Simulator User Group/Transcripts/2011.11.29
| Prev 2011.11.22 | Next 2011.12.02 |
List of Speakers
| Andrew Linden | AnnMarie Otoole | Ardy Lay |
| Fancy Detector | Flip Idlemind | Jonathan Yap |
| JoyofRLC Acker | Kallista Destiny | Kaluura Boa |
| Kelly Linden | Leonel Iceghost | Liisa Runo |
| Pauline Darkfury | Qie Niangao | Rex Cronon |
| Simon Linden | Vincent Nacon |
Transcript
[12:03] Fancy Detector: Kelly Linden has arrived!
[12:03] Andrew Linden: Hello everyone.
[12:03] Leonel Iceghost: hey
[12:04] Ardy Lay: Hello Andrew, Kelly
[12:04] Leonel Iceghost: for some reason I don't read lindens chat blue today
[12:05] Ardy Lay: Alright you Lindens! Get with the program! Back to using blue chat for you! ;-)
[12:05] Andrew Linden: News first...
[12:05] Leonel Iceghost: haha
[12:05] Andrew Linden: I did get a response about the "enable return of encroaching objects on mainland" issue...
[12:05] Rex Cronon: hello everybody
[12:05] Erica3Dunkan: hi! im very sad! :(
[12:06] Andrew Linden: they are still working on it and haven't forgotten
[12:06] Vincent Nacon: we're not doctors
[12:06] Pauline Darkfury: Cool. Any idea at all on possible timescale for it?
[12:06] Andrew Linden: but other things have come up that trumped its priority
[12:06] Andrew Linden: no I didn't get an ETA for it
[12:07] Kaluura Boa: LR stuff coming soon?
[12:07] Pauline Darkfury: Well, thanks for the latest, happy it's not forgotten :)
[12:07] Andrew Linden: Dunno the ETA for Linden Realms (LR).
[12:08] Kaluura Boa: Pity... I can hardly wait for llTeleportAgent()...
[12:08] Fancy Detector: Simon Linden has arrived!
[12:08] Andrew Linden: As to the request for map tiles with predictable asset uuid's... I'm going to re-ask about that, now that I have re-learned what the request was really about.
[12:09] Kaluura Boa: Good to know... This one I'm waiting for too
[12:09] Andrew Linden: honestly, I don't see how it would work well... the textures would get cached by the asset delivery system
[12:09] Andrew Linden: so they wouldn't update in any predictable way
[12:09] Pauline Darkfury: It would need the caching to have a sane TTL (e.g. 24h) on it
[12:09] Andrew Linden: yeah, and there are multiple levels of cache
[12:09] Kaluura Boa: Yeah... 1 tile per day is good enough.
[12:10] Andrew Linden: So I'm going to ask around to see what the expected TTL would be.
[12:11] Pauline Darkfury: Using the HTTP paradigm, the "best" would be to have an If-Modified-Since query
[12:11] Andrew Linden: I suspect one problem is that if the texture is used a lot, the longer it could be cached, and the the TTL would effectively be forever (or Much Longer Than You'd Like)
[12:11] Andrew Linden: because our asset system really was desinged to be "write once, never change"
[12:11] Pauline Darkfury: Right, yeah, LRU expiry doesn't work for that type of asset
[12:11] Andrew Linden: and many of the caching systems know that
[12:12] Andrew Linden: That's all the news I've got.
[12:12] Kaluura Boa: Why can't you just use the same UUID assignment system to the tiles than what currently exists?
[12:12] Pauline Darkfury: There's "easy" ways to get the predictable refresh at the right time, maintain cache benefits, but whether that's "easy" to retrofit onto the current asset caching is another matter
[12:13] Andrew Linden: Kaluura, I can't parse your question. Could you rephrase that?
[12:13] Pauline Darkfury: The main problem with the current system, Kaluura (leaving aside 1.x vs. 2.x stuff), is that it requires a bot and an external database to get the map texture UUID
[12:14] Kaluura Boa: Currently, the tiles have UUIDs... The ones we can get from 3rd parties. These UUIDs must come from somwhere... And they update regularily... So...
[12:15] Pauline Darkfury: The alternative would be to stick with new UUID every day, but give us a llRequestSimulatorData function to get the current UUID
[12:15] Andrew Linden: The UUID's update regularly? or the contents of the corresponding assets update regularly?
[12:15] Kaluura Boa: The UUIDs change.
[12:15] Flip Idlemind: Could the dataserver...
[12:15] Flip Idlemind: Pauline beat me
[12:15] Flip Idlemind: Gyaahaahah ooooh
[12:16] Andrew Linden: Ok good, because that agrees with what I understand how the asset system would work.
[12:16] Pauline Darkfury: Using a magic UUID, if the cache expiry could be dealt with, would avoid the need to even have a daily refresh script in the prim. You could set the texture once, it would always be current
[12:16] Andrew Linden: That is, new assets have to be created to update the map tile.
[12:16] Ardy Lay: Hmm..... Does each region (or region host) still function as a "mini asset service" for skins of avatar that were baked in the associated region?
[12:16] Andrew Linden: And so I guess the problem with those who want to use the map tile assets is that their UUID's are not easily discoverable.
[12:17] Liisa Runo: yup
[12:17] Kaluura Boa: Well, using HTTP isn't easy...
[12:17] Andrew Linden: Yes Ardy, I believe so.
[12:17] Pauline Darkfury: Yes, that's the hassle with the current setup, you need a bot scanning all 30,000 regions, a database storing the results, web server, and llHTTPRequest to get the current UUID. That's a ton of overhead for something that really should be simple.
[12:18] Andrew Linden: The baked texture assets are temp (don't live in the asset system proper) but their UUID's are communicated to the viewer, so it knows which ones to ask for.
[12:18] Liisa Runo: not about easy/hard, it is about realibility, we take a risk when we make something that rely on 3rd party service, so we want LL to deliver the map tile UUID's to us so we know it will still work after 2036
[12:18] Kaluura Boa: And since pretty much nobody uses MOAP, we don't have much alternative than getting UUIDs....
[12:18] Flip Idlemind: I hate (love) bringing this up, but another thing that's a ton of overhead for something that should really be simple is...getting someone's key by their name
[12:18] Andrew Linden: An LSL script that wanted to use map tile assets has no easy way to discover the right asset.
[12:19] Ardy Lay: Are map tiles treated like resident textures?
[12:19] Pauline Darkfury: Yes, name2key is the other thing which requires an absurd level of resource right now
[12:19] Andrew Linden: No, I don't believe map tile texture assets are treated like baked textures.
[12:19] Ardy Lay: Uhm,
[12:20] Ardy Lay: Are map tiles treades like textures that residents uploaded?
[12:20] Ardy Lay: -treades +treated
[12:20] Pauline Darkfury: Map textures are currently like any other texture in the system, from an end user point of view
[12:20] Andrew Linden: Yes, I think they are.
[12:20] Ardy Lay: I guess what I am asking is, do they use the same asset store system?
[12:20] Andrew Linden: Yes.
[12:21] Ardy Lay: Are they expunged regulary?
[12:21] Kaluura Boa: Yeah... I remember that LL wanted to break the system because nobody was paying for these textures... Or something like that.
[12:21] Andrew Linden: No, I don't think map tiles are garbage collected in a special way.
[12:21] Liisa Runo sends another 10L to governor
[12:21] Kallista Destiny: sorry I'm late
[12:21] Andrew Linden: So, if you were to put an old map tile on a box... it should always be available when that box is viewed.
[12:22] Pauline Darkfury: They map textures are just 256x256, so "small" data size compared to other texture assets
[12:22] Kaluura Boa: Yes. There is even a site where you can retrieve years-old maps from all the regions...
[12:22] Ardy Lay: They get updated updated about daily?
[12:22] Kallista Destiny: OH?
[12:22] Andrew Linden: Yes, the map tiles are updated daily.
