KNX node for node-red

[MGK]

Sorry for OT:

Wie hast du das hinbekommen, dass die Bridge nach einem weiteren Deploy aktiv bleibt...

Das ist ein bekannter bug in dem HomeKit Node. siehe https://github.com/NRCHKB/node-red-c...ged/issues/118
trifft aber nicht alle Benutzer, ich habe damit überhaupt keine Probleme bisher. Workaround: node-red neu starten.

Gruss,
Michael

[TheMax74]

Update:

Added some samples in the wiki https://github.com/Supergiovane/node...-ultimate/wiki

Added sample integration with
filedata/fetch?id=1449686&d=1578394049&type=thumb filedata/fetch?id=1449687&d=1578394049&type=thumb filedata/fetch?id=1449688&d=1578394049&type=thumb

[83nny]

Servus zusammen,

bis ich mich auf eine Visu festgelegt hab nutze ich momentan das Node Red Dashboard. Leider hab ich das ganze wohl nicht richtig umgesetzt.

Folgendes Beispiel:

Ich habe einen Slider für die Rollo-Position. Der Eingang des Sliders hört auf den Statuswert vom Jalousieaktor, der Ausgang setzt eine neue absolute Position.

Rollo.JPG

Problem:
Fahre ich das Rollo über den Taster mittels Befehl "AB" nach unten gibt der Jalousieaktor korrekterweise den Statuswert 100% zurück und der Slider fährt auf 100. Dadurch wird allerdings nochmal der Wert 100 auf den Bus gesendet (Payload hat sich ja geändert) und der Rollomotor macht nach der Verfahrzeit nochmal einen ruck.

Die Frage:
Hat jemand eine elegante Lösung um die Position nicht nochmal auf den Bus zu senden wenn der Wert vorher vom Bus kam?

Und Zusatzfrage:
Wieso macht der Aktor (MDT) überhaupt einen Mucks und versucht von 100% auf 100% zu kommen?

Gruß
Benny

[TheMax74]

Hello
if i understand correctly, you can try to activate the rbe filter in the INPUT section too, of your left hand knx-ultimate node in the picture.

[apoc4lyps]

Hi, I added a gate which is triggered by the output and locks the input for a certain amount of time (node-red-contrib-simple-gate + trigger as watchdog).

Annotation 2020-01-11 211227.jpg

[edelb]

Hi,

first-of-all: awesome node. Thanks for the contribution!!
I am using it to log all KNX datagrams to my InfluxDB.

However, there seems to be an issue that comes after rebooting the machine, node-red runs on (it is actually a Debian LXC-container, running in proxmox):
* Upon reboot of the machine, the KNX ultimate node starts in some kind of "waiting"-mode and doesn't read any info from KNX (see attached screenshot).
* Upon reloading all flows, the issue is gone (this is repeatable)
* Upon restarting node-red (using node-red-restart in shell) the node starts in working mode and is fully functional (this is also repeatable)

2020-01-27_22h28_10.png

Attached is my configuration (I edited the "host", "port" and "physAddr"-items as they are obviously not the cause of this issue):

[
{
"id": "4a389c09.f40834",
"type": "knxUltimate",
"z": "db4d9b06.5fe708",
"server": "d7632da7.4c3bf",
"topic": "",
"dpt": "",
"initialread": false,
"notifyreadrequest": false,
"notifyresponse": true,
"notifywrite": true,
"notifyreadrequestalsorespondtobus": false,
"notifyreadrequestalsorespondtobusdefaultvalueifno tinitialized": "0",
"listenallga": true,
"name": "MDT_IPR",
"outputtype": "write",
"outputRBE": false,
"inputRBE": false,
"x": 220,
"y": 140,
"wires": [
[
"55ffa7ad.cb5bb8",
"6ac4822e.75755c",
"cc532ae3.49fe7"
]
]
},
{
"id": "d7632da7.4c3bf",
"type": "knxUltimate-config",
"z": "",
"host": "MY_IP",
"port": "MY_PORT",
"physAddr": "MY_PA",
"suppressACKRequest": false,
"KNXEthInterface": "Auto",
"KNXEthInterfaceManuallyInput": "",
"statusDisplayLastUpdate": true,
"statusDisplayDeviceNameWhenALL": true,
"statusDisplayDataPoint": true
}
]

I am not 100% sure whether this is an issue with KNX Ultimate node at all, 'cause it seems to be somehow configured correctly.
Just want to make sure, that I do not overlook anything.

