Re: unable to boot system! Why?

Re: unable to boot system! Why?

[igueths]

Hi Kenny. Assuming the libs won't work even after I chroot, can I just download 
the base system from the latest stable version onto a Cd, and then untar it 
from the chroot shell? As in chroot /target cd tar zxvf basefile.tgz? 
> Hi.  Not sure what caused your problem, but I'm almost certain that when
> you copied the potato boot disk's /lib to /target/lib you made things
> worse.
> The libraries on the install disk are stripped down versions.  The files
> you see in /lib after an install were unpacked from the base system.
> Also, the files in /lib have likely been updated since you first
> installed the system.  If you've dist-upgraded to Woody since the
> install most of /lib was replaced with newer versions.
> 
> The way to run the lilo on your installed system if you have booted from
> a rescue disk is to chroot into /target first.  That way, the system
> should use the corect lilo and shared libs.  Since you have already done
> the cp command, it might be to late.
> 
> If I'm wrong then hopefully someone will corect me.
>           Kenny
> 	  
> On Sun, Apr 06, 2003 at 08:32:17PM -0700, Igor Gueths wrote:
> > Hi Greg. Well, I can access all the data on the drive just fine. However, I
> > can't boot from the hard disk. Lilo just does what I described earlier in
> > the previous post. I also can't do anything to the already installed lilo
> > from the rescue floppies, because I get that error while loading shared
> > libraries. I do not have another Linux box to aid in this restoration
> > process. And restoring the libs from the rescue disk to the target /lib
> > didn't do anything. The files got coppied, but I got the same error. Does
> > loading shared libraries. ./Lilo: Symbol __cxa_atexit, version glibc_1.2.3,
> > not defined in file libc.so.6 with link time reference. I tried copying the
> > /lib directory from the rescue flopppy (root floppy) to /target/lib, and no
> > effect. I still got the same error. And this box still will not reboot. And
> > I do not want to install my entire system over again. Is there any way I can
> > restore Lilo/all the libs without having to re-install everything? Has
> > anyone encountered this error before and if so how did you fix it? Thanks!
> > 
> 
> _______________________________________________
> Speakup mailing list
> Speakup@braille.uwo.ca
> http://speech.braille.uwo.ca/mailman/listinfo/speakup

Re: unable to boot system! Why?

[Kenny Hitt]

Hi.  No, base-files is a package.  Try downloading the base system and
untaring it into an empty directory.  I'm not sure if base-files is all
that's included in the base system tar ball.  If the-files package is
all that's in a base system tar ball, you can try installing it with

dpkg -i base-files*.deb

This command assumes you are in the same directory as the base-files
package.


          Kenny
	  

n Mon, Apr 07, 2003 at 12:15:50PM +0000, igueths@attbi.com wrote:
> Hi Kenny. Assuming the libs won't work even after I chroot, can I just download 
> the base system from the latest stable version onto a Cd, and then untar it 
> from the chroot shell? As in chroot /target cd tar zxvf basefile.tgz? 
> > Hi.  Not sure what caused your problem, but I'm almost certain that when
> > you copied the potato boot disk's /lib to /target/lib you made things
> > worse.
> > The libraries on the install disk are stripped down versions.  The files
> > you see in /lib after an install were unpacked from the base system.
> > Also, the files in /lib have likely been updated since you first
> > installed the system.  If you've dist-upgraded to Woody since the
> > install most of /lib was replaced with newer versions.
> > 
> > The way to run the lilo on your installed system if you have booted from
> > a rescue disk is to chroot into /target first.  That way, the system
> > should use the corect lilo and shared libs.  Since you have already done
> > the cp command, it might be to late.
> > 
> > If I'm wrong then hopefully someone will corect me.
> >           Kenny
> > 	  
> > On Sun, Apr 06, 2003 at 08:32:17PM -0700, Igor Gueths wrote:
> > > Hi Greg. Well, I can access all the data on the drive just fine. However, I
> > > can't boot from the hard disk. Lilo just does what I described earlier in
> > > the previous post. I also can't do anything to the already installed lilo
> > > from the rescue floppies, because I get that error while loading shared
> > > libraries. I do not have another Linux box to aid in this restoration
> > > process. And restoring the libs from the rescue disk to the target /lib
> > > didn't do anything. The files got coppied, but I got the same error. Does
> > > loading shared libraries. ./Lilo: Symbol __cxa_atexit, version glibc_1.2.3,
> > > not defined in file libc.so.6 with link time reference. I tried copying the
> > > /lib directory from the rescue flopppy (root floppy) to /target/lib, and no
> > > effect. I still got the same error. And this box still will not reboot. And
> > > I do not want to install my entire system over again. Is there any way I can
> > > restore Lilo/all the libs without having to re-install everything? Has
> > > anyone encountered this error before and if so how did you fix it? Thanks!
> http://speech.braille.uwo.ca/mailman/listinfo/speakup

