Linux-Setup Digest #375, Volume #21 Tue, 5 Jun 01 10:13:12 EDT
Contents:
Re: Geforce2 Go and XFree 4.0.3 (Fabrice Colin)
Re: Is this fixable? (Colin Watson)
Re: Dumb question! How to install Rpms without rpm? ("Peter T. Breuer")
HELP! Unable to mount a partition ("Sean Dynan")
Re: Is this fixable? ("Peter T. Breuer")
Re: Geforce2 Go and XFree 4.0.3 (Mike H. Miller)
Re: HELP! Unable to mount a partition ([EMAIL PROTECTED])
redhat rescue disk ("Yaniv Fine")
Re: HELP! Unable to mount a partition ("SpOOk")
Re: partitions (Huw Lynes)
Re: X Windows Clients ("Dave Armbruster")
Re: X Windows Clients (Lew Pitcher)
Re: redhat rescue disk ("Peter T. Breuer")
Re: partitions ("David Dorward")
Re: Ncurses and menuconfig (Thomas Dickey)
Re: Lilo questions ("Eric")
Re: SCSI problems (Trevor Hemsley)
----------------------------------------------------------------------------
Date: Tue, 05 Jun 2001 12:15:10 +0100
From: Fabrice Colin <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.hardware,comp.os.linux.portable,comp.os.linux.x
Subject: Re: Geforce2 Go and XFree 4.0.3
"T. J. Domsalla" wrote:
>
> Hi,
>
> I recently installed SuSE Linux 7.2 (XFree 4.03) on a Dell Inspiron 8000
> with Nvidia Geforce2 Go. Kernel is 2.4.5, Nvidia GLX and kernel driver
> 1.0.1251.
> When starting X I get "Failed to initialize the NVdriver kernel module" and
> "Screen found, but none have a usable configuration".
> Did anyone get it running?
>
I had the same problem this weekend after having installed 1.0-1251, but
with a GeForce2 GTS on RH6.2/XFree86 4.02/kernel 2.4.5.
Unfortunately I don't remember how I fixed it. I wonder if it was
because
I played with the IgnoreEDID option in XF86Config ?
If it says "Failed to initialize the NVdriver kernel module", it's maybe
worth switching to initlevel 3 (that's "multi-user without X"), load
the NVdriver manually with 'modpobe NVdriver' and run X with 'startx'.
See if it makes any difference.
Sorry if this is all very vague, but I am at work and I am hungry ;-)
Fabrice
------------------------------
From: [EMAIL PROTECTED] (Colin Watson)
Subject: Re: Is this fixable?
Date: 5 Jun 2001 11:03:37 GMT
Dave Uhring <[EMAIL PROTECTED]> wrote:
>Yu Di wrote:
>> Hi, I tried to upgrade glibc for my Linux system (I had glibc-2.1.3
>> installed, and I wanted to upgrade to 2.2.3). I got the source code of
>> glibc-2.2.3, compiled it. I first installed it to /usr/local/glibc-new
>> to see if the installation can be finished successfully, and it did,
>> but then I re-configured it to use the prefix /usr (so the libraries
>> go to /lib and /usr/lib), and during the "make install" it breaks
>> down, saying /bin/sh can't find shared object files etc. I made the
>> big mistake of logging out, and cannot get in any more after that. The
>> system refuses to boot, saying init cannot find shared object files
>> etc. Lucky that I had my personal important files backup'ed before
>> this.
>>
>> So is this still fixable? Or should I simply re-install the whole
>> system?
>
>Besides breaking just about every executable that you have if you
>actually do get that new set of libraries installed, what else do you
>hope to accomplish with this exercise?
Why do you think that installing a new glibc will break old executables?
(Hint: it won't, at least not if done properly.) Of course, installing
it by brute force is never a good idea.
>Yes, you are going to have to re-install.
Reinstalling libc6 and related packages should be enough. How easy this
is depends on your distribution (or lack thereof).
--
Colin Watson [[EMAIL PROTECTED]]
Jam wants colonies. I am a jest on clowns.
------------------------------
From: "Peter T. Breuer" <[EMAIL PROTECTED]>
Subject: Re: Dumb question! How to install Rpms without rpm?
