health-manager unable to pick clusters for listening

[Qiming Teng]

due to implementation service status check logic, dead services still have service records in database to be deleted next time when an engine is restarted.

health manager should check if a service is dead when selecting clusters to listen for, instead of filtering out all service engines.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/355743

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/355743
Committed: https://git.openstack.org/cgit/openstack/senlin/commit/?id=ad8a538094652e9f1e722b29c73e4a454d9b9e16
Submitter: Jenkins
Branch: master

commit ad8a538094652e9f1e722b29c73e4a454d9b9e16
Author: tengqm <email address hidden>
Date: Tue Aug 16 02:19:30 2016 -0400

    Fix service filtering when claiming clusters

    The health manager is not claiming clusters accurately when service is
    restarted. The reason is that there are some dead service records to be
    removed next time one of the engines gets restarted. This patch adds a
    check about the service's status and filters the service records based
    on the check result.

    Change-Id: Id482e74068cc1cb4394c79f14f1c6442cf9f717b
    Partial-Bug: #1613552

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/355751
Committed: https://git.openstack.org/cgit/openstack/senlin/commit/?id=5a68ced6d2397a459a1b33b10d8aaebeb7c27460
Submitter: Jenkins
Branch: master

commit 5a68ced6d2397a459a1b33b10d8aaebeb7c27460
Author: tengqm <email address hidden>
Date: Tue Aug 16 02:34:50 2016 -0400

    Fix health manager initialization

    With this patch we postpone health manager initialization until service
    records are cleaned up. Also, we change the executor to 'threading'
    which is recommended by oslo.messaging. A new INFO log is added for
    debugging's purpose.

    Closes-Bug: #1613552
    Change-Id: I761b61ee164fe3532aaaaad5408b68df84d0d7ca

[OpenStack Infra (hudson-openstack)]

This issue was fixed in the openstack/senlin 2.0.0.0b3 development milestone.