×

Prerequisites

  • You must have the oc command-line interface (CLI) tool installed.

  • You must provision a local container image registry with the container images for your update, as described in Mirroring the OKD image repository.

  • You must have access to the cluster as a user with admin privileges. See Using RBAC to define and apply permissions.

  • You must have a recent etcd backup in case your update fails and you must restore your cluster to a previous state.

  • You must ensure that all machine config pools (MCPs) are running and not paused. Nodes associated with a paused MCP are skipped during the update process. You can pause the MCPs if you are performing a canary rollout update strategy.

  • If your cluster uses manually maintained credentials, update the cloud provider resources for the new release. For more information, including how to determine if this is a requirement for your cluster, see Preparing to update a cluster with manually maintained credentials.

  • If you run an Operator or you have configured any application with the pod disruption budget, you might experience an interruption during the upgrade process. If minAvailable is set to 1 in PodDisruptionBudget, the nodes are drained to apply pending machine configs which might block the eviction process. If several nodes are rebooted, all the pods might run on only one node, and the PodDisruptionBudget field can prevent the node drain.

If you run an Operator or you have configured any application with the pod disruption budget, you might experience an interruption during the upgrade process. If minAvailable is set to 1 in PodDisruptionBudget, the nodes are drained to apply pending machine configs which might block the eviction process. If several nodes are rebooted, all the pods might run on only one node, and the PodDisruptionBudget field can prevent the node drain.

Pausing a MachineHealthCheck resource

During the upgrade process, nodes in the cluster might become temporarily unavailable. In the case of worker nodes, the machine health check might identify such nodes as unhealthy and reboot them. To avoid rebooting such nodes, pause all the MachineHealthCheck resources before updating the cluster.

Prerequisites
  • Install the OpenShift CLI (oc).

Procedure
  1. To list all the available MachineHealthCheck resources that you want to pause, run the following command:

    $ oc get machinehealthcheck -n openshift-machine-api
  2. To pause the machine health checks, add the cluster.x-k8s.io/paused="" annotation to the MachineHealthCheck resource. Run the following command:

    $ oc -n openshift-machine-api annotate mhc <mhc-name> cluster.x-k8s.io/paused=""

    The annotated MachineHealthCheck resource resembles the following YAML file:

    apiVersion: machine.openshift.io/v1beta1
    kind: MachineHealthCheck
    metadata:
      name: example
      namespace: openshift-machine-api
      annotations:
        cluster.x-k8s.io/paused: ""
    spec:
      selector:
        matchLabels:
          role: worker
      unhealthyConditions:
      - type:    "Ready"
        status:  "Unknown"
        timeout: "300s"
      - type:    "Ready"
        status:  "False"
        timeout: "300s"
      maxUnhealthy: "40%"
    status:
      currentHealthy: 5
      expectedMachines: 5

    Resume the machine health checks after updating the cluster. To resume the check, remove the pause annotation from the MachineHealthCheck resource by running the following command:

    $ oc -n openshift-machine-api annotate mhc <mhc-name> cluster.x-k8s.io/paused-

Retrieving a release image digest

In order to update a cluster in a disconnected environment using the oc adm upgrade command with the --to-image option, you must reference the sha256 digest that corresponds to your targeted release image.

Procedure
  1. Run the following command on a device that is connected to the internet:

    $ oc adm release info -o 'jsonpath={.digest}{"\n"}' quay.io/openshift-release-dev/ocp-release:${OCP_RELEASE_VERSION}-${ARCHITECTURE}

    For {OCP_RELEASE_VERSION}, specify the version of OKD to which you want to update, such as 4.10.16.

    For {ARCHITECTURE}, specify the architecture of the cluster, such as x86_64, aarch64, s390x, or ppc64le.

    Example output
    sha256:a8bfba3b6dddd1a2fbbead7dac65fe4fb8335089e4e7cae327f3bad334add31d
  2. Copy the sha256 digest for use when updating your cluster.

Updating the disconnected cluster

Update the disconnected cluster to the OKD version that you downloaded the release images for.

