1.0 - Constant crashing on OS X

[DavidHunter]

The iMc I'm using is still on Mavericks. If I was careful I could get as long at 15 mins gametime before it locked up.

I'm back on 0.25 for now so apart from keeping an eye on this thread and the devnotes, I have no involvement with KSP 1 in any of it's iterations.

Any idea on where to download these earlier versions? It's not on the main KSP download website and I was stupid enough to delete my copies of each version since 0.18.

[Jansn]

http://forum.kerbalspaceprogram.com/threads/24858-Old-versions-of-KSP-%28some-versions-still-wanted!%29

[DavidHunter]

http://forum.kerbalspaceprogram.com/threads/24858-Old-versions-of-KSP-%28some-versions-still-wanted!%29

Too bad that's only got up to v0.13. Thanks, anyway...

[sal_vager]

None of the versions after 0.13.3 except the 0.18 and 1.0 demo are free to redistribute yet sorry.

[xerogravity]

I never deleted a version until i was convinced the next one was better! When I got to 0.9, it wasn't as stable (for me) as 0.25 so both stayed on the system. At 0.9 I tried moving KSP to another Mac and it still wasn't as good as 0.25 on my old machine. KSP 1 just made it all worse. Because of that I have a good working copy of 0.25 which, while not having a lot of the fancy toys of the later versions, does allow me to play as long and as complex as I like without locking up. Happy enough with that. I'll keep subscribed to this thread and read the DevNotes. Who knows, one day Squad might fix it up. Fingers crossed for KSP1.1 Unity5 edition. You never know, that might be the one that leapfrogs me back up to date!

[sal_vager]

Well, I expect Unity 5 to have better Mac support, Unity are likely just going to concentrate on U5 and U4 will be deprecated (gotta push people to the new version somehow eh?).

Has anyone tried reverting to Mavericks?

[xerogravity]

I'm still on Mavericks, the iMac I use for KSP hasn't got Yosemite on it. I have to upgrade it at some point. That point would come sooner if KSP behaved better on Yosemite than Mavericks but from what I've read this isn't the case.

[sp1989]

I'm still on Mavericks, the iMac I use for KSP hasn't got Yosemite on it. I have to upgrade it at some point. That point would come sooner if KSP behaved better on Yosemite than Mavericks but from what I've read this isn't the case.

No it isn't Yosemite. It is definitely Squad. There are a ton memory leaks in general but compound that by there being even more for Mac users it really isn't Yosemite. I was running a ton of mods on Yosemite way back in the .24 days. It wasn't that long ago but I had almost no crashes and I built huge things. I built a station from Armageddon with KSO shuttles and KW rocketry. Slight lag but there were a lot of parts on the station but definitely no crashes. At least not every half hour like it is not with the 1.02 update.

Javascript is disabled. View full album

[Jouni]

No it isn't Yosemite. It is definitely Squad. There are a ton memory leaks in general but compound that by there being even more for Mac users it really isn't Yosemite. I was running a ton of mods on Yosemite way back in the .24 days. It wasn't that long ago but I had almost no crashes and I built huge things. I built a station from Armageddon with KSO shuttles and KW rocketry. Slight lag but there were a lot of parts on the station but definitely no crashes. At least not every half hour like it is not with the 1.02 update.

It's a mix of KSP version, mods, OS X version, and Mac hardware. The more parts and other assets you have, both from stock and mods, the more likely it is that the memory issues affect you. KSP 0.25, 0.90, and 1.0 all introduced a lot of new parts. KSP 1.0 also improved memory management, reducing the memory usage of the stock game a bit, despite the new parts. OS X Yosemite changed something for 32-bit apps, making it more likely that KSP runs out of memory. The effect the upgrade to Yosemite has may depend on hardware, especially on the GPU.

For me, KSP ran perfectly fine on Mavericks (and earlier OS X versions) up to 0.25. Version 0.90 had minor issues when I played it with a few mods, but it was definitely playable. Upgrading to Yosemite made KSP 0.90 unplayable, and I had to stop playing it until 1.0 came out. KSP 1.0 has been about as stable as 0.25.

