Recently my Node Red inject node has stopped injecting. Restarting Home Assistant doesn’t change anything. Any tips?
Edit- My debug nodes are also not debugging… But my flows still work.
Remove the nodes and add new ones.
It has only happened with debug nodes for me but I guess it could happen with inject also.
-
Make sure the node is enabled in the bottom of the settings page.
-
Make sure that you deploy all nodes when you press the deploy button.
-
If HA has been updated, then there might be a schema update included, which will present itself with a yellow edge on affected nodes and a “update all nodes”, but the button might disappear if you deploy without pressing it and then only the yellow edge of affected nodes are showing the sign of a needed update. Nodes are updated by editing them and then pressing save and finally deploying.
-
Make sure your palettes are updated.
-
Check the logs for errors.
Tried all that. The issue is intermittent. Inject and Debug nodes were working a few minutes ago, and now they aren’t. The add-on logs don’t show anything that would lead to this.
I’ve tried restarting the add-on.
Restarting Home Assistant,
A different browser.
Nothing seems to work.
Five minutes later, they are working again…
Do you only have only one HA server defined? Any unused config nodes? Are you using direct access ha_ip:1880
Thanks for the reply.
I do have a “sandbox” HA server, but it doesn’t have the Node-Red add-on installed.
I don’t see any unused config nodes.
I’ve tried bot the MDNS and IP address- no difference. I also tried on another browser.
What is most puzzling is the intermittence of the problem. This morning when I first opened the add-on, the inject buttons were non-responsive. I left the window open and read some emails, and when I returned, the inject buttons were live.
I opened the add-on on another PC and it starts with the inject button (and debug output) are also not working there.
Very frustrating.
Next time it happens try disabling a whole flow page and deploying. If it’s a node on the page you disable the inject and debug should come back on deploy.
It has been seen before that with multiple HA servers the mDNS corrupted, especially if you do not have renamed the mDNS name on one of the servers in the HA settings if you run HAOS or in the hostOS, if you run another installation type.
My sandbox is named “sandbox”, and as I mentioned there is no Node-Red add-on on the sandbox instance. But since I rarely play in the sandbox, I shut it down.
I can’t imagine that mDNS is the problem because I see the same problem whether I use .local or IP.
Worse, it’s random. In another thread I am looking at an unexplained call-service warning, and most of the time a deploy didn’t make the inject node (and debug node) problem present again. And the last time I deployed, the inject/debug nodes were unavailable. Ten minutes later, they are back.
It’s not just this page. It’s on every flow that has an inject node, which is most of them. The first indication is the warnings from some of my nodes that come up in the debug window with every deploy. If my deploy finishes, I usually get a handful of warnings in the debug window. When it is blank, I know that the inject nodes are also inactive.
I’ve tried multiple deploys, but no change.
I get that. My thinking is that this is being caused by a misbehaving node, like that call service. If you can narrow down the page. Then isolate the node/s. Which brings me to this
The page/pages these nodes are on, I’d look at first. Does the deploy error give you the option to view nodes? Try using the option, deploy all nodes to trigger the error.
There is no deploy error. When I deploy there are a bunch of specific device warnings and errors like entity not found because, for example, the device happens to be offline or the device only triggers its state every few seconds.
I think we’re talking about two unrelated problems. Or not. The inject/debug unavailable has persisted for a few weeks, and the call-service warning is new. That’s why I made a second thread.
In my experience, a misconfigured node is quite obvious and deploy would then flagged an error.
We are talking about the same problem. The errors that you get correspond with the debug and inject not working, correct? It’s these errors that are knocking out the debug/inject.
Something along the lines of the node with the offline entity is causing problems but only when the device is offline. If that is the problem, if you disable the page containing the node, upon deploy the inject/debug should start working.
If it does then re enable the page, deploy and see if the inject/debug stop working again.
If it were only that binary. A couple of hours ago I was deploying a few times with no problem, then after a deploy the inject/debug nodes were unavailable.
I’ve deleted the node that was giving the warning, so we will see if that was the problem.
Mines the same… anyone made any progress with this?
Thank you?
At least I am not the only one with this problem. It is really frustrating when I can’t use inject or debug nodes to test things.
This is what I get when I first start Node Red:
The buttons are greyed out and unusable.
I’ve tried everything except the nuclear option.
- Forcing a full deploy
- Different browsers
- Different PC
- Deleting any nodes that give a warning in the logs
Maybe today is the day to kaboom.
(That means backing up my flows, removing and reinstalling the add-on then importing my flows back into Node Red.)
Good luck with that, mine is exactly that unless I use port 1880 - then it works. Drives me nuts though! Why did this start happening
Very interesting. I hadn’t tried that, but it does work here as well.
one of my first posts in the thread
UPDATE
Running Node-Red from the “Show in Sidebar” does not work.
Open WebUI from the add-on configuration page does not work.
Running Node-Red as a Web Page Dashboard does work.
what happens if you run your HA in inkognito mode?
Does it work then?