Chat log from the meeting on 2016-06-14

[11:11] Andrew.Hellershanks @hg.osgrid.org: Probably a good time for the first announcement. [11:12] Billy.Bradshaw @hg.viewtwo.net:8600: A little bird has told me that OSG upgrade to 0.9 has been fun and games, thought they already had, so a double surprise [11:12] Andrew.Hellershanks @hg.osgrid.org: The code running osgrid has been updated. It is now running the version of code generally referred to as 0.9 [11:13] Andrew.Hellershanks @hg.osgrid.org: If you run across any important issues or problems in osgrid please report them. [11:15] Billy.Bradshaw @hg.viewtwo.net:8600: Oh the keep alive thing, well spotted Robert [11:15] Andrew.Hellershanks @hg.osgrid.org: Billy, there was one bit of code that had to be reverted which fixed an issue with logins to this grid. [11:15] Andrew.Hellershanks @hg.osgrid.org: hm... I wonder If Kayaker or Gavin will be here today. I had some comments for them. [11:16] Billy.Bradshaw @hg.viewtwo.net:8600: Just looking at the last commit comments [11:16] Andrew.Hellershanks @hg.osgrid.org: Robert, I have one item for you. If Gavin shows up you can get more information from him. [11:17] Robert Adams: ok [11:17] Andrew.Hellershanks @hg.osgrid.org: Robert, he has been reported seeing 10% to 15% higher CPU load with BulletSim in 0.9 vs. 0.8.2 so you might want to add that to your ToDo list. [11:17] Andrew.Hellershanks @hg.osgrid.org: I haven't confirmed that issue so that would be the first step. [11:18] Robert Adams: k... I'veheard a few reports of that... haven't tried to reproduce it yet [11:18] Andrew.Hellershanks @hg.osgrid.org: ok [11:19] Andrew.Hellershanks @hg.osgrid.org: I don't know where it will land on your ToDo list but wanted to make sure you had heard about it. Next time I see Gavin I'll ask him to create a mantis about the issue. [11:19] Andrew.Hellershanks @hg.osgrid.org: Ah, Speaking of Gavin. He must have heard his name being mentioned. :) [11:20] Gavin.Hird @grid.xmir.org:8002: There are some odd issues teleporting here [11:20] Gavin.Hird @grid.xmir.org:8002: everyone showas as Unknown user and [11:22] Billy.Bradshaw @hg.viewtwo.net:8600: (from Billy) @Gavin .. Andrew opened with your comments relating to cpu load with the current 0.9 and bulletsim [11:22] Andrew.Hellershanks @hg.osgrid.org: I'll go ahead with the couple of other announcements. [11:22] George Equus: All look normal to me, only thing is new arrivals take some time to rez properly, Gavin, you were all grey for a couple of minutes, and so was Andrew when he arrived. [11:22] Andrew.Hellershanks @hg.osgrid.org: Gavin, I told Robert about the CPU load issue you have seen with BulletSim. [11:23] Gavin.Hird @grid.xmir.org:8002: right [11:23] Gavin.Hird @grid.xmir.org:8002: I also filed a Mantis the other day with even more CPU use [11:23] Gavin.Hird @grid.xmir.org:8002: when running on mono 4.4 [11:23] Andrew.Hellershanks @hg.osgrid.org: Gavin, I also mentioned that osgrid has been updated to 0.9 so if you run across important issues or problems then please report them. [11:23] Robert Adams: Gavin: is there a Mantis for the discussion? [11:23] Gavin.Hird @grid.xmir.org:8002: the CPU use on mono 4.4 is 3X the use on 4.2.3.4 [11:24] Andrew.Hellershanks @hg.osgrid.org: Ah, the ongoing issues with mono versions. [11:24] Robert Adams: it's somehow related to Mono and not 8.8.2 to 0.9.0? [11:24] Gavin.Hird @grid.xmir.org:8002: reported in 7975 [11:25] Robert Adams: k... I'll look for that [11:25] Gavin.Hird @grid.xmir.org:8002: there is not mantis for the initial observation [11:25] Gavin.Hird @grid.xmir.org:8002: simply because I have not had sufficent time to restore the grid to 0.8.2.1 for proper comparison [11:26] Andrew.Hellershanks @hg.osgrid.org: Good time for my next announcement. [11:26] Robert Adams: k... not sure how Mono version would effect physics CPU use... but stranger things have happened [11:26] Gavin.Hird @grid.xmir.org:8002: I don't think it has to do with physics at all [11:27] Gavin.Hird @grid.xmir.org:8002: as I said before more likely region to region chatter [11:27] Robert Adams: that gives a place to look... will look [11:28] Billy.Bradshaw @hg.viewtwo.net:8600: I have an 0.9 (dev/master) micro grid, I have not seen any appreciable increase in resources.. but I do run on a super slim CoreOS operating system and mono 4.5.2.. maybe that is a difference [11:28] Gavin.Hird @grid.xmir.org:8002: .... when the regions are in different simulators [11:28] Andrew.Hellershanks @hg.osgrid.org: There are plans being made to release the 0.9 version of Open Simulator. A list is being made of any major issues that currently exist in the git master (aka. 0.9) version so they can be worked on before the release. If you know of any important issues that haven't