Tempest tests occasionally failed on tear down

[hongbin]

Here is a snapshot of the log: http://paste.openstack.org/show/673497/ . It looks we need to revisit the cleanup logic on tearDown since it seems to delete 'paused' container.

Logstash search: http://logstash.openstack.org/#dashboard/file/logstash.json?query=message%3A%5C%22Cannot%20delete_force%20container%5C%22%20AND%20tags%3A%5C%22console%5C%22%20AND%20voting%3A1&from=864000s

[OpenStack Infra (hudson-openstack)]

Related fix proposed to branch: master
Review: https://review.openstack.org/545995

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/545995
Committed: https://git.openstack.org/cgit/openstack/zun/commit/?id=403b4cd3b5e09fd4055d620eba73890f8223f9e1
Submitter: Zuul
Branch: master

commit 403b4cd3b5e09fd4055d620eba73890f8223f9e1
Author: Hongbin Lu <email address hidden>
Date: Mon Feb 19 21:55:17 2018 +0000

    Pause container before doing commit

    Docker daemon will pause the container being committed if we don't
    pause the container. However, this will lead to inconsistency
    between Docker daemon and DB (the container's state is Running in
    DB but Paused in Docker). Although the inconsistency will be
    converged in periodic task, we can do better by explicitly pausing
    the container before doing the commit and unpause the container
    after. This will keep the container's state consistent during and
    after commit and avoid bug like #1749879.

    Change-Id: Ie872e71e386714b791f3277fcdb2326e09235128
    Related-Bug: #1749879