Using Liquibase and Docker

Docker is an open platform for developing, shipping, and running applications. Docker provides the ability to package and run an application in a loosely isolated environment called a container. You can also create and use images, networks, volumes, plugins, and other objects. For more information, see Docker's overview.

If you use a virtual machine, it requires a separate copy of the operating system, which needs space. Docker leverages the host system for the operating system to cut down on space. Additionally, Docker helps to start your application more quickly.

Liquibase Docker container image

Liquibase Docker container image includes the Liquibase software, Java, JDBC drivers, and all other dependencies already preconfigured. The image is based on the Eclipse Temurin image eclipse-temurin:17-jre-focal.

Also, there are libraries represented by database driver and client packages that are preinstalled into the container image. The list of available database drivers:

  • DB2
  • Firebird
  • MariaDB
  • MS SQL Server
  • PostgreSQL
  • Snowflake
  • SQLite
  • Sybase

Note: The database driver for MySQL is not preinstalled into the container image and requires additional steps to install. See MySQL driver.

See the Liquibase Dockerfile for more details. You can find the official repository for Liquibase images on the download page.

Docker pull command:

docker pull liquibase/liquibase

Example JDBC URLs

Database JDBC URL
IBM DB2 LUW jdbc:db2://<IP OR HOSTNAME>:50000/<DATABASE>
Microsoft SQL Server jdbc:sqlserver://<IP OR HOSTNAME>:1433;database=<DATABASE>
MySQL jdbc:mysql://<IP OR HOSTNAME>:3306/<DATABASE>
MariaDB jdbc:mariadb://<IP OR HOSTNAME>:3306/<DATABASE>
PostgreSQL jdbc:postgresql://<IP OR HOSTNAME>:5432/<DATABASE>?currentSchema=<SCHEMA NAME>
Snowflake jdbc:snowflake://<IP OR HOSTNAME>/?db=<DATABASE>&schema=<SCHEMA NAME>
Sybase jdbc:jtds:sybase://<IP OR HOSTNAME>:/<DATABASE>
SQLite jdbc:sqlite:/tmp/<DB FILE NAME>.db

Supported tags and image flavors

Tags

The following tags are officially supported:

  • Overall most recent build: The latest tag will be kept up to date with the most advanced Liquibase release: latest.
  • Latest major or minor builds: These tags are kept up to date with the most recent patch release of each stream, such as 4.23.
  • Specific releases: Each specific release has an associated tag, such as 4.23.1.

For a full list of tags, see Docker Hub: liquibase/liquibase.

liquibase:

This liquibase:<version> image is considered the standard choice. If you're uncertain about your specific requirements, it's recommended to opt for this image. It is designed to serve as a disposable container as well as a foundational building block for creating other images.

liquibase:-alpine

The liquibase:<version>-alpine image is a slimmed-down version of the Liquibase Docker container (liquibase:<version>). It is designed to be lightweight and have a smaller footprint, making it suitable for environments with limited resources or when only the essential functionality is required. This image is built upon the popular Alpine Linux project, which can be found in the official Alpine image. Alpine Linux stands out for its significantly smaller size compared to other distribution base images, typically around 5MB. As a result, it enables the creation of overall slimmer images.

If your main concern is minimizing the final image size, this flavor proves to be quite useful. However, it is important to note that certain software may encounter issues depending on their specific libc requirements or assumptions.

To keep the image size to a minimum, additional tools such as git or bash are not commonly included in Alpine-based images. Instead, you can utilize this image as a foundation and add the necessary components in your own Dockerfile. For more information, see the "Create a custom image" section.

Changelog files

The docker image has a /liquibase/changelog volume in which the directory that contains the root of your changelog tree can be mounted. Your --changelog-file attribute should include the path relative to the volume.

You can also use the /liquibase/changelog volume for commands that create output files, such as generate-changelog. In this case, you must specify the absolute path to the changelog; prefix the path with /liquibase/changelog/<PATH TO CHANGELOG FILE>.

Example

If you have a local c:\projects\my-project\src\main\resources\com\example\changelogs\root.changelog.xml file, you can run:

docker run --rm -v c:\projects\my-project\src\main\resources:/liquibase/changelog liquibase/liquibase update
--changelog-file=com/example/changelogs/root.changelog.xml

To generate a new changelog file at this location, run:

docker run --rm -v c:\projects\my-project\src\main\resources:/liquibase/changelog liquibase/liquibase generate-changelog --changelog-file=changelog/com/example/changelogs/root.changelog.xml

Configuration files

If you use a Liquibase properties file (defaults file) like liquibase.properties to specify attributes instead of passing them on the command line, include it in your changelog volume mount and reference it when running commands.

When you specify a custom Liquibase properties file, ensure you include searchPath=/liquibase/changelog so that Liquibase can continue looking for your changelog files there.

Example

If you have a local c:\projects\my-project\src\main\resources\liquibase.properties file, where liquibase.properties represents the Liquibase properties file, run the following command:

docker run --rm -v c:\projects\my-project\src\main\resources:/liquibase/changelog liquibase/liquibase
--defaults-file=/liquibase/changelog/liquibase.properties update

Drivers and extensions

The Liquibase docker container includes drivers for many databases. If your driver is not included or if you have an extension, you can mount a local directory that contains JAR files to /liquibase/classpath and add JAR files to your classpath setting.

