/boot/initrd.img*should be removed from filesystem.squashfs

Bug #64887 reported by Mantas Kriaučiūnas
2
Affects Status Importance Assigned to Milestone
livecd-rootfs
Invalid
Undecided
Unassigned
Baltix
New
Undecided
Unassigned
livecd-rootfs (Ubuntu)
Fix Released
Undecided
Colin Watson

Bug Description

/boot/initrd.img*should be removed from filesystem.squashfs - now it takes about 7 MB space in CD, because it can't be compressed.

from #ubuntu-devel irc log:
Oct 09 15:02:10 Kamion tfheen: ubiquity does call update-initramfs, so removing /boot/initrd* is probably ok

Revision history for this message
Matt Zimmerman (mdz) wrote : Re: [Bug 64887] /boot/initrd.img*should be removed from filesystem.squashfs

On Mon, Oct 09, 2006 at 05:46:53PM -0000, Mantas Kriaučiūnas wrote:
> Public bug reported:
>
> /boot/initrd.img*should be removed from filesystem.squashfs - now it
> takes about 7 MB space in CD, because it can't be compressed.
>
> from #ubuntu-devel irc log:
> Oct 09 15:02:10 Kamion tfheen: ubiquity does call update-initramfs, so removing /boot/initrd* is probably ok

I thought we were already doing this; are we only removing the kernel itself
and not the initramfs?

Colin, does ubiquity call update-initramfs unconditionally? i.e., is that
code already being thoroughly tested? Are we sure it does the right thing if
no initramfs already exists? If it's predictable and safe, we can do it,
otherwise it should wait until after the release.

--
 - mdz

Revision history for this message
Colin Watson (cjwatson) wrote :

On Tue, Oct 10, 2006 at 02:02:06AM -0000, Matt Zimmerman wrote:
> On Mon, Oct 09, 2006 at 05:46:53PM -0000, Mantas Kriaučiūnas wrote:
> > /boot/initrd.img*should be removed from filesystem.squashfs - now it
> > takes about 7 MB space in CD, because it can't be compressed.
> >
> > from #ubuntu-devel irc log:
> > Oct 09 15:02:10 Kamion tfheen: ubiquity does call update-initramfs, so removing /boot/initrd* is probably ok
>
> I thought we were already doing this; are we only removing the kernel itself
> and not the initramfs?

I hope we're not removing the kernel! Ubiquity needs to copy it or the
resulting system won't boot.

> Colin, does ubiquity call update-initramfs unconditionally? i.e., is
> that code already being thoroughly tested?

Yes, it's called unconditionally. The most recent iteration of that code
landed on 16th September, so has been thoroughly tested in the beta.

> Are we sure it does the right thing if no initramfs already exists?

Yes; I just tested with a version of ubiquity modified to remove the
copied initramfs just before calling 'update-initramfs -u', and it
worked fine.

Revision history for this message
Matt Zimmerman (mdz) wrote :

On Tue, Oct 10, 2006 at 03:09:43PM -0000, Colin Watson wrote:
> On Tue, Oct 10, 2006 at 02:02:06AM -0000, Matt Zimmerman wrote:
> > On Mon, Oct 09, 2006 at 05:46:53PM -0000, Mantas Kriaučiūnas wrote:
> > > /boot/initrd.img*should be removed from filesystem.squashfs - now it
> > > takes about 7 MB space in CD, because it can't be compressed.
> > >
> > > from #ubuntu-devel irc log:
> > > Oct 09 15:02:10 Kamion tfheen: ubiquity does call update-initramfs, so removing /boot/initrd* is probably ok
> >
> > I thought we were already doing this; are we only removing the kernel itself
> > and not the initramfs?
>
> I hope we're not removing the kernel! Ubiquity needs to copy it or the
> resulting system won't boot.

Right, of course.

> > Colin, does ubiquity call update-initramfs unconditionally? i.e., is
> > that code already being thoroughly tested?
>
> Yes, it's called unconditionally. The most recent iteration of that code
> landed on 16th September, so has been thoroughly tested in the beta.
>
> > Are we sure it does the right thing if no initramfs already exists?
>
> Yes; I just tested with a version of ubiquity modified to remove the
> copied initramfs just before calling 'update-initramfs -u', and it
> worked fine.

OK, let's go ahead then.

--
 - mdz

Adam Conrad (adconrad)
Changed in ubuntu-cdimage:
assignee: nobody → adconrad
Revision history for this message
Adam Conrad (adconrad) wrote :

livecd.sh removes the initrd from the filesystem before compressing it now, so this should be fixed on the next daily (or manual) run.

Changed in ubuntu-cdimage:
status: Unconfirmed → Fix Released
Revision history for this message
Mantas Kriaučiūnas (mantas) wrote :

I found /boot/initrd.img.bak in 7.10 beta compressed filesystem - it seems livecd.sh doesn't remove /boot/initrd.img* :(

Changed in ubuntu-cdimage:
status: Fix Released → New
Matt Zimmerman (mdz)
Changed in ubuntu-cdimage:
assignee: adconrad → nobody
Revision history for this message
Colin Watson (cjwatson) wrote :

Moving to the livecd-rootfs package in Ubuntu.

Changed in livecd-rootfs:
status: New → Invalid
Revision history for this message
Colin Watson (cjwatson) wrote :

Thanks for pointing out the stray .bak files; will be fixed in Intrepid.

Changed in livecd-rootfs:
assignee: nobody → kamion
status: New → Fix Committed
Revision history for this message
Launchpad Janitor (janitor) wrote :

This bug was fixed in the package livecd-rootfs - 0.59

---------------
livecd-rootfs (0.59) intrepid; urgency=low

  [ Colin Watson ]
  * Remove /boot/initrd.img-*.bak (LP: #64887).

  [ Adam Conrad ]
  * Resurrect the proposed option for livecd.sh/BuildLiveCD, so
    we can build test images that include packages from proposed.
  * Build with updates and security by default, which is a no-op
    on development releases (where these suites are empty), but
    a necessity for point-release image building.
  * Copy bin/true from the chroot, not from the real root, so
    image builds for i386/lpia work on amd64, as intended.

 -- Adam Conrad <email address hidden> Tue, 10 Jun 2008 13:23:18 -0600

Changed in livecd-rootfs:
status: Fix Committed → 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.