Open Stack

The discussed things are nice, but we had to race with the time.

- We should avoid doing any major code change
 in order to let the testr-subunit-testtools run.

- Gate time is continuously increasing and it will
  if we do not act soon.

- The huge test gaps is mainly related to keystone v3, quantum, ceilometer and heat.
 A major design change could cause major merge conflict -> slowing down the test development.

- Temporary we could increase the isolation level to higher than the ideal.
- Temporary we can loose the refstack parallel.
- Temporary we can use even dirty tricks to have the test cases to run parallel
- Temporary we can move the non parallel friendly tests to an another job

- If we cannot do it soon, we might need to limit the number of the gating tests.