Date: Tue, 05 Jun 2001 11:20:23 GMT
ada2 <[EMAIL PROTECTED]> wrote:
> I accidentaly deleted my rpm program. Now, I can not install it back
> because it is packed as rpm. Is there some other way to install
> rpm-4.0-XXX.rpm? or is there exist a rpm-4.0.xxx.src.tgz that I can install
> with ./configure, make , make install?
You need rpm for that :-).
You can use rpm2cpio to unpack the binary rpm, and then cpio to
install the parts. Then you can reinstall rpm using rpm.
I.e. you need a working copy of rpm first. You either compile it,
borrow it, or steal it ... it's called "bootstrapping". I'd just
copy rpm from somewhere else if all you did was remove the executable.
If you rpm -e'ed it, then I'd borrow a rpm2cpio and put it back, then
reinstall it. If you can't get hold of an rpm2cpio, tehn I'd compile
the whole thing.
Peter
------------------------------
From: "Sean Dynan" <[EMAIL PROTECTED]>
Crossposted-To: alt.linux,alt.os.linux
Subject: HELP! Unable to mount a partition
Date: Tue, 5 Jun 2001 12:36:54 +0100
Newbie alert. And it's gonna show.
Redhat 6.2; kernel 2.2.14-5.0; mount version 2.10f.
"mount /dev/hda5 /home"
mount: /dev/hda5 already mounted or /home busy
/etc/mtab
/dev/hda7 / ext2 rw 0 0
none /proc proc rw 0 0
/dev/hda1 /boot ext2 rw 0 0
none /dev/pts devpts rw,gid=5,mode=620 0 0
/etc/fstab
/dev/hda7 / ext2 defaults
1 1
/dev/hda1 /boot ext2 defaults
1 2
/dev/hda5 /home ext2 defaults
1 2
/dev/fd0 /mnt/floppy auto noauto,owner
0 0
none /proc proc defaults
0 0
none /dev/pts devpts gid=5,mode=620
0 0
/dev/hda6 swap swap defaults
0 0
The directory /home exists and is empty.
If I substitute /home for /mnt/home in the mount command I get the same
error message for /mnt/home.
Here's the history:
/ and /home used to be mounted the other way up. In other words, / was
originally mounted on /dev/hda5 and /home was originally mounted on
/dev/hda7. Because of disk space constraints I decided to devote the larger
hda5 partition to / (and /usr and whatever) and dedicate the smaller hda7
partition to my /home directory.
I used "cp -dpR" to copy all the non-mounted directories formerly on hda7 to
hda5 (including /proc, probably my downfall). I didn't remove all the
original directories formerly on hda5 (they're still there).
I edited my new, copied /etc/fstab to reflect the partition change and
rebooted.
Then I remembered to change my new /etc/lilo.conf and re-run lilo. This all
seemed to go according to plan.
But the O/S refuses to mount /dev/hda5 on /home (not even read-only) and
fsck runs every time at boot-up on /dev/hda5.
Please give clueless some clues.
Thanks in advance,
--Sean.
------------------------------
From: "Peter T. Breuer" <[EMAIL PROTECTED]>
Subject: Re: Is this fixable?
Date: Tue, 5 Jun 2001 13:41:30 +0200
Colin Watson <[EMAIL PROTECTED]> wrote:
> Why do you think that installing a new glibc will break old executables?
Eh? Surely in this case it should do. The symbols are versioned, and
a 2.1 to 2.2 change should cause most 2.1 symbols not to be found in
the new 2.2 library, no? Or are you saying that he wants the 2.1
library still around as the principal library, and he should install
the new library in some harmless place, only to be found by a
LD_LIBRARY_PATH directive for particular executables? Yay, I'd second
that!
> (Hint: it won't, at least not if done properly.) Of course, installing
> it by brute force is never a good idea.
Yes. But I really doubt if 2.1 and 2.2 can coexist with equal status
(I'm not about to try it just to find out!). Now libc5 and libc6 we
know have no problem living with each other.
> Reinstalling libc6 and related packages should be enough. How easy this
> is depends on your distribution (or lack thereof).
