* Fix watch cache filtering ([#28968](https://github.com/kubernetes/kubernetes/pull/28968), [@liggitt](https://github.com/liggitt))
* Lock all possible kubecfg files at the beginning of ModifyConfig. ([#28232](https://github.com/kubernetes/kubernetes/pull/28232), [@cjcullen](https://github.com/cjcullen))



# v1.3.1

	if autotune {
		newConcurrent := concurrent + (concurrent+1)/2
		autoTunedCapacityNeeded := uint64(newConcurrent * size)
		if capacity < autoTunedCapacityNeeded {
			// Turn-off auto-tuning if next possible concurrency would reach beyond disk capacity.
			return true, false, ""
		}
	}

	return true, autotune, ""
}

// DriveSpeedtestHandler - reports throughput of drives available in the cluster

    assertThat(data2.streamId).isEqualTo(3)
    assertArrayEquals("fghi".toByteArray(), data2.data)
  }

  /**
   * Confirm that we account for discarded data frames. It's possible that data frames are in-flight
   * just prior to us canceling a stream.
   */
  @Test fun discardedDataFramesAreCounted() {
    // Write the mocking script.
    peer.sendFrame().settings(Settings())

### Container Images

All container images are available as manifest lists and support the described
architectures. It is also possible to pull a specific architecture directly by
adding the "-$ARCH" suffix  to the container image name.

name | architectures
---- | -------------

* Default behavior in Cinder storageclass is changed. If availability is not specified, the zone is chosen by algorithm. It makes possible to spread stateful pods across many zones. ([#44798](https://github.com/kubernetes/kubernetes/pull/44798), [@zetaab](https://github.com/zetaab))

        instead of 2.15, or users could receive an installation error due to
        those missing dependencies.
*   TensorFlow 2.15.0.post1 is being released for Linux x86_64 to resolve this
    issue as quickly as possible.
    *   This version removes the `tensorrt` Python package dependencies from the
        tensorflow[and-cuda] installation method to ensure `pip install

		select {
		case <-mTimer.C:
			saveMRFToDisk()
			mTimer.Reset(mrfSaveInterval)
		case <-p.ctx.Done():
			p.mrfStopCh <- struct{}{}
			xioutil.SafeClose(p.mrfSaveCh)
			// We try to save if possible, but we don't care beyond that.
			saveMRFToDisk()
			return
		case e, ok := <-p.mrfSaveCh:
			if !ok {
				return
			}
			entries[e.versionID] = e

			if len(entries) >= mrfMaxEntries {

		return
	}
	// We do not care if service account is readable or not at this point,
	// since this is a delete call we shall allow it to be deleted if possible.
	svcAccount, _, err := globalIAMSys.GetServiceAccount(ctx, serviceAccount)
	if errors.Is(err, errNoSuchServiceAccount) {
		writeErrorResponseJSON(ctx, w, errorCodes.ToAPIErr(ErrAdminServiceAccountNotFound), r.URL)

- For statically provisioned PVs ensure that we use canonical path whenever possible for attach/detach ([#104485](https://github.com/kubernetes/kubernetes/pull/104485), [@gnufied](https://github.com/gnufied)) [SIG Cloud Provider and Storage]

### Other notable changes

* Bump GLBC version to 0.9.3 ([#45055](https://github.com/kubernetes/kubernetes/pull/45055), [@nicksardo](https://github.com/nicksardo))
* kubeadm: Fix invalid assign statement so it is possible to register the master kubelet with other initial Taints ([#45376](https://github.com/kubernetes/kubernetes/pull/45376), [@luxas](https://github.com/luxas))