Skip to content

Latest commit

 

History

History
196 lines (143 loc) · 8.79 KB

FAQ.md

File metadata and controls

196 lines (143 loc) · 8.79 KB

Frequently Asked Questions

What are the relations with OCP project? Is OKD4 an upstream of OCP?

In 3.x release time OKD was used as an upstream project for Openshift Container Platform. OKD could be installed on Fedora/CentOS/RHEL and used CentOS based images to install the cluster. OCP, however, could be installed only on RHEL and its images were rebuilt to be RHEL-based.

Universal Base Image project has enabled us to run RHEL-based images on any platform, so the full image rebuild is no longer necessary, allowing OKD4 project to reuse most images from OCP4. There is another critical part of OCP - Red Hat Enterprise Linux CoreOS. Although RHCOS is an open source project (much like RHEL8) it's not a community-driven project. As a result, OKD workgroup has made a decision to use Fedora CoreOS - open source and community-driven project - as a base for OKD4. This decision allows end-users to modify all parts of the cluster using prepared instructions.

It should be noted that OKD4 is being automatically built from OCP4 ci stream, so most of the tests are happening in OCP CI and being mirrored to OKD. As a result, OKD4 CI doesn't have to run a lot of tests to ensure the release is valid.

These relationships are more complex than "upstream/downstream", so we use "sibling distributions" to describe its state.

How stable is OKD4?

OKD4 builds are being automatically tested by release-controller. Release is rejected if either installation, upgrade from previous version or conformance test fails. Test results determine the upgrade graph, so for instance, if upgrade tests passed for beta5->rc edge, clusters on beta5 can be directly updated to rc release, bypassing beta6.

The OKD stable version is released bi-weekly, following Fedora CoreOS schedule, client tools are uploaded to Github and images are mirrored to Quay.

Can I run a single node cluster?

Yes.

Set the following in install config, which renders 0 worker nodes and 1 control-plane node:

compute:
- name: worker
  replicas: 0
controlPlane:
  name: master
  replicas: 1

This would inject a non-HA manifest for etcd and run a single ingress pod.

WARNING: this cluster cannot be upgraded or adjusted via MachineConfigs. Adding more masters is not yet supported.

Where can I find upgrades?

https://origin-release.svc.ci.openshift.org/

Note that nightly builds (from 4.x.0-0.okd) are pruned every 72 hours. If your cluster uses these images, consider mirroring these files to a local registry. Builds from the stable-4 stream are not removed.

How can I upgrade my cluster to a new version?

Find a version where a tested upgrade path is available from your version for on

https://origin-release.svc.ci.openshift.org

Upgrade options:

Preferred ways:

  • Web Console: Home -> Overview -> Tab: Cluster, Card: Overview -> View settings -> Update Status

  • Shell: Upgrades to latest available version

    oc adm upgrade

Last resort:

Upgrade to a certain version (will ignore the update graph!)

oc adm upgrade --force --allow-explicit-upgrade=true --to-image=registry.svc.ci.openshift.org/origin/release:4.4.0-0.okd-2020-03-16-105308

This will take a while; the upgrade may take several hours. Throughout the upgrade, kubernetes API would still be accessible and user workloads would be evicted and rescheduled as nodes are updated.

Interesting commands while an upgrade runs

Check overall upgrade status:

oc get clusterversion

Check the status of your cluster operators:

oc get co

Check the status of your nodes (cluster upgrades may include base OS updates):

oc get nodes

How can I find out what's inside of a (CI) release and which commit id each component has?

This one is very helpful if you want to know if a certain commit has landed in your current version:

oc adm release info registry.svc.ci.openshift.org/origin/release:4.4  --commit-urls

Name:      4.4.0-0.okd-2020-04-10-020541
Digest:    sha256:79b82f237aad0c38b5cdaf386ce893ff86060a476a39a067b5178bb6451e713c
Created:   2020-04-10T02:14:15Z
OS/Arch:   linux/amd64
Manifests: 413

Pull From: registry.svc.ci.openshift.org/origin/release@sha256:79b82f237aad0c38b5cdaf386ce893ff86060a476a39a067b5178bb6451e713c

Release Metadata:
  Version:  4.4.0-0.okd-2020-04-10-020541
  Upgrades: <none>

Component Versions:
  kubernetes 1.17.1
  machine-os 31.20200407.20 Fedora CoreOS

Images:
  NAME                                           URL
  aws-machine-controllers                        https://github.com/openshift/cluster-api-provider-aws/commit/5fa82204468e71b44f65a5f24e2675dbfa0f5c29
  azure-machine-controllers                      https://github.com/openshift/cluster-api-provider-azure/commit/832a43a30d7f00cd6774c1f5cd117aeebbe1b730
  baremetal-installer                            https://github.com/openshift/installer/commit/a58f24b0df7e3699b39d4ae1d23c45672706934d
  baremetal-machine-controllers
  baremetal-operator
  baremetal-runtimecfg                           https://github.com/openshift/baremetal-runtimecfg/commit/09850a724d9290ffb05db3dd7f4f4c748b982759
  branding                                       https://github.com/openshift/origin-branding/commit/068fa1eac9f31ffe13089dd3de2ec49c153b2a14
  cli                                            https://github.com/openshift/oc/commit/2576e482bf003e34e67ba3d69edcf5d411cfd6f3
  cli-artifacts                                  https://github.com/openshift/oc/commit/2576e482bf003e34e67ba3d69edcf5d411cfd6f3
  cloud-credential-operator                      https://github.com/openshift/cloud-credential-operator/commit/446680ed10ac938e11626409acb0c076edd3fd52
  ...

How can I enable the (non-community) Red Hat Operators?

If you have installed OKD with an "official" pull secret which contains registry.redhat.io, such as that with which you can install OpenShift, you are entitled to enable the Red Hat operators alongside the default community operators.

One reason for doing so, is to enable the "metering-ocp" operator, as the community operators ships with a deprecated "metering" operator.

Firstly, ensure that you do have a pull secret which contains registry.redhat.io

Then, update the OperatorHub CR:

(
cat <<EOF
apiVersion: config.openshift.io/v1
kind: OperatorHub
metadata:
  name: cluster
spec:
  disableAllDefaultSources: true
  sources:
  - disabled: false
    name: redhat-operators
  - disabled: false
    name: community-operators
EOF
 ) | oc apply -f -

What to do in case of errors?

If you experience problems during installation you must collect the bootstrap log bundle, see instructions

If you experience problems post installation, collect data of your cluster with:

oc adm must-gather

See documentation for more information.

Upload it to a file hoster and send the link to the developers (Slack channel, ...)

During installation the SSH key is required. It can be used to SSH onto the nodes later on - ssh core@<node ip>

Where do I seek support?

OKD is a community-supported distribution, Red Hat does not provide commercial support of OKD installations.

Contact us on Slack:

  • Workspace: Kubernetes, Channel: #openshift-dev (for developer communication)

  • Workspace: Kubernetes, Channel: #openshift-users (for users)

See https://openshift.tips/ for useful Openshift tips