[Bioclusters] how are the Redhat product changes affecting existing and future plans?

Mitchell Skinner bioclusters@bioinformatics.org
05 Nov 2003 06:40:55 -0800


On Tue, 2003-11-04 at 07:57, Chris Dagdigian wrote:
> Another item that has been on my mind recently...
> 
> What are people doing about RedHat deciding to kill off their consumer 
> product line? Are people going to pay the freight for Redhat Enterprise 
> Linux or are people just going to use Suse/Debian/Gentoo etc.

I'm going to push for fedora.

Obviously, people running large clusters that have to cater to more than
one set of users have to be a little more careful in the
change-management department, and for those people the pain of
re-certifying everyone's apps is going to be a concern.  However, I
expect that moving between fedora releases won't be terribly painful
(per fedora objective #7 [1]), and if that's still too much change
there's the fedora legacy project [2].  In his email Chris said fedora
expected 2-3 "major" releases per year, but I don't see the "major" in
any of the actual fedora pages.

If pain caused by change is the problem, then I understand people who
want less change, but there's a lot of room for improvement in making
change less painful, and I would rather expend effort there than in
back-porting errata fixes to old versions of software.

For us, we have a relatively small cluster used by a relatively small
number of people in one main field.  We have source for most of our
apps, and the expertise to fix, recompile, and (to some extent) port
them, so I'm not too concerned about binary compatibility.

Also, there are places where having more recent software would be nice--
lm_sensors, SATA, ACPI, IPMI, there must be others on the hardware
side.  On the software side, is it just me or is the pace of change in
clustering-related software picking up?  It might be worthwhile to live
nearer to the edge of some of those projects.

IMHO, something like ROCKS would work better as a set of packages in
fedora-extras.  There are getting to be more and more third-party
packagers for Red Hat, and you just know that if they all stay separate
then you're going to want packages from more than one, and they're going
to conflict with one another somehow.  Fedora-extras isn't going to
enforce a no-conflict policy, but I would expect better feedback and
communication about conflicts in the fedora-extras case than in the
"we're all going to do our own thing" case.

Mitch

[1] http://fedora.redhat.com/about/objectives.html
[2] https://lists.dulug.duke.edu/pipermail/fedora-legacy-list/