Re: unable to boot system! Why?

[igueths]

Hi Kenny. What I meant is that the base system tarball should contain the 
directory structure necessary 
i.e., /usr/bin/progname, /usr/local/bin/progname, /lib/libfile.so, etc etc. So 
what I meant was let's say I cd'ed to / within my chroot shell and typed tar 
zxvf basefile.tgz. It should theoretically extract to all appropriate 
directories including /lib and /usr? And if base-files isn't all includd in the 
tarball, do you know what other packages comprise it? Because what I figured I 
would do is extract the tarball to / and it should extract to all its 
directories contained in the structure. So my restore process would go 
something like this. 1. boot the rescue/root potato disks (this is all I have 
for the moment). 2. mount /dev/hda2 /target. 3. mount /dev/hda3 /target/usr. 4. 
chroot /target. 5. cd. 6. su -. 7. examine /etc/lilo.conf and make sure things 
are set up properly. 8. cd /. 9. tar zxvf basefile.tgz. 10. /sbin/lilo. 11. 
exit su and exit chroot shell. 12. umount /target/usr. 13. umount /target. 14. 
reboot with ctrl+alt+del. One more thing. I don't have access to another cdrw 
drive in order to transfer the base system to my Linux system. Does anyone know 
of a good file splitter for tat other os that will allow me to split a given 
file into 1.44 mb chunks? Bc I have about 50-100 or so floppies here so I can 
do that. And then I will just cat all the split files together once they are 
all coppied. The only reason I ask is because I know there are numerous file 
splitters out there, I just figured I would ask for recommendations. So does 
the above procedure work for restoring my system? Thanks all again for all the 
help! 
> Hi.  No, base-files is a package.  Try downloading the base system and
> untaring it into an empty directory.  I'm not sure if base-files is all
> that's included in the base system tar ball.  If the-files package is
> all that's in a base system tar ball, you can try installing it with
> 
> dpkg -i base-files*.deb
> 
> This command assumes you are in the same directory as the base-files
> package.
> 
> 
>           Kenny
> 	  
> 
> n Mon, Apr 07, 2003 at 12:15:50PM +0000, igueths@attbi.com wrote:
> > Hi Kenny. Assuming the libs won't work even after I chroot, can I just 
> download 
> > the base system from the latest stable version onto a Cd, and then untar it 
> > from the chroot shell? As in chroot /target cd tar zxvf basefile.tgz? 
> > > Hi.  Not sure what caused your problem, but I'm almost certain that when
> > > you copied the potato boot disk's /lib to /target/lib you made things
> > > worse.
> > > The libraries on the install disk are stripped down versions.  The files
> > > you see in /lib after an install were unpacked from the base system.
> > > Also, the files in /lib have likely been updated since you first
> > > installed the system.  If you've dist-upgraded to Woody since the
> > > install most of /lib was replaced with newer versions.
> > > 
> > > The way to run the lilo on your installed system if you have booted from
> > > a rescue disk is to chroot into /target first.  That way, the system
> > > should use the corect lilo and shared libs.  Since you have already done
> > > the cp command, it might be to late.
> > > 
> > > If I'm wrong then hopefully someone will corect me.
> > >           Kenny
> > > 	  
> > > On Sun, Apr 06, 2003 at 08:32:17PM -0700, Igor Gueths wrote:
> > > > Hi Greg. Well, I can access all the data on the drive just fine. However, 
> I
> > > > can't boot from the hard disk. Lilo just does what I described earlier in
> > > > the previous post. I also can't do anything to the already installed lilo
> > > > from the rescue floppies, because I get that error while loading shared
> > > > libraries. I do not have another Linux box to aid in this restoration
> > > > process. And restoring the libs from the rescue disk to the target /lib
> > > > didn't do anything. The files got coppied, but I got the same error. Does
> > > > loading shared libraries. ./Lilo: Symbol __cxa_atexit, version 
> glibc_1.2.3,
> > > > not defined in file libc.so.6 with link time reference. I tried copying 
> the
> > > > /lib directory from the rescue flopppy (root floppy) to /target/lib, and 
> no
> > > > effect. I still got the same error. And this box still will not reboot. 
> And
> > > > I do not want to install my entire system over again. Is there any way I 
> can
> > > > restore Lilo/all the libs without having to re-install everything? Has
> > > > anyone encountered this error before and if so how did you fix it? Thanks!
> > http://speech.braille.uwo.ca/mailman/listinfo/speakup
> 
> _______________________________________________
> Speakup mailing list
> Speakup@braille.uwo.ca
> http://speech.braille.uwo.ca/mailman/listinfo/speakup

Re: unable to boot system! Why?

[igueths]

Hi Adam. I could get into the system through some rescue disks I had created 
from speakup/disks/debia/potato. I ran fsck and oddly enough, that produced no 
errors at all. I can see all my data htere i.e., I can read textfile susing ae 
which is the editor on that rescue diskset. Fsck ran as if there were no errors 
in the filesystem. And I can't rerun Lilo, I got the error I described even 
before I coppied the libs over. And when I checked /target/lib, which is the 
lib directory on the hdd, I saw all the libs there that I had always seen. I 
didn't see anything that had been modified. Before I powered down the system, I 
checked ps and there was no process called updatedb. I actually set that to run 
at 11:00 every night, and the system went down at about 5:38 PM EDT. So 
therefore, I have no idea how the system slowed down like that and also no idea 
of how to account for that significant disk I/O. I believe what the problem is 
is that all my data is there, however lilo somehow killed itself. I thought of 
running liloconfig, however I can't do that right now as I don't have access to 
my box from here. Anyone have nay other ideas on how to fix this? Thanks again!
> Did you run fsck on your partition from the rescue disk?  Since you gave
> it a rather abrupt power-down, it probably needs to be checked and fixed
> if possible.  Start with "fsck -p /dev/hda2" where you put in the correct
> partition name.  If you have more than one partition, check them all.
> This instructs fsck to fix anything that is safe to fix without causing
> data loss.  If it exits with an error, you are going to have to run it and
> answer "yes" to the repair questions by hand, but there is a high
> possibility in this case that something has gotten damaged somewhere and
> it will be restored to a file with a garbage name in /lost+found.
> Assuming you can repair the damage to the filesystem, try to boot the hard
> drive from the rescue disk.  Better still, did you make a boot disk when
> you installed?  You should try to boot that once you've run fsck.  If you
> can get into the system from the boot disk, you may be able to re-run
> Lilo.  BTW, you probably shouldn't have copied the lib directory from the
> rescue disk as it's likely somewhat incomplete and you may end up
> reinstalling libraries.  I'd sure love to know what crashed on you.  Did
> you see a process called updatedb by any chance?  If so, it's best to let
> it run its course.  It's a package that often gets installed by default
> that indexes all the files on your system and lets you locate a file
> quickly from that index rather than having to do a full search of the hard
> drive.  You can say something like "locate lynx," and it will look in the
> index and tell you where lynx was at the time of its creation.  Most newly
> installed systems have updatedb run from cron daily.  Redhat, and possibly
> Debian, have something called anacron which will run missed cron jobs.
> Hope this helps.
> 
> 
> _______________________________________________
> Speakup mailing list
> Speakup@braille.uwo.ca
> http://speech.braille.uwo.ca/mailman/listinfo/speakup

unable to boot system! Why?

[Igor Gueths]

[-- Attachment #1: Type: text/plain, Size: 1118 bytes --]

Hi all. I recently was using my system, in fact I was working on some source code. All of a sudden, my system started slowing down majorly. I checked ps to see if there were any processes that were writing to disk, and I couldn't find any. Yet I still heard the disk being accessed, and the system kept on slowing down. I then powered it down and tried rebooting. However, the machine failed to reboot. When I did attempt to reboot it, it started printing endless strings of integers. And when I tried running lilo with the Potato rescue floppies, I got the following error: "lilo: error in loading shared libraries. ./Lilo: Symbol __cxa_atexit, version glibc_1.2.3, not defined in file libc.so.6 with link time reference. I tried copying the /lib directory from the rescue flopppy (root floppy) to /target/lib, and no effect. I still got the same error. And this box still will not reboot. And I do not want to install my entire system over again. Is there any way I can restore Lilo/all the libs without having to re-install everything? Has anyone encountered this error before and if so how did you fix it? Thanks! 

