RHEL Copyright Removal
Robert G. Brown
rgb at phy.duke.edu
Fri Nov 21 11:30:36 EST 2003
On Fri, 21 Nov 2003 neil.brown at syngenta.com wrote:
> Hi all,
> We're having a bit of a dilemma here, as I'm sure many others are, about
> what to use as our standard Linux distro with the end of life of the Red Hat
> family. RHEL or SLES are looking favourites in terms of supportability, but
> of course there's the not insignificant problem of cost. The thought of
> having to pay at least $179 per server, with around 50 compute nodes, along
> with various other non-beowulf Linux servers doesn't appeal.
> I've been trying to find out how much effort it takes to strip the RH
> copyrighted bits out of RHEL and compile it for our own use and whether
> doing so reduces it's functionality a great deal. I've trawled the web and
> usenet, but not found much to write home about on the subject.
This will soon be a FAQ. The best solutions kicked around so far (if
you wish to stick with basically free RPM-based full-service
kickstartable/pxe installable distros) are two community supported
This is basically a core that is RH9 with all the logos etc stripped
down to where they inherit GPL (eventually completely, I imagine). It
is a community supported model, where I believe they are looking for
people to take on pieces of the bug triage tree -- Adopt a Package
Today! It is designed in layers, with a "core" that should be fully
functional at the server and workstation level and supported as well as
anything out there, a legacy layer, and a contributed/kitchen sink layer
with less stable but bleeding-edge useful stuff. The project is
yummified from the beginning, which means that it is very simple to
create/rsync your own repository mirror and then use yum to maintain a
LAN or cluster from it.
At a guess, NEARLY anything you have set up for RH 9 will eventually be
quite portable to fedora, although on the yum list I hear of occasional
exceptions, as one might expect until things settle down.
The fedora core is "in production" now at version 1, I believe, although
I expect that only hardy admins and developer types are adopting it at
first until and to help it settle in.
Note well that www.fedora.org is a site that will just ask you to go
away, it is NOT associated with this project...:-(
Note well that Red Hat IS associated with this project. This may or may
not make you feel good about going this route. I personally think they
are strongly committed to it as they rely on SOMETHING to create a
rawhide -> semistable released -> rockstable corporate chain; they damn
well can't unlease barely-out-of-rawhide on people paying big bucks per
seat and disinclined to participate in the debugging process.
This is Community Linux WITHOUT corporate strings, run at least in part
by clustervolken. They too are stripping RH as a base, but plan to
eventually diverge. At a guess, at some point there will be Much
Synthesis and sharing between the two projects as it would be silly not
to. They too are soliciting humans to help out.
I know people heavily engaged in both projects, and expect both projects
to be stable at the starting level of RH9 before RHL support ceases.
One or the other will likely be the most successful at setting up and
organizing the bug triage network and perhaps eventually dominate,
although they are also likely to differentiate in focus (Caos has a very
definite cluster/scientific computing flavor due to the work
environments of a lot of the primary drivers).
HTH. I personally have stopped worrying about the transition, and plan
to convert my personal machines to fedora "soon" to start screwing
around with it prior to a campus conversion likely in the spring.
"Soon" as in my first rsync of the oceanic fedora core to my home
repository server is being slurped through a DSL straw as I type this,
likely done sometime today. I have a totally idle box all lined up to
be first, PXE and kickstart all happy -- I'll cheerfully report my
experiences as soon as I have any.
> Have any of you had experiences with such an exercise? Were they positive?
> How much effort was required? I believe ROCKS is based on RHEL 2.1 (not sure
> whether it's ES or AS) so it surely can't be that bad as a cluster oriented
> Thanks for any suggestions,
> Beowulf mailing list, Beowulf at beowulf.org
> To change your subscription (digest mode or unsubscribe) visit http://www.beowulf.org/mailman/listinfo/beowulf
Robert G. Brown http://www.phy.duke.edu/~rgb/
Duke University Dept. of Physics, Box 90305
Durham, N.C. 27708-0305
Phone: 1-919-660-2567 Fax: 919-660-2525 email:rgb at phy.duke.edu
Beowulf mailing list, Beowulf at beowulf.org
To change your subscription (digest mode or unsubscribe) visit http://www.beowulf.org/mailman/listinfo/beowulf
More information about the Beowulf