Please reopen if this occurs in the current release.

I am seeing this same issue, though in my case it's all Netgear "smart" "semi-managed" switches (supporting SNMP and LLDP, but no CDP/OSPF/IS-IS I believe - ). I have just set up OpenNMS in a VM for testing purposes to give it a try.

Ports which have a single MAC on them still show up as "Multiple Mac Addresses" with a cloud icon. Sometimes both "Shared Segments found by Enhanced Linkd" and "LLDP Remote Table Links found by Enhanced Linkd" will list a MAC, sometimes just "Shared Segments" (not all devices appear to respond to LLDP?), but it doesn't matter, the ports will still show connected to a cloud rather than the single device.

I have exact the same issue:

Core Switch is a Linksys (by Cisco) SRW2024P 24Port-Gbit switch. All end-devices are directly connected + one managed HP switch + Internet Router and 2 Access Points.

The topology map only shows resiliant the shared segments and the LLD-links to the HP switch. On next level (2+) in most cases no connections are shown.

I digged into the bridge-to-bridge information collected by enlinkd. All mac adresses and links discovered are correct and there .

The funny thing is: when I reload the topology map, sometimes 1 or 2 devices directly connected to the bridge are shown, but more with a hazard principle.

Example: I have two exactly the same access-points connected with the same version and configuration; one is shown sometimes, the other never)

As i tried this on 2 different installations (docker, Horizon Core Web V22.0.3, V23.0.0.) with exactly the same behaviour: I assume that there must pe an issue with the instance who renderes the topology map (vertex and edges).

I tried nearly everything to relsolve: deleted the core switch and waited until discovery of other devices is reloaded/calculated; then re-provisioned the switch, tried different endlinkd /snmp-config configurations and time-out sesstings etc.

Nothing helps!

pls help!