container-create fails with ContainerException: Docker internal Error: ('Connection aborted.', gaierror(-2, 'Name or service not known'))

[Ryan Rossiter]

To reproduce:

Follow steps as detailed here: http://docs.openstack.org/developer/magnum/dev/dev-quickstart.html#building-and-using-a-swarm-bay

During the container-create step, the following stacktrace is thrown:

2015-09-15 19:26:23.227 ERROR oslo_messaging.rpc.dispatcher [req-598b04ec-9791-42fe-8781-19ddadc9be21 admin admin] Exception during message handling: Docker internal Error: ('Connection aborted.', gaierror(-2, 'Name or service not known'))
2015-09-15 19:26:23.227 TRACE oslo_messaging.rpc.dispatcher Traceback (most recent call last):
2015-09-15 19:26:23.227 TRACE oslo_messaging.rpc.dispatcher File "/usr/local/lib/python2.7/dist-packages/oslo_messaging/rpc/dispatcher.py", line 142, in _dispatch_and_reply
2015-09-15 19:26:23.227 TRACE oslo_messaging.rpc.dispatcher executor_callback))
2015-09-15 19:26:23.227 TRACE oslo_messaging.rpc.dispatcher File "/usr/local/lib/python2.7/dist-packages/oslo_messaging/rpc/dispatcher.py", line 186, in _dispatch
2015-09-15 19:26:23.227 TRACE oslo_messaging.rpc.dispatcher executor_callback)
2015-09-15 19:26:23.227 TRACE oslo_messaging.rpc.dispatcher File "/usr/local/lib/python2.7/dist-packages/oslo_messaging/rpc/dispatcher.py", line 129, in _do_dispatch
2015-09-15 19:26:23.227 TRACE oslo_messaging.rpc.dispatcher result = func(ctxt, **new_args)
2015-09-15 19:26:23.227 TRACE oslo_messaging.rpc.dispatcher File "/opt/stack/magnum/magnum/conductor/handlers/docker_conductor.py", line 71, in wrapped
2015-09-15 19:26:23.227 TRACE oslo_messaging.rpc.dispatcher "Docker internal Error: %s" % str(e))
2015-09-15 19:26:23.227 TRACE oslo_messaging.rpc.dispatcher ContainerException: Docker internal Error: ('Connection aborted.', gaierror(-2, 'Name or service not known'))

It looks like the problem is here: https://github.com/openstack/magnum/blob/master/magnum/conductor/handlers/docker_conductor.py#L127-L128

docker_utils.parse_docker_image() is expecting an image url:version string is given to it. When following the docs above, the image string given is 'cirros'. This is then parsed to have the image_repo to be 'cirros', which is then given to docker.pull(), where it cannot resolve the url 'cirros', and throws the error above.

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/223815
Committed: https://git.openstack.org/cgit/openstack/magnum/commit/?id=cf2470b7c9486299493af4c6fac11fb6046a385c
Submitter: Jenkins
Branch: master

commit cf2470b7c9486299493af4c6fac11fb6046a385c
Author: Ryan Rossiter <email address hidden>
Date: Tue Sep 15 21:07:41 2015 +0000

    Change swarm_manager to swarm_master

    When swarm bays are being created, the bay api address is not being set.
    This is because in the template definition, the api_address attribute is
    being set on the swarm_manager, not the swarm_master.

    Change-Id: I19c99db02086c19f76f33bce5ff4f2336b244d7b
    Closes-Bug: #1496110

[OpenStack Infra (hudson-openstack)]

Related fix proposed to branch: master
Review: https://review.openstack.org/224315

[OpenStack Infra (hudson-openstack)]

Change abandoned by Adrian Otto (<email address hidden>) on branch: master
Review: https://review.openstack.org/224315
Reason: I am marking this as abandoned so that reviewers are only looking at patches that should be considered for merge. You are welcome to un-abandon this and resubmit a new patch set if you'd like. We would welcome that.