EDIT: I think I fixed the behavior by setting the correct host IP in the binding configuration. But I have no clue why
Hi,
I use several Shelly 1 devices connected to my physical light switches.
In openHAB I make use of the “follow” profile to switch other items, once the Shelly has been turned on or off externally / manually. This worked flawlessly meaning instantly in the past and stopped at an unknown time when I upgraded my openHAB instance. Maybe I switched at the same time from MQTT integration to Shelly Binding. Not sure.
Now, when pushing the light switch manually I see that it takes a random time within approx. 30s for openHAB to detect the change of the state. I tried to disable the ECO feature in the shelly, but this didn’t help.
I don’t use Shelly devices, but “random time within approx. 30s” suggests that the devices are being polled at a 30s interval. This seems relevant from the Discovery section of the docs:
Important for Generation 1 Devices: It’s recommended to enable CoIoT in the device settings for faster response times (event driven rather than polling). Open the device’s Web UI, section “COIOT settings” and select “Enable COCIOT”. It’s recommended to switch the Shelly devices to CoAP peer mode if you have only your openHAB system controlling the device. This allows routing the CoIoT/CoAP messages across multiple IP subnets without special network setup required. You could use Shelly Manager (doc/ShellyManager.md) to easily do the setup (configuring the openHAB host as CoAP peer address). Keep Multicast mode if you have multiple hosts, which should receive the CoAP updates.
Thanks, in deed Gen 1, but COIOT was already enabled. I think I fixed the issue by setting the host IP in the Shelly binding. No clue why this helps as I don’t know how COIOT works
well, I use openHAB in docker container with “host” networking. So I set the “Host Interface IP” in the Shelly binding to my normal ipv4 host ip address. Previously the value was empty.
