Chat log from the meeting on 2014-08-05

From OpenSimulator

Jump to: navigation, search
[11:00]  Nebadon.Izumi hello
[11:00]  BlueWall.Slade hey neb
[11:00]  Dev Random: hello Nebadon. enjoy your sit.
[11:00]  BlueWall.Slade hey RiRa
[11:01]  Richardus Raymaker: hi bluewall all
[11:01]  Nebadon.Izumi :)
[11:01]  Cuteulala Artis is Offline
[11:01]  Justin.Clark-Casey is Online
[11:02]  Mata.Hari hi Justin
[11:02]  Justin.Clark-Casey hi folks
[11:02]  BlueWall.Slade Hi justin
[11:02]  Nebadon.Izumi just got done updating the OSCC grid to master git
[11:02]  Nebadon.Izumi logging in my bots now
[11:02]  Mata.Hari I was looking at Mantis 7259 ( and question whether it really should have been closed. I completely disagree with Sylvia that it's "not a bug" because it *is* a bug, but it's documented. In a 1.0 release either the ideal case: the restriction against two adjacent regions being different sizes should be removed; or there should be some built-in mechanism to prevent two dissimilar-sized regions from being created adjacent to one another.
[11:02]  Mata.Hari The average user isn't going to go plumbing the depths of wiki to find out why they suddenly can't walk between two regions (but *can* tp between them)...they're going to instantly assume it's a bug. And it is. And even if we assume Sylvia's comment about "you should have read the wiki first" isn't intended to be snarky, who's to say that someone won't plunk down a var beside one of my regular regions and then I would somehow magically be expected to have familiarized myself with var regions in order to understand why I can't walk over to visit my new neighbour.
[11:03]  Justin.Clark-Casey nebadon: I just this second may have found out why pCampbot is crashing
[11:03]  Nebadon.Izumi nice
[11:03]  Richardus Raymaker: hi justin.
[11:03]  Nebadon.Izumi something on the client side i hope?
[11:03]  Richardus Raymaker: strange sicnme 0.8 i havenmot got problems with var -> normal var. but this private. hmm
[11:03]  Nebadon.Izumi im using 2 computers for my bots today, 40 on each machine
[11:03]  Justin.Clark-Casey I believe it's connected with seting the connection limit on the service point. It causes a problem in an unepxected place internally within mono
[11:03]  Nebadon.Izumi one is linux one is windows
[11:03]  Justin.Clark-Casey so this may be technically a mono bug
[11:04]  Nebadon.Izumi oh that sucks
[11:04]  Richardus Raymaker: better use windows :)
[11:04]  Justin.Clark-Casey I can get it to stop spewing a certain class of failure if I disable chnging that limit, but then it' sback to only 2 simultaneous connections to a service point until the code is restructed to reintroduce it without causing other issuers
[11:05]  Dahlia Trimble is Online
[11:05]  Justin.Clark-Casey like a massive amount of other timeout I'm seeing on caps - could be that libomv caps system is too aggressive with simultaneous requests - I've already addressed a similar issue on read
[11:05]  Justin.Clark-Casey ho hum
[11:05]  Nebadon.Izumi ok
[11:05]  Nebadon.Izumi well it sounds like progress even if your not sure on a fix
[11:05]  Ubit Umarov is Online
[11:05]  Nebadon.Izumi so that is good
[11:06]  Sheera.Khan Hiyas :-)
[11:06]  BlueWall.Slade o/
[11:06]  Nebadon.Izumi hello
[11:06]  Billy.Bradshaw Hi Sheera
[11:08]  Justin.Clark-Casey mata: maybe, you should ask kenvc
[11:08]  Kayaker Magic: I saw a bunch of mantis emails last week, looked like old issues that had been fixed but never marked closed until now. Someone cleaning house?
[11:08]  Nebadon.Izumi yes
[11:08]  Nebadon.Izumi it was a script that did some cleanup
[11:08]  Nebadon.Izumi nothing to be concerned about
[11:08]  Nebadon.Izumi it closed aprox 900 resolved manti
[11:08]  Latif Khalifa is Online
[11:09]  Kayaker Magic: Ah, but I only got emails for ones I had posted to, so for a minute I though someone was fixing all MY bugs!
[11:09]  Kayaker Magic: LOL
[11:09]  Mata.Hari I guess my point is the whole point of Mantis is to have a place to report issues that shouldn't be there in a 1.0 release and leave them open there until they're fixed
[11:09]  Nebadon.Izumi nope, ya anything you ever commented on
[11:09]  Nebadon.Izumi you will receive any update to the ticket
[11:10]  Dahlia Trimble: hi
[11:10]  Latif Khalifa: hello
[11:10]  BlueWall.Slade Hi Dahlia
[11:10]  BlueWall.Slade Hi Latif
[11:11]  Nebadon.Izumi hello
[11:11]  Richardus Raymaker: hi all
[11:11]  Mata.Hari and when I encounter one, I go to Mantis to find out if someone has reported it; I don't go to see if wiki has a report
[11:11]  Billy.Bradshaw Justin, you disabled AvatarAppearance messages during last weeks load test. Wondered if your findings similar to ours?
[11:11]  Justin.Clark-Casey Billy: I did a range of things.
[11:12]  Justin.Clark-Casey so it's not yet possible to determine what did what
[11:13]  Billy.Bradshaw Our concern was with the viewer, the message sends VisualParams, that for 100+ avi could cause pproblems
[11:13]  Justin.Clark-Casey have you made measurements?
[11:14]  Billy.Bradshaw some, but at a network level. There maybe othe results I will check
[11:15]  Justin.Clark-Casey need published data
[11:15]  Billy.Bradshaw agree
[11:16]  Sarah Kline is Online
[11:18]  Sarah Kline: hi
[11:18]  Nebadon.Izumi anyone have anything else they wanted to talk about this week?
[11:18]  Nebadon.Izumi hello Sarah
[11:18]  Vivian Klees: hi sarah
[11:18]  Nebadon.Izumi i just about have 40 bots logged into load test already
[11:18]  Nebadon.Izumi so much easier using 2 computers
[11:19]  Nebadon.Izumi so quiet this week :)
[11:19]  Richardus Raymaker: vacation time
[11:20]  Billy.Bradshaw abitary question, why does OSGrid avis not include a suitcase?
[11:20]  Nebadon.Izumi because of how we configured the grid
[11:21]  Nebadon.Izumi the suitcae is meant to really provide an extra layer of protection for more private grids
[11:21]  Nebadon.Izumi where the grid operator also manages and operates all the regions themselves
[11:21]  Nebadon.Izumi in an open grid like OSgrid it provides no protection since the grid is open already
[11:22]  Nebadon.Izumi its only really useful in grids that are not open where anyone can host their own region
[11:22]  Mata.Hari @Dahlia: do you have any guestimate on implementing the advanced mats part of get/set params? I have a project I want to work on and if it's imminent I'll put it off until it's working....if it's more likely to be months away then I'll do the project without it
[11:25]  Dahlia Trimble: Mata, sorry, no, there are some issues with it that need to be resolved forst
[11:25]  Dahlia Trimble: *first
[11:25]  Dahlia Trimble: there is a problem with the way materials are stored
[11:26]  Mata.Hari probably better I just go ahead and assume it's a long time away (if ever)
[11:26]  Dahlia Trimble: for now, yes unfortunately
[11:27]  Cuteulala Artis is Online
[11:27]  Andrew Hellershanks is Online
[11:28]  Mata.Hari perhaps that also explains where there are sometimes some odd quirks when editing them with the viewer too
[11:28]  Andrew Hellershanks: I was working on something and almost forgot about the meeting.
[11:28]  Nebadon.Izumi hello
[11:29]  Dahlia Trimble: Mata, materials use some protocol hacks that aren't really well suited for real time editing
[11:29]  Dahlia Trimble blames LL
[11:30]  Dahlia Trimble: so yes I suspect there will bee occasional quirks while editing
[11:30]  Mata.Hari which is funny since real-time editing was one of their big "selling features" :p
[11:30]  Andrew Hellershanks is Offline
[11:30]  Andrew Hellershanks is Online
[11:30]  Mata.Hari it's disappointing but such is life
[11:31]  Dahlia Trimble: why LL did the materials protocol the way they did is beyond me but I dont think it was very well thought out
[11:31]  Andrew Hellershanks: Dang viewer still crashes when switching workspaces. :P
[11:31]  Mata.Hari maybe something they decided to add later in the design process and had to shoehorn in there
[11:32]  Sarah Kline: It was a TPV that worked on it. as an addition
[11:32]  Dahlia Trimble: it was pushed on them by one of the TPV groups but I dont know who was responsible for the protocol
[11:34]  Mata.Hari oh...the issue is with the advanced mats part, rather than just with real-time edits in general....gatcha
[11:34]  Latif Khalifa is Offline
[11:34]  Dahlia Trimble: anyway it's pretty complex and ugly for something that could have been very simple
[11:35]  Mata.Hari I guess that also means bump-mapping, SSS, etc aren't likely to get added anytime either
[11:35]  Latif Khalifa is Online
[11:36]  Dahlia Trimble: dont know wny you need bump if you have normal
[11:36]  Mata.Hari they map differently in a render
[11:36]  Dahlia Trimble: but the materials are extensible so a viewer could add bump if they chose to
[11:38]  Kayaker Magic: Neb, I have a gadget to give away, shows mantis 6743 bug, can I give it to you to put on the table to hive out copies?
