upgrade-reports: system hangs after woody->sarge upgrade

Bug #19848 reported by Debian Bug Importer
6
Affects Status Importance Assigned to Milestone
lilo (Debian)
Fix Released
Unknown
lilo (Ubuntu)
Invalid
High
Unassigned

Bug Description

Automatically imported from Debian bug report #316955 http://bugs.debian.org/316955

Revision history for this message
In , Steve Langasek (vorlon) wrote : reassign 316955 to lilo

# Automatically generated email from bts, devscripts version 2.8.14
reassign 316955 lilo

Revision history for this message
In , Adam Kessel (ajkessel) wrote : More LILO problesm

I've just spent several hours trying to boot with LILO or GRUB after the
woody->sarge upgrade, following all the various troubleshooting guides on
the web. LILO usually stops at L and gives error 99 repeatedly; GRUB
gives error 18.

I understand these are usually related to BIOS issues. I tried every
reported solution (short of reformatting and repartitioning the drives)
and nothing worked.

Then downgraded lilo back to woody, and the machine boots up fine.
--
Adam Rosi-Kessel
http://adam.rosi-kessel.org

Revision history for this message
In , Adam Kessel (ajkessel) wrote : More information re: lilo upgrade

I should also have mentioned that the old (woody) LILO reports the
following warning, while the new (sarge) LILO does not:

Int 0x13 Function 8 and Function 0x48 return different head/sector geometries for BIOS drive 0x83
--
Adam Rosi-Kessel
http://adam.rosi-kessel.org

Revision history for this message
In , ERIIX M. Blaike (eriix-subs) wrote : Re: Bug#316955: upgrade-reports: system hangs after woody->sarge upgrade

Victor Munoz wrote:

SNIP

>Naively maybe, I would have expected lilo to have been re-run if necessary
>during the upgrade. Maybe a mention of lilo issues in the Release Notes
>would have been welcome.
>
>
I believe it was. That's one of the things that happens during the lilo
configuration run that happens during the lilo installation. It just
failed to install correctly, which I've seen many times on my own
machines. The solution is to do just as you did until it actually boots
(which is usually the next time you try). I have very little
information on the boot process and the black magic that entails, but it
seems that the problem is in the copying, I suppose. Perhaps lilo
should do a sync, then read the mbr/boot sector to confirm the copy
succeeded? Any comments on that? Does it already and my thinking is off?

Revision history for this message
In , Francesco Poli (frx) wrote : Duplicate?

Victor Munoz wrote:

> I checked it. I did not upgrade kernel. I thought it was safe to reboot.
> But the system hang, with "LI" being the only text on screen.

Is this a duplicate of bug#306472 ?

--
    :-( This Universe is buggy! Where's the Creator's BTS? ;-)
......................................................................
  Francesco Poli GnuPG Key ID = DD6DFCF4
 Key fingerprint = C979 F34B 27CE 5CD8 DC12 31B5 78F4 279B DD6D FCF4

Revision history for this message
Debian Bug Importer (debzilla) wrote :

Automatically imported from Debian bug report #316955 http://bugs.debian.org/316955

Revision history for this message
Debian Bug Importer (debzilla) wrote :

Message-Id: <email address hidden>
Date: Tue, 05 Jul 2005 15:59:22 +0900
From: Victor Munoz <email address hidden>
To: Debian Bug Tracking System <email address hidden>
Subject: upgrade-reports: system hangs after woody->sarge upgrade

Package: upgrade-reports
Severity: critical
Justification: breaks the whole system

My system would not boot after the woody->sarge upgrade.

I followed the instructions in the Debian Release notes for sarge:

- edit sources.list

deb http://ftp.jp.debian.org/debian/ sarge main non-free contrib
deb http://security.debian.org/ sarge/updates main contrib non-free

- apt-get update
- aptitude install aptitude
- aptitude install doc-base
- aptitude -f --with-recommends dist-upgrade

The only problem until now was that I had to remove festival for the
upgrade to proceed (reported as a separate bug, #316948). The upgrade went
on apparently succesfully, installing and configuring packages as needed.
The release notes only ask to pay attention to the XFree upgrade notes
before rebooting:

/usr/share/doc/xfree86-common/README.Debian-upgrade.gz

I checked it. I did not upgrade kernel. I thought it was safe to reboot.
But the system hang, with "LI" being the only text on screen.

I used a woody bootable CD to check the system. The problem was solved
after I run 'lilo'. However, I understand that is only necessary if
lilo.conf is changed, or a new kernel is installed. Nothing of that happened.

$ ls -l lilo.conf
-rw-r----- 1 root root 4131 2004-02-02 20:04 lilo.conf

Checking the typescript of the upgrade session, I noticed that lilo was
updated:

Preparing to replace lilo 1:22.2-3 (using .../lilo_1%3a22.6.1-6.2_i386.deb)
...
Unpacking replacement lilo ...
Preparing to replace lilo-doc 1:22.2-3 (using
.../lilo-doc_1%3a22.6.1-6.2_all.deb) ...
Unpacking replacement lilo-doc ...

