The following are my current system and installed package versions:
Ubuntu 24.04.3 LTS (noble)
GeForce RTX 4080
Driver Version: 580.105.08
Cuda compilation tools, release 13.0, V13.0.88
Python 3.12.3ubuntu
Docker version 29.0.2, build 8108357
ROS2 jazzy
Package: ros-jazzy-moveit-core Version: 2.12.3-1noble.20251107.235048
NVIDIA Container Toolkit CLI version 1.18.1
NVIDIA Container Runtime version 1.18.1
runc version 1.3.3
Before the update, I was able to plan using Isaac Manipulator, but after the Isaac Manipulator update, I reinstalled isaac_ros_cumotion_moveit and now every Plan & Execute attempt returns Failed. I am unable to identify the cause. The video of the operation process is here. Could you please help me resolve this issue?
Could you run ros2 control list_controllers to check if joint_state_broadcaster,panda_hand_controller and panda_arm_controller are showing as active?
When franka.launch.py is executed, it should bring up all these needed nodes and continuously publish the /joint_states topic. Please verify this first.
Also, it would be helpful if you could upload the log so we can assist you better.
Thank you for your reply. This video shows my workflow when operating cuMotion, and this is the log file.
This log file contains the terminal output after running isaac_ros_cumotion.launch.py. The content includes the output returned from executing ros2 control list_controllers.
I originally upgraded directly from Ubuntu 22.04 to 24.04 and encountered issues. I have now reinstalled the system on a separate hard drive. This log file records the installation process after the system reinstallation, and I hope it will be helpful.
Thank you for providing the details.
I noticed that the error may be due to an incorrect command for launching cumotion_planner, which caused it to fall back to the default path and resulted in an extra “e” (as seen in “panda.urdfe” in your log).
Could you run the following command and share the output? ros2 param get /cumotion_planner urdf_path
Also, please try this command and let me know if the issue persists. ros2 run isaac_ros_cumotion cumotion_planner_node --ros-args -p robot:=$(ros2 pkg prefix --share isaac_ros_cumotion_robot_description)/xrdf/franka.xrdf -p urdf_path:=/opt/ros/humble/share/moveit_resources_panda_description/urdf/panda.urdf
Let me know the results so I can assist you further.
Thank you for your reply. I noticed the issue with the extra “e” at the end of the .urdf filename and have already fixed it. I apologize for posting a silly question earlier — the problem was caused by an incorrect link configuration on my side.
I have re-recorded the video and collected the relevant log files:
During runtime, the command outputs are as follows:
Output of ros2 param get /cumotion_planner urdf_path
String value is: /opt/ros/jazzy/share/moveit_resources_panda_description/urdf/panda.urdf
Output of ros2 control list_controllers
[INFO] [1765954179.326787209] [_ros2cli_5204]: waiting for service /controller_manager/list_controllers to become available...
[WARN] [1765954189.333163560] [_ros2cli_5204]: Could not contact service /controller_manager/list_controllers
[INFO] [1765954189.333729416] [_ros2cli_5204]: waiting for service /controller_manager/list_controllers to become available...
[WARN] [1765954199.339872331] [_ros2cli_5204]: Could not contact service /controller_manager/list_controllers
[INFO] [1765954199.340393282] [_ros2cli_5204]: waiting for service /controller_manager/list_controllers to become available...
[WARN] [1765954209.346250139] [_ros2cli_5204]: Could not contact service /controller_manager/list_controllers
[INFO] [1765954209.346776420] [_ros2cli_5204]: waiting for service /controller_manager/list_controllers to become available...
[WARN] [1765954219.353151875] [_ros2cli_5204]: Could not contact service /controller_manager/list_controllers
Based on my experience using cuMotion over the past year prior to the Manipulator update, I can’t figure out where the problem is coming from. I’m sorry for repeatedly bothering you about this issue.
I’ve also copied the entire robot description package for easier inspection. I’m not sure whether this will be helpful, but please let me know if there is any additional information you need, and I’ll be happy to provide it.
