ManagedOSImage reference

The ManagedOSImage resource is responsible of defining an OS image or image version that needs to be applied to each node in a set of targeted Clusters.
Once created, the ManagedOSImage resource can be updated with a new osImage or managedOSVersionName to trigger a new upgrade.
Similarly, an existing ManagedOSImage can be updated to target new Clusters.

There are several keys that can be configured under a ManagedOSImage resource spec.

upgrade-cluster-target.yaml

apiVersion: elemental.cattle.io/v1beta1
kind: ManagedOSImage
metadata:
  name: my-upgrade
  namespace: fleet-default
spec:
  # Set to the new Elemental version you would like to upgrade to or track the latest tag
  osImage: "registry.suse.com/suse/sle-micro/5.5:latest"
  clusterTargets:
    - clusterName: my-cluster

ManagedOSImageSpec reference

Key	Type	Default value	Description
osImage	string	empty	The fully qualified image to upgrade nodes to. This value has priority over managedOSVersionName if both are configured.
managedOSVersionName	string	empty	The name of a ManagedOSVersion to upgrade nodes to.
cloudConfig	object	null	A cloud-init config to apply to the nodes during upgrades. See reference.
nodeSelector	object	null	This selector can be used to target specific nodes within the clusterTargets. See reference.
concurrency	int	1	How many nodes within the same cluster should be upgraded at the same time.
cordon	bool	true	Set this to true if the nodes should be cordoned before applying the upgrade. Ineffective when drain is also configured.
drain	object	See ref	Configure if and how nodes should be drained before applying the upgrade. See reference.
prepare	object	null	The prepare init container, if specified, is run before cordon/drain which is run before the upgrade container. See reference.
upgradeContainer	object	null	The upgrade container that will run the upgrade on the nodes. See reference.
clusterRolloutStrategy	object	null	RolloverStrategy controls the rollout of the upgrade bundle across clusters. See reference.
clusterTargets	list	null	Declares clusters to deploy the upgrade plan to. See reference.

cloudConfig

This describes a cloud-init config that will be copied to each upgraded node to the /oem/90_operator.yaml path.
This cloud-init config will be applied by the system after reboot.
For more information and examples, see the MachineRegistration spec.config.cloud-config reference.

nodeSelector

This Label Selector can be used to restrict the upgrades to only a certain set of nodes within the targeted Clusters.

Example

nodeSelector:
  matchExpressions:
    - {key: my-node/label, operator: Exists}

drain

Configure if and how nodes should be drained.
To disable drain during upgrades you can configure this field to null.
Drain is enabled by default.

The drain settings directly translates to the kubectl drain command being executed on the node before upgrade.

Key	Type	Default value	Description
timeout	time.Duration	null	The length of time to wait before giving up draining a node, zero means infinite.
gracePeriod	int	null	Period of time in seconds given to each pod to terminate gracefully. If negative, the default value specified in the pod will be used.
deleteEmptydirData	bool	true	Continue even if there are pods using emptyDir (local data that will be deleted when the node is drained).
ignoreDaemonSets	bool	true	Ignore DaemonSet-managed pods.
force	bool	true	Continue even if there are pods that do not declare a controller.
disableEviction	bool	false	Force drain to use delete, even if eviction is supported. This will bypass checking PodDisruptionBudgets, use with caution.
skipWaitForDeleteTimeout	int	60	If pod DeletionTimestamp older than N seconds, skip waiting for the pod. Seconds must be greater than 0 to skip.
podSelector	label selector	null	Label selector to filter pods on the node. Only selected pods will be evicted.

prepare

Defines a prepare Init container that is ran before the upgrade container executing the upgrade job on a node.
The keys directly translate to the container specification.
Note that the node filesystem is mounted at /host inside the container.

Key	Type	Default value	Description
image	string	empty	Container image name.
command	list	empty	Entrypoint array.
args	list	empty	Arguments to the entrypoint.
env	list	empty	List of environment variables to set in the container.
envFrom	list	empty	List of sources to populate environment variables in the container.
volumes	list	empty	List of hostPath volumes. See reference.
securityContext	object	null	The security options the ephemeral container should be run with.

prepare.volumes

Each volume definition will translate to a hostPath volume (source) which will be mounted in the container (destination).
Note that by default the host root filesystem / will always be mounted at /host.

Key	Type	Default value	Description
name	string	empty	Volume name.
source	string	empty	HostPath volume path.
destination	string	empty	HostPath volume mount path.

Example

volumes:
  - name: my-custom-volume
    source: /foo
    destination: /foo

upgradeContainer

Defines the upgrade container executing the upgrade job on a node.
The keys directly translate to the container specification.
Note that the node filesystem is mounted at /host inside the container.

warning

When using any Elemental or Elemental based image you are expected to only edit the env key to optionally set the FORCE, UPGRADE_RECOVERY, or UPGRADE_RECOVERY_ONLY variables.
For more info you can read the upgrade documentation.
Any other change to the upgradeContainer may result in issues during upgrades.

clusterRolloutStrategy

This controls the rollout of the bundle across clusters.
For more information you can read the reference documentation.

clusterTargets

Select Clusters to be targeted for the OS image upgrade.
For more information you can read the reference documentation.

Example

clusterTargets:
  - clusterName: volcano