Updating a cluster using the CLI - Performing a cluster update | Updating clusters

Prerequisites
Pausing a MachineHealthCheck resource
About updating single node OKD
Updating a cluster by using the CLI
Cluster update status using oc adm upgrade status
Changing the update server by using the CLI

You can perform minor version and patch updates on an OKD cluster by using the OpenShift CLI (oc).

Prerequisites

Have access to the cluster as a user with admin privileges. See Using RBAC to define and apply permissions.
Have a recent etcd backup in case your update fails and you must restore your cluster to a previous state.
Have a recent Container Storage Interface (CSI) volume snapshot in case you need to restore persistent volumes due to a pod failure.
Your Fedora7 workers are replaced with Fedora8 or FCOS workers. Red Hat does not support in-place Fedora7 to Fedora8 updates for Fedora workers; those hosts must be replaced with a clean operating system install.
You have updated all Operators previously installed through Operator Lifecycle Manager (OLM) to a version that is compatible with your target release. Updating the Operators ensures they have a valid update path when the default software catalogs switch from the current minor version to the next during a cluster update. See Updating installed Operators for more information on how to check compatibility and, if necessary, update the installed Operators.
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.
Ensure that you address all Upgradeable=False conditions so the cluster allows an update to the next minor version. An alert displays at the top of the Cluster Settings page when you have one or more cluster Operators that cannot be updated. You can still update to the next available patch update for the minor release you are currently on.
If you run an Operator or you have configured any application with the pod disruption budget, you might experience an interruption during the update 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.

When an update is failing to complete, the Cluster Version Operator (CVO) reports the status of any blocking components while attempting to reconcile the update. Rolling your cluster back to a previous version is not supported. If your update is failing to complete, contact Red Hat support.
Using the unsupportedConfigOverrides section to modify the configuration of an Operator is unsupported and might block cluster updates. You must remove this setting before you can update your cluster.

Additional resources

Support policy for unmanaged Operators

Pausing a MachineHealthCheck resource

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

Some MachineHealthCheck resources might not need to be paused. If your MachineHealthCheck resource relies on unrecoverable conditions, pausing that MHC is unnecessary.

Prerequisites

Install the OpenShift CLI (oc).

Procedure

To list all the available MachineHealthCheck resources that you want to pause, run the following command:
```
$ oc get machinehealthcheck -n openshift-machine-api
```

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-

About updating single node OKD

You can update, or upgrade, a single-node OKD cluster by using either the console or CLI.

However, note the following limitations:

The prerequisite to pause the MachineHealthCheck resources is not required because there is no other node to perform the health check.
Restoring a single-node OKD cluster using an etcd backup is not officially supported. However, it is good practice to perform the etcd backup in case your update fails. If your control plane is healthy, you might be able to restore your cluster to a previous state by using the backup.
Updating a single-node OKD cluster requires downtime and can include an automatic reboot. The amount of downtime depends on the update payload, as described in the following scenarios:
- If the update payload contains an operating system update, which requires a reboot, the downtime is significant and impacts cluster management and user workloads.
- If the update contains machine configuration changes that do not require a reboot, the downtime is less, and the impact on the cluster management and user workloads is lessened. In this case, the node draining step is skipped with single-node OKD because there is no other node in the cluster to reschedule the workloads to.
- If the update payload does not contain an operating system update or machine configuration changes, a short API outage occurs and resolves quickly.

There are conditions, such as bugs in an updated package, that can cause the single node to not restart after a reboot. In this case, the update does not rollback automatically.

Additional resources

For information on which machine configuration changes require a reboot, see the note in About the Machine Config Operator.

Updating a cluster by using the CLI

You can use the OpenShift CLI (oc) to review and request cluster updates.

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

Prerequisites

Install the OpenShift CLI (oc) that matches the version for your updated version.
Log in to the cluster as user with cluster-admin privileges.
Pause all MachineHealthCheck resources.

Procedure

View the available updates and note the version number of the update that you want to apply:
```
$ oc adm upgrade recommend
```

Example output

...

Upstream update service: https://amd64.origin.releases.ci.openshift.org/graph
Channel: stable-scos-4

Updates to 4.20:
  VERSION               ISSUES
  4.20.0-okd-scos.ec.14 no known issues relevant to this cluster

Updates to 4.19:
  VERSION            ISSUES
  4.19.0-okd-scos.17 no known issues relevant to this cluster

You can use the --version flag to determine whether a specific version is recommended for your update. If there are no recommended updates, updates that have known issues might still be available.

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)

<version> is the update version that you obtained from the output of the oc adm upgrade recommend command.

If the cluster administrator evaluates the potential known risks and decides it is acceptable for the current cluster, then the administrator can waive the safety guards and proceed with the update by running the following command:

$ oc adm upgrade --allow-not-recommended --to <version>

When using oc adm upgrade --help, there is a listed option for --force. This is heavily discouraged, as using the --force option bypasses cluster-side guards, including release verification and precondition checks. Using --force does not guarantee a successful update. Bypassing guards put the cluster at risk.

Optional: Review the status of the Cluster Version Operator by running the following command:

$ oc adm upgrade status

To monitor the update in real time, run oc adm upgrade status in a watch utility.

Example output

info: An upgrade is in progress. Working towards 4.14.0-0.okd-2024-01-06-084517: 117 of 864 done (13% complete), waiting on etcd, kube-apiserver

Upstream: https://amd64.origin.releases.ci.openshift.org/graph
Channel: stable-4
No updates available. You may still upgrade to a specific release image with --to-image or wait for new updates to be available.

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

$ oc adm upgrade

Example output

Cluster version is 4.14.0-0.okd-2024-01-06-084517

Upstream: https://amd64.origin.releases.ci.openshift.org/graph
Channel: stable-4
No updates available. You may still upgrade to a specific release image with --to-image or wait for new updates to be available.

If you are updating your cluster to the next minor version, such as version X.y to X.(y+1), it is recommended to confirm that your nodes are updated before deploying workloads that rely on a new feature:

$ oc get nodes

Example output

NAME                           STATUS   ROLES    AGE   VERSION
ip-10-0-168-251.ec2.internal   Ready    master   82m   v1.33.4
ip-10-0-170-223.ec2.internal   Ready    master   82m   v1.33.4
ip-10-0-179-95.ec2.internal    Ready    worker   70m   v1.33.4
ip-10-0-182-134.ec2.internal   Ready    worker   70m   v1.33.4
ip-10-0-211-16.ec2.internal    Ready    master   82m   v1.33.4
ip-10-0-250-100.ec2.internal   Ready    worker   69m   v1.33.4

Cluster update status using oc adm upgrade status

When updating your cluster, the oc adm upgrade command returns limited information about the status of your update. The cluster administrator can use the oc adm upgrade status command to decouple status information from the oc adm upgrade command and return specific information regarding a cluster update, including the status of the control plane and worker node updates. Worker is also known as compute.

The oc adm upgrade status command is read-only and does not alter any state in your cluster.

The oc adm upgrade status command can be used for clusters from version 4.12 up to the latest supported release.

The oc adm upgrade status command will output three sections, control plane update, worker nodes update, and health insights.

Control Plane Update: Displays details about the updating cluster control plane, contains a high-level assessment, completion status, duration estimate, or cluster operator health. The section also shows a table with control plane node update information.

The control plane update section can also show an additional table that lists cluster operators being updated if the --details=operators or --details-all flags are used. Please note that due the asynchronous distributed nature of OKD, an operator may appear in this section more than once during the update, or not at all. The section is only shown when a Cluster Operator is observed to be updating. It is normal during an update to observe no updating Cluster Operator at certain periods; not every performed action can be assigned to an observable updating Cluster Operator.
Worker Notes Update: Displays the worker node update information. The worker nodes section starts with a table that displays a summary of information about each worker pool configured in the cluster. Each non-empty worker pool output will show a dedicated table listing update information about nodes that belong to that pool. If a cluster does not have any worker nodes the output will not contain the worker node section. You can make the node tables show all lines by using --details=nodes or --details=all.
Health Insights: displays insights about states and events present in the cluster that may be relevant for the ongoing update. You can use --details=health to expand the items in this section into a more verbose form with more content such as documentation links, longer form descriptions, or cluster resources involved in the insight.

The oc adm upgrade status command is currently not supported on hosted control planes clusters.

The following is an example of the output you will see for an update progressing successfully:

= Control Plane =
Assessment:      Progressing
Target Version:  4.17.1 (from 4.17.0)
Updating:        machine-config
Completion:      97% (32 operators updated, 1 updating, 0 waiting)
Duration:        54m (Est. Time Remaining: <10m)
Operator Status: 32 Healthy, 1 Unavailable

Control Plane Nodes
NAME                                        ASSESSMENT    PHASE      VERSION   EST    MESSAGE
ip-10-0-53-40.us-east-2.compute.internal    Progressing   Draining   4.17.0    +10m
ip-10-0-30-217.us-east-2.compute.internal   Outdated      Pending    4.17.0    ?
ip-10-0-92-180.us-east-2.compute.internal   Outdated      Pending    4.17.0    ?

= Worker Upgrade =

WORKER POOL   ASSESSMENT    COMPLETION   STATUS
worker        Progressing   0% (0/2)     1 Available, 1 Progressing, 1 Draining
infra         Progressing   50% (1/2)    1 Available, 1 Progressing, 1 Draining

Worker Pool Nodes: Worker
NAME                                       ASSESSMENT    PHASE      VERSION   EST    MESSAGE
ip-10-0-4-159.us-east-2.compute.internal   Progressing   Draining   4.17.0    +10m
ip-10-0-99-40.us-east-2.compute.internal   Outdated      Pending    4.17.0    ?

Worker Pool Nodes: infra
NAME                                             ASSESSMENT    PHASE      VERSION   EST    MESSAGE
ip-10-0-4-159-infra.us-east-2.compute.internal   Progressing   Draining   4.17.0    +10m
ip-10-0-20-162.us-east-2.compute.internal        Completed     Updated    4.17.1    -

= Update Health =
SINCE   LEVEL   IMPACT   MESSAGE
54m4s   Info    None     Update is proceeding well

Changing the update server by using the CLI

Additional resources

Changing the update server is optional. If you have an OpenShift Update Service (OSUS) installed and configured locally, you must set the URL for the server as the upstream to use the local server during updates. The default value for upstream is https://api.openshift.com/api/upgrades_info/v1/graph.

Procedure

Change the upstream parameter value in the cluster version:

$ oc patch clusterversion/version --patch '{"spec":{"upstream":"<update-server-url>"}}' --type=merge

The <update-server-url> variable specifies the URL for the update server.

Example output

clusterversion.config.openshift.io/version patched