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

Re: I'll be a son of a bitch.



Steve Langasek wrote:
>
>The previous primary buildd for alpha has been off-line now for about 9
>months.  The porter machine for alpha has been off-line now for 2-3 months,
>IIRC.  That leaves the project with one operational alpha total right now.
>
>The ARM architecture has two buildds currently on-line; the normal
>complement required to keep up with unstable seems to be four.  Efforts are
>being made to get a third buildd up, but until a few weeks ago we were down
>to *one* buildd, so it's no surprise that progress is slow.  At least we
>have a porter machine up, but for a long time we didn't have that either.
>
>Mips and mipsel have each taken turns over the past year suffering from a
>lack of horsepower.  They are both in pretty good shape right now.
>
>Sparc only has one machine on-line right now, vore; although many developers
>do have logins on vore, as usual with buildds, using it for porting work is
>discouraged.  The secondary buildd, auric, is down and most likely is not
>coming back: the local admin for the two machines wants to down vore and
>replace both machines with a single chassis.
>
>I don't know if I would say "dire", but things look pretty messed up to me.

Agreed. I must admit, I was horrified when I first saw the proposal to
drop architectures from the etch release. But recent history shows
that several of the architectures are really struggling, either from
lack of machines or lack of manpower or both. This generally isn't due
to existing people not working hard enough, but rather that there is
often too much for individual people to do, no matter how dedicated
they might be.

As I see it, we need the following for each architecture:

 * Enough machines running:
   + normal buildds (could be hosted anywhere so long as the admins for that
     arch are happy)
   + security buildds (admin by DSA)
   + devel-accessible machines (best done by DSA, but necessarily essential -
     responsive local admins can help a lot here)
 * Enough people:
   + people with deep technical skills on the arch who can track / fix
     kernel, toolchain, glibc, etc.
   + other porters who can help fix / test arch-specific bugs as they arise
   + people to admin/host machines

If we're missing either people or machines, then an arch will hit
trouble. Machines _should_ be available easily enough for most of our
arches, with a couple of obvious exceptions. The problem comes down to
manpower in most situations - getting machines installed and running,
then working on the deep technical bugs as they come up.

This is a place where people should be able to help quite readily if
they want their pet architecture to remain supported and released. We
need a certain level of support from the interested developers to make
a port viable, and leaving all the work to DSA people doesn't work. No
matter how many users an arch may have, we need enough people willing
to step forward and do that work.

I'm happy to be able to host several machines at the moment - one of
the arm buildd machines, plus assorted machines from several of the
other arches. At the moment I'm mainly building and testing my own
packages; I'm happy to give access to those machines where they can
help.

-- 
Steve McIntyre, Cambridge, UK.                                steve@einval.com
  Getting a SCSI chain working is perfectly simple if you remember that there
  must be exactly three terminations: one on one end of the cable, one on the
  far end, and the goat, terminated over the SCSI chain with a silver-handled
  knife whilst burning *black* candles. --- Anthony DeBoer



Reply to: