First time trying out hub mush - using two C7's on 2.3.6.146. Neither hub "sees" the other via "Active Hubs" section or "Linked Devices" tab. I've ensured that both hubs have mesh enabled, and that at least one device on each is "shared". Below is a quick snapshot of my network topo; both hubs are on the same subnet. No VLANs.
What is the mechanism for hub discovery/linking? UDP? TCP? mDNS?
:EDIT: Network config changed temporarily (below reply) to test
I think discovery uses multicast. Do your AP and L2 switches bridge multicast between the segments of your LAN?
Also tagging @gopher.ny
That's a good question - unknown at the AP/switch combo, because it's a TPLink router in "AP mode". No granularity of settings. So I've reconfigured as shown below to eliminate that possibility.
The other "L2 switches" are Netgear smart+ switches; IGMP snooping is enabled, and I've disabled blocking of unknown multi-casts, FWIW.
I've reconfigured for test, below, still no dice:
So it took a reboot of both hubs but they are seeing on another when on the same switch (as I would hope).
I switched back to my original network config, rebooted again, and they are (still) seeing one another. Not very re-assuring. I'm 100% positive I rebooted both hubs originally. So I'm not sure if it's a case of now that they've seen one another, that connection is 'cached', or it took that network config change to update multicast tables somewhere...
The most likely explanation is that the TP-Link AP blocked multicast.
FWIW, hubs look for each other on reboot, so it is reasonable that they found each other after they were on the same switch and you rebooted.
I assume this discovery is no different than performing a "refresh" on the mesh page though correct? Which I did systematically after each config change.
Just playing devil's advocate (really just trying to better understand - I'm not a fan of getting things working without being able to programmatically break it again, otherwise the true issue is elusive).
If the AP is/was indeed blocking multicast, then why are the hubs now still "visible" to each other after reverting to connection through the AP's switch? I even performed another reboot, and refreshed from mesh page, and the hubs remain connected (and functioning).
In my mind either Hubitat has cached (for lack of better words) the connection, or maybe AP/switch routing tables have been updated? Again just trying to understand how it's supposed to work (as I can foresee having this issue down the road randomly).
It seems a manual entry of a target hub's IP would make this a lot more foolproof? If it indeed falls back to TCP for operation anyway. I don't know if this is a direct socket connection or publisher-subscriber type arrangement...
Anyway thanks for your input no less.
This topic was automatically closed 365 days after the last reply. New replies are no longer allowed.
Download the Hubitat app
