"Error: unknown datapoint BidCos-RF.OEQxxxxxxx:1.STATE"

[geekawhyte]

Hi there,

I have been using Homematic thermostats (HM-CC-RT-DN) and window contacts (HM-Sec-SC-2) under Node-RED on a Raspberry Pi 3 Model B with piVCCU as CCU2 and the radio module HM-MOD-RPI-PCB for a long time .

However, it wasn't until this weekend that I switched my flows from the old nodes node-red-contrib-homematic and node-red-contrib-homematic-rpc to node-red-contrib-ccu. That worked very well in itself and I was even able to streamline my flows.

But I still have one problem. When Node-RED is restarted, the current status of the window contacts should be displayed, i.e. whether they are open or closed. To do this, I created a get Value node with the corresponding channel and the STATE datapoint and let it trigger via a time-controlled inject. Then the debug message appears: "Error: unknown datapoint BidCos-RF.OEQxxxxxxx: 1.STATE". If I open and close the window once and execute the node or inject again by hand, the error does not recur. So I would first have to open and close all windows so that the datapoint is known.

Is there a better solution, for example to somehow update the list of datapoints or save it to anywhere? Or is it a bug?

Thanks and best regards

[stuckis6er]

hello, is the problem resolved if you change the start time from 0.1 to 1 in the inject node?

[geekawhyte]

Hello, unfortunately no, because I've already set start time to 5 seconds.

[czacherl]

Hi there,
I just transferred my RaspberryMatic to a cleaner installation of pivccu3 and addes the latest node red installation, not using the RedMatic addon on the ccu installation.
Before the change, my ccu nodes were working fine. Since the change, I have experienced exactly the same errors as mentioned here and in the other tickets.
My BidCos-RF and Hm-IP endpoints don't bring any data as it's failing with "unknown endpoint"
When trying to change the setup, the nodes don't bring up the channel name anymore. I think, that something has changed in the ccu setup, so the node red contrib ccu is not getting the right values any longer and that way, the communication fails.
Could you please continue your work on this, as that ways, the node-red-contrib-ccu is completely unusable.
PS: My ccu-connection is setup like the following picture:

No authentication, etc. I have a working connection. Both the node red and the pivccu installation run on the same raspberry

[ptweety]

Have you seen the documentation re/piVCCU here: https://github.com/rdmtc/node-red-contrib-ccu#pivccu?

What are your IPs for piVCCU and the host system?

[czacherl]

Hi @ptweety, thanks for that link again. I had read it but entered one false port in my firewall settings for the 2010 tcp.
After changing that, I can call the nodes and set values again.

Though, two things still are not working:

• I cannot receive events from my door sensor
• I cannot read any "names", just IDs for all my devices

Just to confirm my setup, is it enough to open the ports 2048, 2049, 2001 and 2010 or do I have to add the pivCCU-IP to the firewall settings, too?

My IPs: 192.168.178.58 ist my pivCCU and 192.168.178.42:1880 is my nodered

[ptweety]

I would recommend to use these IP address settings:

192.168.178.1;
192.168.178.0/24;
fc00::/7;
fe80::/10

and the following port settings:

2000;
2001;
2010;
8181;
9292;
9293

That should cover everything you need.

You can also activate

• ReGaHSS
• VirtualDevices

and change the listen address to 192.168.178.42.

[czacherl]

Thanks @ptweety, that did the trick.
The 9292 and 9293 are not mentioned in the documentation for the pivccu. Is that what made it run or is it the ips?
Anyhow... It took about 10 minutes and two reboots of the nodered until I got the names back again. Now everything is fine, even the rpc events started working again.
Thanks for your help.