Peter
------------------------------
From: [EMAIL PROTECTED] (Mike H. Miller)
Crossposted-To:
comp.os.linux.hardware,comp.os.linux.portable,comp.os.linux.x,de.comp.os.unix.linux.hardware
Subject: Re: Geforce2 Go and XFree 4.0.3
Date: 5 Jun 2001 11:18:28 GMT
On Mon, 4 Jun 2001 18:34:42 +0200, T. J. Domsalla <[EMAIL PROTECTED]> wrote:
>Hi,
>
>I recently installed SuSE Linux 7.2 (XFree 4.03) on a Dell Inspiron 8000
>with Nvidia Geforce2 Go. Kernel is 2.4.5, Nvidia GLX and kernel driver
>1.0.1251.
>When starting X I get "Failed to initialize the NVdriver kernel module" and
>"Screen found, but none have a usable configuration".
>Did anyone get it running?
>
>Thorsten J. Domsalla
You're going to love this ;). Where I work, we've got a few of these laptops...
they're great machines, but the GeForce2 Go chips were giving us problems.
First, you can use the 0.9-769 drivers (just be aware that the kernel module
and the GLX driver must be the same version). The problem with that approach is
that it crashes whenever you exit X (known bug). This means you can never shut
down the machine safely.
But wait! 1.251 will work... if you pass the kernel modules an option of
NVreg_Mobile=1 (1 for a Dell, 2 if you have a Toshiba). To do this, either load
the modules manually using modprobe NVdriver NVreg_Mobile=1, or add the
following to /etc/modules.conf (or /etc/conf.modules, which was the older style)
options NVdriver NVreg_Mobile=1
Once you do this (which, btw, is obscurely documented in /usr/share/doc/NVIDIA_
GLX_1.251/ ... NOT with the kernel modules... duh), you can run X safely. Note
that there still is a bug that prevents you from _starting_ X more than once.
Therefore, I advise you edit inittab and make sure you have a line that reads
id:5:initdefault:
instead of
id:3:initdefault:
The first one will automatically load into X. Good luck!
P.S. Everything I did was on a RH 6.2 and RH7.1 system. YMMV.
------------------------------
From: [EMAIL PROTECTED]
Crossposted-To: alt.linux,alt.os.linux
Subject: Re: HELP! Unable to mount a partition
Date: 5 Jun 2001 12:32:32 GMT
In comp.os.linux.setup Sean Dynan <[EMAIL PROTECTED]> wrote:
> / and /home used to be mounted the other way up. In other words, / was
> originally mounted on /dev/hda5 and /home was originally mounted on
> /dev/hda7. Because of disk space constraints I decided to devote the larger
> hda5 partition to / (and /usr and whatever) and dedicate the smaller hda7
> partition to my /home directory.
I think that you are still booting your root from /dev/hda5, and this
is the origin of the "busy" message you received.
Check if you have a root option in your lilo.conf and use rdev
to change the root directory in the kernel (that shouldn't be
necessary, but the other way never worked for me...), then reboot
your system.
Davide
------------------------------
From: "Yaniv Fine" <[EMAIL PROTECTED]>
Subject: redhat rescue disk
Date: Mon, 4 Jun 2001 17:59:48 +0200
Reply-To: "Yaniv Fine" <[EMAIL PROTECTED]>
hi all
i am using RedHat 6.2
how can i make linux Rescue disk ?
tnx
Yaniv Fine
Israel
------------------------------
From: "SpOOk" <[EMAIL PROTECTED]>
Crossposted-To: alt.linux,alt.os.linux
Subject: Re: HELP! Unable to mount a partition
Date: Tue, 5 Jun 2001 07:41:43 -0500
you have five partitions on your first hard drive?
"Sean Dynan" <[EMAIL PROTECTED]> wrote in message
news:3b1cc45f$0$15023$[EMAIL PROTECTED]...
> Newbie alert. And it's gonna show.
>
> Redhat 6.2; kernel 2.2.14-5.0; mount version 2.10f.