In theory, the built-in robot description is downloaded from NVIDIA. I followed the standard build process without making any other modifications, so intuitively, as long as the package can be found, it should be fine. I also checked the build logs from my own MoveIt build and didn’t notice anything unusual. I was able to run it before NVIDIA’s update, but after the update, this issue appeared, which I find quite confusing.
Thanks for providing me more information to investigate.
That’s a bit weird, it’s likely an issue with the Kinova and ros2_control integration on Jazzy, rather than the cuMotion itself.
It crashes inside hardware_interface::ResourceManager::load_and_initialize_components(...) with a segmentation fault.
Seem like the hardware interface plugin referenced by your Kinova URDF / ros2_control config is not properly available in the current environment.
From your logger_all.txt, your MoveIt + ros2_kortex workspace is under /home/nexcobot/ws_moveit
Could you source your Kinova workspace and jazzy with below:
source /opt/ros/jazzy/setup.bash
source /home/nexcobot/ws_moveit/install/setup.bash
# Then, if needed, the Isaac ROS workspace (inside the isaac_ros-dev container)
# source install/setup.bash
Then in that same terminal, verify that the Kinova / kortex packages are visible: ros2 pkg list | grep -E 'kortex|kinova_gen3'
And then run franka.launch.py and isaac_ros_cumotion.launch.pyin two separate terminals both after sourcing with above to see if it works.
If it’s still not working, it’s also possible that ros2_control_node is launched with the Kinova MoveIt config (ros2_controllers.yaml) but with namespace.
We need to distinguish “no controller manager at all” from “controller manager under a namespace”. After everything is launched, please run below command to check controller_manager:
ros2 service list | grep controller_manager
ros2 node list | grep controller_manager
If you do see something like /something/controller_manager/list_controllers, then try:
ros2 control list_controllers --controller-manager /something/controller_manager
I hope my understanding and operation are correct. I have included a video showing my workflow for reference.
I already have ROS 2 and MoveIt 2 setup.bash sourced in my ~/.bashrc. However, to be safe, I followed your suggestion and sourced them again before launching the container.
Running the command:
admin@nexcobot:/workspaces/isaac_ros-dev$ ros2 pkg list | grep -E 'kortex|kinova_gen3'
It seems that kinova_gen3_7dof_robotiq_2f_85_moveit_config can be found correctly.
Running:
admin@nexcobot:/workspaces/isaac_ros-dev$ ros2 service list | grep controller_manager
ros2 node list | grep controller_manager
returns:
/controller_manager/list_controllers
/moveit_simple_controller_manager/describe_parameters
/moveit_simple_controller_manager/get_parameter_types
/moveit_simple_controller_manager/get_parameters
/moveit_simple_controller_manager/get_type_description
/moveit_simple_controller_manager/list_parameters
/moveit_simple_controller_manager/set_parameters
/moveit_simple_controller_manager/set_parameters_atomically
WARNING: Be aware that there are nodes in the graph that share an exact name, which can have unintended side effects.
/moveit_simple_controller_manager
I assumed that /controller_manager/list_controllers is the one we are looking for, so I proceeded to run the next command.
Running:
admin@nexcobot:/workspaces/isaac_ros-dev$ ros2 control list_controllers --controller-manager /something/controller_manager
returns:
[INFO] [1766132210.028146821] [_ros2cli_4093]: waiting for service /something/controller_manager/list_controllers to become available...
[WARN] [1766132220.034137277] [_ros2cli_4093]: Could not contact service /something/controller_manager/list_controllers
[INFO] [1766132220.034565698] [_ros2cli_4093]: waiting for service /something/controller_manager/list_controllers to become available...
[WARN] [1766132230.040758929] [_ros2cli_4093]: Could not contact service /something/controller_manager/list_controllers
[INFO] [1766132230.041508650] [_ros2cli_4093]: waiting for service /something/controller_manager/list_controllers to become available...
[WARN] [1766132240.048378472] [_ros2cli_4093]: Could not contact service /something/controller_manager/list_controllers
[INFO] [1766132240.048894266] [_ros2cli_4093]: waiting for service /something/controller_manager/list_controllers to become available...
