[release-4.20] OCPNODE-3722: Enforce OCP 4.20 and earlier cluster to have AutoSizingReserved disabled by default

[ngopalak-redhat]

Fixes: #OCPNODE-3722

- What I did
This patch introduces the 50-master-auto-sizing-disabled MachineConfig to OpenShift 4.20 clusters, setting the NODE_SIZING_ENABLED flag to false by default on master and worker nodes.
This change is required as we are making auto sizing enabled by default for cluster created using 4.21 and above.

Summary of changes

• Enforce Default Autosizing: Ensures that clusters created in 4.20 will retain the pre-4.21 behavior of having auto node sizing disabled by default.

• Upgrade Pre-requisite: This patch is a mandatory requirement for upgrading 4.20 clusters to 4.21. Changes to Cincinnati (OCPNODE-3722: Set minimum version of 4.20 required to upgrade to 4.21 cincinnati-graph-data#8277) will enforce that this patch must be present before the upgrade path to 4.21 is started.

• User Override (Priority): The MachineConfig uses the prefix 01- to ensure it sets the initial default. If a user has already created a KubeletConfig to explicitly enable autoSizing (as per the KubeletConfig documentation), that explicit user configuration will take precedence (override this default) and will be retained when upgrading to 4.21.

Reference: This change addresses the shift in default behavior introduced in OpenShift 4.21, where NODE_SIZING_ENABLED is set to true for all new clusters: #5390

Additional Notes for Developers
The approach taken in this PR is patterned after the change implemented in #4715, which was used to modify the default container runtime.

Rejected Alternatives
We explored several alternative solutions, but they were not feasible:

• In-Place Upgrade Handling: We found that direct handling during the 4.21 upgrade was unreliable. After multiple upgrade cycles, there was no consistent mechanism to identify clusters originally provisioned before 4.21.

• Changing the Default File: Switching the default configuration file (e.g., away from /etc/node-sizing-enabled) was overly complex, requiring us to manually manage legacy configuration paths for existing clusters.

• Installer-Created KubeletConfig: Since OpenShift clusters do not contain a default KubeletConfig resource, one option was to have the installer create it. This was rejected because Hypershift deployments may bypass the standard OCP installer.

• Adding a Default KubeletConfig Resource: This approach was dismissed because OpenShift allows only a single KubeletConfig per cluster. Introducing a default resource risks a user's explicit KubeletConfig unintentionally overriding the system default, leading to confusion.

- How to verify it

• Verified the patch on a 4.20 cluster: Created a cluster using ClusterBot, applied the patch via oc adm upgrade, confirmed the new MachineConfig was created, and ensured auto node sizing was disabled.

• Direct Patch Verification: Created a cluster using ClusterBot with the patch applied and confirmed auto node sizing was disabled.

• User Override Test: Created a KubeletConfig to explicitly enable auto sizing and verified that the setting was correctly enabled (overriding the default).

• Upgrade Path Validation: Successfully upgraded the patched cluster to 4.21 (using the above referenced 4.21 PR changes). Confirmed that auto node sizing remained disabled for upgraded clusters that had not been explicitly configured otherwise.

- Description for the changelog

Introduces the auto sizing MachineConfig, ensuring the feature remains disabled by default during upgrade

[openshift-ci]

Skipping CI for Draft Pull Request.
If you want CI signal for your change, please convert it to an actual PR.
You can still manually trigger a test run with /test all

[ngopalak-redhat]

/test all

[openshift-ci-robot]

@ngopalak-redhat: This pull request references OCPNODE-3718 which is a valid jira issue.

Warning: The referenced jira issue has an invalid target version for the target branch this PR targets: expected the epic to target the "4.20.z" version, but no target version was set.

In response to this:

- What I did
This patch introduces the 01-master-auto-sizing-disabled MachineConfig to OpenShift 4.20 clusters, setting the NODE_SIZING_ENABLED flag to false by default on master and worker nodes.
This change is required as we are making auto sizing enabled by default for cluster created using 4.21 and above.

Summary of changes

• Enforce Default Autosizing: Ensures that clusters created in 4.20 will retain the pre-4.21 behavior of having auto node sizing disabled by default.

• Upgrade Pre-requisite: This patch is a mandatory requirement for upgrading 4.20 clusters to 4.21. Changes to Cincinnati (TODO) will enforce that this patch must be present before the upgrade path to 4.21 is started.