[12:23] Pauline Darkfury: gridsurvey.com lets you see ancient map textures, and I don't think they store the actual image, just the UUID, then pull the current from LL's web texture viewing service
[12:23] Kaluura Boa: That's it.
[12:23] Ardy Lay: Storage must be .... uhm
[12:23] Kaluura Boa: Storage is handled by LL...
[12:23] Andrew Linden: the name2key issue... that relates to LSL key discovery?
[12:23] Flip Idlemind: SCR-91
[12:23] JIRA-helper: http://jira.secondlife.com/browse/SCR-91
[#SCR-91] llRequestAgentKey (llName2Key)
[12:24] Pauline Darkfury: Yeah, being able to look it up for various commerce applications where it's required to start with a name, interact by key
[12:24] Andrew Linden: Ah, that is a feature request. Hrm...
[12:24] Pauline Darkfury: Most notable is "buy for a friend" on vendors
[12:24] Leonel Iceghost: does LL ever delete textures not used by a given time?
[12:24] Flip Idlemind: I wanna buy this for a friend, but what on earth is a "key"?
[12:24] Flip Idlemind: For example
[12:24] Andrew Linden: Unfortunately, it may be that the simulator itself does not have the key cached the moment the call is made
[12:25] Rex Cronon: there might be a nice can easy way to get the key for any ave. why can't u guys put the user key here: https://my.secondlife.com/user.name
[12:25] Andrew Linden: so the answer that comes back may arrive some time later
[12:25] Rex Cronon: a nice an easy way*
[12:25] Andrew Linden: and you'd need to catch it in an "event"
[12:25] Jonathan Yap: The viewer can make a call to the name cache and get a key
[12:25] Flip Idlemind: The dataserver could do it?
[12:25] Pauline Darkfury: Right, yes, we need a llRequestAgentKey or something like that which calls the central agent DB
[12:26] Rex Cronon: each one of us already has an page, u just need to list the key in that hlml code
[12:26] Pauline Darkfury: Right now, we have an extremely heavy, HTTP call to external DB, which may then hit vwrsearch.secondlife, etc
[12:26] Andrew Linden: Ah I see. The feature request details the "dataserver()" event that would go with the call.
[12:27] Andrew Linden: What do you think about it, Kelly?
[12:27] Kaluura Boa: At least, doing it for user names shouldn't be that hard...
[12:27] Simon Linden: I just distracted Kelly with a different problem
[12:28] Kelly Linden: I've talked about that fairly extensively at the Scripting user group.
[12:28] Kelly Linden: Is there something new you are aiming to talk about?
[12:28] Liisa Runo: yup, and everybody agree, now we just need the ETA
[12:28] Liisa Runo: :P
[12:28] Kelly Linden: There is no current ETA, it is not in development at this time.
[12:28] Kaluura Boa: hehehe
[12:29] Andrew Linden: Oh I see. WorkingOnIt Linden has that issue assigned to them, which means we think we might be able to do it.
[12:29] Andrew Linden: ok
[12:29] Kelly Linden: To recap previous discussions: It would require a dataserver event and is more feasible these days than it was a year ago but is not in development right now and there is no ETA.
[12:29] Andrew Linden: Thanks Kelly.
[12:30] Meeter: Timecheck : User Group is half over
[12:30] Pauline Darkfury: Dataserver is fine for it, the need is to eliminate expensive and/or unreliable off-world DBs for it
[12:31] Leonel Iceghost: I wonder, you can add things to the dataserver event, but cannot to the changed event because it break stuff?
[12:31] Leonel Iceghost: how are they so different
[12:32] Pauline Darkfury: Look at the params, Leonel
[12:32] Pauline Darkfury: changed is a bitmask, although I do strongly believe that bad code deserves to break, if it doesn't do its bitmasks properly, tbh
[12:33] Pauline Darkfury: The more valid argument with changed is that there's a max of 32 possible change types for it, so it's a very limited facility
[12:33] Pauline Darkfury: Dataserver is unlimited in scope, as it responds with (request_key, data)
[12:33] Rex Cronon: it is possible that there r some scripts that use radom values to test bitmasks?
