kernel-image-2.4.27-2-686: creates bad initrd with complex lvm2 setup

[Debian Bug Importer]

Automatically imported from Debian bug report #291386 http://bugs.debian.org/291386

[Simon Horman (horms)]

reassign 291386 initrd-tools
thanks

On Thu, Jan 20, 2005 at 02:17:34PM +0100, Eric Deplagne wrote:
> Package: kernel-image-2.4.27-2-686
> Version: 2.4.27-7
> Severity: critical
> Justification: breaks the whole system
>
> When the install of kernel-image tries to create initrd,
> those lines go to /var/log/messages:
>
> Jan 20 09:39:39 fregate kernel: device-mapper: unknown block ioctl 0x801c6d02
> Jan 20 09:39:40 fregate last message repeated 14 times
>
> I do have 15 logical volumes, which explains how many times the message is emitted...
>
> I guess I should mention a bit of my partitionning scheme, out of my /etc/fstab:
>
> # system
> /dev/mapper/vg-root / ext3 defaults,errors=remount-ro 0 1
> /dev/mapper/vg-usr_lib /usr/lib ext3 defaults 0 1
> /dev/mapper/vg-usr_share /usr/share ext3 defaults 0 1
> /dev/mapper/vg-doc /usr/share/doc ext3 defaults 0 1
> /dev/mapper/vg-locale /usr/share/locale ext3 defaults 0 1
> /dev/mapper/vg-cache_apt /var/cache/apt/archives ext3 defaults 0 2
>
> maybe my having /usr/lib on a different logical volume is relevant...
> just a guess of mine, of course...
>
> my other logical volumes are one for /home, and some for subdirectories of /home,
> so should not be relevant...
>
> Last thing I guess I can mention is that on another machine, with only one logical volume for /,
> and one for /home, /var/log/messages gets its two lines, but the system boots properly...
>
> you might notice I presently run a 2.4.27-1-686, this is my rescue that I had a hard time to setup
> when the problem first happend to me with the last update of 2.4.27-1-868,
> which really hang the computer, having overwritten my working kernel.
> (that did not happen this time because of the -1-686 to -2-686 move)
>
> -- System Information:
> Debian Release: 3.1
> APT prefers testing
> APT policy: (500, 'testing')
> Architecture: i386 (i686)
> Kernel: Linux 2.4.27-1-686
> Locale: LANG=fr_BE@euro, LC_CTYPE=fr_BE@euro (charmap=ISO-8859-15)
>
> Versions of packages kernel-image-2.4.27-2-686 depends on:
> ii coreutils [fileutils] 5.2.1-2 The GNU core utilities
> ii fileutils 5.2.1-2 The GNU file management utilities
> ii initrd-tools 0.1.76 tools to create initrd image for p
> ii modutils 2.4.26-1.2 Linux module utilities
>
> -- no debconf information
>
>
> --
> To UNSUBSCRIBE, email to <email address hidden>
> with a subject of "unsubscribe". Trouble? Contact <email address hidden>

--
Horms

[Debian Bug Importer (debzilla)]

Automatically imported from Debian bug report #291386 http://bugs.debian.org/291386

[Debian Bug Importer (debzilla)]

Message-Id: <email address hidden>
Date: Thu, 20 Jan 2005 14:17:34 +0100
From: Eric Deplagne <email address hidden>
To: Debian Bug Tracking System <email address hidden>
Subject: kernel-image-2.4.27-2-686: creates bad initrd with complex lvm2 setup

Package: kernel-image-2.4.27-2-686
Version: 2.4.27-7
Severity: critical
Justification: breaks the whole system

When the install of kernel-image tries to create initrd,
those lines go to /var/log/messages:

Jan 20 09:39:39 fregate kernel: device-mapper: unknown block ioctl 0x801c6d02
Jan 20 09:39:40 fregate last message repeated 14 times

I do have 15 logical volumes, which explains how many times the message is emitted...

I guess I should mention a bit of my partitionning scheme, out of my /etc/fstab:

# system
/dev/mapper/vg-root / ext3 defaults,errors=remount-ro 0 1
/dev/mapper/vg-usr_lib /usr/lib ext3 defaults 0 1
/dev/mapper/vg-usr_share /usr/share ext3 defaults 0 1
/dev/mapper/vg-doc /usr/share/doc ext3 defaults 0 1
/dev/mapper/vg-locale /usr/share/locale ext3 defaults 0 1
/dev/mapper/vg-cache_apt /var/cache/apt/archives ext3 defaults 0 2

maybe my having /usr/lib on a different logical volume is relevant...
just a guess of mine, of course...

my other logical volumes are one for /home, and some for subdirectories of /home,
so should not be relevant...

