Comment 6 for bug 1722577

Reviewed: https://review.openstack.org/510951
Committed: https://git.openstack.org/cgit/openstack/tempest/commit/?id=0d4551b367c9d3f37117a5426a9ca500d791ac25
Submitter: Jenkins
Branch: master

commit 0d4551b367c9d3f37117a5426a9ca500d791ac25
Author: Matt Riedemann <email address hidden>
Date: Tue Oct 10 13:00:48 2017 -0400

    Only attempt to detach an in-use volume during cleanup

    test_list_get_volume_attachments intermittently fails
    during cleanup because it tries to detach an already
    detached volume, which results in a 400 response.

    Tempest, as the client, should be checking the volume
    status before making the detach request. The only reason
    this ever worked before Pike was because of some
    (incorrect) ordering of operations in the compute
    service which affected how the API behaved during detach,
    and the compute API would return a 404 rather than a 400.

    That changed with I2581ff9f9c0e7cfc14a25acf45eb1860df69eacf
    in Pike, which exposed the race on the Tempest side by
    deleting the BDM in nova *after* marking the volume as
    'available' in Cinder, and the os-volume_attachments API
    checks for the existence of the BDM and if it exists, attempts
    the detach (which then fails with the 400 from Cinder).

    Change-Id: Id2d22cbb86d8d5fa7f71202b274260c1367e8a0f
    Closes-Bug: #1722577