[11:38]  Nebadon.Izumi sure
[11:38]  Dahlia Trimble: what would be bad is if one viewer added new materials/shader types and another did something different
[11:39]  Kayaker Magic gave you Turning Card Compass.
[11:39]  Mata.Hari yes, that would be a nightmare
[11:39]  albertlr Landar is Online
[11:39]  Nebadon.Izumi it gives an error
[11:39]  Dahlia Trimble: it's possible with the materials system
[11:39]  Nebadon.Izumi looks like its missing content inside
[11:39]  Nebadon.Izumi when you click it
[11:40]  Mata.Hari would be amazing to have Blender patched into a viewer and somehow manage real-time renders :p
[11:40]  Mata.Hari probably have to borrow CERN's computing facilities to achieve that though
[11:41]  Nebadon.Izumi ya that would be very difficult to do something like that
[11:41]  Kayaker Magic: This compass Neb is putting out, it works if you attach it to a vehicle, it works if you wear it as an attachment and walk around, but it fails if you wear it as an attachment and sit in a vehicle (or a chair) in avatar direction mode.
[11:41]  Nebadon.Izumi not so much you need a lot of computing power
[11:41]  Mata.Hari complete with full ray-tracingg!
[11:41]  Nebadon.Izumi would require a lot of man hours in programming
[11:41]  Dahlia Trimble: Mata, well good luck waiting for that :P
[11:41]  Kayaker Magic: This means you cannot use it as a hud in a vehicle.
[11:41]  Mata.Hari lol...yeah....need quantum processors first
[11:41]  Kayaker Magic: Oh, it is missing a help file if you touch it unattached.
[11:42]  Nebadon.Izumi ya no biggie
[11:42]  Kayaker Magic: Touch it while worn as a hud and it switches between two modes: Heading and bearing.
[11:42]  Kayaker Magic: Bearing fails to work because of mantis 6743
[11:43]  Nebadon.Izumi probably something with the linkset order
[11:43]  Nebadon.Izumi things change when avatar sits
[11:43]  mattie mcbride is Online
[11:44]  Nebadon.Izumi does it work if its linked as a non-root prim?
[11:44]  Kayaker Magic: yes, in a vehicle
[11:44]  Nebadon.Izumi k wierd
[11:44]  BlueWall.Slade I think avatars are added to the end of the link-chain
[11:44]  Kayaker Magic: Yes, that was a fix a while ago
[11:44]  BlueWall.Slade so the root prim should stay the same
[11:44]  Andrew Hellershanks: How do I get the compass?
[11:45]  Mata.Hari oh...that reminds me of an odd bug I encountered this week but haven't rested enough yet to report.....I was doing a script to change the colour of the headband that is part of my hair....when I check the link number using FS and set that in the script, it changes on of the hair flexi-prims instead
[11:45]  Nebadon.Izumi there
[11:45]  Nebadon.Izumi i made it copyable
[11:45]  Andrew Hellershanks: ty
[11:45]  Mata.Hari if I use llGetPrimitiveParams to find out the link number if gives a completely different result (which *will* work with the script)
[11:46]  Justin.Clark-Casey is Offline
[11:46]  Justin.Clark-Casey is Online
[11:46]  Kayaker Magic: Is that a FS viewer bug?
[11:46]  Mata.Hari I encountered that on 2 different hairs, but otherwise I've never seen that
[11:46]  Simulator Version v0.5 shouts: OpenSim 0.8.1 Dev          de3421b: 2014-08-02 08:21:25 -0400 (Unix/Mono)
[11:46]  vegaslon.plutonian it is the same in singularity
[11:46]  Mata.Hari could be FS.
[11:47]  vegaslon.plutonian the veiwer reported link number is diffrent then the link number in opensim
[11:47]  Justin.Clark-Casey crashed. Kind of about sums it up right now]
[11:47]  Kayaker Magic: I've had so much trouble with link numbers changing that I always script to search for the prim I want.
[11:47]  TBG Renfold: same here
[11:47]  Kayaker Magic: by name
[11:47]  Richardus Raymaker: thats how i solved it always
[11:47]  Andrew Hellershanks: iOdd. CoolVL doesn't seem to be putting the item in to my inventory.
[11:47]  Mata.Hari yeah, that was my only solution too
[11:47]  Mata.Hari just wondering *why* ut happens
[11:48]  Mata.Hari (and i don't do enough building in SL to know if the bug exists there too)
[11:49]  Richardus Raymaker: sofar i have not seen it. but maby did not do the same with link numbers
[11:49]  Andrew Hellershanks: I haven't had an issue with link numbers changing in SL.
[11:49]  Billy.Bradshaw justin: our code base has been changed to such an extent that our measurement would not be I am told.
[11:49]  Justin.Clark-Casey you can still publish data
[11:49]  Kayaker Magic: The link numbers change every time you add a new prim to a build
[11:49]  Mata.Hari right
[11:49]  Kayaker Magic: that is documented i the wiki
[11:49]  Billy.Bradshaw I will ask
[11:50]  Richardus Raymaker: Technical if its the last prim you add it must be added to the existing link set
[11:50]  Mata.Hari but once you have a static build shouldn't the number be correctly reported in the viewer?
[11:50]  Kayaker Magic: So if you put fixed link numbers in your script, they will fail if you change anything
[11:50]  Andrew Hellershanks: Kayaker, You are finding link numbers changing in recent code? Are you adding the prim to the end of an existing linkset?
[11:50]  BlueWall.Slade if you select the new prim, then the existing linkset - it should add the new prim to the end of the list?
[11:50]  Mata.Hari this isn't a case of the number of links ever changing, though
[11:50]  Nebadon.Izumi grr
[11:50]  Seeress Seidkona is Online
[11:50]  Kayaker Magic: I think new prims go on the start of the chain...
[11:51]  Nebadon.Izumi Radegast doesnt run well under linux
[11:51]  Andrew Hellershanks: You have to select new prim then original root set to make the number of the new prim 1 after the last numbered prim in the original link set.
[11:51]  Nebadon.Izumi all 22 of my radegast viewers crashed
[11:51]  BlueWall.Slade last prim selected should be root.
[11:51]  Nebadon.Izumi oh well
[11:51]  Kayaker Magic: Yes, once you stop adding links, they say constant.
[11:51]  Nebadon.Izumi was worht a try
[11:51]  Richardus Raymaker: yes bluewall, thats how i see it in sl sofar i have watched it
[11:51]  Nebadon.Izumi i can run 80 clients on similar hardware under windows
[11:51]  Latif Khalifa: Nebandon, blame mono
[11:51]  Nebadon.Izumi yea
[11:51]  Richardus Raymaker: mono su*ks
[11:51]  Andrew Hellershanks: Kayaker, I have a prim/script to test link numbers. I'll see if it has broken again.
[11:51]  Richardus Raymaker: that i know for 2 years
[11:52]  Mata.Hari they stay constant but sometimes aren't reported correctly in the viewer if there are a large number....I question is whether that's a viewer bug or an Opensim bug
[11:52]  Nebadon.Izumi man i cant beleive they all crashed
[11:52]  Nebadon.Izumi thats probably going to skew todays results
[11:52]  Nebadon.Izumi sigh
[11:52]  Kayaker Magic: I don't think it is broken, the LL way just changes all the numbers when adding a new prim
[11:52]  Latif Khalifa: Nebandon ,why did you switch from Windows?
[11:52]  Andrew Hellershanks: Kayaker, can you can me a copy of the compass? CoolVL won't let me take a copy of it.
[11:52]  Justin.Clark-Casey if that'swhat the viewer is expecting then that might be what's going wrong
[11:53]  Nebadon.Izumi i wanted to split the load up between computers
[11:53]  Nebadon.Izumi that computer is running linux at the moment
[11:53]  Latif Khalifa: link numbers don't matter to the viewer
[11:53]  BlueWall.Slade if you select the existing work, then select the new prim last, then the new prim will be the root
[11:53]  Latif Khalifa: in fact the viewer doesn't even know link numbers exist
[11:53]  Andrew Hellershanks: Kayaker, ty
[11:53]  Latif Khalifa: it only matters to some poorly written LSL
[11:54]  Kayaker Magic: lol
[11:54]  Latif Khalifa: One should not write scipts depending on the link numbers anyway
[11:54]  Snoopy Pfeffer is Online
[11:54]  Kayaker Magic: If you select a new prim, then an existing build and link, the new prim always becomes link #2. It's in the wiki
[11:54]  Richardus Raymaker: no you select the new prim first and then the old link set
[11:54]  Andrew Hellershanks: Latif, If you think that way, I'll get you to put a number in the description of 254 prims in a link set.
[11:55]  Mata.Hari yeah
[11:55]  Latif Khalifa: how do you know that Kayaker?
[11:55]  Mata.Hari plus have to do a while{} call to find the one you're looking for every time
[11:55]  Latif Khalifa: Andrew, do you have to do different stuff on all 254?
[11:55]  Kayaker Magic: Read about in
[11:55]  Mata.Hari or on attach have it make a list of them all and hold it in memory
[11:55]  Andrew Hellershanks: Kayaker, that is how it used to work. Melanie (IIRC) put in a fix ages ago that addressed that.
[11:56]  Latif Khalifa: Mata, every time? Ever heard of caching a number?
[11:56]  Kayaker Magic: Yes, OpenSim does it right!
[11:56]  Mata.Hari no point caching if the number changes
[11:56]  Kayaker Magic: But the right way does change all the numbers every time you link. Just warning you.
[11:56]  Andrew Hellershanks: huh?
[11:56]  Latif Khalifa: some lazy LSL coders here I see
[11:57]  Andrew Hellershanks: Either the numbers change or they don't. THe new link set used to be inserted after the root then the rest of the child prims were added after the new link set but that was fixed ages ago.
[11:57]  Justin.Clark-Casey I agree I woujldn't use link num as a programmer. BUt unfortunately, if the system provides something ppl end up using it
[11:57]  Latif Khalifa: I think OpenSIm should randomize link numbers just to teach tham a lesson :P
[11:57]  Andrew Hellershanks: I've not seen a problem with link numbers changing for ages.
[11:57]  Kayaker Magic: LOL
[11:57]  Kayaker Magic: Link numbers are SUPPOSED to change when you add new prims.
[11:58]  Andrew Hellershanks: Using link numbers is a very handy feature. I have an object wtih 254 prims. 200 of them need to be in a given order. Before I changed the script to not require the script in the 200 prims I would have had to renumber them evertime I changed the script.
[11:59]  Andrew Hellershanks: I would have given up working on the device if I had to keep doing that.
[11:59]  ladyjo martin is Offline
[11:59]  Latif Khalifa: or you could make a script to change descriptions of each prim
[11:59]  Latif Khalifa: it would have taken you 1 min flat
[11:59]  Latif Khalifa: and solve the problem forever
[12:00]  Justin.Clark-Casey alright, well it's 8pm
[12:00]  Kayaker Magic: I think we have a communication problem between the carbon based units/
[12:00]  Andrew Hellershanks: Hm... that SL wiki page explains why someone implemented linking primsets in OS the way it used to be done. It hasn't worked that way in the years I've been in SL.
[12:00]  Nebs Metal Bar Stool v1.5 (w/sit & launch): Goodbye..
[12:00]  Dahlia Trimble: load test today? linkie for it?
[12:00]  Mata.Hari the ugly bags of mostly water?
[12:00]  Nebadon.Izumi ok i guess lets jump on over
[12:01]  Kayaker Magic: lol
[12:01]  Nebadon.Izumi Keynote 1
[12:01]  Andrew Hellershanks: Mata, :)
[12:01]  Justin.Clark-Casey I have bots that might sort of wrk without crashing, or might be hilariously inefficient
[12:01]  Nebadon.Izumi lol
[12:01]  Nebadon.Izumi kk
[12:01]  Justin.Clark-Casey I just haven't had time to copy that code to your machine unfortunately neb
[12:01]  Nebadon.Izumi no worries
[12:01]  Justin.Clark-Casey will juhst have to try on mine and try to setup on yours in the bg
[12:01]  Nebadon.Izumi ok well I will see you all there
[12:01]  Nebadon.Izumi head on over
[12:02]  Region found!
[12:02]  Region found!
[12:02]  Teleport completed from https://osgrid/region/Wright%20Plaza/146/149/32
[12:02]  Nebadon Izumi is Online
[12:02]  Connecting to in-world Voice Chat...
[12:02]  Connected
[12:02]  Latif Khalifa: hello
[12:02]  acryline erin: hello
[12:02]  Nebadon.Izumi 0
[12:02]  Nebadon.Izumi hello
[12:03]  Justin.Clark-Casey you still have bots logged on neb?
[12:03]  Dahlia Trimble: hi
[12:03]  Andrew Hellershanks: I can't move.
[12:03]  Dahlia Trimble does a vanity rebake...
[12:04]  Nebadon.Izumi give it a second to settle
[12:04]  Cheopss Forlife7: Hello everyone
[12:04]  acryline erin: coucou
[12:04]  Nebadon.Izumi hello
[12:04]  Justin.Clark-Casey I am going to start logging on bots in 6 mins since it now takes longer as I'm running a 10 sec cpsace
[12:04]  Justin.Clark-Casey space
[12:05]  Andrew Hellershanks: Latif, making a script to number the prims wouldn't help without knowing the order of them. I would have to write a script to check their coordinates and set their numbers.
[12:05]  Cheopss Forlife7: coucou Acry
[12:05]  Nebadon.Izumi ok
[12:05]  Ange Menges: Hi there
[12:05]  Nebadon.Izumi im rebooting my spare machine into windows 7
[12:05]  Nebadon.Izumi i'll try to get my missing 20 bots back too
[12:06]  Mata Hari: tp attempt with my OSG avi timed out so all attachments have to clear cache, purge appearance, log to plaza, load inventory and pray nothing is corrupted
[12:06]  Andrew Hellershanks: Man. I don't know if it is my viewer, my network connection, or this region. Lag city
[12:06]  Justin.Clark-Casey all ppl do around here is complain
[12:06]  Nebadon.Izumi Mata when i logged into Wright Plaza
[12:06]  Nebadon.Izumi i saw a bunch of Yellow Pyramids above your
[12:06]  Andrew Hellershanks: I can't move most of the time
[12:06]  Andrew Hellershanks: I'm going to switch viewer
[12:06]  Nebadon.Izumi kk Andrew
[12:07]  Nebadon.Izumi movement seems fluid here
[12:08]  Crista Lopes: hi. testing 123
[12:08]  Crista Lopes: no chat lag
[12:08]  BlueWall.Slade I wonder why the region your avatar is in renders after the neighbors?
[12:08]  BlueWall.Slade ++ chat is great
[12:08]  Billy.Bradshaw must have a duff network Andrew, seems ok to me
[12:08]  Cheopss Forlife7: hello Grim!
[12:09]  Grimley.Graves hi cheops
[12:09]  vegaslon.plutonian you are created quicker as a child then as a root agent?
[12:09]  Nebadon.Izumi its wierd with Radegast
[12:09]  Nebadon.Izumi when there is this many avatars, an object scan takes minutes
[12:09]  Nebadon.Izumi like 3-4 minutes
[12:09]  Justin.Clark-Casey child agent creation is only triggered after root creation, unless one already exists
[12:09]  Nebadon.Izumi nearly impossible to get my radegast viewers seated with this kind of load
[12:09]  Justin.Clark-Casey but there is a wierd delay in my local pcampbot tests - possibly a connection limit is getting maxed out somewhere
[12:10]  BlueWall.Slade do others see the neighbor regions rezzing first?
[12:10]  Nebadon.Izumi BlueWall that is likely normal
[12:10]  Justin.Clark-Casey ok, logging in bots to kn 4 to bring it up to max
[12:10]  BlueWall.Slade I wonder if there is any way to reverse that?
[12:11]  Crista Lopes: everyone please take a sit
[12:11]  Nebadon.Izumi hello Crista
[12:11]  Crista Lopes: hi
[12:11]  Justin.Clark-Casey I restored the ability to temporarily disable physics when bulletsim is running on a diff thread, so that can be tested if ncessary
[12:11]  Latif Khalifa: Crista has two handsome bodyguards :P
[12:11]  Serene.Jewell Hello all.
[12:11]  Nebadon.Izumi nice
[12:12]  Crista Lopes: oh cool. No physics? Does that mean we'll all float?
