[Bug] Some problems of Metis partition

[Ron-Wang]

With an external mesh file (.msh file), I use mpirun Geosx. However, it say there in rank 1 doesn't have some regions. I think it's because the mesh has some many regions (26 regions), and some small regions are not in all partitions (ranks).
I wonder How can I solve this problem? Thanks a lot!

[AntoineMazuyer]

Hi @Ron-Wang, can you please communicate the logs of your simulation ? Thanks !

[Ron-Wang]

Thanks! @AntoineMazuyer The logs of my simulation are as follows: " ***** ERROR ***** LOCATION: /home/jp/GEOSX/GEOSX/src/coreComponents/mesh/CellElementRegion.cpp:51 ***** Controlling expression (should be false): source == nullptr ***** Rank 11: Cell block named Region1_TETRA does not exist

** StackTrace of 5 frames ** Frame 0: geosx::ElementRegionManager::GenerateMesh(geosx::dataRepository::Group*) Frame 1: geosx::ProblemManager::GenerateMesh() Frame 2: geosx::ProblemManager::ProblemSetup() Frame 3: main Frame 4: __libc_start_main Frame 5: _start

" I use the tetra element type, and I think some small regions are not in all partitions (ranks) when mpirun. For example, the partition of "Rank 11" doesn't have the "Region1_TETRA".

[wrtobin]

Please provide any input files if possible.

Documenting that this may be related to #841 and if/when that is resolved this should be checked.

[TotoGaz]

The msh format is no more supported since we no more use PAMELA but vtk instead.