CS & Syrthes Coupling error in CAELinux

This forum is dedicated to Syrthes related issues, as the Syrthes tool does not have its own forum.
Pino
Posts: 7
Joined: Wed Oct 23, 2019 1:53 pm

CS & Syrthes Coupling error in CAELinux

Post by Pino » Wed Oct 23, 2019 3:01 pm

Good afternoon all,

I'm trying to coupling CS & Syrthes.
In particular, I'm using Salome_CFD Vers. 8.5.0 containing Code_Saturne 5.0.4 and Syrthes 4.3.0.
I tried to follow the example "3disks2D", but without success; I tried to create another simple case, always getting the same answer from the terminal; in particular an error in preprocessing stage, as shown in this text file:
Coupling_text_in_terminal.odt
(18.69 KiB) Downloaded 3 times
here attached the other files that might be useful:
preprocessor.log
(5.73 KiB) Downloaded 2 times
compile.log
(698 Bytes) Downloaded 3 times

Thank you all for your support.

Yvan Fournier
Posts: 2774
Joined: Mon Feb 20, 2012 3:25 pm

Re: CS & Syrthes Coupling error in CAELinux

Post by Yvan Fournier » Wed Oct 23, 2019 10:48 pm

Hello,

I have no hands-on experience on the CAELinux distribution, but it seems the install is missing a path information.

As a workaround, can you run the computation from "salome shell" or at least "salome context" in a terminal ?

Regards,

Yvan

Pino
Posts: 7
Joined: Wed Oct 23, 2019 1:53 pm

Re: CS & Syrthes Coupling error in CAELinux

Post by Pino » Thu Oct 24, 2019 2:50 pm

Hi Yvan,
thanks for your help, I have just tried to run the computation from "salome shell" and "salome context" in a terminal.
Now I have some different errors (the old error I think it is solved!).
Please find attached the two text terminal results.
Terminal_text_salome_shell.odt
(19.2 KiB) Downloaded 3 times
Terminal_text_salome_context.odt
(19.28 KiB) Downloaded 3 times
Thanks again for your help.

Best Regards.

Pino

Yvan Fournier
Posts: 2774
Joined: Mon Feb 20, 2012 3:25 pm

Re: CS & Syrthes Coupling error in CAELinux

Post by Yvan Fournier » Thu Oct 24, 2019 5:40 pm

Hello,

Do you have any error message in RESU_COUPLING ?

The warning message relative to using a different MPI version (libmpi.so.1 or libmpi.so.12) may be quite bad... it seems to indicate the MPI available on the machine and the one used in SALOME may be different.

Can you try

Code: Select all

find /home/gabriele/salome-cfd/Salome_CFD-SCFD_V8_5_0-Code_Saturne-5.2-univ_d54f597c -name libmpi.so.1

to check if an MPI library is contained within the SALOME-CFD version (I think in the "universal" version it should be), and then if a path is found, add it to your LD_LIBRARY_PATH before running the case ?

If this is not the case or does not work, it would mean you  may need to recompile code_saturne and Syrthes so as to have the correct "system" MPI library...

Best regards,

	Yvan

Pino
Posts: 7
Joined: Wed Oct 23, 2019 1:53 pm

Re: CS & Syrthes Coupling error in CAELinux

Post by Pino » Fri Oct 25, 2019 9:09 am

Hello Yvan,

thanks again for your answer, I think I'll need to recomplile all, because I don't found anything (both the library and error message).
Next week I won't be able to work on it, but in the next one I'll try to fix it. I'll let you know as soon as possible.
Best regards.

Pino

Yvan Fournier
Posts: 2774
Joined: Mon Feb 20, 2012 3:25 pm

Re: CS & Syrthes Coupling error in CAELinux

Post by Yvan Fournier » Sun Oct 27, 2019 2:51 pm

Hello,

If you recompile all, I recommend switching to version 6.0. You also have a more up to date/easier to install Syrthes sources in the "download/prerequisites" section of the code_saturne web site: https://www.code-saturne.org/cms/sites/ ... _12.tar.gz if you are using Python3 and Qt5 (for Python2 and Qt4, use the older version; Python3/Qt4 and Python2/Qt5 combinations are not handled for Syrthes).

