BUST: 1.1.X Crashes (Your Help Requested!)

[dooots]

So I am in the process of trying to reproduce all the crashes I have seen in an unmodded install and got an interesting crash.  I was testing attaching and detaching a radial decoulper when I could no longer attach the decoupler.  I tried to attach the decoupler a few times and the game crashed.  I don't know if this crash is related to the other VAB crashes but the log actually has some info in it so I figured I would post it here.

KSP.log

output_log.txt

[Claw]

6 hours ago, dooots said:

I tried to attach the decoupler a few times and the game crashed.

I've done this a bunch. So much so, that I didn't bother to count. What else does your ship look like? I've been doing this with a pretty simple ship, though I did fiddle with the staging a bunch on something like the KerbalX.

Your log looks like there's only 30ish parts.The errors thrown are also not consistent with the other reports, but still might be something worth investigating (though might not be related). 

[dooots]

I am just using the stock Kerbal 1-5.  I have been trying to crash in the VAB multiple times with various methods, but for some reason radial with symmetry is the only thing I can get to crash in under 10 minutes.

Here is link to everything I have collected so far testing data

I still have more I want to test but I can only sit and attach and detach parts for so long.

[Moesly_Armlis]

CTD happens when open the editor load a ship and then crash.  Reloaded KSP and the craft and then within a minute of editing another crash.  Sort of repeatable because that happened the same last night with another craft.

There is rumor that there are those who do not experience any CTD.

Hold on keep those pitchforks down for a minute! 

I have a question.

Is there any commonality for those systems?

 

Edit: Worked on same craft today and now no CTD after 2 hours.

Edited May 24, 2016 by MoeslyArmlis

[dooots]

More testing, by far and away the fastest way to crash the VAB for me is to drag a multi-part, surface attach, assembly, across the surface of a rocket with symmetry greater than 1.

For example load the Kerbal 1-5, copy the radial decoupler, srb, and nose cone by alt clicking on the radial decoupler.  Then set the symmetry to 2x and drag the copy across any part of the rocket.  I also make sure the drag/swipe motion starts off of the rocket and ends off of the rocket, but I do not know if this matters yet.  Most of the time KSP will crash for me in less than a minute.

So far I have caused the crash without making a copy of a part already on the rocket, but all crashes have happened when I had a Hammer SRB some place in the assembly.

I will keep testing and see if I can find a way to always crash KSP quickly.

[FullMetalMachinist]

Hi @Claw, unfortunately I can't help much with tracking down reproduction steps, because 1.1.x is rock steady for me. I've never had a crash. I did crash a few times during the prerelease, but that's it.

ASUS motherboard
i7-4770K 3.5Ghz (OC to 4.4Ghz)
EVGA GTX770
16GB RAM
Windows 7
DXDiag
output_log

Hopefully this helps you to find some common ground for what could be behind the crashes.

[dooots]

This has to be a race condition, if my frame rate drops to much from dragging the assembly over the ship it takes much longer to crash.  The sweet spot for me seems to be a simple radial srb booster in 2x symmetry. (radial decoupler, srb, and nose cone)  If I watch the performance monitor in the alt-f12 window when I drag such an assembly over the Kerbal 1-5 my fps is always just below 60 and seems to drop no lower then 57.8 fps.

Other notes:
I have not forced a crash by simply dragging a girder in 2x symmetry across and ship, but I have forced a crash by dragging a vector engine in 2x symmetry.
I have forced crashes with a larger assembly like the Dynawing's fuel tank and boosters in 2x symmetry.
I have forced a crash with a radial decoupler, fuel tank, and nose cone.
I have forced a crash in both radial modes.
Testing with the Kerbal X instead of the Kerbal 1-5 seems to take longer to crash on average.
I have not forced a crash in 1x symmetry.

One thing I don't get is I did not see any such crashes from dragging the radial decoupler over the ship yesterday when I was testing.  All of the crashes I had yesterday happened on a mouse click.

Oh I have vsync on, and task manager says ksp is using close to 50% of my cpu when in the VAB.

If you want any more info or for me to run more tests I am willing to do so, but for now I am out of ideas on what to try next.

[Claw]

22 minutes ago, dooots said:

If you want any more info or for me to run more tests I am willing to do so, but for now I am out of ideas on what to try next.

This is the most comprehensive information I've seen to date. Thank you very much for that!

