Mantis Bug Tracker
Main | My View | View Issues | Vote Tallies | FAQ
  

View Issue Details [ Jump to Notes ] [ Issue History ] [ Print ]
IDProjectCategoryDate SubmittedLast Update
0000780AI War[All Projects] Bug - Crash or ExceptionOct 20, 2010 9:54 pmOct 21, 2010 4:05 pm
Reportermillejoh 
Assigned Tox4000 
Severitycrash 
StatusresolvedResolutionfixed 
Product Version3.719 
Fixed in Version 
Summary0000780: Spinning Beach Ball of Death shortly after starting the program
DescriptionI am having a hard time finding a way to make this reproducible, but it seems to always occur within an hour of launching AIWar on my Macbook Pro (running Snow Leopard 10.6.4). I am attaching the ErrorsReportedByEngine.txt but neither the entries in that file nor in UnhandledErrors.txt have timestamps that correspond with the actual times the application hangs.

Very strange. I will try to update with more info if I find any. Hope I am an isolated case with my 2006 machine. Still enjoying the game quite a bit...
TagsNo tags attached.
Internal Weight
Attached Filestxt file icon ErrorsReportedByEngine.txt [^] (1,005 bytes) Oct 20, 2010 9:54 pm [Show Content]
txt file icon Sample of AI War: Fleet Command.txt [^] (41,175 bytes) Oct 21, 2010 10:46 am [Show Content]
log file icon Player.log [^] (24,058 bytes) Oct 21, 2010 12:13 pm
txt file icon Sample of AI War: Fleet Command - Normal.txt [^] (267,996 bytes) Oct 21, 2010 12:20 pm
txt file icon MacBook Pro.txt [^] (1,445,526 bytes) Oct 21, 2010 12:28 pm

- Relationships
duplicate of 0000644resolvedx4000 OSX crash 
has duplicate 0000803resolvedx4000 game freezes and requires restart (OS X, 3.801) 

-  Notes
(0001675)
x4000 (administrator)
Oct 20, 2010 9:56 pm

 That was with version 3.719, I see. Can you retest on 3.800? We fixed up a number of things there in an attempt to resolve this for the affected OSX users.
(0001754)
millejoh (reporter)
Oct 21, 2010 9:29 am

 I upgraded to 3.801 this morning and just let the program run. At 1:26:12 game time I saw the spinning beach ball and AIWar stopped responding. There is nothing I can see to indicate what was the cause. I was just letting the application run with very little interaction on my part. Again, maybe this is an isolated weirdness with my somewhat aged machine.

Let me know if there is anything else I can do to help.
(0001755)
keith.lamothe (administrator)
Oct 21, 2010 9:30 am

 Do you have any error log files in the game directory or in the RuntimeData subdirectory?
(0001756)
keith.lamothe (administrator)
Oct 21, 2010 9:31 am

 From this last freeze, I mean.
(0001757)
keith.lamothe (administrator)
Oct 21, 2010 9:33 am

 Also, if you have the time to test it, can you try:

1) Start the game.
2) Go to settings, audio, and disable all sound/music.
3) Save settings.
4) Close the game.
5) Start the game again, and repeat the test you did earlier with letting it run.

Just trying to isolate whether this really is a sound-system problem as the errors might indicate, or whether it's a more general threading problem.
(0001767)
millejoh (reporter)
Oct 21, 2010 10:46 am

 Keith,

Darn, I didn't think to look at the logs before restarting. I am attaching a log of the thread state of AI War at the time it locked. I'm not too good at reading these things, so I can't tell if there is anything useful or not. Note that I have left the game running, but Paused, for well over an hour and I have yet to see it lock-up. Pausing the game also pauses the music, so maybe it is a sound-system issue. I will also try to steps you recommended and see what happens.
(0001768)
keith.lamothe (administrator)
Oct 21, 2010 10:54 am

 I don't think I've seen a log like that one before, looks interesting :) Now if I can persuade my OS to recognize the line endings...

Anyway, thanks very much for your help tracking this down; it's very critical that the mac build be not crashing :)
(0001769)
keith.lamothe (administrator)
Oct 21, 2010 10:58 am

 Ok, opened that sampling log in a non-braindead text editor and it reads fine. Very cool info in there; is that from some kind of light profiling tool included with OSX?

Anyway, it didn't look like all that many threads in there; our previous problems were with holding more than 256 semaphores in a process, which generally meant having more than that many threads (due to lots of sounds, each playing from its own thread, or something like that), and I'm only counting 13 threads from that file in a quick glance.
(0001772)
x4000 (administrator)
Oct 21, 2010 11:34 am

 I'm running AI War on my own mac, and am just letting it run with some FRD ships defending me on my homeworld. So far 23 minutes and I'm not seeing anything that looks suspicious -- memory usage seems good, the number of threads in use are pretty low, the garbage collector is counting upwards at a pleasingly slow rate, etc.

