docker services running as detached need better failure checking

[Tim Rozet]

When running an openstack service as detached, a configuration error will cause services to not start correctly. For example, neutron server with missing plugin config will result in neutron not starting correctly. In docker the service will be shown as constantly restarting. However, OOO will continue deployment and think the service deployment was successful. This results in a false positive that the deployment succeeded. We should add checks after issuing the docker run to detached services to ensure they are not stuck in restarting state.

[Bogdan Dobrelya (bogdando)]

There is related spec for health-checks https://review.openstack.org/#/c/465633/

[David Peacock (davidjpeacock)]

Did any progress happen with regards to testing for this? Endpoint testing? Other kinds of validations? Did we decide that running detached like this is the way to go?

Is this bug even valid anymore?