Unless you’ve been living under a rock, you’ve probably heard about Docker. By combining its container engine technology and container platform, Docker enables you to bring traditional and cloud-native applications into an automated and secure supply chain, advancing dev to ops collaboration and reducing time to value.
CAS embraced Docker a while ago by providing a sample
Dockerfile template to kickstart the builds. This template simply wraps the necessary environment and components, such as OS and Java, around an existing WAR Overlay project. There is a fair amount of smarts and creativity that could go into the build to optimize build layers, take advantage of Docker multi-stage builds and more to produce an ideal CAS image for deployment.
This tutorial focuses on an alternative approach to building CAS docker images by using Jib. Jib is an open-source Java containerizer from Google that lets Java developers build containers using the tools they know. It is a container image builder that handles all the steps of packaging your application into a container image. It does not require you to write a Dockerfile or have Docker installed, and it is directly integrated into Maven and Gradle.
Starting with CAS
6, The existing Gradle overlay embraces the Jib Gradle Plugin to provide easy-to-use out-of-the-box tooling for building CAS docker images. Our starting position is based on the following:
Once you have cloned the CAS WAR overlay, a rather simplified workflow for building a CAS Docker image using
jib is as follows:
Assuming you have completed the first two steps, let’s move on to the build step.
If you execute the build shell script, you should be greeted with a new
docker command as such:
./build.sh help Apereo CAS Enterprise Single SignOn for all earthlings and beyond ***************************************************** Usage: build.sh [...] ... docker: Build a Docker image based on the current build and configuration ...
Seems like the one we might need. Let’s run it:
./build.sh docker ... Configuration on demand is an incubating feature. warning: Setting image creation time to current time; your image may not be reproducible. Containerizing application to Docker daemon as org.apereo.cas/cas... Getting base image adoptopenjdk/openjdk11:jdk11-alpine-nightly-slim... Building dependencies layer... Building snapshot dependencies layer... Building resources layer... Building classes layer... The base image requires auth. Trying again for adoptopenjdk/openjdk11:jdk11-alpine-nightly-slim... Finalizing... Container entrypoint set to [docker/entrypoint.sh] Loading to Docker daemon... Built image to Docker daemon as org.apereo.cas/cas
What’s happening here is that the Gradle build is invoking the
jib plugin to fetch a base image, and then build all other required layers on top of it. Our CAS settings and logging configuration, etc as well as the
cas.war are moved into the image to subsequently be handled via the startup shell script, which is the point where we instruct the build and Docker to use an entrypoint allowing our image to turn into a running container.
If you query for available Docker images, you might see:
docker images ... org.apereo.cas/cas latest 34a3d3502970 4 minutes ago 551MB ...
…which may then conveniently be executed via the following:
docker run --name cas -p 8443:8443 -d org.apereo.cas/cas docker logs -f cas
That is all it takes.
It is evidently very convenient to use the same CAS build to generate a Docker image without dabbling too much into the specifics and nuances of a
Dockerfile syntax. There are also few configuration options available in the Gradle build that allow one to decide the base image, expose ports and assign tags and labels. More interestingly and while the CAS build tries to keep things simple,
jib really makes pushing images to remote registries easy by supporting the likes of Docker Hub, Google Container Registry and Amazon Elastic Container Registry all in on spot using a familiar consistent syntax.
In doing so, it should be noted that the
jib plugin and project is relatively new with the initial project announcement dating back to July 2018. Furthermore, support for Spring Boot projects and especially those that build WARs and double-specifically those that depend and work with WAR Overlays such as CAS is extremely brand new. It is completely plausible that future iterations of the plugin allow tighter integrations with the Docker engine to produce better-optimized images, expose more configuration knobs and keep improving WAR support. If the current behavior and capabilities of this plugin do not meet the requirements, you’re welcome to continue using the native
Dockerfile approach and keep an eye towards newer versions of the plugin.
I hope this review was of some help to you and I am sure that both this post as well as the functionality it attempts to explain can be improved in any number of ways. Please feel free to engage and contribute as best as you can.