Revision history [back]

Gazebo uses client-server architecture, namely there are gzserver and gzclient executables. In ROS they can be started by $ rosrun gazebo gazebo and $ rosrun gazebo gui, accordingly.

So, if you need a fresh world you have to restart both server and client. It is also a good idea to restart all other ROS-nodes as well as roscore. In general, the latter is not required, but may save you a lot of time.

Hope it helps

Gazebo uses client-server architecture, namely there are gzserver and gzclient executables. In ROS they can be started by $ rosrun gazebo gazebo and $ rosrun gazebo gui, accordingly.correspondingly.

So, if you need a fresh world you have to restart both server and client. It is also a good idea to restart all other ROS-nodes as well as roscore. In general, the latter is not required, but may save you a lot of time.

Hope it helps

Gazebo uses client-server architecture, namely there are gzserver and gzclient executables. In ROS they can be started by $ rosrun gazebo gazebo and $ rosrun gazebo gui, correspondingly.

So, if you need a fresh world you have to restart both server and client. It is also a good idea to restart all other ROS-nodes as well as roscore. In general, the latter is not required, but may save you a lot of time.

UPDATE:

System restart should definitely wipe any temporary data. So the problem is somewhere in startup scripts.

It is possible that spawner is dying midway and restarts automatically, e.g. the node has attribute respawn="true" in a launch file. Another possibility is that 'spawner script' is just included twice.

Please update the question with contents of your launch and world files. Additionally it would be good to see startup log of your scenario.

Hope it helps

Gazebo uses client-server architecture, namely there are gzserver and gzclient executables. In ROS they can be started by $ rosrun gazebo gazebo and $ rosrun gazebo gui, correspondingly.

So, if you need a fresh world you have to restart both server and client. It is also a good idea to restart all other ROS-nodes as well as roscore. In general, the latter is not required, but may save you a lot of time.

UPDATE:

System restart should definitely wipe any temporary data. So the problem is somewhere in startup scripts.

It is possible that scripts, world or model files.

There are three possibilities:

1. The spawner is dying midway and restarts automatically, e.g. the node has attribute respawn="true" in a launch file. Another possibility is that file
2. 'spawner script' is just included twice.

3. Gazebo world or model has a problem. For example, if your robot utilize the gazebo_ros_controller_manager plugin, it may hang because you did not load the controller, e.g. robot_mechanism_controllers/JointTrajectoryActionController. Controller manager will wait for controller till it loads. In this case actual model will not appear in a scene, but will occupy the name. It not explains why the model is spawned twice, though it can be a part of the problem.

Please update the question with contents of your launch and world files. Additionally it would be good to see startup log of your scenario.scenario. If the listings are too long, please use Pastebin or similar service.

Hope it helps