OCPNODE-3719: Default Enablement of Auto Sizing Reserved in OpenShift 4.21 #5390

ngopalak-redhat · 2025-11-04T02:49:07Z

- What I did
This PR updates the default cluster configuration to enable Auto Sizing Reserved for all new OpenShift clusters created from version 4.21 onward.

New Cluster Default: New 4.21 clusters will now have Auto Sizing Reserved enabled by default

Upgrade Preservation:

Clusters originally created before 4.21 will not be impacted by this default change. The prior behavior (disabled) is maintained unless one of the following occurs:

The MachineConfig introduced by the corresponding 4.20 patch is removed.
The user creates a custom KubeletConfig to explicitly override the default.

If a user defines explicit memory or cpu reservations within a custom KubeletConfig, Auto Sizing Reserved will automatically remain disabled to prevent conflicting configuration logic.

Related 4.20 Patch: For context on how older clusters are transitioned, please refer to the pre-requisite patch:

- How to verify it
The following scenarios were successfully validated:Upgrade Test (4.20 --> 4.21 + PR): A cluster created in 4.20 was upgraded to this 4.21 PR. Auto Sizing Reserved was confirmed to be disabled, preserving the pre-4.21 behavior.

New Cluster Test (4.21 + PR): A new cluster was provisioned in 4.21 with this PR. Auto Sizing Reserved was confirmed to be enabled by default.

Manual Test result

Here's the state of mc after upgrade from [4.20 + patch] to [4.21 + this pr]:

oc get mc                                                                                                       ✔  admin 󱃾  23:17:58
NAME                                               GENERATEDBYCONTROLLER                      IGNITIONVERSION   AGE
00-master                                          cdfd00e2a65bb4c8641900c207c74a3a7fb3bb39   3.5.0             124m
00-worker                                          cdfd00e2a65bb4c8641900c207c74a3a7fb3bb39   3.5.0             124m
01-master-auto-sizing-disabled                                                                3.5.0             124m
01-master-container-runtime                        cdfd00e2a65bb4c8641900c207c74a3a7fb3bb39   3.5.0             124m
01-master-kubelet                                  cdfd00e2a65bb4c8641900c207c74a3a7fb3bb39   3.5.0             124m
01-worker-auto-sizing-disabled                                                                3.5.0             124m
01-worker-container-runtime                        cdfd00e2a65bb4c8641900c207c74a3a7fb3bb39   3.5.0             124m
01-worker-kubelet                                  cdfd00e2a65bb4c8641900c207c74a3a7fb3bb39   3.5.0             124m
97-master-generated-kubelet                        cdfd00e2a65bb4c8641900c207c74a3a7fb3bb39   3.5.0             14m
97-worker-generated-kubelet                        cdfd00e2a65bb4c8641900c207c74a3a7fb3bb39   3.5.0             14m
98-master-generated-kubelet                        cdfd00e2a65bb4c8641900c207c74a3a7fb3bb39   3.5.0             124m
98-worker-generated-kubelet                        cdfd00e2a65bb4c8641900c207c74a3a7fb3bb39   3.5.0             124m
99-master-generated-registries                     cdfd00e2a65bb4c8641900c207c74a3a7fb3bb39   3.5.0             124m
99-master-ssh                                                                                 3.2.0             130m
99-worker-generated-registries                     cdfd00e2a65bb4c8641900c207c74a3a7fb3bb39   3.5.0             124m
99-worker-ssh                                                                                 3.2.0             130m
rendered-master-7978e7fd1b3d8deb9618638ad045a112   cdfd00e2a65bb4c8641900c207c74a3a7fb3bb39   3.5.0             14m
rendered-master-ac5bf6659679f8fb20815992cece3127   ece93ae657e115244dba500093fc78eef0d63241   3.5.0             124m
rendered-worker-b7224764ec4bac392d79170b0ec4d3bc   ece93ae657e115244dba500093fc78eef0d63241   3.5.0             124m
rendered-worker-c6659bc45fd053b3995322d8c0490e0d   cdfd00e2a65bb4c8641900c207c74a3a7fb3bb39   3.5.0             14m

We can see "auto-sizing-disabled" mc

Here's the post upgrade NODE_SIZING_ENABLED:

oc debug node/ip-10-0-6-61.us-east-2.compute.internal                                                           ✔  admin 󱃾  23:24:58
Starting pod/ip-10-0-6-61us-east-2computeinternal-debug-zc2rs ...
To use host binaries, run `chroot /host`
Pod IP: 10.0.6.61
If you don't see a command prompt, try pressing enter.
sh-5.1# chroot /host
sh-5.1# cat /etc/node-sizing-enabled.env
NODE_SIZING_ENABLED=false
SYSTEM_RESERVED_MEMORY=1Gi
SYSTEM_RESERVED_CPU=500m
SYSTEM_RESERVED_ES=1Gi
sh-5.1#

