On Tue, 20 May 2003, John Beamon wrote:
> John wrote:
> > On Fri, 16 May 2003, John Beamon wrote:
> >
> >
> >>This is a tiny bit of a tangent, but it's relevant for me right now,
>
> snip
>
> >
> > I firmly believe there will not be a 9.x: there is 9 and _not_ 9.0. Its
> > successor, I predict, will be 10.
> >
> > If you want the stablility you have been accustomed to in 7.3 then you
> > must pay for one of the ES offering that's not yet available, or discard
> > your RH knowledge and adopt another distribution.
> >
>
> I know that second paragraph is meant as helpful advice, but I've heard
> it from several responders already. "If you want the stability you have
> been accustomed to in 7.3, then you must pay...", message after message.
You should not assume anyone has read the replies you've seen when they
make their reply.
> I have a boxed set of 7, released without a decimal, and I think 6
> was as well. I don't know that the prediction of rolling 9 straight up
> to 10 will prove true, but the context of the message is nonetheless
Then read the announcements, take a look at the RH ftp site and see
directories 7.{0,1,2,3}, 8.0 and 9.
The reasons RH has previously expounded for its dot releases no longer
exist.
> disturbing. No common users are going to be interested in "Linux 14"
> and "Linux 15" coming out eight months apart, let alone admins who got
> their companies involved to ESCAPE Frequent User Fees.
>
> I want stability. I've had at least incremental predictability in minor
> versions since 6. For my experience, that's 6, 6.1, 6.2, 7, 7.1, 7.2,
> and 7.3, over a considerable expanse of time in "Linux years". That's
> gone now. I will not be the first to bash RH for a business plan, but
> the first has long since come and gone on this list. When the kickstart
> list at redhat.com has a thriving thread about automating SuSe and
> Debian, there's a problem.
I agree there's a problem, and I think it's beyond repair. Some of Red
Hat's most experienced unpaid helpers have gone to other distros, and
they won't be coming back. Uses who pay more will expect more, and
probably be less willing to help out on these lists.
>
> KS is a good product, and I'll likely be rolling my next full platform
> out with KS and a heavily customized, barely recognizable RH
> installation. I'm building .rpm's for half the stuff we use already. I
> have all the parts here for a good package manager (don't flame that...
> it won't help the discussion), an automated installer, and a base
> installation upon which I can roll one for the company. And my scripts
> will end with...
It's not hard to write a basic installer: I've done so. Mine boots from
ROM or bootrom floppy, mounts a filesystem by NFS, partitions a drive,
mounts it and untars a tarball on it. It then does some customisation &
switches to running the installed system without reboot. It has nowhere
the flexibility of ks, but it's fine for use where all the target
systems are similar.
In a sense, it's more flexible than ks: it's written entirely using
shell scripts. Change the parition and mount functions and you can have
RAID or LVM. it asks no questions, can run without a console, imposes no
limits: if you want to install to nfs that's fine (that's one thing the
Debian installer does do).
--
Please, reply only to the list.
Join the "Linux Support by Small Businesses" list at
http://mail.computerdatasafe.com.au/mailman/listinfo/lssb