[12:12]  Nebadon.Izumi means you cant walk
[12:12]  Nebadon.Izumi lol
[12:12]  Justin.Clark-Casey no, we'll just be stuck in place :)
[12:12]  Justin.Clark-Casey if only it did mean no grav...
[12:12]  Andrew Hellershanks: I'm now on latest singularity
[12:12]  Crista Lopes: argh, someone should to a no-gravity physics
[12:12]  Crista Lopes: hehe lkalif!
[12:12]  Crista Lopes: I feel safer already
[12:12]  Nebadon.Izumi I think you can
[12:12]  Nebadon.Izumi you can adjust gravity
[12:12]  vegaslon.plutonian yep
[12:12]  Nebadon.Izumi you could in ODE anyway
[12:13]  vegaslon.plutonian can turnoff gravity or reverse it direction
[12:13]  Crista Lopes: cool, I'll have to try it with the oculus rift
[12:13]  Andrew Hellershanks: I still can't move around
[12:13]  Nebadon.Izumi wierd Andrew
[12:13]  Nebadon.Izumi lower draw distance maybe
[12:13]  Nebadon.Izumi lot of avatars here
[12:13]  Andrew Hellershanks: I can't get it any lower. It is already at 64m
[12:14]  Nebadon.Izumi wierd, you havent had trouble like that in the past right?
[12:14]  Andrew Hellershanks: Not like this
[12:14]  Kayaker Magic: Neb, I was looking at one of your bots, and it TP'ed to a spot underwater. There's two of them down there now.
[12:14]  Nebadon.Izumi any viewers hung in memory?
[12:14]  Nebadon.Izumi check task manager
[12:14]  Nebadon.Izumi whenever i get wierd frame rates like that
[12:14]  Nebadon.Izumi i usually have a hung viewer
[12:14]  Billy.Bradshaw Andrew, turn off Netflix
[12:14]  Andrew Hellershanks: Its getting a bit better. Seems Singularity and CoolVL get mostly hung up during download of the prims in the region.
[12:15]  BlueWall.Slade +Netflix, lol
[12:15]  Andrew Hellershanks: Billy, hehe. No netflix here.
[12:15]  Justin.Clark-Casey ok, this is going to take a little too long, I'm just going to start loading up 2, 3 and 4 now. I don't think my crash fix will work
[12:15]  Nebadon.Izumi bummer
[12:15]  Andrew Hellershanks: Its a little laggy but I can move
[12:16]  Justin.Clark-Casey it's a very odd failure - a stream is reported as being disposed where that can't really be the case....
[12:17]  Justin.Clark-Casey ah, I might as well try putting bots into kn1 as well, can always disconnect a few if actually reach 400
[12:17]  Andrew Hellershanks: 113 avatars across the 4 regions
[12:17]  Nebadon.Izumi its very slow going for me at this point
[12:17]  Nebadon.Izumi to get more bots in
[12:17]  Nebadon.Izumi i can get 1 every 3-4 minutes at best right now
[12:18]  Andrew Hellershanks: THat's not good
[12:18]  Nebadon.Izumi well its the radegast object scan
[12:18]  Nebadon.Izumi takes forever with this many avatars
[12:18]  Nebadon.Izumi so i could log them in fast
[12:18]  Nebadon.Izumi i just cant get them sat down fast
[12:18]  Nebadon.Izumi when you have 80 windows open
[12:18]  Nebadon.Izumi you have to go 1 at a time
[12:18]  Nebadon.Izumi lol
[12:19]  Nebadon.Izumi im using 3 computers today though
[12:19]  Nebadon.Izumi so it was a bit faster
[12:19]  Nebadon.Izumi unfortunatley mono barfed with the viewer
[12:19]  Mata Hari: then get all 400 to simultaneously stand up....wonder what the sim would think of the mass of physics calls
[12:19]  Billy.Bradshaw My viewer seems to be receiving more AvatarGroupReply, more than avatar adds.
[12:20]  Justin.Clark-Casey 192
[12:20]  Nebadon.Izumi nice
[12:20]  Kayaker Magic: Neb, there are 3 of your bots underwater, two of them off-sim.
[12:20]  Justin.Clark-Casey well, they're already standing even if you can't see that
[12:20]  Nebadon.Izumi wierd
[12:20]  Grimley.Graves graphics at minimum I get 5 FPS max
[12:20]  Justin.Clark-Casey I will sit them if they appear to be causing complete physics meltdown
[12:20]  Nebadon.Izumi the ones underwater
[12:20]  Nebadon.Izumi are probably sitting on a chair that someone else is already sitting on
[12:20]  Nebadon.Izumi just ignore them
[12:21]  Nebadon.Izumi i cant tell if someone is sitting on a chair already
[12:21]  Nebadon.Izumi without lots of extra work
[12:21]  Nebadon.Izumi i still managed to get 65 in today
[12:21]  Nebadon.Izumi so not too bad
[12:21]  Nebadon.Izumi im still going
[12:21]  Nebadon.Izumi doubt ill actually get to 80 though again this week
[12:22]  Andrew Hellershanks: I've turned graphics quality down as far as it will go and still the viewer is locked up.
[12:22]  Nebadon.Izumi it also seems now
[12:22]  Nebadon.Izumi every new one i bring iis a cloud
[12:22]  Justin.Clark-Casey 211
[12:22]  Nebadon.Izumi that is definitely not good
[12:22]  Nebadon.Izumi so we are already hitting failure point maybe
[12:23]  Justin.Clark-Casey how long have you been seeing that?
[12:23]  Nebadon.Izumi 1 avatar so far
[12:23]  Nebadon.Izumi right around 200
[12:23]  Serene.Jewell Yes, I see all new bots as clouds too. (am on Firestorm)
[12:23]  Justin.Clark-Casey I mean, did you see it last week?
[12:23]  Nebadon.Izumi yes
[12:23]  Justin.Clark-Casey did you see it week before that?
[12:23]  Nebadon.Izumi probably
[12:23]  Nebadon.Izumi hard to say I very much doubt its new
[12:23]  Justin.Clark-Casey well, this may ot may not be an effect of turning off avatar appearance resends
[12:24]  Cheopss Forlife7: no lost of attachements this time or am I dreaming?
[12:24]  Nebadon.Izumi ok
[12:24]  Andrew Hellershanks: 161 avatars across four regions
[12:24]  Nebadon.Izumi last week it definitely happened
[12:24]  Nebadon.Izumi it should be more than 161
[12:24]  Andrew Hellershanks: Keynote 1 only reports 2 avatars. Strange.
[12:24]  Latif Khalifa: 230
[12:24]  BlueWall.Slade any idea how much the bandwidth cost is on the client -side with so many avatars?
[12:24]  Grimley.Graves hey, this test made my skin rot away WTH?!?
[12:24]  Nebadon.Izumi for some reason KN1 is only showing 2 avatars on map
[12:24]  Kayaker Magic: lol
[12:24]  Andrew Hellershanks: Seems it doesn't properly report online avatars for the region I'm in
[12:24]  Nebadon.Izumi for me anyway
[12:25]  Andrew Hellershanks: Nebadon, same issue here
[12:25]  Nebadon.Izumi wierd
[12:25]  Nebadon.Izumi maybe because of my bot crash
[12:25]  Nebadon.Izumi i had 20 avatars here before everyone showed up
[12:25]  Latif Khalifa: it shows 67 "main agents" IN THE STATS FLOATER
[12:25]  Nebadon.Izumi and it crashed hard on my end
[12:25]  Latif Khalifa: oops caps
[12:25]  Billy.Bradshaw map says 64 here
[12:26]  Serene.Jewell I see hair and other attachments! But some avatars, like Avia Bonne, I only see her hat - no body.
[12:26]  Latif Khalifa: my radar show 249
[12:26]  Nebadon.Izumi ya last bot again is cloud
[12:26]  Andrew Hellershanks: nebadon, I TP'ed to 3 and the count of avatars here in 1 went to 1.
[12:26]  Justin.Clark-Casey 251
[12:26]  Nebadon.Izumi Tester6 and Teste7
[12:26]  Nebadon.Izumi in the last row of KN1
[12:26]  Nebadon.Izumi you can see they are stuck
[12:26]  Nebadon.Izumi let me try a rebake
[12:27]  Mata Hari: Serene: that might be LOD bug...try zooming out and in
[12:27]  Billy.Bradshaw 255
[12:27]  Nebadon.Izumi rebake not helping
[12:27]  Mata Hari: Avia was invisible to me until I crossed an LOD camera boundery
[12:27]  Serene.Jewell OK, yes, after zooming in and out I now see Avia.
[12:28]  Serene.Jewell And she is cute.
[12:28]  Mata Hari: that's a viewer side bug...happens quite a lot
[12:28]  Justin.Clark-Casey 266
[12:29]  Mata Hari: can also happen with other assets like animations....the viewer receives it but for some reason never starts to play it
[12:29]  Mata Hari: as soon as you relog the animation will work perfectly instantly
[12:29]  Billy.Bradshaw map shows 70, but only on one corner, color red. Not sure which one that is
[12:29]  Mata Hari: so it's sitting there in viewer cache
[12:29]  Nebadon.Izumi ok im not sure if i should keep loading in more or not at this point
[12:30]  Billy.Bradshaw 280
[12:30]  Nebadon.Izumi since they all come in as cllouds I am pretty sure things are already failing
[12:30]  Justin.Clark-Casey up to you
[12:30]  Grimley.Graves I have an idea, turn build on and we can all rez physical cubes in the air and let em go :D
[12:30]  Mata Hari: but I think if there's a little lag in fetching the data it somehow forgets to trigger showing it on the viewer once it finishes receiving the file
[12:30]  Justin.Clark-Casey I have capabilty to launch bots from your system now but I don't have the ability to abort all the ones coming from my lcoal
[12:30]  Justin.Clark-Casey so I'm just going to wait until this thing crashes or huge numbers fall off
[12:31]  Nebadon.Izumi ok
[12:32]  Billy.Bradshaw 300
[12:33]  Justin.Clark-Casey 309
[12:33]  Nebadon Izumi: I cant move
[12:33]  Justin.Clark-Casey anybody else have chatlag?
[12:33]  Nebadon Izumi: oh there it goes
[12:33]  Justin.Clark-Casey test
[12:35]  Nebadon.Izumi ok I am loosing bots now
[12:35]  Dahlia Trimble: did a bunch of bots just die/
[12:35]  7bruno.Spyker bonsoir a tous pas de crash encore?:)
[12:35]  Latif Khalifa: 1
[12:35]  Latif Khalifa: 2
[12:35]  Latif Khalifa: 3
[12:35]  Nebadon.Izumi just a couple though
[12:35]  Latif Khalifa: 4
[12:35]  Latif Khalifa: 5
[12:35]  Mata Hari: t+1m
[12:35]  acryline erin: Hello 7bruno !
[12:35]  Dahlia Trimble: 2
[12:35]  Dahlia Trimble: 19
[12:35]  Dahlia Trimble: 7
[12:35]  Kayaker Magic: my last chat never appeared
[12:35]  Dahlia Trimble: 3
[12:35]  Justin.Clark-Casey I lost a whole bunch
[12:36]  Nebadon.Izumi ya
[12:36]  Dahlia Trimble: 11.36
[12:36]  Mata Hari: at my "12:33"
[12:36]  Nebadon.Izumi it seems like 200 is the sore point
[12:36]  Nebadon.Izumi thats when i started noticing failures
[12:36]  Justin.Clark-Casey well, about 300
[12:36]  Latif Khalifa: it froze for a while... but now it seems back
[12:36]  Nebadon.Izumi ya
[12:36]  Nebadon.Izumi but things starting declining at 200
[12:36]  Nebadon.Izumi even though it wasnt total fail
[12:36]  Mata Hari: lost some chat but is back to being lag-free chat again now
[12:36]  Justin.Clark-Casey okay, I'm going to disconn everything from this one system and try spreading load between two
[12:36]  Dahlia Trimble: test message sent at 9:32 am PDT
[12:36]  Nebadon.Izumi ok cool Justin
[12:36]  Nebadon.Izumi i'll try to squeeze a few more in
[12:37]  Billy.Bradshaw 313
[12:38]  Grimley.Graves walking and testing... testing and walking
[12:38]  Nebadon.Izumi lol
[12:38]  Billy.Bradshaw clouds look bad on a photo shoot
[12:38]  Nebadon.Izumi heh ya
[12:38]  Nebadon.Izumi some of justins bots will inevitably be clouds
[12:39]  Nebadon.Izumi however none of mine should be, but all the latest ones are
[12:39]  Cuteulala Artis gave you WallPaper Bloom1024.
[12:39]  Grimley.Graves I cant cross the region boundry
[12:40]  Dahlia Trimble: you can cross but it takes a trick
[12:40]  Justin.Clark-Casey pls don't cross the boundary
[12:40]  Dahlia Trimble: lol
[12:41]  Latif Khalifa: are these 4 regions running on 4 physical machines?
[12:41]  Grimley.Graves ok, can't anyway lol
[12:41]  Justin.Clark-Casey hmm, sim thnks 216 avs, which is surprising
[12:41]  Nebadon Izumi: no Larif
[12:41]  Nebadon Izumi: its a 24 core machine
[12:41]  Nebadon Izumi: with 64gb ram
[12:41]  Latif Khalifa: so you could use 1 var
[12:41]  Nebadon Izumi: but each region is its own instance
[12:41]  Billy.Bradshaw I do see an awful lot of AvatarGroupReply
[12:41]  Nebadon Izumi: maybe, maybe not
[12:41]  Nebadon Izumi: i really doubt you can get 400 avatars in 1 region
[12:42]  Nebadon Izumi: its more than just connections
[12:42]  Nebadon Izumi: its physics
[12:42]  Nebadon Izumi: and how much memory 1 process can access
[12:42]  Dahlia Trimble: would be interesting to try 1 large region
[12:42]  Nebadon Izumi: ya we will for sure
[12:42]  Billy.Bradshaw Don't seem to be able to manage 400 on 4 at present
[12:42]  Nebadon Izumi: we have aregion setup for testing
[12:42]  Billy.Bradshaw though with 346 and counting
[12:42]  Nebadon Izumi: but I suspect it wont peform better
[12:43]  Nebadon Izumi: some things may
[12:43]  Nebadon Izumi: physics would not
[12:43]  Nebadon Izumi: also memory could be problematic
[12:43]  Nebadon Izumi: even in 64 bit
[12:43]  Nebadon Izumi: opensim pukes around 6gb being used
[12:43]  Billy.Bradshaw 6gb, I have not seen that
[12:44]  Justin.Clark-Casey crashed
[12:44]  Billy.Bradshaw 401
[12:44]  Nebadon Izumi: I have
[12:44]  Billy.Bradshaw yeah
[12:44]  Serene.Jewell My Firestorm people counter shows 402
[12:44]  Nebadon Izumi: Wright Plaza has used that much before
[12:44]  Dahlia Trimble: 275 for me
[12:44]  Latif Khalifa: 404 here
[12:44]  Billy.Bradshaw 403 at 12:44
[12:45]  Mata Hari: 403 is what mine says
[12:45]  Billy.Bradshaw +me
[12:45]  Cheopss Forlife7: wooo!
[12:45]  Justin.Clark-Casey ok, sims think we are at 405
[12:45]  Latif Khalifa: yes 403 + yourself = 404
[12:45]  Justin.Clark-Casey yes, spreading this bot load between two systems does work a bit better
[12:45]  Billy.Bradshaw lol Latif
[12:45]  Justin.Clark-Casey phew
[12:45]  Dahlia Trimble: my radar says 399
[12:45]  Cheopss Forlife7: thar double more than last year?
[12:45]  Cheopss Forlife7: that's*
[12:45]  Nebadon Izumi: something like that
[12:46]  Justin.Clark-Casey so, now to see if it's reasonably stable
[12:46]  Justin.Clark-Casey i.e. don't suddenly get massive bot falloff
[12:46]  Nebadon Izumi: its a bit laggy for me
[12:46]  Latif Khalifa: what's the load like on the server
[12:46]  Justin.Clark-Casey I',m now on kn4, couldn't get to kn1 because full
[12:46]  Latif Khalifa: chat seems fluid
[12:46]  Latif Khalifa: 1
[12:46]  Latif Khalifa: 2
[12:46]  Latif Khalifa: 3
[12:46]  Latif Khalifa: 45
[12:46]  Latif Khalifa: 5
[12:46]  Mata Hari: very little movement lag for me
[12:46]  Justin.Clark-Casey lag wil prob be worse on kn1
[12:46]  Billy.Bradshaw a tad lag llol
[12:47]  Justin.Clark-Casey serer is about0.49 load av
[12:47]  Justin.Clark-Casey which is ok
[12:47]  Latif Khalifa: that's pretty amazing Justin
[12:47]  Justin.Clark-Casey yes... a little too much perhaps
[12:48]  Serene.Jewell At the end of the test do we all get up and dance to properly crash the servers?
