I am trying to use nvblox and I started following this guide.
I am using docker and I have run run_dev.sh.
I have used both installation from Debian and from source but if I run:
Thank you for your post. Looking at your output, I can see that nvblox is expanding the memory on your computer, and after you sent CTRL-C to the terminal and switched off the launch file.
[component_container_isolated-1] I0514 10:25:42.962843 2347 mapper.cpp:304] Allocating space for last depth image
[component_container_isolated-1] I0514 10:25:42.981626 2347 image_cache_impl.h:45] Allocating a new cached image
[component_container_isolated-1] I0514 10:25:44.441819 2347 block_memory_pool_impl.h:65] Expanding the memory pool with 4096 blocks. Number of allocated blocks: 6144
[component_container_isolated-1] I0514 10:25:44.442116 2347 gpu_layer_view_impl.cuh:100] Resizing GPU hash capacity from 4096 to 8192 in order to accomodate space for 76 new elements.
[component_container_isolated-1] I0514 10:25:44.442312 2347 gpu_hash_interface_impl.cuh:83] Creating a GPUHashImpl with requested capacity of 8192 blocks. Real capacity: 8192
[component_container_isolated-1] I0514 10:25:44.694808 2347 block_memory_pool_impl.h:65] Expanding the memory pool with 4096 blocks. Number of allocated blocks: 6144
[component_container_isolated-1] I0514 10:25:44.695183 2347 gpu_layer_view_impl.cuh:100] Resizing GPU hash capacity from 4096 to 8192 in order to accomodate space for 341 new elements.
[component_container_isolated-1] I0514 10:25:44.695341 2347 gpu_hash_interface_impl.cuh:83] Creating a GPUHashImpl with requested capacity of 8192 blocks. Real capacity: 8192
[component_container_isolated-1] I0514 10:25:58.492200 2347 gpu_layer_view_impl.cuh:100] Resizing GPU hash capacity from 8192 to 16384 in order to accomodate space for 30 new elements.
[component_container_isolated-1] I0514 10:25:58.492799 2347 gpu_hash_interface_impl.cuh:83] Creating a GPUHashImpl with requested capacity of 16384 blocks. Real capacity: 16384
[INFO] [ros2-2]: process has finished cleanly [pid 2259]
[WARNING] [launch]: user interrupted with ctrl-c (SIGINT)
From the terminal, I don’t read any other crash reports from the system.
Try to give more time to the execution to load the NVBLOX properly into the GPU memory.
Thank you for your reply. Yes, there is no crash it seems, apart from rviz2 I think.
I have attached another file with no CTRL+C. The only problem is rviz2 then.
Then I opened a second terminal (cd $ISAAC_ROS_WS && ./src/isaac_ros_common/scripts/run_dev.sh) and I run rviz2 and manually opened /opt/ros/humble/share/nvblox_examples_bringup/config/visualization/isaac_sim_example.rviz.
Then in the first terminal, I run the quickstart example:
Anyway I managed to record all ros2 topics and play the bag file to visualize them in rviz2 with /opt/ros/humble/share/nvblox_examples_bringup/config/visualization/isaac_sim_example.rviz configuration of rviz2.
So I think it is a sort of conflict between rviz2 and nodes launched with isaac_sim_example.launch.py
Thank you for collecting the log. The original issue at this post was that rviz2 got died while the demo was just started. But according to your log, the rviz2 looks running okay. The rviz2 was stoped after you terminated it by Ctrl-C. Can you elaborate a bit more on your issue? If it looks like a different issue, please open a new topic with detailed context.
Im sorry , i did not clarify the issue earlier.
The issue is that the demo gets stuck midway and does not complete the full 360 of the bot as shown in the example. Rviz does not crash , but the demo get stuck.
The rosbag file has been shortened( it’s about 3/4 of what is shown on the documentation ).
So your demo result should be as expected, which is not an issue, thank you.