Please excuse me for checking before I commit to a huge task. But, there is a reason for this. I tried using the z-wave binding long time ago in OH 1.x and a Razberry board and at that time I found out that there were issues with this combo and z-wave did not work very well in OH at that time. I have described in a previous post
that I have tried using z-way for managing the network on my Razberry 2 board and the z-way binding, but the solution is not stable and not performant.
The network I want to control is built up of just short of 60 mains powered devices, and 9 FLIRS (Smoke Detectors). And I have a couple of battery powered sensors too. A blog post http://e-rook.org/2018/01/09/smarthome-server-change-openhab/ suggested that the z-wave binding is much better than the z-way system + binding, but I would like to make sure before I start moving everything out of the system.
So, what is your network size? I if someone is using the z-wave binding with a network that is similar to mine, it would of course increase my eagerness to do the work transferring stuff OH rather than go back to the fibaro controller.
My ZWave network is about the same size as yours (I stopped counting so don’t have the exact numbers).
I’m using the OH2 ZWave binding (the development version, actually) on a Pi3+RaZberry with no performance issues for well over a year.
Note ZWave network healing is not available in standard ZWave2 binding as it used to be in OH1 binding but it’s back in the development version.
I do need to resort to shutdown OH and fire up Z-Way every now and then to clean up some zwave things after I changed/extended the network, but other than that, it’s pretty stable and fast.
Well to speak frankly, the most recent version has a severe problem resulting in lockup after some hours of runtime, but I’m confident @chris will resolve that soon.
Thank you for this information! Very helpful.
Just a couple of quick follow up questions
Are the things you need to fire up z-way for related to network healing / reorganization, or something else? What does the z-wave binding not do thet z-way can do?
Ok, just to make sure - this is the very last version of the development version then right? So the development branch has been stable up until the last version, is that how I should understand the situation?
Occasionally, I had problems to properly include/exclude devices using PaperUI/habmin/Zwave binding and felt more comfortable using Z-Way, but I guess those were corner cases related to specific devices.
What the ZWave binding can not do is to backup the ZWave controller (= RaZberry) configuration and network information stored in there, so I use Z-Way to backup and eventually restore my controller and network config.
Yes, the only one you can download at the moment.
Yes.
And this code is to become the standard binding in 2.4.
My installation includes 118 zWave Devices(about 70 main powered, the rest battery powered) at the moment. And it works rock solid, stable and fast.
I run OH 2.3 stable version (including zWave Binding) on a Mini-Server (Pokini F).
Thanks @michael_sutter Great information!
Pokini is of course a more powerful computer. Did you try with a smaller computer before and then upgraded to this mini-server, or was this your first system?
I started years ago with a RPI 1, updated to a RP3. (on my “old” flat).
In 2016 I renovated a house and renew all my infrastructure. In this step, I decided to use the Pokini instead of an RPI.
The main reason for me was not the performance, but the reliability of the hardware (especially to prevent the common SD Card Problem and its workaround).
A Pi3 is (unlike a Pi1) fast enough to handle that number of devices and the inevitable number of rules such a large installation is coming with (well unless you’re computing something exceptional). Now you’re stuck with Pis as exchanging them would mean to get a new zwave controller.
HW reliability and SD corruption in particular is a topic specific to most Pi-ish single board computers, but it’s a solveable problem (see this thread and others on the forum).
My environment is not comparable but I’ll post anyway. I’ve 5 battery (three smoke alarms, 2 minimotes) and 7 mains powered devices. I’ve two deadbolts that I can add to my network but I’ve been waiting for the development version of the zwave binding to make it to a release. I’m too lazy to mess with installing it by hand and I’m actually a little ambivalent about adding the locks to the network. I’ve no interest in locking/unlocking remotely but I sure would love to get the battery status.
My experience with the Zwave binding has been rock solid. I started with a Gen 2 Aeotech controller and am now running with a HUSBZB-1 Zwave/Zigbee combo controller/coordinator. I’ve been able to do pretty much everything I need to thus far through Habmin or PaperUI, though I’ve a copy of OpenZwaveTools I can bring up in a pinch. My network is so small I’ve not worried too much about backup and restore.
As I use an Aeotec Stick the exchange of the underlaying unit isnt very complicated…
I know, there are solutions But as we had a - lets call it a good calculated - budget for features like this, I’ve lifecycled all my environment, started from Fileserver, Switches and my small little OH Server…
My answer was directed at the OP who is using RaZberry. Yeah you can move that card over to another Pi if your HW breaks or SD gets corrupted, but you cannot attach it to any computer via USB, so he’s stuck with Pis.
But I guess neither you nor him want to ex/re-include all of your devices with another controller, so ensure you can backup and restore your Aeotec stick’s or RaZberry card’s content to another controller of the same type, and have a replacement unit at hand.
Move it to the top of your list, then … if my RaZberry broke and I needed to access all devices again, with my wife waiting for me to finish … ugh, shiver. I feel much more confident now that I know I can just restore and move to another controller unit.
Oh, and no matter if your server is SD based or not, to prepare for a server OS level backup/
HW replacement/OS restore is also very important.
But as we had a - lets call it a good calculated - budget for features like this, I’ve lifecycled all my environment, started from Fileserver, Switches and my small little OH Server…