Re: Question for all candidates: handle debian-admin more openly
On Fri, Mar 10, 2006 at 11:20:47PM +0100, Martin Schulze wrote:
> > > Are bruckner and voltaire overloaded or do they lack services the developers
> > > need?
> >
> > The release team has called for a multi-arch implementation to support
> > powerpc64 userland over the biarch situation. This calls for a machine capable
> > of building *and running* powerpc 64 code, which is not the case of existing
> > powerpc 32bit machines.
>
> Such requests and requirements change the situation. However, I have
> to admit that I first read about this particular requirement here. I
Well, i menitoned you ppc64, but not in detail, and back then i believed that
the biarch path was enough, which didn't need a ppc64 box.
> noticed some babbling about ppc64, sparc64, mips64 and s390x
> architectures but nothing that ended up in "will be included in the
> archive, hence, requres buildd and development machines".
Well, the current state is "will be included if someone makes it happen; but
etch will not be delayed for it". Still having ppc64 buildds is a prerequisite
for it to happen.
During fosdem i was part of a meeting concerning the multi-arch support, and
altough i feel the deadline for etch is very short, the plan is to at least
try for it.
> If this has changed, most probably debian-admin won't deny two
> machines for these purposes.
Cool.
> ----------------------------------------------------------------------
> Question to the release and archive people: Is there such a
> requirement? Will such architectures indeed be included the archive?
> Do we really need machines of the particular 64 bit architectures? If
> so for which architectures exactly?
> ----------------------------------------------------------------------
i guess thye three first are yes, and the last is amd64 and powerpc64 for
etch, and the rest for post etch if someone is willing to do the work,
(sparc64, mips64, hppa64, s390x). maybe others may make it to the first group,
but these are more niche markets.
I believe the powerpc64 is the one which wil benefit most from it, since
multi-arch is less usefull for amd64, where pure-64 is more interesting, and
multi-arch is only useful for running 32bit binaries of openoffice or other
stuff known to be broken on 64bit.
> > > Another question would be if the Debian project should accept every
> > > arbitrary donation and increase their machine pool even when there
> > > is no use the machine?
> >
> > We plan to support powerpc64 userland for etch, as thus this is necessary.
>
> Who is "we" in this case? Is it the Debian project?
I guess we here is the debian/powerpc porters, me among them.
> > Furthermore these machines could be used for other use than just development,
> > but then, i guess it is preferable to get a donation from intel than get a
> > donation from ibm ?
>
> I don't want to judge between donators and I don't see a reason why
> Debian should do so. Instead we should judge between donations and
> use the best that we can get.
Cool :)
> > > > It was my believe that at least one of them should be maintained under the
> > > > umbrella of the DSA team, in order to have it thrusted to be used to upload
> > > > packages, but the DSA team refused to have anything to do with them, which i
> > > > suppose is understandable since they have no time for it. I proposed to handle
> > > > it for them though, or have Bastian do so, and was equally refused.
> > >
> > > I've once told you that currently (one year ago or so) there is no
> > > use for them, since a) voltaire works fine as buildd and bruckner
> > > works fine as developer machine.
> >
> > sure, but we need a powerpc64 autobuilder for etch.
>
> I'm sorry but I don't see powerpc64 mentioned in the list of release
> architectures for etch in the status mail from the release team:
> http://lists.debian.org/debian-devel-announce/2005/12/msg00013.html
> Maybe you could point me to where it is listed?
well, it is a tentative release multi-arch, the plan is to not rebuild the
whole archive, since this would be wastefull and a performance hit in general,
but only those libraries which are needed to support apps needing >4GB of ram
are to be concerned. We spoke this with the RMs back then, and they told it
would be a non-mandatory release goal, not sure if it finally made it in the
list though.
So i ammend my claim, we need a powerpc64buildd if we are going to make it and
get the multiarch ppc64 support done in time for etch. It is probably not a
prerequisite, and can be added after the fact, but it would help to have it.
In the meantime, there are 2 perfectly fine quad power5 machines with huge
amounts of disk space, ram and reasonable bandwidth (the university of
augsbourg's network), sitting mostly idle.
Friendly,
Sven Luther
Reply to: