moby/daemon/containerd/image_exporter.go
Paweł Gronowski 5e13f54f57
c8d/save: Handle digested reference same as ID
When saving an image treat `image@sha256:abcdef...` the same as
`abcdef...`, this makes it:

- Not export the digested tag as the image name
- Not try to export all tags from the image repository

Signed-off-by: Paweł Gronowski <pawel.gronowski@docker.com>
2024-01-29 16:29:05 +01:00

319 lines
9.8 KiB
Go

package containerd
import (
"context"
"fmt"
"io"
"strings"
"github.com/containerd/containerd"
"github.com/containerd/containerd/content"
cerrdefs "github.com/containerd/containerd/errdefs"
containerdimages "github.com/containerd/containerd/images"
"github.com/containerd/containerd/images/archive"
"github.com/containerd/containerd/leases"
"github.com/containerd/containerd/platforms"
"github.com/containerd/log"
"github.com/distribution/reference"
"github.com/docker/docker/api/types/events"
"github.com/docker/docker/container"
"github.com/docker/docker/daemon/images"
"github.com/docker/docker/errdefs"
dockerarchive "github.com/docker/docker/pkg/archive"
"github.com/docker/docker/pkg/streamformatter"
ocispec "github.com/opencontainers/image-spec/specs-go/v1"
"github.com/pkg/errors"
)
func (i *ImageService) PerformWithBaseFS(ctx context.Context, c *container.Container, fn func(root string) error) error {
snapshotter := i.client.SnapshotService(c.Driver)
mounts, err := snapshotter.Mounts(ctx, c.ID)
if err != nil {
return err
}
path, err := i.refCountMounter.Mount(mounts, c.ID)
if err != nil {
return err
}
defer i.refCountMounter.Unmount(path)
return fn(path)
}
// ExportImage exports a list of images to the given output stream. The
// exported images are archived into a tar when written to the output
// stream. All images with the given tag and all versions containing
// the same tag are exported. names is the set of tags to export, and
// outStream is the writer which the images are written to.
//
// TODO(thaJeztah): produce JSON stream progress response and image events; see https://github.com/moby/moby/issues/43910
func (i *ImageService) ExportImage(ctx context.Context, names []string, outStream io.Writer) error {
platform := matchAllWithPreference(platforms.Default())
opts := []archive.ExportOpt{
archive.WithSkipNonDistributableBlobs(),
// This makes the exported archive also include `manifest.json`
// when the image is a manifest list. It is needed for backwards
// compatibility with Docker image format.
// The containerd will choose only one manifest for the `manifest.json`.
// Our preference is to have it point to the default platform.
// Example:
// Daemon is running on linux/arm64
// When we export linux/amd64 and linux/arm64, manifest.json will point to linux/arm64.
// When we export linux/amd64 only, manifest.json will point to linux/amd64.
// Note: This is only applicable if importing this archive into non-containerd Docker.
// Importing the same archive into containerd, will not restrict the platforms.
archive.WithPlatform(platform),
archive.WithSkipMissing(i.content),
}
leasesManager := i.client.LeasesService()
lease, err := leasesManager.Create(ctx, leases.WithRandomID())
if err != nil {
return errdefs.System(err)
}
defer func() {
if err := leasesManager.Delete(ctx, lease); err != nil {
log.G(ctx).WithError(err).Warn("cleaning up lease")
}
}()
addLease := func(ctx context.Context, target ocispec.Descriptor) error {
return leaseContent(ctx, i.content, leasesManager, lease, target)
}
exportImage := func(ctx context.Context, target ocispec.Descriptor, ref reference.Named) error {
if err := addLease(ctx, target); err != nil {
return err
}
if ref != nil {
opts = append(opts, archive.WithManifest(target, ref.String()))
log.G(ctx).WithFields(log.Fields{
"target": target,
"name": ref,
}).Debug("export image")
} else {
orgTarget := target
target.Annotations = make(map[string]string)
for k, v := range orgTarget.Annotations {
switch k {
case containerdimages.AnnotationImageName, ocispec.AnnotationRefName:
// Strip image name/tag annotations from the descriptor.
// Otherwise containerd will use it as name.
default:
target.Annotations[k] = v
}
}
opts = append(opts, archive.WithManifest(target))
log.G(ctx).WithFields(log.Fields{
"target": target,
}).Debug("export image without name")
}
i.LogImageEvent(target.Digest.String(), target.Digest.String(), events.ActionSave)
return nil
}
exportRepository := func(ctx context.Context, ref reference.Named) error {
imgs, err := i.getAllImagesWithRepository(ctx, ref)
if err != nil {
return errdefs.System(fmt.Errorf("failed to list all images from repository %s: %w", ref.Name(), err))
}
if len(imgs) == 0 {
return images.ErrImageDoesNotExist{Ref: ref}
}
for _, img := range imgs {
ref, err := reference.ParseNamed(img.Name)
if err != nil {
log.G(ctx).WithFields(log.Fields{
"image": img.Name,
"error": err,
