Python

Overview

OKD provides S2I enabled Python images for building and running Python applications. The Python S2I builder image assembles your application source with any required dependencies to create a new image containing your Python application. This resulting image can be run either by OKD or by a container runtime.

Versions

Currently, OKD provides versions 2.7, 3.3, 3.4, and 3.5 of Python.

Images

These images come in two flavors, depending on your needs:

  • RHEL 7

  • CentOS 7

RHEL 7 Based Images

The RHEL 7 images are available through the Red Hat Registry:

  1. $ docker pull registry.redhat.io/rhscl/python-27-rhel7
  2. $ docker pull registry.redhat.io/openshift3/python-33-rhel7
  3. $ docker pull registry.redhat.io/rhscl/python-34-rhel7
  4. $ docker pull registry.redhat.io/rhscl/python-35-rhel7

CentOS 7 Based Images

These images are available on Docker Hub:

  1. $ docker pull centos/python-27-centos7
  2. $ docker pull openshift/python-33-centos7
  3. $ docker pull centos/python-34-centos7
  4. $ docker pull centos/python-35-centos7

To use these images, you can either access them directly from these image registries or push them into your OKD container image registry. Additionally, you can create an image stream that points to the image, either in your container image registry or at the external location. Your OKD resources can then reference the ImageStream. You can find example image stream definitions for all the provided OKD images.

Build Process

S2I produces ready-to-run images by injecting source code into a container and letting the container prepare that source code for execution. It performs the following steps:

  1. Starts a container from the builder image.

  2. Downloads the application source.

  3. Streams the scripts and application sources into the builder image container.

  4. Runs the assemble script (from the builder image).

  5. Saves the final image.

See S2I Build Process for a detailed overview of the build process.

Configuration

The Python image supports a number of environment variables which can be set to control the configuration and behavior of the Python runtime.

To set these environment variables as part of your image, you can place them into a .s2i/environment file inside your source code repository, or define them in the environment section of the build configuration’s **sourceStrategy** definition.

You can also set environment variables to be used with an existing image when creating new applications, or by updating environment variables for existing objects such as deployment configurations.

Environment variables that control build behavior must be set as part of the s2i build configuration or in the .s2i/environment file to make them available to the build steps.

Table 1. Python Environment Variables
Variable nameDescription

APP_FILE

This variable specifies the file name passed to the Python interpreter which is responsible for launching the application. This variable is set to app.py by default.

APP_MODULE

This variable specifies the WSGI callable. It follows the pattern $(MODULE_NAME):$(VARIABLE_NAME), where the module name is a full dotted path and the variable name refers to a function inside the specified module. If you use setup.py for installing the application, then the module name can be read from that file and the variable defaults to application. There is an example setup-test-app available.

APP_CONFIG

This variable indicates the path to a valid Python file with a gunicorn configuration.

DISABLE_COLLECTSTATIC

Set it to a nonempty value to inhibit the execution of manage.py collectstatic during the build. Only affects Django projects.

DISABLE_MIGRATE

Set it to a nonempty value to inhibit the execution of manage.py migrate when the produced image is run. Only affects Django projects.

PIP_INDEX_URL

Set this variable to use a custom index URL or mirror to download required packages during build process. This only affects packages listed in the requirements.txt file.

WEB_CONCURRENCY

Set this to change the default setting for the number of workers. By default, this is set to the number of available cores times 4.

Hot Deploying

Hot deployment allows you to quickly make and deploy changes to your application without having to generate a new S2I build. If you are using Django, hot deployment works out of the box.

To enable hot deployment while using Gunicorn, ensure you have a Gunicorn configuration file inside your repository with the **reload** option set to true. Specify your configuration file using the **APP_CONFIG** environment variable. For example, see the oc new-app command. You can use the oc set env command to update environment variables of existing objects.

You should only use this option while developing or debugging; it is not recommended to turn this on in your production environment.

To change your source code in a running pod, use the oc rsh command to enter the container:

  1. $ oc rsh <pod_id>

After you enter into the running container, your current directory is set to /opt/app-root/src, where the source code is located.