GitLab CI Docker WORKDIR not being created - docker

Question

I am trying to deploy my NodeJS repo to a DO droplet via GitLab CI. I have been following this guide to do so. What is odd is that the deployment pipeline seems to succeed but if I SSH into the box, I can see that the app is not running as has failed to find a package.json in /usr/src/app which is the WORKDIR my Dockerfile is pointing to.
gitlab-ci.yml
cache:
key: "${CI_COMMIT_REF_NAME} node:latest"
paths:
- node_modules/
- .yarn
stages:
- build
- release
- deploy
build:
stage: build
image: node:latest
script:
- yarn
artifacts:
paths:
- node_modules/
release:
stage: release
image: docker:latest
only:
- master
services:
- docker:dind
variables:
DOCKER_DRIVER: "overlay"
before_script:
- docker version
- docker info
- docker login -u ${CI_REGISTRY_USER} -p ${CI_BUILD_TOKEN} ${CI_REGISTRY}
script:
- docker build -t ${CI_REGISTRY}/${CI_PROJECT_PATH}:latest --pull .
- docker push ${CI_REGISTRY}/${CI_PROJECT_PATH}:latest
after_script:
- docker logout ${CI_REGISTRY}
deploy:
stage: deploy
image: gitlab/dind:latest
only:
- master
environment: production
when: manual
before_script:
- mkdir -p ~/.ssh
- echo "${DEPLOY_SERVER_PRIVATE_KEY}" | tr -d '\r' > ~/.ssh/id_rsa
- chmod 600 ~/.ssh/id_rsa
- eval "$(ssh-agent -s)"
- ssh-add ~/.ssh/id_rsa
- ssh-keyscan -H ${DEPLOYMENT_SERVER_IP} >> ~/.ssh/known_hosts
script:
- printf "DB_URL=${DB_URL}\nDB_NAME=${DB_NAME}\nPORT=3000" > .env
- scp -r ./.env ./docker-compose.yml root#${DEPLOYMENT_SERVER_IP}:~/
- ssh root#${DEPLOYMENT_SERVER_IP} "docker login -u ${CI_REGISTRY_USER} -p ${CI_REGISTRY_PASSWORD} ${CI_REGISTRY}; docker-compose rm -sf scraper; docker pull ${CI_REGISTRY}/${CI_PROJECT_PATH}:latest; docker-compose up -d"
Dockerfile
FROM node:10
WORKDIR /usr/src/app
COPY package.json ./
RUN yarn
COPY . .
EXPOSE 3000
CMD [ "yarn", "start" ]
docker-compose.yml
version: "3"
services:
scraper:
build: .
image: registry.gitlab.com/arby-better/scraper:latest
volumes:
- .:/usr/src/app
- /usr/src/app/node_modules
ports:
- 3000:3000
environment:
- NODE_ENV=production
env_file:
- .env
I'm using GitLab Shared Runners for my pipeline. My pipeline looks like it executes completely fine but for this symlink failure at the end:
...which I don't think is anything to worry about. If I SSH into my box & go to where the docker compose was copied & inspect:
Docker has not created /usr/src/app.
Versions:
Docker: 19.03.1
Docker-compose: 1.22.0
My DO box is Docker 1-click btw. Any help appreciated!
EDIT
I have altered my Dockerfile to attempt to force the dir creation so have added RUN mkdir -p /usr/src/app before the line declaring it as the working dir. This still does not create the directory...
When I look at the container status' (docker-compose ps), I can see that the containers are in an exit state & have exited with code either 1 or 254...any idea as to why?

Answer 1

Your compose file is designed for a development environment, where the code directory is replaced by a volume mount to the code on the developers machine. You don't have this persistent directory in production, nor should you be depending on code outside of the image in production, defeating the purpose of copying it into your image.
version: "3"
services:
scraper:
build: .
image: registry.gitlab.com/arby-better/scraper:latest
# Comment out or delete these lines, they do not belong in production
#volumes:
# - .:/usr/src/app
# - /usr/src/app/node_modules
ports:
- 3000:3000
environment:
- NODE_ENV=production
env_file:
- .env