decomposePar 4-core warning/error?
 

Hi guys.
I've got a 2D VAWT simulation running on a 6-year-old dual-Xeon processor with hyperthreading (so effectively quad-core) computer.
I've run the very same simulation on my 5-year-old Core 2 Duo laptop, with no hitches. But on the quad-core, I got this weird error in my log.decomposePar file. Any advice as to what it might mean?

I'm not even sure if it's really an error, since decomposePar finishes without a problem and pimpleDyMFoam runs afterward without crashing ...

Greetings Anand,

This is indeed an intriguing warning message... it seems to have loaded the library "libincompressibleRASModels.so" 4 times, where it complains 3 times when it finds the repeated object "alphatJayatillekeWallFunction".

OK, a few questions:

1. How did you install OpenFOAM 2.2.1? Did you use the official Deb packs for Ubuntu?
2. Do you also have OpenFOAM 2.2.0 installed? And if you do, how are you keeping them apart?
3. Does your case have the entry "libs" in the file "system/controlDict"? And if it does, what does it contain?
4. Do any of the boundary conditions in your case use the incompressible "alphatJayatillekeWallFunction"? Possibly 4 boundary conditions?
5. Do you have any customized boundary conditions in binary form? If you do, is there any named "alphatJayatillekeWallFunction"?
6. Have you modified the global or personal "controlDict" files? For more on this topic: http://openfoamwiki.net/index.php/Ho...in_debug_flags

Either way, at first glance, it doesn't seem to be too problematic, but if the solver complains about the same, then it might be a good reason for concern.

Best regards,
Bruno

I'm using Lubuntu 13.04 (raring) (dual-booting with Windows).

Looks like the solver had the same error in the beginning, but it still continued solving. I had stopped the simulation myself after about two hours or so since I had to switch to Windows for a while. After about 2 hours of simulation it had got to approximately 0.8s of simulated time, and I had set it to simulate 8 seconds.

Background information on the case: I basically used the AMI mesh from the tutorial incompressible/pimpleDyMFoam/propeller, converted into a 2D mesh using my own STL files, and basically dumped that mesh into the tutorial incompressible/pimpleDyMFoam/wingmotion2D/pimpleDyMFoam. After a lot of work and a lot of help, I got it solving.
This worked fine on my dual core. The only change I made was in decomposeParDict and the Allrun bash script to break it into 4 instead of 2.

1) Yes, I installed OpenFOAM 2.2.1 using the deb packs.
2) I don't have 2.2.0 installed.

3) The libs entry in controlDict contains:
This should be exactly the same as the wingMotion tutorial.

4), 5), and 6) I'm not familiar enough with OpenFOAM to try any major customizations, I stuck with using the boundary conditions from the wingMotion tutorial with no changes. The only change I made to the controlDict file was the maxCo parameter, from 0.9 to 0.5.

https://www.dropbox.com/s/ip7zqpslwz1zass/VAWT_sim.zip

There's my (cleaned) case files, with the log files from the previous run included. Looks like the error's in all of the files. I'm not sure why it would call it multiple times ... And I have no memory of ever typing in any different wallFunction except what was in the wingMotion tutorial.
Thanks for your advice!

Hi Anand,

I'll try to look at this as soon as I can, possibly only in the coming weekend :(

Best regards,
Bruno

Hi, I have encountered the same issue and narrowed it down to the following (in my case):

forces function in the controlDict:

If you comment the section, the error disappear - but in all likelyhood the forces output disappear as well...

Greetings to all!

I've narrowed it down.
It's as Julien says, the "libforces.so" library helps introduce the problem. I can replicate this issue with the tutorial case "mesh/moveDynamicMesh/simpleHarmonicMotion", while using OpenFOAM 2.2.1 installed from Ubuntu Deb packages on Ubuntu 12.04, both i686 and x86_64. I only need to run blockMesh, in order to trigger the same messages.

Nonetheless, the error doesn't come from "libforces.so" itself, it's because of the following two files:
In 2.2.1, they have the same exact name runtime name "alphatJayatillekeWallFunction", which is why the message is shown. But this has been fixed on the day right after the release: https://github.com/OpenFOAM/OpenFOAM...94a946c945a99b

In theory, you should only have problems if you eventually do need this wall function "alphatJayatillekeWallFunction" in either "incompressible" or "compressible", because only the first one registered is the most likely to be used. If this is the case, you'll have to switch to OpenFOAM 2.2.x or apply manually the commit I mentioned above and rebuild only the library "$FOAM_SRC/turbulenceModels/compressible/RAS/".


I've moved this thread to the bugs section, so that it will make it easier to be found in the future.

Best regards,
Bruno

Much thanks, Bruno!

As far as I can tell, it doesn't affect anything during runtime since I don't actually call on that wallFunction for anything. I think I'll leave it as is since my simulation is running well for now.

hello,
just to confirme that the solution from Bruno works find for me
thanks again Bruno!

Yeah, I haven't had this error since I upgraded to the latest GitHub release. It wasn't bothering the simulation, but it's no longer bothering me :D

I have the same problem and it appears to be triggered whenever the "libforces.so" is included in controlDict. I am running OpenFOAM 2.2.1. So how do I fix it?

Please let me know, thanks!

very simple, just change the files
OpenFOAM-2.2.x / src / turbulenceModels / compressible / RAS / derivedFvPatchFields / wallFunctions / alphatWallFunctions / alphatJayatillekeWallFunction / alphatJayatillekeWallFunctionFvPatchScalarField.H
as describe in the link, and compile
cheers LL

Thankyou verymuch for your response. As I am a newbie when it comes to linux admin, I am not familiar with the steps involved. Are the steps described somewhere that you may know of or will you be able to provide them?

