OpenZWave removed from build checks

[Gravityz]

the writing is on the wall

Code: Select all

14793	20221202	Bui;ld	Removed OpenZWave from build checks as it is deprecated anyway and caused compile errors

for later build they now use ubuntu 20.4 to avoid build errors

this means that any time now something can happen which breaks OpenZWave.

be extremeley carefull and test everything after upgrade.

i am still not moving to MQTT AD because it is not flawless

[gizmocuz]

Well for 99% of the users it will be flawless, and I worked years on the OZW integration and can tell you it is not flawless as well

Besides, ZWaveJS-UI is also much faster, and has more device support

As far as I know (except some exotic devices) there are no issues

And of course, everyone is testing, and a backup is created when updating.

Sure, current OZW users can test at this very moment if ZWaveJS-UI will work for them

They do not have to wait for a new beta or until OZW is removed, in fact, I recommend starting testing now

[Gravityz]

ok sounds promising. regarding the speed, i heared the opposite that commands through MQTT were received slower(slower light reaction) as compared with native zwave

[jake]

With Mqtt I never have any issue with zwave. With ozw I had infrequent errors that a node was dead.

It is a learning curve, but the results are good!

[madpatrick]

Well for 99% of the users it will be flawless, and I worked years on the OZW integration and can tell you it is not flawless as well

Besides, ZWaveJS-UI is also much faster, and has more device support

As far as I know (except some exotic devices) there are no issues

And of course, everyone is testing, and a backup is created when updating.

Sure, current OZW users can test at this very moment if ZWaveJS-UI will work for them

They do not have to wait for a new beta or until OZW is removed, in fact, I recommend starting testing now

@gizmocuz

Does this mean OZW is not working anymore with the new betas in Domoticz?
Even if you have OZW already installed and you perform an upgrade?

[kiddigital]

Currently OpenZwave is still (again) part of the Beta’s but it will be removed in time.

[lost]

As far as I know (except some exotic devices) there are no issues

Hello,

IMO, issue is really having to re-do most of the devices integration work. For sure, inclusion etc should be kept by controller. But having to replace all sensors/switches by new ones when they'll show up when you have ten's of nodes & hundreds of devices will not be a 1 hour job. IMO, not even a single day one.

And that's on top of a heavy layer cake software (& at the foundation, some HW driven in JS that was clearly not designed for this) compared to a clean integration. The kind of SW pile "accidentally working" with too many dependencies you're advised to containerize because that's almost impossible to reach OS/Interpreter/"driver" SW/mqtt/controller SW planet alignment of dependencies over time.

So maybe there is no issue, provided you have the right cookbook & been able to refine it on some test system, but IMO lots of traps & several days of work when you're currently full z-wave and need your working system!

Yesterday, I had to exclude/reinclude a Fibaro "eye" PIR that did not show it's temp sensor since a week and misbehaved showing wrong CO2 etc sensors. Happens sometimes, but his time a wake-up did not lead to a working again state. For this single device, having to replace devices (using an intermediate name, to be able to replace, then rename to former device name to keep all scripts that may use it working) by new ones, with only 3 used (movement detection/tamer/temperature), this was ~30 mn with a bit of testing.

I currently have 9 pages with ~25 devices under zwave category. So that's ~225 devices & even really used ones are between half and 2/3 this figure, that's a huge job & quite repetitive so error prone: For sure, issues will occurs after a very boring job. We're humans after all!

[gizmocuz]

Come on! In halve a day you can do a lot. Even is you do have 200 nodes
Is it a boring job? No, because it has to be done right
Will you benefit from it? Yes, you probably will. Everything is probably going to function a lot better/faster, you might even be able to do OTA's and because you now have everything in MQTT you can do more. (For example use homebright if you like)
You probably noticed there weren't any updates to openzwave for the last year...
Then again, if you are happy with your current system, there is no reason to change it.

And to top it all, If you ever decided to use another HA system, or use it together, you have to use ZWaveJS-UI as well....

[madpatrick]

Then again, if you are happy with your current system, there is no reason to change it.

That's my motto. It is working for me good enough and the issues with OZW i'll take them as they come.
Changing my Domoticz system which is working good is asking for possible problem.
At the moment the only reason to change is whin i need to change my system (OS) and definately will change to ZWaveJS-UI also
My OS is getting out-of-date, but migrating to other OS is a painfull operation.

Keep up the good work ! I'm very pleased with Domoticz and also with the development

[lost]

Come on! In halve a day you can do a lot.

Yes, I for sure noticed for OZW... and know this is the same for other systems & was not planning to change anyway as I really enjoy Domoticz & built a lot on it.

