Trying to couple an in-house potential solver with OpenFOAM. The potential code has been modified using preCICE Fortran interface. The simulation runs fine at beginning till half-way then throws an error message and exits. I have attached the log files for both solvers and the precice configure file. Any comments on possible cause of this error are mostly appreciated! log.HOS.zip (626.7 KB) hosCoupleFoam.zip (2.0 MB) precice-config.xml (3.7 KB) preciceDict.txt (4.9 KB)
*---[precice] ERROR: Data etaVOF on mesh OpenFOAM-Coupling-Mesh didn't contain any data samples while attempting to map to mesh HOS-Coupling-Mesh. Check your exchange tags to ensure your coupling scheme exchanges the data or the pariticipant produces it using an action. The expected exchange tag should look like this: <exchange data="etaVOF" mesh="OpenFOAM-Coupling-Mesh" from=... to=... />.*
*[fdgs00-ws:94894] *** Process received signal ****
*[fdgs00-ws:94894] Signal: Aborted (6)*
*[fdgs00-ws:94894] Signal code: (-6)*
*[fdgs00-ws:94894] [ 0] /lib64/libpthread.so.0(+0xf630)[0x7f6be2ee5630]*
*[fdgs00-ws:94894] [ 1] /lib64/libc.so.6(gsignal+0x37)[0x7f6bc7f81387]*
*[fdgs00-ws:94894] [ 2] /lib64/libc.so.6(abort+0x148)[0x7f6bc7f82a78]*
*[fdgs00-ws:94894] [ 3] /scratcha-fdgs/lux/OpenFOAM/spack/opt/spack/linux-centos7-cascadelake/gcc-10.2.1/precice-develop-gkhj76ntcv2mbgqrj3555kw4nrzz6fhp/lib64/libprecice.so.3(precicef_advance_+0xe6)[0x7f6bded8e297]*
*[fdgs00-ws:94894] [ 4] HOS-ocean[0x49b876]*
*[fdgs00-ws:94894] [ 5] HOS-ocean[0x4acfa8]*
*[fdgs00-ws:94894] [ 6] HOS-ocean[0x409f93]*
*[fdgs00-ws:94894] *** End of error message ****
One thing I noticed is that the simulation always quits at the same time-window (6836th), no matter what timestep size I use. So far I have tried 0.01, 0.02 and 0.05. All got the same error.
Furthermore, I’ve tried parallel-explicit and serial-explicit, which gave me same error at the same time-window.
I also checked the results and everything seems normal at time-window 6835.
Here’s the visualization of the coupling configuration.
Both participants have two meshes (relaxation-mesh and coupling-mesh), and they are mapping with the corresponding meshes from the other participant as follows:
HOS-Relaxation-Mesh <=> OpenFOAM-Relaxation-Mesh
HOS-Coupling-Mesh <=> OpenFOAM-Coupling-Mesh
The Relaxation-Mesh is a volume mesh in 3D and a surface mesh in 2D, and the Coupling-Mesh is a surface mesh in 3D and a single line in 2D.
@Makis@fsimonis@ajaust@uekerman, I apologize for taking the liberty to tag you here, but I have been struggling with a particular issue for quite some time, and the development is urgent for project delivery.
Your expertise and guidance would be greatly appreciated in resolving this matter. Please let me know if you need any additional information or details regarding the development.
Thank you so much for your time and support. I truly appreciate your help!
Thank you for your reply. Although switching the mapping between participants does not seem helpful in the current situation, I am still curious about the general rule of thumb for determining the following:
The order of participants (first and second).
The impact of shifting the mapping operation from one participant to another and consequently changing a “read” operation to a “write” operation.
Just now tried to uninstall the precice and then re-install, here’s the output of Spack. Not sure whether Spack manages to get the latest version source code of precice.
Would it be possible to install Git on this system (in case this is missing)? The no-info [git failed to run] should be a string with a hash.
Not sure whether Spack manages to get the latest version source code of precice.
I think it does not take the latest state of the branch, but uses a cached version. You can always try explicitly installing the latest commit (right now this is 31e7a9e) instead of the branch.
This uncertainty with the state of Git branches is often causing me some stress when trying to ensure everything is up-to-date.
How long did it take you to run the simulation that previously crashed?
[lux@fdgs00-ws openfoam-adapter-HOS]$ git --version
git version 1.8.3.1
No idea why got the [git failed to run] warning. I am not quite familiar with the git and Spack stuff. Could you please advise how to explicitly install the latest commit in Spack?
Quick update, I tried parallel-implicit as well and got the same error at the same time-window.
I am a bit confused on how to configure the “mesh” entries in the coupling-scheme section. How should I determine which mesh to use? I found sometimes the “didn’t contain any data samples” error occurs at the very first step when I put wrong meshes.
Update. On the new workstation (precice was not installed previously), install selected commit in spack failed. But precice@develop seems to take the latest state of the branch (precice-develop-myznc4oooqcly7tonqgmxedhqf7lx4bh). On the old machine, is there a way to clean-up the cached source code in spack and let it download the fresh state from the git repository?
Just managed to compile the latest develop branch from source on the system. Now running, will update soon whether it passes through the crashing point.
