The site is read-only. Please transition to use Robotics Stack Exchange
| ROS Resources: Documentation | Support | Discussion Forum | Index | Service Status | ros @ Robotics Stack Exchange |
 | 1 | initial version |
So, depending on the configuration being used for the action server, there is a way to tell if it crashes, but its not built into any of our tools. The default setting for a server is to publish status about its goals on the action_namespace/status topic at a rate of 5Hz (though it should be noted that there is a parameter that will allow users to run without the heartbeat). So, the easiest way to get the behavior you want at the moment is probably to subscribe to the status topic in your client node and take appropriate action on a crash, when you don't receive status for some amount of time.
| | 2 | No.2 Revision |
So, depending on the configuration being used for the action server, there is a way to tell if it crashes, but its not built into any of our tools. The default setting for a server is to publish status about its goals on the action_namespace/status topic at a rate of 5Hz (though it should be noted that there is a parameter that will allow users to run without the heartbeat). So, the easiest way to get the behavior you want at the moment is probably to subscribe to the status topic in your client node and take appropriate action on a crash, when you don't receive status for some amount of time.
UPDATE: Kai's question in the comments:
There's a camp of people who want to be able to disable sending of status on a heartbeat because they're worried about the network overhead associated. If you really care about a node going down, the proper solution is probably to set up a bond between the two nodes.
| | 3 | No.3 Revision |
So, depending on the configuration being used for the action server, there is a way to tell if it crashes, but its not built into any of our tools. The default setting for a server is to publish status about its goals on the action_namespace/status topic at a rate of 5Hz (though it should be noted that there is a parameter that will allow users to run without the heartbeat). So, the easiest way to get the behavior you want at the moment is probably to subscribe to the status topic in your client node and take appropriate action on a crash, when you don't receive status for some amount of time.
UPDATE: Kai's Kei's question in the comments:
There's a camp of people who want to be able to disable sending of status on a heartbeat because they're worried about the network overhead associated. If you really care about a node going down, the proper solution is probably to set up a bond between the two nodes.
ROS Answers is licensed under Creative Commons Attribution 3.0 Content on this site is licensed under a Creative Commons Attribution Share Alike 3.0 license.
