[Djub]

Hi every body,

may I ask you a naive question?

I have openFoam on my linux server, and I installed ParaView on my windows PC. My linux server has a "samba" protocol file so that my PC can read the results of openFoam.

BUT I don't have the foam reader in ParaView. So I cannot use my windows-paraview to post process my calculations...

Does anybody has a solution?

My wish is to use the graphical enhanced capabilities of my PC (with its tough graphical cards) instead of the CPU time of my linux server (not efficient at all for that: it takes more than 2 minutes to refresh the screen!)

Thanks,

DJU

[wyldckat]

Greetings Julien,

Create a dummy file with the name "case.foam" inside your case folder. The file can be left empty.

Then open that file in ParaView and you're good to go! On the plus side, you can even open a decomposed case when using this internal reader!


I'll take the opportunity to do a shameless plug to blueCFD-SingleCore's "paraFoam.bat" batch file (I'm the guy responsible for this project ): http://openfoam21x.bluecfd-singlecor...n/paraFoam.bat

This batch file assumes that ParaView is visible from the Windows command line.

Best regards,
Bruno

[Djub]

I did not expect a so simple solution !!!

It works fine !

Thanks, and excuse me for not having looked up in the filter list...

[Djub]

Hi Bruno,

In found in ParaView User Guide that there is some parameters for the OpenFOAMReader.
http://paraview.org/Wiki/ParaView/Users_Guide/List_of_readers#OpenFOAMReader

But how to access to these parameters? Hos to change them?
I tried to write them in the*.foam "dummy" file, without success...

-----------------------------------------------------------------------------------------------------------------------
And 10 minutes later:

Stuby boy! You can choose these parameters in the "property inspector" of your *.foam case in your pipeline!

Nevertheless, thanks bruno!

Regards,

Julien (the so impatient)

[Djub]

Hi Everybody,

I have a new problem with ParaView on windows reading files on Linux:
If I try to read a decomposed case, there is an error:

ERROR: In ..\..\..\..\src\VTK\IO\vtkOpenFOAMReader.cxx, line 8481
vtkOpenFOAMReader (000000000AE9BB50):
Can't open Z:\OpenFOAM\julien-2.1.x\run\Bordeaux3\processor0/constant/


I think there is problem with the file separator (be aware about the slash / and antislash \ in the previous error message).

Does anybody have a solution?

[wyldckat]

Greetings Julien,

I suggest you make a copy of the folder "Bordeaux3" to your real Windows machine, because sometimes there is a problem with the virtual file system.

As for the slash and backslashes: they both should be valid in ParaView's perspective, at least when working on Windows.
Nonetheless, you can try and change the base folder in ParaView's Object Inspector to this:

Z:/OpenFOAM/julien-2.1.x/run/Bordeaux3/

Best regards,
Bruno

[Djub]

Hi Bruno,

thanks for help. I was prefering not to move the data, because it is not very convenient. But my I.T. manager suggests me also to make these moves, for traffic issues...
Nevertheless, I tried your solution (replacing \ by / in the Object Inspector): and paraview crashed! First a classic 'Output Massages' window with a red "ERROR :", and then very quickly a Windows window saying paraview.exe has stopped working..

By the way, I do not understand: my_case/processor0 does NOT have a "constant" folder ! It has only time folders. "Constant" is not is a processor, but at the "root" of my case. Why does paraview search for this folder ? This could be the source of the error, no?

[Djub]

Hi ,

I just tried to copy on my local disk before to open the case with paraview. But it does the same! Same error window, and then paraview crashes.
With or without changing the pathname in the Object Properties Inspector...

[wyldckat]

Hi Julien,

Mmm... the contant folder isn't always necessary for parallel cases. Older versions of OpenFOAM always had a "constant" folder, but nowadays it's not obligatory.

Which ParaView version are you using?
And is there any OpenFOAM tutorial in particular with which you get the same issue, so that I can try and replicate the same steps!?

Best regards,
Bruno

[Djub]

My Paraview is version: 3.14.1 64-bit

I tried pitzDailyMapped and it works fine. BUT there IS a constant folder in each processor# directory.

I founded the option -constant for decomposePar programm. But it dosenot work properly:

decomposePar -constant
--> FOAM FATAL IO ERROR:
size 247153 is not equal to the given value of 249000

I should emphasis that decomposePar with no arguments works properly.

Before this, I
- ran blockMesh, (=> 249000 cells)
- edited decomposeParDict to put "ptscotch",
- ran snappyHexMesh (=> 247153 cells)

Thanks for your help,
Djub

[wyldckat]

Hi Djub,

OK, now it's starting to make more sense! Although... there are two possibilities from your description:

1. If you ran snappyHexMesh in normal mode, did you use the "-overwrite" option?
2. When using snappyHexMesh in parallel, the mesh in the processor folders is different from the one in serial mode, which can lead to some applications having problems with agreeing on what should be done.
   The trick here should be to run:
   

   reconstructParMesh

   Run it at least after the mesh has been generated in parallel. Only afterwards will you be able to use reconstructPar... and perhaps in this case, only then will ParaView know what to do.

For more on this, check the reference cases listed here: http://openfoamwiki.net/index.php/Sn...ng_in_parallel

Best regards,
Bruno

[Djub]

By Bruno,

For the moment, I run sHM in normal mode because it is still too difficult in parallel (I am working on it but with no success!).

So, you think that I should run
- sHM with -overwrite
and - decomposePar with -constant

Yes?
I tried like this. There is no more paraView error (and the processor#/constant folders exist), BUT paraView crashes (a window from windows saying "paraView.exe has stopped working"). I precise that running on reconstruct case works (for time 0, I have both original and decomposed cases).

One solution is -of cource!- to reconstruct my case before to view it. But I am a bit disapointed ! It would be very more convenient without this additional step!