On Wed, Sep 30, 2020 at 9:39 AM Jonathan Dowland <jdowl...@redhat.com> wrote:
> Hello there, > > I've got a proof-of-concept of an S2I-enabled Java application source > that injects a Java "agent" into the resulting build images[1]. (A Java > agent is some code that is loaded into the runtime environment and > executed but is otherwise tangential to the application) > > The agent listens on a configurable TCP port. In order for this to be > accessed within the cluster (or without), the image output from the > build would need that port opened. (I think!) > > I am no expert but I believe this is not possible with the S2I build > strategy. It might be possible with the Docker build strategy > (by defining additional LISTEN instructions). Can anyone confirm my > suspicion? > do you mean EXPOSE instructions? > > Part of my uncertainty is whether the containers within OpenShift need > open-port information at the container image level, Docker-style, at > all, or whether this is superceded by OpenShift-level concepts (Routes > etc.) > Openshift does not need the explicit directive at the image level. You can tell openshift to route traffic to any port on a pod/container that you want, regardless of what the image explicitly declares. So you should be fine to do what you are trying to do. > > [1] > https://github.com/jmtd/quarkus-quickstarts/commit/7f9e866aa0dab735ec8d7397f78b40cad46745b5 > > > > Many thanks, > > -- > 👱🏻 Jonathan Dowland <jdowl...@redhat.com> > Senior Software Engineer, OpenJDK, Red Hat > > _______________________________________________ > users mailing list > users@lists.openshift.redhat.com > http://lists.openshift.redhat.com/openshiftmm/listinfo/users > -- Ben Parees | OpenShift
_______________________________________________ users mailing list users@lists.openshift.redhat.com http://lists.openshift.redhat.com/openshiftmm/listinfo/users
