This will **depend heavily** on the way you **deploy your system**, and it would probably be connected to the way you start programs, handling restarts, etc.

Here are some possible ideas:

* An "Init Container" in Kubernetes that runs before your app container
* A bash script that runs the previous steps and then starts your application
    * You would still need a way to start/restart *that* bash script, detect errors, etc.

		// if modification time is the only timestamp being encoded.
		var mbuf [9]byte // 2*SizeOf(uint16) + SizeOf(uint8) + SizeOf(uint32)
		mt := uint32(fh.Modified.Unix())
		eb := writeBuf(mbuf[:])
		eb.uint16(extTimeExtraID)
		eb.uint16(5)  // Size: SizeOf(uint8) + SizeOf(uint32)
		eb.uint8(1)   // Flags: ModTime
		eb.uint32(mt) // ModTime
		fh.Extra = append(fh.Extra, mbuf[:]...)

/// info

如果你使用 Kubernetes，这可能是 <a href="https://kubernetes.io/docs/concepts/workloads/pods/init-containers/" class="external-link" target="_blank">Init Container</a>。

///

如果在你的用例中，运行前面的步骤**并行多次**没有问题（例如，如果你没有运行数据库迁移，而只是检查数据库是否已准备好），那么你也可以将它们放在开始主进程之前在每个容器中。

### 单容器

	if runtime.GOOS == "windows" && len(pathName) > 1024 {
		return errFileNameTooLong
	}

	// On Unix we reject paths if they are just '.', '..' or '/'
	if pathName == "." || pathName == ".." || pathName == slashSeparator {
		return errFileAccessDenied
	}

	// Check each path segment length is > 255 on all Unix
	// platforms, look for this value as NAME_MAX in
	// /usr/include/linux/limits.h
	var count int64

There are several Flask REST frameworks, but after investing the time and work into investigating them, I found that many are discontinued or abandoned, with several standing issues that made them unfit.

### <a href="https://marshmallow.readthedocs.io/en/stable/" class="external-link" target="_blank">Marshmallow</a> { #marshmallow }

/// info | 정보

만약 여러분이 쿠버네티스를 사용하고 있다면, 아마도 이는 <a href="https://kubernetes.io/docs/concepts/workloads/pods/init-containers/" class="external-link" target="_blank">Init Container</a>일 것입니다.

///

만약 여러분의 이용 사례에서 이전 단계들을 **병렬적으로 여러번** 수행하는데에 문제가 없다면 (예를 들어 데이터베이스 이전을 실행하지 않고 데이터베이스가 준비되었는지 확인만 하는 경우), 메인 프로세스를 시작하기 전에 이 단계들을 각 컨테이너에 넣을 수 있습니다.

### 단일 컨테이너

	fn := func(md5sumcurr []byte) []byte {
		return sealETag(key, md5sumcurr)
	}
	return fn
}

// compressOpts are the options for writing compressed data.
var compressOpts []s2.WriterOption

func init() {
	if runtime.GOARCH == "amd64" {
		// On amd64 we have assembly and can use stronger compression.
		compressOpts = append(compressOpts, s2.WriterBetterCompression())
	}
}

	// StopDiskAtLimit will stop listing on each disk when limit number off objects has been returned.
	StopDiskAtLimit bool

	// pool and set of where the cache is located.
	pool, set int
}

func init() {
	gob.Register(listPathOptions{})
}

func (o *listPathOptions) setBucketMeta(ctx context.Context) {
	lc, _ := globalLifecycleSys.Get(o.Bucket)
	vc, _ := globalBucketVersioningSys.Get(o.Bucket)

/// tip

另外，请记住，根据您的设置，在某些情况下，您在开始应用程序之前**可能甚至不需要任何先前的步骤**。

 在这种情况下，您就不必担心这些。 🤷

///

### 前面步骤策略的示例

这将在**很大程度上取决于您部署系统的方式**，并且可能与您启动程序、处理重启等的方式有关。

以下是一些可能的想法：

* Kubernetes 中的“Init Container”在应用程序容器之前运行
* 一个 bash 脚本，运行前面的步骤，然后启动您的应用程序
     * 您仍然需要一种方法来启动/重新启动 bash 脚本、检测错误等。

/// tip

我将在以后的章节中为您提供使用容器执行此操作的更具体示例：[容器中的 FastAPI - Docker](docker.md){.internal-link target=_blank}。

///

	default:
		return fmt.Errorf("dataUsageCache: unknown version: %d", ver)
	}
}

// Trim this from start+end of hashes.
var hashPathCutSet = dataUsageRoot

func init() {
	if dataUsageRoot != string(filepath.Separator) {
		hashPathCutSet = dataUsageRoot + string(filepath.Separator)
	}
}

// hashPath calculates a hash of the provided string.