redistributePar utility does not write pointDisplacement field

redistributePar utility when used to decompose , does not write pointDisplacement field .
testcase : tutorials/multiphase/interDyMFoam/ras/DTCHull

Reassigned to @Mattijs

Ditto at OpenFOAM.org: http://bugs.openfoam.org/view.php?id=2086

By Bruno Santos on 2016-05-14T12:35:40 (imported from GitLab project)

## Reattaching the author to the issue ticket: @wyldckat ##

Related to #62 (closed)

By Bruno Santos on 2016-05-14T12:36:47 (imported from GitLab project)

I've had a look at this.

• implemented pointPatch reordering and trimming (needed in fvMeshDistribute). Needs testing.
• need to test pointFields sending and receiving
• need to test subsetting of pointFields
• need to implement processorPointPatch addition

this is still open

The big problem is the addition and deletion of processor boundaries. The pointMesh is 'owned' by the polyMesh and consequently gets deleted whenever something changes in the polyMesh. This invalidates any pointFields (which are being held on the polyMesh - the pointMesh is not an objectRegistry). This leads to a whole lot of dangling pointers problems.

Possible solutions:

1. have pointMesh inside redistributePar (so not a mesh object). Feed pointMesh as argument into all fvMeshDistribute.
2. store pointFields on pointMesh, not polyMesh

Checked with 1812, the reported issue is present.

present in v2106. checked with tutorials/multiphase/interFoam/RAS/DTCHull.

changed the description

added feature request label

Running tutorials/incompressible/pimpleFoam/laminar/movingCone with

mpirun -np 2 redistributePar -parallel -reconstruct

• seems to work with pointFields
• but does not show any message as in Reconstructing volVectorFields @mark

will turn verbose_ on to make people happy (and not needing to write -verbose).

Time = 0

    Reconstructing volScalarFields

        alpha.water
        k
        nut
        omega
        p_rgh

    Reconstructing volVectorFields

        U

    Reconstructing pointVectorFields

        pointDisplacement

End

Finalising parallel run

mentioned in commit b51e0d4c

@Mattijs - still open/valid?

closed