"Call-Service attempted without connection to server."

[balens]

Hi,

I know there’s another thread that deals with this error but my problem is a bit different.

I’m new to Node-red and have been playing with it. And everything seems to work fine except that this message appears in my debugwindow every now and again;

“Call-Service attempted without connection to server.”

This is often when I use the ‘inject’ to test some sequences. Unlike the other topic, after a few clicks he gets trough and it works. Mainly after clicking deploy it doesn’t work for a bit but also when I start testing after a day.

This is my configuration in Node-red:

certfile: fullchain.pem
credential_secret: ****
http_node:
  password: ''
  username: ''
http_static:
  password: ''
  username: ''
init_commands: []
keyfile: privkey.pem
npm_packages: []
ssl: false
system_packages: []
theme: default

[Mikefila]

In the call service node ther is a drop down called server. Is there more than one home assistant there? Can you recreate the error, then on the nodered addon log page post the output.

[balens]

Hey! Thanks for your reply.

No, there’s only the one. I’ve been trying to recreate it but somehow it didn’t occur anymore… A bit awkward after my call for help. I’ll keep an eye out if it happens again I guess.

[pfurrie (Pat Furrie)]

Okay, here is my “Call-Service attempted without connection to server” problem:

“Call-Service attempted without connection to server.” That’s what I get from one of the nodes in a sequence every time I deploy. Node-Red is set up as an add-on to Home Assistant. The Node-Red version is current as of 10:20 am on 2/26/2022 (I uninstalled and re-installed Node-Red a few minutes ago).

• The only node causing the error is the “light off” call-service node.

• There is only one instance of Home Assistant showing up in the nodes.

• If I click the “off” inject node, the “light off” node does not throw an error, and it does the assigned task (turns off the light). That would suggest that the “light off” call-service node is properly configured.

• Both the light timer node and the “off” inject node are using the same topic and payload strings. (topic = “light” and payload = “off”)

• The timer node output is setup for “when state changes + startup” which is why it is firing everytime I deploy (during the daylight hours).

• “Dusk / Dawn” is set to “Only when dark.”

• The schedule has nothing set on it. Previously I had everything set on the schedule (24/7), but the problem happened in that configuration as well.

• Lat-Lon are set properly to my location, if that even matters to the “off” node (why would it?)

• If I flip the switch node so on-and-off are going to the opposite outputs, the “on” node now throws the error (since it is the one being triggered by the timer). If I trigger it using the inject nodes, the “on” node works perfectly fine with no errors.

What in the world is going on?

[Kermit]

If you have your NR deployment set to “Full” every time you deploy you cause NR to reconnect to HA. This is not instant. If you try and do a service call before NR has connected to HA you get the error “call-service attempted without connection to server” well because NR has not connected to HA yet.

[pfurrie (Pat Furrie)]

Yep, figured that out a little while ago. However note: this wasn’t me (directly) doing a service call, but rather the light scheduler node. It is set to output on changes AND startup. Ideally, NR would delay any reactivation of a sequence until after it (NR) has completed its connection chores. What we currently have here is a race condition.

[Kermit]

That’s why queuing was added to the call service node.

[Daniel_Liebst]

and how am i supposed to use queuing in the case that i get the same error for a call service, try and explain what its supposed to do, so maybe i can resolve my issue.

this is what my error looks like in NR.

and the debug window says “Call-Service attempted without connection to server.”

[Mikefila]

in the call service server drop down list is there more than one entry? there should only be one.

[Daniel_Liebst]

I have only one home assistant instance there yes.

[balens]

Well, for me somehow the problem dissappeared. I didn’t do anything but it just stopped being a problem.

[heastwood]

Did anyone ever figure out the issue or did it just start working? I uninstalled NR and fully re-installed my instance. Fought with it for about an hour, and it did just start working for me. Not sure the magic sequence that got it to work, just wondering if others were/are still having ‘connection to server’ issues.

[balens]

With me it just started working m8! No more issues on my part!

[jwadsley]

I’m having this issue with call service either to turn on or off lights / switches. I only have the one instance of Home Assistant in the Server drop down settings.

I’m pretty new to Node Red so not sure at all why this is occuring for some flows and not others? Seems to only be an issue with lights, not switches

[reypm]

I am running exactly into the same issue and have no clues into what could be wrong since the next node is being executed. This only happens each time I hit the “Deploy” button but is bothering me and adding a lot of non-useful logs to the debug area any ideas?

[MaxK (Mark)]

You need to wait until node red is ready (something like this)

Node Red Integration - how to avoid "NoConnectionError: No connection to Home Assistant"

It appears that an event is fired when HA is ready, so I have a set of nodes that listens for a home_assistant_client events, and then if the payload is services_loaded, HA and Node-RED are ready to go. I use the nodes below and just have a switch on the payload, though I don’t do anything with the switch. It then goes to a link out and I use a link in for other nodes that need to know when HA is ready. [{"id":"f44c410e38cbeeea","type":"switch","z":"a42ea764.b87b98","name":"State","property":"p…

More documentation about client events here: Events: all | node-red-contrib-home-assistant-websocket

And/or set the the queue attribute as described above.

[steerage250 (Ian Steer)]

I have 5 instances of “home assistant” in the Server configuration for Call Service nodes. I’m sure this is not good, but don’t know how to fix it.

[4romany (Romany)]

Old thread I know but Kermit earns a thank you sticker for this posting! I was seeing that error and his explanation explains it all!

[ebeall]

It works when I Change to “first messages only”

[stevemann (Stephen Mann (YAML-challenged))]

@steerage250

I have 5 instances of “home assistant” in the Server configuration for Call Service nodes. I’m sure this is not good, but don’t know how to fix it.

Not a problem. Every time you copy/paste a call-service node or MQTT node, you get a new instance in the server configuration. The fix is easy but annoying. Just edit the call service node, click on “server” and select the first one.

If you open the Configuration Nodes panel in Node Red, you can see how many nodes use the server:

When that number gets to zero you can safely delete the server node:

But it isn’t necessary. In my case I have about two-dozen MQTT server nodes.