Re: [Treebase-devel] Fwd: Treebase Dev problems

[Carl B. <cbo...@gm...>]

Thanks, this sounds like a good place to start.

It isn't common that it does this, only if a user writes a loop that makes
repeated calls to the server.  Most of my use examples show constructing
single queries.  I've just added a function for users to make repeated
calls that should have a bit less handshaking each time, and can pause over
the loop.

-Carl

On Mon, Nov 14, 2011 at 6:16 PM, William Piel <wil...@ya...> wrote:

>
> Yeah, I totally agree -- if it takes down dev, it will also take down
> production. It would be better to point it to a third deployment (e.g.
> perhaps http://treebase-stage.nescent.org/treebase-web/), or better yet,
> one that runs on a separate spare machine so that it won't take down any
> other services -- and meanwhile we try to figure out how to make the API
> more robust. As a stateless API, it ought not to be filling up memory
> faster than it can dispose of it (if that's what's happening), but perhaps
> the problem is something else (corrupt data?).
>
> bp
>
> On Nov 14, 2011, at 7:56 PM, Rutger Vos wrote:
>
> Hey Carl,
>
> production isn't that much more ironclad to be honest. If a user is
> doing some serious harvesting of dozens/hundreds of requests for
> studies there's a good chance some of those are going to be ones that
> will construct long running queries to re-constitute all the data in a
> study.
>
> Until we have some caching strategy there's not really a good
> recommendation to make for rate of calls: even relatively low rates
> (several seconds apart, say) may be able to bring the server to its
> knees.
>
> Rutger
>
>
>


-- 
Carl Boettiger
UC Davis
http://www.carlboettiger.info/