Kubernetes injects a specific hostname that corresponds to the DNS name assigned by Kube (in some cases).
On Wed, Apr 20, 2016 at 10:49 AM, Jan Pazdziora <jpazdzi...@redhat.com> wrote: > On Wed, Apr 20, 2016 at 07:11:10AM -0700, Daniel J Walsh wrote: >> >> >Because the labels are no longer shell-processed, AFAIU. >> >> Is there a different way we could go about doing this? Can we create the >> container >> during the atomic install, and then just start and stop it using atomic? > > We possibly could but I'm not very fond of using docker as the > state-holding mechanism. The whole state of this application container > is in one directory, I now work on verifying patch that would store > the hostname there upon the first run so that we catch mismatches > upon subsequent executions. > > It seems as overkill to force people to have the container created > and lingering around, instead of just running a new one when needed, > with the configuration / data from that single directory. > > By the way -- how do Kubernetes handle hostnames for their containers, > especially when the container needs a specific FQDN hostname? > > -- > Jan Pazdziora | adelton at #ipa*, #brno > Sr. Principal Software Engineer, Identity Management Special Projects, Red Hat >