[12:34] Leonel Iceghost: 32? how many are now, a half used?
[12:34] Andrew Linden: Oh. The dataserver() event already exists. Apparently I don't code much LSL these days.
[12:34] Flip Idlemind: If you wanna get super creative with a timer you could invent your own "changed2()" event (I've done this <.<)
[12:34] Leonel Iceghost: Flip, it was discused to make a "going to restart this sim" changed event
[12:35] Leonel Iceghost: and then it couldn't be done because it breaks scripts
[12:35] Flip Idlemind: Actually...hold on
[12:35] Flip Idlemind: My opinion on that is, llRequestSimulatorData
[12:35] Ardy Lay: A lot of old vendors have a bad comparison in them that make them reset themselves on any changed() event.
[12:36] Kaluura Boa: I'd like that... The restart messages are failing half of the time nowadays... No blue popup...
[12:36] Pauline Darkfury: Frankly, tough, those creators should learn to code
[12:36] Kaluura Boa: These vendors deserve to break... Really crappy scripting...
[12:36] Leonel Iceghost: ++Pauline
[12:36] Pauline Darkfury: If you know what vendors those are, have fun sitting on them ;)
[12:36] Kaluura Boa: hehehe
[12:37] Kallista Destiny: Unfortunatly the people that made the vendors wonlt be the ones who suffer
[12:37] Ardy Lay: I would tend to agree but LL was implored to roll back a service improvement to accomodate the bad LSL script.
[12:37] Kaluura Boa: That sounds very wrong.
[12:38] Pauline Darkfury: Well, the people using those vendors really need to move onto something written by a competent coder. Handling bitmasks cleanly is really coding 101. If they can't get something as simple as that right, trusting them with L$ is a big mistake
[12:38] Ardy Lay: One of the things the vendor did was change textures on a face when the scripts resets.
[12:38] Ardy Lay: LL added a changed() event for texture changing.
[12:38] Kallista Destiny: lols
[12:38] Kaluura Boa: I like this one. =^_^=
[12:38] Fancy Detector: Andrew Linden has arrived!
[12:38] Fancy Detector: Kelly Linden has arrived!
[12:38] Fancy Detector: Simon Linden has arrived!
[12:38] Ardy Lay: The result was regions filled with looping script resets.
[12:38] Flip Idlemind: Got disconnected .()
[12:39] Rex Cronon: when working with bits, it can be quite easy to mess them up. doesn't really mattery how good u r:)
[12:39] Pauline Darkfury: (change & CHANGED_THING) is not difficult
[12:39] Kallista Destiny: That is why Kerrinigan invented #inculdes
[12:39] Pauline Darkfury: (change & (CHANGED_THING1 | CHANGED_THING2)) is also not difficult
[12:39] Leonel Iceghost: so we are stuck with no more new changed events forever because they are bits?
[12:40] Kaluura Boa: No, just because these old venders will break.
[12:40] Kaluura Boa: vendors*
[12:40] Andrew Linden: :-( my primary mouse just died
[12:40] JoyofRLC Acker: why not just add a ling intgere type?
[12:40] Andrew Linden moves to secondary mouse
[12:40] JoyofRLC Acker: cough
[12:40] Pauline Darkfury: Break the vendors, embarass the terrible coders, move forward
[12:40] JoyofRLC Acker: erm is this the sevrer group?
[12:40] Ardy Lay: The scripter used the wrong operator.
[12:41] Andrew Linden: Welcome to Server User Group JoyofRLC. Got a topic?
[12:41] Rex Cronon: i hope u aren't using wireless mouse, andrew
[12:41] Ardy Lay: I have olso been told the scripter no longer gives a $*)#&$ about SL.
[12:41] Ardy Lay: Not sure if that's true.
