Deployments with docker containers in unhealthy state should fail deployment

[Tim Rozet]

When deploying with docker containers in OOO, it is possible that a container could come up in an unhealthy state during deployment. When this happens and the container never transitions to healthy, the deployment should fail. If the container is configured and started, but never goes to healthy, it can be only be interpreted as the service is not up. If the service is not up, the deployment step did not pass and should fail.

[Bogdan Dobrelya (bogdando)]

I think issues like this one worth adding a day2 tag. I'm not sure the failures like described may be detected for the deployment time. So it seems like a subject of either the consequent future runs of 'openstack overcloud deploy' (or undercloud install) OR something like cluster operators coming from Kubernetes with the adoption of the latter as well...

[Bogdan Dobrelya (bogdando)]

...basically, for k8s, the problems like that is already solved by simple health checks and kubelet restarting the failing ones. So cluster operators are mostly the case for HA statefull services to extend that logic with a proper orchestrating/roll-outs of restarts et al.

[Alex Schultz (alex-schultz)]

We had this and then had to revert https://review.openstack.org/#/c/584288/

[David Peacock (davidjpeacock)]

What was the reason for the revert?

[Alex Schultz (alex-schultz)]

because https://bugs.launchpad.net/tripleo/+bug/1782598

[Marios Andreou (marios-b)]

This is an automated action. Bug status has been set to 'Incomplete' and target milestone has been removed due to inactivity. If you disagree please re-set these values and reach out to us on freenode #tripleo

[Launchpad Janitor (janitor)]

[Expired for tripleo because there has been no activity for 60 days.]