Re: custom kernel

[Morgan Terry <mterry@acsbps.com>]

Marcelo Chiapparini wrote:
> 
> Dear debianers,
> 
> I've already compiling (debian way) a 2.2.18pre21 custom kernel for my potato
> box. I have the new /usr/src/kernel-image_*.deb file waiting to be installed.
> I think that if I do "dpkg -i" I will destroy the current 2.2.17 kernel, which
> works perfectly. I want to know if there exists any way to boot with the new
> kernel just to see if it runs OK, keeping the old one just in case something
> goes wrong and I need to boot with the old kernel again. The idea is to define
> the new kernel as the default one after testing.
> Thanks in advance for your time
> Marcelo

Installing the new kernel-image will not do anything to your existing
kernel.  It will, however, make lilo boot the new one.  If you want to
have access to your old kernel (which is always a good idea), all you
need to do is set up a section for it in your lilo.conf.  You can
probably just copy your current kernel's section, and change the label
and image lines.  For example, here is what I have for one of my boxes:

image=/vmlinuz
	label=linux
	root=/dev/hda1
	read-only
image=/vmlinuz.old
	label=linux.old
	root=/dev/hda1
	read-only

Assuming you have /vmlinuz linked to your current kernel (which should
be the case - it should have been set up when you installed), typing
"linux" at the lilo prompt will boot it.  If you want to boot the old
kernel (/vmlinuz.old should point to it automatically), you can type
"linux.old" instead. 

Every time you install a new kernel-image package, it will automatically
rotate these links (/vmlinuz will point to the newly installed kernel,
and /vmlinuz.old will point to the previous kernel), so this will give
you a nice little safety net.  

HTH 

-- 
Morgan Terry
mterry@acsbps.com
  
It is easier to move a problem around (for example, by moving
the problem to a different part of the overall network
architecture) than it is to solve it.  --RFC 1925