Last thing I guess I can mention is that on another machine, with only one logical volume for /,
and one for /home, /var/log/messages gets its two lines, but the system boots properly...

you might notice I presently run a 2.4.27-1-686, this is my rescue that I had a hard time to setup
when the problem first happend to me with the last update of 2.4.27-1-868,
which really hang the computer, having overwritten my working kernel.
(that did not happen this time because of the -1-686 to -2-686 move)

-- System Information:
Debian Release: 3.1
  APT prefers testing
  APT policy: (500, 'testing')
Architecture: i386 (i686)
Kernel: Linux 2.4.27-1-686
Locale: LANG=fr_BE@euro, LC_CTYPE=fr_BE@euro (charmap=ISO-8859-15)

Versions of packages kernel-image-2.4.27-2-686 depends on:
ii coreutils [fileutils] 5.2.1-2 The GNU core utilities
ii fileutils 5.2.1-2 The GNU file management utilities
ii initrd-tools 0.1.76 tools to create initrd image for p
ii modutils 2.4.26-1.2 Linux module utilities

-- no debconf information

[Debian Bug Importer (debzilla)]

Message-ID: <email address hidden>
Date: Fri, 21 Jan 2005 12:26:24 +0900
From: Horms <email address hidden>
To: Eric Deplagne <email address hidden>, <email address hidden>
Subject: Re: Bug#291386: kernel-image-2.4.27-2-686: creates bad initrd with complex lvm2 setup

reassign 291386 initrd-tools
thanks

On Thu, Jan 20, 2005 at 02:17:34PM +0100, Eric Deplagne wrote:
> Package: kernel-image-2.4.27-2-686
> Version: 2.4.27-7
> Severity: critical
> Justification: breaks the whole system
>
> When the install of kernel-image tries to create initrd,
> those lines go to /var/log/messages:
>
> Jan 20 09:39:39 fregate kernel: device-mapper: unknown block ioctl 0x801c6d02
> Jan 20 09:39:40 fregate last message repeated 14 times
>
> I do have 15 logical volumes, which explains how many times the message is emitted...
>
> I guess I should mention a bit of my partitionning scheme, out of my /etc/fstab:
>
> # system
> /dev/mapper/vg-root / ext3 defaults,errors=remount-ro 0 1
> /dev/mapper/vg-usr_lib /usr/lib ext3 defaults 0 1
> /dev/mapper/vg-usr_share /usr/share ext3 defaults 0 1
> /dev/mapper/vg-doc /usr/share/doc ext3 defaults 0 1
> /dev/mapper/vg-locale /usr/share/locale ext3 defaults 0 1
> /dev/mapper/vg-cache_apt /var/cache/apt/archives ext3 defaults 0 2
>
> maybe my having /usr/lib on a different logical volume is relevant...
> just a guess of mine, of course...
>
> my other logical volumes are one for /home, and some for subdirectories of /home,
> so should not be relevant...
>
> Last thing I guess I can mention is that on another machine, with only one logical volume for /,
> and one for /home, /var/log/messages gets its two lines, but the system boots properly...
>
> you might notice I presently run a 2.4.27-1-686, this is my rescue that I had a hard time to setup
> when the problem first happend to me with the last update of 2.4.27-1-868,
> which really hang the computer, having overwritten my working kernel.
> (that did not happen this time because of the -1-686 to -2-686 move)
>
> -- System Information:
> Debian Release: 3.1
> APT prefers testing
> APT policy: (500, 'testing')
> Architecture: i386 (i686)
> Kernel: Linux 2.4.27-1-686
> Locale: LANG=fr_BE@euro, LC_CTYPE=fr_BE@euro (charmap=ISO-8859-15)
>
> Versions of packages kernel-image-2.4.27-2-686 depends on:
> ii coreutils [fileutils] 5.2.1-2 The GNU core utilities
> ii fileutils 5.2.1-2 The GNU file management utilities
> ii initrd-tools 0.1.76 tools to create initrd image for p
> ii modutils 2.4.26-1.2 Linux module utilities
>
> -- no debconf information
>
>
> --
> To UNSUBSCRIBE, email to <email address hidden>
> with a subject of "unsubscribe". Trouble? Contact <email address hidden>

--
Horms

[Debian Bug Importer (debzilla)]

Message-ID: <email address hidden>
Date: Mon, 24 Jan 2005 17:30:56 +0900
From: Horms <email address hidden>
To: Eric Deplagne <email address hidden>, <email address hidden>
Cc: Debian Bug Tracking System <email address hidden>
Subject: Re: Bug#291386: kernel-image-2.4.27-2-686: creates bad initrd with complex lvm2 setup

