Category Archives: Bugs

What happens when Second Life goes all pear-shaped.

BZZZT! WRONG!

Just two days after I posted my call to Linden Lab™ to cut the crap and make the Second Life™ Grid, you know, actually work, they’ve proven they can’t listen worth a damn:

Today we are very happy to share some exciting news with you: Linden Lab has acquired Xstreet SL and OnRez – the two leading Web-based marketplaces for buying and selling creations for Second Life. Over the past few months we’ve been working with the folks at Virtuatrade and the Electric Sheep Company to hammer out the details…

How much of those “last few months” spent in negotiating to take over two services that, unlike the Grid, actually work, could have been spent on, say, making the Grid actually work?

How much effort will integrating these two services into the overall SL environment suck away from making the Grid actually WORK?

And will these two services now quit, you know, actually working once they’re subsumed into LL’s already bursting-at-the-seams infrastructure, thus requiring even more effort to make them work again, effort that could have been devoted to making the Grid actually WORK?

(Are you starting to see a pattern here?  I hope so. 🙂 )

I’ll leave it to others to debate the business aspects of this acquisition.  I’m more interested in having a working environment in SL.

Linden Lab: Does the phrase “fiddling while Rome burns” do anything for you?

UPDATE: Okay…now maybe I can start to believe that LL is taking these problems seriously.  But I’ll refrain from sending the roses until I see some real results.

Advertisements

1 Comment

Filed under Bugs, Business, Downtime

Linden Lab: This has gone FAR ENOUGH. Fix SL *NOW*.

As I write this, Selena has just had to cancel our Sunday night event because Second Life™ is brokenAGAIN.

Most In world services are at reduced functionality at the moment. Please avoid L$ transactions or handling valuable (no-copy) assets until we post an ALL-CLEAR. Regettably, our ability to broadcast a warning in world is also disabled. Please let your friends know if you’re logged in. [emphasis mine]

When the system is so broken that the Lindens can’t even broadcast a message to tell people in-world how broken it is…well, something is rotten in the state of Denmark.  And it sure as hell ain’t Danish blue cheese.

Friday night, we had to cancel our event because the sim on which Solar Moonlight sits (Tyros) suddenly crashed on us, logging us out, fifteen minutes before the event was due to begin…and, upon logging back in, we were unable to TP there.  Thank God Lexxotica still seemed to be up and running, or who the hell knows what would have happened?

And this doesn’t just affect us; Prokofy Neva, one of the few people who tries to run a rental business in a reasonable manner, reports that he’s getting lots of people breaking leases:

I don’t know whether people refund because they can’t log on and get sick and suspicious of SL even when they *can* log on (or perhaps they get mad their friends can’t log on), or whether, more likely, they can log on, but they can’t get me to do something for them because *I* can’t log on.

Either way, bad for business.

Much as Prok’s critics might cheer his business troubles, anything that’s bad for his business is likely to be worse–perhaps fatally so–for other businesses.

Meanwhile, the Lindens issue self-congratulatory blog posts, promise “pie in the sky, by and by” with infrastructure improvements (that have yet to materialize), and continue to chase educators with a platform that can’t seem to even support its present level of use, let alone act as a mission-critical tool for education.  Anyone else have the words “fiddling while Rome burns” coming to mind?

It’s time for the Lindens to start bringing what Jim McCarthy, in his book Dynamics of Software Development, called “radical focus” on the problem of stability of the SL platform.  You can’t call for radical focus too many times over the course of a project, as McCarthy points out, but at this point the Lindens are overdue.  Come on, M Linden, now’s the time to show leadership.  If my own boss in RL can do it, you can do it.  LL’s ability to ship bug-free code has fallen from “average” down to “marginal at best,” and is continuing the spiral towards “complete fiduciary misconduct” at this point.  How much more do they think their paying customers can take?