There are two fundamental issues. Unity 4 doesn't work too well in OS X, causing memory leaks and wasting a lot of memory. Apple on the other hand doesn't really care about memory issues with 32-bit apps. For them, 32-bit software has been obsolete for a long time, and they think that only stupid people try running memory-hungry 32-bit apps on a Mac.

[MaxPeck]

My problems are most definitely not hardware related - running on a brand new macbook pro w/8 gigs of memory. I monitor my memory usage, and the 32-bit program is running out of memory far quicker than my machine is. Generally when it crashes, I still have 2-3 Gb of memory unused, which means its a 32-bit application memory issue, not a hardware issue.

It doesn't really matter if I'm playing Vanilla or with mods, eventually the same thing happens. I can play longer in vanilla, but eventually I click on one thing too many, and *POP*! My current workflow goes something like this: go to mission control, accept a mission, reload. Go to VAB, design a platform for the mission, reload (as needed). Go to launchpad (maybe), launch rocket, watch it go haywire, "revert to VAB", reload. Go back to VAB, tweak rocket, reload (as needed). Repeat above steps. THAT just gets me into space. Quicksaving before any maneuver or UI change is a necessity unless I want to just write off the minutes to hours of my life that it took me to reach that point. Eventually I'll click on the resource tab or the contracts tab or do something really outrageous like try to maneuver my spacecraft... and yahtzee. Reload as needed.

Sorry, that's unacceptable. Yes, I couched my frustration in a snarky request for more funny loading phrases. Squad may not take kindly to sarcastic feature requests, but you know what I don't take kindly to? Paying for something that's broken. I don't mine paying for software, especially when it's as awesome as KSP seems to be, but I expect something for my money, namely a functional product.

I really like this game, which is why I'm willing to deal with the frustration of having to reload as part of my workflow, but I would REALLY like to see Squad invest some time in customer service and give us a game that works.

[sal_vager]

JJE64, post your logs after the next crash and we will try to find the cause.

[SmallFatFetus]

OSX support is lacking at best. I have to tell myself constantly that it's not worth building a computer just to play ksp.

[MaxPeck]

Will do

[sal_vager]

It's possible this issue was not fully resolved, as Unity had to work around an OSX bug, a workaround that may not work on Yosemite.

[MaxPeck]

So here's the tail end of my latest scene switch crash...

[LOG 02:57:01.936] Flight State Captured

[LOG 02:57:01.954] Saving Achievements Tree...

[LOG 02:57:01.954] Saving Achievements Tree...

[LOG 02:57:02.123] Game State Saved to saves/Joe/quicksave

[LOG 02:57:02.286] Game State Saved to saves/Joe/persistent

[LOG 02:57:44.602] Flight State Captured

[LOG 02:57:44.618] Saving Achievements Tree...

[LOG 02:57:44.618] Saving Achievements Tree...

[LOG 02:57:44.783] Game State Saved to saves/Joe/persistent

[WRN 02:57:44.798] [HighLogic]: =========================== Scene Change : From FLIGHT to SPACECENTER (Async) =====================

[EXC 02:57:45.018] OutOfMemoryException: Out of memory

LibNoise.Unity.Generator.VoronoiCached.ConstructCubeArray ()

LibNoise.Unity.Generator.VoronoiCached..ctor (Double frequency, Double displacement, Int32 seed, Boolean distance)

PQSMod_VoronoiCraters.OnSetup ()

PQS.Mod_OnSetup ()

PQS.SetupMods ()

PQS.StartSphere (Boolean force)

PQS+.MoveNext ()

[EXC 02:57:45.057] OutOfMemoryException: Out of memory

LibNoise.Unity.Generator.VoronoiCached.ConstructCubeArray ()

LibNoise.Unity.Generator.VoronoiCached..ctor (Double frequency, Double displacement, Int32 seed, Boolean distance)

PQSMod_VoronoiCraters.OnSetup ()

PQS.Mod_OnSetup ()

PQS.SetupMods ()

PQS.StartSphere (Boolean force)

PQS+.MoveNext ()

[sal_vager]

Was this at the Mun? The voronoi craters are only generated at the Mun, so a lower terrain or render quality setting may help.