[12:41] JoyofRLC Acker: well sort of a question but i was curious about the chat - sorry i arrived a little late
[12:41] Kallista Destiny: That would only double the number of events.... for mucho code breakage
[12:41] Pauline Darkfury: Well, the people using those dreadful scripts need to find a decent scripter and replace them
[12:41] JoyofRLC Acker: i was looking for a sort of server update
[12:41] Rex Cronon: those things can become useless if they batteries die
[12:42] Kallista Destiny: The question is 'Who was the created of these brainbammaged scripts?'
[12:42] Rex Cronon: their*
[12:42] Flip Idlemind: Back in the "old days" there was talk of "script versioning" (But that was back when C# was still on the table)
[12:42] Andrew Linden: Ah, JoyofRLC, you're wondering about region crossing issues, right?
[12:42] Flip Idlemind: Not that the two are necessarily related, just an idea of when it was talked about
[12:43] Ardy Lay: Each affected vendor entered an endless reset loop that continued until the region simulator code was changed to stop it.
[12:43] JoyofRLC Acker: yes andrew. i have some anecodtal observations but wondered if you had some hard facts to update us
[12:43] Pauline Darkfury: Yeah, the ideal thing would be llSetAPIVersion() which could be called after script start to enable stuff which breaks bad code
[12:43] Leonel Iceghost: that would be awersome Flip, we already have two versions.. I wouldn't mean to change the option to LSL 3 for new scripts
[12:44] Andrew Linden: JoyofRLC, the kernel upgrade is supposed to start this week (I think)
[12:44] Pauline Darkfury: Anyway, a more urgent broken scripts thing, we need llPosJump()
[12:44] Pauline Darkfury: SVC-7265
[12:44] JIRA-helper: http://jira.secondlife.com/browse/SVC-7265
[#SVC-7265] PosJump breaks if object-entry is disabled
[12:44] Kaluura Boa: Another week or rolling restarts... Oh joy! -.-
[12:44] JoyofRLC Acker: ok sounds good. could i back up a level though?
[12:44] JoyofRLC Acker: yes, kaluura?
[12:45] JoyofRLC Acker: lol
[12:45] Pauline Darkfury: We also need the current broken behaviour with the current PosJump fixed
[12:45] Andrew Linden: which is the real thing we're hoping will bring region crossings to their previous level of reliability (pre mid Oct 2011)
[12:45] JoyofRLC Acker: yo have made some improvements arelady yes?
[12:45] Ardy Lay: Heh, we could go back to "black Wednesdays".
[12:45] Kallista Destiny: Yeah, no thanks
[12:45] Kaluura Boa: ...and withdrawal symptoms...
[12:45] JoyofRLC Acker: Blake has been much better lately
[12:45] Ardy Lay: Hehe
[12:45] JoyofRLC Acker: btu i gtaher other areas nto so much
[12:46] Kallista Destiny: and a new viewer each week... nevermind
[12:46] JoyofRLC Acker: the changes you made so far were just hardward ... sims on adjacent serves kind of thing?
[12:46] Jonathan Yap: That did make for easy server protocol changes though
[12:46] Andrew Linden: Yes JoyofRLC, but that work didn't fix everything
[12:46] JoyofRLC Acker: nods
[12:46] JoyofRLC Acker: whagt I have been finding is ...
[12:46] JoyofRLC Acker: in blake and round about ...
[12:47] JoyofRLC Acker: sim crssings MUCH better
[12:47] Ardy Lay: Not to mention the old viewer was a 42+ megabyte download because it included all the textures used at the new resident orientation locations.
[12:47] JoyofRLC Acker: but other perfomane issues still around and some new ones
[12:47] JoyofRLC Acker: performance mid sim is an issue sometimes
[12:47] JoyofRLC Acker: sort of lag but with some differences i think, hard to describe
[12:48] JoyofRLC Acker: should we just wait till the new kernal is out and its all finished?
[12:48] Andrew Linden: No. If you could try to describe the problem I'd like to hear it.
[12:48] Andrew Linden: An example region name would be helpful too.
[12:48] JoyofRLC Acker: well theres different issues
[12:49] Andrew Linden: If it is a new problem I wonder if it is associated with a particular server channel.
