A very long time ago, I entered all my KNX addresses in the configuration.yaml file, together with a certain name. At that time, there was no other way. And everything worked fine.
Later on, we could change the name and the entity-id in a visual interface. What I certainly did. And perhaps I have forgotten to do the same in the configuration.yalm file
But now, certain KNX entities don’t work anymore.
I tried already several things. Like putting the KNX entries in comment in the configuration.yaml file + reloading + restoring everything in the configuration file + reloading
I could perhaps delete the KNX integration + adding it again, but I ‘m afraid to lose all my programming in NodeRed.
I suppose HA does not use the KNX addresses as a reference, but ??.
And I suppose my configuration.yalm file does not correspond anymore with an “internal database” used by Home Assistant.
So what can I do? Can I force HA to read anything again from the configuration.yaml file? Or ??
Below an example of an entry in the configuration.yalm file that does not work anymore.
But the KNX integration seems NOT using the name from the configuration.yaml file.
What I did:
I changed the name of 8 KNX lines in the configuration.yaml file + check configuration (no remark) + reboot of my RPI.
→ the KNX integration keeps showing the old names + the delete button is unavailable
I put the the 8 items of the KNX switch in comment in the configuration.yaml file + check configuration (no remark) + reboot RPI
After reboot the old names were shown as unavailable. And this time the delete button is available. So I deleted the 8 items from within the graphical user interface.
After reboot of my RPI, the 8 items were no longer shown in the KNX integration.
I UNcommented the 8 items in the configuration file + check configuration (no remark) + reboot of my RPI.
The 8 items were again visible in the KNX integration but with the OLD name.
Take this example of an entry in the configuration.yaml file. The name is ‘Stekkers 230 tuin 8 KNX’
In the graphical user interface of the KNX integration, I cannot find this name ‘Stekkers 230 tuin 8 KNX’
In reality this switch corresponds still with an old name “stekkers_230V_tuin_knx”
So the KNX integration is NOT using the name from the configuration.yaml file. But from ??
The KNX integration probably uses the KNX address as a reference. Because, when I put the adress in comment → the “old name” is shown as unavailable.
I have read the documentation again.
As you said: “Right, once the entity is created, the name key is ignored. It is only used for initial creation”
What I don’t understand:
After putting a line in comment in the yaml file, this entity becomes unavailable and I can delete the entity in the graphical user interface of Home Assistant.
But when I restore the contents of the line in yaml, the entity becomes available again in the KNX interface. But with the OLD name. So, something of the line/entity remains in the HA database.
My real problem: 4 of 8 KNX switches are not working anymore from HA. I find it hard to believe they are all broken. So I suppose that HA turns a switch on, but that in reality HA sends the control to an unknown device/address.
This is the reason why I want to delete completely this entity. To start again with the entity of the yaml file = with the correct address and a new name. And with no longer any reference to the old entity in HA.
What a great feature the “group monitor”. I was not aware of it’s existence.
Yes, I can see a true (or false) value which is sent to the 4 non-working switches.
But I cannot hear the click of the relay, nor do I see the manual toggle in the upside position. Where I can hear the click for the 4 working switches.
So what’s the verdict? Is the control part of these 4 switches of my Siemens 8 * 16A relays definitely death/broken?
You can check if the command are getting to the bus and not being filtered by using ETS lite (free for 3 months) or find ETS and run it without a dongle/license in demo mode. You only need the diagnostics page anyway
In ETS you can also send commands to the group addresses and see what happens.
If they dont even work then… then its time for a new actor