Install OpenFOAM v1906 on a cluster with CentOS/RHEL 7
 

Hey guys


I am trying to install the openFOAM on the university cluster and i have been following this link for reference.



https://www.openfoam.com/download/install-source.php


and every thing seems right until i run the command: foamInstallationTest and this is the error message displayed saying 'icoFoam not installed: (i apologize for the bad formatting of the post, it is actually my first post)

I am not sure what causes the issue and what should i be doing?

Thank you very much in advance!!!!:):):):)

Quick question: In that page you provided, it gives a link to the Build page: https://www.openfoam.com/code/build-guide.php - did you follow the instructions given on that page?

Hello Bruno
Thanks for replying.
Yes i follow everything on the build page and when i try the command 'foamSystemCheck' it says the following:


And when i try to source the bashrc file, the following is shown:


But, after this i can use the 'foam' command to go to the OpenFoam installation directory.


So i am not sure what is the cause of the problem

Thank you very much !!!

Quick answers:

1. The command foamSystemCheck will only work after you activated the OpenFOAM shell environment, namely after you source the "bashrc" file.
2. The message:
   
   Code:

   ---

   bash: mpicc: command not found...

   ---

   is because there is no Open-MPI ready to be used... check if you have any module with Open-MPI ready to be used by running:
   
   Code:

   ---

   module avail

   ---

   I expect it's something like "openmpi-x86_64" or "mpi/openmpi-x86_64" that is available in your system. If you do have it, you can load it by running, for example:
   
   Code:

   ---

   module load mpi/openmpi-x86_64

   ---

   then source OpenFOAM's "bashrc" file once again.
3. The message:
   
   Code:

   ---

   No completion added for ...

   ---

   is being shown because OpenFOAM-v9106 hasn't been compiled yet.
4. Then to compile OpenFOAM-v1906, use the following commands:
   
   Code:

   ---

   ./Allwmake -j 4 -log=log.make

   ---

   where 4 is the number of cores to use to build/compile OpenFOAM.
   The output is sent to the file "log.make" which you can check its contents on another terminal, e.g. by running:
   
   Code:

   ---

   tail -F log.make

   ---

   This log file is critical for diagnosing any compilation problems.

Hello Bruno



Thank you for your reply !



When i tried your suggestion, i found out that my university's cluster does not have 'openmpi-x86_64' available, so i loaded a one of existing openmpi which is 'openmpi/1.10.7-mlx'



then i tried to compile the openFOAM with ./Allwmake and i get the following warning:




By the looks of error message, i am not sure if it is to do with the openmpi that i did not use the right one or something else i am not doing correctly. So the Openmpis that is available in the cluster is as follows:




And i am so sorry if i am asking very obvious questions, but i have zero experience in doing jobs in cluster previously.


Thank your very much for your patience! :):)

Quick answer: The commands you should execute are as follows:
(I'm assuming the path on the second line works in your installation, but if not, adjust accordingly)
Then check what error messages you get. If you only get the error about "no completion added", then Open-MPI should be ready to be used with OpenFOAM.

If so far so good, then run:

However, if sourcing OpenFOAM's "etc/bashrc" file gave the error about mpicc or some other error, then run:
and tell us what it gives you.

Hi Bruno

Thanks for replying!

When i load the openmpi and source the bashrc, there was nothing happening in the command window so i waited for a hour or two and found out there was still nothing. and i can not type any command either except ctrl+C to terminate the process.


I am not sure if it is suppose to take that long to source the bashrc or anything has gone wrong again?


But the following is the what i get when i call the $MPI_ARCH_PATH

Regards

Hi Bruno



It seems that it was the server to have a technical issue from my last post to not respond.


And this is what i get when i run the process again when the cluster is normal again.

So when i try to load the openmpi-1.10.7 i give the following message:


So i changed to another openmpi (openmpi/1.10.3-gcc5). then i source the bashrc and get the following:


Any suggesting on where should i proceed next?

Thank you very much!! :):):)



Regards