following the new doc on Swarm, why create a ‘local’ machine, then a swarm-master then agents?
since with the flow of the docs there is no mention of changing the env to the swarm master (except later), the agent are in fact joining the ‘local’ machine… and the example breaks.
not sure why there is a need for that ‘local’ machine, when really it seems all is needed is the swarm-master and then the nodes.