• User Override (Priority): The MachineConfig uses the prefix 01- to ensure it sets the initial default. If a user has already created a KubeletConfig to explicitly enable autoSizing (as per the KubeletConfig documentation), that explicit user configuration will take precedence (override this default) and will be retained when upgrading to 4.21.

Reference: This change addresses the shift in default behavior introduced in OpenShift 4.21, where NODE_SIZING_ENABLED is set to true for all new clusters: #5390

Additional Notes for Developers
The approach taken in this PR is patterned after the change implemented in #4715, which was used to modify the default container runtime.

Rejected Alternatives
We explored several alternative solutions, but they were not feasible:

• In-Place Upgrade Handling: We found that direct handling during the 4.21 upgrade was unreliable. After multiple upgrade cycles, there was no consistent mechanism to identify clusters originally provisioned before 4.21.

• Changing the Default File: Switching the default configuration file (e.g., away from /etc/node-sizing-enabled) was overly complex, requiring us to manually manage legacy configuration paths for existing clusters.

• Installer-Created KubeletConfig: Since OpenShift clusters do not contain a default KubeletConfig resource, one option was to have the installer create it. This was rejected because Hypershift deployments may bypass the standard OCP installer.

• Adding a Default KubeletConfig Resource: This approach was dismissed because OpenShift allows only a single KubeletConfig per cluster. Introducing a default resource risks a user's explicit KubeletConfig unintentionally overriding the system default, leading to confusion.

- How to verify it

- Description for the changelog

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the openshift-eng/jira-lifecycle-plugin repository.

[ngopalak-redhat]

/test unit

[openshift-ci-robot]

@ngopalak-redhat: This pull request references OCPNODE-3718 which is a valid jira issue.

Warning: The referenced jira issue has an invalid target version for the target branch this PR targets: expected the epic to target the "4.20.z" version, but no target version was set.

In response to this:

- What I did
This patch introduces the 01-master-auto-sizing-disabled MachineConfig to OpenShift 4.20 clusters, setting the NODE_SIZING_ENABLED flag to false by default on master and worker nodes.
This change is required as we are making auto sizing enabled by default for cluster created using 4.21 and above.

Summary of changes

• Enforce Default Autosizing: Ensures that clusters created in 4.20 will retain the pre-4.21 behavior of having auto node sizing disabled by default.

• Upgrade Pre-requisite: This patch is a mandatory requirement for upgrading 4.20 clusters to 4.21. Changes to Cincinnati (TODO) will enforce that this patch must be present before the upgrade path to 4.21 is started.

• User Override (Priority): The MachineConfig uses the prefix 01- to ensure it sets the initial default. If a user has already created a KubeletConfig to explicitly enable autoSizing (as per the KubeletConfig documentation), that explicit user configuration will take precedence (override this default) and will be retained when upgrading to 4.21.

Reference: This change addresses the shift in default behavior introduced in OpenShift 4.21, where NODE_SIZING_ENABLED is set to true for all new clusters: #5390

Additional Notes for Developers
The approach taken in this PR is patterned after the change implemented in #4715, which was used to modify the default container runtime.

Rejected Alternatives
We explored several alternative solutions, but they were not feasible:

• In-Place Upgrade Handling: We found that direct handling during the 4.21 upgrade was unreliable. After multiple upgrade cycles, there was no consistent mechanism to identify clusters originally provisioned before 4.21.

• Changing the Default File: Switching the default configuration file (e.g., away from /etc/node-sizing-enabled) was overly complex, requiring us to manually manage legacy configuration paths for existing clusters.

• Installer-Created KubeletConfig: Since OpenShift clusters do not contain a default KubeletConfig resource, one option was to have the installer create it. This was rejected because Hypershift deployments may bypass the standard OCP installer.

• Adding a Default KubeletConfig Resource: This approach was dismissed because OpenShift allows only a single KubeletConfig per cluster. Introducing a default resource risks a user's explicit KubeletConfig unintentionally overriding the system default, leading to confusion.

- How to verify it

• Verified the patch on a 4.20 cluster: Created a cluster using ClusterBot, applied the patch via oc adm upgrade, confirmed the new MachineConfig was created, and ensured auto node sizing was disabled.

