I reinstalled OH to upgrade to v4.x and then reconnected the z-wave nodes to the system. I have done this before to other systems with no problems. In this case however, the communication from the nodes (according to the network map) is only one-way. This seems to be confirmed by some communication problems we’re having.
I checked the parameters for the Aeotec z-stick and one of the nodes, which is powered by USB, against another system and I don’t see any differences.
All nodes have dotted lines with single-direction arrows rather than the two-way solid lines I usually see:
Nothing to be concerned about. The neighbor map is created / modified during a network or node “Heal”. It is not diagnostic and doesn’t show actual ZW routings. It is based on the zwave_neighbours property on the device UI page. The main change was that with 700 and 800 chip controllers the controller was unable to “Heal” itself, so was excluded from the process (starting with OH4). 500 chip controllers somehow managed to heal themselves but took a long time. The assumption, verified with zniffer, is if a node has the controller as a neighbor the controller “sees” the node also. You could manually add the controller neighbors using the API explorer, for a prettier picture.
As to the battery nodes, I suspect that they have not been woken to activate a heal, but are not involved with routing anyway (except to route their own messages to the controller).
Although not definitive, your network looks good to me (except 26 which is offline) with so many of the nodes having the controller as a neighbor.
