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

Re: Debian services and Debian infrastructure

On Tue, Jan 21, 2014 at 03:07:09PM +0000, Stephen Gran wrote:
> This one time, at band camp, Lucas Nussbaum said:
> > On 21/01/14 at 13:11 +0000, Stephen Gran wrote:
> > > This one time, at band camp, Lucas Nussbaum said:
> > > 
> > > [...] single mail shot with no follow up.
> > > 
> > > Well, it's been a bit over 2 weeks, and you haven't posted a follow up.
> > > This doesn't feel like a conversation to me.  I understand you're busy,
> > > but it feels very much like you're not actually interested in engaging.
> > > I find that slightly demoralizing.
> > 
> > I'm not sure of what you would like me to follow up on? I've just
> > re-checked, and emails from you in that thread did not include a single
> > question.
> I didn't think it was a Q&A session, I thought it was a discussion.  You
> asked several questions, none of which had a completely concrete answer in
> the thread, although several people offered opinions.  Even a mail saying,
> "ok, I take this discussion to mean I should stop talking to DSA and pay for
> a development environment somewhere" would have been useful.  I for one
> didn't take this mail thread to have reached that conclusion.

Indeed.  A response would have been appropriate had you reached a decision on
how you'd proceed, Lucas.  We asked for an open discussion on a public list yet
you've proceeded in a non-transparent manner.

> > Generally, I agree with everything that has been said in that thread (minor
> > snarky comments from Joerg on m68k and 'real unrealistic requirements', but
> > I didn't think it was worth replying to them).
> > 
> > For example, I fully agree that we should try hard to host 'important'
> > services on Debian hardware.  Also, I don't think that anybody disagrees
> > that we have a problem with 'speculative' services (as you put it, or
> > 'services development' as I put it). DSA cannot and should not provide
> > hosting for all Debian-related services being developed, as you wrote
> > yourself:
> > > I don't think jumping straight to a solution that puts all of the
> > > responsibility for every idea for a service in Debian on DSA shoulders is
> > > either the only way to go or even a good way to go.  There are plenty of
> > > bad ideas that should be allowed to wither on the vine, and there are
> > > always going to be services that have been designed in such a way as to
> > > be difficult to integrate into DSA-managed infrastructure.  We are, after
> > > all, a reasonably small team of volunteers.  Pretending that we can
> > > support an infinite number of services or an infinite variety of designs
> > > is just going to end in disappointment for someone.
> I think there's quite a range of options between "DSA can't host everything
> under the sun" and "I'll go set up a private parallel development environment
> out of project funds without any further discussion".
> > Now, of course, I'm very disappointed that nobody from DSA is interested in
> > acting as a gateway between service developers and hosting solutions
> > outside of Debian infrastructure that would be suitable for services in
> > development and experimental/maturing services. In my eyes, that would have
> > been a win-win situation, by putting DSA in the perfect position to be
> > aware of emerging services, and to interact early with service maintainers.
> > But, well, I cannot force anyone to do work that they don't want to do.
> So, I offered at one point to set up an openstack private cloud for DDs to
> use for service development and so on.  I got almost as enthusiastic a
> response to that as we got to kerberos, AFS, and now MQ.  I decided to let it
> go instead of putting lots of energy into something that no one would use.
> That sort of thing can be revisited if it's actually interesting for people.
> I'm not sure what you picture when you talk about us acting as a gateway.
> Perhaps you could elaborate on that.  I'm not keen on playing script monkey
> to set up machines for people - I'd much rather that interested people be
> able to do that for themselves.  If you just want us to be a point of contact
> for people developing new services, I think we've said several times that
> we'd like to be just that.
> > However, it sounded pointless to argue on that if there is no concrete
> > offer to host Debian's services being developed outside of Debian
> > infrastructure.  So, since that discussion, I've been talking to a few
> > hosting providers, and two of them have offered to support Debian with free
> > resources (on their clouds) for Debian development. Since I think that
> > avoiding vendor lock-in is a must, I'd like to make sure that we can get a
> > third one on board before working out further details. That will include
> > deciding how allocation of such resources happen, and where discussion
> > about this should happen. My first choice would be to use
> > debian-services-admin@ for that, but of course that will be your decision
> > as I don't want to 'pollute' the list with traffic you are not interested
> > in.
> No, that's precisely the sort of thing the list is for, I thought - it's not
> a private list for DSA or anything.  Not sure where the word pollute or its
> scare quotes have come from, but it sure feels hostile.  I'll assume you
> don't mean it that way.
> I have some operational questions about this cloud setup, since it seems
> you've delegated running Debian owned machines to us and then gone and got
> some that you don't want us to run.  I'm not sure what to do with that
> disjuncture.

We, we haven't said that we wouldn't talk to VPS providers.  We've said that
we're more interested in a bytemark-style donation of (real) hardware rather
than virtual machines (since an insecure hypervisor is an insecure virtual
machine - see openssl).  That said, vendors like rackspace have "private cloud"
offerings that might be the best of both worlds (we'd have to learn more about
their configuration).  And I actively engaged other DDs regarding their VPS
opportunities (although no response so far in some cases).

So, to me at least, the DPL engaging with equipment donors (real or virtual),
beyond the initial securing of a positive relationship, for the possible
development of debian.org services is not kosher.  It actively fractures the

My thoughts,


Luca Filipozzi

Attachment: signature.asc
Description: Digital signature

Reply to: