Building a custom Docker container is easy ... basically - a case for using phusion/baseimage

What challenges have I experienced?

1. A Dockerfile is quickly written to contain the base image and application files you need to get started.
2. Once built, the container runs smoothly when starting it with /bin/bash or sh as it is the default for many base images. From there, running a #tomcat application via .sh is really easy.
3. I could use the official tomcat image, however I needed Oracles Metadata Management application to run and I had no idea how to make it work with the tomcat image. So I tried alpine (java missing), oraclelinux (not working either for some reason) and finally, the good old ubuntu base image.
4. However, users of my Dockerfile (my colleagues) should not have to manually execute the shell scripts. So I want the Docker to (re)start the application server once it is created and up and running.

Why was I challenged?

• once executed, bash as a process ends and since it is the primary process (PID 1), the container ceases to run...
• but I need that tomcat server!
• Okay, basically ENTRYPOINT sets the command to be executed when a container is run. This can also be combined with custom input arguments for the command (which can also be extra commands executed when ENTRYPOINT is /bin/bash).
• However, tomcat is running as process java, a child process of the bash process doing the startup.sh
• So, again, the container stops.

How do I keep my tomcat application container running after starting?

• Some solutions are complex and hard to realize for a non-Linux-administrator
• So I read about the image phusion/baseimage
• It offers starting up services when running a container (and daemon services)
• Adding a small script to /etc/my_init.d/ did the trick
• Every time the container starts, the tomcat application is started and I can access it via the browser