Post
A recurring OpenArm problem is that controller_manager changes look fine in logs but leave the robot in an unsafe or confusing state after a switch or restart.
How are you handling ROS 2 controller switching and safe restart order on OpenArm when you move between bringup, teleop, and task-specific controllers?
Please share the restart order, stale-state checks, and how you confirm the arm is ready before operators continue.
If you reply, include one exact controller-switch symptom and one exact restart or validation step that fixed it.