[keycloak-user] kubernetes

Stephen Henrie stephen at saasindustries.com
Wed Jan 17 12:33:59 EST 2018


I have a keycloak cluster running on Openshift, which of course runs
kubernetes under the covers.

I pass in the name space, pod selector and other configuration using
environment variables in the deployment descriptor like this:

          - name: OPENSHIFT_KUBE_PING_NAMESPACE
            valueFrom:
              fieldRef:
                fieldPath: metadata.namespace
          - name: OPENSHIFT_KUBE_PING_LABELS
            value:  name=mylabel
          - name: OPENSHIFT_KUBE_PING_SERVER_PORT
            value: 47600
          - name: OPERATING_MODE
            value: clustered

I think that the environment variables are used up by the kubernetes-0.9.3
code to configure the cluster and find the pods that belong to it.

In particular, the name space id dynamically provided from the meta data
for the pod set by the kubernetes controller.

Good luck!

Stephen



On Wed, Jan 17, 2018 at 4:03 AM, Simon Payne <simonpayne58 at gmail.com> wrote:

> Hi all,
>
> i'm trying to get keycloak clustered on google cloud using KUBE_PING.
>
> i have a starting keycloak server using docker based on the latest keycloak
> and using kubernetes-0.9.3
>
> however, i get the message:
>
> [org.jgroups.protocols.kubernetes.KUBE_PING] (ServerService Thread Pool --
> 51) namespace not set; clustering disabled
>
> i cant figure out how to add the namespace - all example are using
> infinispan which uses different markup to keycloak.
>
> my standalone-ha uses <protocol type="kubernetes.KUBE_PING"/>
>
> if i add any additional attributes on this tag then keycloak fails to start
>
> any help would be appreciated.
>
>
> thanks
>
> Simon.
> _______________________________________________
> keycloak-user mailing list
> keycloak-user at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/keycloak-user
>


More information about the keycloak-user mailing list