• Direct Patch Verification: Created a cluster using ClusterBot with the patch applied and confirmed auto node sizing was disabled.

• User Override Test: Created a KubeletConfig to explicitly enable auto sizing and verified that the setting was correctly enabled (overriding the default).

• Upgrade Path Validation: Successfully upgraded the patched cluster to 4.21 (using the above referenced 4.21 PR changes). Confirmed that auto node sizing remained disabled for upgraded clusters that had not been explicitly configured otherwise.

- Description for the changelog

Introduces the auto sizing MachineConfig, ensuring the feature remains disabled by default during upgrade

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the openshift-eng/jira-lifecycle-plugin repository.

[openshift-ci-robot]

@ngopalak-redhat: This pull request references OCPNODE-3722 which is a valid jira issue.

In response to this:

- What I did
This patch introduces the 01-master-auto-sizing-disabled MachineConfig to OpenShift 4.20 clusters, setting the NODE_SIZING_ENABLED flag to false by default on master and worker nodes.
This change is required as we are making auto sizing enabled by default for cluster created using 4.21 and above.

Summary of changes

• Enforce Default Autosizing: Ensures that clusters created in 4.20 will retain the pre-4.21 behavior of having auto node sizing disabled by default.

• Upgrade Pre-requisite: This patch is a mandatory requirement for upgrading 4.20 clusters to 4.21. Changes to Cincinnati (TODO) will enforce that this patch must be present before the upgrade path to 4.21 is started.

• User Override (Priority): The MachineConfig uses the prefix 01- to ensure it sets the initial default. If a user has already created a KubeletConfig to explicitly enable autoSizing (as per the KubeletConfig documentation), that explicit user configuration will take precedence (override this default) and will be retained when upgrading to 4.21.

Reference: This change addresses the shift in default behavior introduced in OpenShift 4.21, where NODE_SIZING_ENABLED is set to true for all new clusters: #5390

Additional Notes for Developers
The approach taken in this PR is patterned after the change implemented in #4715, which was used to modify the default container runtime.

Rejected Alternatives
We explored several alternative solutions, but they were not feasible:

• In-Place Upgrade Handling: We found that direct handling during the 4.21 upgrade was unreliable. After multiple upgrade cycles, there was no consistent mechanism to identify clusters originally provisioned before 4.21.

• Changing the Default File: Switching the default configuration file (e.g., away from /etc/node-sizing-enabled) was overly complex, requiring us to manually manage legacy configuration paths for existing clusters.

• Installer-Created KubeletConfig: Since OpenShift clusters do not contain a default KubeletConfig resource, one option was to have the installer create it. This was rejected because Hypershift deployments may bypass the standard OCP installer.

• Adding a Default KubeletConfig Resource: This approach was dismissed because OpenShift allows only a single KubeletConfig per cluster. Introducing a default resource risks a user's explicit KubeletConfig unintentionally overriding the system default, leading to confusion.

- How to verify it

• Verified the patch on a 4.20 cluster: Created a cluster using ClusterBot, applied the patch via oc adm upgrade, confirmed the new MachineConfig was created, and ensured auto node sizing was disabled.

• Direct Patch Verification: Created a cluster using ClusterBot with the patch applied and confirmed auto node sizing was disabled.

• User Override Test: Created a KubeletConfig to explicitly enable auto sizing and verified that the setting was correctly enabled (overriding the default).

• Upgrade Path Validation: Successfully upgraded the patched cluster to 4.21 (using the above referenced 4.21 PR changes). Confirmed that auto node sizing remained disabled for upgraded clusters that had not been explicitly configured otherwise.

- Description for the changelog

Introduces the auto sizing MachineConfig, ensuring the feature remains disabled by default during upgrade

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the openshift-eng/jira-lifecycle-plugin repository.

[openshift-ci-robot]

@ngopalak-redhat: This pull request references OCPNODE-3722 which is a valid jira issue.

In response to this:

Fixes: #OCPNODE-3722

- What I did
This patch introduces the 01-master-auto-sizing-disabled MachineConfig to OpenShift 4.20 clusters, setting the NODE_SIZING_ENABLED flag to false by default on master and worker nodes.
This change is required as we are making auto sizing enabled by default for cluster created using 4.21 and above.

Summary of changes

• Enforce Default Autosizing: Ensures that clusters created in 4.20 will retain the pre-4.21 behavior of having auto node sizing disabled by default.

• Upgrade Pre-requisite: This patch is a mandatory requirement for upgrading 4.20 clusters to 4.21. Changes to Cincinnati (TODO) will enforce that this patch must be present before the upgrade path to 4.21 is started.

• User Override (Priority): The MachineConfig uses the prefix 01- to ensure it sets the initial default. If a user has already created a KubeletConfig to explicitly enable autoSizing (as per the KubeletConfig documentation), that explicit user configuration will take precedence (override this default) and will be retained when upgrading to 4.21.

Reference: This change addresses the shift in default behavior introduced in OpenShift 4.21, where NODE_SIZING_ENABLED is set to true for all new clusters: #5390

Additional Notes for Developers
The approach taken in this PR is patterned after the change implemented in #4715, which was used to modify the default container runtime.

Rejected Alternatives
We explored several alternative solutions, but they were not feasible:

• In-Place Upgrade Handling: We found that direct handling during the 4.21 upgrade was unreliable. After multiple upgrade cycles, there was no consistent mechanism to identify clusters originally provisioned before 4.21.

• Changing the Default File: Switching the default configuration file (e.g., away from /etc/node-sizing-enabled) was overly complex, requiring us to manually manage legacy configuration paths for existing clusters.

• Installer-Created KubeletConfig: Since OpenShift clusters do not contain a default KubeletConfig resource, one option was to have the installer create it. This was rejected because Hypershift deployments may bypass the standard OCP installer.

• Adding a Default KubeletConfig Resource: This approach was dismissed because OpenShift allows only a single KubeletConfig per cluster. Introducing a default resource risks a user's explicit KubeletConfig unintentionally overriding the system default, leading to confusion.

- How to verify it

• Verified the patch on a 4.20 cluster: Created a cluster using ClusterBot, applied the patch via oc adm upgrade, confirmed the new MachineConfig was created, and ensured auto node sizing was disabled.

• Direct Patch Verification: Created a cluster using ClusterBot with the patch applied and confirmed auto node sizing was disabled.

• User Override Test: Created a KubeletConfig to explicitly enable auto sizing and verified that the setting was correctly enabled (overriding the default).

• Upgrade Path Validation: Successfully upgraded the patched cluster to 4.21 (using the above referenced 4.21 PR changes). Confirmed that auto node sizing remained disabled for upgraded clusters that had not been explicitly configured otherwise.

- Description for the changelog

Introduces the auto sizing MachineConfig, ensuring the feature remains disabled by default during upgrade

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the openshift-eng/jira-lifecycle-plugin repository.

[ngopalak-redhat]

/test bootstrap-unit

[ngopalak-redhat]

/test all

[openshift-ci-robot]

@ngopalak-redhat: This pull request references OCPNODE-3722 which is a valid jira issue.

In response to this:

Fixes: #OCPNODE-3722

- What I did
This patch introduces the 01-master-auto-sizing-disabled MachineConfig to OpenShift 4.20 clusters, setting the NODE_SIZING_ENABLED flag to false by default on master and worker nodes.
This change is required as we are making auto sizing enabled by default for cluster created using 4.21 and above.

Summary of changes

• Enforce Default Autosizing: Ensures that clusters created in 4.20 will retain the pre-4.21 behavior of having auto node sizing disabled by default.

• Upgrade Pre-requisite: This patch is a mandatory requirement for upgrading 4.20 clusters to 4.21. Changes to Cincinnati (OCPNODE-3722: Set minimum version of 4.20 required to upgrade to 4.21 cincinnati-graph-data#8277) will enforce that this patch must be present before the upgrade path to 4.21 is started.

• User Override (Priority): The MachineConfig uses the prefix 01- to ensure it sets the initial default. If a user has already created a KubeletConfig to explicitly enable autoSizing (as per the KubeletConfig documentation), that explicit user configuration will take precedence (override this default) and will be retained when upgrading to 4.21.

Reference: This change addresses the shift in default behavior introduced in OpenShift 4.21, where NODE_SIZING_ENABLED is set to true for all new clusters: #5390

Additional Notes for Developers
The approach taken in this PR is patterned after the change implemented in #4715, which was used to modify the default container runtime.

Rejected Alternatives
We explored several alternative solutions, but they were not feasible:

• In-Place Upgrade Handling: We found that direct handling during the 4.21 upgrade was unreliable. After multiple upgrade cycles, there was no consistent mechanism to identify clusters originally provisioned before 4.21.

• Changing the Default File: Switching the default configuration file (e.g., away from /etc/node-sizing-enabled) was overly complex, requiring us to manually manage legacy configuration paths for existing clusters.

• Installer-Created KubeletConfig: Since OpenShift clusters do not contain a default KubeletConfig resource, one option was to have the installer create it. This was rejected because Hypershift deployments may bypass the standard OCP installer.

• Adding a Default KubeletConfig Resource: This approach was dismissed because OpenShift allows only a single KubeletConfig per cluster. Introducing a default resource risks a user's explicit KubeletConfig unintentionally overriding the system default, leading to confusion.

- How to verify it

• Verified the patch on a 4.20 cluster: Created a cluster using ClusterBot, applied the patch via oc adm upgrade, confirmed the new MachineConfig was created, and ensured auto node sizing was disabled.

• Direct Patch Verification: Created a cluster using ClusterBot with the patch applied and confirmed auto node sizing was disabled.

• User Override Test: Created a KubeletConfig to explicitly enable auto sizing and verified that the setting was correctly enabled (overriding the default).

• Upgrade Path Validation: Successfully upgraded the patched cluster to 4.21 (using the above referenced 4.21 PR changes). Confirmed that auto node sizing remained disabled for upgraded clusters that had not been explicitly configured otherwise.

- Description for the changelog

Introduces the auto sizing MachineConfig, ensuring the feature remains disabled by default during upgrade

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the openshift-eng/jira-lifecycle-plugin repository.

[ngopalak-redhat]

@haircommander @sairameshv Can you please review?

[ngopalak-redhat]

/hold Until rebasing and testing the rebase

[ngopalak-redhat]

/unhold All tests are passing with rebase

[haircommander]

/retest

LGTM, @sairameshv @QiWang19 PTAL

[sairameshv]

Left a few nits to be addressed.
The changes LGTM

[sairameshv]

nit: ignConfig gets replaced in the further steps. So, why can't we directly create the rawAutoSizingIgn and pass the same while creating a machine config?

Suggested change

	autoSizingMC, err := ctrlcommon.MachineConfigFromIgnConfig(pool.Name, autoSizingDisabledMCName, ignConfig)
	autoSizingMC, err := ctrlcommon.MachineConfigFromIgnConfig(pool.Name, autoSizingDisabledMCName, rawAutoSizingIgn)

[sairameshv]

Does this sound more appropriate?

Suggested change

	// createAutoSizingIgnConfig creates the Ignition config with environment variables
	// createDefaultAutoSizingIgnConfig creates the Ignition config with the default auto sizing environment variable contents

[ngopalak-redhat]

@sairameshv Can you review again? Only change is the removal of the unit test

[sairameshv]

LGTM apart from minor nits

/lgtm

[ngopalak-redhat]

/retest-required

[openshift-ci]

@ngopalak-redhat: The following tests failed, say /retest to rerun all failed tests or /retest-required to rerun all mandatory failed tests:

Test name	Commit	Details	Required	Rerun command
ci/prow/e2e-aws-ovn-upgrade	8f0b0a8	link	true	/test e2e-aws-ovn-upgrade
ci/prow/e2e-gcp-op-2of2	8f0b0a8	link	true	/test e2e-gcp-op-2of2
ci/prow/e2e-gcp-op-1of2	8f0b0a8	link	true	/test e2e-gcp-op-1of2
ci/prow/e2e-aws-ovn	8f0b0a8	link	true	/test e2e-aws-ovn
ci/prow/e2e-gcp-op-single-node	8f0b0a8	link	true	/test e2e-gcp-op-single-node
ci/prow/bootstrap-unit	8f0b0a8	link	false	/test bootstrap-unit

Full PR test history. Your PR dashboard.

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository. I understand the commands that are listed here.

[umohnani8]

LGTM

/approve
/label backport-risk-assessed

Looks like it is unhappy with the jira card though

[openshift-ci]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: ngopalak-redhat, sairameshv, umohnani8

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

• OWNERS [umohnani8]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment