Infinite loop if we change the URL of the consolePublicURL and MAsterPublicURL to use the route address

[Charles Moulliard]

When we create an Openshift route to expose the WebConsole, in order to resolve the address using an external DNS, as such

apiVersion: route.openshift.io/v1
kind: Route
metadata:
  labels:
    app: openshift-web-console
  name: admin
  namespace: openshift-web-console
spec:
  host: admin-openshift-web-console.192.168.99.50.nip.io
  port:
    targetPort: https
  tls:
    termination: passthrough
  to:
    kind: Service
    name: webconsole
    weight: 100
  wildcardPolicy: None

next reconfigure the configMap of the webconsole to use the Route Address

oc get configmap/webconsole-config -o yaml
apiVersion: v1
data:
  webconsole-config.yaml: |
    apiVersion: webconsole.config.openshift.io/v1
    clusterInfo:
      consolePublicURL: https://admin-openshift-web-console.192.168.99.50.nip.io/console/
      loggingPublicURL: ''
      logoutPublicURL: ''
      masterPublicURL: https://admin-openshift-web-console.192.168.99.50.nip.io
      metricsPublicURL: ''
    extensions:
      properties: {}
      scriptURLs: []
      stylesheetURLs: []
    features:
      clusterResourceOverridesEnabled: false
      inactivityTimeoutMinutes: 0
    kind: WebConsoleConfiguration
    servingInfo:
      bindAddress: 0.0.0.0:8443
      bindNetwork: tcp4
      certFile: /var/serving-cert/tls.crt
      clientCA: ''
      keyFile: /var/serving-cert/tls.key
      maxRequestsInFlight: 0
      namedCertificates: null
      requestTimeoutSeconds: 0
kind: ConfigMap

modify the Oauthclient resource

[root@cloud ~]# oc get OAuthClient/openshift-web-console -o yaml
apiVersion: oauth.openshift.io/v1
grantMethod: auto
kind: OAuthClient
metadata:
  name: openshift-web-console
redirectURIs:
- https://admin-openshift-web-console.192.168.99.50.nip.io/console

and recreate a new pod

oc scale --replicas=0 deploy/webconsole
oc scale --replicas=1 deploy/webconsole

Then, when we access the web console, the screen is looping continusouly

and the log of the webconsole's pod reports this error

W0518 08:10:08.940019       1 start.go:93] Warning: config.clusterInfo.loggingPublicURL: Invalid value: "": required to view aggregated container logs in the console, web console start will continue.
--
  | W0518 08:10:08.940336       1 start.go:93] Warning: config.clusterInfo.metricsPublicURL: Invalid value: "": required to view cluster metrics in the console, web console start will continue.
  | I0518 08:10:09.256195       1 start.go:201] OpenShift Web Console Version: v3.9.0+b600d46-dirty
  | I0518 08:10:09.256502       1 serve.go:89] Serving securely on 0.0.0.0:8443
  | I0518 08:10:19.139673       1 logs.go:41] http2: server: error reading preface from client 10.128.0.1:33750: read tcp4 10.128.0.17:8443->10.128.0.1:33750: read: connection reset by peer
  | I0518 08:10:19.342687       1 logs.go:41] http2: server: error reading preface from client 10.128.0.1:33754: read tcp4 10.128.0.17:8443->10.128.0.1:33754: read: connection reset by peer
  | I0518 08:10:19.704623       1 logs.go:41] http2: server: error reading preface from client 10.128.0.1:33746: read tcp4 10.128.0.17:8443->10.128.0.1:33746: read: connection reset by peer
  | I0518 08:10:37.052718       1 logs.go:41] http2: server: error reading preface from client 10.128.0.1:33796: read tcp4 10.128.0.17:8443->10.128.0.1:33796: read: connection reset by peer

Web Console - Javascript reports API discovery error

[Sam Padgett]

You need to set up the corsAllowedOrigins to allow access from the console route's domain. See the corsAllowedOrigins discussion in

https://docs.openshift.org/latest/architecture/infrastructure_components/web_console.html#overview

We should do better than loop infinitely, but this will fix the problem.

[Sam Padgett]

The infinite loop problem is tracked by openshift/origin-web-console#2758

/close