Yes this called router-sharding.
I have tried to add some pr user doc but I stopped due to the fact that it takes too long.
You can run a router just for internal usage.
When you search for sharding in the doc you will finde some infos
maybe this doc will also help
One important info. When you decide to use router labels you should also use it for internl & external.
Von: users-bounces lists openshift redhat com [mailto:users-bounces lists openshift redhat com]
Im Auftrag von Den Cowboy
Gesendet: Mittwoch, 20. Juli 2016 15:27
An: aleks <al-openshiftusers none at>
Cc: users lists openshift redhat com
Betreff: RE: OpenShift origin: internal routing with services
I read the documentation about it.
It's not very clear for me but it seems to be something that you can deploy multiple routers en routerA will handle the routes of project A, B and C and router B will handle the routes of project D,E,F or something?
I don't really see or know how I can create a router which will handle routes internally (without going to the outside)
> Date: Mon, 18 Jul 2016 19:04:22 +0200
> From: al-openshiftusers none at
> To: dencowboy hotmail com
> CC: users lists openshift redhat com
> Subject: Re: OpenShift origin: internal routing with services
> Am 14-07-2016 09:27, schrieb Den Cowboy:
> > Hi,
> > At the moment we have a setup like this:
> > project A
> > project B
> > project A contains a pod A which needs an API which is running in pod B
> > in project B.
> > Pod A has an environment variable: "api-route.xxx.dev/api"
> > So when I'm going to that route in my browser I'm able to see the API
> > and this works fine (okay we're able to configure https route etc)
> > But we'd like to keep everything internally. So without using routes.
> > So thanks to the ovs-multitenant-pluging we're able to "join" the
> > networks of our projects (namespaces). And I'm able to ping to from
> > inside pod A to the service of my pod B in project B.
> > ping api-service.project-b
> > api-service.project-b.svc.cluster.local (172.30.xx.xx) 56(84) bytes of
> > data.
> > So we're able to access the pod from its service without using an
> > external route.
> > But like I told in the beginning. Our API is on api-route.xxx.dev/api
> > so I have to go to something like 172.30.xx.xx:8080/api.
> > Is there a way to obtain this goal? So we try to connect to a 'subpath'
> > of our service without using routes.
> > Is this possible?
> I think you can go another way and use a internal router with
> and deploy the internal api on the internal router.
> Best regards