Isaac Sim crashes when running ROS examples

[nathanshankar]

Isaac Sim Version

4.5.0
4.2.0
4.1.0
4.0.0
4.5.0
2023.1.1
2023.1.0-hotfix.1
Other (please specify):

Operating System

Ubuntu 22.04
Ubuntu 20.04
Windows 11
Windows 10
Other (please specify):

GPU Information

• Model: RTX 6000
• Driver Version: 535.183.01

Topic Description

Detailed Description

I created my own docker container with isaac sim 4.5.0 and ros2 humble. I have this running on an ML computer with 3 NVIDIA RTX 6000 GPUs, and I’m connect to this remote machine and use webrtc streaming client to see the isaac sim GUI on my base system. Whenever I run ROS2>Navigation>Carter it runs for a minute or two and then process memory overflows and crashes. I did not have this issue when I used the same container on my base machine.

Steps to Reproduce

1. Clone GitHub - nathanshankar/docker_isaac_ros2
2. Build and run the container on your remote machine
3. Run ./runheadless.sh on your remote machine and set up your web streaming client on your base machine
4. Run the robotics examples>ros2>navigation>Carter
5. run a ros2 node to teleop_twist_keyboard and move it around for a minute or two.

Error Messages

I don’t seem to be getting any errors, just that it’s Killed:
2025-04-04 11:20:35 [132,821ms] [Info] [rtx.rtxsensor.plugin] Initialized sensor model 0x79233964f190 on device 2
2025-04-04 11:20:35 [132,821ms] [Info] [rtx.rtxsensor.plugin] Using User-specified coordinate frame quaternion (0.000000, 0.000000, 0.000000, 1.000000) in RTXSensor
2025-04-04 11:20:35 [132,821ms] [Info] [carb.scenerenderer-rtx.plugin] RTXSensor Model produced RtxSensorCpu
2025-04-04 11:20:35 [132,821ms] [Info] [carb.scenerenderer-rtx.plugin] RTXSensor Model produced RtxSensorMetadata
2025-04-04 11:20:35 [133,070ms] [Info] [omni.kit.menu.utils.utils] omni.kit.menu.utils.refresh_menu_items Create
2025-04-04 11:20:35 [133,152ms] [Info] [omni.kit.menu.utils.utils] omni.kit.menu.utils.refresh_menu_items Edit
2025-04-04 11:20:35 [133,158ms] [Warning] [omni.timeline.plugin] Deprecated: direct use of ITimeline callbacks is deprecated. Use ITimeline::getTimeline (Python: omni.timeline.get_timeline_interface) instead.
2025-04-04 11:20:35 [133,178ms] [Info] [isaacsim.ros2.bridge.plugin] isaac_ros_nitros_bridge_interfaces NitrosBridgeImage message type not found. The NITROS bridge publisher was not created
2025-04-04 11:20:35 [133,291ms] [Info] [lidar.lidarprofilereader] Reader is not an array or member channelId not found
2025-04-04 11:20:35 [133,291ms] [Info] [lidar.lidarprofilereader] Reader is not an array or member rangeId not found
2025-04-04 11:20:35 [133,291ms] [Info] [lidar.lidarprofilereader] Reader is not an array or member vertOffsetM not found
2025-04-04 11:20:35 [133,291ms] [Info] [lidar.lidarprofilereader] Reader is not an array or member horOffsetM not found
2025-04-04 11:20:35 [133,291ms] [Info] [lidar.lidarprofilereader] Reader is not an array or member distanceCorrectionM not found
2025-04-04 11:20:35 [133,291ms] [Info] [lidar.lidarprofilereader] Reader is not an array or member focalDistM not found
2025-04-04 11:20:35 [133,291ms] [Info] [lidar.lidarprofilereader] Reader is not an array or member focalSlope not found
2025-04-04 11:20:35 [133,291ms] [Info] [lidar.lidarprofilereader] Reader is not an array or member reportRateDiv not found
2025-04-04 11:20:35 [133,291ms] [Info] [lidar.lidarprofilereader] Reader is not an array or member bank not found
2025-04-04 11:20:35 [133,291ms] [Info] [lidar.lidarprofilereader] Reader is not an array or member roi not found
2025-04-04 11:20:59 [156,844ms] [Warning] [omni.graph.core.plugin] /World/Nova_Carter_ROS/differential_drive/differential_controller_01: [/World/Nova_Carter_ROS/differential_drive] invalid dt 0.000000, cannot check for acceleration limits, skipping current step
2025-04-04 11:20:59 [156,845ms] [Info] [omni.usd.audio] resetting the animation timeline
Killed

Screenshots or Videos

Additional Information

What I’ve Tried

I’m not really sure where the issue seems to happen, since it works fine on my base system but doesn’t on a super powerful ML computer, I checked the forum but couldn’t find any available solutions.

[nathanshankar]

Here’s the simulation before I start it:

[VickNV]

Could you confirm whether the officially supported Docker container running on the remote ML system would reproduce the issue?

[nathanshankar]

Yes I get the same issue with the official supported Docker Container as well. It works fine when I run it on my base system, but crashes on the ML system.