[12:48]  Justin.Clark-Casey still at 402
[12:48]  Billy.Bradshaw lost 2 according to my counter
[12:48]  Justin.Clark-Casey heh, yes
[12:48]  Justin.Clark-Casey just want to hold this fora bit longer
[12:48]  Justin.Clark-Casey yeah, there is some falloff
[12:48]  Billy.Bradshaw makes tea
[12:49]  Justin.Clark-Casey there's a long way to go, but sustaining 400, ieven if unrealistic bots is a start
[12:49]  Billy.Bradshaw anyone want a cup
[12:49]  Nebadon.Izumi
[12:49]  Mata Hari: yes please...with milk if you have some
[12:49]  Nebadon.Izumi there you go lkalif
[12:49]  Serene.Jewell Do you have any biscuits, darling?
[12:49]  Billy.Bradshaw Coming right up Mata
[12:50]  Dahlia Trimble: cookies? o.0
[12:50]  Billy.Bradshaw dang, forgot the biscuits
[12:50]  Kayaker Magic: Now you've done it, I'm going AFK to make lunch
[12:50]  Justin.Clark-Casey still at 402
[12:50]  Justin.Clark-Casey but as you can see from viewer stats, not all child agents are properly logged in
[12:50]  Serene.Jewell At what number do you declare an official miracle?
[12:51]  Billy.Bradshaw 402 at 12:51
[12:51]  Nebadon.Izumi well 400 is our goal
[12:51]  Latif Khalifa: 400
[12:51]  Nebadon.Izumi right now things start getting troublesome around 200
[12:51]  Justin.Clark-Casey even in very low load conditions, for some reason pCampbot often doesn't connect children in time for some reason, whic needs investigation
[12:51]  Nebadon.Izumi atleast with appearance
[12:51]  Nebadon.Izumi things get very cloudy around 200
[12:51]  Justin.Clark-Casey I could try turning on appearance resending ;)
[12:51]  Mata Hari: ouch
[12:51]  Justin.Clark-Casey yeah, there are myriad issues
[12:52]  Nebadon.Izumi it must be something with the amount of appearance updates being sent
[12:52]  Mata Hari: set the timer for that appearance resend to something really short, too, like 10 seconds
[12:52]  Nebadon.Izumi at some point it becomes to much
[12:52]  Justin.Clark-Casey yeah, bots are maxing out your machine too neb
[12:52]  Justin.Clark-Casey sitting at 90% on 4 cores
[12:52]  Nebadon.Izumi perhaps takes longer to complete than the cycle of appearance updates
[12:52]  Justin.Clark-Casey but bots aren't falling off, which is something
[12:52]  Nebadon.Izumi so you never get a full update
[12:53]  Nebadon.Izumi just guessing
[12:53]  Justin.Clark-Casey neb: these are bots in kn1?
[12:53]  Nebadon.Izumi someone mentioned something with appearance updates
[12:53]  Nebadon.Izumi starting to fail around 100
[12:53]  Nebadon.Izumi who was that?
[12:53]  Billy.Bradshaw I cannot see why AvatarAppearance packet sizes vary, especially with the b0b0b0 bots
[12:53]  Nebadon.Izumi yes they are
[12:53]  Nebadon.Izumi the 3 or 4 stuck as clouds are all KN1
[12:53]  Nebadon.Izumi and logged in very late to the game
[12:53]  Nebadon.Izumi somewhere around the 200-300 avatar mark
[12:54]  Nebadon.Izumi i stoipped logging them after that
[12:54]  Billy.Bradshaw ok, now i see avatarappearance
[12:54]  Justin.Clark-Casey neb: any diff? I just manually triggered some resends
[12:54]  Nebadon.Izumi let me look
[12:54]  Nebadon.Izumi no
[12:55]  Justin.Clark-Casey which bot in particular?
[12:55]  Nebadon.Izumi looks like one of them got logged out
[12:55]  Nebadon.Izumi its Nebadon Tester6, and Tester8
[12:55]  Nebadon.Izumi last row of KN1
[12:55]  Nebadon.Izumi Tester7 got logged out
[12:55]  Justin.Clark-Casey I can'#t lpook atm since if I turn it crashes my viewer
[12:55]  Justin.Clark-Casey give me a name
[12:55]  Nebadon.Izumi hes coing back in
[12:56]  Nebadon.Izumi Nebadon Tester6
[12:56]  Justin.Clark-Casey yeah, I'm way overlaoded with tjhis bot stuff
[12:56]  Justin.Clark-Casey load average of 42
[12:56]  Nebadon.Izumi Nebadon Tester8
[12:56]  Nebadon.Izumi seems to have froze again when i logged him in
[12:56]  Justin.Clark-Casey kn1 is having issues
[12:56]  Nebadon.Izumi oh dang viewer crashed hard on reconnect
[12:56]  Nebadon.Izumi indeed
[12:57]  Nebadon.Izumi i reconnected a bot and it froze the region
[12:57]  Latif Khalifa: 1
[12:57]  Nebadon.Izumi possible others were logging in at same time
[12:57]  Nebadon.Izumi multiple logins are nasty
[12:57]  Justin.Clark-Casey yeah, 7 has no appaerance uploaded
[12:57]  Justin.Clark-Casey it could be that the messages are simply getting lost, even with retries
[12:57]  Justin.Clark-Casey I think there is an issue with overloading the capabiltiy system
[12:58]  Justin.Clark-Casey 401 still
[12:58]  Justin.Clark-Casey I'm going to risk moving around, and iof I crash I will comein with my linux system viewer
[12:59]  Nebadon.Izumi well thats good
[12:59]  Nebadon.Izumi its a bit laggy, were certainly close to keeping the simulator up without too much disruption
[12:59]  Mata Hari: I now can't move at all
[13:00]  Mata Hari: and probably crashed since chat isn't working either
[13:00]  Nebadon.Izumi heh huge delay there
[13:00]  Nebadon.Izumi hmm major lag nothing i am saying is coming through
[13:00]  Mata Hari: oh.///mpw suddenly working again
[13:00]  Mata Hari: and now not
[13:00]  Nebadon.Izumi ya I think Justin crashed
[13:00]  Nebadon.Izumi good test though
[13:00]  Nebadon.Izumi no doubt got a lot of good data to mull over again this week
[13:01]  Nebadon.Izumi as bad as it was today, i definitely saw some improvement
[13:01]  Nebadon.Izumi bot crashes were not nearly as bad this week
[13:01]  Mata Hari: and now movement is smooth again
[13:01]  Nebadon.Izumi last week i lost 50% of my bots by this time
[13:01]  Nebadon.Izumi this week i lost maybe 5
[13:01]  Nebadon.Izumi and most reconnected no problem
[13:01]  Nebadon.Izumi lots of lag though
[13:01]  Mata Hari: seems to come and go
[13:02]  Nebadon.Izumi ya typical
[13:02]  Nebadon.Izumi probably GC
[13:02]  Cheopss Forlife7: the best is if you don't lose your ideas Neb
[13:02]  Mata Hari: at the moment it's silky smooth
[13:02]  Mata Hari: but for a moment there it was lagging by close to a minute
[13:02]  Nebadon.Izumi ya
[13:03]  Justin.Clark-Casey ok,. I forgot my linyux box is also running all these bots....
[13:03]  Nebadon.Izumi welcome back
[13:03]  Nebadon.Izumi i was saying justin even though this week was far from perfect
[13:04]  Billy.Bradshaw Maybe you need to distribute these bots sources
[13:04]  Nebadon.Izumi it feels improved to me
[13:04]  Justin.Clark-Casey well, one of my bot programs finally crashed
[13:04]  Nebadon.Izumi i didnt loose nearly as many bots
[13:04]  Justin.Clark-Casey there's no change server side
[13:04]  Grimley.Graves gotta run, storm coming in
[13:04]  Nebadon.Izumi and the ones that did drop off got back on fairly quickly
[13:04]  Nebadon.Izumi lots of lag though
[13:04]  Grimley.Graves good luck to you all
[13:04]  Nebadon.Izumi see you Grimley
[13:04]  Justin.Clark-Casey a lot of issues to look at but I'm quite pleased was able to sustain 400 (mainly bots) for a period today
[13:04]  Nebadon.Izumi say hi to the wizard of oz for us
[13:04]  Nebadon.Izumi lol
[13:04]  Cheopss Forlife7: bye Grim
[13:05]  Justin.Clark-Casey now I can make them all start jumping around
[13:05]  Justin.Clark-Casey and the sims should crash :)
[13:05]  Justin.Clark-Casey or maybe not
[13:05]  Nebadon.Izumi heh give it a shot
[13:05]  Justin.Clark-Casey I think I'll start all the kn1 going
[13:05]  Mata Hari puts on crash helmet
[13:05]  Justin.Clark-Casey lol
[13:05]  Serene.Jewell So this is the dancing part?
