dockerpull命令实现与镜像存储（1）

我们在《docker命令解析》篇章我们了解了命令的解析过程&＃xff0c;所以不再赘述。我们直接看执行命令任务的代码。
定位到docker\cli\command\commands\commands.go的AddCommands函数&＃xff0c;我们容易找到pull命令的实现函数 在hide(image.NewPullCommand(dockerCli))注册。我们进入该函数&＃xff1a;

// NewPullCommand creates a new &＃96;docker pull&＃96; command
func NewPullCommand(dockerCli *command.DockerCli) *cobra.Command {var opts pullOptionscmd :&＃61; &cobra.Command{Use: "pull [OPTIONS] NAME[:TAG|&＃64;DIGEST]",Short: "Pull an image or a repository from a registry",Args: cli.ExactArgs(1),RunE: func(cmd *cobra.Command, args []string) error { //镜像名字&＃xff0c;如&＃xff1a;docker pull ubuntu,则args[0]就是ubuntuopts.remote &＃61; args[0]return runPull(dockerCli, opts)},}flags :&＃61; cmd.Flags()flags.BoolVarP(&opts.all, "all-tags", "a", false, "Download all tagged images in the repository")command.AddTrustedFlags(flags, true)return cmd
}

我们了解了命令的解析过程&＃xff0c;容易知道将执行函数runPull&＃xff0c;同时将拉取的镜像参数传入&＃xff08;镜像名&＃xff0c;版本&＃xff0c;是否所有tag等&＃xff09;&＃xff0c;我们看下函数runPull&＃xff1a;

func runPull(dockerCli *command.DockerCli, opts pullOptions) error {//从参数中解析出带镜像仓库地址等信息的镜像引用&＃xff0c;如果参数中没有仓库地址信息&＃xff0c;则使用默认的docker.iodistributionRef, err :&＃61; reference.ParseNamed(opts.remote)if err !&＃61; nil {return err}// -a, --all-tags Download all tagged images in the repository//如果使用了all选项&＃xff0c;但是又不是只有镜像名&＃xff08;包含tag&＃xff09;&＃xff0c;则报错处理if opts.all && !reference.IsNameOnly(distributionRef) {return errors.New("tag can&＃39;t be used with --all-tags/-a")}//如果没有使用all选项&＃xff0c;且只有镜像名&＃xff0c;则添加一个默认的tag&＃xff08;latest&＃xff09;if !opts.all && reference.IsNameOnly(distributionRef) {distributionRef &＃61; reference.WithDefaultTag(distributionRef)fmt.Fprintf(dockerCli.Out(), "Using default tag: %s\n", reference.DefaultTag)}var tag stringswitch x :&＃61; distributionRef.(type) {//标准的case reference.Canonical:tag &＃61; x.Digest().String()//name:tag形式case reference.NamedTagged:tag &＃61; x.Tag()}registryRef :&＃61; registry.ParseReference(tag)// Resolve the Repository name from fqn to RepositoryInforepoInfo, err :&＃61; registry.ParseRepositoryInfo(distributionRef)if err !&＃61; nil {return err}ctx :&＃61; context.Background()authConfig :&＃61; command.ResolveAuthConfig(ctx, dockerCli, repoInfo.Index)requestPrivilege :&＃61; command.RegistryAuthenticationPrivilegedFunc(dockerCli, repoInfo.Index, "pull")//如果没有添加disable-content-trust,而且没有附带数字摘要,则不对镜像进行校验if command.IsTrusted() && !registryRef.HasDigest() {// Check if tag is digesterr &＃61; trustedPull(ctx, dockerCli, repoInfo, registryRef, authConfig, requestPrivilege)} else {//向dockerd发送拉取镜像请求err &＃61; imagePullPrivileged(ctx, dockerCli, authConfig, distributionRef.String(), requestPrivilege, opts.all)}if err !&＃61; nil {if strings.Contains(err.Error(), "target is a plugin") {return errors.New(err.Error() &＃43; " - Use &＃96;docker plugin install&＃96;")}return err}return nil
}

