[Bug]: New install cannot log in over local network and frontend seems incomplete.

root-reindeer-flotilla commented 1 month ago

What happened?

I installed quivr on my local windows machine with docker desktop. I got everything to work yesterday after a few hours. I decided to move my install to my server so I can access it over my network. I'm using proxmox to which I installed a debian container that has docker installed on it. I set up everything mostly the same but I'm getting this very weird frontend bug. I try to enter admin@quivr.app admin and login and nothing happens at all.

The supabase install works well, and even 10.10.10.101:5050/docs#/ works as I can visit them like normal.

Does anyone have any ideas?

Relevant log output

No response

Twitter / LinkedIn details

No response

linear[bot] commented 1 month ago

CORE-197 [Bug]: New install cannot log in over local network and frontend seems incomplete.

dosubot[bot] commented 1 month ago

Hello @root-reindeer-flotilla! I'm here to assist you with any bugs or questions you have about Quivr. Let's get this sorted out together!

I found a similar open issue that might help you resolve your problem: [Bug]: Not able to login. Dead home page. The suggested solution involves several steps:

Environment Configuration: Ensure that the NEXT_PUBLIC_FRONTEND_URL in your environment configuration matches your application's public URL, including the correct protocol and port. For your setup, this means including http:// and specifying port 3000 if that's what you're using.
Browser Settings: Make sure JavaScript is enabled in your browser. Check the browser's console for any JavaScript errors that might reveal more about the issue.
Network Requests: Use your browser's developer tools to monitor network requests when attempting to log in. Look for any requests that fail or return error responses.
Service Status: Confirm that all related services, including Docker containers and Supabase, are running correctly. This involves checking that Docker containers are up and that Supabase services are accessible.

Additionally, another issue might be related to the configuration of your environment variables in the .env file. Specifically, the NEXT_PUBLIC_BACKEND_URL and NEXT_PUBLIC_SUPABASE_URL should be set to the IP address of your server instead of localhost. Here is the corrected configuration:

NEXT_PUBLIC_BACKEND_URL=http://10.10.10.101:5050
NEXT_PUBLIC_SUPABASE_URL=http://10.10.10.101:54321

Make sure to update these values in your .env file and restart the services for the changes to take effect ^[1]^[2].

For the frontend appearing incomplete, ensure that all necessary environment variables are correctly set and that your Docker configuration is accurate. Verify the dockerfile path and build context in the docker-compose.yml file. Also, check the deployment logs for any errors during the build and deployment process ^[3].

By following these steps, you should be able to identify and resolve the issues preventing login and causing the frontend to appear incomplete.

To continue talking to Dosu, mention @dosu.

root-reindeer-flotilla commented 1 month ago

@dosu (I am aware this is an AI)

Changing those did not do anything.

I think you misunderstood. It's still on my local network and the internal IP address is 10.10.10.101. The backend/supabase IP is still local to the frontend. eg http://10.10.10.101:5050 aka http://localhost:5050.

My working theory is that docker isn't properly forwarding ports from the internal containers but I could be completely wrong. Any other ideas to try?

dosubot[bot] commented 1 month ago

The issue with Docker not properly forwarding ports from the internal containers, preventing login and causing the frontend to appear incomplete, could be due to several reasons. Here are some potential causes and solutions:

Port Configuration: Ensure that the ports are correctly mapped and not being blocked by any firewall or network settings. The docker-compose.yml file shows the following port mappings:
- Frontend: 3000:3000
- Backend API: 5050:5050
- Flower: 5555:5555
- Redis: 6379:6379
- Debug Port: 5678:5678 (from docker-compose.dev.yml)
Environment Variables: The frontend service relies on several environment variables (NEXT_PUBLIC_BACKEND_URL, NEXT_PUBLIC_SUPABASE_URL, etc.). Ensure these variables are correctly set and accessible.
Network Configuration: The services are connected via a custom bridge network (quivr-network). Ensure that this network is properly created and that there are no conflicts with other networks.
Service Dependencies: The frontend service depends on the backend API service. Ensure that the backend API service is running and healthy. The health check for the backend API is configured to check http://localhost:5050/healthz.
Docker Host Configuration: The extra_hosts configuration is used to map host.docker.internal to host-gateway. Ensure that this mapping is correct and that the host gateway is accessible.
Volume Mounts: Ensure that the volume mounts are correctly set up, especially for services like backend-api, notifier, worker, beat, and flower, which have volumes mounted from the host to the container. This ensures that the latest code changes are reflected inside the containers.

Here is the relevant part of the docker-compose.yml and docker-compose.dev.yml files for reference:

services:
  frontend:
    image: quivr-frontend-prebuilt
    pull_policy: never
    build:
      context: frontend
      dockerfile: Dockerfile
      args:
        - NEXT_PUBLIC_ENV=local
        - NEXT_PUBLIC_BACKEND_URL=${NEXT_PUBLIC_BACKEND_URL}
        - NEXT_PUBLIC_SUPABASE_URL=${NEXT_PUBLIC_SUPABASE_URL}
        - NEXT_PUBLIC_SUPABASE_ANON_KEY=${NEXT_PUBLIC_SUPABASE_ANON_KEY}
        - NEXT_PUBLIC_CMS_URL=${NEXT_PUBLIC_CMS_URL}
        - NEXT_PUBLIC_FRONTEND_URL=${NEXT_PUBLIC_FRONTEND_URL}
        - NEXT_PUBLIC_AUTH_MODES=${NEXT_PUBLIC_AUTH_MODES}
    container_name: web
    depends_on:
      - backend-api
    restart: always
    ports:
      - 3000:3000

  backend-api:
    image: quivr-backend-api:latest
    pull_policy: if_not_present
    env_file:
      - .env
    build:
      context: backend
      dockerfile: Dockerfile
    container_name: backend-api
    extra_hosts:
      - "host.docker.internal:host-gateway"
    healthcheck:
      test: [ "CMD", "curl", "http://localhost:5050/healthz" ]
    command: >
      /bin/bash -c "python -m uvicorn quivr_api.main:app --host 0.0.0.0 --log-level info --reload --port 5050 --loop uvloop"
    restart: always
    ports:
      - 5050:5050

  notifier:
    image: backend-api:latest
    extra_hosts:
      - "host.docker.internal:host-gateway"
    env_file:
      - .env
    container_name: notifier
    volumes:
      - ./backend/:/app/
    command: >
      /bin/bash -c "python /app/worker/quivr_worker/celery_monitor.py"
    restart: always
    depends_on:
      - redis
      - worker

  worker:
    image: backend-api:latest
    extra_hosts:
      - "host.docker.internal:host-gateway"
    env_file:
      - .env
    container_name: worker
    volumes:
      - ./backend/:/app/
    command: >
      /bin/bash -c "python -m celery -A quivr_worker.celery_worker worker -l info -E"
    restart: always
    depends_on:
      - redis

  redis:
    image: redis:latest
    container_name: redis
    extra_hosts:
      - "host.docker.internal:host-gateway"
    restart: always
    ports:
      - 6379:6379

  beat:
    image: backend-api:latest
    extra_hosts:
      - "host.docker.internal:host-gateway"
    env_file:
      - .env
    container_name: beat
    command: >
      /bin/bash -c "python -m celery -A quivr_worker.celery_worker beat -l info"
    volumes:
      - ./backend/:/app/
    restart: always
    depends_on:
      - redis

  flower:
    image: backend-api:latest
    extra_hosts:
      - "host.docker.internal:host-gateway"
    env_file:
      - .env
    container_name: flower
    command: >
      /bin/bash -c "python -m celery -A quivr_worker.celery_worker flower -l info --port=5555"
    volumes:
      - ./backend/:/app/
    restart: always
    depends_on:
      - redis
      - worker
      - beat
    ports:
      - 5555:5555

Check these configurations and ensure that all services are running correctly and that the ports are accessible from your local network setup with IP 10.10.10.101 ^[1]^[2].

To continue talking to Dosu, mention @dosu.

aishock commented 1 month ago

Hi, you mentioned that you once got everything worked on your local machine, do you mind sharing more details of what models you used (such as ChatGPT, Ollama) and how you add in the models to Quivr?

QuivrHQ / quivr