reassign 291386 initrd-tools
thanks

On Thu, Jan 20, 2005 at 02:17:34PM +0100, Eric Deplagne wrote:
> Package: kernel-image-2.4.27-2-686
> Version: 2.4.27-7
> Severity: critical
> Justification: breaks the whole system
>
> When the install of kernel-image tries to create initrd,
> those lines go to /var/log/messages:
>
> Jan 20 09:39:39 fregate kernel: device-mapper: unknown block ioctl 0x801c6d02
> Jan 20 09:39:40 fregate last message repeated 14 times
>
> I do have 15 logical volumes, which explains how many times the message is emitted...
>
> I guess I should mention a bit of my partitionning scheme, out of my /etc/fstab:
>
> # system
> /dev/mapper/vg-root / ext3 defaults,errors=remount-ro 0 1
> /dev/mapper/vg-usr_lib /usr/lib ext3 defaults 0 1
> /dev/mapper/vg-usr_share /usr/share ext3 defaults 0 1
> /dev/mapper/vg-doc /usr/share/doc ext3 defaults 0 1
> /dev/mapper/vg-locale /usr/share/locale ext3 defaults 0 1
> /dev/mapper/vg-cache_apt /var/cache/apt/archives ext3 defaults 0 2
>
> maybe my having /usr/lib on a different logical volume is relevant...
> just a guess of mine, of course...
>
> my other logical volumes are one for /home, and some for subdirectories of /home,
> so should not be relevant...
>
> Last thing I guess I can mention is that on another machine, with only one logical volume for /,
> and one for /home, /var/log/messages gets its two lines, but the system boots properly...
>
> you might notice I presently run a 2.4.27-1-686, this is my rescue that I had a hard time to setup
> when the problem first happend to me with the last update of 2.4.27-1-868,
> which really hang the computer, having overwritten my working kernel.
> (that did not happen this time because of the -1-686 to -2-686 move)
>
> -- System Information:
> Debian Release: 3.1
> APT prefers testing
> APT policy: (500, 'testing')
> Architecture: i386 (i686)
> Kernel: Linux 2.4.27-1-686
> Locale: LANG=fr_BE@euro, LC_CTYPE=fr_BE@euro (charmap=ISO-8859-15)
>
> Versions of packages kernel-image-2.4.27-2-686 depends on:
> ii coreutils [fileutils] 5.2.1-2 The GNU core utilities
> ii fileutils 5.2.1-2 The GNU file management utilities
> ii initrd-tools 0.1.76 tools to create initrd image for p
> ii modutils 2.4.26-1.2 Linux module utilities
>
> -- no debconf information
>
>
> --
> To UNSUBSCRIBE, email to <email address hidden>
> with a subject of "unsubscribe". Trouble? Contact <email address hidden>

--
Horms

[Eric Deplagne (eric-deplagne-info)]

Hi...

  I'm comming with news that might explain things...

  First I've diffed the initrds on the two machines I talked about in my initial report...

  The one that breaks is "1" and the one that works is "2"...
  I attach the diff file...

  Another interesting detail is dpkg -l lvm\*...
  On 1 it gives me:

  un lvm <none> (no description available)
  un lvm-binaries <none> (no description available)
  ii lvm-common 1.5.17 The Logical Volume Manager for Linux (common files)
  ii lvm10 1.0.8-8 The Logical Volume Manager for Linux
  ii lvm2 2.00.24-1 The Linux Logical Volume Manager
  un lvm5 <none> (no description available)
  un lvm6 <none> (no description available)

  On 2 it gives me (I guess the french locales won't be a problem):

  un lvm <néant> (aucune description n'est disponible)
  un lvm-binaries <néant> (aucune description n'est disponible)
  ii lvm-common 1.5.17 The Logical Volume Manager for Linux (common files)
  un lvm10 <néant> (aucune description n'est disponible)
  ii lvm2 2.00.24-1 The Linux Logical Volume Manager
  un lvm5 <néant> (aucune description n'est disponible)
  un lvm6 <néant> (aucune description n'est disponible)

  Now I've a guess on my mistake...

  <after a little test>

  Good guess...

  After removing lvm10, reinstalling kernel-image does give a good initrd...

  I attach the diff between "2" and the new one "3" for info...
  I guess the diffs are easily explained by the differences between the two boxes...

  Now the problem seems clear to me... Remains to choose how to solve it...

--
  Eric Deplagne

[Debian Bug Importer (debzilla)]

Message-ID: <20050124175212.GA1488@fregate>
Date: Mon, 24 Jan 2005 18:52:12 +0100
From: Eric Deplagne <email address hidden>
To: <email address hidden>
Subject: Re: Bug#291386: kernel-image-2.4.27-2-686: creates bad initrd with complex lvm2 setup

--aT9PWwzfKXlsBJM1
Content-Type: multipart/mixed; boundary="i0/AhcQY5QxfSsSZ"
Content-Disposition: inline

--i0/AhcQY5QxfSsSZ
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

Hi...

  I'm comming with news that might explain things...

  First I've diffed the initrds on the two machines I talked about in my in=
itial report...

  The one that breaks is "1" and the one that works is "2"...
  I attach the diff file...

