[PLUG] Linux distro recommendation as a replacement for SuSE

[someone]

On Sat, 2006-11-04 at 15:56 -0800, Wil Cooley wrote:
> On Sat, 2006-11-04 at 15:40 -0800, M. Edward (Ed) Borasky wrote:
> 
> > I'd recommend staying away from CentOS. I've used it, and I haven't had
> > any problems whatsoever with it, and it seems to be what it claims to be
> > -- a community-driven rebuild of Red Hat Enterprise Linux. My hesitation
> > to recommend it is based on concerns about the strength of the community
> > relative to the strength of the Fedora community. CentOS is *competing*
> > with Red Hat and Fedora is *co-operating* with Red Hat. That's a *big*
> > difference!
> 
> I disagree.  Regardless of the relationship between CentOS and RHEL,
> CentOS is a very strong choice--it has the stability and familiarity of
> RHEL without the price tag (which honestly isn't that bad at $300/yr).
> If you want to get, say, your RHCE you can learn it on CentOS.
> 
> The additional benefit is that RHEL remains an easy option--if you have
> customers who want to pay for support or want to run a proprietary
> application on a certified platform in production, you can use RHEL and
> use CentOS for dev and QA.
> 
> If you have a customer with a RHEL server which they no longer want to
> pay for support, you can just switch update sources to CentOS and not
> worry about it.
> 
> If you maintain a collection of locally-built packages, you need only
> build them once (per major version).
> 
> Fedora isn't a good choice for commercial uses like this because its
> lifespan is frightfully short and fedora-legacy near death.  I like it
> on my own workstations because I always want the latest there, but my
> servers are all CentOS.
> 
> Wil

I'll second Wil on this so far.  I seem to be moving in this direction
myself.  CentOS is a good system if you want a clean Linux system that
doesn't come with any proprietary software.  It is seems to be a good
choice for stability.  CentOS is not for supporting the latest
and greatest.  The update repository seems very limited.  On the other
hand, upgrades often impact system stability in a negative way.
Theoretically if you don't install experimental compilers etc., you
don't need to update as frequently.

I think all of the major Linux distributions could be pushed 
aside by one that is: documented better, current, and designed to
run in a small footprint.  I wonder what Puppylinux is based on
and if any of the small Linux distributions will compete with 
the RHEL clones?  We are starting to see embedded Linux distributions,
but they seem to be mostly specialty distros that are hard to find out
about.  These distros appear to be very tightly tied to specific
hardware for the most part.  A linux distro that targets certain Linksys
routers which has been mentioned recently on this list comes to mind.

I've tried gentoo and I don't like it.  If you don't like the poor
documentation that comes with the 2.6 kernel, I seriously doubt that
you will like gentoo much.  I'm sorry, but I don't understand portage
very well and I worry that a system that trusts it's package management
system and recompiling software to stay current is going to be nothing
more than a maintenance nightmare.  In a world where no upgrade can
cause loss of functionality or break programs, gentoo might shine.  In
reality, making a mistake in how you compile one package or upgrading
something that can't be without replacing everything is not the sort of
situation you want on a production server.  I think gentoo magnifies the
confusion of updates that people already feel on yum and apt based Linux
systems.  Updating in general in my opinion hasn't been implemented well
in any popular operating system at this time.

     --  Michael C. Robinson