Here are the nodes post upgrade, showing successful upgrade:

 oc get no                                                                                              ✔  1m 35s  admin 󱃾  23:26:53
NAME                                         STATUS   ROLES                  AGE    VERSION
ip-10-0-108-148.us-east-2.compute.internal   Ready    worker                 123m   v1.34.1
ip-10-0-113-32.us-east-2.compute.internal    Ready    control-plane,master   137m   v1.34.1
ip-10-0-49-43.us-east-2.compute.internal     Ready    control-plane,master   137m   v1.34.1
ip-10-0-6-61.us-east-2.compute.internal      Ready    worker                 123m   v1.34.1
ip-10-0-80-128.us-east-2.compute.internal    Ready    control-plane,master   137m   v1.34.1
ip-10-0-99-28.us-east-2.compute.internal     Ready    worker                 123m   v1.34.1

Now I can apply kubeletconfig to enable auto sizing reserved:

oc apply -f 80-worker-auto.yaml                                                                                 ✔  admin 󱃾  23:26:57
kubeletconfig.machineconfiguration.openshift.io/80-worker-auto created

After applying the NODE_SIZING_ENABLED changes:

oc debug node/ip-10-0-6-61.us-east-2.compute.internal                                                           ✔  admin 󱃾  23:32:43
Starting pod/ip-10-0-6-61us-east-2computeinternal-debug-f8jnj ...
To use host binaries, run `chroot /host`
Pod IP: 10.0.6.61
If you don't see a command prompt, try pressing enter.
sh-5.1# chroot /host
sh-5.1# cat /etc/node-sizing-enabled.env
NODE_SIZING_ENABLED=true
SYSTEM_RESERVED_MEMORY=1Gi
SYSTEM_RESERVED_CPU=500m
SYSTEM_RESERVED_ES=1Gi
sh-5.1#

Also we can see the KubeletConfig showing correct auto sizing reserved:

oc get kubeletconfig -o yaml                                                                            ✔  1m 6s  admin 󱃾  23:34:02
apiVersion: v1
items:
- apiVersion: machineconfiguration.openshift.io/v1
  kind: KubeletConfig
  metadata:
    annotations:
      machineconfiguration.openshift.io/mc-name-suffix: ""
    creationTimestamp: "2025-11-09T17:57:52Z"
    finalizers:
    - 99-worker-generated-kubelet
    generation: 1
    name: 80-worker-auto
    resourceVersion: "83327"
    uid: 0868236a-47c3-4d5e-be99-9d5511cda70e
  spec:
    autoSizingReserved: true
    machineConfigPoolSelector:
      matchLabels:
        pools.operator.machineconfiguration.openshift.io/worker: ""
  status:
    conditions:
    - lastTransitionTime: "2025-11-09T17:57:52Z"
      message: Success
      status: "True"
      type: Success
    observedGeneration: 1
kind: List
metadata:
  resourceVersion: ""

Updates:

The hypershift environments have been excluded from the PR due to high resource usage during the hypershift tests.

- Description for the changelog

Enable auto sizing reserved by default on new clusters

openshift-ci · 2025-11-04T02:49:11Z

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

openshift-ci-robot · 2025-11-07T10:38:21Z

@ngopalak-redhat: This pull request references OCPNODE-3719 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 story to target the "4.21.0" version, but no target version was set.

Details

In response to this:

… true is created

- What I did
This PR updates the default cluster configuration to enable Auto Sizing Reserved for all new OpenShift clusters created from version 4.21 onward.

New Cluster Default: New 4.21 clusters will now have Auto Sizing Reserved enabled by default

Upgrade Preservation:

Clusters originally created before 4.21 will not be impacted by this default change. The prior behavior (disabled) is maintained unless one of the following occurs:

The MachineConfig introduced by the corresponding 4.20 patch is removed.

The user creates a custom KubeletConfig to explicitly override the default.

If a user defines explicit memory or cpu reservations within a custom KubeletConfig, Auto Sizing Reserved will automatically remain disabled to prevent conflicting configuration logic.

Related 4.20 Patch: For context on how older clusters are transitioned, please refer to the pre-requisite patch:

- How to verify it
The following scenarios were successfully validated:Upgrade Test (4.20 --> 4.21 + PR): A cluster created in 4.20 was upgraded to this 4.21 PR. Auto Sizing Reserved was confirmed to be disabled, preserving the pre-4.21 behavior.

New Cluster Test (4.21 + PR): A new cluster was provisioned in 4.21 with this PR. Auto Sizing Reserved was confirmed to be enabled by default.
- Description for the changelog

Enable auto sizing reserved by default on new clusters

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 · 2025-11-07T10:38:37Z

@ngopalak-redhat: This pull request references OCPNODE-3719 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 story to target the "4.21.0" version, but no target version was set.

Details

In response to this:

Fixes #OCPNODE-3719

- What I did
This PR updates the default cluster configuration to enable Auto Sizing Reserved for all new OpenShift clusters created from version 4.21 onward.

New Cluster Default: New 4.21 clusters will now have Auto Sizing Reserved enabled by default

Upgrade Preservation:

Clusters originally created before 4.21 will not be impacted by this default change. The prior behavior (disabled) is maintained unless one of the following occurs:

The MachineConfig introduced by the corresponding 4.20 patch is removed.

The user creates a custom KubeletConfig to explicitly override the default.

If a user defines explicit memory or cpu reservations within a custom KubeletConfig, Auto Sizing Reserved will automatically remain disabled to prevent conflicting configuration logic.