If you have a local OpenShift Update Service, you can update by using the connected web console or CLI instructions instead of this procedure.

Prerequisites
  • You mirrored the images for the new release to your registry.

  • You applied the release image signature ConfigMap for the new release to your cluster.

    The release image signature config map allows the Cluster Version Operator (CVO) to ensure the integrity of release images by verifying that the actual image signatures match the expected signatures.

  • You obtained the sha256 digest for your targeted release image.

  • You installed the OpenShift CLI (oc).

  • You paused all MachineHealthCheck resources.

Procedure
  • Update the cluster:

    $ oc adm upgrade --allow-explicit-upgrade --to-image <defined_registry>/<defined_repository>@<digest>

    Where:

    <defined_registry>

    Specifies the name of the mirror registry you mirrored your images to.

    <defined_repository>

    Specifies the name of the image repository you want to use on the mirror registry.

    <digest>

    Specifies the sha256 digest for the targeted release image, for example, sha256:81154f5c03294534e1eaf0319bef7a601134f891689ccede5d705ef659aa8c92.

    • See "Mirroring OKD images" to review how your mirror registry and repository names are defined.

    • If you used an ImageContentSourcePolicy or ImageDigestMirrorSet, you can use the canonical registry and repository names instead of the names you defined. The canonical registry name is quay.io and the canonical repository name is openshift-release-dev/ocp-release.

    • You can only configure global pull secrets for clusters that have an ImageContentSourcePolicy object. You cannot add a pull secret to a project.

Additional resources

Configuring image registry repository mirroring

Setting up container registry repository mirroring enables you to do the following:

  • Configure your OKD cluster to redirect requests to pull images from a repository on a source image registry and have it resolved by a repository on a mirrored image registry.

  • Identify multiple mirrored repositories for each target repository, to make sure that if one mirror is down, another can be used.

The attributes of repository mirroring in OKD include:

  • Image pulls are resilient to registry downtimes.

  • Clusters in disconnected environments can pull images from critical locations, such as quay.io, and have registries behind a company firewall provide the requested images.

  • A particular order of registries is tried when an image pull request is made, with the permanent registry typically being the last one tried.

  • The mirror information you enter is added to the /etc/containers/registries.conf file on every node in the OKD cluster.

  • When a node makes a request for an image from the source repository, it tries each mirrored repository in turn until it finds the requested content. If all mirrors fail, the cluster tries the source repository. If successful, the image is pulled to the node.

Setting up repository mirroring can be done in the following ways:

  • At OKD installation:

    By pulling container images needed by OKD and then bringing those images behind your company’s firewall, you can install OKD into a datacenter that is in a disconnected environment.

  • After OKD installation:

    Even if you don’t configure mirroring during OKD installation, you can do so later using the ImageContentSourcePolicy object.

The following procedure provides a postinstallation mirror configuration, where you create an ImageContentSourcePolicy object that identifies:

  • The source of the container image repository you want to mirror.

  • A separate entry for each mirror repository you want to offer the content requested from the source repository.

You can only configure global pull secrets for clusters that have an ImageContentSourcePolicy object. You cannot add a pull secret to a project.

Prerequisites
  • Access to the cluster as a user with the cluster-admin role.

Procedure
  1. Configure mirrored repositories, by either:

    • Setting up a mirrored repository with Red Hat Quay, as described in Red Hat Quay Repository Mirroring. Using Red Hat Quay allows you to copy images from one repository to another and also automatically sync those repositories repeatedly over time.

    • Using a tool such as skopeo to copy images manually from the source directory to the mirrored repository.

      For example, after installing the skopeo RPM package on a Red Hat Enterprise Linux (RHEL) 7 or RHEL 8 system, use the skopeo command as shown in this example:

      $ skopeo copy \
      docker://registry.access.redhat.com/ubi8/ubi-minimal@sha256:5cfbaf45ca96806917830c183e9f37df2e913b187adb32e89fd83fa455ebaa6 \
      docker://example.io/example/ubi-minimal

      In this example, you have a container image registry that is named example.io with an image repository named example to which you want to copy the ubi8/ubi-minimal image from registry.access.redhat.com. After you create the registry, you can configure your OKD cluster to redirect requests made of the source repository to the mirrored repository.

