- **[alpha]** Represent node problems "not ready" and "unreachable" using `NoExecute` taints. In combination with `tolerationSeconds` described below, this allows per-pod specification of how long to remain bound to a node that becomes unreachable or not ready, rather than using the default of 5 minutes. You can enable this alpha feature by including `TaintBasedEvictions=true` in `--feature-gates`,...

  @Test
  fun tlsTimeoutsNotRetried() {
    enableTls()
    server.enqueue(MockResponse.Builder().onResponseStart(Stall).build())
    server.enqueue(MockResponse(body = "unreachable!"))
    client =
      client
        .newBuilder()
        .readTimeout(Duration.ofMillis(100))
        .build()
    val request = Request.Builder().url(server.url("/")).build()

- The deprecated taints `node.alpha.kubernetes.io/notReady` and `node.alpha.kubernetes.io/unreachable` are no longer supported or adjusted. These uses should be replaced with `node.kubernetes.io/not-ready` and `node.kubernetes.io/unreachable`
 ([#73001](https://github.com/kubernetes/kubernetes/pull/73001), [@shivnagarajan](https://github.com/shivnagarajan))

    }

    @Override
    public Set<Entry<K, V>> entries() {
      return map.entrySet();
    }

    @Override
    Collection<Entry<K, V>> createEntries() {
      throw new AssertionError("unreachable");
    }

    @Override
    Multiset<K> createKeys() {
      return new Multimaps.Keys<K, V>(this);
    }

    @Override
    Iterator<Entry<K, V>> entryIterator() {

    }

    @Override
    public Set<Entry<K, V>> entries() {
      return map.entrySet();
    }

    @Override
    Collection<Entry<K, V>> createEntries() {
      throw new AssertionError("unreachable");
    }

    @Override
    Multiset<K> createKeys() {
      return new Multimaps.Keys<K, V>(this);
    }

    @Override
    Iterator<Entry<K, V>> entryIterator() {

    or force deletion of the pod from the api-server).
    This helps prevent "split brain" scenarios in clustered applications by
    ensuring that unreachable pods will not be presumed dead unless some
    "fencing" operation has provided one of the above indications.
  * For all other existing controllers except StatefulSet, this has no effect on

#### Node controller
* Bug fixes:

  * Fix [transition between NotReady and Unreachable taints](https://github.com/kubernetes/kubernetes/issues/43444). ([#44042](https://github.com/kubernetes/kubernetes/pull/44042), [@gmarek](https://github.com/gmarek))


#### Node Components

* Features

- Reduced the 5s delay before tainting `node.kubernetes.io/unreachable:NoExecute` when a Node became unreachable. ([#120816](https://github.com/kubernetes/kubernetes/pull/120816), [@tnqn](https://github.com/tnqn)) [SIG Apps and Node]

- Performance of the anti-affinity predicate of the default scheduler has been improved. ([#66948](https://github.com/kubernetes/kubernetes/pull/66948), [@mohamed-mehany](https://github.com/mohamed-mehany))
- The unreachable taint gets applied to a node when it loses its network connection. ([#67734](https://github.com/kubernetes/kubernetes/pull/67734), [@Huang-Wei](https://github.com/Huang-Wei))

        && network.edges().size() > network.asGraph().edges().size()) {
      return true;
    }
    return hasCycle(network.asGraph());
  }

  /**
   * Performs a traversal of the nodes reachable from {@code startNode}. If we ever reach a node
   * we've already visited (following only outgoing edges and without reusing edges), we know
   * there's a cycle in the graph.
   */