Dockerized Wildfly with Oracle driver developing environment, setting up a data source (using credential store, keystore) and basic auth. Autodeployment and Debugging.

Thanks to SleepingTalent github for the oracle driver solution.

Table of contents:

• docker-wildfly-oracle
  • Pre-reqs
  • Configure environment
  • Database
  • Application server
    • BASIC AUTH
    • Deployment and logs
    • Logging
    • Self-signed certificates
    • Remote debugging and admin console
  • Recommended tools
  • Setup remote java application debugging

• Install Docker for Windows

Note! As Docker for Windows 7 is running under Boot2Docker (a lightweight linux VM), container's URLs will be accessible through 192.168.99.100 while for Windows 10 will just be localhost as it runs "natively". You can run docker-machine ip to figure the dockerhost out anyway.

Open bash in this project directory, then execute the following commands to create and run both containers Oracle-XE and Wildfly in Docker:

Create oracle-xe docker image:

cd oracle
docker build -t oracle-xe .
cd ..

Create Wildfly docker image:

cd wildfly
docker build -t wildfly .
cd ..

Once the images are created, only this step will be necessary to bootstrap the whole environment in the future:

docker-compose up

Oracle-XE should be already running:

• host localhost/192.168.99.100:1521
• SID XE
• user dockercrud
• password password

Some example scripts (oracle/*.sql) to be executed at the creation time of the database are provided, have a look at them, they are self-explaining.

Note! user and password + DB initialization configured in oracle/00_create_database_schema.sql. Note 2! a datasource has already been created in the wildfly image to access this database service in the file oracle-driver-commands.cli.

Wildfly should also be running in 192.168.99.100:8080 (exposed in docker-compose). The most tricky part is probably in the config-commands.cli file, where both the ORACLE DATASOURCE (password added to a credential store, looks pretty okay) and the BASIC AUTH.

Basic auth, as well as the datasource, is configuered as an ELYTRON (new security framework for wildfly) subsystem; config-commands.cli file contains some comments to help you adapt it to your special needs. I point out there three different ways of getting basic auth users created (filesystem-realm, properties-realm or using wildfly's cli tool add-user.sh).

Everything is thought to keep all the "configurable values" in the cli.properties file, so you can maybe use environment variables when deploying or something like that.

Volume ~/webapp:/opt/jboss/wildfly/standalone/deployments/ has been created so built wars can be placed in c:/users/<user>/projects/docker-wildfly-oracle locally and will be deployed in Wildfly container. The idea is you can point your IDE's target dir here so everything gets deployed as you build, zero effort.

A second volume ~/webapp:/opt/jboss/wildfly/standalone/log/ has been created so the logs from wildfly will be placed in c:/users/<user>/projects/docker-wildfly-oracle/logs locally and can be accessed. Some IDEs can configure directly this log into IDE's console.

Some workaround for logging into the application-server is configuered as a login module in wildfly, have a look at line 29 in the /wildfly/Dockerfile. File module.xml (uncomment if necessary) defines it, which gets copied over, and added to the standalone.xml file in case you hae a logging.jar for example.

In case you need self-signed certificates, add to the root of the project a keystore.jks with your certificates, then uncomment lines 25 and 26 in wildfly/Dockerfile and 34,35 and 36 in standalone.xml.

Remote debugging is using port 8787 (also exposed in docker-compose)

Admin console will be served in 192.168.99.100:9990 (user admin, password admin)

IDE -> Eclipse Oxygen/IntelliJ DB management -> Oracle SQL developer, DBeaver Git management -> Git Bash

In your IDE, set the application as a remote java application.

In Eclipse, go to Debug Configurations -> Remote Java Application:

• Pick the project
• host: localhost/192.168.99.100 (windows 10/windows 7 docker)
• port: 8787

In IntelliJ, go to Run -> Debug -> Edit Configurations:

• click on + icon and select Remote
• specify a name
• specify port 8787
• in the Use module classpath drop down, select you application
• click Apply and OK
• select the Debug configuration and click on Debug icon.

At this point you should see in the console: Connected to the target VM, address: 'localhost:8787', transport: 'socket'