Too much downtime

and we usually get the standard we're working on the problem. Why not just pony up and admit this version and probably the next one will be utter disasters? The major problem I can see is an absolutely psychotic upgrade schedule and a totyal lack of attention to the existing bugs. I am sure that residents will not mind if you take the time to FULLY test releases for bugs before releasing them. As it is all you are accomplishing is adding to your plate. Due to the huge increase in downtime customer confidence is fast failing and sales for in world merchants are in the toilet. People won't buy stuff if they think they can't get in world and use it. Take some major strong advice..fix the problems before adding anymore new ones. Otherwise you could find SL going the way of the Dodo bird as soon as someone who has taken the time to adress issues adn does enofrce ToS violations in thier world comes around. And don't overlook the possibility. I know of several projects already in the works. You guys gotta start worrying about customer service or it's gonna wind up hurting the company, and the customer base overall.

Past a certain point, the techies at LL should focus on stabilizing the game..making the servers/grid(s) more secure. I love and use most all the new features in every update, but would gladly forgo many if, in return, I got a less laggy, more secure virtual space here.

For some reason some of the smartest tech monkeys among us are unhappy enough with LL to cause disruptions. And with the exploding SL population, it may mean the more the 'merrier' in regards to who can bring a sim down. So, while it's not exciting or flashy..security may be the BEST new feature to add to SL.

Oh there are plenty of reasons to be disgruntled and discouraged about the state of SL. It seems that sim-crashing is nearly the only feature LL felt was worth adding to the script language (see my other post on this forum to get an idea of what I mean).

In my case, rather than resorting to such juvenile behavior as crashing the grid, I just started writing my own SL-like-system instead. At the current rate of "progress", I'll have the features I listed before LL will. Heh.

No use MY reading your post about coding. I hate tech stuff. Sell your intelligence to LL...as a security consultant of sorts. I'd ask for at least 60k real dollars/per annum for starters. ONE thing's for sure. Apparently the smartest techies in SL are NOT LL employees

Yeah too much for sure. Not good when you`re investing RL money into shops, etc.

A real problem lately is that, I suspect, the grid attacks are being launched by people who don't realize what they're responsible for.

How many times has someone in a sandbox handed you something questionable? And then you rez it?

We just saw this a couple days ago - someone came on the forum and claimed they had no idea how this happened, they were a newbie. And several people backed them up on it. The real attacker, most likely, got off scott free, because they had actually made the attack out of Linden-owned content. The resulting attack objects were flagged as "created by eric linden, owned by [possibly-innocent newbie]".

So now that it's possible to forge both fields, you just have to hope that the asset server logs who makes changes to objects and/or inventory transfers, or you can't even track down the perp.

I'd say a first step to unravelling this mess is simply to add a "recall user's items" function, capable of returning or deleting objects, across the grid, owned by a specific user. For good measure, set it so that user can access it as well. Most of us scripters have run into cases where one of our scripted items goes berzerk and runs off across 3 sims, never to be seen again. If we're unlucky, it falls on land with no autoreturn, and just sits there forever. And an unwitting newbie launching a grid attack has no recourse - there is no way to shut down an attack that you launch, even if you DON'T want to launch it. Besides, again, lots of scripters run into a misbehaving self-replicating item - one of my own projects inadvertently filled a sim (although, as a precaution, I had set a 10-minute lifespan on the items, which promptly deleted themselves when the time expired). There's no easy way to clean up these sorts of messes, and that's a big problem.

Its not like this is a fault in the code, or a bad server...

This is an _attack_ on the grid.. Someone made a malicious object for the purpose of taking down the grid. Dont complain like this is the Linden's fault, cause its not.Its a resident, like you and me doing this.

Blaming the lindens is like sueing the DMV cause someone rammed their licensed and insured car into yours.

The attack is caused by a user. The hours of downtime after each one is caused by insufficient tools on LL's side to clean it up.

A major grid attack should take 5 to 10 minutes to clean up.

kinda, sorta always knew it

Solar...for some reason, I don't think SL is profitable as yet. Could these tools be expensive software?

Not really. Unless their code is hopelessly broken (which, granted, some of it surely is), the tools should take about a single developer-week to build, at least to a primitive level that would allow the grid to be cleaned up quickly.

And you got this number.. how? I'm sure you think you know what your talking about, but 5 to 10 minutes? Puhlease.. Takes that long just to verify something is going on. Its not a simple matter of clicking a few buttons. If people didnt minded a few dozen sims being whiped clean of prims, it might take 5 ot 10 minutes.. But not if you are trying to save your good customers data, and verify that its all good, and that no more of the problem object are around.

By no means am I an expert either, but c'mon, 5 ot 10 minutes? That's dreaming.

Educated guess as a programmer of both 3D and database systems.

As long as the system tags objects in a way that allows them to be tracked and deleted easily (which, granted, SL does not seem to do), then it's just a matter of:
1) finding the offending object (which takes only a couple minutes, when using an application that can scan the database looking for such objects).
2) find the tag or tags for the objects (which should be only a few seconds, once an object has been isolated).
3) remove those objects.

The current numbers suggest that they are doing the cleanup largely by hand at the moment, without the benefit of such automated tools.

Well, between the person that has to actually solve the problem and the person that only has to imagine how long it would take them to solve it, the one doing the imagining invariably gets done much, much faster.

Wouldn't you think the LL employees would dust up at least a BIT faster after all the crashes the game has had...with as much experience as they've had at this point? Because it seems that Solar's ideas could be easily had and implemented. (Of course one would have to assume some sort of halfway decent work ethic at LL.)
So..somewhere there seems to be a weak LL link?

Am interested in Solar's ideas. They don't just explain away a problem as being part of life online, and are quite constructive.

If they're doing the removal by hand, then honestly, they're doing a good job of doing it as quickly as they are now.

The fact that they're doing it by hand is the problem.

Without knowing just how broken the SL codebase is, it's hard to estimate how long it would take to build the tools to automate this sorta stuff. A good, clean, manageable codebase would allow these tools to be written by one person in a matter of days. A snarly ball of spaghetti code could prevent writing such tools *EVER*.

I still stand by my estimate that a cleanup should take 5 to 10 minutes. The hours of manually looking through the database can be done in a matter of milliseconds by a computer program, leaving the only manual task of actually isolating the offending items.