How to run Lagom in development mode and in the background

To start a Lagom or Play project in development mode for testing, you start everything with sbt runAll. You can stop everything by hitting Enter (or any other key).

Now you might think: cool, Lagom is starting everything for me, so why not use the same approach in my CI pipeline to run integration tests in a clean environment. Well, good luck!

One way or another, no matter what, the server is stopped immediately after startup. This happens

• in the container
• as background process
• with and without “</dev/null sbt runAll &”
• …

If you trace down the issue in the Lagom source, you’ll end up in ConsoleHelper.scala, which uses System.in.read() to wait for “Enter”:

def blockUntilExit() = {
  // blocks until user presses enter
  System.in.read()
}

This piece of code also triggers on other input, even if you use /dev/null as standard input. In case of /dev/null, read() returns an end of file and will not block.

Luckily there is a solution for it. To figure out what to do, you have to return to the basics. The question is: “how to deal with System.in.read()” and not “how to run Lagom in the background”. For the first question StackOverflow knows the answer.

Put into action, you have to create a bash script, e.g. run_in_bg.sh with the following contents:

#!/usr/bin/env bash

(while true; do sleep 10000; done) | sbt runAll >lagom.log &

This will allow you to run Lagom (or Play) in the background.