I've actually been trying your Kerbal1-5 "drag the srb" for a while here and I'm still not hitting any issues. I've sped my processor up and slowed it down with no change. Mostly been in x64, but getting ready to switch to 32-bit next.

I've also been watching the in-game perf monitor, without noticing any particular change. What's your vsync set to? Every frame, or every other?

[dooots]

vsync is every frame

My guess is you may need a bigger booster to get it to crash.  As far as I could tell the rate of crash for me was pretty sensitive to the number of parts on the booster.  This also seems to align with the attach and detach tests I was doing before.  Using 2x symmetry for the attach and detach tests felt a bit smoother then using 4x, and I would not be surprised to find out if I did more of the tests that the 2x symmetry crashes happen faster then the 4x ones.  I also did an attach detach test today that replaced the radial decoupler of the booster with a modular girder segment.  This allowed me to attach and detach the booster faster but also resulted in small pauses in the game.  I think those small pauses is why only two of the three test managed to crash the game in less then ten minutes.

[Kwadima]

Hey guys! Great thread, I haven't had time to read it all through, maybe I will find a solution to whatever is causing my crash here but I will just write down what happens to me first!

KSP started through Steam client.

Specs copied from current KSP.log:

OS: Windows 10  (10.0.0) 64bit
CPU: Intel(R) Core(TM) i5-4210H CPU @ 2.90GHz (4)
RAM: 8116
GPU: NVIDIA GeForce GTX 860M  (1995MB)
SM: 30 (Direct3D 9.0c [nvd3dum.dll 10.18.13.5354])
RT Formats: ARGB32, Depth, ARGBHalf, Shadowmap, RGB565, Default, ARGB2101010, DefaultHDR, ARGBFloat, RGFloat, RGHalf, RFloat, RHalf, R8

Last lines of log:
.....
[LOG 04:32:15.698] commDish added to ship - part count: 41
[LOG 04:32:28.549] deleting part solarPanels4
[LOG 04:32:33.315] SnapMode
[LOG 04:33:41.035] KSPediaController: Compiling KSPedia databases
.....
The game crashes. This issue is mostly VAB-related, and it happens on either version (KSP/KSP_x64).

It often happens right after I've deleted a piece from my construction, within a few seconds or so. I take the unwanted item and drag it to the components list on the left, let go, and bam - Crash. This time it was a solar panel (1x6 fixed) but it seems to happen with any item.

The other times I've experienced crashes outside of the VAB is when I was constructing a Mun-base, docking two vessels together. Right after the dock, the game lagged for a few seconds and then crashed. I didn't save the log from that session unfortunately. That happened twice, and it seemed to be hard for my PC to compute whatever needed to be computed in order to re-orient the view, adjust to one vessel instead of two etc (and this PC can handle heavy games, thank you very much ;) ). This crash is not my main concern though, as the VAB-crash is most annoying.. and most frequent. It happens at random times too, or not at all, but i'm guessing that about of the KSP-runs crash  

This post is now long enough! Now i shall read all your fantastic posts and try to figure out if I can do something about this.. Any tips specific to my system are appreciated!

Cheers o7

[Alshain]

Claw, I don't know if this will help but it is possible one of the crashes are something saved with the craft.  I wish now I had saved the craft file, but sadly I did not have the foresight.  I had one lifter I was working on, and I grabbed an SRB by it's decoupler.  The game crashed.  I started the game again, loaded the auto-save from the test I had just done, tried to grab the exact same decoupler, crashed again.  A third attempt, once again crash, just from picking up the decoupler with the SRB attached.  After the third time I grabbed by a different decoupler in the symmetry set and did not produce a crash.  I'm really kicking myself for not saving a copy of that broken lifter file (this was before the thread was posted here).

 

I'm running an Intel 2500k, Nvida 950, Windows 7 x64, 8GB RAM, 64-bit KSP.

Edited May 25, 2016 by Alshain

[MarvinKosh]

@Alshain that sounds like a good theory. Bad craft data in might not do anything when you load in the craft, but if you then do something in the VAB or subsequently on a mission, that data is transformed. How exactly I don't know, I would assume it is copied and/or referenced in memory. However, craft validation (to allow old craft from previous versions of KSP to work) may already be a thing and would blow this theory out of the water. :/

[sormi]

I realized something that somewhat works for me. If I'm launching the game from Steam it won't take 5 minutes before I'm back to desktop from VAB. However, if I use the launcher from the KSP folder I crash far less frequently, say once or twice in an hour.

