I imported a few nodes that had a Home Assistant node in it to help someone understand his flow. Did that, then deleted the nodes. But now, I am getting this error: Is there ANY way to determine what node could be trying to connect to duckdns? I have looked at every node in every flow and can't find the offending node. I even tried rebooting the Raspberry this instance of Node-red is on. I went to "Selected Nodes" in the Debug tab and unticked everything, but I am still getting this error. Debug is useless as I am getting this error every second.
8/4/2018, 11:34:44 PM Home Assistant
msg : string
"Home assistant connection failed with error: Connection to home assistant could not be established with config: https://MyDomain.duckdns.org:8123 <password redacted>"
When you deploy do you get a message about unused configuration nodes? If so then that is the problem.
As Colin suggests, you may still have a HA config node lurking in the background.
If you open the configuration nodes sidebar from the drop-down menu, look for any home assistant nodes in there. If there is one, double click it and hit the delete button to remove it.
Thanks Colin and Knolleary- This is one of the first tings I thought of, but there is no 'unused configuration nodes' warning on deploy, and the only 'HomeAssistant' nodes are my own:
I could try removing all 32 of my Home Assistant nodes and reinstall them, but there's no guarantee that this would work. The worst case would be to remove Node-Red from Home Assistant and reinstall NR from scratch. But if I just export/import my flows, will this phantom come back?
Where else can I look?
You could search your nodes (using menu option) for MyDomain to see if any of your nodes contain that. Or open your flow file in a text editor and do likewise.
Ah, I misunderstood, I thought you had deleted all HA nodes. I presume therefore that you did have your domain at duckdns configured at one point.
Have you tried restarting node-red? Perhaps deleting a node has left something running.
First, I did reboot the Raspberry at the command line - no change.
I did export everything and look at the flows text file for the offending link- nothing there. (Also, when I selected "All Nodes" and deselected everything, the error continued).
One last try- I did a cold-boot. Shutdown from the command line then unplug the power to the Pi.
Plug-in, reboot, and now, the offending message is gone. Apparently there was something in RAM that survived a warm-reboot.
Thanks for the assistance- it is greatly appreciated. Note that this Node-red is running on hass.io, not Raspian. The last time this happened with Node-red on my Raspian, deleting the offending flow or node always fixed it. So you can appreciate why this one was so confounding.
(And, BTW, I've never used DuckDNS).
When you said the error contains "MyDomain.duckdns.org:8123" was that literally MyDomain or was it your domain? I assumed the latter. I now suspect it was the former. I imagine this is some default in something somewhere.
I have great difficulty understanding how a reboot without power cycle had a different effect in this respect to a power cycle. I think I will steer clear of hass.io if (as you suggest) it has a habit of presenting you with such issues.
Easy mistake. This was the address in the other user's Home Assistant node. Why it didn't go away when I deleted his flow mysterifies me. I have had other "oddities" in hass.io that only cleaned up with a cold boot, so it will likely be my default tool in that Pi.