Tobias Florek

Sorry, that was a little terse. I will write tomorrow if time permits, but in short, you don't have to change the deployment (or stop the running pod) at all....

Note, that I won't be able to take care of any support requests for at least a month starting tomorrow. I am sorry.

You are right.

I agree to have more explicit error handling. I don't like calling watch_routes twice though.

Yes ScheduledJob is the right way to replace cron, and Job might be the right way to decouple the controller from the requesting of certificates. Unfortunately ScheduledJobs are only supported...

Oh, I misremembered. Fortunately the differences between CronJob and ScheduledJob won't matter much (only the `openshift-letsencrypt` template would be have to change, not the code to manage jobs). @caruccio Can...

Yeah, I know about kube 1.3 (I checked this exact swagger spec.) But sometimes openshift origin disables certain features even if they are in their swagger spec, e.g. ThirdPartyResources. Thank...

Huh? That did not work for me. See openshift/origin#8622. Maybe I am missing something, but my testing did confirm that they would not be supported until they are stable.

Yes, that's right. Any idea how an api should look like? I prefer being explicit which domains to combine.

Oh, you mean routes with paths. Yeah, that's a bug, I introduced when going from secrets to just storing the certificates in the route.