![]() |
paraview/paraFoam crash "ill defined primitiveEntry"
Hello ParaViewers,
I just walked into a problem with running paraview/paraFoam for multi-region cases. OpenFoam 2.1.0 paraview version 3.12.0 (64-bit); Both installed using apt-get (...) Running on Ubuntu 10.04 (lucid) 64-bit When the problem appears: We have a multi region case (like cthMultiRegionFoam tutorial). We prep the case, run the simulation and now want to view the results. Run: Code:
paraFoam -touchAllAfter that, paraview crashes, giving me the error message: Code:
--> FOAM FATAL IO ERROR: Do you know what may be the problem, and can you please help me overcome it? Best, Pawel |
Greetings Pawel,
What does the file "system/controlDict" have in the "FoamFile" zone, respectively in the line 10? Specifically for the line "version"? Does it look like this: Code:
FoamFileBest regards, Bruno |
Hello Bruno,
controlDict first lines, taken from tutorial case 2.1.0, no changes made: Code:
FoamFileI played a bit with this problem and got some interesting results. First, if I change "version 2.0;" to "version 2;" the error disappears from line 10 and moves to another entry with a dot, like the timeStep. Strangely enough, removing all dots does not help and the program crashes anyway. Secondly, I managed to open the cases, but using cunning trickstery. Lest suppose we have a pure multi region case, with no polyMesh directly in constant folder, and no fvFiels directly in system folders (there are separate polyMesh'es in constant/regionX, similar with system/regionX/fvFiels). We run: Code:
paraFoam -touchAllWe close the error box, delete the "builtin" artificial object, and try to open our regions (not the full case). This time everything works. It seems to me, that there is some misspell or small bug in paraFoam script, or within foamReader. Best, Pawel |
Hi Pawel,
Ah HA! I now know what the problem really is: you have your Linux installation in a language that isn't the standard English ;) Allow me to explain:
Code:
export LC_ALL=CThis is something that was detected some time ago and was meant to be fixed already directly in ParaView. But do to some other stupid bug, the internal fix in ParaView doesn't fix for all Linux installations, only some of them. Therefore, the only fix known to be always efficient is that magic LC_ALL variable :) So, you have (at least) two choices:
Bruno |
Good morning Bruno!
This is the reason why I did not put this thread to "bug" section- the magic line did the trick :D Just in my defence I will add that I tried changing dots to comas- did not work. Probably because of the conflict with linux language and the "fix" in paraview. Anyway, with the LC_ALL in .bashrc the problem is gone. Thank you very much for your help! With best regards, Pawel |
| All times are GMT -4. The time now is 02:34. |