How does slave computer behave after master crashes?
Could anyone give some explanation or point out some references about the behaver of slave ROS program if there are some issues in the master-slave setup? For example, the master computer crashes or the wifi connection of the master computer drops off. I have been wondering this for a while but couldn't find the answer online.
One of the scenarios is as follows:
I have the master computer running on a drone, and for safety reasons, a slave computer is running on the ground. The idea is that if the slave node detects something wrong on the master side, it will do something to land the drone. I tested this system by using rosnode kill on the master computer. Once some critical nodes are killed on the master computer, the slave one works as expected.
Then I tried to reboot the master while the ros program is running (to mimic the master computer crashes). In this test, the slave doesn't behave properly. Maybe this is related to the reboot command. It will be helpful if someone can give more insight into the master-slave mechanics.
Thanks in advance.
There are no "slave programs" in ROS. I don't really get where people are finding this.
Yes, there is a ROS Master, but that does not make any of the other nodes slaves -- or at least not in any useful sense of those words, as nodes do not "answer" to the master in any way (ie: it cannot control nodes directly).
Communication is peer-to-peer.