Orbiter crashes without reason

Disconnect

New member
Joined
May 5, 2008
Messages
98
Reaction score
0
Points
0
Location
Budapest
Hi

Today my orbiter started crashing, without any change.

Now some scenarios crash in 3 second after start, (of course orbiter.log writes nothing helpful), some scenarios work. Now i started a new scenario and tried to create vessels with scenario editor. What i see, if i create more than 6 individual vessels, the game will crash in seconds.
If i create multiple instances of the same vessel, then i can make more then 6, it won't crash.

It happens with default embedded 3d engine, and with d3d9 too.

Why is this?
 
Last edited:
Post your .log and .cfg file and we'll see if we can help you out.

Computer specs too, that might help as well.
 
As i wrote nothing helpful is written to log (never saw that it can be used for debugging, because of the lack of debug informations), but whatever, i've attached both.

Same happens with every graphics clients (dx7, dx9, ogla)

specs: quad core athlon II@ 3,1ghz, ATI radeon 3850 512mb GDDR3, 4 gb DDR2 800mhz ram.

I worked few days ago, now it isn't. My orbiter beta installation from SVN is works without problem.
 

Attachments

the log isn't for debugging it's for troubleshooting crashes.

Can you post the scenario that doesn't work?
 
As i wrote every scenario that have more than 5 types of vessels, crashes. Same if i adding the vessels with scenario editor, few seconds after adding the 6th vessel-> crash.

But a particular scenario which is crashes is the default Delta-Glider/DG Mk4 in orbit. Actually the 70% of scenarios crashing.

No any debugging tools for orbiter that writes a helpful error message?
 
As i wrote nothing helpful is written to log (never saw that it can be used for debugging, because of the lack of debug informations), but whatever, i've attached both.

Could you do the same for standard orbiter.exe, please? The log contains an NG run, and shows what appears to be a D3D9Client error message ("Failed to create vertex/index buffers. MaxVert=0, MaxFace=0, MeshOpt=0x8, nGrp=0"). So at least this is something helpful.

---------- Post added at 19:02 ---------- Previous post was at 18:47 ----------

Yep, it is definitely a D3D9Client message: https://bitbucket.org/face/ovp/src/...rsdk/D3D9Client/Mesh.cpp?at=D3D9Client#cl-107

Seems like the mesh handle is invalid. A log of the standard client would certainly be interesting.
 
As i wrote, this happens with every graphics engines, even with the default, non-ng orbiter (there nothing is written to log except that the simulation is started successfully).
I've also wrote that with the clean install from svn it works without problem, but i'm getting tired of reinstalling every addons in every month because orbiter starts doing this, and can't write out something helpful about why it's crashing... And it's more annoying that i've tried every vessels, and none of them causing this alone... just when more than 5 types are in the scenario.
 
Last edited:
so why not help us to help you and post things we've asked you to post? without delving in to the files and testing we cannot help you. Just because you think that there is nothing helpful in the log file or in posting the scenario it doesn't mean that someone won't spot an issue.

Help us to help you.
 
Now i've compared with e 100830 version files with my installation. The notable differences were that the deltaglider.dll was older the, and the base.cfg that had extra tile references (however it isn't changed since 2 weeks, and but the game have worked few days ago). Now i've restored these to default, and now it isn't crashing, no thanks to the non existing debug features of the game.

And by the way the log now looks exactly the same as when it was crashed.
 
Last edited:
Yep, it is definitely a D3D9Client message: https://bitbucket.org/face/ovp/src/...rsdk/D3D9Client/Mesh.cpp?at=D3D9Client#cl-107

Seems like the mesh handle is invalid. A log of the standard client would certainly be interesting.
So in this case Modules/D3D9Client/D3D9ClientLog.html file should tell which mesh file caused the error (may need changing the debug verbosity level for the D3D9Client).

The notable differences were that the deltaglider.dll was older
Did you install OGLA client? If so, that's a known issue - you shouldn't replace the deltaglider.dll with the one from OGLA.
 
yes i did install ogla. sadly it's abandoned, but still that is the best looking graphics client, the raytraced atmosphere is awesome.
 
yes i did install ogla. sadly it's abandoned
It wasn't abandoned as of Jan 29, 2013 14:21 UTC, but just waiting for the next major release of Orbiter to be updated for it. Where did you get the information it was abandoned?

From Orbiter Beta Team forum:
OGLAClient is the mess it is since i was chasing the changing specifications as the betas evolved, producing a nice, but indigestible bowl of stirred spaghetti.

