- The "extraArgs" component construct is now a list of "name"/"value" pairs instead of a string/string map. This has been done to support duplicate args where needed.
  - The "JoinConfiguration.discovery.timeout" field has been replaced by "JoinConfiguration.timeouts.discovery".
  - The "ClusterConfiguration.timeoutForControlPlane" field has been replaced by "{Init|Join}Configuration.timeouts.controlPlaneComponentHealthCheck".

	//	$sym±offset(SB)(reg)(index*scale)
	// Exceptions are:
	//
	//	R1
	//	offset
	//	$offset
	// Every piece is optional, so we scan left to right and what
	// we discover tells us where we are.

	// Prefix: $.
	var prefix rune
	switch tok := p.peek(); tok {
	case '$', '*':
		prefix = rune(tok)
		p.next()
	}

	// Symbol: sym±offset(SB)
	tok := p.next()

   * since the concurrency control must manage multiple Conditions. So this method must generally be
   * called when waits are interrupted.
   *
   * <p>On the other hand, if a signalled thread wakes up to discover that its guard is still not
   * satisfied, it does *not* need to call this method before returning to wait. This can only
   * happen due to spurious wakeup (ignorable) or another thread acquiring the lock before the

    cache.close()
    createNewCache()

    // The journal has no record that 'a' and 'b' were removed. It will have an entry for both, but
    // when it tries to read the cache files for either entry, it will discover the cache files are
    // missing and remove the entries from the cache.
    assertThat(cache.size()).isEqualTo(4)
    assertThat(cache["a"]).isNull()
    assertThat(cache["b"]).isNull()

                        @Override
                        protected void configure() {
                            try {
                                container.lookup(Injector.class).discover(extension.getClassRealm());
                            } catch (Throwable e) {
                                // ignore
                                e.printStackTrace();
                            }

* Make discovery summarizer call servers in parallel ([#26705](https://github.com/kubernetes/kubernetes/pull/26705), [@nebril](https://github.com/nebril))

labels.chat_phase_evaluate=Reviewing search results...
labels.chat_phase_fetch=Retrieving documents...
labels.chat_phase_answer=Generating answer...
labels.chat_messages_area=Chat messages
labels.chat_welcome_title=Ask and Discover.
labels.chat_welcome_description=Ask questions and I'll search through documents to find the answers you need.
labels.chat_step_intent=Analyze
labels.chat_step_search=Search
labels.chat_step_evaluate=Evaluate

- client-go discovery clients constructed using `NewDiscoveryClientForConfig` or `NewDiscoveryClientForConfigOrDie` default to rate limits that cause normal discovery request patterns to take several seconds. This is fixed in https://issue.k8s.io/86168 and will be resolved in v1.17.1. As a workaround, the `Burst` value can be adjusted higher...

- Resolves a spurious "Unknown discovery response content-type" error in client-go discovery requests by tolerating extra content-type parameters in API responses ([#117571](https://github.com/kubernetes/kubernetes/pull/117571), [@seans3](https://github.com/seans3)) [SIG API Machinery]

- Aggregated Discovery v2beta1 fixture is removed in `./api/discovery`. Please use v2 ([#127008](https://github.com/kubernetes/kubernetes/pull/127008), [@Jefftree](https://github.com/Jefftree)) [SIG API Machinery]