you would need to register the cluster vip and each node ip. so you would run jobs against the cluster vip object and that would always goto the active node. If you needed to do something to each node, regardless if it was active or not, you’d use the node objects.
Yeah, basically, have a target that is mapped to the IP of the cluster, and one for each o fthe nodes as well mapped to their individual primary ips.
If you need to execute a job against whatever the active node is, you execute it against the target representing the cluster's vip, and otherwise, you execute against the individual node's target.
Just make sure you run an update server properties job against the target representing the cluster's vip before running any job on it however, to make sure that the properties of the target truly represent the node it's pointing to. This won't be done automatically.