saying :
>
>Overflow in line No line number in module FINDPART at address 1634:410C
>
>the result fp.txt is the attached
>
>
>Any help will be appreciated
>
>Marc
>Findpart, version 3.6.
>Copyright Svend Olaf Mikkelsen, 2000.
>Searches for partitions type 01, 04, 06, 07, 0B, 0C, 0E, 83,
>logical partitions type 82, plus Fdisk F6 and Lilo sectors.
>Information based on bootsectors is marked B (ID guessed).
>If the disk is larger than supported by BIOS, the supported
>part of the disk is examined. Disks are numbered from 1.
>OS: DOS 7.10
>Disk: 2 Cylinders: 299043 Heads: 1 Sectors: 60 MB: 8761
>-PCyl N ID -----Rel -----Num ---MB -Start CHS- --End CHS-- BS CHS
> 204 1 83 60 377880 184 205 0 1 6502* 0 60 NB OK
> 204 1 83 60 377880 184 205 0 1 6502* 0 60 NB OK
It seems as something is wrong with the BIOS setting. 299043
cylinders, 1 head and 60 sectors is something that I definitely did
not see before.
Try this (assuming the disk is an IDE disk):
Set the disk to none in BIOS. Then boot do DOS and do if the disk is
primary slave:
findpart ps fp-a.txt
If the disk is not primary slave, then use in stead of ps:
pm for primary master
sm for secondary master
ss for secondary slave
Depending on which BIOS settings was used when the partitions were
created, it might be nessesary with further searches after this.
Note that some ISP's (including my primary) filter messages containing
attached files. It is better to insert the output files as text.
--
Svend Olaf
------------------------------
From: Jean-Philippe Desbiems <[EMAIL PROTECTED]>
Subject: Seaching for Matlab for linux ... somebody has it..
Date: Tue, 06 Jun 2000 07:55:11 -0400
Hi ,
I'm searching for a linux version of Matlab for my University work
and I can't find any. Is there anybody who has it or can tell me where i
can get it...this is the last software I need to get rid of Windose on
my computer ...
So waiting for advice or answer
bye !
------------------------------
From: =?iso-8859-1?Q?Beno=EEt?= Smith <[EMAIL PROTECTED]>
Subject: WANTED Epson Printer Driver
Date: Tue, 06 Jun 2000 14:39:51 +0200
Greetings,
I am looking for the 'stc600p.upp' driver, for my Epson Stylus Color 640
printer. Please could someone tell me where to find it ?
Thanks in advance !
--
Beno�t Smith
Just a Rhyme Without a Reason
------------------------------
From: Geoff Sullivan <[EMAIL PROTECTED]>
Subject: Re: samba troubles
Date: Tue, 06 Jun 2000 12:41:19 GMT
Akira Yamanita <[EMAIL PROTECTED]> wrote:
> Cool. Can you see Penguin in the Network Neighborhood too?
No :-( , but I can browse sunfish from the file manager in penguin. That's
a partial improvement.
Now, however, there are two other problems. If I boot penguin and sunfish
is shut off (network down), penguin takes *forever* to boot. It grinds to
a halt when it encounters the modules automount, smb server, and one of
the mail deamons. It litteraly takes a half hour to figure out that there
is no network available before it completes the boot! The remedy was to
remove the automount and smb modules. Forget about automount and load the
smb deamons when I need them.
The other complication is that Kppp no longer works. I only would use it
when the network is shut down to dial out from the modem in penguin. It
does dial the phone, but no ppp connection is made and it immediately
hangs up.
I *can* get pppd to work though! Go figure.
(why couldn't I just keep it simple?) Right now I'm using my trusty old
Commodore 128 to do email, newsgroups, etc. It never fails me!
--
** **
G. Sullivan sunfishATshell.gis.net.invalid http://www.gis.net/~sunfish
------------------------------
From: [EMAIL PROTECTED] (David Efflandt)
Subject: Re: Mandrake 7 setup question
Date: 6 Jun 2000 12:43:07 GMT
Reply-To: [EMAIL PROTECTED]
On 5 Jun 2000 23:36:28 GMT, Karel Jansens <jansens_at_ibm_dot_net> wrote:
>If I let Mandrake set itself up semi-automatically (i.e. no "expert"
>mode), will I still be able to stick lilo somewhere else than on the
>MBR of the root partition, like /dev/hda6? I use OS/2's Boot Manager
>as my primary booting tool, and it is now set up to boot lilo off the
>linux partition. I would be seriously p*ssed off if Mandrake were to
>overwrite Boot Manager.
