Qie Niangao Posted May 22, 2018 Share Posted May 22, 2018 23 minutes ago, Kyrah Abattoir said: If they move regions to an "on demand" kind of system i wonder how this is going to affect scripts, since a lot of them are supposed to run all the time. I mean... what about my prime numbers? Interesting. A lot of scripts burn a tremendous amount of CPU between interactions with the rest of the simulation, and even more burn even more between effects on the simulation state. Theoretically, all that could be offloaded from the sim completely, although that would be a massive rewrite of the way LSL compiles to Mono. Currently it seems the vast majority of script computation is to make breedables hungry (which is, after all, the whole purpose of breedables). If all those miserable critters could get hungry and die offline, Schrödinger's cat-like, until there's somebody to observe their starved corpses, that could improve real time performance by quite a lot. 4 Link to comment Share on other sites More sharing options...
KT Kingsley Posted May 22, 2018 Share Posted May 22, 2018 There was this thing a while ago where empty regions would get put to sleep. I can't remember the term used for the process and I don't know if it's still in operation. It did, apparently, cause some distress to Meeroos, the breedable de jour. I did publish a script (a dummy HTTP call every 5 minutes, or something), that circumvented the region's automatic downtime. Seeing as how Meeroos are now, apparently, extinct, maybe my script turned out to be ineffective, wasn't taken up by breeders, was in its turn circumvented by LL, or people just got bored with Meeroos. But let this be a lesson from history. 1 Link to comment Share on other sites More sharing options...
Drake1 Nightfire Posted May 22, 2018 Share Posted May 22, 2018 5 hours ago, Kyrah Abattoir said: If they move regions to an "on demand" kind of system i wonder how this is going to affect scripts, since a lot of them are supposed to run all the time. I mean... what about my prime numbers? What if someone goes to a store where the merchant has caspervend vendors that use a dropbox system and the drop box it at their home sim and they arent online. The home sim would be in pause mode but the store sim would be active. Would the vendors even work if they cant talk to the dropbox? Is it really that much of a cost drain to leave them up? I mena, teir isnt cheap. You cant tell me it really costs 300 usd a month to keep one sims worth of info online. Link to comment Share on other sites More sharing options...
Kyrah Abattoir Posted May 22, 2018 Share Posted May 22, 2018 42 minutes ago, Drake1 Nightfire said: What if someone goes to a store where the merchant has caspervend vendors that use a dropbox system and the drop box it at their home sim and they arent online. The home sim would be in pause mode but the store sim would be active. Would the vendors even work if they cant talk to the dropbox? Is it really that much of a cost drain to leave them up? I mena, teir isnt cheap. You cant tell me it really costs 300 usd a month to keep one sims worth of info online. It never did but I've never seen SL as some kind of "low cost" service. Even today they essentially offer a rather unique product that simply doesn't exist anywhere else because any company looking at creating a virtual world is purposely trying to AVOID the madness that LindenLab created. Sure there are opensim grids but none of them has attained such a longevity, popularity, or reliability. Despite all the problems SL had and is having it still outperforms all competition. I always considered that the costs of owning land are basically subsidizing all the infrastructures nobody is paying directly, like the centralised asset storage which appears to (almost) never delete anything that has ever been created in the last 14 years. 4 Link to comment Share on other sites More sharing options...
Phil Deakins Posted May 22, 2018 Share Posted May 22, 2018 1 hour ago, Kyrah Abattoir said: I always considered that the costs of owning land are basically subsidizing all the infrastructures nobody is paying directly, like the centralised asset storage which appears to (almost) never delete anything that has ever been created in the last 14 years. Very good point. 1 Link to comment Share on other sites More sharing options...
Kyrah Abattoir Posted May 22, 2018 Share Posted May 22, 2018 To clarify my point, SL indentifies everything on a 128bit UUID system so LindenLab will never have to worry about filling it or even random UUID collisions (I've been running a script that tries to create a collision, for years now). The asset server is terrifyingly large (and today probably spans across many physical machines), every time you upload something, everytime you copy an item, everytime you compile a script or save a notecard, a new entry is created and is kept essentially for ever, and deleting something simply removes that reference from your inventory. The asset remains and you can still access it by script if you know the UUID. Then all of this has to be backed up regularly because hard drives do fail. Sure, we aren't at the "Google class" when it comes to size, but this is pretty massive for a company like LindenLab.. 1 Link to comment Share on other sites More sharing options...
