Cleaning error handler not always called during heartbeats

[Josh Gachnang]

If a node is in cleaning and an error occurs during the heartbeat handler, the cleaning_error_handler is not always called (e.g. if Ironic cannot reach the agent). This means the network is not torn down properly and the clean_step is not reset. There are some specific checks in the heartbeat handler, but in the general case the node is just put into CLEANFAIL.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/213769

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/213769
Committed: https://git.openstack.org/cgit/openstack/ironic/commit/?id=d1ba1861a862ed49bd301b987ffe1a0b5b6851be
Submitter: Jenkins
Branch: master

commit d1ba1861a862ed49bd301b987ffe1a0b5b6851be
Author: Josh Gachnang <email address hidden>
Date: Mon Aug 17 09:06:06 2015 -0700

    Clean up cleaning error handling on heartbeats

    If a node is in cleaning and an error occurs during the heartbeat
    handler, the cleaning_error_handler is not always called (e.g. if
    Ironic cannot reach the agent). This means the network
    is not torn down properly and the clean_step is not reset. Fix the
    error message that is set in last_error when an error occurs during
    cleaning.

    Change-Id: I22af61b641a49b48b36c8e57750d8db8a9b8cc94
    Closes-Bug: #1485667