[Opensim-dev] multiple, empty, root inventory folders

[Kyle Hamilton]

On Tue, Apr 22, 2008 at 1:56 AM, Ryan McDougall <sempuki1 at gmail.com> wrote:
>
>  On Mon, 2008-04-21 at 10:28 -0700, Kyle Hamilton wrote:
>  > On Mon, Apr 21, 2008 at 8:34 AM, Justin Clark-Casey
>  > <jjustincc at googlemail.com> wrote:
>  > >
>  > >  Multiple roots were the result of various code bugs.  I actually hope
>  > >  now that we have these nailed, and that all the current multiple
>  > >  inventories are now historical stock.
>  > >
>  > >  If this doesn't prove to be the case, I think it would be worth adding a
>  > >  multiple inventory check on inventory service startup.  I'm loathe to
>  > >  add anything more than this due to the code and startup cost, which
>  > >  would only be going to counter a situation which should in theory never
>  > >  occur.
>  >
>
>  >
>
> > I've been called at 3am -- and called others at 3am -- too many times
>  > (due to unexpected sanity violations on databases and poor handling of
>  > it by codebases).  I personally don't like unplanned failure modes.
>  > I'd like to see some sort of degradation strategy built into the
>  > infrastructure to help provide a highly-available system.
>  >
>  > -Kyle H
>
>  Indeed. However its important to note that since OpenSim is alpha
>  quality software, there is just no way anyone should be thinking OpenSim
>  will give them high availability -- yet.
>
>  Alpha-time is time to make a clean base that can be used without need
>  for hacking about for years to come. Lets make our designs clean while
>  its possible.

I wholeheartedly agree.  The two goals (cleanliness and insight) are
not mutually incompatible.

Alpha-time is the time when the greatest insight is needed into how
the designs behave.  If a given design causes problems that violate
the constraints of the environment, some aspect of the design needs to
be changed BEFORE the code goes to beta (or worse, release).

If you like, you can think of it as a means of obtaining test
information from long-running tests, as opposed to unit-testing single
discrete units of functionality.  (Not that OpenSim has much of that,
either.)  If you can create a clean design for integrating run-time
gathering of statistics, then you've solved one of the main problems
that eventually integrates alpha-quality code and design into beta or
release software.

I'm specifically thinking OSgrid with my suggestion -- but there's no
reason it couldn't be made into a developmental tool to figure out
when and how things break when the code is changed.  It also allows
for useful identification of when problems were introduced.  (In fact,
since OSgrid is OpenSim's largest-scale and longest-running tester,
such insight being made available can only help speed development.)

-Kyle H