[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: .d.o machines which are down (Re: Questions for the DPL candidates)

On Tue, Mar 15, 2005 at 02:28:13PM +0000, Martin Michlmayr - Debian Project Leader wrote:
> * Julien BLACHE <jblache@debian.org> [2005-03-15 11:44]:

> >  - auric: RAID is dead (and auric is basically demilitarized since the
> >    compromise -- not even running a buildd, although I'm not sure
> >    about that)

> As I understand it, the plan was to convert auric into a buildd but
> the RAID needs to be fixed.  Ben Collins was looking into this but I
> don't know about the status.  I've also heard discussions several
> months ago about using one of Ben's really fast machines.

> This is based on what I've heard at some point and may not be
> accurate.  I'm CCing Ben so he can comment.

> If we do need auric and if we need resources to fix the RAID, Debian
> can make funding available.

auric had been running a buildd, until its disk went south (fairly
recently).  We are down to one buildd for sparc now.

> > That would resurrect quite a few machines, including an alpha buildd
> > (escher).

> I'm CCing Thimo Neubauer who is in charge for escher.  What I said
> above applies to escher as well.  Thimo, can you let us know what the
> status is and whether it would be good to have escher back?  BTW, Noah
> Meyerhans at MIT has two fairly large Alpha machines he can bring up,
> but so far we've been told they are not needed.

escher (21164) is not fast enough to keep up with the demands of
unstable on its own; until lully died, it was really only used as a
porter machine, and only got pulled in as a stopgap buildd at that

It would be good and useful to have escher back as a porter machine, but
for etch we would need something a bit beefier for the N+1 buildd
requirement -- e.g., lully, or one of Noah's machines.  If it wasn't
clear, the "N+1" means "any one buildd can fail and Debian is still able
to keep up".

Steve Langasek
postmodern programmer

Attachment: signature.asc
Description: Digital signature

Reply to: