[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: RTL kernels - Re: Latest UTB Newsletter

I also have to question my own intelligence, for sending this message off
the first time without fixing the subject (digest gets me now and then...:).

> Date: Fri, 14 Mar 2003 19:41:26 -0500
> From: Audioslave - 7M3 - Live <creed7m3live columbus rr com>
> To: phoebe-list redhat com
> Subject: Re: RTL kernels - Re: Latest UTB Newsletter
> Reply-To: phoebe-list redhat com

> About the 2.5 kernel being a lot better with speed and such, I'd like to see
> it included and integrated within the distribution. I don't see any conflict
> with releasing an odd numbered distribution with an odd numbered kernel
> version. The whole backporting features seem to be like trying to satisfy all
> of the depedecies needed for certain programs to work. If a crucial system
> factor was not listed in the dependacies, figuring out the needed libs or base
> programs would be very hard to accomplish.

Um, fat chance Red Hat would put a development kernel in a stable
release. I see a whole LOT of conflict there. Stable releases should
only contain stable kernels. I'll take stability over speed, thank you.
You DO understand that odd-numbered kernels are DEVELOPMENT kernels,
while all Red Hat's x.x releases are stable releases, right?

PERHAPS you'd see a late 2.5-series kernel in an 8.1.92 beta or similar,
leading up to a 2.6 kernel in Red Hat Linux 8.2, but for obvious
reasons, Red Hat isn't going to put a relatively untested dev kernel in
a release.

And back-porting is a good thing. It isn't just trying to satisfy
dependencies (spell check, please) needed for certain programs to work.
It provides enhancements from the development kernels that have been
tested, deemed stable and worth adding to the already stable kernel
series. What good is that enhancement if you're running it on an
otherwise unstable kernel? Back-porting makes more sense until all of
the dev kernel features are frozen and deemed suitable to migrate to the
next stable kernel series.

While maybe you can afford a bit of instability as a trade-off for speed
on the system you use at home, many people are using Red Hat Linux in
mission-critical situations, where stability is of far greater
importance than a bit of a speed boost. Hell, I still run 7.3 on many of
my mission-critical systems. There are plenty of people of the opinion
that you shouldn't ever run x.0 releases on a production system. I know
I wouldn't want to have to support systems with unstable kernels, so do
you think Red Hat would?!?

In summary, putting a dev kernel in a stable release is a really stupid
idea. I have to question the intelligence of anyone who doesn't see the
problem there.
Jarod Wilson, RHCE
<jcw wilsonet com>
"A wise man once said nothing at all"

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]