Related 4.20 Patch: For context on how older clusters are transitioned, please refer to the pre-requisite patch:

- How to verify it
The following scenarios were successfully validated:Upgrade Test (4.20 --> 4.21 + PR): A cluster created in 4.20 was upgraded to this 4.21 PR. Auto Sizing Reserved was confirmed to be disabled, preserving the pre-4.21 behavior.

New Cluster Test (4.21 + PR): A new cluster was provisioned in 4.21 with this PR. Auto Sizing Reserved was confirmed to be enabled by default.
- Description for the changelog

Enable auto sizing reserved by default on new clusters

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 · 2025-11-07T10:39:29Z

/test all

ngopalak-redhat · 2025-11-07T10:40:56Z

/hold Wait for openshift/cincinnati-graph-data#8277 to merge

ngopalak-redhat · 2025-11-07T12:51:42Z

@haircommander @sairameshv Please review. I have kept it on hold so that the 4.20 and the cincinnati change can merge

pkg/controller/kubelet-config/helpers.go

openshift-ci-robot · 2025-11-09T18:07:12Z

@ngopalak-redhat: This pull request references OCPNODE-3719 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 story to target the "4.21.0" version, but no target version was set.

Details

In response to this:

Fixes #OCPNODE-3719

- What I did
This PR updates the default cluster configuration to enable Auto Sizing Reserved for all new OpenShift clusters created from version 4.21 onward.

New Cluster Default: New 4.21 clusters will now have Auto Sizing Reserved enabled by default

Upgrade Preservation:

Clusters originally created before 4.21 will not be impacted by this default change. The prior behavior (disabled) is maintained unless one of the following occurs:

The MachineConfig introduced by the corresponding 4.20 patch is removed.
The user creates a custom KubeletConfig to explicitly override the default.

If a user defines explicit memory or cpu reservations within a custom KubeletConfig, Auto Sizing Reserved will automatically remain disabled to prevent conflicting configuration logic.

Related 4.20 Patch: For context on how older clusters are transitioned, please refer to the pre-requisite patch:

- How to verify it
The following scenarios were successfully validated:Upgrade Test (4.20 --> 4.21 + PR): A cluster created in 4.20 was upgraded to this 4.21 PR. Auto Sizing Reserved was confirmed to be disabled, preserving the pre-4.21 behavior.

New Cluster Test (4.21 + PR): A new cluster was provisioned in 4.21 with this PR. Auto Sizing Reserved was confirmed to be enabled by default.

Manual Test result

Here's the state of mc after upgrade from [4.20 + patch] to [4.21 + this pr]:

oc get mc                                                                                                       ✔  admin 󱃾  23:17:58
NAME                                               GENERATEDBYCONTROLLER                      IGNITIONVERSION   AGE
00-master                                          cdfd00e2a65bb4c8641900c207c74a3a7fb3bb39   3.5.0             124m
00-worker                                          cdfd00e2a65bb4c8641900c207c74a3a7fb3bb39   3.5.0             124m
01-master-auto-sizing-disabled                                                                3.5.0             124m
01-master-container-runtime                        cdfd00e2a65bb4c8641900c207c74a3a7fb3bb39   3.5.0             124m
01-master-kubelet                                  cdfd00e2a65bb4c8641900c207c74a3a7fb3bb39   3.5.0             124m
01-worker-auto-sizing-disabled                                                                3.5.0             124m
01-worker-container-runtime                        cdfd00e2a65bb4c8641900c207c74a3a7fb3bb39   3.5.0             124m
01-worker-kubelet                                  cdfd00e2a65bb4c8641900c207c74a3a7fb3bb39   3.5.0             124m
97-master-generated-kubelet                        cdfd00e2a65bb4c8641900c207c74a3a7fb3bb39   3.5.0             14m
97-worker-generated-kubelet                        cdfd00e2a65bb4c8641900c207c74a3a7fb3bb39   3.5.0             14m
98-master-generated-kubelet                        cdfd00e2a65bb4c8641900c207c74a3a7fb3bb39   3.5.0             124m
98-worker-generated-kubelet                        cdfd00e2a65bb4c8641900c207c74a3a7fb3bb39   3.5.0             124m
99-master-generated-registries                     cdfd00e2a65bb4c8641900c207c74a3a7fb3bb39   3.5.0             124m
99-master-ssh                                                                                 3.2.0             130m
99-worker-generated-registries                     cdfd00e2a65bb4c8641900c207c74a3a7fb3bb39   3.5.0             124m
99-worker-ssh                                                                                 3.2.0             130m
rendered-master-7978e7fd1b3d8deb9618638ad045a112   cdfd00e2a65bb4c8641900c207c74a3a7fb3bb39   3.5.0             14m
rendered-master-ac5bf6659679f8fb20815992cece3127   ece93ae657e115244dba500093fc78eef0d63241   3.5.0             124m
rendered-worker-b7224764ec4bac392d79170b0ec4d3bc   ece93ae657e115244dba500093fc78eef0d63241   3.5.0             124m
rendered-worker-c6659bc45fd053b3995322d8c0490e0d   cdfd00e2a65bb4c8641900c207c74a3a7fb3bb39   3.5.0             14m

We can see "auto-sizing-disabled" mc

Here's the post upgrade NODE_SIZING_ENABLED:

oc debug node/ip-10-0-6-61.us-east-2.compute.internal                                                           ✔  admin 󱃾  23:24:58
Starting pod/ip-10-0-6-61us-east-2computeinternal-debug-zc2rs ...
To use host binaries, run `chroot /host`
Pod IP: 10.0.6.61
If you don't see a command prompt, try pressing enter.
sh-5.1# chroot /host
sh-5.1# cat /etc/node-sizing-enabled.env
NODE_SIZING_ENABLED=false
SYSTEM_RESERVED_MEMORY=1Gi
SYSTEM_RESERVED_CPU=500m
SYSTEM_RESERVED_ES=1Gi
sh-5.1#

Here are the nodes post upgrade, showing successful upgrade:

oc get no                                                                                              ✔  1m 35s  admin 󱃾  23:26:53
NAME                                         STATUS   ROLES                  AGE    VERSION
ip-10-0-108-148.us-east-2.compute.internal   Ready    worker                 123m   v1.34.1
ip-10-0-113-32.us-east-2.compute.internal    Ready    control-plane,master   137m   v1.34.1
ip-10-0-49-43.us-east-2.compute.internal     Ready    control-plane,master   137m   v1.34.1
ip-10-0-6-61.us-east-2.compute.internal      Ready    worker                 123m   v1.34.1
ip-10-0-80-128.us-east-2.compute.internal    Ready    control-plane,master   137m   v1.34.1
ip-10-0-99-28.us-east-2.compute.internal     Ready    worker                 123m   v1.34.1

Now I can apply kubeletconfig to enable auto sizing reserved:

oc apply -f 80-worker-auto.yaml                                                                                 ✔  admin 󱃾  23:26:57
kubeletconfig.machineconfiguration.openshift.io/80-worker-auto created

After applying the NODE_SIZING_ENABLED changes:

oc debug node/ip-10-0-6-61.us-east-2.compute.internal                                                           ✔  admin 󱃾  23:32:43
Starting pod/ip-10-0-6-61us-east-2computeinternal-debug-f8jnj ...
To use host binaries, run `chroot /host`
Pod IP: 10.0.6.61
If you don't see a command prompt, try pressing enter.
sh-5.1# chroot /host
sh-5.1# cat /etc/node-sizing-enabled.env
NODE_SIZING_ENABLED=true
SYSTEM_RESERVED_MEMORY=1Gi
SYSTEM_RESERVED_CPU=500m
SYSTEM_RESERVED_ES=1Gi
sh-5.1#

Also we can see the KubeletConfig showing correct auto sizing reserved:

oc get kubeletconfig -o yaml                                                                            ✔  1m 6s  admin 󱃾  23:34:02
apiVersion: v1
items:
- apiVersion: machineconfiguration.openshift.io/v1
 kind: KubeletConfig
 metadata:
   annotations:
     machineconfiguration.openshift.io/mc-name-suffix: ""
   creationTimestamp: "2025-11-09T17:57:52Z"
   finalizers:
   - 99-worker-generated-kubelet
   generation: 1
   name: 80-worker-auto
   resourceVersion: "83327"
   uid: 0868236a-47c3-4d5e-be99-9d5511cda70e
 spec:
   autoSizingReserved: true
   machineConfigPoolSelector:
     matchLabels:
       pools.operator.machineconfiguration.openshift.io/worker: ""
 status:
   conditions:
   - lastTransitionTime: "2025-11-09T17:57:52Z"
     message: Success
     status: "True"
     type: Success
   observedGeneration: 1
kind: List
metadata:
 resourceVersion: ""

- Description for the changelog

Enable auto sizing reserved by default on new clusters

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 · 2025-11-10T03:07:54Z

/retest-required

ngopalak-redhat · 2025-11-10T06:54:58Z

/retest-required

ngopalak-redhat · 2025-11-10T12:10:32Z

/test e2e-hypershift

ngopalak-redhat · 2025-11-10T14:21:23Z

/retest-required

pkg/controller/kubelet-config/helpers.go

pkg/controller/kubelet-config/helpers_test.go

pkg/controller/template/template_controller_test.go

ngopalak-redhat · 2025-11-19T03:10:06Z

@sairameshv I made the changes and retested setting system reserved manually. When system reserved is set via a kubelet config the auto-sizing-reserved is set to false as expected.

Test Result

Created the cluster with the PR using ClusterBot and check the NODE_SIZING_ENABLED flag:

oc debug node/ip-10-0-64-18.us-east-2.compute.internal 
Starting pod/ip-10-0-64-18us-east-2computeinternal-debug-4lthc ...
To use host binaries, run `chroot /host`
Pod IP: 10.0.64.18
If you don't see a command prompt, try pressing enter.
sh-5.1#
sh-5.1# chroot /host
sh-5.1# cat /etc/node-sizing-enabled.env
NODE_SIZING_ENABLED=true
SYSTEM_RESERVED_MEMORY=1Gi
SYSTEM_RESERVED_CPU=500m
SYSTEM_RESERVED_ES=1Gish-5.1# exit