[-- Attachment #2: Type: text/html, Size: 1459 bytes --]

Re: unable to boot system! Why?

[Gregory Nowak]

I had a problem not exactly like yours, but with the same beginning symptoms.

I had just rebooted my server. As I logged in, the drive was making a whole bunch of noise, as if there was some big accessing going on. 
Also, when I was typing on the keyboard, speakup and the doubletalk had a bit of a delay in echoing the key strokes back, and commands took forever to execute.

Then I did reboot, and the system went through the post, but didn't boot.

I was also unable to boot from floppies or cd-rom, until I cleared CMOS, and reconfigured my BIOS. 
By the time I was able to boot from a floppy, I discovered that there was no partition table on my 2 drives which make up the raid 0 array.
It was as if I had a brand new drive.

Fortunately, I had a complete tarred backup of the server on another box, so for me, it was a matter
of getting networking working under zipslack running from a zip disk, and untarring the backup from the other machine to the server.

Hope you didn't have the same thing happen to you.

Greg


On Sun, Apr 06, 2003 at 07:43:49PM -0700, Igor Gueths wrote:
> Hi all. I recently was using my system, in fact I was working on some source code. All of a sudden, my system started slowing down majorly. I checked ps to see if there were any processes that were writing to disk, and I couldn't find any. Yet I still heard the disk being accessed, and the system kept on slowing down. I then powered it down and tried rebooting. However, the machine failed to reboot. When I did attempt to reboot it, it started printing endless strings of integers. And when I tried running lilo with the Potato rescue floppies, I got the following error: "lilo: error in loading shared libraries. ./Lilo: Symbol __cxa_atexit, version glibc_1.2.3, not defined in file libc.so.6 with link time reference. I tried copying the /lib directory from the rescue flopppy (root floppy) to /target/lib, and no effect. I still got the same error. And this box still will not reboot. And I do not want to install my entire system over again. Is there any way I can restore Lilo/all the libs without having to re-install everything? Has anyone encountered this error before and if so how did you fix it? Thanks!

Re: unable to boot system! Why?

[Igor Gueths]

Hi Greg. Well, I can access all the data on the drive just fine. However, I
can't boot from the hard disk. Lilo just does what I described earlier in
the previous post. I also can't do anything to the already installed lilo
from the rescue floppies, because I get that error while loading shared
libraries. I do not have another Linux box to aid in this restoration
process. And restoring the libs from the rescue disk to the target /lib
didn't do anything. The files got coppied, but I got the same error. Does
anyone have any other ideas as to what I should do next?
----- Original Message -----
From: "Gregory Nowak" <greg@romuald.net.eu.org>
To: <speakup@braille.uwo.ca>
Sent: Sunday, April 06, 2003 5:21 PM
Subject: Re: unable to boot system! Why?


I had a problem not exactly like yours, but with the same beginning
symptoms.

I had just rebooted my server. As I logged in, the drive was making a whole
bunch of noise, as if there was some big accessing going on.
Also, when I was typing on the keyboard, speakup and the doubletalk had a
bit of a delay in echoing the key strokes back, and commands took forever to
execute.

Then I did reboot, and the system went through the post, but didn't boot.

I was also unable to boot from floppies or cd-rom, until I cleared CMOS, and
reconfigured my BIOS.
By the time I was able to boot from a floppy, I discovered that there was no
partition table on my 2 drives which make up the raid 0 array.
It was as if I had a brand new drive.

Fortunately, I had a complete tarred backup of the server on another box, so
for me, it was a matter
of getting networking working under zipslack running from a zip disk, and
untarring the backup from the other machine to the server.

Hope you didn't have the same thing happen to you.

Greg


On Sun, Apr 06, 2003 at 07:43:49PM -0700, Igor Gueths wrote:
> Hi all. I recently was using my system, in fact I was working on some
source code. All of a sudden, my system started slowing down majorly. I
checked ps to see if there were any processes that were writing to disk, and
I couldn't find any. Yet I still heard the disk being accessed, and the
system kept on slowing down. I then powered it down and tried rebooting.
However, the machine failed to reboot. When I did attempt to reboot it, it
started printing endless strings of integers. And when I tried running lilo
with the Potato rescue floppies, I got the following error: "lilo: error in
loading shared libraries. ./Lilo: Symbol __cxa_atexit, version glibc_1.2.3,
not defined in file libc.so.6 with link time reference. I tried copying the
/lib directory from the rescue flopppy (root floppy) to /target/lib, and no
effect. I still got the same error. And this box still will not reboot. And
I do not want to install my entire system over again. Is there any way I can
restore Lilo/all the libs without having to re-install everything? Has
anyone encountered this error before and if so how did you fix it? Thanks!

_______________________________________________
Speakup mailing list
Speakup@braille.uwo.ca
http://speech.braille.uwo.ca/mailman/listinfo/speakup

Re: unable to boot system! Why?

[Adam Myrow]

Did you run fsck on your partition from the rescue disk?  Since you gave
it a rather abrupt power-down, it probably needs to be checked and fixed
if possible.  Start with "fsck -p /dev/hda2" where you put in the correct
partition name.  If you have more than one partition, check them all.
This instructs fsck to fix anything that is safe to fix without causing
data loss.  If it exits with an error, you are going to have to run it and
answer "yes" to the repair questions by hand, but there is a high
possibility in this case that something has gotten damaged somewhere and
it will be restored to a file with a garbage name in /lost+found.
Assuming you can repair the damage to the filesystem, try to boot the hard
drive from the rescue disk.  Better still, did you make a boot disk when
you installed?  You should try to boot that once you've run fsck.  If you
can get into the system from the boot disk, you may be able to re-run
Lilo.  BTW, you probably shouldn't have copied the lib directory from the
rescue disk as it's likely somewhat incomplete and you may end up
reinstalling libraries.  I'd sure love to know what crashed on you.  Did
you see a process called updatedb by any chance?  If so, it's best to let
it run its course.  It's a package that often gets installed by default
that indexes all the files on your system and lets you locate a file
quickly from that index rather than having to do a full search of the hard
drive.  You can say something like "locate lynx," and it will look in the
index and tell you where lynx was at the time of its creation.  Most newly
installed systems have updatedb run from cron daily.  Redhat, and possibly
Debian, have something called anacron which will run missed cron jobs.
Hope this helps.

Re: unable to boot system! Why?

[Kenny Hitt]

Hi.  Not sure what caused your problem, but I'm almost certain that when
you copied the potato boot disk's /lib to /target/lib you made things
worse.
The libraries on the install disk are stripped down versions.  The files
you see in /lib after an install were unpacked from the base system.
Also, the files in /lib have likely been updated since you first
installed the system.  If you've dist-upgraded to Woody since the
install most of /lib was replaced with newer versions.

The way to run the lilo on your installed system if you have booted from
a rescue disk is to chroot into /target first.  That way, the system
should use the corect lilo and shared libs.  Since you have already done
the cp command, it might be to late.

If I'm wrong then hopefully someone will corect me.
          Kenny
	  
On Sun, Apr 06, 2003 at 08:32:17PM -0700, Igor Gueths wrote:
> Hi Greg. Well, I can access all the data on the drive just fine. However, I
> can't boot from the hard disk. Lilo just does what I described earlier in
> the previous post. I also can't do anything to the already installed lilo
> from the rescue floppies, because I get that error while loading shared
> libraries. I do not have another Linux box to aid in this restoration
> process. And restoring the libs from the rescue disk to the target /lib
> didn't do anything. The files got coppied, but I got the same error. Does
> loading shared libraries. ./Lilo: Symbol __cxa_atexit, version glibc_1.2.3,
> not defined in file libc.so.6 with link time reference. I tried copying the
> /lib directory from the rescue flopppy (root floppy) to /target/lib, and no
> effect. I still got the same error. And this box still will not reboot. And
> I do not want to install my entire system over again. Is there any way I can
> restore Lilo/all the libs without having to re-install everything? Has
> anyone encountered this error before and if so how did you fix it? Thanks!
>