>
> "mount /dev/hda5 /home"
> mount: /dev/hda5 already mounted or /home busy
>
> /etc/mtab
> /dev/hda7 / ext2 rw 0 0
> none /proc proc rw 0 0
> /dev/hda1 /boot ext2 rw 0 0
> none /dev/pts devpts rw,gid=5,mode=620 0 0
>
> /etc/fstab
> /dev/hda7 / ext2 defaults
> 1 1
> /dev/hda1 /boot ext2 defaults
> 1 2
> /dev/hda5 /home ext2 defaults
> 1 2
> /dev/fd0 /mnt/floppy auto noauto,owner
> 0 0
> none /proc proc defaults
> 0 0
> none /dev/pts devpts gid=5,mode=620
> 0 0
> /dev/hda6 swap swap defaults
> 0 0
>
>
> The directory /home exists and is empty.
>
> If I substitute /home for /mnt/home in the mount command I get the same
> error message for /mnt/home.
>
> Here's the history:
> / and /home used to be mounted the other way up. In other words, / was
> originally mounted on /dev/hda5 and /home was originally mounted on
> /dev/hda7. Because of disk space constraints I decided to devote the
larger
> hda5 partition to / (and /usr and whatever) and dedicate the smaller hda7
> partition to my /home directory.
>
> I used "cp -dpR" to copy all the non-mounted directories formerly on hda7
to
> hda5 (including /proc, probably my downfall). I didn't remove all the
> original directories formerly on hda5 (they're still there).
>
> I edited my new, copied /etc/fstab to reflect the partition change and
> rebooted.
>
> Then I remembered to change my new /etc/lilo.conf and re-run lilo. This
all
> seemed to go according to plan.
>
> But the O/S refuses to mount /dev/hda5 on /home (not even read-only) and
> fsck runs every time at boot-up on /dev/hda5.
>
> Please give clueless some clues.
>
> Thanks in advance,
> --Sean.
>
>
------------------------------
From: [EMAIL PROTECTED] (Huw Lynes)
Subject: Re: partitions
Date: 5 Jun 2001 12:37:22 GMT
Reply-To: [EMAIL PROTECTED]
In article <9fi7rl$cdb$[EMAIL PROTECTED]>, Mike91 wrote:
>I am intending to empty a 6.5gig partition on my HD and install linux on it,
>however I have been reading that I will need a swap partition and maybe a
>boot partition, so I am unsure how to sub-divide the 6.5gig. Also, the
>6.5gig is in FAT32, so I guess it would be better to change that to linux
>native? And if so, how?
Most distributions will allow you to do re-partitioning of your
hard drive. Alternatively you can use a third-party utility like
Partition magic, this can be useful as windows can object to the
state in which linux leaves the partition table. However as
Win is already installed on your system this is unlikely to be a
problem.
You can run linux off a fat32 partition but its bad for your
karma :)
You will most likely need to chop up that 6.5G into a swap
partition and an ext2 (linux native) partition to mount as
the root filesystem. A separate boot partition is a matter
of personal choice but it can be useful. At home I use
a boot partition on my ide drive to mount a root filesystem
on a scsi drive becuase I cannot boot from the scsi drive.
>
>I am intending to install Mandrake 7.2 (as I have been able to borrow a
>copy) ....am I right in thinking the setup procedure will be able to do any
>partition manipulation I will need?
Should do. You will probably have to explicitly tell the
install program that you want to partition the drives
manually else it'll just format the whole thing.
Have fun
Huw
--
To reply you know what "dot" should really be.
------------------------------
From: "Dave Armbruster" <[EMAIL PROTECTED]>
Subject: Re: X Windows Clients
Date: Tue, 5 Jun 2001 08:54:32 -0400
No error message other than the client reports the connection refused
yet the same client attaches fine to HP Unix systems......Dave
> > I have installed Redhat 7 on a new server. Everything works great
> >except I cannot access the system via an X Client. I can telnet, ftp,
> >and rlogin no problem. I have network access enabled and even have
> >run the xhost command to authorize the access. Redhat still refuses the
> >network connection. Anyone else run into this? Thanks in Advance.
>
> It helps if you post your exact error message.
