Z-Wave - Secondary controller

Hi all,
I read around some people added a secondary controller, using either another Vera, or an Aeotec USB controller with different software.
I did not understand some key points:

  1. in case of failure of the primary, is it possible to “promote” it as master?
  2. if using a different platform, like OZW o Z-way that support “more features” (like S2) can you add device from the secondary controller using the new features?
  3. Usually when you add a new device you need it to be close to the controller: can it be the secondary controller?

I’m sure I’ll find many other question, but this is a good start…

Thx

First question as to why you are interested would be to ask yourself what problem you are trying to solve…

You might want to investigate the controller shift function of zway. It enables shifting the master role from one controller to another. It does not cover the failure scenario though and this is when the backup comes into play: You can restore any master controller from backup to replace the master.

These features will only work on the controller which supports them. So if you included with z-way, S2 will work from z-way but the vera won’t be able to control that device at all since it does not support S2.

This is where the SIS/SUC scheme comes into play. It was an improvement to the zwave topology which muddied the definition of “primary” a little.
Traditionally you would have one primary controller and only that one controller could include devices. The secondary could only control them.
With a SUC, you would still have one controller with a primary role but now it could be mobile and you would keep the Static Update Controller (SUC) static. As its name indicate, it is static location wise and is the reference point of the mesh topology, recording the routing table of the mesh. It is a controller which is supposed to be always on and in your application likely should be the vera… though the vera does not support it. Ha!
Enters the SIS, Static ID Server, which needs to be the SUC of your network if you want a SIS. The SIS records not just the routing table but also manages the inclusions from multiple primary controllers on the network. Yes this opens up quite a bit of possibilities but also confusion. The SIS would be called now the “true primary” and the others would be secondary with primary capability. It means that any of these secondary controllers can include devices but that they report any new inclusion to the SIS.
Again the vera doesn’t really support this even though it is a zwave level firmware capability… So it actually works but you can’t enable it with the vera. Almost any other zwave controller can do it.

More info here:

2 Likes

It remains a miracle to me that people can learn this complicated stuff. I don’t know how or where, but since I haven’t run across this detailed info on my own using Google or YouTube by now, it must be well hidden on the Dark Web somewhere.

Thx for you reply. You are as usual very fast and technologically deep with your answers .

The reason why I’m looking for a secondary is that reading the forum I’m not very confident in Vera and with the future development.
I have a small zwave network (about 12 devices) and everything is running well, but I’m a bit lazy and I would like to avoid opening all the wall switches to reset and re-pair all the devices. I’m moving all the logic to Home Assistant + NodeRed, so I’m using the Vera Plus more and more as a bulk controller. Originally it was tought to be the brain of my home automation, but I quickly realized it won’t fit my zigbee needs and the bluetooth is not supported at all (although it is still advertized!).

I read many posts of yours about Z-way and it really look promising but unfortunately there is no (native) integration with Home Assistant and there is no Docker container (I would really like to find a solution that could coexist with deConz for zigbee). But if a secondary controller could help the setup, it could be still a viable solution.

1 Like

The whole zway discussion is related to having first shifted to openLuup and therefore staying within a vera environment. If you have already moved to Home Assistant for your automation environment, then you may want to either stick with what is already integrated there and add a secondary stick to your network first. I don’t recommend the aeotec stick because it uses a very old SDK (no S2, smartstart) and does not support any manual firmware upgrade. Secure inclusions are also very tricky with it. Either a generic uzb like the one from homeseer or any other brand or the uzb for zway would be better even for Home Assistant as they will be more future proof. Home Assistant is shifting towards a much friendlier platform in zwave QT and away from zwave2mqtt. I regularly follow its author’s and openzwave’s author, fishwaldo, insightful posts.
I started zway as you probably read in many of my older posts, as a secondary controller and zwave tool. I used to stick the uzb into a windows laptop running zway and walked around the house to include devices or to fix their configurations when the vera would mess them up or failed inclusion. I used it also to set the vera as the SUC/SIS in my network and therefore the uzb as a secondary with primary capability.
After using it as a secondary, you can also clone the vera’s zwave chip onto the usb dongle and move it to the new controller when ready to be rid of the vera.

PS: Note that the only two controllers I know of which support S2 security are Homeseer and Z-way. openzwave is still far from doing it and the reason why it is important is because S2 significantly decreases the amount of chattiness on the network while improving security. S0 security adds quite a bit of frame exchanges, more than doubling the network traffic for doing anything.

Thank you @rafale77, you look to me as the most relevant z-wave evangelist around!

Thank you for the info about the Aeotec stick. I’ll try to sell it and go for a UZB or a Razberry. Do you think a Raspberry Pi 3 B will be enought to support z-way properly?
I saw a couple of github mqtt plugins for zway: I would prefer to have a direct support for Home Assistant (it means one less component in the chain, one single point of failure less), but mqtt will do.

When talking about home automation, expecially if using it as home security too (just think about covers!!!), security is a must. I probably do not have enough technical background to understand why, but I expected all my devices to be added in secure mode, while they arent.

It’s what I use. Perfectly adequate, IMHO, for running openLuup with the ZWay plugin.

Yes a rPi 3B should do. The only downside of the rPi is… the sdcard storage. I highly recommend to boot it off an external drive and leave the sd card alone.

I got a couple on sale at the time the rPi4 was released and they are nifty little things. Security is indeed quite important.

This is an interesting article about it:

I have just noticed that even openhab has a bridge plugin to integrate z-way as a zwave controller in spite of natively supporting openzwave… this says a lot…

Might be worth posting a request to the folks in home assistant to see if anyone would want to pick it up.

@rafale77, why dont you pick it up? You have the knowledge and HomeAssistant…

I am fully focused on openLuup at the moment. I dabbled in python but am nowhere close to be able to write something like this on it.

Would it be possible to get a backup of zwave network from Vera Plus and restore it on Z-Way?
Or eventually, could I add Z-Way as a secondary controller, setup mqtt and then reconfigure HA to move configurations to the new settings (mqtt), and finally remove Vera Plus and “promote” Z-Way as primary?

One more question: do you know if there is an official Docker container for Z-Way?

I provided many answers to your questions here.

1 Like

Thx a lot. You wrote there were problems with the razberry shield: is this still the case?

Not with the razberry shield but with the rPi version of z-way. You can have a razberry shield and send its zwave port into a VM running on ubuntu like @DesT does. But I think the versions released yesterday 3.0.5 resolved that problem.

I’m using debian by the way, with no Desktop, just command line in a VM and it’s working perfectly stable since day 1 using socat to get the data from the rPI!

This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.