ParaView 6.2.0 complains about not having any points for plot over line on an MPI distributed dataset

In Situ Support
1

Hello,

I am getting these warning since I updated from ParaView 6.2.0 RC1 to 6.2.0:

(   1.403s) [pvbatch.0       ]  vtkPolyDataReader.cxx:267   WARN| vtkPolyDataReader (0x7cd01787c980): No points read!
(   1.403s) [pvbatch.0       ]  vtkPolyDataReader.cxx:267   WARN| vtkPolyDataReader (0x7cd01787ae80): No points read!
(   1.404s) [pvbatch.0       ]  vtkPolyDataReader.cxx:267   WARN| vtkPolyDataReader (0x7cd01787ba80): No points read!
(   1.404s) [pvbatch.0       ]  vtkPolyDataReader.cxx:267   WARN| vtkPolyDataReader (0x7cd017884480): No points read!
(   1.404s) [pvbatch.0       ]  vtkPolyDataReader.cxx:267   WARN| vtkPolyDataReader (0x7cd017882f80): No points read!
(   1.404s) [pvbatch.0       ]  vtkPolyDataReader.cxx:267   WARN| vtkPolyDataReader (0x7cd017880b80): No points read!
(   1.452s) [pvbatch.0       ]  vtkPolyDataReader.cxx:267   WARN| vtkPolyDataReader (0x7cd017cb1080): No points read!
(   1.452s) [pvbatch.0       ]  vtkPolyDataReader.cxx:267   WARN| vtkPolyDataReader (0x7cd017cb2580): No points read!
(   1.452s) [pvbatch.0       ]  vtkPolyDataReader.cxx:267   WARN| vtkPolyDataReader (0x7cd017cb3a80): No points read!
(   1.453s) [pvbatch.0       ]  vtkPolyDataReader.cxx:267   WARN| vtkPolyDataReader (0x7cd017cb4f80): No points read!
(   1.453s) [pvbatch.0       ]  vtkPolyDataReader.cxx:267   WARN| vtkPolyDataReader (0x7cd017cb6480): No points read!
(   1.453s) [pvbatch.0       ]  vtkPolyDataReader.cxx:267   WARN| vtkPolyDataReader (0x7cd017cb7980): No points read!
(   1.453s) [pvbatch.0       ]  vtkPolyDataReader.cxx:267   WARN| vtkPolyDataReader (0x7cd017cb8e80): No points read!

These warning are thrown when using the PlotOverLine filter on an MPI distributed dataset (CFD code + ParaView Catalyst).

I suspect this is thrown by the filter on ranks that do not have data that intersects with the line.

This may not seem like a big deal, but in a ParaView Catalyst context, we are talking about hundreds of ranks throwing warnings to the screen log everytime Catalyst is called, which is potentially often. This renders the screen log almost unreadable.

If you confirm what I am saying, then I would consider this a regression and would appreciate a quick resolution. I do not think we can use ParaView 6.2 in this state.

Many thanks,
Alexandre

2

Is that catalyst or can this be reproduced by just opening a specific file and adding PlotOverLine ?
If that is so, please share that file.

3

I have not tried to reproduce it in ParaView, as I am not well versed in creating distributed datasets in ParaView.

4

@Alexandre_Minot is it convenient for you to try reverting https://gitlab.kitware.com/vtk/vtk/-/commit/a78669d689337d262ceac319a36878c8dedc3668 locally and see if the warning messages disappear?

By the way, IMO reading zero points is fine and should not produce a warning in any case, so we should remove that message.

6

Yes, reverting to the following commit removes the warning:

VTK$ git status
HEAD detached at 848cd66c88
nothing to commit, working tree clean
7

Note that the released code seems fine functionally, it’s just the warnings that are an issue.

8

What is the path forward? Can I expect a 6.2.1 soon? Thanks

9

Look for a ParaView 6.2.1 in early November. We’ll address the warnings by then.

10

Great, thanks

11

Do you need me to open an issue?

12

No, I’ve got it on my todo list. Thanks for the offer, though.