You can update, or upgrade, an OKD cluster within a minor version by using the OpenShift CLI (oc).


About the OKD update service

The OKD update service is the hosted service that provides over-the-air updates to both OKD and Red Hat Enterprise Linux CoreOS (RHCOS). It provides a graph, or diagram that contain vertices and the edges that connect them, of component Operators. The edges in the graph show which versions you can safely update to, and the vertices are update payloads that specify the intended state of the managed cluster components.

The Cluster Version Operator (CVO) in your cluster checks with the OKD update service to see the valid updates and update paths based on current component versions and information in the graph. When you request an update, the OKD CVO uses the release image for that update to upgrade your cluster. The release artifacts are hosted in Quay as container images.

To allow the OKD update service to provide only compatible updates, a release verification pipeline exists to drive automation. Each release artifact is verified for compatibility with supported cloud platforms and system architectures as well as other component packages. After the pipeline confirms the suitability of a release, the OKD update service notifies you that it is available.

Because the update service displays all valid updates, you must not force an update to a version that the update service does not display.

During continuous update mode, two controllers run. One continuously updates the payload manifests, applies them to the cluster, and outputs the status of the controlled rollout of the Operators, whether they are available, upgrading, or failed. The second controller polls the OKD update service to determine if updates are available.

Reverting your cluster to a previous version, or a rollback, is not supported. Only upgrading to a newer version is supported.

During the upgrade process, the Machine Config Operator (MCO) applies the new configuration to your cluster machines. It cordons the number of nodes that is specified by the maxUnavailable field on the machine configuration pool and marks them as unavailable. By default, this value is set to 1. It then applies the new configuration and reboots the machine. If you use Red Hat Enterprise Linux (RHEL) machines as workers, the MCO does not update the kubelet on these machines because you must update the OpenShift API on them first. Because the specification for the new version is applied to the old kubelet, the RHEL machine cannot return to the Ready state. You cannot complete the update until the machines are available. However, the maximum number of nodes that are unavailable is set to ensure that normal cluster operations are likely to continue with that number of machines out of service.

OKD Upgrade Channels and Releases

In OKD 4.1, Red Hat introduced the concept of channels for recommending the appropriate release versions for cluster upgrade. By controlling the pace of upgrades, these upgrade channels allow users to choose an upgrade strategy. Upgrade channels are tied to a minor version of OKD. For instance, OKD Latest upgrade channels will never include an upgrade to a 4.5 release. This ensures administrators make an explicit decision to upgrade to the next minor version of OKD. Upgrade channels only control release selection and have no impact on the version of the cluster you install; the openshift-install binary for a given patch level of OKD always installs that patch level.

OKD Latest has the following upgrade channels to choose from:

  • candidate-Latest

  • fast-Latest

  • stable-Latest

candidate-Latest Channel

The candidate-Latest channel will contain candidate builds for a z-stream (Latest.z) release. Release candidates contain all the features of the product but they are not supported and should only be used to test feature acceptance and assist in qualifying the next version of OKD. A release candidate is any build (e.g. 4.3.0-rc.3, 4.3.0) that is available in the candidate channel. After a version lands in the candidate channel, it goes through more quality checks and if it meets the quality standard it is promoted to fast-Latest or stable-Latest channels. So if a given release is available in the candidate-Latest channel and also in the fast-Latest or stable-Latest channels, it is a Red Hat supported version. Additionally candidate-Latest may include dead end releases from which there are no or ever be recommended upgrades.

The candidate-Latest channel also allows upgrading from a previous minor version of OKD to Latest.

Release candidates differ from the nightly builds found on Nightly builds are available for early access to features but updating to or from nightly builds is neither recommended nor supported. Nightly builds are not avilable in any upgrade channel.

fast-Latest Channel

The fast-Latest channel is updated with new Latest patch versions as soon as Red Hat declares the given patch as a general availability release. As such, these releases are fully supported, are production quality and have performed well while available as a release candidate in the candidate-Latest channel from where they were promoted. Some time after a release appears in fast-Latest it will also appear in stable-Latest. Releases will never appear in stable-Latest before they appear in fast-Latest.

The fast-Latest channel also allows upgrading from a previous minor version of OKD.

stable-Latest Channel

Like the fast-Latest channel, the stable-Latest channel will only contain releases that have been declared general availability are therefore fully supported. However the stable-Latest channel will gradually roll out releases to customers based on data from our SRE teams, support services, and pre-production and production environments that participate in our connected customer program, rather than being immediately available as they are in the fast-Latest channel. For patch and CVE fixes this delay can range from several hours to a day and allows an extra period of assessment in how the software performs.

The stable-Latest channel also allows upgrading from a previous minor version of OKD.

Upgrade Version Paths

OKD maintains an upgrade recommendation service that understands the version of OKD you have installed as well as the path to take within the channel you choose to get you to the next release. You can imagine seeing the following in the fast-Latest channel:

  • Latest.0

  • Latest.1

  • Latest.3

  • Latest.4

The service only recommends upgrades that have been tested and have no serious issues. If you are on Latest.1 and OKD is allowing you to select Latest.4, then it is safe for you to go from .Latest.1 to .Latest.4. Likewise, the absence of Latest.2 may be due to a CVE that was fixed in Latest.3 and Red Hat no longer suggests upgrading to a known vulnerable version.

