Re: .d.o machines which are down (Re: Questions for the DPL candidates)
Ben Collins <bcollins@debian.org> writes:
> Stop chasing red herrings, and just get back to work. Sparc has always
> been and always will be a maintained architecture.
Actually, work right now consists of answering paniced emails from my
students worried about their test on Friday, and waiting for a compile
of ec-fonts-mftraced to complete (which takes several hours).
Other people often say that sparc has historically been one of the
more problematic buildd's because of bad handling of dep-wait and so
forth. I don't know how much credence to give this.
One of my hopes for the new ftp/scc criteria is that buildds will be
much better maintained than they are currently, with a single set of
clear standards, which every single buildd is required to follow, and
with extra buildd capacity.
Saying "trust me, nothing bad will ever happen" is really
unimpressive. s390 is supposedly one of the fastest Debian archs, but
gee?! I've seen enterprise servers collapse overnight, and what keeps
the enterprises using them reliable is that they have a guy on a
beeper-leash who comes in at 3am and gets it going again. Yes, all
the things you indicate are wonderful reliability tools, *if* you have
the guy who is on a beeper-leash and will come in at 3am. If you
don't is it too much trouble to ask for a second machine so that
progress can still be made while you take a day or two to fix a
problem?
I have no objection to you taking a day or two, I'm just asking for a
backup so that the project doesn't have to wait for it.
Thomas
Reply to: