Cleanup ATs after unsuccessful run

[Albertoimpl]

Sometimes it's inevitable that ATs will fail, leaving things in an inconsistent state that can not be cleaned up.
Can we think of an approach to always clean things up without re-creating the world?

[ojhughes]

Scheduled clean up task perhaps?

[royclarkson]

I just got bit again by this today and lost half an hour. 🙃

[spikymonkey]

Can we do this by moving to the RGB pool with laundromat?

[gberche-orange]

FYI, in my app broker acceptance fork at I've added the following related elements

• ci launches a pre-test clean script
• added display of deployed broker logs, and asserts no errors, see https://github.com/orange-cloudfoundry/osb-cmdb/blob/84ae9244cdce28d050022bd91f987ce147ca0852/spring-cloud-app-broker-acceptance-tests/src/test/java/org/springframework/cloud/appbroker/acceptance/fixtures/cf/CloudFoundryService.java#L183-L208
• some tests cleanup stalled service instances by purging them

I'd be happy to contribute such enhancements to a standalone acceptance project discussed in #244 with @Albertoimpl

[spikymonkey]

Closing this since we now use a fresh environment for each AT run