该函数做了两件事&＃xff1a;解析输入参数填充Named结构对象&＃xff0c;用字符串化的Named对象拉取镜像。在详细说明之前&＃xff0c;我们有必要讲一下Named这个接口。

// Named is an object with a full name
type Named interface {// Name returns normalized repository name, like "ubuntu".Name() string// String returns full reference, like "ubuntu&＃64;sha256:abcdef..."String() string// FullName returns full repository name with hostname, like "docker.io/library/ubuntu"FullName() string// Hostname returns hostname for the reference, like "docker.io"Hostname() string// RemoteName returns the repository component of the full name, like "library/ubuntu"RemoteName() string
}

Named接口有两个子接口带数字摘要的Canonical和带tag的NamedTagged&＃xff1a;

//带数字摘要的形式
// Canonical reference is an object with a fully unique
// name including a name with hostname and digest
type Canonical interface {NamedDigest() digest.Digest
}
//带tag的形式
// NamedTagged is an object including a name and tag.
type NamedTagged interface {NamedTag() string
}

我们知道拉取镜像的命令&＃xff1a;docker pull NAME[:TAG|&＃64;DIGEST] &＃xff0c;TAG代表标签&＃xff0c;DIGEST代表数字摘要&＃xff0c;意思就是我们拉取镜像参数可以附带TAG或数字摘要&＃xff0c;或者只带镜像名&＃xff08;系统会提供一个默认的标签latest&＃xff09;。如果我们提供的参数带TAG则使用NamedTagged描述 &＃xff0c;如果我们提供的参数带DIGEST则使用Canonical 描述。现在我们简单分析下这个解析过程&＃xff0c;函数调用过程&＃xff1a;
reference.ParseNamed(opts.remote)–>distreference.ParseNamed(s)–> Parse(s)
函数reference.ParseNamed(opts.remote)实现在docker\reference\reference.go

函数distreference.ParseNamed(s)和函数Parse(s)都是定义在文件docker\vendor\src\github.com\docker\distribution\reference\reference.go
可以发现文件名都为reference.go,感觉起来就有点蹊跷&＃xff0c;事实上感觉是对的。我们看下两个文件的结构&＃xff08;上面是docker\reference\reference.go&＃xff09;

两相对比&＃xff0c;可以发现两个文件都定义了Named&＃xff0c;Canonical&＃xff0c;NamedTagged三个接口&＃xff0c;而且接口间的关系也是一样的。实际上参数的正则匹配是在后者的Parse函数完成&＃xff0c;一切做好之后&＃xff0c;才在前者的reference.ParseNamed(opts.remote)函数中做一个转化&＃xff08;暂时还不了解为何要这样写代码&＃xff09;&＃xff0c;看reference.ParseNamed(opts.remote)函数&＃xff1a;

// ParseNamed parses s and returns a syntactically valid reference implementing
// the Named interface. The reference must have a name, otherwise an error is
// returned.
// If an error was encountered it is returned, along with a nil Reference.
func ParseNamed(s string) (Named, error) {named, err :&＃61; distreference.ParseNamed(s)if err !&＃61; nil {return nil, fmt.Errorf("Error parsing reference: %q is not a valid repository/tag: %s", s, err)}// If no valid hostname is found, the default hostname is used./如果没有有效的主机名&＃xff0c;则使用默认的主机名docker.io//将distreference.Namded转化为reference.Namedr, err :&＃61; WithName(named.Name())if err !&＃61; nil {return nil, err}if canonical, isCanonical :&＃61; named.(distreference.Canonical); isCanonical {//将distreference.Canonical转化为reference.Canonicalreturn WithDigest(r, canonical.Digest())}if tagged, isTagged :&＃61; named.(distreference.NamedTagged); isTagged {//将distreference.NamedTagged转化为reference.NamedTaggedreturn WithTag(r, tagged.Tag())}return r, nil
}

reference.ParseNamed(opts.remote)函数不过是个马甲&＃xff0c;实际工作并不是自己做的&＃xff0c;看下完成正则匹配的Parse函数&＃xff1a;

