cinder.volume.manager ValueError: incomplete format exception in cheesecake on Failing back to default backend

[Michael Denny]

Scenario to reproduce
enable replication 2.1

cinder failover-host --backend_id eos16-sync cld5b9@3pariscsirep
good status failed-over

cinder failover-host --backend_id default cld5b9@3pariscsirep
bad status failing-over

2016-03-10 11:31:32.669 ^[[01;31mERROR cinder.volume.manager [^[[01;36mreq-918d1414-8041-46c0-a776-b759704a85ec ^[[00;36mNone None^[[01;31m] ^[[01;35m^[[01;31mDriver reported error during replication failover.^[[00m
Traceback (most recent call last):
  File "/usr/lib/python2.7/logging/__init__.py", line 851, in emit
    msg = self.format(record)
  File "/usr/local/lib/python2.7/dist-packages/oslo_log/handlers.py", line 73, in format
    return logging.StreamHandler.format(self, record)
  File "/usr/lib/python2.7/logging/__init__.py", line 724, in format
    return fmt.format(record)
  File "/usr/local/lib/python2.7/dist-packages/oslo_log/formatters.py", line 267, in format
    return logging.Formatter.format(self, record)
  File "/usr/lib/python2.7/logging/__init__.py", line 464, in format
    record.message = record.getMessage()
  File "/usr/lib/python2.7/logging/__init__.py", line 328, in getMessage
    msg = msg % self.args
  File "/usr/local/lib/python2.7/dist-packages/oslo_i18n/_message.py", line 185, in __mod__
    unicode_mod = self._safe_translate(six.text_type(self), params)
  File "/usr/local/lib/python2.7/dist-packages/oslo_i18n/_message.py", line 158, in _safe_translate
    translated_message = translated_message % translated_params
ValueError: incomplete format
Logged from file manager.py, line 3322

there needs to be a way for the host to stay in a failed-over state if there is an error

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/291449

[OpenStack Infra (hudson-openstack)]

Change abandoned by Alex O'Rourke (<email address hidden>) on branch: master
Review: https://review.openstack.org/291449
Reason: I agree this is not the right approach. Eventually we should add the ability to recover from a failover error state. In the mean time, I think we should treat this as an error.