[12:49] JoyofRLC Acker: some of the things i thought i only saw at a sim border seem to happen mid sim sometimes
[12:49] JoyofRLC Acker: submarining for example
[12:49] JoyofRLC Acker: rubber banding
[12:50] Kallista Destiny: Ardy latency from FSO to DCA can be horrible
[12:50] JoyofRLC Acker: things are a bit more 'jerky' mid sim than they sued to be
[12:50] JoyofRLC Acker: also
[12:50] Ardy Lay: JoyofRLC Acker, try disabling evelope > Network > Velocity iInterpolate Objects and see if you still submarine mid-region.
[12:50] JoyofRLC Acker: some sort of time dialtion i think (cant be sure dont have the stats bar up usauly when racing)
[12:50] Andrew Linden: Ok. Rubber banding can becaused by several things. The two that come to mind are: (1) lag spikes in the simulator and (2) network problems between you and SL servers
[12:51] JoyofRLC Acker: but basically boats slow down ... but other things seem quite responsive
[12:51] Liisa Runo: (multi gigabit optical network is not really that fast cause all the data is routed trought the old 80286 operating as rounding error server)
[12:51] Kallista Destiny: Sounds like Physics issus
[12:51] Andrew Linden: Really? Other things are reponsive? You mean, like scripts? other moving objects? or something else?
[12:51] JoyofRLC Acker: im am not up on the tech or terminolgy
[12:52] JoyofRLC Acker: the lag im used to ... everything slows or stops
[12:52] JoyofRLC Acker: you hist a key
[12:52] JoyofRLC Acker: nothinghappens
[12:52] Kallista Destiny: Joy have you tryed watching the Stats meter <CTRL+SHIFT+1>
[12:52] JoyofRLC Acker: but a few times now ... the boat or boats slow
[12:53] JoyofRLC Acker: but you can move the sails etc just fine
[12:53] JoyofRLC Acker: its very weird
[12:53] AnnMarie Otoole: My experience is the physics engine goes on holidays when the sim has excessive physical objects.
[12:53] Andrew Linden: Ah, that "everything stops" thing sounds like viewer lag.
[12:53] JoyofRLC Acker: as i said i dont usually have that up when sailing - i dont have 52" scfreen LOL
[12:53] JoyofRLC Acker: perhpas i should do some test8ng
[12:54] AnnMarie Otoole: My vehicles run off the road more in sims overloaded with horses or chickens.
[12:54] Kallista Destiny: Neither do I, but it's hard to diagnose problems without solid data
[12:54] JoyofRLC Acker: anyway .... id summarise it as saying --- sim borders - crash rates waay down
[12:54] JoyofRLC Acker: lag - def patchy
[12:54] Flip Idlemind: "Lag" is so poorly defined / understood. Sometimes, someone's viewer framerate will be low, so they'll ask people to take scripts off .()
[12:54] Andrew Linden: If everything stops while you're actually on something that is moving (boat) or your avatar is moving (walking/flying/falling) then it is probably viewer lag.
[12:54] Flip Idlemind: Not relevant, just saying
[12:54] Andrew Linden: Otherwise, things rubberband when moving and it is caused by server lag, or lost packets.
[12:54] JoyofRLC Acker: i think its server lag
[12:54] JoyofRLC Acker: forone thing other boats report the same phenomena
[12:55] Andrew Linden: Hrm... ok, and these lag events are "new" (not noticed as often before).
[12:55] Meeter: Timecheck : User Group is almost over
[12:55] Andrew Linden: got an example region name?
[12:55] JoyofRLC Acker: well... thats hard to answer
[12:55] Jonathan Yap: I would like to see a limit to how long you can rubberband for. I have watched myself get to strange coordinates like 1002, -655 when there is server packet issues, etc
[12:55] JoyofRLC Acker: hmm
[12:55] JoyofRLC Acker: let me put it this way ...
