I have a hub v4 running on its own. I’m considering adding my old hub v3 in a hub group for redundancy, not planning to use it.
Obviously this adds redundancy and failover, but does it bring disadvantages?
Eg if I merge thread networks presumably I’d end up with an extra zigbee network possibly adding interference.
Does the merged thread network perform better than a single one, without any overheads?
I could of course ask all these questions from AI but I’m not sure I’d trust the answer.
I believe when you use hub groups only the main hub has devices connected and the other hubs work as repeaters. So I don’t think you will get an extra ZigBee network.
I have 6 different ZigBee networks in my small house and it didn’t have any problems with interference. And 4 of the hubs sit side by side on a small shelf.
Ok, so if I create a hub group and add a v3 hub it will just act as a repeater for zigbee and thread? Does that mean the zwave is disabled (would suit me as I’ve no plans to use it)
the thread networks will be unified when you create the hub group.
z-wave will still be active on the v3
I had a hub group of three V3 hubs for a while. I don’t have one any more.
One problem I have with hub groups is that they recommend you have the hubs at most ten feet apart which just seems ridiculous as even in the UK that can be the same room, but actually I still had problems with the hubs about six feet apart, and indeed two of them about a foot apart.
I frequently used to look at the hub group and discover the primary and secondaries had decreed that they had lost connectivity, and I had no idea why. They’ve never said how they determine this. I always suspected Zigbee was rather too involved somehow which was annoying as actually I didn’t need the Zigbee routing from the secondary hubs. Why would I want the secondary hubs routing? What would happen if you had a primary hub at capacity and then it had to failover to a secondary hub which was already handling devices? They’ve got to go somewhere.
This was after I’d disabled the automatic hub replacement. That was absolutely ridiculous. On more than one occasion I found my primary hub had changed. It didn’t ask me if I was OK with that, it didn’t warn me at all, and indeed it never even notified me after the event. There was also no obvious reason why it did it as everything always seemed to be working fine. It just decreed that as it considered the primary couldn’t contact the secondaries it must be the primary that had a problem. On what basis?
Then there was the problem with restoring the original primary. I’d see a message acknowledging that the original primary was fine and that it would be restored in a few minutes but only once did it ever happen. The rest of the time the message stayed up for hours until I lost patience and reconstructed the hub group.
There was also the problem that things didn’t always come back quickly after a replace, especially Matter devices. Those could take twenty minutes or so. Sometimes the Zigbee devices would come back. Other times they’d all be offline. I often needed to intervene.
My eventual conclusion was that a hub replacement needs manual supervision and once you’ve removed automatic failover there isn’t a lot worth having in a hub group. I don’t want my hubs acting as Zigbee repeaters but I do want the extra TBRs. If a hub goes down I don’t want it replaced by one in the same place, I want it replaced by one somewhere else. So I just have three hubs distributed across the home and just have them all on the same Thread network. If I feel I need to replace a hub I’ll do it myself.
So in this scenario do the other hubs act as thread repeaters , and if so are they using intermediate thread devices to talk to the main hub in use or do they use Ethernet/wifi for back haul?
Based on your comments you’re not selling the idea of creating a hub group with failover for me. I’d thought of placing the secondary on a different floor with two brick walls in between but plugged into Ethernet. Having the 10ft recommended distance would be impossible unless they were in the same room.
I certainly don’t think they can do anything that isn’t native behaviour of Thread.
I don’t believe my hubs are acting as repeaters in my case but whether they would do if required is another matter.
Similarly I believe the ability to use the ethernet or wi-fi as an alternative to the native thread network is a relatively recent enhancement for ST. I don’t really know enough about it.
I didn’t even realise for some time that there is ‘normally’ only one active TBR on a thread network. I thought they’d be sharing the load. Regardless, all I really want my extra hubs to do is to act as the TBR if chosen and to take over as my main hub if I request it. I’d actually rather they didn’t even have Zigbee or Z-Wave turned on.
It is rather how you’d hope it would work isn’t it? You’d expect a wired connection to negate range issues. However even without wired connections I don’t really get why there are range considerations at all as long as the hubs can see each other somewhere on the mesh.