Updating this:
At least my zigbee-stick using /dev/USB0 (a CC2652RB) works nicely along with the zwave-binding in OH3.
I’m currently using zigbee2mqtt, but am wondering if the zigbee-binding of OH3 supports my stick directly…
1 Like
Ok, that sounds good !
how did you get the stick to work without the serial binding in OH3 ?
Or is there allready the v3-version of the serial-Binding available ?
I might need to clarify that:
- I’m running the CC2652RB stick on zigbee2mqtt, which relays the data to OH3 via mqtt
- I tried to get it running natively with the zigbee binding, but it’s not supported and does not run as a CC2531, either.
However, I can reproduce that zigbee2mqtt crashes whenever I restart openhab. So there’s still some connection between the two.
Unfortunately I have to confirm this 
openHAB3 Stable Release, Zwave Binding, zigbee2mqtt 0.17.0-dev, CC2652RB firmware from 20201026.
Runs nicely … until you reboot openHAB.
I did a simple workaround: whenever openHAB needs to be restarted I restart the zigbee2mqtt.service either manually or via System started trigger:
rule "restart zigbee2mqtt"
when
Item Z2M_restart received command ON
then
executeCommandLine(Duration.ofSeconds(10),"sudo", "-u", "openhab", "sudo", "/usr/bin/systemctl", "restart", "zigbee2mqtt.service")
end
Seems like it’s not only a restart of OH3 that crashes z2m, but when I recently pulled my zwave-stick, the same happened. And for some reason it crashed tonight at 2:30 am, when I definitively didn’t mess with the system.
So my approach is a bit different:
I have two CC2530 routers, and those are quite chatty regarding their link quality, posting updates every few minutes or so. I feed these into a time-based database (influxb), and my z2m-restart-script pulls the average of the link qualities for the timeframes 0-5min, 5-10min and 10-15min. If those three values are the same in both routers, it’s safe to assume that no updates have come from z2m for 15 min and that it needs to be restarted.
That are really bad news. If the crash occurs more frequently I probably stick to my old controller CC2531.
I had plans to extend my zigbee network, but I think I have to rethink this …
I had exactly these phenomena with OH2. I also had the problem, for example, when changing the “items” or “sitemap” files. After making changes and saving the files, I had to restart Z2M again and again.
With OH3 at least the phenomenon is gone. Z2M only hangs when I restart the machine or when the Zwave controller is restarted. I think you could avoid the whole thing by letting the Z2M service start with a little delay (i.e. when Zwave is already running).
But I don’t have the problem with hanging up at 2:30. I also use serial binding v3 with standard configuration.
After tinkering around for a few days the CC2652RB works fine now.
Only thing I have to do is restart Z2M after a short delay during startup of openHAB.
Serial Binding is not installed on my system, though.
1 Like
I have the same issue.
OH 2.5.11 Stable
Z-Wave (binding-zwave - 2.5.11) connected to Aeotec Z-Stick Gen5
Z2M (1.17.0) switched from CC2531 to Slae’s CC2652RB
After switching to CC2652RB I get the following message without the chance to add new devices to the network:
Jan 10 10:59:56 openhab npm[32223]: (node:32238) UnhandledPromiseRejectionWarning: Error: SRSP - ZDO - mgmtPermitJoinReq after 6000ms
Jan 10 10:59:56 openhab npm[32223]: at Timeout._onTimeout (/opt/zigbee2mqtt/node_modules/zigbee-herdsman/dist/utils/waitress.js:46:35)
Jan 10 10:59:56 openhab npm[32223]: at listOnTimeout (internal/timers.js:554:17)
Jan 10 10:59:56 openhab npm[32223]: at processTimers (internal/timers.js:497:7)
Jan 10 10:59:56 openhab npm[32223]: (node:32238) UnhandledPromiseRejectionWarning: Unhandled promise rejection. This error originated either by throwing inside of an async function without a catch block, or by rejecting a promise which was not handled with .catch(). To terminate the node process on unhandled promise rejection, use the CLI flag `--unhandled-rejections=strict` (see https://nodejs.org/api/cli.html#cli_unhandled_rejections_mode). (rejection id: 25)
You think upgrading to OH3 and delaying the start of Z2M might solve the issue?
This looks like a Z2M problem, not related to openHAB.
I agree. You can test this by temporarily disabling the zwave stick in OH and see if z2m runs then.
I have the same exact issue with OH3 (3.1) and zigbee2mqtt… both in Docker containers… I know it has been 6 months here but any updates on a working solution?
1 Like
Unfortunately I still have to use the workaround (restarting Z2M) also with the latest openHAB 3.2 stable 
Changing the zigbee coordinator solved the issue, no restarting of zigbee2mqtt required anymore.
old: Slaesh’s CC2652RB
new: SONOFF Zigbee 3.0 USB Dongle Plus
Change could be achieved without repairing 
1 Like
I have the same problem with the modbus binding and slaes stick for zigbee2mqtt. I need to manually restart zigbee2mqtt After Openhab starts.
Apart from programming a restart of zigbee2mqtt or change the stick, is the not a solution?
It seems to be related with serial communications in Openhab.
Has someone tried the alternative firmware in the slaes website?
Hi @sihui I also switched to a new SONOFF Zigbee 3.0 USB Dongle Plus (Version E, Based on EFR32MG21). Zigbee2mqtt works fine but zwave, running with OH 3.1.1 says in PaperUI it’s online but all the devices lost the connection to my zwave controller “Aeotec Z-Stick Gen 5”. Did you configure something that both systems work?
Nope, I did not touch the Zwave network when changing the Zigbee coordinator.
Your Zwave controller stick shows online in MainUI?
If you changed something in your zwave setup, verify that the controller identifier is still the same. If not, you need to adapt all your channel definitions.