## Version 4.4.1

_2020-03-08_

 *  Fix: Don't reuse a connection on redirect if certs match but DNS does not. For better
    locality and performance OkHttp attempts to use the same pooled connection across redirects and
    follow-ups. It independently shares connections when the IP addresses and certificates match,
    even if the host names do not. In 4.4.0 we introduced a regression where we shared a connection

		}
	}()
}

// newPeerRestClients creates new peer clients.
// The two slices will point to the same clients,
// but 'all' will contain nil entry for local client.
// The 'all' slice will be in the same order across the cluster.
func newPeerRestClients(endpoints EndpointServerPools) (remote, all []*peerRESTClient) {
	if !globalIsDistErasure {
		// Only useful in distributed setups
		return nil, nil
	}

- DRA: Starting Kubernetes 1.33, only users with access to an admin namespace with the `kubernetes.io/dra-admin-access` label are authorized to create ResourceClaim or ResourceClaimTemplate objects with the `adminAccess` field in this admin namespace if they want to and only they can reference these ResourceClaims or ResourceClaimTemplates...

index*       (string)             Elasticsearch index to store/update events, index is auto-created
format*      (namespace*|access)  'namespace' reflects current bucket/object list and 'access' reflects a journal of object operations, defaults to 'namespace'
queue_dir    (path)               staging dir for undelivered messages e.g. '/home/events'

- For volumes that allow attaches across multiple nodes, attach and detach operations across different nodes are now executed in parallel. ([#88678](https://github.com/kubernetes/kubernetes/pull/88678), [@verult](https://github.com/verult)) [SIG Storage]

  /*
   * The basic strategy is to subdivide the table among Segments, each of which itself is a
   * concurrently readable hash table. The map supports non-blocking reads and concurrent writes
   * across different segments.
   *
   * The page replacement algorithm's data structures are kept casually consistent with the map. The
   * ordering of writes to a segment is sequentially consistent. An update to the map and recording

 * A lock's behavior when a cycle is detected, however, is defined by the {@code Policy} of the
 * factory that created it. This allows detection of cycles across components while delegating
 * control over lock behavior to individual components.
 *
 * <p>Applications are encouraged to use a {@code CycleDetectingLockFactory} to create any locks for

performing a GET then PUT, the kube-apiserver must write the canonical representation of the object to etcd if the current value does not match. That allows external agents to migrate content in etcd from one API version to another, across different storage types, or across varying encryption levels. This fixes a bug introduced in 1.5 where we unintentionally stopped writing the newest data. ([#48394](https://github.com/kubernetes/kubernetes/pull/48394), [@smarterclayton](https://github.com/smarterclayton))...

	// Since server side copy with same source and dest just replaces the ETag, we save
	// encrypted content MD5Sum as ETag for both SSE-C and SSE-KMS, we standardize the ETag
	// encryption across SSE-C and SSE-KMS, and only return last 32 bytes for SSE-C
	if (crypto.SSEC.IsEncrypted(objInfo.UserDefined) || crypto.S3KMS.IsEncrypted(objInfo.UserDefined)) && !copySource {
		return objInfo.ETag[len(objInfo.ETag)-32:]
	}

	Versioning *versioning.Versioning `msg:"-"`

	// Lifecycle performs filtering based on lifecycle.
	// This will filter out objects if the most recent version should be deleted by lifecycle.
	// Is not transferred across request calls.
	Lifecycle *lifecycle.Lifecycle `msg:"-"`

	// Retention configuration, needed to be passed along with lifecycle if set.
	Retention lock.Retention `msg:"-"`

	// Replication configuration