Klytyna Posted May 22, 2018 Share Posted May 22, 2018 2 hours ago, Kyrah Abattoir said: The asset server is terrifyingly large (and today probably spans across many physical machines) Six years ago, I remember an SL wiki page referring to "Asset servers" (yes plural) being grouped into "clusters" and there were multiple clusters, possibly with some kind of mirroring system, since while there are almost certainly regular full backups, and incremental backups of the data, the sheer volume means that allowing an un mirrored server to fail and then trying to find out what was on it when it died and restore that from removable media is a massive PITA... You also hear references to SL having well over a Petabyte of asset data, thats well over 1000 TB. You don't keep that kind of data on "a single physical machine" unless you are the kind of people who buy "Blue Waters" machines from Cray... Back in 2012/2013 somebody cut the cable between the server farms with a backhoe, and half the grid vanished offline, not just sims but asset server clusters too as they were on the wrong side of a hole in the ground. changing clothes was a no no that day. By way of a comparison, a petabyte of mp3 music tracks would have lasted Chesus Toast from his alleged death to well past his alleged return date 18 years ago, or allowed him to do the special effects for the movie "Avatar", and the human brain is ESTIMATED to have about 2.5 petabytes of data storage... 1 Link to comment Share on other sites More sharing options...
Kyrah Abattoir Posted May 22, 2018 Share Posted May 22, 2018 I know you don't keep this on a single machine, but remember that SL is 5+ years old now. Pretty sure the asset server ran for a while on a single machine (if all the asset server scaling problems we had in the first couple years are a hint of that) 1 Link to comment Share on other sites More sharing options...
Tripp Foxtail Posted October 22, 2020 Share Posted October 22, 2020 (edited) I'm so, SO SORRY I posted this. Nevermind. You can unbunch your panties now. Edited October 22, 2020 by Drew Mysterious 1 Link to comment Share on other sites More sharing options...
Sassy Kenin Posted October 22, 2020 Share Posted October 22, 2020 16 minutes ago, Drew Mysterious said: I'm not Happy, Karen. Is there a Karen running this project? I had no idea!.. its a disaster waiting to happen!....Quietly updating my passport!😳 4 Link to comment Share on other sites More sharing options...
Tripp Foxtail Posted October 22, 2020 Share Posted October 22, 2020 (edited) ... Edited October 22, 2020 by Drew Mysterious Link to comment Share on other sites More sharing options...
Theresa Tennyson Posted October 22, 2020 Share Posted October 22, 2020 2 hours ago, Drew Mysterious said: As a Builder - I've found this to have negatively impacted me BIG TIME. Objects no longer behave in expected manners, scripts need to be reloaded and reset over and over and textures applied to surfaces suddenly go BLANK on saving and have to be reapplied. I'm not Happy, Karen. Where are you seeing this? Because your store is on a region that isn't on the cloud. Link to comment Share on other sites More sharing options...
Drake1 Nightfire Posted October 22, 2020 Share Posted October 22, 2020 1 hour ago, Drew Mysterious said: Yes Karen, it's THAT bad. It's like having an extra 1,000 avatars jammed up my pixel a** and they are ALL having a RAVE. 1 Link to comment Share on other sites More sharing options...
Drake1 Nightfire Posted October 22, 2020 Share Posted October 22, 2020 3 hours ago, Drew Mysterious said: As a Builder - I've found this to have negatively impacted me BIG TIME. Objects no longer behave in expected manners, scripts need to be reloaded and reset over and over and textures applied to surfaces suddenly go BLANK on saving and have to be reapplied. I'm not Happy, Karen. 1. I have been making clothing pretty much daily for the past 10 years. Have yet to see what you are describing. 2. Perhaps the sim you are on is laggy and heavily populated. 3. You necroed a thread about something that was started about 3 years ago. You cant possibly be saying this has just started to impact you. Something that was put in place by LL over 3 years ago is just now hitting your sim? Seriously? Link to comment Share on other sites More sharing options...
bigmoe Whitfield Posted October 22, 2020 Share Posted October 22, 2020 34 minutes ago, Theresa Tennyson said: Where are you seeing this? Because your store is on a region that isn't on the cloud. this is what I'm curious about, I think only the beta grid currently is on aws. 1 Link to comment Share on other sites More sharing options...
