neutron db migration fails on openstack-origin upgrade

[Corey Bryant]

This can be recreated by deploying with the default openstack-origin on trusty and then setting openstack-origin=ppa:ubuntu-cloud-archive/juno-staging for the neutron-api charm.

I'm using our next.yaml and then upgrading in this order:
    keystone swift-storage-z1 swift-storage-z2 swift-storage-z3 \
    swift-proxy cinder glance ceilometer ceilometer-agent \
    neutron-api neutron-gateway openstack-dashboard \
    nova-cloud-controller nova-compute

Here's the traceback:

2014-10-14 17:29:57 INFO config-changed Traceback (most recent call last):
2014-10-14 17:29:57 INFO config-changed File "/var/lib/juju/agents/unit-neutron-api-0/charm/hooks/config-changed", line 400, in <module>
2014-10-14 17:29:57 INFO config-changed main()
2014-10-14 17:29:57 INFO config-changed File "/var/lib/juju/agents/unit-neutron-api-0/charm/hooks/config-changed", line 394, in main
2014-10-14 17:29:57 INFO config-changed hooks.execute(sys.argv)
2014-10-14 17:29:57 INFO config-changed File "/var/lib/juju/agents/unit-neutron-api-0/charm/hooks/charmhelpers/core/hookenv.py", line 502, in execute
2014-10-14 17:29:57 INFO config-changed self._hooks[hook_name]()
2014-10-14 17:29:57 INFO config-changed File "/var/lib/juju/agents/unit-neutron-api-0/charm/hooks/charmhelpers/core/host.py", line 268, in wrapped_f
2014-10-14 17:29:57 INFO config-changed f(*args)
2014-10-14 17:29:57 INFO config-changed File "/var/lib/juju/agents/unit-neutron-api-0/charm/hooks/config-changed", line 122, in config_changed
2014-10-14 17:29:57 INFO config-changed do_openstack_upgrade(CONFIGS)
2014-10-14 17:29:57 INFO config-changed File "/var/lib/juju/agents/unit-neutron-api-0/charm/hooks/neutron_api_utils.py", line 213, in do_openstack_upgrade
2014-10-14 17:29:57 INFO config-changed migrate_neutron_database()
2014-10-14 17:29:57 INFO config-changed File "/var/lib/juju/agents/unit-neutron-api-0/charm/hooks/neutron_api_utils.py", line 227, in migrate_neutron_database
2014-10-14 17:29:57 INFO config-changed subprocess.check_output(cmd)
2014-10-14 17:29:57 INFO config-changed File "/usr/lib/python2.7/subprocess.py", line 573, in check_output
2014-10-14 17:29:57 INFO config-changed raise CalledProcessError(retcode, cmd, output=output)
2014-10-14 17:29:57 INFO config-changed subprocess.CalledProcessError: Command '['neutron-db-manage', '--config-file', '/etc/neutron/neutron.conf', '--config-file', '/etc/neutron/plugins/ml2/ml2_conf.ini', 'upgrade', 'head']' returned non-zero exit status 1
2014-10-14 17:29:57 ERROR juju.worker.uniter uniter.go:486 hook failed: exit status 1

Running the db upgrade manually, I hit:

sqlalchemy.exc.OperationalError: (OperationalError) (1050, "Table 'agents' already exists") '\nCREATE TABLE agents (\n\tid VARCHAR(36) NOT NULL, \n\tagent_type VARCHAR(255) NOT NULL, \n\t`binary` VARCHAR(2
55) NOT NULL, \n\ttopic VARCHAR(255) NOT NULL, \n\thost VARCHAR(255) NOT NULL, \n\tadmin_state_up BOOL NOT NULL, \n\tcreated_at DATETIME NOT NULL, \n\tstarted_at DATETIME NOT NULL, \n\theartbeat_timestamp
DATETIME NOT NULL, \n\tdescription VARCHAR(255), \n\tconfigurations VARCHAR(4095) NOT NULL, \n\tPRIMARY KEY (id), \n\tCHECK (admin_state_up IN (0, 1))\n)ENGINE=InnoDB\n\n' ()

But I'm not sure if that is the original error or not. Perhaps we're conflicting with the nova-cc db upgrade?