------------------------------
From: [EMAIL PROTECTED] (Lew Pitcher)
Subject: Re: X Windows Clients
Reply-To: [EMAIL PROTECTED]
Date: Tue, 05 Jun 2001 13:03:38 GMT
On Tue, 5 Jun 2001 08:54:32 -0400, "Dave Armbruster"
<[EMAIL PROTECTED]> wrote:
>No error message other than the client reports the connection refused
>yet the same client attaches fine to HP Unix systems......Dave
On your RH7 system, before you start the X client on the remote
system, enter the xhost command
xhost +<name-or-IP-of-remote-system>
to allow your RH7 X server to accept X conversations from the remote
system. For better security, use the xauth command instead.
>> > I have installed Redhat 7 on a new server. Everything works great
>> >except I cannot access the system via an X Client. I can telnet, ftp,
>> >and rlogin no problem. I have network access enabled and even have
>> >run the xhost command to authorize the access. Redhat still refuses the
>> >network connection. Anyone else run into this? Thanks in Advance.
>>
>> It helps if you post your exact error message.
>
>
>
Lew Pitcher, Information Technology Consultant, Toronto Dominion Bank Financial Group
([EMAIL PROTECTED])
(Opinions expressed are my own, not my employer's.)
------------------------------
From: "Peter T. Breuer" <[EMAIL PROTECTED]>
Subject: Re: redhat rescue disk
Date: Tue, 5 Jun 2001 14:51:29 +0200
Yaniv Fine <[EMAIL PROTECTED]> wrote:
> how can i make linux Rescue disk ?
mke2fs /dev/fd0
mount /dev/fd0 /mnt/floppy
cd /mnt/floppy
mkdir boot etc dev lib bin
cp /boot/vmlinuz boot/
cp /boot/*.b boot/
cp /etc/lilo.conf etc/
vi etc/lilo.conf
cp -a /dev/hda* dev/
cp -a /dev/fd0 dev/
cp -a /lib/libc***.so lib/
...
awww shucks, why am I bothering? There must be tons of HOWTOs on this
around! You don't need the kernel on the floppy, btw. The one on your
disk will do if you still have access to it.
/usr/share/doc/HOWTO/en-txt/Bootdisk-HOWTO.txt.gz
Peter
------------------------------
From: "David Dorward" <[EMAIL PROTECTED]>
Subject: Re: partitions
Date: Tue, 05 Jun 2001 14:10:58 +0100
It seems that on Tue, 05 Jun 2001 10:13:51 +0100, someone claiming to be
"Mike91" <[EMAIL PROTECTED]> typed this:
> I'm planning to install linux for the 1st time, but I'm a little
> uncertain what I need to do about partitions....
>
> I am intending to empty a 6.5gig partition on my HD and install linux on
> it, however I have been reading that I will need a swap partition and
> maybe a boot partition, so I am unsure how to sub-divide the 6.5gig.
> Also, the 6.5gig is in FAT32, so I guess it would be better to change
> that to linux native? And if so, how?
>
> I am intending to install Mandrake 7.2 (as I have been able to borrow a
> copy) ....am I right in thinking the setup procedure will be able to do
> any partition manipulation I will need?
> cheers, Mike
Yes it will. Although I would go for Mandrake 8 or Red Hat 7.1, they
both come with a series 2.4 kernel by default which will save you a lot
of messing around.
You don't really need a boot partition with recent versions of lilo,
although I tend to use one out of habit. It should be able 10 Megs.
The swap partition is generally suggested to be double your RAM.
I would also have a seperate partitions for /home (where you keep all your
datafiles (like c:\my documents\) and user config info (like
c:\windows\profiles\). How big this needs to be depends on how much you
plan to put on the system.
Both Mandrake and Red Hat have partition editors and formatting tools as
part of the installer.