Example

If you have a local c:\projects\my-project\lib\my-driver.jar file, you run the following command:

docker run --rm -v c:\projects\my-project\src\main\resources:/liquibase/changelog -v c:\projects\my-project\lib:/liquibase/classpath liquibase/liquibase
--classpath=liquibase/changelog:liquibase/classpath/my-driver.jar update

MySQL driver

The JDBC driver for MySQL has licensing restrictions that prevent it from being preinstalled into the container image. You must load the driver in one of the following ways:

Create a new container

Dockerfile:

FROM liquibase/liquibase
RUN lpm add mysql --global

Then enter this command in the CLI:

docker build . -t liquibase/liquibase-mysql

For more information, see docker build.

Set environment variable at runtime

Alternatively, enter this command in the CLI:

docker run -e INSTALL_MYSQL=true liquibase/liquibase update

For more information, see Docker run reference § ENV (environment variables).

Complete examples

Command line

Specifying everything using arguments:

docker run --rm -v <PATH TO CHANGELOG DIR>:/liquibase/changelog liquibase/liquibase
  --license-key="<PASTE LB PRO LICENSE KEY HERE>"
update 
  --url=jdbc:postgresql://<IP OR HOSTNAME>:5432/<DATABASE>?currentSchema=<SCHEMA NAME> 
  --changelog-file=com/example/changelog.xml --username=<USERNAME> 
  --password=<PASSWORD>

Using with Liquibase Environment Variables example:

docker run --env LIQUIBASE_COMMAND_USERNAME --env LIQUIBASE_COMMAND_PASSWORD --env LIQUIBASE_COMMAND_URL --env LIQUIBASE_PRO_LICENSE_KEY --env LIQUIBASE_COMMAND_CHANGELOG_FILE --rm -v <PATH TO CHANGELOG DIR>/changelogs:/liquibase/changelog liquibase/liquibase --log-level=info update

Properties file

Use the following example for configuring the liquibase.docker.properties file

classpath: /liquibase/changelog
url: jdbc:postgresql://<IP OR HOSTNAME>:5432/<DATABASE>?currentSchema=<SCHEMA NAME>
changelog-file: changelog.xml
username: <USERNAME>
password: <PASSWORD>
licenseKey=<PASTE LB PRO LICENSE KEY HERE>

Use the --defaults-file argument to evoke the liquibase.docker.properties file when running commands in the CLI:

docker run --rm -v <PATH TO CHANGELOG DIR>:/liquibase/changelog liquibase/liquibase
--defaults-file=/liquibase/changelog/liquibase.docker.properties update

Using Liquibase and PostgreSQL with Docker

When you start a PostgreSQL database using docker run postgres, the command looks at the local system to see if the PostgreSQL Docker container already exists. If the container doesn’t exist, the command will refer to Docker Hub and download it.

To run PostgreSQL and Liquibase, your Liquibase changelog needs to be accessible to the Docker container. Mount the directory that includes the changelog file with the -v option:

docker run -v /home/changelog:/liquibase/changelog liquibase/liquibase 
--url=jdbc:postgresql://<DATABASE_IP>:5432/postgres 
--changelog-file=/liquibase/changelog/changelog.xml --username=postgres 
--password=postgres

Note: Enter the directory that contains your changelog in place of /home/changelog. Also, enter the hostname/ip of your database. If you are running PostgreSQL as a Docker container, use docker inspect <CONTAINER_NAME> to find the IP address of your PostgreSQL database.

Create a custom image

If you cannot use the default Liquibase Docker container image, you can create a custom image. You can either extend an existing Dockerfile or create a completely new image.

Extend Dockerfile

The best way to create a custom Docker image is to extend the Dockerfile to add whatever tools or drivers you want, such as a native executor. Then, publish your extended Dockerfile internally in your organization. For example, see /docker/examples, which contains both Alpine and non-Alpine Dockerfile examples.

To add a driver, see the "Drivers and extensions" section. Alternatively, you can use the Liquibase Package Manager (lpm) to inject the driver into your container:

FROM liquibase/liquibase
RUN lpm add mssql --global

Then publish this custom Docker image internally.

New image

To produce a new Liquibase Docker image, you must run a few Docker commands to inject the driver into the container. You can do this with any database. The following example is for a Microsoft SQL Server database running with the latest version of Liquibase:

  1. Download the latest Microsoft SQL Server driver, for example 12.2.0
  2. Start a Docker container: docker run -it liquibase/liquibase:4.23.2 sh
  3. Do not exit out of this container. In a new terminal window, obtain the Container ID: docker ps
  4. Copy the Microsoft SQL Server 12.2.0 JAR file into the running container: docker cp mssql-jdbc-12.2.0.jre11.jar <CONTAINER ID>:/liquibase/internal/lib
  5. Create a new Docker image: docker commit <CONTAINER ID> liquibase/liquibase:4.23.2_w_mssql_12.2.0
  6. Publish this custom Docker image internally.

Troubleshooting issues with Liquibase and Docker

If you use the Liquibase Docker image 4.3.5 with Aurora MySQL RDS Cluster and have issues with the connection link, add ?autoReconnect=true"&"useSSL=false to the end of the database name in your URL:

--url=jdbc:mysql://devops-test.us-east-2.rds.amazonaws.com:3306/auditsource?autoReconnect=true"&"useSSL=false

Related links