[...]

Setting up lilo (22.6.1-6.2) ...

Setting up lilo-doc (22.6.1-6.2) ...

Naively maybe, I would have expected lilo to have been re-run if necessary
during the upgrade. Maybe a mention of lilo issues in the Release Notes
would have been welcome.

Regards,
      Victor

-- System Information:
Debian Release: 3.1
Architecture: i386 (i686)
Kernel: Linux 2.4.18-bf2.4
Locale: LANG=es.ISO-8859-1, LC_CTYPE=es.ISO-8859-1 (charmap=ISO-8859-1) (ignored: LC_ALL set to en_US)

Revision history for this message
Debian Bug Importer (debzilla) wrote :

Message-Id: <email address hidden>
Date: Tue, 5 Jul 2005 03:24:22 -0700
From: Steve Langasek <email address hidden>
To: <email address hidden>
Subject: reassign 316955 to lilo

# Automatically generated email from bts, devscripts version 2.8.14
reassign 316955 lilo

Revision history for this message
Debian Bug Importer (debzilla) wrote :

Message-ID: <email address hidden>
Date: Tue, 5 Jul 2005 22:35:03 -0400
From: Adam Kessel <email address hidden>
To: <email address hidden>
Subject: More LILO problesm

I've just spent several hours trying to boot with LILO or GRUB after the
woody->sarge upgrade, following all the various troubleshooting guides on
the web. LILO usually stops at L and gives error 99 repeatedly; GRUB
gives error 18.

I understand these are usually related to BIOS issues. I tried every
reported solution (short of reformatting and repartitioning the drives)
and nothing worked.

Then downgraded lilo back to woody, and the machine boots up fine.
--
Adam Rosi-Kessel
http://adam.rosi-kessel.org

Revision history for this message
Debian Bug Importer (debzilla) wrote :

Message-ID: <email address hidden>
Date: Tue, 5 Jul 2005 22:39:01 -0400
From: Adam Kessel <email address hidden>
To: <email address hidden>
Subject: More information re: lilo upgrade

I should also have mentioned that the old (woody) LILO reports the
following warning, while the new (sarge) LILO does not:

Int 0x13 Function 8 and Function 0x48 return different head/sector geometries for BIOS drive 0x83
--
Adam Rosi-Kessel
http://adam.rosi-kessel.org

Revision history for this message
Debian Bug Importer (debzilla) wrote :

Message-ID: <email address hidden>
Date: Tue, 05 Jul 2005 20:57:23 -0700
From: ERIIX Blaike <email address hidden>
To: Victor Munoz <email address hidden>,
 <email address hidden>
Subject: Re: Bug#316955: upgrade-reports: system hangs after woody->sarge
 upgrade

Victor Munoz wrote:

SNIP

>Naively maybe, I would have expected lilo to have been re-run if necessary
>during the upgrade. Maybe a mention of lilo issues in the Release Notes
>would have been welcome.
>
>
I believe it was. That's one of the things that happens during the lilo
configuration run that happens during the lilo installation. It just
failed to install correctly, which I've seen many times on my own
machines. The solution is to do just as you did until it actually boots
(which is usually the next time you try). I have very little
information on the boot process and the black magic that entails, but it
seems that the problem is in the copying, I suppose. Perhaps lilo
should do a sync, then read the mbr/boot sector to confirm the copy
succeeded? Any comments on that? Does it already and my thinking is off?

Revision history for this message
Debian Bug Importer (debzilla) wrote :

Message-Id: <email address hidden>
Date: Thu, 7 Jul 2005 01:06:03 +0200
From: Francesco Poli <email address hidden>
To: <email address hidden>
Subject: Duplicate?

--Signature=_Thu__7_Jul_2005_01_06_03_+0200_he4IxUu0/r1lqniR
Content-Type: text/plain; charset=US-ASCII
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

Victor Munoz wrote:

> I checked it. I did not upgrade kernel. I thought it was safe to reboot.
> But the system hang, with "LI" being the only text on screen.

Is this a duplicate of bug#306472 ?

--=20
    :-( This Universe is buggy! Where's the Creator's BTS? ;-)
......................................................................
  Francesco Poli GnuPG Key ID =3D DD6DFCF4
 Key fingerprint =3D C979 F34B 27CE 5CD8 DC12 31B5 78F4 279B DD6D FCF4

--Signature=_Thu__7_Jul_2005_01_06_03_+0200_he4IxUu0/r1lqniR
Content-Type: application/pgp-signature

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)

iD8DBQFCzGPmePQnm91t/PQRAtVVAKC4ihy1UwqCLAOa89jygu4mF7lWlACeLlfA
dabPV7OuLphDNfeOZtdxfzE=
=/Y7J
-----END PGP SIGNATURE-----

--Signature=_Thu__7_Jul_2005_01_06_03_+0200_he4IxUu0/r1lqniR--

Revision history for this message
In , Paul Smith (psmith-nortel) wrote : Found why LILO failed for me ...