Mandrake 7 lets you put LILO where you want to. When I got a 27.3G hard
drive I specifically put a couple of 16MB /boot partitions on it before
Win98se. I have the LILO for RH 6.1 in hda1 and LILO for Mandrake 7.0 in
hda2. Then I can make either the active boot partition and can totally
reinstall anything without having to worry about anything (like Windows or
virus detection) messing with the MBR.
--
David Efflandt [EMAIL PROTECTED] http://www.de-srv.com/
http://www.autox.chicago.il.us/ http://www.berniesfloral.net/
http://hammer.prohosting.com/~cgi-wiz/ http://cgi-help.virtualave.net/
------------------------------
From: Guy-Armand Kamendje <[EMAIL PROTECTED]>
Subject: loading module on red-hat 6.2
Date: Tue, 06 Jun 2000 12:43:18 +0000
Reply-To: [EMAIL PROTECTED]
I'm trying to let my kernel load the imm driver for my Zip250 at boot
time. I use following line in my fstab and for the case mount the disk
inserted in the drive.
/dev/sbd4 /mnt/zip auto notauto,rw 0 0
unfortunately, this does not work and I have to do it manualy using
modprobe imm
and
mount /dev/sbd4 /mnt/zip
I have also tried some entries like
probe imm
in conf.module but without succes
Is there any way to automate this job?
thanks G-A
--
Guy-Armand Kamendje || Tel +43 316 873 55 51
Technical University Graz || www.sbox.tu-graz.ac.at/home/g/gaillard/
I.A.I.K || www.iaik.at/people/gkamendje/gkamendje.html
------------------------------
Subject: Hideous X Fonts
From: Neurocrat <[EMAIL PROTECTED]>
Date: 06 Jun 2000 23:54:57 +1000
100dpi fonts in Debian Potato (X 3.3.6) don't seem to be working
correctly. Whenever I run an app using qt or tk widgets, the fonts
are very ugly (thin, squashed, inconsistent in thickness).
If I remove the 100dpi fonts package (leaving xfonts-75dpi and
xfonts-scalable) things look a lot better, except that all GUI toolkit
fonts are fixed pitch fonts. (Something like Courier or Courier
New). When I write a qt app explicitly requesting a "Times" font, I'm
out of luck. Same plain fixed pitch font.
Can anybody suggest a package I might need to install or reconfigure?
I know it is not an inherent weakness in XFree86-3.3.6, because the
same qt and tk apps look perfectly normal in FreeBSD.
Thanks in advance for any suggestions.
------------------------------
From: Stephen Cornell <[EMAIL PROTECTED]>
Crossposted-To: linux.redhat.install,alt.os.linux.slackware,comp.os.linux.x
Subject: Re: Where can i get XBF-neomagic-glibc-1.0.0-1.i386.rpm ?
Date: 06 Jun 2000 13:47:57 +0100
[EMAIL PROTECTED] writes:
> I am having problems with X-Windows.
>
> I have installed RedHat 6.1 workstation.
> I am using the X-Server : XBF-i740-libc5-1.0.0-1.i386.rpm
> I have a Dell Inspiron 3000
> with Pentium MMX 266 MHZ and 64MB RAM
> Video Controller is NeoMagic 2160 with 2MB memory
> (128 bit accelerated...),Type - TFT SVGA with maximum resolution
> at 800x600 pixels;16 million colors.
>
> My X-Window looks fine.However it occupies only 3/4th of my screen
> giving me an 8 bpp depth.When i run it under 16bpp it does not
> work.
>
> I keep reading about success with the server which comes with
> XBF-neomagic-glibc-1.0.0-1.i386.rpm.
I seem to recall that the XBF servers were supplied with older Red Hat
distributiuons for use with cards that were not supported by XFree86.
According to the XFree86 website, your chipset is now supported, and
you should use the SVGA server. If you install the XFree86-SVGA RPM
that came with your distribution and run Xconfigurator, it should set
things up for you.
I installed Red Hat 6.2 on a Dell Latitude XPi, and I initially had a
problem getting 800x600 screen resolution. If you encounter this
problem, feel free to email me and I'll tell you how I fixed this.
The problem with X only ocuppying a small part of the screen is
probably due to the wrong virtual desktop size. Again, email me if
the problem persists.
