Z-Wave repair re-creates routes for nodes on your mesh. There is no magic behind this process, and it will not revive failed devices.
Most of this process is for legacy devices, but it can help to occasionally run repair on a Z-Wave plus network as well. Z-Wave Plus was designed to not need repairs and will self-heal.
What happens during a Z-Wave repair:
-
The Hubitat Hub will do a quick assessment of your network and create a repair plan that organizes the nodes to repair. It does this to maximize efficiency and reduce potential failures that can come from having no routes available to reach a node. It will repair nodes closest to the hub first and then move out hop, by hop. Only mains powered devices can currently be repaired.
-
Once a node is selected for the start of repair, the hub will ping the node (NOP) to determine if the node is reachable. If the node is reachable it will go on the next step, if the node is not reachable it is marked as failed and will retry in the last step.
-
The hub will request the node look for new neighbors. During this process, the node will ping all nodes in its range and report back to the hub.
-
The hub will then assign new SUC return routes, route back to the hub, based on the updated internal routing table the hub maintains.
-
The hub then requests additional routes assigned to the node for each of the virtual node ids assigned to the hub to maximize request and reporting efficiency.
Z-Wave repair is very network intensive as every device is re-routed. If you are not currently experiencing delays or problems, I would not run it.
For more details on mesh performance expectations, check out this training module by Silicon Labs.