So this time i'll wait for the release before adapting it to the new version.


Anyway, OGLA replaces DeltaGlider with a version using older VESSEL3 implementation from Orbiter 100606, and that's the reason for crashes. You shouldn't replace it, unless you only want to use Orbiter 100606, not patched to 100830.
 
As i wrote, this happens with every graphics engines, even with the default, non-ng orbiter (there nothing is written to log except that the simulation is started successfully).

This is not the full story. The log posted can't happen with every engine, because the error message is specific to D3D9Client. Skipping such important detail makes us fail in nailing down the real reason... to your disadvantage.

I've also wrote that with the clean install from svn it works without problem, but i'm getting tired of reinstalling every addons in every month because orbiter starts doing this, and can't write out something helpful about why it's crashing... And it's more annoying that i've tried every vessels, and none of them causing this alone... just when more than 5 types are in the scenario.

It writes out something helpful: the error message that is right there in the log. Why do you discuss it away?

Now i've compared with e 100830 version files with my installation. The notable differences were that the deltaglider.dll was older the, and the base.cfg that had extra tile references (however it isn't changed since 2 weeks, and but the game have worked few days ago). Now i've restored these to default, and now it isn't crashing

So you had a faulty installation, right? You had a faulty installation and got an error message about failing to load a mesh in the log.

And by the way the log now looks exactly the same as when it was crashed.

So you still get the D3D9Client error message?


On a side note, we are all trying to help, while you come across like someone DEMANDING a solution for his problems, and getting all annoyed and angry for not getting it ASAP.

Orbiter is free software, and we all offer help in our spare time. If you keep up with your attitude here, chances are high that you will sooner or later get no help at all. Please tone it down a little.
 
that error is not related to the crash as it happens before the crash, and before the D3D9Client: [Scene Initialized]

And as i wrote it was crashing with any graphics plugins. ogla and the default graphics isn't written that message.

That message appears now too, while now it isn't crashing. That's something non critical error.
 
Help from the community only depends on the information about the problem you provided.

Nobody is able to say why Orbiter is crashing if you only say that without any change to Orbiter it started crashing without a reason. There must be a reason, and it's something that changed either in Orbiter or in the environment (like installed software, hardware), like for example an add-on which replaced Orbiter stock file with its own.


As i wrote nothing helpful is written to log
And the reason for your crashes is indeed listed in the Orbiter.log file:
Code:
**** Orbiter.log
Build Aug 30 2010 [[b]v.100830[/b]]

{...}

Module DeltaGlider.dll ....... [Build 100607, [highlight]API 100606[/highlight]]
 
that error is not related to the crash as it happens before the crash, and before the D3D9Client: [Scene Initialized]

And as i wrote it was crashing with any graphics plugins. ogla and the default graphics isn't written that message.

That message appears now too, while now it isn't crashing. That's something non critical error.

Well, that's your interpretation of it. Maybe it is an indication that there is something wrong with the mesh constructor call. Whatever that error is, it is an important detail that should not be ignored in detecting what happens on your side.

Your previous posts suggest that you have the impression that Orbiter will stop working "after some time". This is certainly not the case in the strict sense of Orbiter having some code that gets executed after a while. Code is not rotting on its own, most of the time the execution environment just changed.

I strongly suspect that you do all your Orbiter tests/installations into the same Orbiter folder. This practice often leads to the impression that "it worked yesterday, now it crashes, but I didn't do anything", because an intermediate add-on installation/deinstallation is perceived as a no-op on the Orbiter installation. This is unfortunately not the case most of the time.

Therefore, many Orbiter users developed the habit of keeping a pristine copy (maybe with all the must-have add-ons installed and tested), while "working" into another one. If something "suddenly" stops working, the step back to a new copy is not so painful. Some even go so far as to version control their Orbiter directory in order to deal with the various add-on troubles that may happen.

I suggest you do the same, especially if you work with beta releases. This will not only help you to keep your installation free from bugs, but will also benefit the developers, as you will then be able to give precise bug reports, e.g. "I installed OGLA Deltaglider.dll, now more than 5 vessels crash my Orbiter". Who knows, maybe that info would have rung a bell for Artlav and/or other OVP developers.

regards,
Face
 
It wasn't abandoned as of Jan 29, 2013 14:21 UTC, but just waiting for the next major release of Orbiter to be updated for it. Where did you get the information it was abandoned?

I thought it's abandoned since wasn't new version was released since about 3 years.
 
Last edited:
Back
Top