But even if OZW had it's flaws, it allowed much much better integration. Over time, I'm not really confident that ZwaveJS will last as long as OZW: This kind of SW design is good for web stuff designed quickly over what exists (& is actively developed) at the time with TTM as a 1st criteria + no-one plans it will stay online many years without a (almost) full rebuild over the new fancy framework of the time...

I'll probably move someday, at least try... But my system does not grow anymore since 2 years & fit my needs so there is no emergency until current stable does not run over an updated raspbian. When I said "boring", this was not intended to be rude but because it's a job already done progressively: Node after node over several months/years, this can be fancy. All at same time, that's for sure headaches & tablets of aspirin to plan!!!

All in all, if I must extend my system this may no more be using zwave (to avoid new/OZW unsupported devices) but zigbee: For a single hosted system (what the need for MQTT here? Domoticz have it's own event system!), there is a all in one python plugin option.

[waltervl]

Just to mention, the OZW project is no longer actively maintained and will be archived soon.
https://github.com/OpenZWave/open-zwave/issues/2646

So trying to keep the code running in Domoticz will be very difficult. It possibly will be in Domoticz as long as it will not give issues. As operating systems develop, older sources will not compile anymore. If you want OZW to be alive you can step up and become a maintainer of that project. Else we have to deal with what is available now (currently only ZwaveJS).

[Gravityz]

a good first step would be switching to docker.
with docker you are way more flexible and you do not need to deal with operating systems/patches/drivers anymore.

also it is a rather failsafe migration when you are upgrading versions(you basically clone the container as a new version but still keep the old one)
upgrading domoticz will take approx 1 minute and if something breaks(eg ozw) then switching back is as easy as copy the old database back and fire up the old container

i am doing this for years and bought an intel based synology box for that purpose

Then again, if you are happy with your current system, there is no reason to change it.

That's my motto. It is working for me good enough and the issues with OZW i'll take them as they come.
Changing my Domoticz system which is working good is asking for possible problem.
At the moment the only reason to change is whin i need to change my system (OS) and definately will change to ZWaveJS-UI also
My OS is getting out-of-date, but migrating to other OS is a painfull operation.

Keep up the good work ! I'm very pleased with Domoticz and also with the development

[rrozema]

That will be time I will have to leave Domoticz for openHAB then... Hope it will take some more time as I'm not done setting up my test system yet. But it gets forced upon me like this, I will have to say "so long and thanks for all the fish"...

[waltervl]

OpenHab is still using OpenZwave?

[rrozema]

I'm not sure if it's ozw, but their z-wave is still within the same executable. But what's more important (to me at least) is that their implementation is more advanced. In openhab you can not only control z-wave devices, you can also have the controller behave as a z-wave device. This enables me f.e. to use any -even non z-wave- temperature sensor as an external sensor for my eurotronic spiritZ thermostatic valves. Z-wave has many features that we haven't been able to use from Domoticz and I'm still finding more in openHAB.
There were 3 things keeping me with Domoticz: 1 ease of installation, 2 the easy user interface and 3 dzVents. 1 is now gone with the introduction of Java script and seperate processes, i have gone through that hell already with zigbee2mqtt: after a lot of time spent it worked finally, but I'm to afraid to change anything and break it.. 2 I am most likely going to find a workaround for in the form of some sort of dashboard. So it's actually only dzVents that has kept me on the side of Domoticz and that's the part I'm really sorry to say goodbye to....

[lost]

i have gone through that hell already with zigbee2mqtt: after a lot of time spent it worked finally, but I'm to afraid to change anything and break it..

For zigbee, at least, there is the single plugin option with ZigBeeForDomoticZ. This saves separate MQTT broker setup etc... Maybe this should be the way to go with zwave also, but the way it's designed, adding JS that was not really designed to drive some hardware etc may not allow this easily.

But I agree z-wave in Domoticz is now a problem with this zwaveJS implementation. At least, if they didn't wanted to cross-compile, they should have used Python instead of JS to avoid another interpreter + python could interact with C/C++ easily to provide a good old API (that could have been designed close to OZW to ease integration) to avoid MQTT hassle.

[Micha123]

I was also working with open-zwave, all was working fine, without any problems, and i won't to go to other setup

But now i switched to zwave-js-ui (so i can use it in the Future if OZW wil be removed) and i need to say, its working much better and much faster then OZW.


The only think it was much better if i wil add an device by myself in domoticz. Like OZW, so it wil be appear in the device tab and i wil add it myself what i need to add. For now it wil be automatically added in domoticz and every device i need to remove in the device's tab.

I have a lot of remote controlers and every time i pressed a button it was added in Domoticz, and i had to remove it in the device's tab..


It wil be much better if new z-wave devices wil appear only in the devices tab in domoticz and not automatically added.