[WARN] [1766132250.055418153] [_ros2cli_4093]: Could not contact service /something/controller_manager/list_controllers
[INFO] [1766132250.056035357] [_ros2cli_4093]: waiting for service /something/controller_manager/list_controllers to become available...
[WARN] [1766132260.062587352] [_ros2cli_4093]: Could not contact service /something/controller_manager/list_controllers
[INFO] [1766132260.063079622] [_ros2cli_4093]: waiting for service /something/controller_manager/list_controllers to become available...
[WARN] [1766132270.069753707] [_ros2cli_4093]: Could not contact service /something/controller_manager/list_controllers
[INFO] [1766132270.070330405] [_ros2cli_4093]: waiting for service /something/controller_manager/list_controllers to become available...
[WARN] [1766132280.076787357] [_ros2cli_4093]: Could not contact service /something/controller_manager/list_controllers
[INFO] [1766132280.077449255] [_ros2cli_4093]: waiting for service /something/controller_manager/list_controllers to become available...
[WARN] [1766132290.083924867] [_ros2cli_4093]: Could not contact service /something/controller_manager/list_controllers
[INFO] [1766132290.084399324] [_ros2cli_4093]: waiting for service /something/controller_manager/list_controllers to become available...
I hope the above information helps identify the root cause of this issue.
When you ran the command ros2 service list | grep controller_manager ros2 node list | grep controller_manager, the first line of the return is already the one we expected **/controller_manager/list_controllers. **
So ros2 control list_controllers --controller-manager /something/controller_manager, /something/is not necessary unless your controller manager is under a different namespace. In that case, replace /something/ with the appropriate namespace you found from the commands above. Sorry for the confusion.
Thanks for the clarification.
I’ve already verified the URDF path and the ros2_controllers.yaml exists.
However, ros2_control_node is still crashing (segfault), and Plan & Execute always returns failed.
Could you please advise what I should check or try next to resolve the ros2_control_node crash?
Could you confirm if your ros-jazzy-controller-manager is also version 4.39.2? We have seen similar issues when these packages are not all on the same version recently. If any package is outdated, please use sudo apt install to update to the latest version.
Additionally, I have reported this issue to our internal team and will keep you updated if there are any developments.
It appears that the version is consistent with ros-jazzy-controller-interface, ros-jazzy-hardware-interface, and ros-jazzy-ros2-control.
I would like to ask: if this issue is handled internally, will it be released in the next update? Because we are going to exhibit at GTC, and if the new version cannot run properly, I will need to immediately switch back to using the older version of the Isaac Manipulator.
Currently, we cannot guarantee when a fix will be released, as our team is still investigating the issue. For the exhibition, you may want to consider preparing a backup plan with the older first, just in case. We’ll keep you updated on any progress.
Hi,
I am new to nvidia issac sim/ issac-ros and was starting to use the cumtion and I am stuck on this same issue as well. Since this issue was opened in Dec’25 were you guys able to resolve it.
I am using the latest isaac-ros docker image (ros2 jazzy).
By the latest Isaac ROS Docker, do you mean Isaac ROS 4.0 or 4.1?
We just released Isaac ROS 4.1, which includes some updates to franka.launch.py and cumotion package that should have fixed the issue above. Feel free to try it out and let us know if you’re still experiencing the same problem.
I’m running into an issue that looks somewhat similar to a previous post, but this time RViz shows “no planning library available.”
When I run:
ros2 service list | grep controller_manager
I can see /controller_manager/list_controllers.
However, when I run:
ros2 node list | grep controller_manager
there is no output at all. In other words, the service exists but the corresponding node does not.
How can I resolve it? I would be very happy if I can get some advice.
Hello @g2739741069,
Since this is a separate topic, would you mind starting a new thread? It helps us better track and support issues on the forum.
Also, in order to assist you better, please let us know which version of Isaac ROS you’re using and what system you’re on. Thanks!