[MaxPeck]

It wasn't... it was switching from a station in LKO ~150km back to the space center.

[sal_vager]

Okay next time please post the whole logs, the KSP.log and Player.log contain mostly the same info but there's some stuff that is only in one or the other.

[MaxPeck]

I can't post the full log... the forum won't let me. Says it's 393000 characters with a max of 100000 allowed. Is there a way to attach a file to a post?

Also, you might find this interesting... I just had a new crash, but this one was different from the others. This time the framerate went to hell, and the game slowed down until I got the OS X beachball. When I look at this logs it's thousands of lines of this:

[LOG 17:18:20.278] Look rotation viewing vector is zero

[LOG 17:18:20.442] Look rotation viewing vector is zero

[LOG 17:18:20.592] Look rotation viewing vector is zero

[LOG 17:18:20.759] Look rotation viewing vector is zero

[LOG 17:18:20.908] Look rotation viewing vector is zero

[LOG 17:18:21.061] Look rotation viewing vector is zero

[LOG 17:18:21.206] Look rotation viewing vector is zero

[LOG 17:18:21.359] Look rotation viewing vector is zero

[LOG 17:18:21.508] Look rotation viewing vector is zero

[LOG 17:18:21.642] Look rotation viewing vector is zero

[LOG 17:18:21.792] Look rotation viewing vector is zero

[LOG 17:18:21.941] Look rotation viewing vector is zero

[LOG 17:18:22.107] Look rotation viewing vector is zero

[LOG 17:18:22.257] Look rotation viewing vector is zero

[LOG 17:18:22.408] Look rotation viewing vector is zero

[LOG 17:18:22.541] Look rotation viewing vector is zero

[LOG 17:18:22.694] Look rotation viewing vector is zero

[LOG 17:18:22.844] Look rotation viewing vector is zero

[LOG 17:18:22.987] Look rotation viewing vector is zero

[LOG 17:18:23.139] Look rotation viewing vector is zero

[LOG 17:18:23.291] Look rotation viewing vector is zero

[LOG 17:18:23.443] Look rotation viewing vector is zero

[LOG 17:18:23.575] Look rotation viewing vector is zero

[LOG 17:18:23.710] Look rotation viewing vector is zero

[LOG 17:18:23.875] Look rotation viewing vector is zero

[LOG 17:18:24.027] Look rotation viewing vector is zero

[LOG 17:18:24.173] Look rotation viewing vector is zero

[Master Tao]

There's no way to attach anything, but you can upload it to a file hosting site like Dropbox or Google Drive and then post a link here.

[fluxx23]

Been having the same issue. Can't pin it down but I doubt submitting all the "required information" will result in anything so I'm posting the crash log. Yes I'm pessimistic about a fix because it seems many OSX users have been having this issue for quite some time and no solution has been found.