Last week I went through a woody -> sarge upgrade myself and got the
original problem reported here: the system booted to "LI" then stopped.

I went in with tomsrtbt (this system is headless AND has no CDROM, so I
can only boot off of a floppy and I have to tote around a monitor and
keyboard to plug into it... ugh!) and found the problem. Maybe this
will help someone fix the issue, since it seems too me there are LOTS of
reports of booting only to "LI" then hanging out there on the 'net:

This is what I found: in my /boot directory I had a symlink boot.b which
pointed to boot-menu.b... but that file does not exist:

    # ls -al /boot/boot*
    -rw-r--r-- 1 root root 512 Mar 22 2003 /boot/boot.0300
    lrwxrwxrwx 1 root root 11 Mar 22 2003 /boot/boot.b -> boot-menu.b

This was the cause of all my problems; apparently LILO was being
invoked, couldn't find the MBR file (I think that's what this is), and
somehow trashed my boot. I don't understand why LILO broke my system
when the file clearly doesn't exist, or why my upgrade succeeded with no
errors if LILO failed. I didn't even notice any problems with my system
until I tried to reboot it a week or so after I did the upgrade.

I also don't know where the file "boot-menu.b" came from or why I have a
symlink to it (I don't remember doing that myself but maybe I did: it's
been years since I messed with that box). I suppose whatever package it
was in got removed, but the symlink was not "fixed".

Anyway, once I fixed the symlink and worked out how to run LILO from
tomsrtbt (they use an old kernel so I couldn't chroot until I set an
environment variable to convince GLIBC to use the old kernel syscall
interface), my system came up OK again.

Hopefully this will help tracking down the problems.

--
-------------------------------------------------------------------------------
 Paul D. Smith <email address hidden> HASMAT: HA Software Mthds & Tools
 "Please remain calm...I may be mad, but I am a professional." --Mad Scientist
-------------------------------------------------------------------------------
        These are my opinions--Nortel takes no responsibility for them.

Revision history for this message
Debian Bug Importer (debzilla) wrote :

Message-ID: <email address hidden>
Date: Mon, 12 Sep 2005 14:31:39 -0400
From: "Paul Smith" <email address hidden>
To: <email address hidden>
Subject: Found why LILO failed for me ...

Last week I went through a woody -> sarge upgrade myself and got the
original problem reported here: the system booted to "LI" then stopped.

I went in with tomsrtbt (this system is headless AND has no CDROM, so I
can only boot off of a floppy and I have to tote around a monitor and
keyboard to plug into it... ugh!) and found the problem. Maybe this
will help someone fix the issue, since it seems too me there are LOTS of
reports of booting only to "LI" then hanging out there on the 'net:

This is what I found: in my /boot directory I had a symlink boot.b which
pointed to boot-menu.b... but that file does not exist:

    # ls -al /boot/boot*
    -rw-r--r-- 1 root root 512 Mar 22 2003 /boot/boot.0300
    lrwxrwxrwx 1 root root 11 Mar 22 2003 /boot/boot.b -> boot-menu.b

This was the cause of all my problems; apparently LILO was being
invoked, couldn't find the MBR file (I think that's what this is), and
somehow trashed my boot. I don't understand why LILO broke my system
when the file clearly doesn't exist, or why my upgrade succeeded with no
errors if LILO failed. I didn't even notice any problems with my system
until I tried to reboot it a week or so after I did the upgrade.

I also don't know where the file "boot-menu.b" came from or why I have a
symlink to it (I don't remember doing that myself but maybe I did: it's
been years since I messed with that box). I suppose whatever package it
was in got removed, but the symlink was not "fixed".

Anyway, once I fixed the symlink and worked out how to run LILO from
tomsrtbt (they use an old kernel so I couldn't chroot until I set an
environment variable to convince GLIBC to use the old kernel syscall
interface), my system came up OK again.

Hopefully this will help tracking down the problems.

--
-------------------------------------------------------------------------------
 Paul D. Smith <email address hidden> HASMAT: HA Software Mthds & Tools
 "Please remain calm...I may be mad, but I am a professional." --Mad Scientist
-------------------------------------------------------------------------------
        These are my opinions--Nortel takes no responsibility for them.

Revision history for this message
Carthik Sharma (carthik) wrote :

This is a Debian upgrade specific bug. Does not affect Ubuntu. Closing.

Changed in lilo:
status: Unconfirmed → Rejected
Revision history for this message
In , Marc 'HE' Brockschmidt (marc-marcbrockschmidt) wrote : Bug not present in etch

close 316955 1:22.6.1-7.1
reopen 316955
thanks

This bug is not relevant for sarge -> etch upgrades.

Marc
--
BOFH #26:
first Saturday after first full moon in Winter

Revision history for this message
In , Martin Zobel-Helas (zobel-spohr) wrote : closing 316955

close 316955

Revision history for this message
In , Martin Zobel-Helas (zobel) wrote :

# Automatically generated email from bts, devscripts version 2.9.21
close 316955 1:22.6.1-7.1

Changed in lilo:
status: Unconfirmed → Fix Released
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.