Thanks.
Felix


Edit: I found a workaround.
Added a watchdog logic which reloads the flows using the http API if no messages arrive from KNX:
2020-01-28_00h32_38.png
Seems to be working so far.
However, I would still be interested in finding the root cause of the behavior I observed.

Thanks again.
Felix

[TheMax74]

Hello Felix
thank you for using my node.
Are you using a knx/ip router or knx/ip interface?
How many eth interfaces have you configured in your container?
Anyway, it seems a problem with the ethernet port not coming up so quickly as the container.
Do you use PM2 to launch nodered? If so, there should be an option to delay the startup of nodered.
Try to delay it to, for example, 30 seconds after container start.

[edelb]

Hi TheMax74,

Anyway, it seems a problem with the ethernet port not coming up so quickly as the container.

Sounds plausible to me.

Just for documentation:
* I am using the MDT IP Router (nmot sure, which gen), but in ip tunneling mode (don't have so many connections yet).
* there is only one eth connection on the container (and actually nothing else besides nodered itself)
* I am starting NodeRed as a service using systemd (like described here: https://nodered.org/docs/getting-started/raspberrypi#running-as-a-service)
--> Not sure whether there is an option to start nodered with a delay (without some self-written scripting).

I would leave it that way now.
Thanks for your feedback

[TheMax74]

Disable the autostart on node-red.
Restart the container and leave it running alone for 2-3 minutes.
Start node-red manually.
If it works, my idea is right.

Then...
Please use ever IP Routing, because it’s the best way and the more reliable. You need to activate the routing option in the MDT configuration in ETS. Then set the multicast IP address in knx-ultimate config (instead of IP address of your MDT). In this manner, you’ll solve all your problems with ethernet device, because multicast routing uses UDP connectionless transport.
Let me know! ;-)

[edelb]

Disable the autostart on node-red.
Restart the container and leave it running alone for 2-3 minutes.
Start node-red manually.
If it works, my idea is right.

You were right, that worked.

Edit:

Please use ever IP Routing, because it’s the best way and the more reliable.

Did also try that for documentation purposes, didn't solve the issue.

However, the watchdog logic I implemented a few days ago seems to be solid. It is also watching for other kinds of errors (e.g. non node-red related), and can inform me via Email or so. So I am definitely going for that now.

[TheMax74]

New v.1.1.26 is out, containing the new WatchDog node.
https://github.com/Supergiovane/node...-Configuration
Ciao.

[Alloc]

Hi TheMax74 ,

is there some way to create a list of GAs that are used in the Node-RED flows? Ideally in some way that can be imported with a dummy node in ETS so that the IP router's filter tables can automatically be filled and I don't have to set it to forward everything?

Other than that so far it's pretty nice. Tried the ioBroker adapter first but had some troubles, this one works just flawlessly so far
Thanks for your hard work on this node.

Regards,
Chris

[TheMax74]

Hello Chris
thank you for using my node.
Yes it’s possible to “fake” the export list.
Please click on the github link below. Respect the tabulation of the fake list.
Expand the “Sample ETS csv file to paste into the ETS field of your config node.” below the Youtube image, and copy/paste it into kns-ultimate config node. Then change what you need
The only issue, could be the datapoint setting. You must adhere with ETS csv export standard notation.

[Alloc]

Hi Massimo,

thanks for the quick reply. I think you misunderstood my question though
I know I can get the GAs from my ETS project loaded into the knx-ultimate config so knx-ultimate knows about the project's GAs. What I want is the other direction: I have set up flows that use just a part of the GAs from my project and I want a list of those GAs used in Node-RED so I can set up the GA filters of the IP router.

Currently my IP router is set to forward *everything* which I think isn't ideal. But in order to set up a proper filter list I need to know what GAs are used in the Node-RED flows. Of course I can manually dig through all my flows and check every single KNX-node and write down the GAs used, but that can take quite some time obviously

So a button somewhere on the knx-ultimate config to get a list of used GAs that I can then feed to the ETS somehow to configure a dummy device would be nice.

Regards,
Chris

[TheMax74]

Ops... sorry! I’ve not understood.
So.... the config node has the visibility of all nodes, so this is possible.
What type of output do you need? A GA each row, a comma separated list of GA or what?