### Um Processo por Contêiner { #one-process-per-container }

Nesse tipo de cenário, provavelmente você desejará ter **um único processo (Uvicorn) por contêiner**, pois já estaria lidando com a replicação no nível do cluster.

	associations := []clause.Association{
		{Association: "Orders", Type: clause.OpDelete},
		{Association: "Profiles", Type: clause.OpUpdate},
	}

	// In practice, each Association would be processed individually
	// since []clause.Association doesn't implement AssociationAssigner directly
	for i, assoc := range associations {
		assigns := assoc.AssociationAssignments()
		if len(assigns) != 1 {

import org.slf4j.Logger;
import org.slf4j.LoggerFactory;

/**
 * The concern of the project is provide runtime values based on the model.
 * <p>
 * The values in the model remain untouched but during the process of building a project notions like inheritance and
 * interpolation can be added. This allows to have an entity which is useful in a runtime while preserving the model so

n := 0
for i := 0; i < m; i++ {
	n += *shared
}
</pre>

into:

<pre>
n := 0
local := *shared
for i := 0; i < m; i++ {
	n += local
}
</pre>

<p>
provided it can be proved that <code>*shared</code> will not fault on access,
because the potential added read will not affect any existing concurrent reads or writes.
On the other hand, the rewrite would not be valid in a source-to-source translator.
</p>

	}
	word, cond = p.lex.Text(), ""
	operands = scratch[:0]
	// Zero or more comma-separated operands, one per loop.
	nesting := 0
	colon := -1
	for tok != '\n' && tok != ';' {
		// Process one operand.
		var items []lex.Token
		if cap(operands) > len(operands) {
			// Reuse scratch items slice.
			items = operands[:cap(operands)][len(operands)][:0]
		} else {
			items = make([]lex.Token, 0, 3)

		}, index)
	}

	for index, err := range g.Wait() {
		if err != nil {
			sys.addNodeErr(&reply[index], sys.peerClients[index], err)
		}
	}
	return reply
}

// GetProcInfo - Process information
func (sys *NotificationSys) GetProcInfo(ctx context.Context) []madmin.ProcInfo {
	reply := make([]madmin.ProcInfo, len(sys.peerClients))

	g := errgroup.WithNErrs(len(sys.peerClients))

このマネージャー・ プロセスは、おそらくIPの**ポート**でリッスンしているものです。そして、すべての通信をワーカー・プロセスに転送します。

これらのワーカー・プロセスは、アプリケーションを実行するものであり、**リクエスト**を受けて**レスポンス**を返すための主要な計算を行い、あなたが変数に入れたものは何でもRAMにロードします。

<img src="/img/deployment/concepts/process-ram.drawio.svg">

そしてもちろん、同じマシンでは、あなたのアプリケーションとは別に、**他のプロセス**も実行されているでしょう。

興味深いことに、各プロセスが使用する**CPU**の割合は時間とともに大きく**変動**する可能性がありますが、**メモリ（RAM）**は通常、多かれ少なかれ**安定**します。

	}
	if err != nil {
		return err
	}
	t := time.Now().UTC()
	c.lastConnect.Store(&t)
	// Signal that we are reconnected, update state and handle messages.
	// Prevent other connections from connecting while we process.
	c.reconnectMu.Lock()
	if c.remoteID != nil {
		c.reconnected()
	}
	rid := uuid.UUID(req.ID)
	c.remoteID = &rid

	// Handle incoming messages until disconnect.
	c.handleMessages(ctx, conn)

### `tf.data`:

*   Adds new `tf.data.experimental.service.register_dataset` and
    `tf.data.experimental.service.from_dataset_id` APIs to enable one process to
    register a dataset with the tf.data service, and another process to consume
    data from the dataset.
*   Adds support for dispatcher fault tolerance. To enable fault tolerance,
    configure a `work_dir` when running your dispatcher server and set

			if err := p.queueMRFHeal(); err != nil && !osIsNotExist(err) {
				replLogIf(p.ctx, err)
			}
			pTimer.Reset(mrfQueueInterval)
		case <-p.ctx.Done():
			return
		}
	}
}

// process sends error logs to the heal channel for an attempt to heal replication.
func (p *ReplicationPool) queueMRFHeal() error {
	p.mrfMU.Lock()
	defer p.mrfMU.Unlock()

	if !p.initialized() {