“…I warned the distributor I’m a Hershey bar…The Hershey bar gets smaller and smaller to stay the same price.  But it can only get so small.  I can shrink myself only so small before I’m nothing, a man without quality or quantity.” – Mort Lesser, “Mouthpiece,” by Edward Wellen

UPDATE: FJ Linden has posted a big, semi-technical explanation of what’s been going on and how LL is moving to fix it.  All well and good, FJ, but, as we say in America, “Talk is cheap.”  If you want to convince me, and other dissatisfied Residents, that you mean business, here’s the way to go about it:

  • Your timeframe for the rollout of these fixes is WAY too long.  Think “days,” not “months.”
  • What about manpower to meet that timeframe?  Easy: Every Linden who can code should be working on stability fixes right now.  Every Linden who can’t code should be working on testing said stability fixes. It’s “crash priority” time.  You guys’ future is at stake.
  • Forget all those other side projects, like building more mainland sims, or replacing the browser engine in the client, or other such foolishness.  All other considerations must be secondary to stabilizing the Second Life Grid and making it so people can actually USE it. I remind you: Your future is at stake here.

In other words, LL:  It’s time to shit or get off the pot.  Go big, or go home.

20 Comments

Filed under Bugs, Downtime

Report From The Aftermath

1.13 Upgrade StatementFour days ago, on the 29th of November, Linden Labs released the new version of Second Life to the world, version 1.13.

That version has turned out to be every bit as much “bad luck” as its version number would suggest.

Now, before we go into the litany of woes the Grid has experienced since then, I do want to highlight some of the things this new version seems to have gotten right. The new Friends List functionality, for instance, is far more flexible than it used to be, and should hopefully reduce, if not eliminate entirely, the problem of friends TP’ing in on you at the wrong moment. (This happened to me once…a young newbie TP’d in on me as Danielle and I were sharing, ahem, a tender moment.) My old friend Vertex Zenith is no doubt pleased to see that his feature proposal, allowing emotes in IMs, has come to fruition at last; that and the “typing…” indicators do a lot to help bring SL IMs into the 21st century. And the “Web” tab in the profile is, of course, the shiznit; not many people seem to be making use of it yet, but you may now read a “stripped-down” version of Evans Avenue Exit straight from my profile in-world. I have not explored some of the more esoteric things, like the new parcel info functions in LSL, but more-experienced scripters than me are actively doing so.

Are these things enough to compensate for the living hell the Grid has gone through these past few days? I don’t know…especially when you consider what has happened since then…

Basically, most of the problems can be traced to the main database, which keeps track of pretty much everything about the world. Database failures cause all kinds of things to happen:

  • Teleports start failing
  • People’s avatars fail to appear on their screen, or their cash balances fail to show up
  • Search no longer works
  • Money and inventory transfers in-world fail
  • Profile and group information fails to come up
  • The new Friends List functionality fails to work, in that permissions cannot be changed
  • Things fail on the secondlife.com Web site itself, mainly the Friends Online indicator, but also things like the Partners page and even the LindeX, on occasion.

These problems were all amply documented in various posts to the Grand Unified Linden Blog between then and now; in particular, Torley seems to have been run ragged trying to cope. Other bizarre issues have cropped up, such as missing water in some sims. A couple of temporary workarounds have been posted, too.

As I write this, seated at my computer desk in the main house at the family compound, the Grid seems quiet right now…but Search Places is still nonfunctional, as are Classified and Event icons on the map. This hits the Gin Rummy and Don’t Panic! Designs right where it hurts, as we depend on Event listings for nightly traffic, and on Classifieds to attract new potential employees, which we sorely need right now. It gets worse, too; money transfer failures can inhibit our ability to conduct events at all. (We were literally about to call off an event two nights ago, until the money transfers mysteriously started working again.)

So are people packing it in? Bailing out of SL and going back to whatever else they did before they found this world? Not bloody likely, says Tateru:

Ultimately the litmus test of the issues that Second Life is having right now is in what we do, not what we say. We might say it stinks, and howl, scream or grumble about it, but what we’re doing seems to indicate no especial lack of enthusiasm.

[…]

You might be unhappy about glitches, performance or whatever, but you’re still logging in for about the same amount of time on the same days. Or, if you’re cutting back, more people are logging in in your place. That suggests you’ve either got confidence, or that you don’t, but something keeps you coming back anyway.