[12:56] JoyofRLC Acker: lag is worse than it used to be
[12:56] JoyofRLC Acker: but prob no woirse than it has been the last few weeks
[12:56] Andrew Linden: Jonathan, I thought Simon had added a rubberband limit. What viewer are you using/
[12:56] Jonathan Yap: Not the current viewer Andrew
[12:56] Kallista Destiny: laughs I've been to 5K on occasion recovered too
[12:57] JoyofRLC Acker: ... i see strange coords quite often
[12:57] Flip Idlemind: I think Viewer 2 has had such a limit for quite a while now
[12:57] Jonathan Yap: If it was a server change though, would my viewer version make a difference?
[12:57] Andrew Linden: Yeah, it was added to V2 several months, or a year, ago.
[12:57] Flip Idlemind: And you can't "rubberband" through terrain, that's cool
[12:57] Kallista Destiny: I so hated being a Uboat
[12:58] Andrew Linden: Well, JoyofRLC I'd recommend waiting for the kernel upgrade to arrive.
[12:58] JoyofRLC Acker: we can go miles underground lol
[12:58] JoyofRLC Acker: i hear you
[12:58] Andrew Linden: If you do see it bad, the best thing to do is to note the time + region name, and bring those here.
[12:58] JoyofRLC Acker: the hardward changes andrew - they were only for some areas?
[12:58] Pauline Darkfury: I think my record was Z=-10,000 ;)
[12:58] JoyofRLC Acker: not grid wide?
[12:58] Liisa Runo: and watch the stats next time you go sailing, gives you lil idea of the cause and the type of the lag you are experiencing
[12:59] Qie Niangao: (Before time runs out, just a reminder about the Google Map API thing, and whether the Lab will be getting billed in 2012.)
[12:59] Andrew Linden: I'm unlikely to dig if I don't have a good approximate time and region name... there are just too many logs to dig through otherwise.
[12:59] JoyofRLC Acker: we are trying to get some info together for you andrew
[12:59] Kallista Destiny: yeah 30+K regions is a LOT of logs
[12:59] JoyofRLC Acker: of course
[12:59] Andrew Linden: We tried to do some hardware jiggling for the whole world at once... and hit some problems.
[13:00] Andrew Linden: So we made some smaller conversions, Blake Sea in particular.
[13:00] Kallista Destiny snickers "Nice understatement"
[13:00] JoyofRLC Acker: i think fanci maybe filing a toicket in regard to some of the SC simes yestefay - i have to get her a notecrd with the detauils though
[13:00] Andrew Linden: And then have been working to finish the juggling in stages.
[13:00] Meeter: Thank you for coming to the Server User Group
[13:00] JoyofRLC Acker: yes its blake where the dichotomy between sim xong and general lag is most marked
[13:00] Andrew Linden: I don't know if they are done, or still in progress.
[13:01] Jonathan Yap: Andrew, were you able to see improvements once Blake Sea was rearranged?
[13:01] Andrew Linden: Yes Jonathan. According to annecdotal evidence (such as JoyofRLC stated here) there were improvements in Blake Sea.
[13:02] JoyofRLC Acker: Jonathan - Id say xings much better. "lag" not much better.
[13:02] JoyofRLC Acker: unsure about precise nagture of "lag" atm tho.
[13:02] Jonathan Yap: Crossings are what you would notice being improved
[13:02] Andrew Linden: I've got to go. Thanks for coming everyone.
[13:02] JoyofRLC Acker: Yes. Not crashing is very noticeable.
[13:02] JoyofRLC Acker: roflol
[13:03] Kallista Destiny: THanks andrew
[13:03] Pauline Darkfury: Thanks, Lindens, have a good afternoon :)
[13:03] Rex Cronon: tc andrew
[13:03] JoyofRLC Acker: Seriously, Thank you!
[13:03] Qie Niangao: Thanks Andrew, all.
[13:03] Liisa Runo: thanks all
[13:03] JoyofRLC Acker: To be clear, tho, there are reports from elsewhere that things are still bad.
[13:03] Rex Cronon: tc everybody
| Prev 2011.11.22 | Next 2011.12.02 |