I need to explain the topic summary: I have an existing HASS setup running HAOS as a VM in Xen. It’s working fine, but I need to repurpose that VM as something else. So, I’ve set up a new Dockerized HA in my kubernetes cluster. Since I can’t use AddOns ( since I’ve no supervisord, now), I also have a Node-Red container set up in the same namespace. So, big change in infrastructure.
My issue is: I have a bunch of NR flows that I’ve exported from my OLD HA environment and have imported into the new NR pods successfully (I think), at least all the flows are there and appear intact. I HAVE installed the HA community palette items and they all are there. I’ve updated the HA config in NR from “I’m using the AddOn” to a defined the external HA server via URL/TOKEN. In the new setup, NR can’t see any HA domains/entities/devices/services/etc. As in I can search for something simple like “binary_sensor.back_deck_motion” which DOES exist in HA, but in NR when trying to add that sensor to a flow, NR says it can’t find the sensor, so the flow is always invalid. Back in HA, I’ve tried installing the Node-Red Companion via HACS, but it shows nothing; no name, no devices or entities.
I don’t see any errors in the HASS raw logs, nor anything (other than the heartbeat output to the HA instance every 30 seconds) in the Node Red logs.
Is this expected behavior? Or, any suggestions on where to look, or what to even look for?
Is Node Red also running inside your K8S cluster, and in the same namespace?
If not, did you expose HA to outside your cluster (Nodeport / ingress / …)
If so, are you using the proper hostname / port in the NR config of HA?
Yes, both are in the same namespace AND exposed outside my cluster via both Ingress and LoadBalancer services (using MetalLB), so both have discreet LAN addresses.
I’ve tried using both internal cluster addressing (i.e. nodered.hass.service.cluster.local & hass.hass.service.cluster.local) and also via external DNS names issued to the MetalLB IPs. In both cases, nothing shows up in NodeRed from HomeAssistant.
Well, I’ll be honest… it WAS flooding lots of "Connecting to: https://hass..." up to about 5 minutes ago (I killed an restarted the pod), but now since restarting the pod, I get a single startup log: "Connecting to: https://hass..." followed by a single "Connected to: https://hass..." message. And I now see SOME of the HASS content, but it’s missing a majority of them. I might need to go check my discovery settings in HASS, since it seems to be missing entities and devices as well.
NR config is mounted from a ConfigMap, and that hasn’t changed once since deploying the pod; so I’m not sure what might have changed at this point.
OMFG!!! Some times I can be as dense as a concrete block! I just figured out what wasn’t working… I hadn’t opened up traffic in my firewall to allow my IoT VLAN access to the new HASS instance (although, not sure why NodeRed wasn’t able to retrieve entities from HASS due to that, since they are on the same subnet)… but once I enabled traffic, not only did MQTT populate like hell, it also started letting HA objects pop up in NR!!!