On 07/15/2013 11:55 AM, lee issued this missive:
James Hogarth <james.hogarth@xxxxxxxxx> writes:
Unless I missed it, nobody has described a particular use case yet in
which it is obvious that it is good to use CentOS. Upgrading holds its
risks as well as using software that cannot be upgraded. The future
cannot be predicted. So how do you make a decision like between using
Fedora and CentOS?
As much as I disagree with Harald on certain issues he did give you the
answer to this... You just choose not to hear it.
I'm hearing it; I'm merely asking for a particular use case.
The key risk for upgrades in business is security fixes and not features -
these upgrades are present... Within the lifetime of a project you are
highly unlikely to be ripping out and replacing key bits of the application
stack.
Vendors (including pure open source solutions backed by a vendor) will
support a long term distribution but not something like fedora where the
increase of cost for them to support it due to potentially massive changes
every 6 months is significant.
So you are saying that, for some good reasons, you can generally assume
right away that in business applications the risks involved with changes
greatly outweigh the potential advantage of having new features and that
you need something that doesn't change for at least three to five years?
Fedora is a bleeding-edge testbed. Massive changes every six months is
what Fedora is all about. We are the lab rats for Red Hat's engineering
efforts. We test the stuff, figure out what's good, bad, indifferent,
stupid or brilliant and they use that to guide their research.
Once enough changes have been made and those changes prove stable, then
Red Hat snapshots Fedora, does some additional tuning and tweaking and
it becomes the next major RHEL release. A bit later on, CentOS (built
from RHEL's source RPMs) becomes available.
In business, stability is the watchword. Using something like Fedora for
any business-critical operation is asking for trouble. That is what RHEL
(if you need support and are willing to pay for it) and/or CentOS (if
you are willing to live with community-based support) are for. Those are
stable platforms and won't undergo "major" changes for several years at
a time (e.g. CentOS 5 is still around three years later, although
updated; CentOS 6 has been around at least 18 months).
Does that make it worthwhile to use software that cannot be upgraded
when it reaches its end of life (either because support ceases or
because new features are needed)?
It depends on what the software is. If your business depends on it and
it's critical to your processes, then you really have no choice but to
use it. That's not uncommon, but it can be dangerous. I collect and
restore old cars (mostly Jaguars). Parts are often NLA and I have to
make my own or adapt others. No, the cars aren't critical to my
survival so the analogy isn't exact, but you get the point. If you can
transition from something with no support to something that does have
support, then you should do it. But only _you_ can make that
determination. Don't blindly upgrade something because "Gee! It's a new
release...it's gotta be better!" You have to look at the release notes
and see if the patches and such are something you need and it won't
break what you have. Due diligence, my friend!
----------------------------------------------------------------------
- Rick Stevens, Systems Engineer, AllDigital ricks@xxxxxxxxxxxxxx -
- AIM/Skype: therps2 ICQ: 22643734 Yahoo: origrps2 -
- -
- A day for firm decisions!!! Well, then again, maybe not! -
----------------------------------------------------------------------
--
users mailing list
users@xxxxxxxxxxxxxxxxxxxxxxx
To unsubscribe or change subscription options:
https://admin.fedoraproject.org/mailman/listinfo/users
Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Have a question? Ask away: http://ask.fedoraproject.org
