Estimated Reading Time: 5 mins
Any Java framework or tool that can run inside a Docker container will run on Codeship Pro. This documentation article will highlight simple configuration files for a Java-based Dockerfile with Maven and Gradle build steps.
When accessing other containers please be aware that those services do not run on
localhost, but on a different host, e.g.
mysql. If you reference
localhost in any of your configuration files you will have to change that to point to the service name of the service you want to access. Setting them through environment variables and using those inside of your configuration files is the cleanest approach to setting up your build environment.
project_name: build: image: organisation_name/project_name dockerfile: Dockerfile links: - redis - postgres environment: - DATABASE_URL=postgres://postgres@postgres/YOUR_DATABASE_NAME - REDIS_URL=redis://redis redis: image: redis:2.8 postgres: image: postgres:9.4
The following is an example of a Codeship Steps file.
Note that every step runs on isolated containers, so changes made on one step do not persist to the next step. Because of this, any required setup commands, such as migrating a database, should be done via a custom Dockerfile, via a
entrypoint on a service or repeated on every step.
- name: ci service: project_name command: mvn test -B - name: ci service: project_name command: gradle build
Following are two example Dockerfiles, one for using Maven and one for using Gradle, with inline comments describing each step in the file. The Dockerfiles show the different ways you can install extensions or dependencies so you can extend it to fit exactly what you need.
# We're using the official Maven 3 image from the Dockerhub (https://hub.docker.com/_/maven/). # Take a look at the available versions so you can specify the Java version you want to use. FROM maven:3 # INSTALL any further tools you need here so they are cached in the docker build WORKDIR /app # Copy the pom.xml into the container to install all dependencies COPY pom.xml ./ # Run install task so all necessary dependencies are downloaded and cached in # the Docker container. We're running through the whole process but disable # testing and make sure the command doesn't fail. RUN mvn install clean --fail-never -B -DfailIfNoTests=false # Copy the whole repository into the container COPY . ./
# We're using the official OpenJDK image from the Dockerhub (https://hub.docker.com/_/java/). # Take a look at the available versions so you can specify the Java version you want to use. FROM java:openjdk-8-jdk # Set the WORKDIR. All following commands will be run in this directory. WORKDIR /app # Copying all gradle files necessary to install gradle with gradlew COPY gradle gradle COPY \ ./gradle \ build.gradle \ gradle.properties \ gradlew \ settings.gradle \ ./ # Install the gradle version used in the repository through gradlew RUN ./gradlew # Run gradle assemble to install dependencies before adding the whole repository RUN gradle assemble ADD . ./
Using Docker’s multi-stage build feature, you can implement some changes to your Dockerfile to allow you to build and use a Java binary from a single Dockerfile, outputting a Docker image with the JAVA binary but none of the build tools - meaning a smaller and more efficient image with a less complex setup.
Multi-stage builds allow you to specify multiple
FROM lines in a Dockerfile, where each
FROM line begins a new stage. The final image is the result of the last stage, which means any previous stages are not saved in the final image. This is great for creating “builder” workflows easily.
Here’s an example using Java in a Dockerfile:
# phase one, labeled as build-stage # first stage does the building FROM maven:latest as build-stage COPY src /usr/src/app/src COPY pom.xml /usr/src/app RUN mvn -f /usr/src/app/pom.xml clean package # phase two, which uses the java binary produced above FROM org/app:latest COPY --from=BUILD /usr/src/app/target/binary.war /opt/org/app/path/deployments/binary.war
Notice that the second
FROM line begins the second stage, and this second stage is what the final image will consist of.
Because of version and test dependency issues, it is advised to try using the Jet CLI to debug issues locally via
You can enable caching per service in your Services file.