Re: initrd-tools and booting root on lvm with new kernel

[Maximilian Attems <maks@sternwelten.at>]

On Thu, Oct 06, 2005 at 05:57:08PM +0200, Sven Luther wrote:
> On Thu, Oct 06, 2005 at 09:29:06AM -0500, Manoj Srivastava wrote:
> > On Mon, 3 Oct 2005 13:03:45 +0200, Maximilian Attems
> > <maks@sternwelten.at> said:  
> > 
> > > naa initramfs-tools are in the archive an working for all arch but
> > > sparc (klibc ftbfs there with gcc3.4/4.0).
> > 
> > > the kernel-package does not yet use the initramfs-tools. you need to
> > > invoke update-initramfs like this: sudo update-initramfs -c -k
> > > 2.6.13-1-686
> > 
> >         Err. putting 
> >   ramdisk = /usr/sbin/mkinitramfs
> >  in /etc/kernel-img.conf does seem to work. Or are you talking about
> >  using mkinitramfs by default? mkinitramfs fails if the kernel version
> >  is not at least 2.6.12, so it can't be used unilaterally. 
> 
> Would a solution to this not to have the kernel provide a
> /etc/kernel/postinst.d/<version>/kernel.sh script, which would in turn call
> mkinitramfs or mkyaird (or whatever it is called), or alternatively have
..
the patch that uses mkinitramfs-tools conditionaly in ubuntu is very
small and easy to integrate.

> >         I guess we could add a version specific check into the
> >  postinst to default to using yaird or mkinitramfs , if installed, in
> >  preference to mkinitrd, though I am usually hesitant to add in
> >  version dependencies into kernel-package in general, I could be
> >  persuaded that an exception is justified in this case.
> 
> Would a solution as proposed above be more to your liking :) ?

naa, working with kernel-package is very fine.
my proposal is just to localize changes, so that we clearly know,
what broke?

releasing a new upstream version and changing the early boot stuff
in the same time needs to be avoided.

--
maks