--
David Dorward http://david.us-lot.org/
The only way to keep your health is to eat what you don't want, drink
what you don't like, and do what you'd rather not. -- Mark Twain
------------------------------
From: Thomas Dickey <[EMAIL PROTECTED]>
Subject: Re: Ncurses and menuconfig
Date: 5 Jun 2001 13:17:28 GMT
James Fraser <[EMAIL PROTECTED]> wrote:
>>snip<
> make -C scripts/lxdialog all
> make[1]: Entering directory `/usr/src/linux-2.4.5/scripts/lxdialog'
> /usr/i386-slackware-linux/bin/ld: cannot open crt1.o: No such file or
crt1.o is the entrypoint for the C runtime (not related to ncurses)
--
Thomas E. Dickey <[EMAIL PROTECTED]> <[EMAIL PROTECTED]>
http://dickey.his.com
ftp://dickey.his.com
------------------------------
From: "Eric" <[EMAIL PROTECTED]>
Subject: Re: Lilo questions
Date: Tue, 5 Jun 2001 15:36:39 +0200
> >> 1) When I finished installing my PC, and boot, it displays "LI" and
then
> >> stops. Whether I put lilo in /dev/hda or /dev/hda1. (/dev/hda1 active
> >> partition or not).
> >> Why is this?
> >
> >the files needed to boot are not at the expected location.
>
> So should I create a small (+-10MB) /boot partition at the start of the
> disk? Will this solve the problem? It's a rather old machine, installing
the
> minimal system takes about 2 hours, so I'd rather get as much info as
> possible then do it.
>
not if it already stays under 1024 cylinders.
It won't change a thing.
Does `/sbin/lilo -v -v` give more info?
> >> 3) Should lilo be in /dev/hda or /dev/hda1 (generally speaking, and
why?)
> >
> >Whatever you prefer. I always put it in the MBR.
> >
> >> It's not the 1024 Cylinder problem, my disk doesn't have that many.
It's
> >an
> >> old, 512MB HDD (working condition, used it to boot another machine)
> >
> >What does fdisk -l /dev/hda tell you about the number of cylinders?
> >The size of the disk is no guarantee that the number of cylinders is less
> >than 1024. (eg. you would be using 16 heads/63 sectors per track)
>
> I know, it's got 512 cylinders according to fdisk.
do both fdisk and the BIOS agree on the geometry?
Eric
------------------------------
From: [EMAIL PROTECTED] (Trevor Hemsley)
Crossposted-To: comp.os.linux.hardware
Subject: Re: SCSI problems
Date: 05 Jun 2001 13:45:51 GMT
On Mon, 4 Jun 2001 15:46:39, John English <[EMAIL PROTECTED]> wrote:
> Thanks to those who replied, and to those who helped by private email.
> I seem to be suffering from a bad case of versionitis...
>
> Trevor Hemsley wrote:
> > There's a known problem with FAT formatted media on MO disks (or
> > anything which uses a 2K sector size) when run with a 2.4.x kernel.
>
> Eric DESHAYES wrote:
> > for writing cd, i have been told that you need the kernel 2.4.5(maybe
> > 2.4.4).
>
> Does anyone know whether these problems exist with earlier kernels?
> Is there anything in RH7.1 which *requires* a 2.4 kernel, or could
> I get away with a kernel "downgrade" to fix my problems?
>
> I've certainly used the same M-O drive model on another box, many
> moons ago (kernel version < 2, but I can't remember exactly) and
> am quite happily burning CDs on a RH6.2 setup (but using a SCSI
> burner)...
>
> Any thoughts, anyone?
It's certainly possible to run a newer kernel on a machine based on an
older distribution - as long as you update the bits that need updating
(see /usr/src/linux/Documentation/Changes). I'm not sure if you can
run an older kernel on a distribution that is specifically built
around a 2.4.x kernel though. Try it and see is about the best answer
- you can set up a new lilo stanza to boot a different kernel
alongside the existing one.
Does RH 7.1 not supply 2.2.x and 2.4.x kernels? I run SuSE 7.1 and
that gives you a choice of 2.2.18 and 2.4.something.
There are patches for the MO 2KB sector problem and I think they are
included in the latest 2.4.x-acn series of kernels which will be on
ftp.[country].kernel.org in an Alan cox specific directory somewhere.
Writing CD's does not require 2.4.5 - at least not generically. I've
written CD's on my SCSI Plextor 12/10/32 with lots of different kernel
releases up to and including 2.4.3. The ide-scsi emulation layer may
be a different thing to factor in but I do not remember reading that
it was terribly broken at any recent time.
--
Trevor Hemsley, Brighton, UK.
[EMAIL PROTECTED]
------------------------------
** 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 by posting to comp.os.linux.setup.
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-Setup Digest
******************************
