Frequently Asked Questions

[Clayton Coleman]

A set of frequently asked questions about the developer preview:

1. What is required for the preview?

For the preview you must have access to an AWS account that can provision EC2 machines, create load balancers, and create buckets.  Future iterations of the preview may alter how credentials are used against the cluster.  We use m2.large instances at the current time while we are assessing the appropriate default tuning for a small, highly loaded cluster.  The created cluster with the instance sizes we chose costs about $14 USD/day, so please don't forget to spin down clusters you don't need once you're done trying them!

2. What about platform X?

Other platforms will be added to the installer as we verify the Kubernetes machine API against them and add support to Red Hat CoreOS.  For now, AWS is supported as an adequate stand-in for other cloud environments while we stress test the infrastructure APIs in Kubernetes.  Expect to see more platforms added, including bare metal.

3. Where is the source code for the preview?

The preview is based on a large number of projects in the OpenShift and Kubernetes ecosystem.  The source code for all the consumed projects reside in the open source OpenShift organization on GitHub as part of OKD or other projects.  You can see which projects are part of a release by running 

oc adm release info --commits

while connected to a cluster. 

4. What data is collected by the built in telemetry?

Telemetry data is collected from OpenShift preview clusters as described in the github.com/openshift/telemeter project.  The data collected is uniquely identified by your clusterID and contains a limited set of metrics that help us assess usage, the health of the created preview clusters, and whether updates delivered to those clusters succeeded.  Telemetry is an important step to ensure that updates are successfully rolled out across the OpenShift fleet across a wide variety of configurations and setups.  That data will be a critical component to ensuring the success of automated updates and catching bugs in prerelease, before they are deployed to production clusters.  Future preview releases will add additional data anonymization (currently partially implemented) and configuration to disable telemetry.

To see the telemetry value reported by a cluster, run the following command against your cluster:

oc rsh -n openshift-monitoring deploy/telemeter-client curl http://localhost:8080/federate

This will show the metrics that are sent to the remote telemetry endpoint.

 

5. Can I update a preview cluster?

At the current time preview clusters are install only.  A future preview release will introduce rolling updates, once the necessary testing has been completed to ensure updates are truly zero downtime.  You can "force" update a cluster to a slightly newer release with the following command:

oc adm upgrade --to-image quay.io/openshift-release-dev/ocp-release:4.0.0-5

Please do not run this command unless you are prepared for your cluster to malfunction.  Once an update stream is defined, updates will appear when you run "oc adm upgrade" and we will update this note.

6. Will I be able to upgrade a preview cluster to GA?

Not at the current time.  From now until OpenShift 4 GA significant API changes to config and cluster behavior are expected that will require new clusters to be created.  Once OpenShift 4 reaches a release candidate status, updates to GA may be possible, but we expect the possibility even then of changes that require a forced reinstall.

7. How do I report issues?

Please capture the result of "oc get clusterversion -o yaml" and include it with an issue in this forum.  Please search before posting to ensure the issue has not already been reported.