It shows [ros2run]: Segmentation fault when running in ROS2 in local environment

[shavargo109]

Describe the bug Following the 'Installation' and 'Getting started' guide, everything installed OK. When running "ros2 run glim_ros glim_rosnode", it shows[ros2run]: Segmentation fault, and cannot proceed afterward

To Reproduce Steps to reproduce the behavior:

1. ros2 run glim_ros glim_rosnode in the workspace

Expected behavior works normally followed the guide

Screenshots There is some problem on uploading screenshot, I will post the log in terminal instead

ros2 run glim_ros glim_rosnode [2024-12-11 17:51:58.173] [glim] [info] config_path: /home/asd/glim_ws/install/glim/share/glim/config [2024-12-11 17:51:58.173] [glim] [info] load libodometry_estimation_gpu.so [2024-12-11 17:51:58.353] [glim] [info] load libsub_mapping.so [2024-12-11 17:51:58.354] [glim] [info] load libglobal_mapping.so [2024-12-11 17:51:58.354] [glim] [info] load libmemory_monitor.so [2024-12-11 17:51:58.355] [glim] [info] load libstandard_viewer.so [2024-12-11 17:51:58.358] [glim] [info] load librviz_viewer.so [ros2run]: Segmentation fault

Environment (please complete the following information):

• CPU/GPU: Core i9-13900HX/RTX4070
• OS: Ubuntu 22.04
• ROS: ROS2 humble
• CUDA: CUDA 12.3
• Build with CUDA: OFF

Additional context Thanks

[wsung1]

I’m experiencing the same issue.
After completing a normal installation and using it without problems for a while, this issue suddenly occurred one day, and I still don’t know how to resolve it.

[wsung1]

A bit different error I ran into was terminate called after throwing an instance of 'thrust::system::system_error' what(): This program was not compiled for SM 75
: cudaErrorInvalidDevice: invalid device ordinal

My environment is CPU/GPU: Core i7-9700T/GTX 1650 OS: Ubuntu 22.04 ROS: Humble CUDA: CUDA 12.3 Build with CUDA: ON

[Efesendil]

Hello, I have resolved this issue by installing a recent CUDA version, for me 12.6 works without any problem.

[shavargo109]

Hello, I have resolved this issue by installing a recent CUDA version, for me 12.6 works without any problem.

Installed 12.6 version but still run into the same issue, do you have any other potential solution? Thanks

[shavargo109]

I tried running gdb on glim_rosnode and found this issue Thread 18 "glim_rosnode" received signal SIGSEGV, Segmentation fault. [Switching to Thread 0x7fffb9fff000 (LWP 42506)] 0x00007fffbab9d12b in std::vector<std::filesystem::__cxx11::path::_Cmpt, std::allocatorstd::filesystem::__cxx11::path::_Cmpt >::~vector() () from /usr/local/lib/libiridescence.so

it maybe something due to iridescence.so, but this file is generated by the installation guideline, I'm afraid it would be a bit difficult for me to solve this