GenerateGlobalId filter crashes in parallel
We have a crash in the GenerateGlobalId filter. Here is how to replicate:
- 64 remote servers (sorry Utkarsh), Linux, master (v5.7.0-RC2-359-g5c062439). Note that 32 remote servers does NOT show this bug.
- Load the Watney Exodus dataset. All vars on. Apply.
- GenerateGlobalIds.
Server side crash.
You get lots of:
pvserver: ////snip///VTK/ThirdParty/diy2/vtkdiy2/include/vtkdiy2/detail/algorithms/kdtree-sampling.hpp:429: float vtkdiy2::detail::KDTreeSamplingPartition<Block, Point>::find_split(const Bounds&, const Bounds&) const [with Block = {anonymous}::PointBlockT; Point = {anonymous}::PointTT; vtkdiy2::detail::KDTreeSamplingPartition<Block, Point>::Bounds = vtkdiy2::Bounds<float>]: Assertion `0' failed.
Loguru caught a signal: SIGABRT
This is a Watney bug.
Edited by Mathieu Westphal (Kitware)