[arieljeds]

Hi all,

This is quite a simple question and I'm surprised this isn't working but I am running a large simulation on a cluster and then attempting to transfer the .OpenFOAM case to my local directory using sftp to view it on paraview.

I am running a case in parallel and then running reconstructPar -latestTime as I hope to view just the latest time step in paraview on my local system before removing all of the processor directories in the remote directory, deconstructing again and running again from the latest time.

I am using the command:

paraFoam -touch

While inside the directory, or else, I also tried:

touch case1.OpenFOAM

Both of these commands successfully created the case1.OpenFOAM file. However, when I transfer the file to my local system and try to open in paraview, only the mesh and geometry information is available and not the fields.

The fields are definitely being saved correctly in the time steps (and the last time step) and this has worked for me before so I am very confused what the problem is now.

Any help is greatly appreciated!

[arieljeds]

Anyone? I'm still having the same problem with this...

[alexeym]

Hi,

Not quite sure, I understand your message correctly but both commands (paraFoam -touch and touch case1.OpenFOAM) just create empty file named case1.OpenFOAM. It is necessary for paraview and just an indication that current folder is OpenFOAM case and corresponding reader should be used.

And do I get it right: you copy this empty file from cluster to your local file system?

[arieljeds]

Hi @alexeym,

Thanks for responding to this.

First of all, yes I am trying to create the file on the cluster and transfer it to my own system to view in paraview.

The problem that I'm having is that I do create the file case1.OpenFOAM, and I can then transfer this my own local system and view it in paraview. However, when I do this, none of the information from the fields is saved and I can only view the mesh/geometry. I thought this would work and then I could view the fields and the time step information, but these are not saved.

Does this explanation of the problem make a bit more sense?

Thanks,
Ariel

[alexeym]

Hi,

Let me ask one more clarifying question (cause still what your are doing looks quite surprising to me). From above posts I deduced this sequence:

1. You run simulation on cluster.
2. You run "reconstructPar -latestTime" to reconstruct fields for the latest time (for convenience lets denote it as 10)
3. Fields are reconstructed into folder 10.
4. You run "paraFoam -touch", it creates case1.OpenFOAM file (EMPTY file).
5. You copy case1.OpenFOAM to your local filesystem, and keep folder 10 on the cluster.
6. Locally you have case folder (guess you prepared case locally and then copied it to cluster).
7. You put case1.OpenFOAM into this folder and open in paraview.

Does the above describe what your are doing or I am mistaken?

[arieljeds]

Yes those are the steps I am taking.. And then I would like to view the case for, say, t = 10 in paraview on my local machine.

Note: I have not copied the directory for t = 10 to my local machine.

[alexeym]

Originally Posted by arieljeds

Note: I have not copied the directory for t = 10 to my local machine.

But you have to Folder 10 contains the data (fields for t=10, also mesh if you run dynamic mesh simulation). As I said case1.OpenFOAM is EMPTY file, it is created for paraview to know that current folder is OpenFOAM case and load OpenFOAM reader. And this reader will look for folder 10 (for t = 10), where real data is located.

[arieljeds]

As I started responding to your previous message, I began suspecting this was my very silly problem... I'm guessing I should also have the 0 folder (with all of the fields initialised here) as well?

Strangely I managed to get this to work before, so I'm not sure how I would have realised to do this before and forgotten now...


I will try this now, thank you for your help!!

[arieljeds]

Hi alexeym, as you probably expected, this works now. Thanks a lot!

[arieljeds]

Hi a@alexeym, I seem to have spoken to soon. It is working for the initial time step, but I ran a little test case and after doing reconstructPar -latestTime and then transferring the folder for t = 10 (and t = 0, and case1.OpenFOAM), I am getting the following error when I go to view the final time step in paraview:

--> FOAM FATAL IO ERROR: 
size 1873760 is not equal to the given value of 1870560

file: /media/Ariels2TB/OpenFOAM_Cases/cylinderCases/linearWaveCases/nonSlipTurb/case1/10/nut from line 18 to line 1930440.

    From function Field<Type>::Field(const word& keyword, const dictionary&, const label)
    in file lnInclude/Field.C at line 292.

FOAM exiting

Inconsistency detected by ld.so: dl-close.c: 759: _dl_close: Assertion `map->l_init_called' failed!

Any ideas what this is about?? I am guessing it's something to do with the 10/nut file not being written to or closed properly.. but I have no idea really.