// token will receive heal status records from the running heal
// sequence.
//
// If no client token is provided, and a heal sequence is in progress
// an error is returned with information about the running heal
// sequence. However, if the force-start flag is provided, the server
// aborts the running heal sequence and starts a new one.

        replicas taking part in sync training.
    *   Performance improvements for GPU multi-worker distributed training using
        `tf.distribute.experimental.MultiWorkerMirroredStrategy`
    *   Update NVIDIA `NCCL` to `2.5.7-1` for better performance and performance
        tuning. Please see

The recommended way to handle the *startup* and *shutdown* is using the `lifespan` parameter of the `FastAPI` app as described above. If you provide a `lifespan` parameter, `startup` and `shutdown` event handlers will no longer be called. It's all `lifespan` or all events, not both.

You can probably skip this part.

///

There's an alternative way to define this logic to be executed during *startup* and during *shutdown*.

to the network can execute arbitrary code with the privileges of the user
running the `tf.train.Server`.

## Untrusted inputs during training and prediction

TensorFlow supports a wide range of input data formats. For example it can
process images, audio, videos, and text. There are several modules specialized
in taking those formats, modifying them, and/or converting them to intermediate

  }

  /**
   * This covers a bug where if a listener was installed that would stop the manager if any service
   * fails and something failed during startup before service.start was called on all the services,
   * then awaitStopped would deadlock due to an IllegalStateException that was thrown when trying to
   * stop the timer(!).
   */
  public void testFailStart_stopOthers() throws TimeoutException {
    Service a = new FailStartService();

    String result;
    if (state == DONE) {
      result = "running=[DONE]";
    } else if (state instanceof Blocker) {
      result = "running=[INTERRUPTED]";
    } else if (state instanceof Thread) {
      // getName is final on Thread, no need to worry about exceptions
      result = "running=[RUNNING ON " + ((Thread) state).getName() + "]";
    } else {
      result = "running=[NOT STARTED YET]";
    }

And the **container** itself (in contrast to the **container image**) is the actual running instance of the image, comparable to a **process**. In fact, a container is running only when it has a **process running** (and normally it's only a single process). The container stops when there's no process running in it.

## Container Images

	openNetns := s.currentPodSnapshot.Get(string(pod.UID))
	if openNetns != nil {
		return openNetns, nil
	}
	log.Debug("pod netns was not found, trying to find it using procfs")
	// this can happen if the pod was dynamically added to the mesh after it was created.
	// in that case, try finding the netns using procfs.
	if err := s.rescanPod(pod); err != nil {
		log.Errorf("error scanning proc: error was %s", err)
		return nil, err
	}

* There can be **multiple processes** of the **same program** running at the same time.

If you check out the "task manager" or "system monitor" (or similar tools) in your operating system, you will be able to see many of those processes running.

     * <p>The default {@link AbstractFuture} implementation throws {@code InterruptedException} if
     * the current thread is interrupted before or during the call, even if the value is already
     * available.
     *
     * @throws InterruptedException if the current thread was interrupted before or during the call
     *     (optional but recommended).
     * @throws CancellationException {@inheritDoc}
     */
    @CanIgnoreReturnValue