Question regarding Linux kernels 2.2 and 2.4

[Thomas Mondoshawan Tate]

--3Gf/FFewwPeBMqCJ
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Fri, Aug 17, 2001 at 07:48:21AM -0700, David A. Sinck wrote:
>=20
>=20
> \_ SMTP quoth Thomas Mondoshawan Tate on 8/16/2001 16:58 as having spake =
thusly:
> \_
> \_ Anybody notice the abundance of people holding on and still patching t=
he
> \_ older 2.2 kernels? I've been rather curious about this for a while -- =
didn't
> \_ say much about it because I thought it was a fluke or something. Isn't=
 the
> \_ existance of the 2.4 kernels saying that the experimental 2.3 kernels =
have
> \_ reached a stable version?
>=20
> 1) Why install when it's working?

Because the 2.2 kernels may have security flaws and other issues that were
overlooked before, but fixed now. Besides, the abundance of new patches to =
the
v2.2 kernel would imply that it _isn't_ working for some cases.

> 2) How keen are you to take down a production machine just for a silly
> kernel upgrade?  What if it goes bad?  Horribly bad?

I would be quite keen for the interests of security and hardware driver
issues. If the install goes bad, I can still revert back to the older
installed kernel (now if I wanted to be completely asinine, I'd delete all
my older compiled kernels and set LILO to boot only to the newer 2.4 kernel
-- but then that's why we have boot disks, huh?). Also, at this point, the
2.4 kernel series is considered _stable_. This means, Linus is confident
that it will not kill your HD's filesystem in case of a crash, etc, and that
it is safe to be used on a production machine.

> 3) Sometimes kernel upgrades have lots of other dependencies that you
> may be unwillling to fight.  (glibc, etc)

Erm... 2.4's kernel dependencies haven't changed since the older 2.0 kernels
(besides modutils, but there's nothing functionally different -- only the
modules directory layout has changed). There are no such dependencies to
change between 2.2 and 2.4.

> 4) Look: a new client want something in the water in 4 days *counting*
> the weekend!  Fire! Fire! Turkey Lurkey, the sky is falling!

Eh? I'm confused... =3Dop

> 5) Laziness.  Never overlookup apathy. =20

Agreed. This may be a big reason, but it still doesn't explain why people
are still building new patches against the older 2.2 kernels.

> 6) Never overlookup apathy's friend: ignorance.

Good point.

> I've turned down significant hardware upgrades based on not wanting to
> burn in the box to my comfort.  Of course, then when they offered me
> the mem, mem's easy, right?  No problems there.  Except the HP Vectra
> XA's Quantum Fireball...lived down to it's name.  How apt.  I wonder if
> they gave the naming engineer a bonus for accuracy.  So I got the
> hardware anyway, but it was a pain to recover from backup to a new
> archetecture.

Ick... Sorry about that... I've done that a few times myself -- ironicially
with the same make and model HD. Conspiracy, maybe? j/k

--
Thomas "Mondoshawan" Tate
phoenix@psy.ed.asu.edu
http://tank.dyndns.org/presence

--3Gf/FFewwPeBMqCJ
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE7fVH9Yp5mUsPGjjwRAiTnAJ0dyTkKa0XU+/jf9RZsnPINF4//lwCbBbop
xVbTkNy5rxfWk/XdU+tZhgw=
=abNs
-----END PGP SIGNATURE-----

--3Gf/FFewwPeBMqCJ--