OpenHAB + RPi2/RPi3 = work or buggy?

Folks, as you know OpenHAB seems have some problem to run on RPi1, the bug is related with duplicated message, I just wondering what’s your experience with RPi2+?

Also, does USB Z-Wave stable enough? (e.g Aeotec) or razberry board is more reliable?

I have OH2 running on a couple of Pi2s without much trouble at all. I have not tried a Pi3 yet, but there is another problem with 64-bit ARM.

What is OH2?
Which Z-wave module are you using? what model is it?

the problem seems related with Debian itself, have you run into this problem or you are suggesting it might happen in the future?

OH2 = openHAB2. I did have it running with openHAB 1 as well though. The link was to my post about the 64bit serial native code not being available. It is a real problem, but seems to have a path forward. Suffice it to say, if you install on a 64 bit ARM machine today, the USB z-wave stick will not work without the extra work documented in that thread.

I am using the Aeotec USB dongle, both the S2 and the Gen 5.

well I think RPi2 is running 32bit code, so I am not worrying too much.

so Aeotec USB dongle seems stable enough for production use.

I have OH2 running on Pi3 with AEON GEN5 USB Stick. I just migrated within the past few days, and outside of the recent zwave binding issue, all is up and running as of last night for me. I was able to add the controller and OH2 auto detected my 5 zwave devices. 4 updated correctly and are controllable. 1(ge zwave power outlet) is being detected, but the type is not updating and is not controllable at this time. Will be checking on the zwave database for that one.

Time will tell on the reliability aspect…


Never heard of anyone really encountering this RPi1 bug. It might have existed in specific OH software revision, but hasn’t surfaced anytime recently. I guess someone should better remove the pointer from the Wiki.
RPi2s are fine, lots of people here have been using it with either the Aeon stick or RaZberry board (I use the latter).
Modules? Short answer is they all work fine, but of course that’s a question too broad to answer quickly. Browse the forum. Most popular ‘general purpose’ ones are the Fibaro modules, followed by some Aeon and Philio products, I’d say.

Since you asked in another thread about HK frequencies, now that’s a specific problem. HK seems to use a specific band (919MHz) that is possibly not covered by most vendors (not sure, you need to check. At least I have not seen any vendor explicitly offer it).
So if I were you, I’d choose either US or EU frequency modules (and controller, of course). They’ll work fine unless HK is a hotspot to use the same frequencies for other purposes.

I use a Pi3 and a Pi2 to run many things largely successfully.
The only reliability problems I have is the occasional file can’t be read problem on startup, largely but not totally alleviated by loading the rules one at a time using a linux batch file to move them out of the way at startup, before copying them back one by one.
One of the biggest problems as a system grows seems to be with the CPU being maxed as a rule file is read which most of the time gives a stack dump when it reads the next rule file, my workaround is to load each rule file individually by copying them from a command prompt with relevant delays inserted by trial and error if needed.
I have a Pi3 as my main system with a Pi2 connected via Ethernet and MQTT.

I have been running Openhab on 3 RPI 2’s since last autumn and it is very stable. I have 2 x Aeon Z-Stick GEN5 and 1 x usb controller and both work well.

My installation of OH1 on Rpi2 + Aeon Z-wave stick runs already for 3 month without any single reboot or hangup. During this time I added about 10 of different z-wave devices online and programmed couple of rules.
I would say this is quite stable for opensource software.

I see a good many of you posting that you use a RPi2. Is this because it is more stable than the RPi1/3? I am thinking of moving my OH install from Windows to RPi3 but the bit about Zwave problems from the wiki is causing me to rethink the move. Can someone confirm that Zwave problems have been resolved?

I’m currently using a Zstick (black one not the latest version) on OH 1.3.8

dinki, this is just because RPi2 was available at the moment, when I started with my HA. Right now I would select RPi3 off-course. I have read about RPi + Z-Wave Problems on Wiki, but really don’t understand about what those are - my install works pretty well, and I’m moving slowly to 20 z-wave devices in my network.

Same for me. And as I already wrote somewhere up, I have never actually met anyone to encounter these zwave problems in recent times.

Pi3, AEON Gen5, 5 zwave devices and no issues.

Just need to take my words back. Couple of days ago I started to notice that sometimes some of my Z-wave devices don’t reacting instantaneously and whole Z-wave service hangs up for 10ths of seconds. This happens occasionally with my roller shutter relays. I will investigate this more.

So this is just a slight delay but the devices do continue to work, right?

Pi2, latest 1.x OH, AEON Gen5, 20 z-wave devices. My only issue is homebridge/HomeKit doesn’t always work for the first request but that could be iOS or homebridge… Other than that, my setup will run for months without issue.

No. It’s absolutely freezing. E.g I have a Z-wave rollershutter controller, controlled by wall control panel. My wife comes and presses button to close the shutters. They start closing. After some delay she presses stop button. The shutters continue to move. She presses up button, since shutters are too low. No reaction. She starts to press stop, up, down button in different order trying to do something. No reaction. Suddenly shutter stops. After a while (10-15 seconds) shutters starts quickly execute movements, similar to wife’s button press sequence - stir up and down, stop etc, until situation normalizes. This looks like all button presses were stored in queue and then flushed to z-wave device in just seconds.
I checked the logs - this definitely lies in z-wave communication - for some reason second frame was lost and then the circus started. The problem is that it takes up to a minute to restore everything, and during this time all other z-wave devices affected too - for example I can not turn lights during that time.

I see the same problem with zwave rollershutters and also with zwave light switches.
This is on a RPi2 using the Razberry card as the controller.
A press of the “down” button just now had a delay of a slow count to 10 and the the stop command just afterwards did not respond until a count of 30.
Does anyone know of a way to analyse this?

Well, delays are not the original problem, so don’t mix it.
Start by enabling zwave debugging and compare timestamps of zwave messages with events.log (when you pushed the button), check for retransmit and error messages in general.
Use habmin or Z-Way software to visualize and cross-check zwave routes and controller view on nodes. Not unlikely something’s bad in there. And please open a new thread if you need help there.