[13:05]  Justin.Clark-Casey I've just realized that connected bots don't take on new behaviours...
[13:06]  Justin.Clark-Casey damn
[13:06]  Nebadon.Izumi lol
[13:06]  Nebadon.Izumi CPU is 1000%
[13:06]  Nebadon.Izumi oh there its 350% now
[13:06]  Nebadon.Izumi 270%
[13:06]  Nebadon.Izumi did you stop them?
[13:06]  Justin.Clark-Casey oh, maybe they are sort of moving
[13:06]  Justin.Clark-Casey wierd
[13:06]  Nebadon.Izumi cpu hit 1300% at one point
[13:06]  Billy.Bradshaw I'm seeing 364 at present
[13:06]  Mata Hari: they all did move
[13:07]  Mata Hari: or in KN1 they seemed to
[13:07]  Nebadon.Izumi its odd CPU went that high
[13:07]  Justin.Clark-Casey they don't seem to be moving out of place though
[13:07]  Nebadon.Izumi technically Bullet should only be able to use 1 core
[13:07]  Nebadon.Izumi but seperated from the main thread
[13:07]  Nebadon.Izumi its single thread
[13:07]  Nebadon.Izumi strange it jumped so high
[13:07]  Nebadon.Izumi mybe it was UDP updates
[13:07]  Nebadon.Izumi for movement
[13:07]  Nebadon.Izumi ?
[13:07]  Justin.Clark-Casey alright, back in kn1
[13:08]  Justin.Clark-Casey ah man, they'vemanaged to mainly jump out
[13:08]  Justin.Clark-Casey this is only 50 though
[13:09]  Justin.Clark-Casey yeah, sim is not liking this
[13:09]  Nebadon.Izumi cpu is way high again
[13:10]  Nebadon.Izumi heh ya
[13:10]  Justin.Clark-Casey I'm now attempting to stop them jumping again....
[13:10]  Nebadon.Izumi k
[13:10]  Justin.Clark-Casey which has revealed a slight problem with that command in that it's doing it very very slowly
[13:10]  Nebadon.Izumi I think there are just too many updates occuring
[13:10]  Nebadon.Izumi we dont see it actually happening
[13:10]  Justin.Clark-Casey yeah, could be
[13:10]  Nebadon.Izumi more than the viewer can process probably
[13:10]  Justin.Clark-Casey it would be a massive peak of updates
[13:11]  Nebadon.Izumi which probably leads to more issues
[13:11]  Nebadon.Izumi resending of packets
[13:11]  Justin.Clark-Casey I think this is going to have to be another sit-down conf
[13:11]  Justin.Clark-Casey if even 50 bots can do that
[13:11]  Nebadon.Izumi oh absolutely
[13:11]  Nebadon.Izumi i never doubted that
[13:11]  Nebadon.Izumi im not even sure we intended otherwise
[13:11]  Nebadon.Izumi i sure didnt
[13:12]  Justin.Clark-Casey it would be nice to have complete freedom... but not gonna happen
[13:12]  Nebadon.Izumi lets get as close as we can though
[13:12]  Nebadon.Izumi ya next year :)
[13:12]  Nebadon.Izumi its good to have a reason to improve things
[13:12]  Nebadon.Izumi honestly I feel like this week was a big improvement over last week
[13:13]  Mata Hari: I wonder what the outbound traffic has to be like to send all those updates to every single viewer
[13:13]  Justin.Clark-Casey there are physical limits....
[13:13]  Cheopss Forlife7: oh we can give you reasons every day if you need!
[13:13]  Nebadon.Izumi even though its not glaringly obvious improvement
[13:13]  Justin.Clark-Casey funiily enough there have been no affecting server side changes
[13:13]  Nebadon.Izumi i see improvement
[13:13]  Justin.Clark-Casey but as pCampbot is crashing much less frequently that might be the factor
[13:13]  Billy.Bradshaw I can bring a bot army in next time, maybe enable justin to have more control of his desktops?
[13:13]  Nebadon.Izumi last week by now i lost over 50% of my bots
[13:13]  Justin.Clark-Casey the thing with my bot army is that I am in control of them, which helps
[13:13]  Mata Hari: would be 400^2 updates that have to get sent I suppose
[13:13]  Justin.Clark-Casey yeah, mass simultaenous client logout may be causing issues
[13:14]  Justin.Clark-Casey mata: yeah, it doesn't scale linearly, which is a big problem
[13:14]  Nebadon.Izumi i maybe lost 5% this week
[13:14]  Nebadon.Izumi no doubt me spreading bot load over 3 computers helped
[13:14]  Justin.Clark-Casey getting from 300 to 400 is vastly harder than 100 to 200
[13:14]  Nebadon.Izumi instead of all 80 on 1 box
[13:14]  Nebadon.Izumi i suspect you will see same improvement with pcampbot
[13:14]  Justin.Clark-Casey yeah, my lcao machine is still way overloaded. Your's is loaded but only a little over load factor
[13:15]  Nebadon.Izumi great
[13:15]  Mata Hari: maybe some sort of capping mechanism by proximity like the viewer uses for avatar imposters
[13:15]  Justin.Clark-Casey well, having said that, 5 min average is 11
[13:15]  Justin.Clark-Casey er, 15 min
[13:15]  Nebadon.Izumi even though technically your CPU is faster than my server is
[13:15]  Nebadon.Izumi and its a true Quad core
[13:15]  Justin.Clark-Casey my system is running a desktop and other crap as well
[13:15]  Mata Hari: so only send updates "frequently" for something nearby
[13:15]  Justin.Clark-Casey bah :)
[13:15]  Nebadon.Izumi my server doesnt have the GUI overhead
[13:15]  Nebadon.Izumi probably faster bus too on mine
[13:15]  Mata Hari: and far less frequently for distant
[13:15]  Justin.Clark-Casey mata: fancy submitting any patches?
[13:15]  Nebadon.Izumi its 1333 mhz bus
[13:15]  Mata Hari: though then you have to make those calculations
[13:16]  Mata Hari: I don't know C#, sorry
[13:16]  Nebadon.Izumi while yours has a better capacity for any single process
[13:16]  Nebadon.Izumi my cpu has much better throughput
[13:16]  Justin.Clark-Casey mata: it's not hard t olearn
[13:16]  Nebadon.Izumi anyway if it helps i can give you control of my core2quad machine as well
[13:17]  Nebadon.Izumi boot into opensuse 13.1 console only no gui
[13:17]  Justin.Clark-Casey we'll see, I have turned down some bot activities which makes them less realistic
[13:17]  Nebadon.Izumi i had to boot it into windows 7 today for radegast
[13:17]  Nebadon.Izumi ok
[13:17]  Mata Hari: if I didn't have other stuff on my plate I probably would
[13:17]  Justin.Clark-Casey so yeah, might be necessary to spread even further. I guess it's not so surprising - after all it's hard enough to run 1 full client on a system
[13:17]  Mata Hari: but there are only so many hours in the day
[13:17]  Justin.Clark-Casey mata: the same is true for me
[13:17]  Nebadon.Izumi yea indeed Justin
[13:18]  Nebadon.Izumi i see similar issues with Radegast too
[13:18]  Dahlia Trimble: bye all :)
[13:18]  Nebadon.Izumi at some point your run out of IO
[13:18]  Kayaker Magic: Bye
[13:18]  Justin.Clark-Casey bye dahlia
[13:18]  Nebadon.Izumi no matter how fast your cpu is and how much memory you have
[13:18]  Justin.Clark-Casey I have already put in some experimental stuff to turn down agent udpates
[13:18]  Justin.Clark-Casey particularly from child regions, which is currently active, and one reason why cpu loads are lower now
[13:19]  Nebadon.Izumi we could get very experimental here
[13:19]  Justin.Clark-Casey as well as agent update sending loads
[13:19]  Mata Hari: yes, if child updates were way lower that would help a ton
[13:19]  Nebadon.Izumi since no is really moving around much
[13:19]  Nebadon.Izumi make it something tuneable that only this region would run
[13:19]  Nebadon.Izumi hwile the other regions have a bit more available
[13:19]  Billy.Bradshaw 397
[13:19]  Justin.Clark-Casey there is a siwtch to also start cutting out root agent updates
[13:20]  Justin.Clark-Casey that I haven't turned on, since it's significantly more noticeable even if the ones going to your own avatar are unfiltered
[13:20]  Justin.Clark-Casey ok, bot physics has been off for a while
[13:20]  Nebadon.Izumi it also seems like when logins/logouts slow down or stop
[13:20]  Nebadon.Izumi things become relatively stable
[13:21]  Nebadon.Izumi no lag spikes or long pauses
[13:21]  Justin.Clark-Casey logins are high stree periods because they trigger a very large amount of packet send to the client
[13:21]  Mata Hari: I was thinking about the inventory fetch issue and the utter lack of any messages in log or console about it possible they're timing out and not getting reported?
[13:21]  Nebadon.Izumi yea
[13:21]  Justin.Clark-Casey amongst other things
[13:21]  Justin.Clark-Casey logouts are more of a mystery
[13:21]  Nebadon.Izumi progressively gets worse as more avatars are present and receiving updates
[13:21]  Justin.Clark-Casey yes, new avatar ends up triggering data send to all existing avatars
[13:21]  Justin.Clark-Casey attachments will play a role in that
[13:22]  Nebadon.Izumi yes I have always noticed that logouts are just as expensive as logins
[13:22]  Nebadon.Izumi ive noticed that for years at WP meetings
[13:22]  Justin.Clark-Casey it is odd, and one area I hope to look at in detail
[13:22]  Nebadon.Izumi when everyone leaves at the end
[13:22]  Justin.Clark-Casey it shouldn't be so expensive
[13:22]  Nebadon.Izumi the region freezes almost same as during logins
[13:22]  Justin.Clark-Casey I presume avatars having attachments doesn't make a difference?
[13:22]  Serene.Jewell you could schedule us human subjects to log in at different times to test the login thing.
[13:22]  Nebadon.Izumi i cant say
[13:23]  Nebadon.Izumi i bet it does make a big difference
[13:23]  Justin.Clark-Casey co-ordinating humans is horribly messy :()
[13:23]  Serene.Jewell lol
[13:23]  Nebadon.Izumi reality is chaos
[13:23]  Mata Hari: easier with cattle :p
[13:23]  Nebadon.Izumi not order
[13:23]  Nebadon.Izumi these tests need to remain chaotic
[13:23]  Justin.Clark-Casey whichis the big problem, since conf has to be for reals
[13:23]  Nebadon.Izumi othewise its unreastic
[13:24]  Justin.Clark-Casey but the more chaotic you get the harder it is to work out what's happening or control the inputs
[13:24]  Nebadon.Izumi yes
[13:24]  Nebadon.Izumi it is
[13:24]  Nebadon.Izumi you and I should have some bot only tests
[13:24]  Nebadon.Izumi just you and I
[13:24]  Justin.Clark-Casey yeah, I'm going to do some myself, hope to get a bit more time now
[13:24]  Nebadon.Izumi k cool
[13:24]  Billy.Bradshaw I will prepare an army should it be required, say 50... bye for now.. watch me logout.. well tp
[13:24]  Nebadon.Izumi let me know if I can help
[13:24]  Nebadon.Izumi see you Billy
[13:24]  Justin.Clark-Casey but still need to work out why this pCampbot is crashing. All I've done here is reduce the frequency, though that might be a clue
[13:24]  Justin.Clark-Casey bye billy
[13:24]  Nebadon.Izumi yea
[13:25]  Nebadon.Izumi as frustrating as this is, i really do feel like there has been imrprovement
[13:25]  Nebadon.Izumi the numbers will be the real truth though
[13:25]  Justin.Clark-Casey it seems reducing the number of concurrent connections allowed to an endpoint reduces crashing
[13:25]  Nebadon.Izumi be interesting to see the trend this week
[13:25]  Nebadon.Izumi if the numbers reflect what I feel
[13:25]  Justin.Clark-Casey so it's either something to do with number of concurrenct connections or setting the paramater. Much more likely the former
[13:26]  Nebadon.Izumi it wont be drastic either way
[13:26]  Justin.Clark-Casey yeah, I kleep running out of time to write more reports. Hope to do one for this week (!)
[13:26]  Justin.Clark-Casey graph production is relatively easy since I have most of it scripted now
[13:27]  Justin.Clark-Casey ok, count is still 397 which is pretty good
[13:27]  Justin.Clark-Casey unless there's anything else I'm going to take down the bots and stop stats recording
[13:27]  Justin.Clark-Casey neb: ?
[13:28]  Justin.Clark-Casey yeah, myu singu isn't crashing on linux. Must be drivers or similar on my crappy old win xp laptop :)
[13:29]  Nebadon.Izumi sure
[13:29]  Nebadon.Izumi thats fine by me
[13:29]  Justin.Clark-Casey ok, thanks a lot for coming folks, it's really helpful
[13:29]  Mata Hari: okies, I"ll be off then too
[13:29]  Nebadon.Izumi ya thanks everyone
[13:29]  Nebadon.Izumi same time next week
[13:29]  Nebadon.Izumi tell more friends
[13:29]  Nebadon.Izumi lol
[13:29]  Mata Hari avaes
[13:30]  Justin.Clark-Casey I will probably have to fether bot numbers a bit. When at 400 nobody can get it
[13:30]  Nebadon.Izumi ya spreading that load might help too
[13:30]  Nebadon.Izumi even at 350+
[13:30]  Nebadon.Izumi i could still get my avatars in
[13:30]  TBG Renfold: np, night al. outta here
[13:30]  Nebadon.Izumi its just with Radegast the object scan is horribly slow
[13:30]  Nebadon.Izumi so i cant sit down fast
[13:30]  Nebadon.Izumi takes forever
[13:31]  Justin.Clark-Casey I'm not sure sittinf down is too crucial as long as you aren't moving
[13:31]  Nebadon.Izumi ya
[13:31]  Justin.Clark-Casey the bots you can see are standing. It's only when they start jumping around that things really go haywire
[13:31]  Nebadon.Izumi i suspect though big avatar piles at landing zone
[13:31]  Nebadon.Izumi probably does a number on physics
[13:31]  Nebadon.Izumi increasing cpu load
[13:31]  Justin.Clark-Casey a lod of my bots overflowed to landing one 1
[13:31]  Justin.Clark-Casey lz1
[13:31]  Nebadon.Izumi interesting
[13:31]  Nebadon.Izumi thats good
[13:31]  Nebadon.Izumi thats what should happen
[13:32]  Justin.Clark-Casey I think it's only happening on direct login
[13:32]  Justin.Clark-Casey when I tried to HG to kn1 earlier I just got the "region is full" msg
[13:32]  Nebadon.Izumi id be more upset if people couldnt log in at all
[13:32]  Nebadon.Izumi that is strange
[13:32]  Justin.Clark-Casey ok, disconn
[13:32]  Nebadon.Izumi that is says region is full
[13:32]  Nebadon.Izumi kk
[13:32]  Nebadon.Izumi ya i need to kill mine
[13:33]  Justin.Clark-Casey nebadon: why strange? Because I'm admin? ( I think)
[13:33]  Nebadon.Izumi was the region actually full?
[13:33]  Nebadon.Izumi guess its not strange if region was really full
[13:33]  Nebadon.Izumi i was assuming it wasnt
[13:33]  Justin.Clark-Casey yeah, it was full
[13:34]  Nebadon.Izumi ah ok
[13:34]  Nebadon.Izumi then nevermind
[13:34]  Nebadon.Izumi hehe
[13:34]  Justin.Clark-Casey ok, didn't seem too bad just now on mass bot logout
[13:34]  Nebadon.Izumi ya
[13:35]  Nebadon.Izumi the good news is
[13:35]  Nebadon.Izumi we havent had a single hard region crash this year
[13:35]  Nebadon.Izumi not a one
[13:35]  Nebadon.Izumi thats quite impressive
[13:35]  Justin.Clark-Casey yeah, even throughout these load tests it didn' happen
[13:35]  Nebadon.Izumi ya
[13:35]  Justin.Clark-Casey and regions seem to return to normal once bot load was removed
[13:35]  Justin.Clark-Casey which are very good things
[13:35]  Nebadon.Izumi yep
[13:35]  Joyce.Bettencourt great
[13:36]  Nebadon.Izumi ok well if I am not needed I am going to hop out
[13:36]  Nebadon.Izumi and find some food to chew on
[13:36]  Justin.Clark-Casey yep, thanks neb
[13:36]  Justin.Clark-Casey me too
[13:36]  Joyce.Bettencourt I was nibbling during
[13:36]  Nebadon.Izumi ya thank you
[13:36]  Justin.Clark-Casey do you have chatlog from prvious meeting btw
[13:36]  Joyce.Bettencourt though didnt have to run bots like you all did
[13:36]  Nebadon.Izumi and thanks everyone who came
[13:36]  Nebadon.Izumi good stuff
[13:36]  Justin.Clark-Casey I lost the easy to access one in my crash
[13:36]  Nebadon.Izumi i do
[13:36]  Nebadon.Izumi i can handle it
Personal tools
About This Wiki