- A new feature was enabled to improve the performance of the iptables mode of `kube-proxy` in large clusters. No action was required, however:
  

- Kubeadm: amend the node kernel validation to treat CGROUP_PIDS, FAIR_GROUP_SCHED as required and CFS_BANDWIDTH, CGROUP_HUGETLB as optional ([#96378](https://github.com/kubernetes/kubernetes/pull/96378), [@neolit123](https://github.com/neolit123)) [SIG Cluster Lifecycle and Node]

	return pp, nil
}

// PushPromiseParam are the parameters for writing a PUSH_PROMISE frame.
type http2PushPromiseParam struct {
	// StreamID is the required Stream ID to initiate.
	StreamID uint32

	// PromiseID is the required Stream ID which this
	// Push Promises
	PromiseID uint32

	// BlockFragment is part (or all) of a Header Block.
	BlockFragment []byte

	// Note the pod will be crashlooping, so this may take a few minutes to become fully functional based on when the retry occurs.
	// This requires no RBAC privilege, but will require the CNI agent to run as a privileged pod.
	RepairPods bool `protobuf:"varint,11,opt,name=repairPods,proto3" json:"repairPods,omitempty"`
	// No longer used.
	//

- Fixed two regressions introduced by the PodDisruptionConditions feature (on by default in 1.26):
  - pod eviction API calls returned spurious precondition errors and required a second evict API call to succeed

    /**
     * Get the value for the key 'api.access.token.required'. <br>
     * The value is, e.g. false <br>
     * @return The value of found property. (NotNull: if not found, exception but basically no way)
     */
    String getApiAccessTokenRequired();

    /**
     * Is the property for the key 'api.access.token.required' true? <br>
     * The value is, e.g. false <br>

}

func autoConvert_v1_VolumeNodeAffinity_To_core_VolumeNodeAffinity(in *v1.VolumeNodeAffinity, out *core.VolumeNodeAffinity, s conversion.Scope) error {
	out.Required = (*core.NodeSelector)(unsafe.Pointer(in.Required))
	return nil
}

// Convert_v1_VolumeNodeAffinity_To_core_VolumeNodeAffinity is an autogenerated conversion function.

 - Scheduler's CycleState now embeds internal read/write locking inside its `Read()` and `Write()` functions. Meanwhile, `Lock()` and `Unlock()` function are removed. Scheduler plugin developers are now required to remove `CycleState#Lock()` and `CycleState#Unlock()`. Just simply use `Read()` and `Write()` as they're natively thread-safe now. ([#101542](https://github.com/kubernetes/kubernetes/pull/101542), [@Huang-Wei](https://github.com/Huang-Wei))

## Changelog since v1.27.0

## Urgent Upgrade Notes 

### (No, really, you MUST read this before you upgrade)

- Action required for the custom scheduler plugin developers. 
  Here's the breaking change in `EnqueueExtension` in the scheduling framework. 

li=Object.defineProperty;var bb=Object.getOwnPropertyDescriptor;var vb=Object.getOwnPropertyNames;var xb=Object.getPrototypeOf,kb=Object.prototype.hasOwnProperty;var au=i=>li(i,"__esModule",{value:!0});var ou=i=>{if(typeof require!="undefined")return require(i);throw new Error('Dynamic require of "'+i+'" is not supported')};var C=(i,e)=>()=>(i&&(e=i(i=0)),e);var v=(i,e)=>()=>(e||i((e={exports:{}}).exports,e),e.exports),Ae=(i,e)=>{au(i);for(var t in e)li(i,t,{get:e[t],enumerable:!0})},Sb=(i,e,t)=>{if(e&&typeof...