[clug] Fedora and the future of RedHat

[Andrew Pollock]

Anyone at my work will tell you I'm a rabid Debian supporter. I'm also a
Debian Developer and a RedHat Certified Engineer. I personally believe
that Debian is a technologically superior Linux distribution to Red Hat.
That said, I don't use Red Hat on a day to day basis, so I don't compare 
them regularly.

I've previously successfully lobbied to run Debian on all our 
infrastructure. Life has been reasonably good. We're now switching to Red 
Hat. Why?

Commercial Support.

Unfortuantely, management don't care about "technological superiority" or
"dependency hell".

At the end of the day, we have to run some proprietary commercial software
on Linux, and those vendors mandate what platforms they will support it
on.  They almost always mandate Red Hat version x, and pretty much never
mention support for Debian.

Sure, you can (and we have) make something that claims it will only work 
on Red Hat work on Debian. Usually there's a bit of work involved. You 
then have to lie to the vendor's tech support when you want some support, 
otherwise they'll wash their hands of you as soon as they discover it 
isn't running on Red Hat (for example). Management tend not to like how an 
off-the-shelf product has required manual friggery to get it to work. No 
longer turnkey.

The killer for us is we're currently replacing our Nagios monitoring with
CA Unicenter (another bright decision I won't get worked up about here)
and CA will only support Red Hat. It's now critical to the success of the 
implementation of Unicenter that we throw out about 60 servers running 
Debian and replace them with Red Hat. I actually question how hard CA has 
tried to get things to work with Debian, but I wasn't involved in the 
project...

The next problem is when Vendor A supports Red Hat X but Vendor B only 
supports Red Hat Y, and you want both products to run on the same Linux 
box. It's a no win situation, which is why the Linux Standard Base is so 
utterly important to end distro wars in terms of third party vendor 
support once and for all. That's why a vendor can say they support Solaris 
8, because Solaris 8 is Solaris 8. They can't say they support Linux 2.4.x 
because the actual operating environment can be so wildly variable.

So I don't think that my management's decision to switch to Red Hat over 
Debian is actually going to solve anything in the long term, but there's 
not a lot more that I can do about it unfortunately.

In the long term significant vendor education of the LSB and getting
vendors to think "distribution agnostic" instead of "Red Hat specific"  
will be all that allows good distributions like Debian to make it in the
"enterprise" space.

Andrew

On Wed, Oct 22, 2003 at 09:09:22AM +1000, Matthew Hawkins wrote:
> 
> Michael Cohen said:
> > maybe i have a big misunderstanding on the whole issue. opinions anyone?
> 
> Friends don't let friends use Redhat.
> 
> Seriously... what you need to do is get away from the "product" mentality
> and think more about the business solution.  If your company's priority is
> to run Redhat, they need a huge cluebat swung at them before they make a
> balls-up of any Linux installation, Redhat included.  The business drive
> for an operating system should be something that is reliable, easy to
> administrate, has decent support if required and can achieve an expected
> level of operation/performance (though this also depends on hardware...). 
> You should be able to demonstrate this with *any* Linux distribution
> that's worth looking at.  If you prefer Debian to Redhat (and who doesn't?
> ;) then demonstrate how Debian meets (exceeds?) your companies business
> requirements.  Focus on the features unique to Debian that will save your
> company money and make you a happier and more productive person to work
> with.  The bottom line isn't that you use product X, the bottom line is a)
> things work so b) employees and the company are productive.
> 
> 1) deploy decent system designed to exceed business reqs
> 2) ???
> 3) Profit
> 
> -- 
> Matt