// Parse parses s and returns a syntactically valid Reference.
// If an error was encountered it is returned, along with a nil Reference.
// NOTE: Parse will not handle short digests.
func Parse(s string) (Reference, error) {////matches :&＃61; ReferenceRegexp.FindStringSubmatch(s)if matches &＃61;&＃61; nil {if s &＃61;&＃61; "" {return nil, ErrNameEmpty}// TODO(dmcgowan): Provide more specific and helpful errorreturn nil, ErrReferenceInvalidFormat}if len(matches[1]) > NameTotalLengthMax {return nil, ErrNameTooLong}ref :&＃61; reference{name: matches[1],tag: matches[2],}//带数字摘要&＃xff0c;有SHA256, SHA384, SHA512&＃xff0c;一般为SHA256if matches[3] !&＃61; "" {var err error//主要是校验ref.digest, err &＃61; digest.ParseDigest(matches[3])if err !&＃61; nil {return nil, err}}//这里根据解析参数是否包含镜像名&＃xff0c;是否包含标签&＃xff0c;以及是否包含数字摘要来决定返回引用的类型r :&＃61; getBestReferenceType(ref)if r &＃61;&＃61; nil {return nil, ErrNameEmpty}return r, nil
}

ReferenceRegexp匹配规则定义docker\vendor\src\github.com\docker\distribution\reference\regexp.go

ReferenceRegexp &＃61; anchored(capture(NameRegexp),optional(literal(":"), capture(TagRegexp)),optional(literal("&＃64;"), capture(DigestRegexp)))

可以看到跟我们的命令的形式是对应的&＃xff0c;如果带镜像名&＃xff0c;则matches[1]不为空&＃xff0c;如果带tag&＃xff0c;则matches[2]不为空&＃xff0c;如果带数字摘要&＃xff0c;则matches[3]不为空。getBestReferenceType根据各个matchs是否为空&＃xff0c;返回对应的引用Reference。我们接着分析下函数getBestReferenceType&＃xff1a;

func getBestReferenceType(ref reference) Reference {//只带数字摘要if ref.name &＃61;&＃61; "" {// Allow digest only referencesif ref.digest !&＃61; "" {return digestReference(ref.digest)}return nil}//带数字摘要和镜像名if ref.tag &＃61;&＃61; "" {if ref.digest !&＃61; "" {return canonicalReference{name: ref.name,digest: ref.digest,}}return repository(ref.name)}//带标签和镜像名if ref.digest &＃61;&＃61; "" {return taggedReference{name: ref.name,tag: ref.tag,}}return ref
}

函数逻辑很简单&＃xff0c;就是根据是否带相应的部分返回不同类型的Reference 。
好了&＃xff0c;我们把上面的过程梳理下&＃xff1a;
第一&＃xff0c;我们传入拉取镜像的参数&＃xff0c;如我们执行docker pull ubuntu:latest,则“ubuntu:latest”将被Parse解析为三个部分matches[1]&＃61;ubuntu,matches[2]&＃61;latest,matches[3]&＃61;”“&＃xff0c;并返回NamedTagged类型的Reference对象&＃xff08;distreference.Named为Reference的子接口&＃xff0c;也即是返回distreference.Named对象&＃xff09;

第二&＃xff0c;reference.ParseNamed(opts.remote)将Reference(distreference.Named)对象转化为reference.Named对象
第三&＃xff0c;reference.ParseNamed(opts.remote)返回reference.Named给pull函数使用
转了那么多圈&＃xff0c;也就干了这么点事情。