Not that it’s helpful in this case, but on my network, I do not see a “-1” or “-2” suffix’d listing.
My only mDNS entry is of the Bond Pro ID directly, and it resolves correctly.
@endy - is this something you’ve seen in any of your test setups or do you otherwise have an idea why Alex might be seeing this in his environment?
@merck - another instance where ethernet seemed to be the fix for 107 (but I think you’re already tracking this on the 2.17.2 upgrade)
Ah, that’s a good point. This could cause some trouble if one of the interfaces goes down and then the mDNS name is stuck with the -2 on it. That could make it harder for the app or other API clients to connect. All the more reason to use BPUP discovery
We cannot fix the mDNS standard, which is braindead in a few ways, in particular not handling multiple interfaces well.
We could discourage users from setting up both Wi-Fi and Ethernet, or forcibly disable this via the firmware.
But it’s a quite minor issue I think. There’s only an issue if users (a) use both Wi-Fi and Ethernet simultaneously and (b) one of these interfaces has a connectivity issue during discovery.
The only reason why I had to setup Ethernet - because I couldn’t update FW via WiFi (error 107).
And I can’t understand the mDNS problem with two interfaces - if each interface has its own IP - then mDNS could report two different services on two different endpoints?
And I know what BPUP is (I use in HomeSeer integration) - I just can’t see how it can be used for discovery (BPUP discovery)?
I guess we don’t document the discovery use case.
Basically, you can send a BPUP packet to the broadcast addr (255.255.255.255) and all Bonds on the local network will reply with their Bond IDs (and you’ll get the IP addr from those replies’ metadata).
And, in firmware v2.18+, we’ve added d (discoverable) and v (version) fields on mDNS, BPUP, and MQTT, to save some requests in discovery. (This is discussed in the latest API docs.)