[Paraview] parallel file strangeness.

[David Thompson]

> I am having some problems with paraview when I read in .pvti files.
> First off, I have a structured points .vtk file that I read in with
> paraview. I then save the file data to local disk on the cluster as a
> .vti file ( the other option is a .pvd file and I get the same
> behavior with that). I use one ghost level when I save. The read of
> the .vti file goes fine. I have problems when I try to do a
> celldatatopoint data filter followed by a coutour filter. I get an
> error from process 0 that says The UpdateExtent cannot be filled. Then
> the client cant talk to the server any more and the game is over. Now
> I can do all of this if I read in the .vtk file instead of the .vti
> file. Everything works great in that situation. I do have to wait a
> bit longer for the read though. So I am wondering if others have had
> similar experience. It is not a problem with file size as the failure
> is repeatable across several file sizes. Anyone got any ideas? By the
> way the threshold filter works just fine regardless of how I read the
> data.
It sounds to me like you don't have enough ghost levels...
CellDataToPointData uses one ghost level and isocontouring requires
another.