long term support release

Ralf Corsepius rc040203 at freenet.de
Fri Jan 25 11:24:25 UTC 2008


On Fri, 2008-01-25 at 11:29 +0100, Michael Schwendt wrote:
> On Fri, 25 Jan 2008 10:47:11 +0100, Emmanuel Seyman wrote:
> 
> > * Ralf Corsepius [25/01/2008 10:31] :
> > >
> > > Do you have the TBytes to host it, do you have dozens of mirrors, do you
> > > have the bandwidth, do you have the man-power to administrate the
> > > buildsystem?
> > 
> > Were I interested in Fedora LTS, I'll start looking for server space and
> > bandwidth (man-power isn't exactly a technical ressource). And, yes,
> > I'm certain I could find enough to get started.
> > 
> > > Have a look at rpmfusion - Despite they are working on it for months,
> > > AFAICT, trying to copy/reuse the fedoraproject.org infrastructure it's
> > > still not up.
> > 
> > Is copying the Fedora infrastructure really the showstopper ?
> > Last I heard, it was hardware problems and lack of man-power.
All I know is what had been posted to the rpmfusion lists.

I recall complaints about setting up FAS, I recall complaints on lack of
stability, I recall pleas for help due to lack of knowledge and of
man-power. 

> Do you expect anyone outside the Fedora Project to duplicate existing
> Fedora Infrastructure, such as bugzilla, CVS+FAS, lookaside cache,
> buildsys for multiple architectures, master repos, introduce a new GPG
> key, switch mailing-lists, advertise the whole thing as a non-Fedora
> project?
Exactly this had been the reasons

>  That would make it much more difficult.


> Well, I'm sceptical that there would be enough volunteers to offer Fedora
> LTS free of charge, even if the Fedora Project permitted use of their
> infrastructure.
Which costs would 
* extending FC7's life time (and may-be re-branding/re-labeling it)
* combined with lifting all ACLs (in particular "core"/RH ACLs such that
community folks can intervene)
* and keeping alive the infrastructure
introduce?

No doubt, it would introduce some costs, ... but most of it could be
"bought-in" as side-effects of what already exists and needs to be
maintained anyway (buildsys, build machines, bugzilla, mirrors etc.).

>  Fedora LTS would compete with CentOS+EPEL and would lead
> to even more mass-builds.

We are back to what IMO, this all obits around: questioning the
"will" :/

> Still, a discussion of Fedora LTS should start
> with outlining guarantees (or less strictly, the concrete project
> goals/promises), policies, procedures, and a gathering of people with
> strong interest in such a project.
Agreed, this would be the actually critical, and the actually
interesting part.

>  Long before determining whether
> infrastructure is the last thing that holds up the project.
... but this is what would cause the real costs and what is building up
the barriers preventing 3rd parties from jumping "onto this train".

Ralf





More information about the devel mailing list