UnloadTime: 37.262978 ms
KSP(9491,0xa10021d4) malloc: *** mach_vm_map(size=4194304) failed (error code=3)
*** error: can't allocate region
*** set a breakpoint in malloc_error_break to debug
Receiving unhandled NULL exception
Obtained 28 stack frames.
#0  0x0000000134ae3d in array_push_back
#1  0x0000000134b651 in mono_add_process_object
#2  0x0000000134b28b in mono_unity_liveness_calculation_from_statics
#3  0x000000002ea948 in GarbageCollectSharedAssets(bool)
#4  0x0000000030b902 in CleanupAfterLoad()
#5  0x000000002efe45 in LevelLoading::LoadLevel(int, std::string const&, AwakeFromLoadQueue&)
#6  0x000000002efad7 in PlayerLoadLevelFromThread(int, std::string const&, AwakeFromLoadQueue&)
#7  0x000000002f4c91 in PreloadLevelOperation::IntegrateMainThread()
#8  0x000000002f3cdb in PreloadManager::UpdatePreloadingSingleStep(bool)
#9  0x000000002f4308 in PreloadManager::WaitForAllAsyncOperationsToComplete()
#10 0x000000002f1e14 in PlayerLoop(bool, bool, IHookEvent*)
#11 0x0000000065e65c in -[PlayerAppDelegate UpdatePlayer]
#12 0x00000090b176cf in __NSFireTimer
#13 0x0000009b0d0006 in __CFRUNLOOP_IS_CALLING_OUT_TO_A_TIMER_CALLBACK_FUNCTION__
#14 0x0000009b0cfab4 in __CFRunLoopDoTimer
#15 0x0000009b14b54f in __CFRunLoopDoTimers
#16 0x0000009b087531 in __CFRunLoopRun
#17 0x0000009b086aa6 in CFRunLoopRunSpecific
#18 0x0000009b08690b in CFRunLoopRunInMode
#19 0x0000009ac648f8 in RunCurrentEventLoopInMode
#20 0x0000009ac64503 in ReceiveNextEventCommon
#21 0x0000009ac6442c in _BlockUntilNextEventMatchingListInModeWithFilter
#22 0x000000939aa721 in _DPSNextEvent
#23 0x000000939a9dc5 in -[NSApplication nextEventMatchingMask:untilDate:inMode:dequeue:]
#24 0x0000009399e77c in -[NSApplication run]
#25 0x00000093913bc0 in NSApplicationMain
#26 0x0000000065e3ab in PlayerMain(int, char const**)
#27 0x00000000003095 in start
Stacktrace:




Native stacktrace:


    0   libsystem_kernel.dylib              0x98c0769a __pthread_kill + 10
    1   libsystem_c.dylib                   0x96e86eee abort + 156
    2   KSP                                 0x0034a822 _Z12HandleSignaliP9__siginfoPv + 34
    3   libmono.0.dylib                     0x0134981a mono_chain_signal + 76
    4   libmono.0.dylib                     0x01293ae2 mono_sigsegv_signal_handler + 234
    5   libsystem_platform.dylib            0x9054603b _sigtramp + 43
    6   ???                                 0xffffffff 0x0 + 4294967295
    7   libmono.0.dylib                     0x0134b651 mono_add_process_object + 115
    8   libmono.0.dylib                     0x0134b28b mono_unity_liveness_calculation_from_statics + 516
    9   KSP                                 0x002ea948 _Z26GarbageCollectSharedAssetsb + 3128
    10  KSP                                 0x0030b902 _Z16CleanupAfterLoadv + 18
    11  KSP                                 0x002efe45 _ZN12LevelLoading9LoadLevelEiRKSsR18AwakeFromLoadQueue + 869
    12  KSP                                 0x002efad7 _Z25PlayerLoadLevelFromThreadiRKSsR18AwakeFromLoadQueue + 39
    13  KSP                                 0x002f4c91 _ZN21PreloadLevelOperation19IntegrateMainThreadEv + 65
    14  KSP                                 0x002f3cdb _ZN14PreloadManager26UpdatePreloadingSingleStepEb + 299
    15  KSP                                 0x002f4308 _ZN14PreloadManager35WaitForAllAsyncOperationsToCompleteEv + 136
    16  KSP                                 0x002f1e14 _Z10PlayerLoopbbP10IHookEvent + 708
    17  KSP                                 0x0065e65c -[PlayerAppDelegate UpdatePlayer] + 252
    18  Foundation                          0x90b176cf __NSFireTimer + 119
    19  CoreFoundation                      0x9b0d0006 __CFRUNLOOP_IS_CALLING_OUT_TO_A_TIMER_CALLBACK_FUNCTION__ + 22
    20  CoreFoundation                      0x9b0cfab4 __CFRunLoopDoTimer + 1316
    21  CoreFoundation                      0x9b14b54f __CFRunLoopDoTimers + 351
    22  CoreFoundation                      0x9b087531 __CFRunLoopRun + 2081
    23  CoreFoundation                      0x9b086aa6 CFRunLoopRunSpecific + 390
    24  CoreFoundation                      0x9b08690b CFRunLoopRunInMode + 123
    25  HIToolbox                           0x9ac648f8 RunCurrentEventLoopInMode + 262
    26  HIToolbox                           0x9ac64503 ReceiveNextEventCommon + 192
    27  HIToolbox                           0x9ac6442c _BlockUntilNextEventMatchingListInModeWithFilter + 99
    28  AppKit                              0x939aa721 _DPSNextEvent + 742
    29  AppKit                              0x939a9dc5 -[NSApplication nextEventMatchingMask:untilDate:inMode:dequeue:] + 350
    30  AppKit                              0x9399e77c -[NSApplication run] + 907
    31  AppKit                              0x93913bc0 NSApplicationMain + 2082
    32  KSP                                 0x0065e3ab _Z10PlayerMainiPPKc + 731
    33  KSP                                 0x00003095 start + 53


