Temporal worker exits | Unable to load configuration: config file corrupted: Persistence.DataStores[default](value).Cassandra.Hosts: zero value

Community Support
1

Hi.
I am using postgres, not cassandra. The temporal_server container of my docker compose loads successfully. All services load successfully with the exception of temporal worker. I exec into temporal_server container and there is indeed config/docker.yaml file that looks fine, and specifies “postgres” plugin name under visibility:sql. Nowhere in my setup cassandra is mentioned. Here is my docker-compose.yml file. Thank you

version: '3.8'
services:
  postgres:
    container_name: tempo_db
    image: postgres:13
    environment:
      POSTGRES_USER: postgres
      POSTGRES_PASSWORD: postgres
      POSTGRES_DB: tempodb
    healthcheck:
      test: pg_isready -U postgres
      interval: 15s
      timeout: 5s
      retries: 5
    ports:
      - "5432:5432"
    volumes:
      - postgres_data:/var/lib/postgresql/data

  backend:
    container_name: tempo_backend
    build: ./backend
    command: uvicorn app.main:app --host 0.0.0.0 --reload
    environment:
      - DATABASE_URL=postgresql://postgres:postgres@postgres:5432/tempodb
    ports:
      - "8000:8000"
    depends_on:
      - postgres

  frontend:
    container_name: tempo_frontend 
    build: ./frontend
    ports:
      - "3000:3000"
    volumes:
      - ./frontend/src:/app/src
    depends_on:
      - backend

  temporal:
    command: schema-upgrade autosetup
    container_name: temporal_server
    depends_on:
      postgres:
        condition: service_healthy
    image: temporalio/auto-setup:1.21.5
    environment:
      DB: "postgres"
      DB_PORT: "5432"
      POSTGRES_USER: "postgres"
      POSTGRES_PWD: "postgres"
      POSTGRES_SEEDS: "postgres"
    ports:
      - "7233:7233"

  worker:
    container_name: tempo_worker
    image: temporalio/auto-setup:1.21.5
    command: python ./backend/app/workers/worker.py
    environment:
      - DATABASE_URL=postgresql://postgres:postgres@postgres:5432/tempodb
      - TEMPORAL_TASK_QUEUE=csv-process
      - TEMPORAL_NAMESPACE=test
      - TEMPORAL_ADDRESS=temporal:7233
    depends_on:
      - temporal

volumes:
  postgres_data:
2

The culprit was the image for worker service.

I used the same image for temporal service (correct) and the worker (incorrect). The worker needed its own image, in this case an image created from a simple Dockerfile so that python worker script can run.