already been filed in mantis, please do so. [11:28] Gavin.Hird @grid.xmir.org:8002: with region in the same simulator there is no additional CPU use to speak of [11:28] Billy.Bradshaw @hg.viewtwo.net:8600: noted Gavin [11:29] Andrew.Hellershanks @hg.osgrid.org: The mantis reports will be looked at and may be added to the list of things to fix for a 0.9 release. [11:30] Andrew.Hellershanks @hg.osgrid.org: The last item I had was for Kayaker but he isn't here. I'll pass along the information in case others have had the problem. [11:31] Robert Adams: we're keeping the list of release 'show stoppers' at http://opensimulator.org/wiki/0.9_Bug_List [11:32] Andrew.Hellershanks @hg.osgrid.org: Kayaker had reported problems with being able to cross region boundaries multiple times (mantis #7862). There was a change made to OpenSim in the last 24 hours that appears to have fixed that issue. If any others have seen issues crossing back and forth between regions please update your copy of OS and retest. [11:33] Andrew.Hellershanks @hg.osgrid.org: That's the last of the informatin I had written down to pass along for today. [11:33] Billy.Bradshaw @hg.viewtwo.net:8600: I saw that, problem with 0.9 regions connected to 8.2 robust? [11:33] Andrew.Hellershanks @hg.osgrid.org: Billy, yes [11:34] vegaslon plutonian: is anyone running multiple instances of OpenSim 0.9 from the same folder? [11:34] Gavin.Hird @grid.xmir.org:8002: no [11:34] Billy.Bradshaw @hg.viewtwo.net:8600: not me [11:34] Gavin.Hird @grid.xmir.org:8002: different folders [11:37] Robert Adams: go ahead, Andrew... sorry to jump ahead [11:38] Andrew.Hellershanks @hg.osgrid.org: np. I just wanted to make those announcements. [11:38] Gavin.Hird @grid.xmir.org:8002: The parcel crossing issue that was mentioned last meeting is now fixed [11:38] Andrew.Hellershanks @hg.osgrid.org: Anyone else have something for todays meeting? [11:38] Andrew.Hellershanks @hg.osgrid.org: Gavin, have you tested it already? [11:39] Gavin.Hird @grid.xmir.org:8002: yes [11:39] Andrew.Hellershanks @hg.osgrid.org: ok, good. [11:39] Gavin.Hird @grid.xmir.org:8002: it is live on XMIR [11:39] George Equus: Someone having trouble emptying the Recycle bin? [11:40] Billy.Bradshaw @hg.viewtwo.net:8600: No George, but then I don't have an OSG avi [11:40] vegaslon plutonian: OSG only lets you empty the recycling bin from the OSG website [11:40] George Equus: I delete items, full procedure, when logging back on they return and I delete again. Eventually will stay away, after several days. [11:41] George Equus: Had the issue for a month or so. [11:41] Sheera Khan: Chat log of last weeks meeting is posted now :-) [11:42] Lucy Afarensis: yay ! [11:42] George Equus: Most do stay deleted, but sometimes certain objects refuse to... seem to be items from outside osgrid [11:42] Andrew.Hellershanks @hg.osgrid.org: George, is that an issue only in osgrid or do you see that issue in other grids running 0.9? [11:43] George Equus: this avatar here now [11:43] George Equus: osgrid [11:44] Billy.Bradshaw @hg.viewtwo.net:8600: George, maybe you should try again now OSG robust has been updated, that is if you are running 0.9 [11:44] George Equus: Hardly ever use other grids [11:44] Simulator Version v0.5 ruft: OpenSim 0.9.0.0 Dev       4be4e98: 2016-06-13 20:40:47 -0700 (Unix/Mono) [11:45] George Equus: Just deleted two stubborn items for the fourth time, in as many days [11:46] George Equus: they do rez normally when "returning"  so nothing obviously wrong with the prims/ textures [11:46] Lucy Afarensis: I must return to rl I am glad to see that that meetings continue. got worried when no log was posted last week [11:46] vegaslon plutonian: anyone using the -inifile parameter of opensim.exe with OpenSim 0.9? I suspect it is no longer working but I can not be 100% sure [11:46] Billy.Bradshaw @hg.viewtwo.net:8600: I don't see any issue like that on our 0.9 grid, maybe OSG specific [11:47] Andrew Hellershanks: vegaslon, I'm using the -inifile parameter [11:47] vegaslon plutonian: George Equus are you using the OSgrid website to empty your trash? [11:47] George Equus: Tomorrow will tell, or later today, if remain deleted. [11:47] George Equus: Of course Vega :) [11:48] vegaslon plutonian: when I tried to use it, it only loaded the opensim.ini that was in the bin folder [11:49] Gavin.Hird @grid.xmir.org:8002: Portugal - Iceland is about to kick off, so will return to RL [11:50] Andrew Hellershanks: George, the website queues a request to clear the recycle bin. It isn't instant. [11:50] George Equus: 12 hours should do it Andrew [11:51] Billy.Bradshaw @hg.viewtwo.net:8600: vegaslon, I use for robust mono --server --debug Robust.exe -inifile=Robust.HG.ini [11:51] George Equus: normally is done within 15-20 min [11:51] vegaslon plutonian: I stand corrected I was using -inidirectory [11:51] Andrew Hellershanks: George, yea, I would think so [11:51] Billy.Bradshaw @hg.viewtwo.net:8600: running 0.9 no problems [11:51] George Equus: A bug somewhere...?  but so far I seem alone with this issue [11:52] Andrew Hellershanks: ah, -inidirectory is a different beast. [11:52] Andrew Hellershanks: bye, Gavin. [11:52] Andrew Hellershanks: vegaslon, I was trying to use -inidirectory and found its behaviour unexpected. It tells OpenSim to read any .ini file it can find in that directory. [11:53] Andrew Hellershanks: I expected it to just tell OS where to find the files it wanted to read. [11:54] Robert Adams: by default, OS reads all the INI files in the directory "bin/config"... that's where I put my overrides [11:55] Robert Adams: that said, because of the odd way 'include's are done, you can't override everything [11:56] Robert Adams: looking at the code, 'inidirectory' sets the 'bin/config' directory name... the 'd' in 'directory' is lower case [11:56] vegaslon plutonian: I was just attempting to run multiple instances of opensim from the same directory to make it easier to git bisect [11:57] Andrew Hellershanks: Robert, I didn't dig in to the code. That point should be made more clear in the wiki. [11:57] Robert Adams: I haven't tried that... I think it is possible but it would be tricky. [11:58] vegaslon plutonian: people have done it and I found a tutorial but did not work [11:58] vegaslon plutonian: ended up loading the same opensim.ini file no matter what [11:59] Andrew Hellershanks: George, I'm being told that if you clear the recycle bin via the website you could wait 10 to 15 minutes then do it in world too. It may be that the data is being cached and the items eventually go away when the cache expires. [12:00] Andrew Hellershanks: vegaslon, do you need to read different bin/config directories as well as different OpenSim.ini? [12:00] George Equus: So, Andrew, reverse the procedure so to speak then. I'll try that [12:00] vegaslon plutonian: mostly just need to read a diffrent opensim.ini and region.ini [12:01] Andrew Hellershanks: George, the info mentioned to me seemed to be do via the website, wait, then do it in world. I don't think that is the reverse of what you were doing. [12:01] Andrew Hellershanks: or was it? [12:01] Andrew Hellershanks: vegaslon, to read a different OpenSim.ini you use -inifile. [12:02] George Equus: Andrew, I normally empty inworld, then immediately log on to web site and repeat [12:02] vegaslon plutonian: how to read a different region.ini? [12:02] Andrew Hellershanks: ok [12:02] Sheera.Khan @hypergrid.org:8002: In Metropolis it works when you first delete your stuff inworld, then purge it via web site, then empty trash [12:03] George Equus: Always worked fine up till a few months ago? when this all started [12:03] George Equus: so many changes in code? past weeks, something changed [12:05] George Equus: Can't really compare to other grids, hardly ever log on other than to osgrid these days [12:05] George Equus: got accounts on Metro, Kitely, and a few others [12:05] Andrew Hellershanks: vegaslon, if you are running from Linux you can use a symlink to set which Robust.ini file is to be read, or you could just copy or rename the file you want to use. [12:06] George Equus: Never conduct any business, picking up stuff there. [12:10] vegaslon plutonian: https://gwynethllewelyn.net/2012/06/01/opensimulator-multiple-instances-and-monitoring/ that was what I was trying to achieve [12:11] Andrew Hellershanks: Give me a sec to pull up that page [12:11] Andrew Hellershanks: [and then there were three] [12:11] vegaslon plutonian: very long winded, tells whole story [12:12] Andrew Hellershanks: regionload_regionsdir doesn't work for you? [12:13] Andrew Hellershanks: yea, I'm skimming over the page. grey text on white isn't a good colour combination [12:13] vegaslon plutonian: could not even get it to change the http_listener_port, maybe forgot one more port that needs to change as well [12:16] Andrew Hellershanks: vegaslon, Based on what the person is trying to do in the article I would use -inifile instead to point to the OpenSim.ini file. If you need to change to different config-includes then add -inidirectory to point to the different config-include directories. [12:17] vegaslon plutonian: ya realized that now, might actually work that way [12:17] Andrew Hellershanks: I have one small grid set up that uses -inifile for different regions but using a common config-includes directory. [12:17] Andrew Hellershanks: I'm also using it in a large grid. [12:18] Andrew Hellershanks: What I haven't done yet is to set different asset cache directories for each instance to point to a given instances root directory.