On Tue, Sep 19, 2006 at 02:44:04PM +0100, Dave Page wrote:
>  
> > Dave, the questions I raised on -core were the result of concerns
> > and discussion of people on IRC. Fact is that the problematic
> > point of my question was simply ignored and I kept getting no
> > answers at all until there was a positive answer available.  Now
> > people, who don't even answer questions unless they can answer
> > them positively, aren't usually those who point out existing
> > problems ahead of time.
> 
> I agree with that point, and have spoken with Marc to try and
> prevent the situation occuring again.  In partial defence of myself,
> I could only answer when I knew for sure what was going on myself.

That Marc kept *you* waiting like this and made you look flaky is
simply not acceptable.

> > Second, maybe people need to know what those provisions, that are
> > put into place, are.  Right now, as I understood it, the plan is
> > that you have full access to everything.  But Marc made clear that
> > this is only for emergencies and as long as he is somehow
> > reachable, he's going to do the work.  Which means, that the only
> > backup person we have is kept away as far as possible, instead of
> > actually making sure he/she can perform the task in case of
> > emergency just as swift as the main person.  This strikes me as
> > dumb of a plan as having no plan at all, sorry.  This isn't
> > designed to serve as a backup in case of emergency, this is
> > designed to reinforce the status quo.
> 
> No, it's common sense.  Marc's day job is running his servers which
> he knows inside out.

With due respect, his day job is working at the university, and he
does not make his living running servers.  In fact, he loses money at
that.

> Mine is running Vale Housing's servers.  I'll back up Marc as far as
> the postgresql.org resources are concerned but neither I, nor anyone
> else can be expected to know his systems as well as he does and it
> is inevitable

No, it's not.

> that it will remain quicker and easier for him to fix most issues -
> which in almost every case he does (I woke him at 3AM yesterday for
> example).  Similarly I wouldn't expect anyone but a member of my
> staff to be able to fix an issue on a Vale server as quickly as I
> might.

It's not about "as quickly," it's about Marc's refusal to take the
appropriate steps now that the project is much bigger than he has any
idea how to deal with.  He consistently refuses to delegate, and does
not answer hard questions.  This is the behavior of a petulant spoiled
brat, not an infrastructure provider, and it's important to the
PostgreSQL project in general and Slony-I in particular that our
stability not be subject to his whims.

> > I originally asked for developing a policy or a guideline how our
> > mission critical services and people _should_ be backed up.  Must
> > have been my German accent, but instead of getting anywhere into
> > that direction, everyone was explaining that others don't do
> > better either, how we are doing things and then went off changing
> > details.  Everyone got lost in the details again, instead of
> > thinking about the big picture.
> 
> Well your accent can be kinda hard to understand sometimes :-) but
> the point I was trying to make was that whilst it might be an
> interesting exercise to draw up such guidelines, we can't expect
> *any* of our providers to actually guarantee anything

In fact, we can do just that.  Part of the deal of becoming a provider
is that you agree to certain terms, which we can spell out.

> (especially as we don't pay them) and that our time was better spent
> improving our own redundancy, server, people and software-wise.

Our redundancy will get precisely nowhere so long as we continue to
permit Marc to make himself the frequent single point of failure.  We
need to grasp this nettle and we need to do in now.

Regards,
Dave.
-- 
David Fetter <david at fetter.org> http://fetter.org/
phone: +1 415 235 3778        AIM: dfetter666
                              Skype: davidfetter

Remember to vote!