Zones remain in ERROR indefinitely with the noop pool manager cache

[Paul Glass]

Reproduction:
Configure designate to use the noop cache. In designate.conf,

    [service:pool_manager]
    cache_driver = noop
    ...

Then create a zone through Designate's api. You should see quickly see the zone created on the nameserver, but the zone will move to ERROR in Designate. The zone seems to keep its ERROR state forever, even after periodic recovery runs multiple times.

Expected:
The zone should eventually go to ACTIVE.

Note: This was on commit cb37ee959dcbf3f1347f2142eb59843f09d01b8e.

[Kiall Mac Innes (kiall)]

More data points - I've also just spotted this with the SQLA cache + PowerDNS..

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/162754

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/162754
Committed: https://git.openstack.org/cgit/openstack/designate/commit/?id=9c6636b6f87aa1fdd13220b8601d28f6219deb88
Submitter: Jenkins
Branch: master

commit 9c6636b6f87aa1fdd13220b8601d28f6219deb88
Author: Vinod Mangalpally <email address hidden>
Date: Tue Mar 10 17:06:03 2015 -0500

    Call central's update_status correctly

    * Cleaned up the way pool manager calls central's
    update_status for create, delete and update.
    * Cleaned up pool manager's update_status
    * Added some more information to logging statements
    when central's update_status and pool manager's
    update_status are called
    * When a domain is deleted, I see SERVFAIL too being
    returned by BIND. Added that to the list of error codes
    to check
    * memcache does not seem to store a value of None.
    Worked around that bug.
    * Updated unit tests
    Closes-Bug: 1427433
    Closes-Bug: 1427425
    Change-Id: Id33381a967afd724ed56ac1d8df648b05a55516c