trove smoke test broken with services not running (blocked charm)

[Alex Kavanagh]

When attempting to run a smoke test (on a lxd based local juju - but suspect that it will fail on all of them), the smoke test can't run because the trove charm gets blocked due to services not running (at least that's the error message):

Unit Workload Agent Machine Public address Ports Message
keystone/0* active idle 0 172.16.1.249 5000/tcp Unit is ready
mysql/0* active idle 1 172.16.1.178 3306/tcp Ready
rabbitmq-server/0* active idle 2 172.16.1.101 5672/tcp Unit is ready
trove/0* blocked idle 3 172.16.1.140 8779/tcp Services not running that should be:
 trove-api, trove-taskmanager, trove-conductor

It looks like the issue is with the an update to trove that's taken place, but something isn't working properly:

trove-api.log:

2017-07-20 09:49:57.378 24006 ERROR root NoSuchTableError: `instances`
..
2017-07-20 09:49:58.436 24273 ERROR root NoSuchTableError: `datastores`
..
2017-07-20 09:49:59.473 24316 ERROR root NoSuchTableError: `datastore_version_metadata`

trove-conductor.log:
2017-07-20 09:49:54.201 17091 WARNING oslo_db.sqlalchemy.engines [-] SQL connection failed. 3 attempts left.
2017-07-20 09:49:57.382 24024 CRITICAL root [-] NoSuchTableError: `instances`
..
2017-07-20 09:49:58.436 24277 ERROR root NoSuchTableError: `datastores`
..
2017-07-20 09:49:59.437 24324 ERROR root NoSuchTableError: `datastore_version_metadata`

So it's either a MySQL initialisation error, credentials, or a change in trove version. Google search didn't yield anything obvious.