<img class="shadow" src="/img/deployment/concepts/image01.png">

---

Now that we know the difference between the terms **process** and **program**, let's continue talking about deployments.

## Running on Startup

	_         int32 // For 64 bits alignment
	lastConn  int64

	// HealthCheckFn is the function set to test for health.
	// If not set the client will not keep track of health.
	// Calling this returns true or false if the target
	// is online or offline.
	HealthCheckFn func() bool

	// HealthCheckRetryUnit will be used to calculate the exponential
	// backoff when trying to reconnect to an offline node

   * k <= i < n}.
   *
   * <p>After this method returns, {@code elements} will contain no duplicates, but {@code elements}
   * may be the real array backing the returned set, so do not modify it further.
   *
   * <p>{@code elements} may contain only values of type {@code E}.
   *
   * @throws NullPointerException if any of the first {@code n} elements of {@code elements} is null

   *   multisetToModify.remove(e);
   * }
   * }</pre>
   *
   * @return {@code true} if {@code multisetToModify} was changed as a result of this operation
   * @since 10.0 (missing in 18.0 when only the overload taking an {@code Iterable} was present)
   */
  @CanIgnoreReturnValue
  public static boolean removeOccurrences(
      Multiset<?> multisetToModify, Multiset<?> occurrencesToRemove) {
    checkNotNull(multisetToModify);

 * deterministic.
 *
 * This class ensures that at most one thread is running at a time. This is initially the JUnit test
 * thread, which yields its execution privilege while calling [runTasks], [runNextTask], or
 * [advanceUntil]. These functions don't return until the task threads are all idle.
 *
 * Task threads release their execution privilege in these ways:
 *

   * one from the pool. We must take the pooled connection because by taking it from the pool, we've
   * fully acquired it.
   *
   * This test yields threads to force the decision of plan1 to be deliberate and not lucky. In
   * particular, we set up this sequence of events:
   *
   *  1. take plan 0
   *  3. plan 0 connects
   *  4. finish taking plan 1
   *
   * https://github.com/square/okhttp/issues/7152
   */

	err = addPodToHostNSIpset(pod, podIPs, &s.hostsideProbeIPSet)
	if err != nil {
		log.Errorf("failed to add pod to ipset: %s/%s %v", pod.Namespace, pod.Name, err)
		return err
	}

	log.Debug("calling CreateInpodRules")
	if err := s.netnsRunner(openNetns, func() error {
		return s.iptablesConfigurator.CreateInpodRules(&HostProbeSNATIP)
	}); err != nil {

Having another process manager inside the container (as would be with Gunicorn or Uvicorn managing Uvicorn workers) would only add **unnecessary complexity** that you are most probably already taking care of with your cluster system.

### Containers with Multiple Processes and Special Cases

    }

    if (queue.futureTasks.isNotEmpty()) {
      readyQueues.add(queue)
    }
  }

  /**
   * Returns an immediately-executable task for the calling thread to execute, sleeping as necessary
   * until one is ready. If there are no ready queues, or if other threads have everything under

 * streams.
 *
 * <p><i>Performance notes:</i> Unless otherwise noted, all of the iterables produced in this class
 * are <i>lazy</i>, which means that their iterators only advance the backing iteration when
 * absolutely necessary.
 *
 * <p>See the Guava User Guide article on <a href=
 * "https://github.com/google/guava/wiki/CollectionUtilitiesExplained#iterables">{@code
 * Iterables}</a>.
 *