yum wants to remove my kernels, why?

[Jeff Spaleta]

On 1/22/06, Timothy Murphy <tim at birdsnest.maths.tcd.ie> wrote:
> I still think it is a bad idea to install the new kernel automatically.
> The worst thing that can happen for a newbie
> is that he turns on his laptop and it doesn't work.


> I regard kernel and distribution as orthogonal,
> and would rather keep them separate.
> I don't find it very onerous to go through the Grub menu,
> and choose the kernel (or OS) I want.

If you don't find it difficult to use the grub menu to find the kernel
you want.. why is it so bad if the new updates become the default?

new kernel comes out with security fixes... it gets installed and
becomes the default
it has a regression.. user notices the regression and then boots into
the backup kernel from the grub menu... user then reports bugs about
the regression. I don't see the problem with this picture. 
Regressions suck.. but the backup kernel is still on the system to use
if someone runs into regressions.  Unless people hit those regressions
and report those regressions back to the maintainer there's very
little hope that those functionality regressions will be fixed.
Security kernel updates are very important, I don't see it as an
appropriate trade off to make the security update kernels optional to
avoid potential regressions.  Anyone running a system which needs to
avoid reboots into kernel updates because of ciritical production
situations should configure their system appropriately at install time
and should be aware of the impact on security by not using the
security update kernels.

if the new kernel comes out and isn't the default.. how many users
will "remember" to use the grub menu to select the new kernel..
instead of just booting in the kernel that was the default?

>
> The most important issue for a newbie (and for me)
> is that whatever OS I am using
> should work with the least possible trouble.
> Everything else - including security - can come later.

Its interesting that you don't include closing known security
vulnerabilities in your "least possible trouble" definition.  From
where I sit, known security vulnerabilities are trouble.

I disagree with your ranking of priorities... its very easy to reboot
into the backup kernel if there is a regression. its not so easy for
users to understand the implications of security and how much
"trouble" vulnerabilities mayb cause. I think its in the userbase's
best interest for the updates to prefer known security updates over
unknown stability issues.

-jef