Docker ignoring entrypoint scripts

Build Environment
1

Hello, I’m new to CircleCI (and this community), thanks for having me!

I have a docker-compose.yml that, among other things, spins up a postgres database and then runs an initialization script, which is mounted in /docker-entrypoint-initdb.d. In that directory is a shell script that creates some roles and extensions, and then invokes the SQL files mounted as below. This works beautifully on my local Linux laptop (and on a coworkers mac for example), but in CircleCI it fails because the script is ignored:

…
“kalidasa-db-1 | /usr/local/bin/docker-entrypoint.sh: ignoring /docker-entrypoint-initdb.d/*”
…

I’ve spent quite a while looking for reasons that Docker would ignore scripts in that directory, and have explored things like:

  1. Explicitly setting execute permission
  2. Using an “ENTRYPOINT” directive in the config to run the script explicitly
  3. Using a machine executor instead of docker
  4. Mounting the script itself instead of the directory it lives in
  5. Different versions of docker compose etc.

I’ve been through the docs and have searched discussions..could be I’ve missed something but I’m not sure if the discrepancy is more Docker or CircleCI related. Again, it has consistently worked locally. Here is the snippet from the docker-compose.yml

kalidasa-db:
    image: postgis/postgis:13-3.5
    ports:
      - "5434:5432"
    volumes:
      - ./kalidasa-db:/docker-entrypoint-initdb.d
      - ./db/kalidasa-roles.sql:/tmp/01-roles.sql
      - ./data/kalidasa-full.sql:/tmp/kalidasa-full.sql
    environment:
      - POSTGRES_USER=postgres
      - POSTGRES_PASSWORD=postgres
      - POSTGRES_DB=postgres
      - POSTGRES_INITDB_ARGS=--encoding=UTF8 --auth-local=trust --auth-host=trust
      - INIT_CHECK_TABLE=clients
    healthcheck:
      test: ["CMD-SHELL", "pg_isready -h localhost -p 5432 -U understory -d kalidasa && psql -U understory -d kalidasa -c \"SELECT 1 FROM information_schema.tables WHERE table_name = 'clients'\" | grep -q 1"]
      interval: 5s
      timeout: 5s
      retries: 10

…and the relevant snippet from the CircleCI config:

jobs:
  docker-purge:
      docker:  
        - image: cimg/base:2020.01  
      parallelism: 30
      steps:
        - setup_remote_docker
        - run: docker volume prune -f

  integration_test:
    docker:  
      - image: cimg/openjdk:8.0
    steps:
      - checkout
      - restore_cache:
          keys:
          - v2-dependencies-{{ checksum "pom.xml" }}
      - setup_remote_docker:
          docker_layer_caching: true
      - run:
          name: Install Docker Compose
          command: |
            sudo curl -L "https://github.com/docker/compose/releases/download/v2.23.3/docker-compose-$(uname -s)-$(uname -m)" -o /usr/local/bin/docker-compose
            sudo chmod +x /usr/local/bin/docker-compose
            docker-compose --version
      - run:
          name: Install Docker Compose plugin
          command: |
            DOCKER_CONFIG=${DOCKER_CONFIG:-$HOME/.docker}
            mkdir -p $DOCKER_CONFIG/cli-plugins
            curl -SL "https://github.com/docker/compose/releases/download/v2.23.3/docker-compose-$(uname -s)-$(uname -m)" -o $DOCKER_CONFIG/cli-plugins/docker-compose
            chmod +x $DOCKER_CONFIG/cli-plugins/docker-compose
            docker compose version
      - run:
          name: Run integration tests
          command: docker compose up
      - run:
          name: Cleanup
          command: docker compose down -v
          when: always

Any insight or tips would be very appreciated!