Python 3.4 Docker容器中的AWS Elastic Beanstalk container_commands

Question

I have a Django 1.6 application running on Python 3.4 that I'd like to deploy on AWS Elastic Beanstalk , so I've selected “Preconfigured - Docker/Python” (ie, Python 3.4) as my configuration but deployments fail on the container_command to run django-admin.py . The same placeholder app bundle runs as expected on the “Preconfigured/Python” configuration (ie, Python 2.7) without error.

According to the documentation .ebextension config container_commands are run after the application and web server have been set up and the application version file has been extracted, but before the application is deployed. That appears to be the case for the Python 2.7 container but not the Python 3.4 Docker container.

To test the theory, I added a script to my bundle to peek at the environment and executed it as a container_command . The Python 2.7 container uses Python 2.7, has pip packages installed from requirements.txt , and finds django-admin.py as expected. The Python 3.4 Docker container, however, uses Python 2.6 during the container_command phase and thus doesn't have the environment set up to find django-admin.py or installed pip packages.

How can I get the the Preconfigured - Docker/Python (3.4) container to execute container_commands in the intended environment? Is there a step missing to activate the container and/or it's virtualenv before triggering container_commands ?

Answer 1

AWS Support pointed me to adding a custom Dockerfile. I used my Dockerfile's ENTRYPOINT to override the default from the base image so I could slip in my own script to trigger the syncdb command.

ref: https://forums.aws.amazon.com/message.jspa?messageID=595630