Debug info from gdb:


KSP(9491,0xa10021d4) malloc: *** error for object 0xee7fab04: incorrect checksum for freed object - object was probably modified after being freed.
KSP(9491,0xa10021d4) malloc: *** set a breakpoint in malloc_error_break to debug

[MaxPeck]

I agree with fluxx above. This isn't a new or isolated problem, it's a well-known one. Fix this.

So here's one from last night. I had just shot a rocket at my orbiting station and was on final approach to dock, I selected a docking port to attach to, and as soon as I moused over the docking port - blamo. The whole works froze up and I got the spinning beachball. I haven't been able to find "player.log", so here is the tail end of my KSP.log. It's thousands and thousands of lines of

[LOG 13:57:26.551] Look rotation viewing vector is zero

Followed by:

[LOG 13:57:36.609] Look rotation viewing vector is zero[LOG 13:57:36.751] Look rotation viewing vector is zero

[LOG 13:57:36.891] Look rotation viewing vector is zero

[LOG 13:57:44.644] Unpacking Reach Station

[LOG 13:58:47.645] [Orbit Targeter]: Target is probeCoreOcto (Reach Station)

[LOG 13:58:59.294] [CurrencyConverter for Outsourced R&D]: 0 Funds taken, yields 0 Science

[LOG 13:58:59.294] [CurrencyConverter for Unpaid Research Program]: 0 Reputation taken, yields 0 Science

[ERR 13:59:08.963] Could not allocate memory: System out of memory!

Trying to allocate: 26272B with 32 alignment. MemoryLabel: VertexData

Allocation happend at: Line:122 in

Memory overview

[ ALLOC_DEFAULT ] used: 397077722B | peak: 0B | reserved: 441696388B

[ ALLOC_GFX ] used: 288787364B | peak: 0B | reserved: 337004496B

[ ALLOC_CACHEOBJECTS ] used: 90088B | peak: 0B | reserved: 12582912B

[ ALLOC_TYPETREE ] used: 5703B | peak: 0B | reserved: 4194304B

[ ALLOC_PROFILER ] used: 71604B | peak: 0B | reserved: 4194304B

[MaxPeck]

You know what would be awesome, since many of us are having this issue?

Any acknowledgment from Squad whatsoever that they're aware of it and working on a fix, instead of just letting us twist in the wind. I don't often buy computer games, but this game was so compelling to me that I just had to plunk down my money. I absolutely feel like I got sold something that should be a release candidate, not a production game. This is a major issue affecting a large segment of your user base. At least address it, let us know you're working on it, something.

It's called customer service, and it's not exactly a new concept.

[Frank_G]

It is indeed the most unstable version of KSP so far. I am really looking forward to the Unity 5 update... Better memory usage and multi threading... As Macs love using plenty of RAM, my hopes are pretty high that it will change a lot..

[Zucal]

Any acknowledgment from Squad whatsoever that they're aware of it and working on a fix, instead of just letting us twist in the wind. I don't often buy computer games, but this game was so compelling to me that I just had to plunk down my money. I absolutely feel like I got sold something that should be a release candidate, not a production game. This is a major issue affecting a large segment of your user base. At least address it, let us know you're working on it, something.

This. I love this game, so I endure the crashes and freezes that happen every 20 minutes, spend hours going through threads like this one, and yet there has never been any acknowledgment that this game is unplayable for a good chunk of their user base. This game is supposedly in a finished state, and yet it's so bugged for me I can barely use it. Squad, just say something that indicates you are trying to fix this. I didn't pay 35 dollars to not play the game.