cannot create flow.vtk files after parallel computation
 

I am trying to get solution output from an unsteady simulation that I have run using the mpirun command. In steady cases I can use SU2_SOL with the config file and get the flow.vtk file but the same process doesn't seem to work for the unsteady case. I am getting a segmentation fault error and no file is created.
Additionally, I realized that the python script "parallel_computation.py" isn't working. I think I may have installed the code incorrectly but I would really appreciate it if anyone could help me on this matter.

Here is the error I am getting from SU2_SOL:
------------------------- Solution Postprocessing -----------------------
[omdemircan:09268] *** Process received signal ***
[omdemircan:09268] Signal: Segmentation fault (11)
[omdemircan:09268] Signal code: Address not mapped (1)
[omdemircan:09268] Failing at address: (nil)
[omdemircan:09268] [ 0] /lib/x86_64-linux-gnu/libpthread.so.0(+0x11630)[0x7fb7df0a7630]
[omdemircan:09268] [ 1] SU2_SOL(+0x2e9bc)[0x561308ab19bc]
[omdemircan:09268] [ 2] /lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xf1)[0x7fb7decef3f1]
[omdemircan:09268] [ 3] SU2_SOL(+0xeb63a)[0x561308b6e63a]
[omdemircan:09268] *** End of error message ***
Segmentation fault (core dumped)

and the error when I try to sun parallel_computation.py:
Traceback (most recent call last):
File "/usr/local/bin/parallel_computation.py", line 112, in <module>
main()
File "/usr/local/bin/parallel_computation.py", line 63, in main
options.compute )
File "/usr/local/bin/parallel_computation.py", line 90, in parallel_computation
info = SU2.run.CFD(config)
File "/usr/local/bin/SU2/run/interface.py", line 89, in CFD
auto_diff = konfig.MATH_PROBLEM == 'DISCRETE_ADJOINT'
File "/usr/local/bin/SU2/io/config.py", line 130, in __getattr__
raise AttributeError , 'Config parameter not found'
AttributeError: Config parameter not found

did you remember to type .cfg after the config file name?

Like so:

python parallel_computation.py -f inv_ONERAM6.cfg -n 1

It's been some time but I remember the error was gone after I reinstalled SU2, I guess something went wrong during the installation.

Hi, thanks a lot for using SU2!

Not 100% sure but... unfortunately, I think that the unsteady SU2 output was never implemented for unsteady problems. In fact, the Paraview output doesn't have the maintenance that we would like (there is no a binary version!).

Have you tried TecPlot output? I think it is possible to have a free license for 1 month..

Anyway, could you please add this problem to the list of issues

https://github.com/su2code/SU2/issues

And, if you are interested in working on the Paraview output, you are more than welcome!

Thanks

Francisco

Hi,
The source of the segmentation error that I mentioned in this thread was a bad configuration during the intallation. I resolved this issue by reinstalling SU2 and the script ran without giving any errors.
However, getting the Paraview output from the solution data with the SU2_SOL command doesn't seem to be working as it seems to be written only for steady simulations. For solving this problem, I have found a work-around that gives me the Paraview output for each solution data that the code outputs. I have written a very simple python script and shared it in the following thread: https://www.cfd-online.com/Forums/su...t-working.html
I don't know if I am doing something wrong when calling SU2_SOL or is it actually as you said, that SU2 unsteady output was never implemented to the code.
Secondly, I didn't quite understand what you meant by trying Tecplot and I haven't used it with SU2 before. The output from the python script that I have written can be opened with Paraview like an unsteady case. I even used it to get pressure data at a point with respect to time/iteration. The problem here is that the solution files tend to be very big since they store all information about the unsteady case. So I think the user should be able to specify different intervals for saving flow data and the surface data.

Thanks for the clarification.
We would appreciate if you add this issue to

https://github.com/su2code/SU2/issues

including your python script

It is the best way to capture the attention of the unsteady solver developers. I know most of them use TecPlot http://www.tecplot.com which is an excellent post-processing tool... but, proprietary.

Thanks!
Francisco

I checked the link you provided and searched for a similar issue, it seems someone else pointed the same issue here: https://github.com/su2code/SU2/issues/133. I added my thoughts as you requested.

Thank you.
Osman Mirza Demircan