1. Takeover and giveback have multiple options available, however we'll focus on the basic.
this option performs a graceful takeover of the target nodes storage, starting with the SFO (data) aggregates, and then the CFO (root) aggregate. The target node then reboots and will go into waiting for giveback (WFG) state.
2. system node reboot, is still from the cluster prompt but uses a command subset of "system" instead of "storage failover" as in a takeover/giveback operation. This performs a reboot of the target node, which will boot into WFG as well. This as well has multiple options available to it, such as preventing a takeover by the partner node (default is to allow the partner to takeover).
3. Halting a node (system node halt) will halt the target node at the loader prompt. there are options here as well to prevent a takeover by the partner node.
4. boot_ontap from loader simply boots the node into ONTAP or into a WFG state depending on previous status and state of the partner.