Best regards,

Yvan

Pino
Posts: 7
Joined: Wed Oct 23, 2019 1:53 pm

Re: CS & Syrthes Coupling error in CAELinux

Post by Pino » Thu Oct 31, 2019 3:27 pm

Hello Yvan,

I decided to switch to Ubuntu 18.04.3 LTS, I just installed CS 6.0 successfully, now I'm installing Syrthes 4.3.5 downloaded from the link you suggested.
I installed the prerequisites, filled out the setup.ini file below, and started the python installation command.
setup.ini
(4.32 KiB) Downloaded 3 times
It gives me a series of errors that I cannot understand.
Attached you will find the .log files.
ERREUR.log
(648 Bytes) Downloaded 3 times
file_log.zip
(2.48 KiB) Downloaded 1 time
Thank you so much for your help.

Best regards.
Pino

Yvan Fournier
Posts: 2774
Joined: Mon Feb 20, 2012 3:25 pm

Re: CS & Syrthes Coupling error in CAELinux

Post by Yvan Fournier » Thu Oct 31, 2019 6:29 pm

Hello,

I won't be able to check this before next Monday, so in the meantime I can only recommend double checking the paths and other settings (I see no obvious error in the setup.ini).

There might also be secondary logs to check (answering from my phone, not my computer, so I did not grep the zip contents).

We should also move this thread to the installation section...

Best regards,

Yvan

Pino
Posts: 7
Joined: Wed Oct 23, 2019 1:53 pm

Re: CS & Syrthes Coupling error in CAELinux

Post by Pino » Fri Nov 08, 2019 5:55 pm

Hi Yvan,

Did you have time to open the zip file and check if I made any mistakes?

Having a dual boot I reinstalled Caelinux and Salome_cfd in the other partition, but I still have problems.
In particular I followed the 3Disks2D example, but the analysis arrives at step 120 and generates error:
listing_fluid.txt
(3.15 MiB) Downloaded 3 times
listing_solid.txt
(328.17 KiB) Downloaded 2 times
So I decided to try another model, and this time the analysis returns another error:
listing2_fluid.txt
(17.42 KiB) Downloaded 2 times
and the last two line in listing file are:

CFD COUPLING INITIALIZATIONS
============================
coupling 0: initializing MPI communication ...



while in the terminal these writings appear:

**********************
Starting calculation
**********************

[gabriele-Lenovo-Z70-80:20386] *** An error occurred in MPI_Intercomm_create
[gabriele-Lenovo-Z70-80:20386] *** reported by process [1542193153,0]
[gabriele-Lenovo-Z70-80:20386] *** on communicator MPI COMMUNICATOR 3 SPLIT FROM 0
[gabriele-Lenovo-Z70-80:20386] *** MPI_ERR_INTERN: internal error
[gabriele-Lenovo-Z70-80:20386] *** MPI_ERRORS_ARE_FATAL (processes in this communicator will now abort,
[gabriele-Lenovo-Z70-80:20386] *** and potentially your MPI job)
--------------------------------------------------------------------------
MPI_ABORT was invoked on rank 0 in communicator MPI COMMUNICATOR 3 SPLIT FROM 0
with errorcode 1.

NOTE: invoking MPI_ABORT causes Open MPI to kill all MPI processes.
You may or may not see output from other processes, depending on
exactly when Open MPI kills them.
--------------------------------------------------------------------------
-------------------------------------------------------
Primary job terminated normally, but 1 process returned
a non-zero exit code.. Per user-direction, the job has been aborted.
-------------------------------------------------------
--------------------------------------------------------------------------
mpiexec.openmpi detected that one or more processes exited with non-zero status, thus causing
the job to be terminated. The first process to do so was:

Process name: [[23532,1],0]
Exit code: 17
--------------------------------------------------------------------------
solver script exited with status 17.

Error running the coupled calculation.

Either Code_Saturne or SYRTHES may have failed.

Check Code_Saturne log (listing) and SYRTHES log (syrthes.log)
for details, as well as error* files.



Thanks for your time.

Pino

Yvan Fournier
Posts: 2774
Joined: Mon Feb 20, 2012 3:25 pm

Re: CS & Syrthes Coupling error in CAELinux

Post by Yvan Fournier » Sun Nov 10, 2019 9:25 pm

Hello,

In the first case, your computation seems to run OK up to iteration 600 (the last one) but not stop correctly.
Normally, either code_saturne or Syrthes should be able to stop the computation cleanly (whichever stops first), but it seems this is not the case here.

If you set the number of iterations to the same number (600) on the Syrthes side, how does it go ?

With the other model, the error might be a mesh mismatch. For this, the Syrthes listing would help also, as the unknown status flag seems strange (Syrthes probably did not reach that step).

Note: I moved this topic to the "Syrthes related questions" forum section, and will remove the "shadow" topic from the general section soon.

Best regards,

Yvan

Post Reply