Hello,
Expected behaviour:
dig +short dashboard-db-stage.ci.stream.centos.org router.ocp.ci.centos.org. 8.43.84.248
Failure:
oc login --token=sha256~XXX --server=https://api.ocp.ci.centos.org:6443 oc rsh <ANY POD>
then:
curl -vvv dashboard-db-stage.ci.stream.centos.org curl: (6) Could not resolve host: dashboard-db-stage.ci.stream.centos.org or MongoServerSelectionError: getaddrinfo ENOTFOUND dashboard-db-stage.ci.stream.centos.org
@bstinson @siddharthvipul1 @mobrien Is this something that can be fixed on the cluster? Thank you.
Well, the ocp cluster domain is ocp.ci.centos.org and so all deployed apps are available under <app>.apps.ocp.ci.centos.org If you create a specific route that doesn't match with the openshift cluster, you have to ask us to create a A record in the public ci.centos.org zone. But the problem is that external and internal are also different things. So can you explain why you deployed with a non default (and known) route ?
@arrfab Hello:
Q> So can you explain why you deployed with a non default (and known) route ?
Because I am not aware about successful updates of Openshift Platfrom to new minor versions. Only to patches update: (major.minor.patch) are successful, that I am aware of. Example: https://console.apps.ci.centos.org:8443/ (the 3.9)
And yes, I am not aware about successful updates for minor even for OpenShift 4.x. I highly doubt that https://console-openshift-console.apps.ocp.ci.centos.org will be updated to new OpenShift.
That is why we requested distinct domain name, to make service independent from hosted platform.
@arrfab I do not quite get your answer.
Openshift router listens to: <app>.apps.ocp.ci.centos.org
dashboard-db-stage.ci.stream.centos.org -- is CNAME configure in DNS:
dashboard-db-stage.ci.stream.centos.org. 595 IN CNAME router.ocp.ci.centos.org. router.ocp.ci.centos.org. 595 IN A 8.43.84.248
I know this schema works at other clusters (CNAME). Wonder why it doesn't work here.
I verified and yes it seems there was a CNAME for external, so it works from outside. Now to also verify how from inside openshift the resolver should be aware about it
Metadata Update from @arrfab: - Issue tagged with: centos-ci-infra, low-gain, medium-trouble
Hello, I found a workaround by using services directly. I am going to close this ticket.
Metadata Update from @astepano: - Issue close_status updated to: Insufficient Data - Issue status updated to: Closed (was: Open)
Login to comment on this ticket.