  2. Log in to your OKD cluster.

  3. Create an ImageContentSourcePolicy file (for example, registryrepomirror.yaml), replacing the source and mirrors with your own registry and repository pairs and images:

    apiVersion: operator.openshift.io/v1alpha1
    kind: ImageContentSourcePolicy
    metadata:
      name: ubi8repo
    spec:
      repositoryDigestMirrors:
      - mirrors:
        - example.io/example/ubi-minimal (1)
        - example.com/example/ubi-minimal (2)
        source: registry.access.redhat.com/ubi8/ubi-minimal (3)
      - mirrors:
        - mirror.example.com/redhat
        source: registry.redhat.io/openshift4 (4)
      - mirrors:
        - mirror.example.com
        source: registry.redhat.io (5)
      - mirrors:
        - mirror.example.net/image
        source: registry.example.com/example/myimage (6)
      - mirrors:
        - mirror.example.net
        source: registry.example.com/example (7)
      - mirrors:
        - mirror.example.net/registry-example-com
        source: registry.example.com (8)
    1 Indicates the name of the image registry and repository.
    2 Indicates multiple mirror repositories for each target repository. If one mirror is down, the target repository can use another mirror.
    3 Indicates the registry and repository containing the content that is mirrored.
    4 You can configure a namespace inside a registry to use any image in that namespace. If you use a registry domain as a source, the ImageContentSourcePolicy resource is applied to all repositories from the registry.
    5 If you configure the registry name, the ImageContentSourcePolicy resource is applied to all repositories from a source registry to a mirror registry.
    6 Pulls the image mirror.example.net/image@sha256:…​.
    7 Pulls the image myimage in the source registry namespace from the mirror mirror.example.net/myimage@sha256:…​.
    8 Pulls the image registry.example.com/example/myimage from the mirror registry mirror.example.net/registry-example-com/example/myimage@sha256:…​. The ImageContentSourcePolicy resource is applied to all repositories from a source registry to a mirror registry mirror.example.net/registry-example-com.
  4. Create the new ImageContentSourcePolicy object:

    $ oc create -f registryrepomirror.yaml

    After the ImageContentSourcePolicy object is created, the new settings are deployed to each node and the cluster starts using the mirrored repository for requests to the source repository.

  5. To check that the mirrored configuration settings, are applied, do the following on one of the nodes.

    1. List your nodes:

      $ oc get node
      Example output
      NAME                           STATUS                     ROLES    AGE  VERSION
      ip-10-0-137-44.ec2.internal    Ready                      worker   7m   v1.25.0
      ip-10-0-138-148.ec2.internal   Ready                      master   11m  v1.25.0
      ip-10-0-139-122.ec2.internal   Ready                      master   11m  v1.25.0
      ip-10-0-147-35.ec2.internal    Ready                      worker   7m   v1.25.0
      ip-10-0-153-12.ec2.internal    Ready                      worker   7m   v1.25.0
      ip-10-0-154-10.ec2.internal    Ready                      master   11m  v1.25.0

      The Imagecontentsourcepolicy resource does not restart the nodes.

    2. Start the debugging process to access the node:

      $ oc debug node/ip-10-0-147-35.ec2.internal
      Example output
      Starting pod/ip-10-0-147-35ec2internal-debug ...
      To use host binaries, run `chroot /host`
    3. Change your root directory to /host:

      sh-4.2# chroot /host
    4. Check the /etc/containers/registries.conf file to make sure the changes were made:

      sh-4.2# cat /etc/containers/registries.conf
      Example output
      unqualified-search-registries = ["registry.access.redhat.com", "docker.io"]
      short-name-mode = ""
      
      [[registry]]
        prefix = ""
        location = "registry.access.redhat.com/ubi8/ubi-minimal"
        mirror-by-digest-only = true
      
        [[registry.mirror]]
          location = "example.io/example/ubi-minimal"
      
        [[registry.mirror]]
          location = "example.com/example/ubi-minimal"
      