  Another interesting detail is dpkg -l lvm\*...
  On 1 it gives me:

  un lvm <none> (no description available)
  un lvm-binaries <none> (no description available)
  ii lvm-common 1.5.17 The Logical Volume Manager for Linux (comm=
on files)
  ii lvm10 1.0.8-8 The Logical Volume Manager for Linux
  ii lvm2 2.00.24-1 The Linux Logical Volume Manager
  un lvm5 <none> (no description available)
  un lvm6 <none> (no description available)

  On 2 it gives me (I guess the french locales won't be a problem):

  un lvm <n=E9ant> (aucune description n'est disponible)
  un lvm-binaries <n=E9ant> (aucune description n'est disponible)
  ii lvm-common 1.5.17 The Logical Volume Manager for Linux (comm=
on files)
  un lvm10 <n=E9ant> (aucune description n'est disponible)
  ii lvm2 2.00.24-1 The Linux Logical Volume Manager
  un lvm5 <n=E9ant> (aucune description n'est disponible)
  un lvm6 <n=E9ant> (aucune description n'est disponible)

  Now I've a guess on my mistake...

  <after a little test>

  Good guess...

  After removing lvm10, reinstalling kernel-image does give a good initrd...

  I attach the diff between "2" and the new one "3" for info...
  I guess the diffs are easily explained by the differences between the two=
 boxes...

  Now the problem seems clear to me... Remains to choose how to solve it...
 =20
--=20
  Eric Deplagne

--i0/AhcQY5QxfSsSZ
Content-Type: text/plain; charset=us-ascii
Content-Disposition: attachment; filename=diff_initrds
Content-Transfer-Encoding: quoted-printable

Binary files /mnt/initrd1/bin/awk and /mnt/initrd2/bin/awk differ
Only in /mnt/initrd2/bin: mkdir
Only in /mnt/initrd2/bin2: mkdir
File /mnt/initrd1/dev/console is a character special file while file /mnt/i=
nitrd2/dev/console is a character special file
Only in /mnt/initrd1/dev: lvm
File /mnt/initrd1/dev/null is a character special file while file /mnt/init=
rd2/dev/null is a character special file
Only in /mnt/initrd2/etc: lvm
diff -ru /mnt/initrd1/etc/modules.conf /mnt/initrd2/etc/modules.conf
--- /mnt/initrd1/etc/modules.conf 1970-01-01 01:00:00.000000000 +0100
+++ /mnt/initrd2/etc/modules.conf 1970-01-01 01:00:00.000000000 +0100
@@ -95,6 +95,24 @@
=20
 ### update-modules: end processing /etc/modutils/paths
=20
+### update-modules: start pr...

[Joey Hess (joeyh)]

# Automatically generated email from bts, devscripts version 2.8.10
retitle 291386 creates initrd that cannot use lvm 2 volumes if both lvm2 and lvm10 are installed

[Debian Bug Importer (debzilla)]

Message-Id: <email address hidden>
Date: Tue, 25 Jan 2005 15:50:10 -0500
From: Joey Hess <email address hidden>
To: <email address hidden>
Subject: retitle 291386 to creates initrd that cannot use lvm 2 volumes if both lvm2 and lvm10 are
 installed

# Automatically generated email from bts, devscripts version 2.8.10
retitle 291386 creates initrd that cannot use lvm 2 volumes if both lvm2 and lvm10 are installed

[Jeff Bailey (jbailey)]

Closing this bug

[Steve Langasek (vorlon)]

tags 291386 unreproducible moreinfo
thanks

Hi Eric,

I've tried to pin down this bug, but I find that the current version of
initrd-tools builds correct (matching) initrds whether or not the lvm10
package is installed. Is it possible that the broken initrd was built with
an old version of initrd-tools, or was somehow built on a system that did
not have lvm2 installed? Can you still recreate this broken initrd problem
if you install lvm10 on the system? What version of initrd-tools do you
currently have installed?

Thanks,
--
Steve Langasek
postmodern programmer

[Debian Bug Importer (debzilla)]

Message-ID: <email address hidden>
Date: Tue, 5 Apr 2005 11:28:22 -0700
From: Steve Langasek <email address hidden>
To: <email address hidden>
Subject: Re: creates initrd that cannot use lvm2 volumes if both lvm2 and lvm10 are installed

--Dxnq1zWXvFF0Q93v
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

tags 291386 unreproducible moreinfo
thanks

Hi Eric,

I've tried to pin down this bug, but I find that the current version of
initrd-tools builds correct (matching) initrds whether or not the lvm10
package is installed. Is it possible that the broken initrd was built with
an old version of initrd-tools, or was somehow built on a system that did
not have lvm2 installed? Can you still recreate this broken initrd problem
if you install lvm10 on the system? What version of initrd-tools do you
currently have installed?

Thanks,
--=20
Steve Langasek
postmodern programmer

--Dxnq1zWXvFF0Q93v
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: Digital signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.5 (GNU/Linux)

iD8DBQFCUtjCKN6ufymYLloRAk20AJ4n/hxZhHyOOWHzFmA4XRWEVZN/NwCgnLrC
Zs1d+hmKBsFT9DPYurfGf3M=
=INIm
-----END PGP SIGNATURE-----

--Dxnq1zWXvFF0Q93v--

[Eric Deplagne (eric-deplagne-info)]

On Tue, 05 Apr 2005 11:28:22 -0700, Steve Langasek wrote:
> tags 291386 unreproducible moreinfo
> thanks
>
> Hi Eric,
>
> I've tried to pin down this bug, but I find that the current version of
> initrd-tools builds correct (matching) initrds whether or not the lvm10
> package is installed. Is it possible that the broken initrd was built with
> an old version of initrd-tools, or was somehow built on a system that did
> not have lvm2 installed? Can you still recreate this broken initrd problem
> if you install lvm10 on the system? What version of initrd-tools do you
> currently have installed?
>
> Thanks,
> --
> Steve Langasek
> postmodern programmer

Hi,

  It does occur all the same for me with initrd-tools 0.1.77...

