[ostf] Improve logging in OSTF tests

[Vladimir Kuklin]

Currently, traceback of failed OSTF tests is lost because of incorrect raising of exceptions and use of logging.

We need to use traceback python module in all logging cases and also pass exception information to wrapper methods, e.g. 'fail' method, which sends fail message:

    def fail(self, msg=None):
        """Fail immediately, with the given message."""
        raise self.failureException(msg)

This one should also receive traceback from upper exception and log this information. Otherwise, we only have information that an error happened and no traceback.

E.g.:
       except Exception as exc:
            cls.error_msg.append(exc)
            LOG.debug(exc)
 ...

       except Exception as exc:
            cls.error_msg.append(exc)
            LOG.debug(traceback.format_exc())

except Exception as e:
    exc_type, exc_value, exc_tb = sys.exc_info()
    exc_info=traceback.format_exception(exc_type, exc_value, exc_tb)
    self.fail(msg,exc_info)
    ....
    def fail(self, msg=None, exc_info=None):
        """Fail immediately, with the given message."""
        raise self.failureException(msg)

[Tatyanka (tatyana-leontovich)]

add traceback.fornat_exc() in logging. As about self.fail - do not agree to change it, according we show on UI all data that we push in fail() method

[Tatyanka (tatyana-leontovich)]

https://review.openstack.org/#/c/60555/