I'd run it for several hours in the past without incident on my mac, but that was with sound/music disabled, so now I have those on to see what happens.

Next time you run into this issue, if you do, can you look in /Users/YourUserName/Library/Logs/Unity/Player.log and post that file? You have to get at that one before you start the game again, though, or else it gets overwritten. This is very curious! But hopefully something we can resolve soon. :/
(0001786)
millejoh (reporter)
Oct 21, 2010 12:12 pm

 Gents,

Happened to me again. I am basically doing the same thing as Chris- just letting it run with some turrets and FRD ships. Got to 1:12:35 game time before the AI get quite impatient with me and sent 1000 ships plus 4 or 5 hybrids to my homeworld. Shortly after that I got the spinning beach ball of death. I am attaching the player.log for your reading enjoyment. Think I'll try again at a lower difficulty and see how long things last this time around...
(0001787)
millejoh (reporter)
Oct 21, 2010 12:20 pm
edited on: Oct 21, 2010 12:21 pm

 And I forgot to mention- I obtained the thread state log from Apple's Activity Monitor.app, doing a "Sample Process" on the AI War process once it deadlocked. I just took another sample with everything "working". Things look a bit different, but I'm not sure if the delta is really all that informative. I will upload it as well; it will be the attachment that has "Normal" at the end of the name.
(0001788)
x4000 (administrator)
Oct 21, 2010 12:21 pm

 Hummmmm.... Nothing interesting in that log, unfortunately. How much ram do you have, out of curiosity?

I'm currently at 1:12 on my game, as fate would have it, and so far no issues.
(0001789)
keith.lamothe (administrator)
Oct 21, 2010 12:21 pm

 Nothing looks askance in that log, but thanks for including it, it would have told us something in the case of most types of errors.

Did you have sound on in this case?
(0001790)
x4000 (administrator)
Oct 21, 2010 12:22 pm

 I was referring to the player.log with that last.

I'll check out the other one. One other question, are you running fullscreen or windowed? You might try windowed for a while if it's doing this on fullscreen. And also, if you disable sound and music, does it still do this? We're trying to at least isolate where the error might be...
(0001792)
x4000 (administrator)
Oct 21, 2010 12:26 pm

 One thing that is interesting is that it's got a bunch of RakNet calls taking up half of the Normal log. RakNet is used by Unity under the hood for its internal networking, but we use Lidgren for actual networking, so the RakNet stuff really ought to be disabled. I'll have to look at that and see what I can see.
(0001794)
millejoh (reporter)
Oct 21, 2010 12:28 pm

 2GB. I don't see the memory usage on AI War get any higher than 900MB. I am having System Profiler do a dump of my system's configuration. Once that is done I will upload it for ever more reading enjoyment.
(0001797)
millejoh (reporter)
Oct 21, 2010 12:29 pm

 oooh, that file was HUGE. Sorry.
(0001799)
x4000 (administrator)
Oct 21, 2010 12:30 pm

 No worries, thanks. And 2gb should be more than enough -- 900mb at max seems right for the moment.
(0001801)
x4000 (administrator)
Oct 21, 2010 12:31 pm

 Just so that I'm clear:

The "Normal" one was from during a deadlock, and the other one was not, correct? If so, looks like it might be related to RakNet, mysteriously.
(0001803)
millejoh (reporter)
Oct 21, 2010 12:33 pm

 Note the latest crash for me was Windowed and with sound and music disabled. I went to windowed after the first time AI War crashed because in Fullscreen it locks up my Mac and I have to reboot!
(0001810)
x4000 (administrator)
Oct 21, 2010 12:37 pm

 ...Wow. :(
(0001813)
x4000 (administrator)
Oct 21, 2010 12:40 pm

 Question: are you playing with hybrids?
(0001814)
millejoh (reporter)
Oct 21, 2010 12:48 pm

 Yes, I am playing with hybrids. The file "Sample of AI War: Fleet Command.txt" is from an AI War instance that had deadlocked. The file "Sample of AI War: Fleet Command - Normal.txt" is from an instance that I am running (right now, actually) and has yet to deadlock- but I am only about 30 minutes into this current game.
(0001815)
keith.lamothe (administrator)
Oct 21, 2010 12:51 pm

 As much as I hate to think my fancy death-machines (hybrids) are causing the problem, could you try a test without any of them?