LittleMe Jewell Posted October 22, 2020 Share Posted October 22, 2020 (edited) 43 minutes ago, bigmoe Whitfield said: this is what I'm curious about, I think only the beta grid currently is on aws. There are main grid regions in the cloud - over 1000 now, per this blog post. The latest Deploy threads in Technology also talk about main grid regions being in the cloud. Edited October 22, 2020 by LittleMe Jewell Fixing the horrid typo from my coffee-needing brain 3 Link to comment Share on other sites More sharing options...
bigmoe Whitfield Posted October 22, 2020 Share Posted October 22, 2020 11 minutes ago, LittleMe Jewell said: There are main grid regions in the cloud - over 1000 now, per this blog post. The latest Deploy threads in Technology also talk about main grid regions being in the cloud. Thank you :3 I never got that email notice about it, which is weird, but thank you ♥ Link to comment Share on other sites More sharing options...
Tripp Foxtail Posted October 22, 2020 Share Posted October 22, 2020 (edited) I am SO, SO ,SO Very Sorry I posted something to the wrong thread that you all feel I am actually NOT experiencing. Let's all take a DEEP BREATH and try to go on with our day without being quite so judgy. M'Kay, Karen? Edited October 22, 2020 by Drew Mysterious Link to comment Share on other sites More sharing options...
Theresa Tennyson Posted October 22, 2020 Share Posted October 22, 2020 1 hour ago, Drew Mysterious said: I am SO, SO ,SO Very Sorry I posted something to the wrong thread that you all feel I am actually NOT experiencing due to the move the Cloud that began approx 14 hours ago. Let's all take a DEEP BREATH and try to go on with our day without being quite so judgy. M'Kay, Karen? Unless you're talking about earth hours, that is... It's been going on for months. Annnnnddddd now we see why the Lab didn't generally say that they they moved various services to the Cloud until after they moved them. Link to comment Share on other sites More sharing options...
Tripp Foxtail Posted October 22, 2020 Share Posted October 22, 2020 (edited) Please ... just stop. I apologised to all of you and took down my post and corrected the other one. Find someone else to pick on. Edited October 22, 2020 by Drew Mysterious 1 1 1 Link to comment Share on other sites More sharing options...
Qie Niangao Posted October 22, 2020 Share Posted October 22, 2020 Over the years I've seen more or less what was described. Which is to say, I don't doubt for a minute that @Drew Mysteriousexperienced it, but it isn't necessarily a result of moving some services and/or regions to the cloud. One reason we should take reports like this seriously as possible cloud migration artifacts is that some operations will be problematic when performed on a region that's still hosted on Lab hardware when the operation relies on cloud-hosted services, and vice versa. That's simply gonna happen during the transition, for which April Linden begged our forbearance yesterday, as noted above. The tricky bit here is knowing when to file a jira, when we also know everything is in flux. Just for sanity, though, I might give a pass to intermittent performance problems: we already know they're going to happen while servers and clients are hosted on opposite sides of the cloud "membrane." 1 Link to comment Share on other sites More sharing options...
Rowan Amore Posted October 22, 2020 Share Posted October 22, 2020 The OP date is actually 3 years ago so it's not really much of a surprise anymore. 1 Link to comment Share on other sites More sharing options...
Solar Legion Posted October 22, 2020 Share Posted October 22, 2020 Right .... you're being picked on .... Link to comment Share on other sites More sharing options...
Rowan Amore Posted October 23, 2020 Share Posted October 23, 2020 Didn't notice anyone being picked on except maybe KAREN. 2 Link to comment Share on other sites More sharing options...
Tripp Foxtail Posted October 23, 2020 Share Posted October 23, 2020 By the way Karen - I just logged-in and got this message. So, it seems that OTHERS are also noticing that something is wrong and it's not just me after all. Good Luck Dogpiling on all of them too. **MWAW** PS- https://outpost6sl.wixsite.com/outpost6/post/a-letter-to-karen 1 2 Link to comment Share on other sites More sharing options...
Facebook
Instagram
Twitter
YouTube
Flickr
Recommended Posts
Please take a moment to consider if this thread is worth bumping.
Please sign in to comment
You will be able to leave a comment after signing in
Sign In Now