Now applied a KubeletConfig with system reserved:

oc apply -f systemReserved.yaml                                                                                 ✔  admin 󱃾  08:23:14
kubeletconfig.machineconfiguration.openshift.io/80-custom-system-reserved created
 ~/redhat/autoreserved  oc get kubeletconfig                                                                                            ✔  admin 󱃾  08:23:33
NAME                        AGE
80-custom-system-reserved   15s
 ~/redhat/autoreserved  oc get kubeletconfig -o yaml                                                                                    ✔  admin 󱃾  08:23:48
apiVersion: v1
items:
- apiVersion: machineconfiguration.openshift.io/v1
  kind: KubeletConfig
  metadata:
    annotations:
      machineconfiguration.openshift.io/mc-name-suffix: ""
    creationTimestamp: "2025-11-19T02:53:33Z"
    finalizers:
    - 99-worker-generated-kubelet
    generation: 1
    name: 80-custom-system-reserved
    resourceVersion: "35491"
    uid: bd02641e-4e8e-4ccb-a98e-7f4e9c919d3a
  spec:
    kubeletConfig:
      apiVersion: kubelet.config.k8s.io/v1beta1
      kind: KubeletConfiguration
      systemReserved:
        cpu: 500m
        memory: 1Gi
    machineConfigPoolSelector:
      matchLabels:
        pools.operator.machineconfiguration.openshift.io/worker: ""
  status:
    conditions:
    - lastTransitionTime: "2025-11-19T02:53:33Z"
      message: Success
      status: "True"
      type: Success
    observedGeneration: 1
kind: List
metadata:
  resourceVersion: ""

Now checked if NODE_SIZING_ENABLED is set to false after the node restarts:

oc debug node/ip-10-0-64-18.us-east-2.compute.internal                                                          ✔  admin 󱃾  08:33:31
Starting pod/ip-10-0-64-18us-east-2computeinternal-debug-fljdx ...
To use host binaries, run `chroot /host`
Pod IP: 10.0.64.18
If you don't see a command prompt, try pressing enter.
sh-5.1# chroot /host
sh-5.1# cat /etc/node-sizing-enabled.env
NODE_SIZING_ENABLED=false
SYSTEM_RESERVED_MEMORY=1Gi
SYSTEM_RESERVED_CPU=500m
SYSTEM_RESERVED_ES=1Gi
sh-5.1#

haircommander · 2025-11-19T14:24:24Z

/skip
/lgtm

haircommander · 2025-11-19T14:24:42Z

@sairameshv @ngopalak-redhat still needs the hold?

openshift-ci-robot · 2025-11-27T07:08:11Z

@ngopalak-redhat: This pull request references OCPNODE-3719 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 story to target the "4.21.0" version, but no target version was set.

Details

In response to this:

Fixes #OCPNODE-3719

- What I did
This PR updates the default cluster configuration to enable Auto Sizing Reserved for all new OpenShift clusters created from version 4.21 onward.

New Cluster Default: New 4.21 clusters will now have Auto Sizing Reserved enabled by default

Upgrade Preservation:

Clusters originally created before 4.21 will not be impacted by this default change. The prior behavior (disabled) is maintained unless one of the following occurs:

The MachineConfig introduced by the corresponding 4.20 patch is removed.
The user creates a custom KubeletConfig to explicitly override the default.

If a user defines explicit memory or cpu reservations within a custom KubeletConfig, Auto Sizing Reserved will automatically remain disabled to prevent conflicting configuration logic.

Related 4.20 Patch: For context on how older clusters are transitioned, please refer to the pre-requisite patch:

- How to verify it
The following scenarios were successfully validated:Upgrade Test (4.20 --> 4.21 + PR): A cluster created in 4.20 was upgraded to this 4.21 PR. Auto Sizing Reserved was confirmed to be disabled, preserving the pre-4.21 behavior.

New Cluster Test (4.21 + PR): A new cluster was provisioned in 4.21 with this PR. Auto Sizing Reserved was confirmed to be enabled by default.

Manual Test result

Here's the state of mc after upgrade from [4.20 + patch] to [4.21 + this pr]:

oc get mc                                                                                                       ✔  admin 󱃾  23:17:58
NAME                                               GENERATEDBYCONTROLLER                      IGNITIONVERSION   AGE
00-master                                          cdfd00e2a65bb4c8641900c207c74a3a7fb3bb39   3.5.0             124m
00-worker                                          cdfd00e2a65bb4c8641900c207c74a3a7fb3bb39   3.5.0             124m
01-master-auto-sizing-disabled                                                                3.5.0             124m
01-master-container-runtime                        cdfd00e2a65bb4c8641900c207c74a3a7fb3bb39   3.5.0             124m
01-master-kubelet                                  cdfd00e2a65bb4c8641900c207c74a3a7fb3bb39   3.5.0             124m
01-worker-auto-sizing-disabled                                                                3.5.0             124m
01-worker-container-runtime                        cdfd00e2a65bb4c8641900c207c74a3a7fb3bb39   3.5.0             124m
01-worker-kubelet                                  cdfd00e2a65bb4c8641900c207c74a3a7fb3bb39   3.5.0             124m
97-master-generated-kubelet                        cdfd00e2a65bb4c8641900c207c74a3a7fb3bb39   3.5.0             14m
97-worker-generated-kubelet                        cdfd00e2a65bb4c8641900c207c74a3a7fb3bb39   3.5.0             14m
98-master-generated-kubelet                        cdfd00e2a65bb4c8641900c207c74a3a7fb3bb39   3.5.0             124m
98-worker-generated-kubelet                        cdfd00e2a65bb4c8641900c207c74a3a7fb3bb39   3.5.0             124m
99-master-generated-registries                     cdfd00e2a65bb4c8641900c207c74a3a7fb3bb39   3.5.0             124m
99-master-ssh                                                                                 3.2.0             130m
99-worker-generated-registries                     cdfd00e2a65bb4c8641900c207c74a3a7fb3bb39   3.5.0             124m
99-worker-ssh                                                                                 3.2.0             130m
rendered-master-7978e7fd1b3d8deb9618638ad045a112   cdfd00e2a65bb4c8641900c207c74a3a7fb3bb39   3.5.0             14m
rendered-master-ac5bf6659679f8fb20815992cece3127   ece93ae657e115244dba500093fc78eef0d63241   3.5.0             124m
rendered-worker-b7224764ec4bac392d79170b0ec4d3bc   ece93ae657e115244dba500093fc78eef0d63241   3.5.0             124m
rendered-worker-c6659bc45fd053b3995322d8c0490e0d   cdfd00e2a65bb4c8641900c207c74a3a7fb3bb39   3.5.0             14m

We can see "auto-sizing-disabled" mc

Here's the post upgrade NODE_SIZING_ENABLED:

oc debug node/ip-10-0-6-61.us-east-2.compute.internal                                                           ✔  admin 󱃾  23:24:58
Starting pod/ip-10-0-6-61us-east-2computeinternal-debug-zc2rs ...
To use host binaries, run `chroot /host`
Pod IP: 10.0.6.61
If you don't see a command prompt, try pressing enter.
sh-5.1# chroot /host
sh-5.1# cat /etc/node-sizing-enabled.env
NODE_SIZING_ENABLED=false
SYSTEM_RESERVED_MEMORY=1Gi
SYSTEM_RESERVED_CPU=500m
SYSTEM_RESERVED_ES=1Gi
sh-5.1#

Here are the nodes post upgrade, showing successful upgrade:

oc get no                                                                                              ✔  1m 35s  admin 󱃾  23:26:53
NAME                                         STATUS   ROLES                  AGE    VERSION
ip-10-0-108-148.us-east-2.compute.internal   Ready    worker                 123m   v1.34.1
ip-10-0-113-32.us-east-2.compute.internal    Ready    control-plane,master   137m   v1.34.1
ip-10-0-49-43.us-east-2.compute.internal     Ready    control-plane,master   137m   v1.34.1
ip-10-0-6-61.us-east-2.compute.internal      Ready    worker                 123m   v1.34.1
ip-10-0-80-128.us-east-2.compute.internal    Ready    control-plane,master   137m   v1.34.1
ip-10-0-99-28.us-east-2.compute.internal     Ready    worker                 123m   v1.34.1

Now I can apply kubeletconfig to enable auto sizing reserved:

oc apply -f 80-worker-auto.yaml                                                                                 ✔  admin 󱃾  23:26:57
kubeletconfig.machineconfiguration.openshift.io/80-worker-auto created

After applying the NODE_SIZING_ENABLED changes:

oc debug node/ip-10-0-6-61.us-east-2.compute.internal                                                           ✔  admin 󱃾  23:32:43
Starting pod/ip-10-0-6-61us-east-2computeinternal-debug-f8jnj ...
To use host binaries, run `chroot /host`
Pod IP: 10.0.6.61
If you don't see a command prompt, try pressing enter.
sh-5.1# chroot /host
sh-5.1# cat /etc/node-sizing-enabled.env
NODE_SIZING_ENABLED=true
SYSTEM_RESERVED_MEMORY=1Gi
SYSTEM_RESERVED_CPU=500m
SYSTEM_RESERVED_ES=1Gi
sh-5.1#

Also we can see the KubeletConfig showing correct auto sizing reserved:

oc get kubeletconfig -o yaml                                                                            ✔  1m 6s  admin 󱃾  23:34:02
apiVersion: v1
items:
- apiVersion: machineconfiguration.openshift.io/v1
 kind: KubeletConfig
 metadata:
   annotations:
     machineconfiguration.openshift.io/mc-name-suffix: ""
   creationTimestamp: "2025-11-09T17:57:52Z"
   finalizers:
   - 99-worker-generated-kubelet
   generation: 1
   name: 80-worker-auto
   resourceVersion: "83327"
   uid: 0868236a-47c3-4d5e-be99-9d5511cda70e
 spec:
   autoSizingReserved: true
   machineConfigPoolSelector:
     matchLabels:
       pools.operator.machineconfiguration.openshift.io/worker: ""
 status:
   conditions:
   - lastTransitionTime: "2025-11-09T17:57:52Z"
     message: Success
     status: "True"
     type: Success
   observedGeneration: 1
kind: List
metadata:
 resourceVersion: ""

Updates:

The hypershift environments have been excluded from the PR due to high resource usage during the hypershift tests.

- Description for the changelog

Enable auto sizing reserved by default on new clusters

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 · 2025-11-27T07:28:47Z

@sairameshv Rebased. @haircommander has reviewed the hyper-shift exclusion, but the required tests for the PR failed on TestKubeletConfigMaxPods. I’ve pushed a fix for that—please review.
Reason for the test fix: The test asserts allocatable memory availability, which proved flaky due to reliance on memory values that can fluctuate.I updated the test logic to toggle true --> false --> true (previously false --> true --> false) to ensure we correctly verify the state change.
I did not to fix the test itself to make it less flaky as it was out of scope here. I can take that up as an enhancement.

ngopalak-redhat · 2025-11-27T07:29:44Z

/retest-required

sairameshv · 2025-11-27T07:50:51Z

templates/common/_base/files/kubelet-auto-node-sizing-enabled.yaml

 contents:
  inline: |
-    NODE_SIZING_ENABLED=false
+    NODE_SIZING_ENABLED={{if eq .Infra.Status.ControlPlaneTopology "External"}}false{{else}}true{{end}}


sairameshv · 2025-11-27T07:51:48Z

/verified later @asahay19

openshift-ci · 2025-11-27T07:51:55Z

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: haircommander, ngopalak-redhat, sairameshv, yuqi-zhang

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

The pull request process is described here

Details

Needs approval from an approver in each of these files:

~~OWNERS~~ [yuqi-zhang]

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

openshift-ci-robot · 2025-11-27T07:51:59Z

@sairameshv: This PR has been marked to be verified later by @asahay19.

Details

In response to this:

/verified later @asahay19

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 · 2025-11-27T08:25:30Z

/retest-required

Remaining retests: 0 against base HEAD 275774a and 2 for PR HEAD 18dbd1a in total

sairameshv · 2025-11-27T09:34:14Z

@ngopalak-redhat , Could you re-visit the description and update it accordingly?

ngopalak-redhat · 2025-11-27T10:41:22Z

/test e2e-hypershift

openshift-ci · 2025-11-27T12:32:52Z

@ngopalak-redhat: The following test 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/bootstrap-unit	`18dbd1a`	link	false	`/test bootstrap-unit`

Full PR test history. Your PR dashboard.

Details

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.

… nodes OCPNODE-3719: Default Enablement of Auto Sizing Reserved in OpenShift 4.21 - Enable AutoSizingReserved by default for worker nodes - Disable AutoSizingReserved for master/control-plane nodes - Disable AutoSizingReserved for arbiter nodes - Disable AutoSizingReserved for Hypershift clusters - Add corresponding tests for the new behavior This combines changes from PR openshift#5390.

ngopalak-redhat · 2025-12-17T11:58:03Z

/payload-job periodic-ci-openshift-machine-config-operator-release-4.21-periodics-e2e-aws-mco-disruptive-techpreview-1of2

openshift-ci · 2025-12-17T11:58:06Z

@ngopalak-redhat: trigger 1 job(s) for the /payload-(with-prs|job|aggregate|job-with-prs|aggregate-with-prs) command

periodic-ci-openshift-machine-config-operator-release-4.21-periodics-e2e-aws-mco-disruptive-techpreview-1of2

See details on https://pr-payload-tests.ci.openshift.org/runs/ci/a517fa40-db3f-11f0-9c58-bc086c06b73a-0

… nodes OCPNODE-3719: Default Enablement of Auto Sizing Reserved in OpenShift 4.21 - Enable AutoSizingReserved by default for worker nodes - Disable AutoSizingReserved for master/control-plane nodes - Disable AutoSizingReserved for arbiter nodes - Disable AutoSizingReserved for Hypershift clusters - Add corresponding tests for the new behavior This combines changes from PR openshift#5390.

openshift-ci bot added the do-not-merge/work-in-progress Indicates that a PR should not merge because it is a work in progress. label Nov 4, 2025

ngopalak-redhat force-pushed the ngopalak/4.21-auto-sizing branch 2 times, most recently from 8b9e810 to 4ff1f91 Compare November 7, 2025 02:28

ngopalak-redhat mentioned this pull request Nov 7, 2025

[release-4.20] OCPBUGS-65777: Enforce OCP 4.20 and earlier cluster to have AutoSizingReserved disabled by default #5387

Merged

ngopalak-redhat changed the title ~~From 4.21 to ensure that KubeletConfig with AutoSizingReserved set to…~~ OCPNODE-3719: Default Enablement of Auto Sizing Reserved in OpenShift 4.21 Nov 7, 2025

openshift-ci-robot added the jira/valid-reference Indicates that this PR references a valid Jira ticket of any type. label Nov 7, 2025

openshift-ci bot added the do-not-merge/hold Indicates that a PR should not merge because someone has issued a /hold command. label Nov 7, 2025

ngopalak-redhat marked this pull request as ready for review November 7, 2025 12:50

openshift-ci bot removed the do-not-merge/work-in-progress Indicates that a PR should not merge because it is a work in progress. label Nov 7, 2025

openshift-ci bot requested review from dkhater-redhat and yuqi-zhang November 7, 2025 12:50

haircommander reviewed Nov 7, 2025

View reviewed changes

pkg/controller/kubelet-config/helpers.go Show resolved Hide resolved

ngopalak-redhat requested a review from haircommander November 10, 2025 06:56

sairameshv reviewed Nov 12, 2025

View reviewed changes

ngopalak-redhat force-pushed the ngopalak/4.21-auto-sizing branch from 2658574 to a3b1c01 Compare November 19, 2025 01:22

ngopalak-redhat requested a review from sairameshv November 19, 2025 03:10

ngopalak-redhat force-pushed the ngopalak/4.21-auto-sizing branch from b818cfd to 74f6c96 Compare November 27, 2025 01:51

openshift-ci-robot removed the verified Signifies that the PR passed pre-merge verification criteria label Nov 27, 2025

openshift-ci bot removed the lgtm Indicates that a PR is ready to be merged. label Nov 27, 2025

openshift-ci-robot removed the verified-later label Nov 27, 2025

Fix AutoSizingReserved test

18dbd1a

ngopalak-redhat force-pushed the ngopalak/4.21-auto-sizing branch from 74f6c96 to 18dbd1a Compare November 27, 2025 03:55

sairameshv reviewed Nov 27, 2025

View reviewed changes

openshift-ci bot added the lgtm Indicates that a PR is ready to be merged. label Nov 27, 2025

openshift-ci-robot added verified-later verified Signifies that the PR passed pre-merge verification criteria labels Nov 27, 2025

openshift-merge-bot bot merged commit 59f330e into openshift:main Nov 27, 2025
13 of 14 checks passed

ngopalak-redhat mentioned this pull request Dec 1, 2025

OCPNODE-3720: Add auto-sizing-reversed test to origin openshift/origin#30465

Merged

ngopalak-redhat mentioned this pull request Dec 10, 2025

OCPNODE-3720: Add a test for NodeSizing default change to OCP 4.21 (With cleanup fix) #5470

Open

neisw mentioned this pull request Dec 11, 2025

OCPBUGS-66420: Revert "Default Enablement of Auto Sizing Reserved in OpenShift 4.21" #5489

Merged

ngopalak-redhat mentioned this pull request Dec 16, 2025

OCPNODE-3973: Default CPU/Memory changes to Workers - AutoSizingReserved #5491

Merged

OCPNODE-3719: Default Enablement of Auto Sizing Reserved in OpenShift 4.21 #5390

OCPNODE-3719: Default Enablement of Auto Sizing Reserved in OpenShift 4.21 #5390

Uh oh!

Conversation

ngopalak-redhat commented Nov 4, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Upgrade Preservation:

Manual Test result

Uh oh!

openshift-ci bot commented Nov 4, 2025

Uh oh!

openshift-ci-robot commented Nov 7, 2025 • edited by openshift-ci bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Upgrade Preservation:

Uh oh!

openshift-ci-robot commented Nov 7, 2025 • edited by openshift-ci bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Upgrade Preservation:

Uh oh!

ngopalak-redhat commented Nov 7, 2025

Uh oh!

ngopalak-redhat commented Nov 7, 2025

Uh oh!

ngopalak-redhat commented Nov 7, 2025

Uh oh!

Uh oh!

openshift-ci-robot commented Nov 9, 2025 • edited by openshift-ci bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Upgrade Preservation:

Manual Test result

Uh oh!

ngopalak-redhat commented Nov 10, 2025

Uh oh!

ngopalak-redhat commented Nov 10, 2025

Uh oh!

ngopalak-redhat commented Nov 10, 2025

Uh oh!

ngopalak-redhat commented Nov 10, 2025

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

ngopalak-redhat commented Nov 19, 2025

Test Result

Uh oh!

haircommander commented Nov 19, 2025

Uh oh!

haircommander commented Nov 19, 2025

Uh oh!

openshift-ci-robot commented Nov 27, 2025 • edited by openshift-ci bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Upgrade Preservation:

Manual Test result

Uh oh!

ngopalak-redhat commented Nov 27, 2025

Uh oh!

ngopalak-redhat commented Nov 27, 2025

Uh oh!

sairameshv Nov 27, 2025

Choose a reason for hiding this comment

Uh oh!

sairameshv commented Nov 27, 2025

Uh oh!

openshift-ci bot commented Nov 27, 2025

Uh oh!

openshift-ci-robot commented Nov 27, 2025

Uh oh!

openshift-ci-robot commented Nov 27, 2025

Uh oh!

sairameshv commented Nov 27, 2025

Uh oh!

ngopalak-redhat commented Nov 27, 2025

Uh oh!

openshift-ci bot commented Nov 27, 2025

Uh oh!

Uh oh!

ngopalak-redhat commented Dec 17, 2025

Uh oh!

openshift-ci bot commented Dec 17, 2025

ngopalak-redhat commented Nov 4, 2025 •

edited

Loading

openshift-ci-robot commented Nov 7, 2025 •

edited by openshift-ci bot

Loading

openshift-ci-robot commented Nov 7, 2025 •

edited by openshift-ci bot

Loading

openshift-ci-robot commented Nov 9, 2025 •

edited by openshift-ci bot

Loading

openshift-ci-robot commented Nov 27, 2025 •

edited by openshift-ci bot

Loading