after upgrading to the new HA version I get the following KNX errors after boot:
Error: KNX bus did not respond in time to GroupValueRead request for: 1/6/1
Error: KNX bus did not respond in time to GroupValueRead request for: 1/6/2
Error: KNX bus did not respond in time to GroupValueRead request for: 1/6/3
Error: KNX bus did not respond in time to GroupValueRead request for: 0/7/2
Error: KNX bus did not respond in time to GroupValueRead request for: 1/5/2
The addresses have set the flag for reading and the errors did not occur in the previous version. Does anyone have the same problems as me?
I also have this errors after 0.115 update. I already changed the knx configuration to the new format and own files for sensors, binary_sensors and so on. Still the error is similar:
Error: KNX bus did not respond in time to GroupValueRead request for: 2/1/13
and a bit later:
Could not sync group address '2/1/13' for Dach - Value
I think it could has something todo with bus load (avg of 2% here), this was not a problem before 0.115 but the ETS Group Monitor shows (as expected, because of sync_state:true) loads of read requests for switches and window/binary_sensors (over 60 here). So the read request is sent and the response is in most cases sent back after 0.3 seconds. Is this timegap to big and raises the error in XKNX?
Default timeout is one second (but not exactly from sending the telegram, but from queueing it).
This will be handled differently in the next version of xknx.
You don’t need to worry about these log entries for now - incoming telegrams will be processed, even if they come a bit too late.
I also get these warnings for exposed knx items.
E.g. I am using KNX Climate where the temperature sensor is exposed to KNX, and then again read from KNX by the climate integration.
In ETS bus monitor I can see HA answers readings of the exposed addresses, so this part works fine.
However, KNX Climate does not show the current temperature any more.
Thats a known but that should be fixed in the next update. Outgoing telegrams were not processed by other devices (with tunnelling connection - I think routing never did this)
Unfortunately I have the same problem. starting the old version 0.114 currently helps.
I think, the bug cames with the breaking changes in the KNX Integration with version 0.115:
Unfortunately, reading exposed values via KNX group addresses (back-reading to HA) does not work also with 0.115.3.
Is there maybe a workaround for this issue, e.g.
can the temperature_address item of HA’s KNX Climate use an HA internal value of the sensor instead of a group address in some way?
could there be a way to establish multiple KNX tunnels? (one for the exposed values, one for the readings etc… just an idea, not sure if useful)
Do you know a certain date or HA version when this will be fixed (or is it maybe already fixed in beta channel)?
(I had to replace my previous KNX gateway (Loxone Miniserver gen.1) with the MDT KNX Ethernet Interface at about the same time when upgrading from HA 0.114 to 0.115, so I am not quite sure if it’s just an integration bug or the device might be not fully compatible… which would be unlikely, I think)
IMO mentioned “next update” might have been understood as incl patch releases. TBH it was valid presumption since bugs brought by major releases are fixed in patches. Since that 115.3 has been released without expected fix
devices not processing outgoing telegrams (eg from expose) -> solved in 0.116
states not updated (error logs “could not sync”/“did not respond in time”) -> not solved; we’ll take another shot in 0.117
but afaik it works correctly most of the time (for small to medium sized installations), it just falsely logs errors.
My switches work on 0.118, but very often I can they don’t update switch state to off when I switch light off from physical sensor. It worked fine on 0.117.6.