[Opensim-dev] Concerning recent grid instability and inconsistency

[Brian Wolfe]

Well, so far as one of those many that's handled the packet code, I've
been chasing down some inefficiencies in the login sequence. I could use
some more people trying out the UserThreadPoolBranch of the repo.

If you understand git then you can fetch a more recent version that was
just merged against trunk yesterday at
http://brianw.terrabox.com/OpenSim/OpenSim.git  as branch UTPBranch (or
was it UserTHreadPool, brain wonky before first coffee)

It replaces the thread per user creation with a dedicated threadpool.
Some people report one impression, others another, andit works "great"
for me personally. So more eyes on what it's doign to syetems would be
appreciated. :)


On Fri, 2008-02-29 at 08:24 -0600, James Stallings II wrote:
> I'm glad to see I'm getting some response to this, it's quite
> encouraging!
> 
> That said, I have to ask for all of your indulgence; as you're all
> aware, I haven't been around the project for near as long as most of
> you, and know much less about how things have been done, by who, and
> why.
> 
> What I do see that causes me to write things like the original post,
> is that we are all embedded in this sort of defacto organic process
> that is admittedly producing profound results, but that on occasion
> has us making mistakes that are, in a more directed effort, outpaced
> with a bit of planning.
> 
> Dont get me wrong ;) I have no interest in seeing some big highly
> articulated roadmap, it would be stale before it could be edited up
> and posted. Nor the sort of monolithic top-down control that would be
> required to produce and implement such.
> 
> However, I *do* think we need to try to adapt effectively in some way
> to some of the circumstances that arise from our somewhat undirected
> workflow as developers and testers.
> 
> My #1 concern as I tried to raise awareness on some of these
> occurrences (my apologies for being so coarsely informed about some of
> the deeper details of grid communications) was to draw attention to
> the possibility that a lot of the bug hunting we were doing was
> perhaps misguided, and the result of providing completely uncontrolled
> access to the osgrid infrastructure to potential region operators. To
> be clear, I do not have philosophical issues with this wrt osgrid,
> indeed, I am very much in favor of letting everyone connect up who
> wants to - with a few very specific and definable exceptions:
> 
> 1> the pathologically misconfigured
> 2> the pathologically ignorant
> 
> The former we can address in a variety of ways - tools are the key.
> The latter, not so much. But if we take care of #1, it pretty much
> covers #2 as well.
> 
> MW, I'm sorry if I gave you the wrong impression about getting
> commitments and producing new work on the 'packet handling code' (a
> label I'm already regretting my use of ;)  Of course many have
> contributed to it and no one 'owns' it. I think what I was after there
> was to try and form up a concerted effort to revisit the work thats
> been done, in the interest of freshening it up and ensuring that no
> opportunities had been missed to make it more stable and/or efficient.
> Any such effort is necesarily informal, but I think it would be far
> more effective if all concerned agree in advance that its an effort
> worthy of undertaking. Also, I don't think I envisioned compelling
> anything of anyone, so much as getting confirmation they were either
> (dis)interested/just had other things to do or, prepared to come focus
> together with us on this specific area of endeavor. Sure it is
> *possible* for anyone to just dive in and start working on the code;
> but it isn't *practical*, per se.
> 
> Thanks so much for taking the time to read my spiel and contribute
> your thoughts. I have every confidence that mere awareness of these
> concerns will have a profound impact on what happens afterwards :)
> 
> 
> Cheers!
> James
> :D
> 
> 
> -- 
> ===================================
> The wind
> scours the earth for prayers
> The night obscures them 
> _______________________________________________
> Opensim-dev mailing list
> Opensim-dev at lists.berlios.de
> https://lists.berlios.de/mailman/listinfo/opensim-dev