And yes, the crashes certainly have something to do with deleting parts and undoing.

Edited May 25, 2016 by sormi

[SAI Peregrinus]

I have NOT been having this issue. Placing and discarding parts causes no crashes. Symmetry mode makes no difference. I've got 229 mods with the following non-CKAN installed mods: FShangarExtender 3.4.7, Hooligan Labs Airships 5.2, OPT Spaceplane parts fork 1.8.1.1, S.A.V.E. 1.3.1-199, and UbioWeldContinued 2.3.0
And the CKAN mods as found at: http://puu.sh/p5bMw/4662daee3f.ckan

On load to the VAB the game takes 8710 MB ram, messing around spamming parts and clicking through all the tabs for 10 increases that to 8725 MB or so, but it stops increasing and stays there. Entering all the buildings in sequence brings the memory usage up to 8895 MB.

My Settings.cfg, mind the strange keybindings, I use a colemak keyboard: http://puu.sh/p5c3q/2d7f0fc414.cfg

My Physics.cfg: http://puu.sh/p5c4s/c4774ca386.cfg

My DxDiag: http://puu.sh/p5cgq/5ef2bbc16d.txt

My GPU is an R9 290X, since DxDiag only knows it's an R9 200 series. Windows 7, 64-bit.

[MarvinKosh]

Another log file.

In this instance I had crashes when I was trying to stage (dropping booster rockets) in atmosphere.

I encountered the same problem on the same rocket design on a later stage in space on a previous try.

Earlier in the evening I had a crash in the VAB.

Can't think of anything that ties them together, sorry.

 

[autumnalequinox]

My crashes decreased in frequency when I switched to my new rig with a different processor.  My old had a phenom 3.0 ghz quad and from what I read those are problematic. I have a fresh, totally clean OS and it was a good test point.  Heavily modded (70+).

Crashes are semi-frequent, sometimes I can go hours and hours.  Sometimes it happens quickly.  Always VAB, and always related to symmetry.  Going slowly in the VAB and not leaving symmetry on seems to help.  I used to have crashes with decoupling, but now the screen simply blinks and there is a second of lag and no problems.

No crash has produced a report.  No exceptions or other spam just normal game activity and then "crash!".

I need to regain access to my dropbox and I'll post a log tomarrow of my last few on this rig (only about 3 so far).  

[Claw]

Well, we found some potential issues that have to do with deleting parts. The problem now is trying to confirm if that has anything to do with this issue at hand. At face value (due to where the problem was), it might be related.

Work forges on, and I'm still reading (even if I don't respond to each individual comment). Thanks for both the "I've seen this" or "I've not had this issue." Adding the settings.cfg was a nice touch, since it could be a variety of things. 

Thanks!

[MarvinKosh]

Glad to hear you've got some new leads to follow.

[LTDominator]

game crashing randomly in VAB while building, first that damn microstuttering now that repurposed bovine waste comon squad fix that please...

dxdiag: http://pastebin.com/rXKj5qL9

ksplogfile: http://www.file-upload.net/download-11615910/KSP.log.html

Edit: game is modded pls move to ksp modded thread...

Edited May 26, 2016 by LTDominator
wrong thread oops...

[PhilMcgroin]

My game has been consistently crashing in the VAB since 1.1.x. I can't recall it crashing once in any scene outside that (maybe once when I tried to start time warp just as a 600 part station just came into physics load range, oops). It most commonly crashes, as often reported, when playing with symmetries in the editor. From what I can gather, usually when hovering a radial part onto something with symmetry already set, or when a part is in single mode hovered about to be placed, then the symmetry is increased. I've also often had it happen when picking up a part and placing it down to the side of the rocket (so its transparent and red) waiting to be re-attached after.

My crash just now occurred when I was moving a 4-way symmetried Thud engine around on a X32 fuel tank. Fairly heavily modded install at the moment, but I've had it happen with a stock install too.

 

Output Log https://drive.google.com/open?id=0B4sRKGVEZfYWYzMxY21ZT0pxb0U

KSP log https://drive.google.com/open?id=0B4sRKGVEZfYWUzdLS1pqSGpZYU0

Whats interesting, from looking at the KSP log, everything seems fine while I was building up to part 8 at 15:54:52.697, then it suddenly started throwing Input Null errors up at seemingly everything. But its interesting to note, that all the KKAOSS parts, then the roverwheels that follow, added to the ship fine with no crashes, mini hang ups, or anything seemingly threatening. Then the errors seem to come in with no parts being added. Then there is the big long list of:

[LOG 15:56:27.446] MGPlus.Start(): v01.03
[LOG 15:56:27.447] MGPlus StockPlus Enabled
[LOG 15:56:30.025] MGPlus.Start(): v01.03
[LOG 15:56:30.026] MGPlus StockPlus Enabled

events, I presume this is me rotating 4 x Thud engines around the fuel tank, dragging them off, putting them back on, etc, with what I guess is the GimbalPlus stock fix plus Module being called as the engines pop in and out of existence. Though, I never actually clicked to place any of the engines before it crashed, they were still waiting to be placed.

As the crash seemed to happen during these module calls, I wonder if anyone has ever had any of these VAB crashes when manipulating parts that have no modules. As 'Dooots' posted, they couldn't get a crash with a simple girder. Perhaps the problems are linked more specifically to Unity's memory handling when dealing with modules in the editor, rather than just parts, or rendering?

 

Editing in system specs:

• Win7 64 bit
• Intel core i5 661
• Gigabyte Mboard
• 8GB RAM
• AMD Radeon R9 200

Edited May 27, 2016 by PhilMcgroin
Adding system specs

[Claw]

1 hour ago, PhilMcgroin said:

Then the errors seem to come in with no parts being added.

Hmm, this may actually be pointing at something. This might be a useful lead, thanks for posting.

If I understand you right, the MGPlus log entries are streaming out after a certain point, but you aren't doing anything? They will certainly come up as you drag parts around (when they pop in and out of existence), but shouldn't be coming up if parts aren't being dragged around. So if they are, that's good to know.

[PhilMcgroin]

No, those MGPlus log entries - I imagine - would have been coming up exactly as I was dragging the engine around, as in the middle of all that dragging was when it crashed, and one of those entries is the last line. So, I guess that is fairly normal then, those weren't coming up out of thin air.

The part I was referring to, is these three entries:

[ERR 15:56:08.526] Input is null
   at System.Environment.get_StackTrace()
   at ConfigNode.CleanupInput(System.String value)
   at ConfigNode.AddValue(System.String name, System.String value)
   at BaseFieldList.Save(.ConfigNode node)
   at PartModule.Save(.ConfigNode node)
   at ShipConstruct.SaveShip()
   at ShipConstruction.CreateBackup(.ShipConstruct ship)
   at EditorLogic.SetBackup()
   at EditorLogic.<SetupFSM>m__101()
   at KerbalFSM.RunEvent(.KFSMEvent evt)
   at KerbalFSM.updateFSM(KFSMUpdateMode mode)
   at KerbalFSM.UpdateFSM()
   at EditorLogic.Update()

[ERR 15:56:21.607] Input is null
   at System.Environment.get_StackTrace()
   at ConfigNode.CleanupInput(System.String value)
   at ConfigNode.AddValue(System.String name, System.String value)
   at BaseFieldList.Save(.ConfigNode node)
   at PartModule.Save(.ConfigNode node)
   at ShipConstruct.SaveShip()
   at ShipConstruction.CreateBackup(.ShipConstruct ship)
   at EditorLogic.onSaveConfirm()
   at EditorLogic.saveShip(.Callback afterSave)
   at EditorLogic.saveShip()
   at UnityEngine.Events.InvokableCall.Invoke(System.Object[] args)
   at UnityEngine.Events.InvokableCallList.Invoke(System.Object[] parameters)
   at UnityEngine.Events.UnityEventBase.Invoke(System.Object[] parameters)
   at UnityEngine.Events.UnityEvent.Invoke()
   at UnityEngine.UI.Button.Press()
   at UnityEngine.UI.Button.OnPointerClick(UnityEngine.EventSystems.PointerEventData eventData)
   at UnityEngine.EventSystems.ExecuteEvents.Execute(IPointerClickHandler handler, UnityEngine.EventSystems.BaseEventData eventData)
   at UnityEngine.EventSystems.ExecuteEvents.Execute(UnityEngine.GameObject target, UnityEngine.EventSystems.BaseEventData eventData, UnityEngine.EventSystems.EventFunction`1 functor)
   at UnityEngine.EventSystems.StandaloneInputModule.ProcessMousePress(UnityEngine.EventSystems.MouseButtonEventData data)
   at UnityEngine.EventSystems.StandaloneInputModule.ProcessMouseEvent(Int32 id)
   at UnityEngine.EventSystems.StandaloneInputModule.ProcessMouseEvent()
   at UnityEngine.EventSystems.StandaloneInputModule.Process()
   at UnityEngine.EventSystems.EventSystem.Update()

