On 3/27/24 11:25, Pierre-Elliott Bécue wrote:
Hi,
Thomas Goirand <z...@debian.org> wrote on 27/03/2024 at 00:24:30+0100:
Hi,
As you know, there's a large amount of money sleeping in SPI account
for Debian. Do you have ideas on how to spend it?
Would you be ok spending 100k USD on buying hardware for a new Debian
cloud, for example? I've always volunteered to operate it for Debian,
but it never went through, because I haven't spent time to find where
to host it and so on, but highvoltage liked the idea. Do you like this
idea? Do you think it'd be useful for Debian?
Please, let's take some time to think about the implications of spending
a shitload of money to buy hardware that we wouldn't know where to host,
and that would require a load of maintenance and time.
FYI, if I didn't go forward with the project, that we've been discussing
with Jonathan over at least 3 years, is because I have no idea where to
host it. I have clearly stated that having this hosted at *my* company
(ie: Infomaniak) is *not* what I want to do to avoid any type of
conflict of interest. I am staying firm with the idea that I shouldn't
do that. Though if anyone has a clue where we should do it, we can open
such a discussion (yeah, after the vote: my question was if Andrea was
happy with this type of project, not how and where we should to it).
Since I have failed to find where to host, maybe I should open the topic
with a wider audience, in the hope to have some suggestions.
So, let's not get further on this topic and let's reopen it soon.
If any discussion should arise on these matters, I'd rather them to
occur not as a platform for a DPL candidate but after a reasonable
discussion with the concerned parties, eg, DSA.
No. The DSA people have already stated multiple times that they do not
want to be integrated in such a project. Let's do this without them if
this is still the case. If you asked me, I'd say I would regret not to
do it with the DSA, but it's entirely their call, and it's also probably
good to try not to be overloaded with complicated clusters like OpenStack.
Also, I found very annoying that we don't have enough buildd, or that
the reproducible build project doesn't have as much hardware as they
would like. Would it be ok to spend another 100k USD for this kind of
things?
Same, with slightly less concern regarding hardware volume and
maintenance.
Yeah, DSA and other teams (ports maintainers?) should be involved. From
a DD perspective, it's just very annoying to have to wait days, and
sometimes even weeks, to have a package built on all arch. IMO, we
should address this, in some way or another.
Cheers,
Thomas Goirand (zigo)
