I'm new to docker swarm and am able to deploy my services on various nodes, however, the environment variables that are exported from a dockerfile ENTRYPOINT script are not setting for tasks that are deployed in the docker swarm cluster.
Setup
- docker version 18.09.1, build 4c52b90
- docker-compose version 1.23.2, build 1110ad01
- Django 2.1.5
- PosgresSQL 10
Trying to do a one-off command inside a django task, using docker exec -t CONTAINER_ID sh to get into the container, then executing python manage.py migrate, I get the following error:
Error
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "manage.py", line 38, in <module>
execute_from_command_line(sys.argv)
File "/usr/local/lib/python3.6/site-packages/django/core/management/__init__.py", line 381, in execute_from_command_line
utility.execute()
File "/usr/local/lib/python3.6/site-packages/django/core/management/__init__.py", line 375, in execute
self.fetch_command(subcommand).run_from_argv(self.argv)
File "/usr/local/lib/python3.6/site-packages/django/core/management/__init__.py", line 211, in fetch_command
settings.INSTALLED_APPS
File "/usr/local/lib/python3.6/site-packages/django/conf/__init__.py", line 57, in __getattr__
self._setup(name)
File "/usr/local/lib/python3.6/site-packages/django/conf/__init__.py", line 44, in _setup
self._wrapped = Settings(settings_module)
File "/usr/local/lib/python3.6/site-packages/django/conf/__init__.py", line 107, in __init__
mod = importlib.import_module(self.SETTINGS_MODULE)
File "/usr/local/lib/python3.6/importlib/__init__.py", line 126, in import_module
return _bootstrap._gcd_import(name[level:], package, level)
File "<frozen importlib._bootstrap>", line 994, in _gcd_import
File "<frozen importlib._bootstrap>", line 971, in _find_and_load
File "<frozen importlib._bootstrap>", line 955, in _find_and_load_unlocked
File "<frozen importlib._bootstrap>", line 665, in _load_unlocked
File "<frozen importlib._bootstrap_external>", line 678, in exec_module
File "<frozen importlib._bootstrap>", line 219, in _call_with_frames_removed
File "/app/config/settings/production.py", line 15, in <module>
DATABASES['default'] = env.db('DATABASE_URL') # noqa F405
File "/usr/local/lib/python3.6/site-packages/environ/environ.py", line 202, in db_url
return self.db_url_config(self.get_value(var, default=default), engine=engine)
File "/usr/local/lib/python3.6/site-packages/environ/environ.py", line 275, in get_value
raise ImproperlyConfigured(error_msg)
django.core.exceptions.ImproperlyConfigured: Set the DATABASE_URL environment variable
So, the DATABASE_URL is not set as an environment variable inside my docker container. As stated before, this is exported from an ENTRYPOINT script which is invoked in the dockerfile.
Dockerfile
FROM python:3.6-alpine
ENV PYTHONUNBUFFERED 1
RUN apk update \
# psycopg2 dependencies
&& apk add --virtual build-deps gcc python3-dev g++ musl-dev \
&& apk add postgresql-dev \
# Pillow dependencies
&& apk add jpeg-dev zlib-dev freetype-dev lcms2-dev openjpeg-dev tiff-dev tk-dev tcl-dev \
# CFFI dependencies
&& apk add libffi-dev py-cffi \
# Translations dependencies
&& apk add gettext \
# https://docs.djangoproject.com/en/dev/ref/django-admin/#dbshell
# Requirements are installed here to ensure they will be cached.
COPY ./requirements /requirements
RUN pip install -r /requirements/production.txt \
&& rm -rf /requirements
COPY ./compose/production/django/entrypoint /entrypoint
RUN sed -i 's/\r//' /entrypoint
RUN chmod +x /entrypoint
RUN chown django /entrypoint
COPY ./compose/production/django/start /start
RUN sed -i 's/\r//' /start
RUN chmod +x /start
RUN chown django /start
COPY ./compose/production/django/celery/worker/start /start-celeryworker
RUN sed -i 's/\r//' /start-celeryworker
RUN chmod +x /start-celeryworker
RUN chown django /start-celeryworker
COPY ./compose/production/django/celery/beat/start /start-celerybeat
RUN sed -i 's/\r//' /start-celerybeat
RUN chmod +x /start-celerybeat
RUN chown django /start-celerybeat
COPY ./compose/production/django/celery/flower/start /start-flower
RUN sed -i 's/\r//' /start-flower
RUN chmod +x /start-flower
COPY . /app
RUN chown -R django /app
USER django
WORKDIR /app
ENTRYPOINT ["/entrypoint"]