[ERR 15:56:21.616] Input is null
   at System.Environment.get_StackTrace()
   at ConfigNode.CleanupInput(System.String value)
   at ConfigNode.AddValue(System.String name, System.String value)
   at BaseFieldList.Save(.ConfigNode node)
   at PartModule.Save(.ConfigNode node)
   at ShipConstruct.SaveShip()
   at ShipConstruction.CreateBackup(.ShipConstruct ship)
   at EditorLogic.onSaveConfirm()
   at EditorLogic.saveShip(.Callback afterSave)
   at EditorLogic.saveShip()
   at UnityEngine.Events.InvokableCall.Invoke(System.Object[] args)
   at UnityEngine.Events.InvokableCallList.Invoke(System.Object[] parameters)
   at UnityEngine.Events.UnityEventBase.Invoke(System.Object[] parameters)
   at UnityEngine.Events.UnityEvent.Invoke()
   at UnityEngine.UI.Button.Press()
   at UnityEngine.UI.Button.OnPointerClick(UnityEngine.EventSystems.PointerEventData eventData)
   at UnityEngine.EventSystems.ExecuteEvents.Execute(IPointerClickHandler handler, UnityEngine.EventSystems.BaseEventData eventData)
   at UnityEngine.EventSystems.ExecuteEvents.Execute(UnityEngine.GameObject target, UnityEngine.EventSystems.BaseEventData eventData, UnityEngine.EventSystems.EventFunction`1 functor)
   at UnityEngine.EventSystems.StandaloneInputModule.ProcessMousePress(UnityEngine.EventSystems.MouseButtonEventData data)
   at UnityEngine.EventSystems.StandaloneInputModule.ProcessMouseEvent(Int32 id)
   at UnityEngine.EventSystems.StandaloneInputModule.ProcessMouseEvent()
   at UnityEngine.EventSystems.StandaloneInputModule.Process()
   at UnityEngine.EventSystems.EventSystem.Update()

 

Above that chunk, the events all seem to relate to placing the rover wheels, and below it is the MGPlus entries from playing with the engine symmetry, but I didn't do anything between that. So that's what I meant by errors without part placement.

[Padishar]

13 hours ago, LTDominator said:

game crashing randomly in VAB while building, first that damn microstuttering now that repurposed bovine waste comon squad fix that please...

First, I understand your frustrations but use of that sort language to describe issues is not constructive and doesn't really encourage anyone to help you (e.g. if you posted anything like it in any mod thread that I develop then you wouldn't get help from me).

Secondly, the garbage collection stutter is not the same thing as "microstutter" and I really wish people would stop confusing the two (both here and on the Unity forums).  Microstutter refers to the very slight stuttering that you can see in the animation in some Unity games due to aliasing effects between the physics update rate and the frame rendering rate not being correctly accounted for.  The garbage collection stutter, as it can often pause the game for up to half a second for some people, really can't be described as "micro" without the application of an entire lorry-load of salt...

Third, well done for actually posting a log and dxdiag rather than just an unhelpful rant as so many are fond of doing...

[Sacnoth]

The crashes I've been experiencing in the VAB seem to be exclusively related to deleting parts. I've tried both deleting by dragging them over the leftside bar and by pressing the delete key, and both have caused my game to CTD without generating a crash log in the KSP directory.

I've had some other rare crashes I think, but deleting parts is the only reliable way to crash my game. It will happen sooner or later, And it doesn't seem to matter if I use high or low part count ships.

 

My Setup:

Windows 7  64-bit

Intel 2600K

Geforce 970

Gigabyte z68X UD3h

16GB Ram

[AlamoVampire]

Potential Data Point: I know, one trial a data point does not make, but, that is why I bring this up here, has anyone who has had these crashes tried running in Admin mode? I did tonight before I wandered off to play a diff game, and got a decent 53 minutes of solid stable play. any other thoughts? I posted this same thing into my own thread so, its in 2 locations now.