The way the installer automation currently works is that it assumes each component is in its default state. This can be a problem if a component was changed before the run as the first test won't be testing much and the component will be changed after the automation runs. Ideally the automation so leave the cluster in the state it was originally in.

In the automation, we need to first get the state of the component and change the order of the tests cases during the run to first test the opposite, then change back to the original state.