Unhandled exception during periodic recovery (_get_failed_domains is returning domains instead of status objects)

[Paul Glass]

Reproduction: Wait for periodic recovery to occur. You may need a zone in an error state. You should see an exception in the logs that looks like: http://paste.openstack.org/show/182105/

File "/root/designate/designate/pool_manager/service.py", line 423, in _periodic_update_domains_that_failed
    context, update_status.domain_id)
AttributeError: 'Domain' object has no attribute 'domain_id'

Discussion in IRC suggested this is because _get_failed_domains returns domains, but should return status objects.

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/159202
Committed: https://git.openstack.org/cgit/openstack/designate/commit/?id=bb86fca692bf32683a202d8a5e3ae228c87fb939
Submitter: Jenkins
Branch: master

commit bb86fca692bf32683a202d8a5e3ae228c87fb939
Author: Vinod Mangalpally <email address hidden>
Date: Wed Feb 25 16:07:11 2015 -0600

    Add error state transition for creation of domains

    * If a domain creation is not successful on the backend
    servers, then the domain transitions to an error state.
    * Rewrote _clear_cache to skip an explicit retrieve.
    For memcache no retrieve is necessary. So no additional
    action is needed.
    For sqlalchemy, an id is needed for a delete, so a retrieve
    is done on a clear if needed.
    * periodic_sync now sync's domains that are not in an
    error state (periodic_recovery handles domains that are
    in error)
    * Added/modified unit tests

    Change-Id: I375262b3ccf4d1d9da1c8954f4768a55955a0129
    Closes-Bug: 1413800
    Closes-Bug: 1425695