Update stability depends on your channel. The presence of an update recommendation in the candidate-Latest channel does not imply that the update is supported. It means that no serious issues have been found with the update yet, but there may not be significant traffic through the update to suggest stability. The presence of an update recommendation in the fast-Latest or stable-Latest channels is a declaration that the update is fully supported while it is in the channel. While releases will never be removed from a channel, update recommendations which exhibit serious issues will be removed from all channels. Updates initiated after the update recommendation has been removed may not be supported.

Red Hat will eventually provide supported update paths from any supported (fast-Latest or stable-Latest) release to the latest release in Latest.z, although there may be delays while safe paths away from troubled releases are constructed and verified.

Fast and Stable Channel Usage and Strategies

The fast-Latest and stable-Latest channels present a choice between receiving general availability releases as soon as they are available or allowing Red Hat to control the rollout of those updates. If issues are detected during rollout or at a later time, upgrades to that version may be blocked in both the fast-Latest and stable-Latest channels, and a new version may be introduced that will be the new preferred upgrade target.

Customers can improve this process by configuring pre-production systems on the fast-Latest channel, production systems on the stable-Latest channel, and participating in Red Hat’s connected customer program. This program allows Red Hat to observe the impact of updates on your specific hardware and software configurations. Future releases may improve or alter the pace at which updates move from the fast-Latest to the stable-Latest channel.

Restricted Network Clusters

Customers who have chosen to not be connected to Red Hat and are controlling their own OKD container image content manually should consult the Red Hat errata associated with product releases and note any comments impacting upgrades. During upgrade the user interface may caution about switching between these versions and it is up to the customer to ensure they have correctly selected the appropriate version before bypassing those cautions.

Switching Between Channels

It is supported for customers to switch from the stable-Latest channel to the fast-Latest at any time. Customers can also switch to the candidate-Latest at any time but should be aware that some releases in the candidate-Latest channel may be release candidates and therefore not supported. A customer can switch from candidate-Latest to fast-Latest if their current release is a general availability release. Customers can always switch from fast-Latest to stable-Latest, although there may be a delay of up to a day if their current release was recently promoted to fast-{product-stable} while they wait for the release to be promoted to stable-Latest. If you change to a channel that does not include your current release, an alert will fire, no updates will be recommended, and you can safely change back to your original channel at any point.

Please refer to the candidate-Latest Channel section above to understand the distinction between a release candidate and a general avilability release.

Updating a cluster by using the CLI

If updates are available, you can update your cluster by using the OpenShift CLI (oc).

You can find information about available OKD advisories and updates in the errata section of the Customer Portal.

  • Install the version of the OpenShift Command-line Interface (CLI), commonly known as oc, that matches the version for your updated version.

  • Log in to the cluster as user with cluster-admin privileges.

  • Install the jq package.

  1. Ensure that your cluster is available:

    $ oc get clusterversion
    version   4.3.0     True        False         158m    Cluster version is 4.3.0
  2. Review the current update channel information and confirm that your channel is set to stable-4.3:

    $ oc get clusterversion -o json|jq ".items[0].spec"
      "channel": "stable-4.3",
      "clusterID": "990f7ab8-109b-4c95-8480-2bd1deec55ff",
      "upstream": ""

    For production clusters, you must subscribe to a stable-* channel.

  3. View the available updates and note the version number of the update that you want to apply:

    $ oc adm upgrade
    Cluster version is 4.1.0
  4. Apply an update:

    • To update to the latest version:

      $ oc adm upgrade --to-latest=true (1)
    • To update to a specific version:

      $ oc adm upgrade --to=<version> (1)
      1 <version> is the update version that you obtained from the output of the previous command.
  5. Review the status of the Cluster Version Operator:

    $ oc get clusterversion -o json|jq ".items[0].spec"
      "channel": "stable-4.3",
      "clusterID": "990f7ab8-109b-4c95-8480-2bd1deec55ff",
      "desiredUpdate": {
        "force": false,
        "image": "",
        "version": "4.3.1" (1)
      "upstream": ""
    1 If the version number in the desiredUpdate stanza matches the value that you specified, the update is in progress.
  6. Review the cluster version status history to monitor the status of the update. It might take some time for all the objects to finish updating.

    $ oc get clusterversion -o json|jq ".items[0].status.history"
        "completionTime": null,
        "image": "",
        "startedTime": "2019-06-19T20:30:50Z",
        "state": "Partial",
        "verified": true,
        "version": "4.1.2"
        "completionTime": "2019-06-19T20:30:50Z",
        "image": "",
        "startedTime": "2019-06-19T17:38:10Z",
        "state": "Completed",
        "verified": false,
        "version": "4.1.0"

    The history contains a list of the most recent versions applied to the cluster. This value is updated when the CVO applies an update. The list is ordered by date, where the newest update is first in the list. Updates in the history have state Completed if the rollout completed and Partial if the update failed or did not complete.

    If an upgrade fails, the Operator stops and reports the status of the failing component. Rolling your cluster back to a previous version is not supported. If your upgrade fails, contact Red Hat support.

  7. After the update completes, you can confirm that the cluster version has updated to the new version:

    $ oc get clusterversion
    version   4.1.2       True        False         2m        Cluster version is 4.1.2