Hello Michael,

Dirk Reiners wrote:
> I've got a crash when running OpenSG in cluster mode. I'm creating and 
> destroying some ChunkMaterials during runtime. This causes a crash of the 
> server 
> application. Using the debug libraries of OpenSG the server runs fine but I 
> noticed some of the following lines in the log:
> 
> WARNING:  FieldContainerFactory::unregisterFieldContainer:id -572662307 
> inconsistent with store size 8034!

hm, this almost looks as if the data is corrupted or the network 
protocol got out of sync.

> Looking into the code this message is only generated in debug mode

yes, unfortunately debug mode will then just never release the 
_pStoreLock :-/  I'm surprised this does not just hang afterwards?

> and the 
> origin of the crash (most likely an out of bounds array access) is avoided, 
> see:

yes, the question is where that odd value for the container id comes from?
Where does the call to unregisterFieldContainer come from, i.e. can you 
get a backtrace - perhaps by setting a conditional breakpoint in the 
function that triggers only if pFieldContainer.getFieldContainerId() is 
unusually large?

> I'll attached my full trace for your information. So am I doing something 
> wrong 
> or is this a bug in OpenSG? (The trace is at
> http://drop.io/OpenSG_Cluster_Problem, file server.zip.)

hm, I'd say it looks like a bug or at the least some subtle assumption 
in OpenSG. I'm just not quite sure how to track it down, so far my 
attempts to reproduce it have failed. The log unfortunately does not 
give a good idea where the unusual container id values come from.

        Cheers,
                Carsten


------------------------------------------------------------------------------
Download Intel® Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_______________________________________________
Opensg-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/opensg-users

Reply via email to