They certainly have enough special logic that it's possible they're doing something that would never happen in games lacking them.

Similarly, Chris, if you test on your mac with hybrids, that could be interesting. In multiple senses.
(0001816)
x4000 (administrator)
Oct 21, 2010 12:53 pm

 Indeed. It's also simply possible that hybrids are just making the game hit a certain memory threshold or something. Right now I don't have hybrids on, and the game is at 1:43 with no apparent aberrations. I'm going to give it until 3 hours, then try with hybrids if I don't have the issue.
(0001817)
keith.lamothe (administrator)
Oct 21, 2010 12:54 pm

 Ok; once I clean up a few other things I might try memory profiling a game with them on, just to see if they're doing anything untoward.
(0001818)
x4000 (administrator)
Oct 21, 2010 12:55 pm

 I wouldn't spend too much time on that until I or the other players can isolate it. We don't have time to spend on wild goose chases like that at the moment, it could be something completely else.
(0001819)
keith.lamothe (administrator)
Oct 21, 2010 12:57 pm

 Ok, that's true. Let me know if there are any particularly non-wild-geese to chase ;)
(0001826)
millejoh (reporter)
Oct 21, 2010 1:52 pm

 At 1:33:30 game time and still going strong. I set the AI level to Sandbox, but Hybrids are still enabled. Could it be that the deadlock is occurring at the end of game condition? I will let this run for a little longer and see if I can get past the 2 hour mark, then do a new game with the AI level at 10. My AI War kung-fu is weak, so I should quickly see if this problem is occurring when the AI blows up my command station.
(0001828)
keith.lamothe (administrator)
Oct 21, 2010 1:54 pm

 Oh, were the other crashes during some kind of assualt on your home system? Forgive me if I just wasn't reading closely enough before, but I want to make sure :)
(0001831)
x4000 (administrator)
Oct 21, 2010 2:01 pm

 VERY interesting -- thanks for that. My game is now at 2:51, and I'm cutting it off for now. Memory usage has been consistent and working as expected, so it doesn't seem to be that.

I need my mac for a bit to do the steamworks integration work, but then I'll run a test with hybrids to see if I can duplicate the problem with that.
(0001833)
millejoh (reporter)
Oct 21, 2010 2:03 pm

 No problem- this thread is getting a bit long. I am pretty sure all my crashes since upgrading to 3.801 have been when the AI blows up my command station. I am at 1:45 in game time so I think I will try restarting and see how quickly I can aggro the AI into destroying my home world and see if the deadlock happens again (maybe we should make an achievement "AggroMeister - Home Base Destroyed by AI in less than 5 minutes"). May be a little bit, though, I have to go pick up my 4 year old from pre-school. I've tried to get him to play AI War, but he seems more interested in Lego Star Wars :).
(0001837)
x4000 (administrator)
Oct 21, 2010 2:08 pm

 No worries -- and thanks. :)

Yeah, my son is 7 weeks old and won't even look at the screen when AI War comes on. I think Keith's two year old likes watching it a bit, though.
(0001839)
keith.lamothe (administrator)
Oct 21, 2010 2:08 pm

 Well, it's certainly a relief that this isn't a guaranteed crash on mac, at least, and that we may be able to pin down some relationship to the timing, etc.
(0001840)
keith.lamothe (administrator)
Oct 21, 2010 2:09 pm

 And she's three years old now, and calls AI War either the "Planet Game" or "The game with the 'tackers".
(0001841)
x4000 (administrator)
Oct 21, 2010 2:10 pm

 Yes, quite so. It's not like memory is always spiraling out of control or something. I'm thinking more and more that this is some sort of infinite-loop condition that we're reaching, but it's hard to be sure where it is: in the sound or music handling, or in the hybrid code, etc. I went through the music code pretty heavily this morning (and there's not really that much of it, honestly), and I'm pretty sure the problem is not there.
(0001842)
keith.lamothe (administrator)
Oct 21, 2010 2:12 pm

 Hmm, maybe when playing a taunt?
(0001843)
x4000 (administrator)
Oct 21, 2010 2:13 pm

 Ooh! That could be it, actually. That would make a lot of sense. And incidentally, I haven't been hearing the AI taunt me on the windows side at all, so that could be another sign that something is not right there in general.

Can you look at that, and get it working on the windows side at least?
(0001850)
x4000 (administrator)
Oct 21, 2010 4:05 pm

 Yay!

* Fixed a bug that was causing AI taunts not only to not play, but to create a memory leak (slash file-handle leak) especially on the OSX side. This was also the cause of sound effects suddenly cutting out on the OSX side in the cases where it would do that.

