Development Usage

Please note! This Docker setup is not for production!

Quickstart

See Quickstart.

Intro

Assuming you have Docker, Docker Desktop, Moby or some remote Docker host configured, up and running from here on. Also assuming you have Java and Maven installed, as you are at least about to develop code changes.

To test drive these local changes to the Dataverse codebase in a containerized application server (and avoid the setup described in Development Environment), you must a) build the application and b) run it in addition to the necessary dependencies. (Which might involve building a new local version of the Config Baker Image.)

Building

To build the application and config baker image, run the following command:

mvn -Pct clean package

Once this is done, you will see images gdcc/dataverse:unstable and gdcc/configbaker:unstable available in your Docker cache.

Note: This will skip any unit tests. If you have built the code before for testing, etc. you might omit the clean to avoid recompiling.

Note: Also we have a docker-compose-dev.yml file, it’s currently not possible to build the images without invoking Maven. This might change in the future.

Running

After building the app and config baker image containing your local changes to the Dataverse application, you want to run it together with all dependencies. There are four ways to do this (commands executed at root of project directory):

Cheatsheet: Running Containers

Using Maven

Using Compose

In foreground

mvn -Pct docker:run

docker compose -f docker-compose-dev.yml up

In background

mvn -Pct docker:start

docker compose -f docker-compose-dev.yml up -d

Both ways have their pros and cons:

Decision Helper: Fore- or Background?

Pros

Cons

Foreground

Logs scroll by when interacting with API / UI
To stop all containers simply hit Ctrl+C
Lots and lots of logs scrolling by
Must stop all containers to restart

Background

No logs scrolling by
Easy to replace single containers
No logs scrolling by
Stopping containers needs an extra command

In case you want to concatenate building and running, here’s a cheatsheet for you:

Cheatsheet: Building and Running Containers

Using Maven

Using Compose

In foreground

mvn -Pct package docker:run

mvn -Pct package && docker compose -f docker-compose-dev.yml up

In background

mvn -Pct package docker:start

mvn -Pct package && docker compose -f docker-compose-dev.yml up -d

Once all containers have been started, you can check if the application was deployed correctly by checking the version at http://localhost:8080/api/info/version or watch the logs.

Note: To stop all containers you started in background, invoke mvn -Pct docker:stop or docker compose -f docker-compose-dev.yml down.

Check that you can log in to http://localhost:8080 using user dataverseAdmin and password admin1.

You can also access the Payara Admin Console if needed, which is available at http://localhost:4848. To log in, use user admin and password admin. As a reminder, the application container is for development use only, so we are exposing the admin console for testing purposes. In a production environment, it may be more convenient to leave this console unopened.

Note that data is persisted in ./docker-dev-volumes in the root of the Git repo. For a clean start, you should remove this directory before running the mvn commands above.

Viewing Logs

In case you started containers in background mode (see Running), you can use the following commands to view and/or watch logs from the containers.

The safe bet for any running container’s logs is to lookup the container name via docker ps and use it in docker logs <name>. You can tail logs by adding -n and follow them by adding -f (just like tail cmd). See docker logs --help for more.

Alternatives:

  • In case you used Maven for running, you may use mvn -Pct docker:logs -Ddocker.filter=<service name>.

  • If you used Docker Compose for running, you may use docker compose -f docker-compose-dev.yml logs <service name>. Options are the same.

Redeploying

The safest and most reliable way to redeploy code is to stop the running containers (with Ctrl-c if you started them in the foreground) and then build and run them again with mvn -Pct clean package docker:run. Safe, but also slowing down the development cycle a lot.

Triggering redeployment of changes using an IDE can greatly improve your feedback loop when changing code. You have at least two options:

  1. Use builtin features of IDEs or IDE plugins from Payara.

  2. Use a paid product like JRebel.

The main differences between the first and the second options are support for hot deploys of non-class files and limitations in what the JVM HotswapAgent can do for you. Find more details in a blog article by JRebel.

IDE Triggered Code Re-Deployments

To make use of builtin features or Payara IDE Tools (option 1), please follow steps below. Note that using this method, you may redeploy a complete WAR or single methods. Redeploying WARs supports swapping and adding classes and non-code materials, but is slower (still faster than rebuilding containers). Hotswapping methods requires using JDWP (Debug Mode), but does not allow switching non-code material or adding classes.

