Reuse Scenario Runners and scenarios in context

We should adopt current benchmarks scenarios:

1) Do not randomize users in scenario runners
(keep order) https://github.com/stackforge/rally/blob/6acc580ee124c6ea30303fc6183381f9d9526c7e/rally/benchmark/runners/base.py#L41

2) Add support to "export" created resources from scenario method:
self.export(dict_of_values)
e.g. here:
https://github.com/stackforge/rally/blob/6acc580ee124c6ea30303fc6183381f9d9526c7e/rally/benchmark/scenarios/keystone/basic.py#L24-L25

3) Export data only if base.Scenario was inited with export=True (by default should be false)

4) Export data in all benchmarks scenarios (that creates resources and could be used in nova)

5) Do something with cleanup mechanism (and resource naming)

6) Switch all context classes to use Scenarios runners

7) Do something with naming of resource (to be able to use generic cleanup)