--
Stephen Cornell [EMAIL PROTECTED] Tel/fax +44-1223-336644
University of Cambridge, Zoology Department, Downing Street, CAMBRIDGE CB2 3EJ
------------------------------
From: Floyd Davidson <[EMAIL PROTECTED]>
Subject: Re: Serious fragmentation under Linux
Date: 06 Jun 2000 03:56:06 -0800
MH <[EMAIL PROTECTED]> wrote:
>"Peter T. Breuer" wrote:
>>
>> MH <[EMAIL PROTECTED]> wrote:
>> : On a recent reboot, I noticed that I had 11.1%, 15.4%, and 19.8%
>> : "non-contiguous" files. Since this workstation is NOT used as a server,
>> : and since it has only been up for a few weeks, I found this level of
>> : fragmentation more than a little surprising. Even more so, given the
>>
>> Well (a) who cares and (b) why do you think it's bad?
>>
>> Clue: it isn't. You should recall that you have about 100 processes
>> running in your system at one time, thus the heads are running between
>> what those processes want to see. They're not waiting for you to scan
>> linearly over one file.
>>
>> It sounds like your partitions are in rw mode and are too small for
>> their task.
>>
>> Peter
>
>Actually, maybe a few dozen processes, most of which have nothing to do
>with reading files from disk. See response to Dances with Crows--for a
>clue.
Peter has a bag full of clues. He's a very knowledgeable person,
and you should listen closely to what he has said.
It simply does not make any great difference. Can you actually
figure out when and where it would cause some difference???
On a multitasking multiuser system (whether it is running just a
few dozens of processes or the 100 number that includes those
not permanently in memory and thus not shown by the ps command)
the heads are randomly jumping all over the disk anyway. Each
time slice that a process gets is very small compared to the
time for disk io, and hence context switching multiple processes
that do file io causes the exact same effect as disk
fragmentation to exist all of the time, even with 0% disk
fragmentation. Hence even 50% fragmentation would have little
*added* effect.
All of the binaries and configuration files that you installed
to start with are totally unfragmented. Only files written
after that can possibly be fragmented. Hence 99% of what you
want to read from disk quickly and repeatedly in a sequential
pattern is unaffected by whatever happens to files later on.
Additionally, many files such as logs and configuration files
are not read sequentially anyway, hence fragmentation once again
is of no consequence.
But regardless of what I've pointed out, what Peter has pointed
out, and the extensive answers of others who gave loads of
information... even if none of that existed, it still would
make little difference because Linux uses disk caching. Any
file that is read often is read from RAM, not from disk.
Fragmentation has absolutely no effect once the file is cached,
and any file reads that are repeated often enough to affect
system performance are read from disk exactly one time only.
As to your partitions and how to reduce the fragmentation, there
are some things you can do (just be warned that this affects
only fragmentation, not system performance!). One is to isolate
all directories that will be fragmentated. That would be spool
and tmp dirs, for example. Put them all on one partition and
use symbolic links to access them. For example you have, if I
remember right, /var and /tmp partitions. Who cares if /var is
fragmented??? or /tmp. Those two partitions are where you want
to put other directories that might have a great deal of
write/read/remove activity. Combine the two of them into one
partition, mount it as /var and have /tmp and /usr/tmp be
symlinks to /var/tmp. Any spool files you find in /usr (though
in modern systems there should not be any) should also be
symlinked to a directory in /var. Likewise any src directories,
such as /usr/src, should be symlinked and physically on the
fragmented partition.
Then make /home a separate partition too, or if space is tight
put it on the same partition as /var. One way to do this is on
a small disk is to mount boot and root partitions (and a /usr
too if you choose) plus one other on something like /u, /u1,
/usr1 or whatever. Then symlink /var and /home and everything
else mentioned above to that partition. Basically there are
non-fragmenting partitions and one fragmenting partition. That
means installing a new binary in /sbin or whatever will not be a
fragmented file.
It is an interesting exercise to set up the above, but it is
also a grand waste of time on a modern system. (Years ago it
made a great deal of sense. And before symlinks were available
it was an art to preconceive of a partitioning scheme that would
last for as long as possible before a new scheme to be designed
and implemented.).
There are a couple of partitioning tricks that might actually
help system performance, though I suspect only so slightly as
to be measurable but not noticable. Either use one large root
partition, or put /usr and /var partitions right next to the
root partition. For example, on a 30Gb disk, don't put root
first, /usr in the middle of the disk, and /var at the end of
the disk. That will cause the worst case head seek times for
most of the disk activity. Swapping is something you obviously
want to avoid if at all possible, but if it is not totally
impossible (a 20Mb system where you actually want to run X),
put the swap partition next to root too.
Or best of all, use several disks and put all of these things
on different disks. Put the most active two partitions on
the two different IDE channels too.
And when all is said and done you'll feel good about it, but
you won't be able to measure the difference. :-)
Floyd
--
Floyd L. Davidson [EMAIL PROTECTED]
Ukpeagvik (Barrow, Alaska)
------------------------------
** 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
******************************