      [[registry]]
        prefix = ""
        location = "registry.example.com"
        mirror-by-digest-only = true
      
        [[registry.mirror]]
          location = "mirror.example.net/registry-example-com"
      
      [[registry]]
        prefix = ""
        location = "registry.example.com/example"
        mirror-by-digest-only = true
      
        [[registry.mirror]]
          location = "mirror.example.net"
      
      [[registry]]
        prefix = ""
        location = "registry.example.com/example/myimage"
        mirror-by-digest-only = true
      
        [[registry.mirror]]
          location = "mirror.example.net/image"
      
      [[registry]]
        prefix = ""
        location = "registry.redhat.io"
        mirror-by-digest-only = true
      
        [[registry.mirror]]
          location = "mirror.example.com"
      
      [[registry]]
        prefix = ""
        location = "registry.redhat.io/openshift4"
        mirror-by-digest-only = true
      
        [[registry.mirror]]
          location = "mirror.example.com/redhat"
    5. Pull an image digest to the node from the source and check if it is resolved by the mirror. ImageContentSourcePolicy objects support image digests only, not image tags.

      sh-4.2# podman pull --log-level=debug registry.access.redhat.com/ubi8/ubi-minimal@sha256:5cfbaf45ca96806917830c183e9f37df2e913b187adb32e89fd83fa455ebaa6
Troubleshooting repository mirroring

If the repository mirroring procedure does not work as described, use the following information about how repository mirroring works to help troubleshoot the problem.

  • The first working mirror is used to supply the pulled image.

  • The main registry is only used if no other mirror works.

  • From the system context, the Insecure flags are used as fallback.

  • The format of the /etc/containers/registries.conf file has changed recently. It is now version 2 and in TOML format.

Widening the scope of the mirror image catalog to reduce the frequency of cluster node reboots

You can scope the mirrored image catalog at the repository level or the wider registry level. A widely scoped ImageContentSourcePolicy resource reduces the number of times the nodes need to reboot in response to changes to the resource.

To widen the scope of the mirror image catalog in the ImageContentSourcePolicy resource, perform the following procedure.

Prerequisites
  • Install the OKD CLI oc.

  • Log in as a user with cluster-admin privileges.

  • Configure a mirrored image catalog for use in your disconnected cluster.

Procedure
  1. Run the following command, specifying values for <local_registry>, <pull_spec>, and <pull_secret_file>:

    $ oc adm catalog mirror <local_registry>/<pull_spec> <local_registry> -a <pull_secret_file> --icsp-scope=registry

    where:

    <local_registry>

    is the local registry you have configured for your disconnected cluster, for example, local.registry:5000.

    <pull_spec>

    is the pull specification as configured in your disconnected registry, for example, redhat/redhat-operator-index:v4.12

    <pull_secret_file>

    is the registry.redhat.io pull secret in .json file format. You can download the pull secret from the Red Hat OpenShift Cluster Manager.

    The oc adm catalog mirror command creates a /redhat-operator-index-manifests directory and generates imageContentSourcePolicy.yaml, catalogSource.yaml, and mapping.txt files.

  2. Apply the new ImageContentSourcePolicy resource to the cluster:

    $ oc apply -f imageContentSourcePolicy.yaml
Verification
  • Verify that oc apply successfully applied the change to ImageContentSourcePolicy:

    $ oc get ImageContentSourcePolicy -o yaml
    Example output
    apiVersion: v1
    items:
    - apiVersion: operator.openshift.io/v1alpha1
      kind: ImageContentSourcePolicy
      metadata:
        annotations:
          kubectl.kubernetes.io/last-applied-configuration: |
            {"apiVersion":"operator.openshift.io/v1alpha1","kind":"ImageContentSourcePolicy","metadata":{"annotations":{},"name":"redhat-operator-index"},"spec":{"repositoryDigestMirrors":[{"mirrors":["local.registry:5000"],"source":"registry.redhat.io"}]}}
    ...

After you update the ImageContentSourcePolicy resource, OKD deploys the new settings to each node and the cluster starts using the mirrored repository for requests to the source repository.