Well, glitches or not, I still have responsibilities, to the club, to our employees, especially to Danielle (still trapped in RL after her house fire)–and not least to my own sense of self-respect. But aside from that, I still believe in the potential of Second Life. We’ve seen upgrade-related glitches before, albeit not of this magnitude, and likely we will again. This is, as I have pointed out time and again, both on this blog and in-world (including to at least one Linden), the single most complex software system I have ever used or heard of. That it even works at all is something of a miracle in and of itself. Plus, I have friends here…friends I would miss if I decided to abandon the Grid, and that hopefully would miss me.

Does that keep me from feeling a little screwed, like lots of people? Oh, hell no. But I’m not giving up hope.

“Never give up, never surrender!” — Jason Nesmith as Commander Peter Quincy Taggart, GalaxyQuest

(Special offer: For a limited time, I’m giving away copies of the T-shirt you see me wearing in the picture above, “I Survived The 1.13 Upgrade.” Come find me in-world for your copy…)

Leave a comment

Filed under Bugs, Current Events, Downtime

Simquake!!!

Just this evening, we encountered some really odd behavior in SL, or at least in the South Sunset region.  Whenever people walked around, they started bouncing, as if they were falling through the floor and then getting booted back up.  Sitting or standing very still seemed to alleviate the problem, but it was still nerve-racking.  It felt almost like an earthquake…and, being originally from California, I have reason to respect the power of earthquakes. 🙂

It lasted long enough for me to call Live Help and report it.  George Linden, one of the newer Lindens, came by to check it out, and it lasted long enough for him to feel the effects himself.  Then, suddenly, while he was there, it stopped.  All of us were at a loss for words, and George was reluctant to restart the sim without there being a real problem (understandably), so the event, hopefully, will simply fade into history.  But that’s a failure mode I’ve never run into…

1 Comment

Filed under Bugs

Self-Replication

Four times this week the Grid has been attacked by a plague of self-replicating objects, as evidenced by this report on the Grand Unified Linden Blog and these two reports over on Second Life Insider. These objects spawn copies of themselves, kind of like the classic UNIX “fork bomb” that spawns new processes until the system is overwhelmed, and quickly overwhelm, not just the parcel or reagion that they’re in, but the asset server itself, which causes a world of hurt for Residents, breaking Search, the map, and teleportation. (The objects may also exhibit other annoying behavior.) The last couple of times this happened yesterday, LL actually had to pull the plug on every single freaking script on the Grid to be able to stop these things. This causes a hell of a lot of stuff to stop working…and nearly caused Danielle to lose L$1000 in a texture vendor.

So, my question is: What is Linden Labs doing to find the perpetrators of these acts, to get them off the Grid, and to ensure that this sort of thing can’t happen again?

Creating objects that interfere with the functioning of the Grid to such a degree that LL has to disable all scripting functionality in order to clean the mess up…that sounds like a pretty damned clear violation of TOS 4.1(viii) as well as “Disturbing The Peace” under the Community Standards. Any user doing this really ought to be kicked off the Grid, for good. But…will they stay kicked off? Thanks to the removal of account verification on the beastly day of 6/6/06, a user whose account has been terminated “with extreme prejudice” can simply create another account and be back online in mere minutes. LL has supposedly implemented a “machine hash” so they can block entire systems from accessing SL…but I don’t know how well that’s working. Even so–and this is a point many critics of the open-registration policy overlook–there’s nothing stopping a kicked-off griefer from contacting his buddies via AIM (or what have you) and passing along his griefing techniques to ten of his closest friends. Result: up to ten more attacks. Theoretically, this could replicate in much the same way that these objects themselves replicate. Lack of open registration would slow down the onslaught, but would probably not eliminate it entirely.

So…how could LL prevent this from happening altogether? I don’t know if there’s a good way to do so without breaking script behaviors that many existing objects rely on. Some have glibly talked of restricting the use of llRezObject(), or even scripting itself, until an account becomes verified and/or has been in-world long enough. Aside from the fact that it still wouldn’t stop a sufficiently-determined griefer, I think that would add a slowdown to the operation of scripts that need to rez objects, by introducing a whole bunch of extra security checks into the process. Is this a necessary evil? I don’t know.

One thing I do know: Danielle said, in response to news of the latest attack, “The griefers got smart.” Here’s hoping LL gets smarter in response.

3 Comments

Filed under Bugs