  1. Download the version of Payara shown in Install Payara and unzip it to a reasonable location such as /usr/local/payara6.
    - Note that Payara can also be downloaded from Maven Central.
    - Note that another way to check the expected version of Payara is to run this command:
    mvn help:evaluate -Dexpression=payara.version -q -DforceStdout
  2. Install Payara Tools plugin in your IDE:

    This step is not necessary for Netbeans. The feature is builtin.

  3. Configure a connection to Payara:

    Launch Netbeans and click “Tools” and then “Servers”. Click “Add Server” and select “Payara Server” and set the installation location to /usr/local/payara6 (or wherever you unzipped Payara). Choose “Remote Domain”. Use the settings in the screenshot below. Most of the defaults are fine.

    Under “Common”, the username and password should be “admin”. Make sure “Enable Hot Deploy” is checked.

    ../_images/netbeans-servers-common.png

    Under “Java”, change the debug port to 9009.

    ../_images/netbeans-servers-java.png

    Open the project properties (under “File”), navigate to “Compile” and make sure “Compile on Save” is checked.

    ../_images/netbeans-compile.png

    Under “Run”, under “Server”, select “Payara Server”. Make sure “Deploy on Save” is checked.

    ../_images/netbeans-run.png
  4. Start all the containers, but take care to skip application deployment.

    mvn -Pct docker:run -Dapp.skipDeploy

    Run above command in your terminal to start containers in foreground and skip deployment. See cheat sheet above for more options. Note that this command either assumes you built the Dataverse Application Image first or will download it from Docker Hub.

    Note: the Admin Console can be reached at http://localhost:4848 or https://localhost:4949

  5. To deploy the application to the running server, use the configured tools to deploy. Using the “Run” configuration only deploys and enables redeploys, while running “Debug” enables hot swapping of classes via JDWP.

    Click “Debug” then “Debug Project”. After some time, Dataverse will be deployed.

    Try making a code change, perhaps to Info.java.

    Click “Debug” and then “Apply Code Changes”. If the change was correctly applied, you should see output similar to this:

    Classes to reload:
    edu.harvard.iq.dataverse.api.Info
    
    Code updated
    

    Check to make sure the change is live by visiting, for example, http://localhost:8080/api/info/version

    See below for a video demonstrating the steps above but please note that the ports used have changed and now that we have the concept of “skip deploy” the undeployment step shown is no longer necessary.

Note: in the background, the bootstrap job will wait for Dataverse to be deployed and responsive. When your IDE automatically opens the URL a newly deployed, not bootstrapped Dataverse application, it might take some more time and page refreshes until the job finishes.

IDE Triggered Non-Code Re-Deployments

Either redeploy the WAR (see above), use JRebel or look into copying files into the exploded WAR within the running container. The steps below describe options to enable the later in different IDEs.

This imitates the Netbeans builtin function to copy changes to files under src/main/webapp into a destination folder. It is different in the way that it will copy the files into the running container deployment without using a bind mount.

  1. Install the File Watchers plugin

  2. Import the watchers.xml file at File > Settings > Tools > File Watchers

  3. Once you have the deployment running (see above), editing files under src/main/webapp will be copied into the container after saving the edited file. Note: by default, IDE auto-saves will not trigger the copy.

  4. Changes are visible once you reload the browser window.

IMPORTANT: This tool assumes you are using the IDE Triggered Code Re-Deployments method to run Dataverse.

IMPORTANT: This tool uses a Bash shell script and is thus limited to Mac and Linux OS.

Using a Debugger

The Application Base Image enables usage of the Java Debugging Wire Protocol for remote debugging if you set ENABLE_JDWP=1 as environment variable for the application container. The default configuration when executing containers with the commands listed at Running already enables this.

There are a lot of tutorials how to connect your IDE’s debugger to a remote endpoint. Please use localhost:9009 as the endpoint. Here are links to the most common IDEs docs on remote debugging: Eclipse, IntelliJ

Building Your Own Base Image

If you find yourself tasked with upgrading Payara, you will need to create your own base image before running the Quickstart. For instructions, see Application Base Image.