desired_configuration needs to be sync between metacontrol nodes

[meta-control]

desired_configuration needs to be sync between metacontrol nodes

[ipa-nhg]

to me the best solution will be to set a new parameter to publish (from the metacontrol, right?) the new desired configuration. For example:

rosparam set /current_configuration "f1_v1_r1"
(obviously from code and not from command line)

ros_graph_parser (and rosgraph_monitor? ) use it to know what .rossystem to compare with the one extracted from the rosgraph

Here is not needed. We can't modify dynamically the desired.rossystem that the rosgraph_monitor takes as input. The main purpose of the rosgraph_monitor for this models experiment is verify that all the required nodes to run my application are up. The current desired.rossystem looks -> https://github.com/rosin-project/rosgraph_monitor/blob/observers/resources/metacontrol_desired.rossystem

[chcorbato]

We can't modify dynamically the desired.rossystem that the rosgraph_monitor takes as input. The main purpose of the rosgraph_monitor for this models experiment is verify that all the required nodes to run my application are up.

For the immediate future, I am very interested in being able to change it dynamically, in order for the metacontrol to manage the nodes on the fly, have a dynamic rosgraph (e.g. kill and launch new nodes as required by the mission goals)

[chcorbato]

to me the best solution will be to set a new parameter to publish (from the metacontrol, right?) the new desired configuration. For example:

rosparam set /current_configuration "f1_v1_r1"
(obviously from code and not from command line)

Agreed! I think we should name that rosparam desired_configuration, so as not to confuse it with the one the rosgraph_parser generates

• @marioney has an issue for metacontrol_experiments
• @chcorbato needs to do it for mros1_reasoner

[chcorbato]

Since it is under control and there is one issue that will finalize the implementation of this, I close this "design issue"