Sorry for the trouble.

Thanks
Musa

Hi Musa,

I have personally gone the way of not trying to fix it. I am just waiting for the next oF version to be released. The issue is "cosmetic" for me as it only affect the screen output.

Julien

Understood. Thankyou for your response.

In case you still wanted to know, these are the steps I took to recompile:

1. 'sudo bash' and login

2. Replace the 'alphatJayatillekeWallFunctionFvPatchScalarField.H ' file linked above with yours in opt/openfoam221/src/turbulenceModels/... etc (same path); you can get to 'opt/openfoam221/src' by 'cd $FOAM_SRC'

3. Go to '/opt/openfoam221/src/turbulenceModels/compressible/RAS'

4. 'wmake'

5. It will compile in a few minutes or so, and then the errors should be gone.

Robert:

Thankyou very much. Much appreciated!!

swak4foam patch give the same error
 

Hi everyone,

I have a problem similar than you. But the issue appeared after I applied a patch for swak4foam given by one post in this forum. The patch make ale to have libsimpleSwakFunctionObjects.so and libswakFunctionObjects.so. Now, I didn't have error message concerning that lib but error message similar than you : Duplicate entry .......

If, I don't use libs ( libswakFunctionObjects.so .. No problem ...

Just to inform all of us. Also, I use LES, and all Duplicate entry have a matter with LES ... I don't know what is the problem

Greetings zarox,

Sorry, but I could not understand what is the specific problem you're having, since you did not provided any specific information. For example:

1. Quote:

   Originally Posted by zarox (Post 473778) But the issue appeared after I applied a patch for swak4foam given by one post in this forum. The patch make ale to have libsimpleSwakFunctionObjects.so and libswakFunctionObjects.so.

   What specific patch are you referring to and which specific version of swak4Foam are you using?
2. Quote:

   Originally Posted by zarox (Post 473778) Also, I use LES, and all Duplicate entry have a matter with LES ... I don't know what is the problem

   If you don't provide the output given by the solver, we cannot ascertain what exactly is wrong.
3. What specific version of OpenFOAM are you using?

Best regards,
Bruno

Informations
 

Hi Wyldckat,

I have read many post from you, your are a master! In order to try to give you some information, the patch was given by you in that post :


Before using this patch libsimpleSwakFunctionObjects.so and libswakFunctionObjects.so were not in the lib directory if memory serves well!

In addition the swak4foam version is 0.2.4 and the OpenFOAM version 2.2.2


Concerning the output log, I give a part in the attached file.

Thanks to your reply.

Best Regards


Emeline

Hi Emeline,

According to the output and from your description, the problem does not seem to be related to swak4Foam. It seems to me that you're using a custom solver which might be the one responsible for this problem in particular.

My guess is that if you run this command:
it will give you a list of libraries it links to and among them are some of your own custom libraries, which very likely were copied from the original turbulence models. Since you're using the installation at "/opt", you can run this command to see a short list, among which are your custom libraries:
I'm guessing that you have a library named "libMyincompressibleLESModels.so", which is nearly identical to the official "libincompressibleLESModels.so", which is why you're getting the error messages you're getting!

At least, this is my guess, without more information.

Best regards,
Bruno

But without some swak4foam librairies
 

Hi Bruno,

Thanks to your answer! Very interresting, now I know a new command ldd very useful to show all librairies link. But to come back to the probleme, Yes I have libMyincompressibleLESModels.so, just a change in the genEddy, if memories serve, in order to add a sponge layer at the outlet (i.e. increase viscosity). So the librairy is very similar than the original. So Your answer seems to sound good, but the strange thing is, in the controlDict file I have some librairies add libsimpleSwakFunctionObjects.so and libswakFunctionObjects.so in order to do some cutting plane and free surface evolution partial solution. That librairy did'nt work first I add your patch, know it work but a Duplicate status is given. If I don't add these librairy not error is print at the beginning of the execution relating to the duplicate process... So, It seems to me to be related to the swak4foam, but also Your explaination seems good but don't take account for the result without the swak librairy. So, I don't know, How can I remove the duplicate status, Is it a true issue?

Nice to have some exchange with you..


Greeting!


Emeline

Hi Emeline,

If you run this command:
you'll see that it indicates that the library "libswakFunctionObjects.so" depends on "libincompressibleLESModels.so".

There are a few possible solutions you can use:

1. You can edit all ".H" files in your customized turbulence libraries and renamed the entries like this one:
   
   Code:

   ---

   TypeName("laplace");

   ---

   to something like this:
   
   Code:

   ---

   TypeName("laplaceMine");

   ---

2. Or you can have only the classes you strictly need in your own library.
3. Or you can embed the classes you need into your solver!

Best regards,
Bruno

Great!
 

Hi Bruno,

You give me the ldd command and I was not able to re-use to have more informations! What kind of poor little cdf girl, I done! I promise, I would do better next time!

Thank for solutions, I try to do the second one, but i don't find the proper way in my mind, because, the class I changed is LESModel and It's use by all LESModel, so i don't puzzle out how to extract only this class, I think I have to change all the Model LESclass like Smagorinsky and other... I should say, I am new in the openFOAM world, and particulary in the C++ world.

So I done the first way, and it's wonderful, no more duplicate status, swak4foam work well!

Thank you very must guy!

:)
See you next time on CFD-online!