It could be a network problem but in your first post you said it looked like the UPnP problem started after the latest update – so it’s difficult to know where we should focus.
Regarding your possible explanations:
- I’d place this last on the list of possibilities, though never say never.
- That is possible. I remember a similar issue that involved Avahi mDNS broadcasts not propagating between wired and wireless parts of a local network. Zeroconf between rasbain and osmc - #12 by dillthedog The test results indicated that the router wasn’t bridging multicast traffic, though the reason was never established.
- I double-checked and the consensus seems to be that if they’re fully unmanaged, it should work. If they’re “slightly” managed, you might need to enable IGMP snooping. What model(s) are the Netgear switches?
I’ll run a few more tests to see if I can understand what might be going on.
Update: On my LAN (actually a VLAN segment), everything works as I believe it should and there is no need to run tshark on the OSMC box for the multicast packets to arrive at other boxes on the (V)LAN. (You need to double-check if this particular problem can be reliably reproduced.) If you’re not running the 4.9 kernel and your Vero4K+ is up to date, the only significant difference on the OSMC side is that I’m using a non-plus Vero4K.
Placing a cheapo unmanaged switch (TP-Link TL-SF1008D) on the network also doesn’t stop the multicast packets from getting through.
From my limited experience of trying DD-WRT firmware, it can often have a few “rough edges”. If your router chipset can support it, OpenWRT can often be a better choice, IMO.