enable users to group deployments/testcases so that server can be restarted between the execution of each group

When running many tests with embedded derby databases and a single glassfish instance(forkmode=none), the tests end up failing with java heap space error.
This can be solved by setting forkmode=pertest.

This made me think that there might be a use for something inbetween one instance per test and one instance for all tests.

A nice feature could be to enable some mechanism for the grouping of testcases. This way we could restart the application server after a set of deployments/testcases have run.

I added a link to a forumpost which describes the issue that inspired this feature request.