  Do you have the root filesystem on lvm2 ?

  I attach the output of "diff -rub /mnt/initrd1 /mnt/initrd2",
  with on /mnt/initrd1 a mount of the initrd without lvm10,
  and on /mnt/initrd2 a mount of the initrd with lvm10

  I also verified that the initrd with lvm10 still does not boot my box properly...

Regards.

--
  Eric Deplagne

[Debian Bug Importer (debzilla)]

Message-ID: <20050407100530.GA1943@fregate>
Date: Thu, 7 Apr 2005 12:05:31 +0200
From: Eric Deplagne <email address hidden>
To: <email address hidden>
Cc: Steve Langasek <email address hidden>
Subject: Re: Bug#291386: creates initrd that cannot use lvm2 volumes if both lvm2 and lvm10 are
 installed

--6sX45UoQRIJXqkqR
Content-Type: multipart/mixed; boundary="lrZ03NoBR/3+SXJZ"
Content-Disposition: inline

--lrZ03NoBR/3+SXJZ
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Tue, 05 Apr 2005 11:28:22 -0700, Steve Langasek wrote:
> tags 291386 unreproducible moreinfo
> thanks
>=20
> Hi Eric,
>=20
> I've tried to pin down this bug, but I find that the current version of
> initrd-tools builds correct (matching) initrds whether or not the lvm10
> package is installed. Is it possible that the broken initrd was built wi=
th
> an old version of initrd-tools, or was somehow built on a system that did
> not have lvm2 installed? Can you still recreate this broken initrd probl=
em
> if you install lvm10 on the system? What version of initrd-tools do you
> currently have installed?
>=20
> Thanks,
> --=20
> Steve Langasek
> postmodern programmer

Hi,

  It does occur all the same for me with initrd-tools 0.1.77...

  Do you have the root filesystem on lvm2 ?

  I attach the output of "diff -rub /mnt/initrd1 /mnt/initrd2",
  with on /mnt/initrd1 a mount of the initrd without lvm10,
  and on /mnt/initrd2 a mount of the initrd with lvm10

  I also verified that the initrd with lvm10 still does not boot my box pro=
perly...

Regards.
=20
--=20
  Eric Deplagne

--lrZ03NoBR/3+SXJZ
Content-Type: text/plain; charset=us-ascii
Content-Disposition: attachment; filename=diff_initrds_new

Only in /mnt/initrd1/bin: mkdir
Only in /mnt/initrd1/bin2: mkdir
diff: /mnt/initrd1/dev/cciss: No such file or directory
diff: /mnt/initrd2/dev/cciss: No such file or directory
File /mnt/initrd1/dev/console is a character special file while file /mnt/initrd2/dev/console is a character special file
diff: /mnt/initrd1/dev/ida: No such file or directory
diff: /mnt/initrd2/dev/ida: No such file or directory
diff: /mnt/initrd1/dev/ide: No such file or directory
diff: /mnt/initrd2/dev/ide: No such file or directory
Only in /mnt/initrd2/dev: lvm
diff: /mnt/initrd1/dev/mapper: No such file or directory
diff: /mnt/initrd2/dev/mapper: No such file or directory
diff: /mnt/initrd1/dev/md: No such file or directory
diff: /mnt/initrd2/dev/md: No such file or directory
File /mnt/initrd1/dev/null is a character special file while file /mnt/initrd2/dev/null is a character special file
diff: /mnt/initrd1/dev/scsi: No such file or directory
diff: /mnt/initrd2/dev/scsi: No such file or directory
diff: /mnt/initrd1/dev/vg: No such file or directory
diff: /mnt/initrd2/dev/vg: No such file or directory
Only in /mnt/initrd1/etc: lvm
Only in /mnt/initrd1/lib: libdevmapper.so.1.01
Only in /mnt/initrd1/lib: libdl.so.2
Only in /mnt/initrd2/lib: liblvm-10.so.1
Only in /mnt/initrd2/lib: lvm-10
Only in /mnt/initrd1/lib: lvm-200
diff -rub /mnt/initrd1/loadmodules /mnt/initrd2/loadmodules
--- /mnt/initrd1/loadmodules 1970-01-01 01:00:00.000...

[Steve Langasek (vorlon)]

# Automatically generated email from bts, devscripts version 2.8.14
tags 291386 - unreproducible moreinfo

[Steve Langasek (vorlon)]

On Thu, Apr 07, 2005 at 12:05:31PM +0200, Eric Deplagne wrote:
> On Tue, 05 Apr 2005 11:28:22 -0700, Steve Langasek wrote:

> > I've tried to pin down this bug, but I find that the current version of
> > initrd-tools builds correct (matching) initrds whether or not the lvm10
> > package is installed. Is it possible that the broken initrd was built with
> > an old version of initrd-tools, or was somehow built on a system that did
> > not have lvm2 installed? Can you still recreate this broken initrd problem
> > if you install lvm10 on the system? What version of initrd-tools do you
> > currently have installed?

> It does occur all the same for me with initrd-tools 0.1.77...

Hmm.

> Do you have the root filesystem on lvm2 ?

Yes, of course; there's no reason to expect any lvm support in the initrd
otherwise.

> I attach the output of "diff -rub /mnt/initrd1 /mnt/initrd2",
> with on /mnt/initrd1 a mount of the initrd without lvm10,
> and on /mnt/initrd2 a mount of the initrd with lvm10

> I also verified that the initrd with lvm10 still does not boot my box
> properly...

Ok, got it -- this is only a problem with 2.4 kernels, where lvm-mod.o is
available, and did not happen when testing against a 2.6 kernel.

Now to figure out how to fix the check, so that it can distinguish between a
volume that requires lvm2 and one that works with lvm1.

Thanks,
--
Steve Langasek
postmodern programmer

[Debian Bug Importer (debzilla)]

Message-ID: <email address hidden>
Date: Thu, 7 Apr 2005 21:59:40 -0700
From: Steve Langasek <email address hidden>
To: Eric Deplagne <email address hidden>
Cc: <email address hidden>
Subject: Re: Bug#291386: creates initrd that cannot use lvm2 volumes if both lvm2 and lvm10 are
 installed

--RIYY1s2vRbPFwWeW
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Thu, Apr 07, 2005 at 12:05:31PM +0200, Eric Deplagne wrote:
> On Tue, 05 Apr 2005 11:28:22 -0700, Steve Langasek wrote:

> > I've tried to pin down this bug, but I find that the current version of
> > initrd-tools builds correct (matching) initrds whether or not the lvm10
> > package is installed. Is it possible that the broken initrd was built =
with
> > an old version of initrd-tools, or was somehow built on a system that d=
id
> > not have lvm2 installed? Can you still recreate this broken initrd pro=
blem
> > if you install lvm10 on the system? What version of initrd-tools do you
> > currently have installed?

> It does occur all the same for me with initrd-tools 0.1.77...

Hmm.

> Do you have the root filesystem on lvm2 ?

Yes, of course; there's no reason to expect any lvm support in the initrd
otherwise.

> I attach the output of "diff -rub /mnt/initrd1 /mnt/initrd2",
> with on /mnt/initrd1 a mount of the initrd without lvm10,
> and on /mnt/initrd2 a mount of the initrd with lvm10

> I also verified that the initrd with lvm10 still does not boot my box
> properly...

Ok, got it -- this is only a problem with 2.4 kernels, where lvm-mod.o is
available, and did not happen when testing against a 2.6 kernel.

Now to figure out how to fix the check, so that it can distinguish between a
volume that requires lvm2 and one that works with lvm1.

Thanks,
--=20
Steve Langasek
postmodern programmer

--RIYY1s2vRbPFwWeW
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: Digital signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.5 (GNU/Linux)

iD8DBQFCVg+3KN6ufymYLloRAqP+AKCNTN506j1keMXc3eqJ6Ve4gCMOAwCgulyb
yIRXBrL2Pyk3ES5hTx+v3Lw=
=Mrkt
-----END PGP SIGNATURE-----

--RIYY1s2vRbPFwWeW--

[Debian Bug Importer (debzilla)]

Message-Id: <email address hidden>
Date: Thu, 7 Apr 2005 21:59:26 -0700
From: Steve Langasek <email address hidden>
To: <email address hidden>
Subject: tagging 291386

# Automatically generated email from bts, devscripts version 2.8.14
tags 291386 - unreproducible moreinfo

[Steve Langasek (vorlon)]

tags 291386 patch
thanks

Hi all,

After poking at this, I believe the attached patch represents the best
available heuristic for choosing lvm10 vs. lvm2 for the initrd, given that
(AIUI) there is no apparent way to query the type of an individual volume
from userspace.

Cheers,
--
Steve Langasek
postmodern programmer

[Debian Bug Importer (debzilla)]

Message-ID: <email address hidden>
Date: Fri, 8 Apr 2005 05:17:04 -0700
From: Steve Langasek <email address hidden>
To: <email address hidden>
Subject: Re: creates initrd that cannot use lvm 2 volumes if both lvm2 and lvm10 are installed

--LyciRD1jyfeSSjG0
Content-Type: multipart/mixed; boundary="VS++wcV0S1rZb1Fb"
Content-Disposition: inline

--VS++wcV0S1rZb1Fb
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

tags 291386 patch
thanks

Hi all,

After poking at this, I believe the attached patch represents the best
available heuristic for choosing lvm10 vs. lvm2 for the initrd, given that
(AIUI) there is no apparent way to query the type of an individual volume
=66rom userspace.

