Sai Kishor Kothakota

@firesurfer thanks for confirming it. Looking at the code last night, I think I might know the root cause of the issue. In 1 hour, I can send you another...

Hello @firesurfer, Could you test this branch: https://github.com/saikishor/ros2_control/tree/fix/iron/missing/resources_lock and let me know if this fixes your issue. Ideally, it should as the issue might be coming from some mutexes

Lol, so now it works once in every 2 runs. Right before this segmentation fault, do you have more logs on activation of the hardware components?

@firesurfer I've pushed another similar commit to the same branch, At least with this the read and write cycles shouldn't run unless the hardware components are loaded and initialized, I've...

then the problem might be coming from here https://github.com/ros-controls/ros2_control/blob/20e01d867fd14c1a3a0b8ea999d41d9d27dcf3bb/controller_manager/src/controller_manager.cpp#L339-L357 https://github.com/ros-controls/ros2_control/blob/20e01d867fd14c1a3a0b8ea999d41d9d27dcf3bb/hardware_interface/src/resource_manager.cpp#L335 This is executed right after the hardware components are loaded from the URDF method in the `load_urdf` method.

@firesurfer I've pushed another commit to the same branch. Can you test it please?

Awesome. I'm glad that the changes worked. I would then try to open a PR to have the fix upstream

@firesurfer If you have time, please review the changes in the above linked PR. Thank you

Hello! I just looked into this issue. It seems like, this is happening due to the the limit of the CPU cores. On my PC, the test runs for 33...

I've also tried to reduce the number of CPUs used by my docker container recommended by as @moriarty. I also see that the runtime of the tests is nearly double.