ENTRYPOINT Script
#!/bin/sh
set -o errexit
set -o pipefail
set -o nounset
# N.B. If only .env files supported variable expansion...
export CELERY_BROKER_URL="${REDIS_URL}"
if [ -z "${POSTGRES_USER}" ]; then
base_postgres_image_default_user='postgres'
export POSTGRES_USER="${base_postgres_image_default_user}"
fi
export DATABASE_URL="postgres://${POSTGRES_USER}:${POSTGRES_PASSWORD}@${POSTGRES_HOST}:${POSTGRES_PORT}/${POSTGRES_DB}"
postgres_ready() {
python << END
import sys
import psycopg2
try:
psycopg2.connect(
dbname="${POSTGRES_DB}",
user="${POSTGRES_USER}",
password="${POSTGRES_PASSWORD}",
host="${POSTGRES_HOST}",
port="${POSTGRES_PORT}",
)
except psycopg2.OperationalError:
sys.exit(-1)
sys.exit(0)
END
}
until postgres_ready; do
>&2 echo 'Waiting for PostgreSQL to become available...'
sleep 1
done
>&2 echo 'PostgreSQL is available'
exec "$@"
This was taken from pydanny's django-cookie-cutter project. Everything works using a normal non-swarm setup: docker-compose -f production.yml build and docker-compose -f production.yml up for single instance production deployment.
Last here is what my docker-compose file looks like for the swarm:
Docker-compose.yml
version: '3.6'
volumes:
production_postgres_data: {}
production_postgres_data_backups: {}
production_caddy: {}
node-modules:
networks:
webnet:
driver: overlay
attachable: true
services:
django: &django
image: registry:image
depends_on:
- postgres
- redis
env_file:
- PATH to .env
command: /start
deploy:
mode: replicated
replicas: 2
restart_policy:
condition: on-failure
delay: 5s
networks:
- webnet
postgres:
image: registry:image
volumes:
- production_postgres_data:/var/lib/postgresql/data
- production_postgres_data_backups:/backups
env_file:
- PATH to .env
deploy:
restart_policy:
condition: on-failure
delay: 5s
placement:
constraints:
- node.role == manager
networks:
- webnet
frontend:
image: registry:image
command: /start
volumes:
- node-modules:/app/node_modules
ports:
- "3000:3000"
deploy:
mode: replicated
replicas: 2
restart_policy:
condition: on-failure
delay: 5s
networks:
- webnet
caddy:
image: registry:image
depends_on:
- django
- frontend
volumes:
- production_caddy:/root/.caddy
env_file:
- PATH to .env
ports:
- "0.0.0.0:80:80"
- "0.0.0.0:443:443"
deploy:
placement:
constraints:
- node.role == manager
networks:
- webnet
redis:
image: redis:3.2
deploy:
mode: replicated
replicas: 2
networks:
- webnet
I'm not sure why exporting the environment variables from the entrypoint script are not setting when the tasks are deployed to nodes using docker stack deploy --with-registry-auth -c production.yml my_swarm.
Any help would be appreciated with this or an alternative solution to setting env variables. I could not find documentation that links dockerfile entrypoint scripts to docker swarm tasks / services.
EDIT:
I'm assuming I have to somehow utilize https://docs.docker.com/engine/swarm/secrets/, but would like to be able to keep the entrypoint script.
EDIT 2: Found the resource, need to adapt my process. https://docs.docker.com/engine/swarm/secrets/#build-support-for-docker-secrets-into-your-images
EDIT 3:
After more inspection, all other environment variables, except the ones in the entrypoint script carried over to each task. I was able to get into a django container using docker exec and run the same commands to create the DATABASE_URL, as well as, the CELERY_BROKER_URL, as shown in the script. However, still do not know why entrypoint scripts can't be used to create the environment variables.