Cheers,
--=20
Steve Langasek
postmodern programmer

--VS++wcV0S1rZb1Fb
Content-Type: text/plain; charset=us-ascii
Content-Disposition: attachment; filename="initrd-tools-291386.diff"
Content-Transfer-Encoding: quoted-printable

diff -Nru /var/tmp/FR9bz2IRIA/initrd-tools-0.1.77/debian/changelog /var/tmp=
/cb9cg6WG67/initrd-tools-0.1.77/debian/changelog
--- /var/tmp/FR9bz2IRIA/initrd-tools-0.1.77/debian/changelog 2005-01-23 10:=
37:41.000000000 -0800
+++ /var/tmp/cb9cg6WG67/initrd-tools-0.1.77/debian/changelog 2005-04-08 05:=
08:42.000000000 -0700
@@ -1,3 +1,15 @@
+initrd-tools (0.1.77-0.1) unstable; urgency=3Dhigh
+
+ * Non-maintainer upload.
+ * High-urgency upload for sarge-targetted RC bugfix
+ * Steve Langasek
+ - Don't be so hasty to use lvm10: only use it if lvmiopversion
+ says we're lvm10-only, *or* there's no dm-mod for the kernel;
+ this avoids breaking lvm2 volumes in 2.4 when lvm10 is installed.
+ Closes: #291386.
+
+ -- Steve Langasek <email address hidden> Fri, 8 Apr 2005 05:06:21 -0700
+
 initrd-tools (0.1.77) unstable; urgency=3Dhigh
=20
   * Martin Michlmayr
diff -Nru /var/tmp/FR9bz2IRIA/initrd-tools-0.1.77/mkinitrd /var/tmp/cb9cg6W=
G67/initrd-tools-0.1.77/mkinitrd
--- /var/tmp/FR9bz2IRIA/initrd-tools-0.1.77/mkinitrd 2005-01-23 10:37:41.00=
0000000 -0800
+++ /var/tmp/cb9cg6WG67/initrd-tools-0.1.77/mkinitrd 2005-04-08 05:08:52.00=
0000000 -0700
@@ -266,7 +266,8 @@
  fi
  LVM=3Dyes
=20
- if module_exists drivers/md/lvm-mod && [ -d /lib/lvm-10 ]; then
+ if ([ "$iopver" -lt 200 ] || ! module_exists drivers/md/dm-mod) \
+ && module_exists drivers/md/lvm-mod && [ -d /lib/lvm-10 ]; then
   echo lvm-mod
   echo /dev/lvm >&4
   {

--VS++wcV0S1rZb1Fb--

--LyciRD1jyfeSSjG0
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: Digital signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.5 (GNU/Linux)

iD8DBQFCVnY8KN6ufymYLloRAuFhAKC+6INxjW9ZPq/EYSct2qQjwc446gCgp2Ds
gAMSdPAgMmQ5LqB+fo+JJQ4=
=oqPw
-----END PGP SIGNATURE-----

--LyciRD1jyfeSSjG0--

[Maks Attems (debian-sternwelten)]

tags 291386 pending
tags 296894 pending
tags 306219 pending
tags 266591 pending
tags 298727 pending
tags 303401 pending
stop

[Steve Langasek (vorlon)]

tag 266591 + fixed
tag 279382 + fixed
tag 285767 + fixed
tag 290329 + fixed
tag 291386 + fixed
tag 292080 + fixed
tag 295412 + fixed
tag 295422 + fixed
tag 296894 + fixed
tag 297724 + fixed
tag 298727 + fixed
tag 301455 + fixed
tag 303281 + fixed
tag 303401 + fixed
tag 306219 + fixed

quit

This message was generated automatically in response to a
non-maintainer upload. The .changes file follows.

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Format: 1.7
Date: Fri, 13 May 2005 21:25:52 -0700
Source: initrd-tools
Binary: initrd-tools
Architecture: source all
Version: 0.1.79
Distribution: unstable
Urgency: high
Maintainer: Debian kernel team <email address hidden>
Changed-By: Steve Langasek <email address hidden>
Description:
 initrd-tools - tools to create initrd image for prepackaged Linux kernel
Closes: 266591 279382 285767 290329 291386 292080 295412 295422 296894 297724 298727 301455 303281 303401 306219
Changes:
 initrd-tools (0.1.79) unstable; urgency=high
 .
   * Steve Langasek
     - Don't be so hasty to use lvm10: only use it if lvmiopversion
       says we're lvm10-only, *or* there's no dm-mod for the kernel;
       this avoids breaking lvm2 volumes in 2.4 when the lvm10 package
       is installed. Closes: #291386.
     - Add myself to uploaders.
   * GOTO Masanori
     - Make mkinitrd work with new ldd format which change is introduced
       in glibc 2.3.4. (Closes: #301455, #303281)
       This change also fixes amd64 mkinitrd breakage.
       (Closes: #279382, #292080, #295412, #295422, #297724)
   * Maximilian Attems
     - Resynchronise with Ubuntu:
       * Error message should refer to mdadm, not raidtools2.
       * Take /etc/mkinitrd/DSDT and put it in the initrd if it exists.
       * Use mktemp, not $$ for creating temporary directories.
     - Function module_exist recognizes modules in subdirs. Closes: #296894
     - Fix upgrade path from woody for sparc arch. thanks to
       Jurij Smakov <email address hidden> Closes: #306219
     - Use libdevmapper1.01 Closes: #266591, #298727, #303401
     - Fix permissions of the initrd files. thanks to patch for debian/rules
       from Horms <email address hidden> Closes: #285767
   * Sven Luther
     - Powerpc now defaults to MODULES=dep. (Closes: #290329)
Files:
 d55fad2edd46b490476c6e640037e622 608 utils optional initrd-tools_0.1.79.dsc
 048265f342b32440466b256296878af8 27103 utils optional initrd-tools_0.1.79.tar.gz
 6180afbac9c9516ec2ce696b4c8b237d 29558 utils optional initrd-tools_0.1.79_all.deb

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

iD8DBQFChYpTKN6ufymYLloRArmxAJ96ibimA1H8yQI3sjosX6nQyoRCngCglJ9/
NJ1vAGYLxonkH4h3dY1TPis=
=gt9k
-----END PGP SIGNATURE-----

[Steve Langasek (vorlon)]

These bugs were all closed in the most recent initrd-tools upload, but owing
to pilot error, were not marked as such. The original closure message was
as follows.

Cheers,
--
Steve Langasek
postmodern programmer

----- Forwarded message from Steve Langasek <email address hidden> -----

X-Original-To: <email address hidden>
From: Steve Langasek <email address hidden>
To: <email address hidden>
Cc: Steve Langasek <email address hidden>,
 Debian kernel team <email address hidden>
X-Katie: $Revision: 1.55 $
Subject: Fixed in NMU of initrd-tools 0.1.79
Date: Sat, 14 May 2005 01:32:13 -0400
X-Spam-Checker-Version: SpamAssassin 3.0.2 (2004-11-16) on norad.dodds.net
X-Spam-Level:
X-Spam-Status: No, score=-2.6 required=5.0 tests=BAYES_00 autolearn=ham
 version=3.0.2

tag 266591 + fixed
tag 279382 + fixed
tag 285767 + fixed
tag 290329 + fixed
tag 291386 + fixed
tag 292080 + fixed
tag 295412 + fixed
tag 295422 + fixed
tag 296894 + fixed
tag 297724 + fixed
tag 298727 + fixed
tag 301455 + fixed
tag 303281 + fixed
tag 303401 + fixed
tag 306219 + fixed

quit

This message was generated automatically in response to a
non-maintainer upload. The .changes file follows.

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Format: 1.7
Date: Fri, 13 May 2005 21:25:52 -0700
Source: initrd-tools
Binary: initrd-tools
Architecture: source all
Version: 0.1.79
Distribution: unstable
Urgency: high
Maintainer: Debian kernel team <email address hidden>
Changed-By: Steve Langasek <email address hidden>
Description:
 initrd-tools - tools to create initrd image for prepackaged Linux kernel
Closes: 266591 279382 285767 290329 291386 292080 295412 295422 296894 297724 298727 301455 303281 303401 306219
Changes:
 initrd-tools (0.1.79) unstable; urgency=high
 .
   * Steve Langasek
     - Don't be so hasty to use lvm10: only use it if lvmiopversion
       says we're lvm10-only, *or* there's no dm-mod for the kernel;
       this avoids breaking lvm2 volumes in 2.4 when the lvm10 package
       is installed. Closes: #291386.
     - Add myself to uploaders.
   * GOTO Masanori
     - Make mkinitrd work with new ldd format which change is introduced
       in glibc 2.3.4. (Closes: #301455, #303281)
       This change also fixes amd64 mkinitrd breakage.
       (Closes: #279382, #292080, #295412, #295422, #297724)
   * Maximilian Attems
     - Resynchronise with Ubuntu:
       * Error message should refer to mdadm, not raidtools2.
       * Take /etc/mkinitrd/DSDT and put it in the initrd if it exists.
       * Use mktemp, not $$ for creating temporary directories.
     - Function module_exist recognizes modules in subdirs. Closes: #296894
     - Fix upgrade path from woody for sparc arch. thanks to
       Jurij Smakov <email address hidden> Closes: #306219
     - Use libdevmapper1.01 Closes: #266591, #298727, #303401
     - Fix permissions of the initrd files. thanks to patch for debian/rules
       from Horms <email address hidden> Closes: #285767
   * Sven Luther
     - Powerpc now defaults to MODULES=dep. (Closes: #290329)
Files:
 d55fad2edd46b490476c6e640037e622 608 utils optional initrd-tools_0.1.79.dsc
 048265f342b32440466b256296878af8 27103 utils optional initrd-tools_0.1.79.tar.gz...