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

Re: Debian needs more buildds. It has offers. They aren't being accepted.

* Nathanael Nerode <neroden@twcny.rr.com> [2004-02-09 05:46]:
> Re: Debian needs more buildds.  It has offers.  They aren't being
> accepted.

> This has been going on for a long time.  Apparently everything is
> bottlenecked either by Ryan Murray or by James Troup.

Maintaining a buildd system for 11 architectures is a fairly complex
task, especially with architectures that require many machines to keep
up since it's likely that some machines will break.  There have been
various discussions about problems in the past, some have been fixed,
other problems came up.  Your mail basically says "fix this", without
actually saying what is wrong; just accusing various people based on
second hand knowledge.  So, for the benefit of everyone, let me give
an overview of what has happened and what is going on, as well as what
the status is and what needs to be done.

Number of out of dates holding up testing:

    i386 (8), hppa (65), ia64 (66), sparc (83), powerpc (92), s390 (92),
    m68k (100), alpha (111), arm (161), mipsel (184), mips (200)

Clearly, higher numbers are bad.  From this you can see that the
obvious problems are currently related to arm (161), mipsel (184), and
mips (200).


ARM's problems mainly stem from the fact that 2 buildd machines in
Canada (unpredictably) lost their hosting.  Fortunately, new hosting
has been found in the meantime (see http://www.debian.org/ports/arm/),
so the ARM situation should improve soon.  In the long run, some
improvements are planned to add more redundancy.  Simtec has donated 2
StrongARM 110 ATX boards (http://www.simtec.co.uk/products/EB110ATX)
to Debian, and I have just found hosting for them.  The plan is to
setup one as buildd and have the other as developer accessible machine
to replace debussy.d.o which crashes often due to hardware problems.


mipsel has 2 LASAT machines which have easily been able to keep up by
themselves.  mipsel is the only architecture which doesn't have any
Developer accessible machines (they are some machines available on
request, but there is no .d.o box).  Because of this, I contacted
Sun/Cobalt a few months ago, and they donated 2 Cobalt RaQ2 to use
which were supposed to be hosted at MIT.  Unfortunately, it turned out
that one of them was DOA, and the other box cannot be used because the
kernel was very unstable.  Nobody has worked on the Cobalt kernel for
a long time, so this is hardly surprising.  However, some patches
appeared recently and the box might run with them in a stable way.
Anyway, back to buildd.  The 2 LASAT machines can handle the mipsel
port just fine.  Unfortunately, one of them recently became unstable.
As Karsten pointed out, it was first assumed to be a problem with the
fan, but there are still problems after fixing that.  All the recent
problems of mipsel you've heard about are due to this machine being
down.  It was first assumed to be a temporary, fixable problem, but it
seems this might not be the case.  (On the other hand, the box has
been stable for a few days now.)  Hence, I've offered another mipsel
box to Ryan.  MIPS UK donated 2 P6064 development boards to Debian;
one is currently with Tollef Fog Heen in Norway, the other one was
with MIPS UK until a few days ago when we put it in an ATX case and
it's sitting in my room now.  The plan is to bring up one as a
developer accessible machine, and to use the other one as buildd if
the LASAT remains unstable.  Ryan is aware of the offer of this new
machine, and I think he's currently waiting to find out whether the
LASAT is stable or not.


MIPS is the architecture where we had a flamewar a few weeks ago.
Some machines were offered during that time and refused; I don't know
details about this, but I suppose Ryan had reasons for turning down
the offer (and to put this into perspective, Ryan does accept
machines; the most powerful mips buildd used to be one from a
non-developer who responded to Ryan's request for machines on
debian-mips).  I can only say that I suggested building XFree86 (as
per the request of the maintainer who wanted feedback for his
experimental package) on this machine which had been offered as a
buildd and was told that it didn't have enough disk space.  Whether
the machine would have been useful as a buildd is therefore
questionable, but I cannot tell for sure without knowing details of
the machine.  However, Ryan has mentioned before that he needs a fast
mips box, and not another slow one.  Which almost brings me to the
current situation.

Just to conclude what happened after the last flamewar.  The problem
with mips back then was that the fastest mips buildd (sgi.spamo.org)
had hardware problems.  The owner thought it was a disk problem, so I
told him Debian can buy new disks; he volunteered to get new ones
himself.  He did, but as it turns out it was a different problem which
he promptly fixed by using a component from his spare machine.  mips
was working again without any problems after this.

Unfortunately, sgi.spamo.org developed problems again.  The current
owner of the box said it can be fixed, but he'd rather not do it: but
Debian can have his box (plus his spare, slower machine) and get it
repaired.  In the meantime, I found out that Stephen Frost has a fast
mips box as well (to put this into terms for people who know about SGI
mips machines; slow is R4K and fast is R5K).  So my idea was to ship
the spamo.org boxen (1 R4K and 1 R5K) to Stephen who has 1 R4K and 1
R5K already, and then build 2 R5K machines and use the R4K as spare
machines for parts.  This is very recent; i.e. Ryan was just informed
about this possibility yesterday.  (Also, an O2 with 2 10K CPUs was
donated to Debian and is being set up, so that can be used as well in
the future if we need it.)

(In the meantime, to make the problem worse, casals.d.o needs a new
kernel and cannot be used as buildd in the meantime.  This should
hopefully be fixed soon, though.)

So, while mips has been problematic for a while due to unfortunate
circumstance, it's on the best way of getting fixed again.

In summary, the currently problematic architectures are being worked
on.  It should be noted, though, that most recent "cannot get
wanna-build access" and other discussions are not about these
currently problematic architectures, but about m68k - which, as you
can see in the statistics above, is doing pretty well with the 10 (!)
buildds they already have.  Two more are currently being build by Ingo
Juergensmann, and they'll probably be added once they become
available.  As pointed out in Ingo Juergensmann's message in this
thread, some m68k systems are also needed for d-i work.  I suggest
that Goswin's two boxes which have not been added to wanna-build
should be used for d-i work (especially given that Goswin expressed
interested in doing d-i m68k work int he past), and the two machines
Ingo is working on will become part of the w-b infrastructure when
they're ready.

Anyway, I am in contact with Ryan, James and others to get arm, mips
and mipsel addressed asap.

Martin Michlmayr

Reply to: