* HPA: ignore scale targets whose replica count is 0 ([#29212](https://github.com/kubernetes/kubernetes/pull/29212), [@sjenning](https://github.com/sjenning))
* Add total inodes to kubelet summary api ([#30231](https://github.com/kubernetes/kubernetes/pull/30231), [@derekwaynecarr](https://github.com/derekwaynecarr))

    checkNotNull(remappingFunction);

    if (range.isEmpty()) {
      return;
    }
    split(range.lowerBound);
    split(range.upperBound);

    // Due to the splitting of any entries spanning the range bounds, we know that any entry with a
    // lower bound in the merge range is entirely contained by the merge range.
    Set<Entry<Cut<K>, RangeMapEntry<K, V>>> entriesInMergeRange =

  private val taskRunner = taskFaker.taskRunner

  /**
   * Note that we don't use the same [TaskFaker] for this factory. That way off-topic tasks like
   * connection pool maintenance tasks don't add noise to route planning tests.
   */
  private val routePlanner = FakeRoutePlanner(taskFaker = taskFaker)
  private val finder = FastFallbackExchangeFinder(routePlanner, taskRunner)

  @AfterEach
  fun tearDown() {
    taskFaker.close()

# v1.3.5

[Documentation](http://kubernetes.github.io) & [Examples](http://releases.k8s.io/release-1.3/examples)

## Downloads

binary | sha256 hash
------ | -----------

  //   have observed 12 micros on 64-bit linux systems to wake up a parked thread). So if the
  //   timeout is small we shouldn't park(). This needs to be traded off with the cpu overhead of
  //   spinning, so we use SPIN_THRESHOLD_NANOS which is what AbstractQueuedSynchronizer uses for
  //   similar purposes.
  // * We want to behave reasonably for timeouts of 0

#### kubelet

* [alpha] Kubelet now supports alternative container-level CPU affinity policies by using the new CPU manager. ([#375](https://github.com/kubernetes/enhancements/issues/375), [@sjenning](https://github.com/sjenning), [@ConnorDoyle](https://github.com/ConnorDoyle))

  //   have observed 12 micros on 64-bit linux systems to wake up a parked thread). So if the
  //   timeout is small we shouldn't park(). This needs to be traded off with the cpu overhead of
  //   spinning, so we use SPIN_THRESHOLD_NANOS which is what AbstractQueuedSynchronizer uses for
  //   similar purposes.
  // * We want to behave reasonably for timeouts of 0

<a href="https://fastapi.tiangolo.com/fastapi-people/#sponsors" class="external-link" target="_blank">Other sponsors</a>

## Opinions { #opinions }

"_[...] I'm using **FastAPI** a ton these days. [...] I'm actually planning to use it for all of my team's **ML services at Microsoft**. Some of them are getting integrated into the core **Windows** product and some **Office** products._"

    private int linkSpeed;  // In units of 1 Mbps
    private byte[] sockaddrStorage;  // Socket address
    private InetAddress address;
    private boolean ipv6;
    private boolean rssCapable;  // Receive Side Scaling
    private boolean rdmaCapable;
    
    // Capability flags
    public static final int NETWORK_INTERFACE_CAP_RSS = 0x00000001;
    public static final int NETWORK_INTERFACE_CAP_RDMA = 0x00000002;
    

<!-- /sponsors -->

<a href="https://fastapi.tiangolo.com/fastapi-people/#sponsors" class="external-link" target="_blank">Other sponsors</a>

## Opinions

"_[...] I'm using **FastAPI** a ton these days. [...] I'm actually planning to use it for all of my team's **ML services at Microsoft**. Some of them are getting integrated into the core **Windows** product and some **Office** products._"