Issue Community Support
This issue is already marked as resolved.
If you feel that is not the case, please reopen it and explain why.
Supporters: No one explicitly supports this issue yet.
Opponents: No one explicitly opposes this issue yet.

- Issue History
Date Modified Username Field Change
Oct 20, 2010 9:54 pm millejoh New Issue
Oct 20, 2010 9:54 pm millejoh File Added: ErrorsReportedByEngine.txt
Oct 20, 2010 9:56 pm x4000 Note Added: 0001675
Oct 20, 2010 9:56 pm x4000 Assigned To => x4000
Oct 20, 2010 9:56 pm x4000 Status new => feedback
Oct 20, 2010 9:57 pm x4000 Product Version 3.800 => 3.719
Oct 20, 2010 10:33 pm x4000 Relationship added duplicate of 0000644
Oct 21, 2010 9:29 am millejoh Note Added: 0001754
Oct 21, 2010 9:29 am millejoh Status feedback => assigned
Oct 21, 2010 9:30 am keith.lamothe Note Added: 0001755
Oct 21, 2010 9:31 am keith.lamothe Note Added: 0001756
Oct 21, 2010 9:33 am keith.lamothe Note Added: 0001757
Oct 21, 2010 10:46 am millejoh Note Added: 0001767
Oct 21, 2010 10:46 am millejoh File Added: Sample of AI War: Fleet Command.txt
Oct 21, 2010 10:54 am keith.lamothe Note Added: 0001768
Oct 21, 2010 10:58 am keith.lamothe Note Added: 0001769
Oct 21, 2010 11:34 am x4000 Note Added: 0001772
Oct 21, 2010 11:35 am x4000 Relationship added has duplicate 0000803
Oct 21, 2010 12:12 pm millejoh Note Added: 0001786
Oct 21, 2010 12:13 pm millejoh File Added: Player.log
Oct 21, 2010 12:20 pm millejoh Note Added: 0001787
Oct 21, 2010 12:20 pm millejoh File Added: Sample of AI War: Fleet Command - Normal.txt
Oct 21, 2010 12:21 pm x4000 Note Added: 0001788
Oct 21, 2010 12:21 pm millejoh Note Edited: 0001787 View Revisions
Oct 21, 2010 12:21 pm keith.lamothe Note Added: 0001789
Oct 21, 2010 12:22 pm x4000 Note Added: 0001790
Oct 21, 2010 12:26 pm x4000 Note Added: 0001792
Oct 21, 2010 12:28 pm millejoh Note Added: 0001794
Oct 21, 2010 12:28 pm millejoh File Added: MacBook Pro.txt
Oct 21, 2010 12:29 pm millejoh Note Added: 0001797
Oct 21, 2010 12:30 pm x4000 Note Added: 0001799
Oct 21, 2010 12:31 pm x4000 Note Added: 0001801
Oct 21, 2010 12:33 pm millejoh Note Added: 0001803
Oct 21, 2010 12:37 pm x4000 Note Added: 0001810
Oct 21, 2010 12:40 pm x4000 Note Added: 0001813
Oct 21, 2010 12:48 pm millejoh Note Added: 0001814
Oct 21, 2010 12:51 pm keith.lamothe Note Added: 0001815
Oct 21, 2010 12:53 pm x4000 Note Added: 0001816
Oct 21, 2010 12:54 pm keith.lamothe Note Added: 0001817
Oct 21, 2010 12:55 pm x4000 Note Added: 0001818
Oct 21, 2010 12:57 pm keith.lamothe Note Added: 0001819
Oct 21, 2010 1:52 pm millejoh Note Added: 0001826
Oct 21, 2010 1:54 pm keith.lamothe Note Added: 0001828
Oct 21, 2010 2:01 pm x4000 Note Added: 0001831
Oct 21, 2010 2:03 pm millejoh Note Added: 0001833
Oct 21, 2010 2:08 pm x4000 Note Added: 0001837
Oct 21, 2010 2:08 pm keith.lamothe Note Added: 0001839
Oct 21, 2010 2:09 pm keith.lamothe Note Added: 0001840
Oct 21, 2010 2:10 pm x4000 Note Added: 0001841
Oct 21, 2010 2:12 pm keith.lamothe Note Added: 0001842
Oct 21, 2010 2:13 pm x4000 Note Added: 0001843
Oct 21, 2010 4:05 pm x4000 Note Added: 0001850
Oct 21, 2010 4:05 pm x4000 Status assigned => resolved
Oct 21, 2010 4:05 pm x4000 Resolution open => fixed

Copyright © 2000 - 2011 MantisBT Group
Powered by Mantis Bugtracker