There are intermittent test failures on the QIT Jenkins machine because the broker under test fails to start as expected. Normally, running the installed artemis_service script gives the following:

$ ./artemis-service start
Starting artemis-service
artemis service is now running (<PID>)

but occasionally, the following is observed, which results in test failure:

$ ./artemis-service start
Starting artemis-service
Already running
ERROR: Broker failed to start

There is a possibility that if the previous test using this docker image (it is shared by all artemis tests) fails to stop the broker, that the docker container will remain open and the following test, using it, will hit this error.

Each test needs to start with a new/fresh container, even though it is using the same image.

Whatever the cause, this needs investigating and fixing.