EMHASS Docker persistently fetching HA data from old DuckDNS URL despite correct config

[BarneyJing]

Hello everyone,

I'm setting up EMHASS (Energy Management Optimization for Home Assistant) in a Docker container on a Raspberry Pi (Raspberry Pi OS) alongside my Home Assistant Docker container (on the same Pi, using --network host).

[Full disclosure: I am not that savvy with coding so have been using ChatGPT and Google Gemini to help which has got me a long way!]

My goal is to use EMHASS for energy optimization, including pre-conditioning my heating/cooling based on PJM day-ahead pricing (the Comed next day pricing API doesn't seem to work).

I'm encountering a very persistent and frustrating issue where EMHASS consistently tries to fetch Home Assistant data from an old, external DuckDNS URL (https://myhass.duckdns.org/), causing my PJM forecast price sensor to remain unavailable. This happens despite all my configuration files correctly pointing to the local IP.

My Setup:

• Home Assistant: Docker container on Raspberry Pi (my Pi's local IP:8123)
• EMHASS: Docker container on the same Raspberry Pi (my Pi's local IP:5000), using --network host

Problem:

1. My PJM forecast price sensor in Home Assistant is unavailable. I have confirmed that the PJM data is flowing through to my Pi.
2. The EMHASS Docker logs consistently show this line during startup and operation:
   [INFO] Home Assistant data fetch will be performed using url: https://myhass.duckdns.org/
3. This is causing EMHASS to fail connecting to Home Assistant locally.

What I've already tried (extensive troubleshooting):

• Verified config.json content:

  • Manually created/edited /path/to/emhass_config/config.json.
  • Confirmed it is syntactically valid JSON.
  • Confirmed it explicitly contains "hass_url": "http://my Pi's local IP:8123" and my correct long_lived_token.
  • Example config.json section (note: actual token string is present, not placeholder):

    "hass_url": "http://my Pi's local IP:8123",
    "long_lived_token": "MY TOKEN STRING",
    "time_zone": "America/MY LOCATION",
    "Latitude": X.XX,
    "Longitude": X.XX,
    "Altitude": X.X,
    "solar_forecast_kwp": X.XX

• Verified secrets_emhass.yaml content (and then consolidated):

  • Initially, secrets_emhass.yaml also contained the correct hass_url and token.
  • To eliminate any potential conflict or precedence issues, I have since consolidated all relevant parameters into config.json and deleted secrets_emhass.yaml.

• Docker Container Management:

  • Always perform sudo docker stop emhass_optimizer && sudo docker rm emhass_optimizer before restarting.

• Docker Image Freshness:

  • Deleted the local Docker image (sudo docker rmi ghcr.io/davidusb-geek/emhass:latest) to force a fresh download.

• Host Volume Permissions & Content:

  • Ensured /path/to/emhass_config/ directory has drwxrwxrwx (777) permissions.
  • Confirmed config.json is a file (-rw-r--r--) and owned by the user running Docker.
  • Performed full content wipe of emhass_config directory (sudo rm -rf /path/to/emhass_config/*) and re-created contents from scratch (manual config.json).

• EMHASS UI Save Attempts:

  • Attempted to save the configuration from the EMHASS web UI (http://my Pi's local IP:5000/configuration) multiple times, both with and without config.json initially mounted. The config.json file was either not created persistently or created as a directory. Manually creating the file was the only way to get it as a file.

• Home Assistant Restart: Performed full Home Assistant Core restarts.

• Browser Cache: Tried hard refreshing browser cache and using Incognito mode.

Relevant Log Snippet (from sudo docker logs emhass_optimizer --tail 100):

   Building emhass @ file:///app
     Built emhass @ file:///app
Uninstalled 1 package in 50ms
Installed 1 package in 2ms
[2025-07-29 02:06:39 +0000] [22] [INFO] Starting gunicorn 23.0.0
[2025-07-29 02:06:39 +0000] [22] [INFO] Listening at: http://0.0.0.0:5000 (22)
[2025-07-29 02:06:39 +0000] [22] [INFO] Using worker: gthread
[2025-07-29 02:06:39 +0000] [23] [INFO] Booting worker with pid: 23
[2025-07-29 02:06:48 +0000] [23] [INFO] Obtaining parameters from config.json:
[2025-07-29 02:06:48 +0000] [23] [WARNING] Some secret parameters values are still matching their defaults
[2025-07-29 02:06:48 +0000] [23] [INFO] Launching the emhass webserver at: http://0.0.0.0:5000
[2025-07-29 02:06:48 +0000] [23] [INFO] Home Assistant data fetch will be performed using url: https://myhass.duckdns.org/
[2025-07-29 02:06:48 +0000] [23] [INFO] The data path is: /data
[2025-07-29 02:06:48 +0000] [23] [INFO] The logging is: INFO
[2025-07-29 02:06:48 +0000] [23] [INFO] Using core emhass version: 0.13.3

This persistent https://myhass.duckdns.org/ URL is the main blocker. Has anyone encountered a similar issue with EMHASS Docker latest image (version 0.13.3) where it seems to ignore hass_url in config.json or falls back to an old URL?

Any insights or suggestions would be greatly appreciated!

---

[radjanoonan]

I had the same issue. I got it running using this commandline: docker run -it --restart always -p 5000:5000 --name emhass-container -v ~/emhass/config.json:/share/config.json -v ~/emhass/secrets_emhass.yaml:/app/secrets_emhass.yaml ghcr.io/davidusb-geek/emhass

I subsituted "./emhass" with "~/emhass"