}).Warn("couldn't parse image name as a valid named reference")
continue
}
if err := exportImage(ctx, img.Target, ref); err != nil {
return err
}
}
return nil
}
for _, name := range names {
target, resolveErr := i.resolveDescriptor(ctx, name)
// Check if the requested name is a truncated digest of the resolved descriptor.
// If yes, that means that the user specified a specific image ID so
// it's not referencing a repository.
specificDigestResolved := false
if resolveErr == nil {
nameWithoutDigestAlgorithm := strings.TrimPrefix(name, target.Digest.Algorithm().String()+":")
specificDigestResolved = strings.HasPrefix(target.Digest.Encoded(), nameWithoutDigestAlgorithm)
}
log.G(ctx).WithFields(log.Fields{
"name": name,
"resolveErr": resolveErr,
"specificDigestResolved": specificDigestResolved,
}).Debug("export requested")
ref, refErr := reference.ParseNormalizedNamed(name)
if refErr == nil {
if _, ok := ref.(reference.Digested); ok {
specificDigestResolved = true
}
}
if resolveErr != nil || !specificDigestResolved {
// Name didn't resolve to anything, or name wasn't explicitly referencing a digest
if refErr == nil && reference.IsNameOnly(ref) {
// Reference is valid, but doesn't include a specific tag.
// Export all images with the same repository.
if err := exportRepository(ctx, ref); err != nil {
return err
}
continue
}
}
if resolveErr != nil {
return resolveErr
}
if refErr != nil {
return refErr
}
// If user exports a specific digest, it shouldn't have a tag.
if specificDigestResolved {
ref = nil
}
if err := exportImage(ctx, target, ref); err != nil {
return err
}
}
return i.client.Export(ctx, outStream, opts...)
}
// leaseContent will add a resource to the lease for each child of the descriptor making sure that it and
// its children won't be deleted while the lease exists
func leaseContent(ctx context.Context, store content.Store, leasesManager leases.Manager, lease leases.Lease, desc ocispec.Descriptor) error {
return containerdimages.Walk(ctx, containerdimages.HandlerFunc(func(ctx context.Context, desc ocispec.Descriptor) ([]ocispec.Descriptor, error) {
_, err := store.Info(ctx, desc.Digest)
if err != nil {
if errors.Is(err, cerrdefs.ErrNotFound) {
return nil, nil
}
return nil, errdefs.System(err)
}
r := leases.Resource{
ID: desc.Digest.String(),
Type: "content",
}
if err := leasesManager.AddResource(ctx, lease, r); err != nil {
return nil, errdefs.System(err)
}
return containerdimages.Children(ctx, store, desc)
}), desc)
}
// LoadImage uploads a set of images into the repository. This is the
// complement of ExportImage. The input stream is an uncompressed tar
// ball containing images and metadata.
func (i *ImageService) LoadImage(ctx context.Context, inTar io.ReadCloser, outStream io.Writer, quiet bool) error {
decompressed, err := dockerarchive.DecompressStream(inTar)
if err != nil {
return errors.Wrap(err, "failed to decompress input tar archive")
}
defer decompressed.Close()
opts := []containerd.ImportOpt{
// TODO(vvoland): Allow user to pass platform
containerd.WithImportPlatform(platforms.All),
containerd.WithSkipMissing(),
// Create an additional image with dangling name for imported images...
containerd.WithDigestRef(danglingImageName),
// ... but only if they don't have a name or it's invalid.
containerd.WithSkipDigestRef(func(nameFromArchive string) bool {
if nameFromArchive == "" {
return false
}
_, err := reference.ParseNormalizedNamed(nameFromArchive)
return err == nil
}),
}
imgs, err := i.client.Import(ctx, decompressed, opts...)
if err != nil {
log.G(ctx).WithError(err).Debug("failed to import image to containerd")
return errdefs.System(err)
}
progress := streamformatter.NewStdoutWriter(outStream)
for _, img := range imgs {
name := img.Name
loadedMsg := "Loaded image"
if isDanglingImage(img) {
name = img.Target.Digest.String()
loadedMsg = "Loaded image ID"
} else if named, err := reference.ParseNormalizedNamed(img.Name); err == nil {
name = reference.FamiliarString(reference.TagNameOnly(named))
}
err = i.walkImageManifests(ctx, img, func(platformImg *ImageManifest) error {
logger := log.G(ctx).WithFields(log.Fields{
"image": name,
"manifest": platformImg.Target().Digest,
})
if isPseudo, err := platformImg.IsPseudoImage(ctx); isPseudo || err != nil {
if err != nil {
logger.WithError(err).Warn("failed to read manifest")
} else {
logger.Debug("don't unpack non-image manifest")
}
return nil
}
unpacked, err := platformImg.IsUnpacked(ctx, i.snapshotter)
if err != nil {
logger.WithError(err).Warn("failed to check if image is unpacked")
return nil
}
if !unpacked {
err = platformImg.Unpack(ctx, i.snapshotter)
if err != nil {
return errdefs.System(err)
}
}
logger.WithField("alreadyUnpacked", unpacked).WithError(err).Debug("unpack")
return nil
})
if err != nil {
return errors.Wrap(err, "failed to unpack loaded image")
}
fmt.Fprintf(progress, "%s: %s\n", loadedMsg, name)
i.LogImageEvent(img.Target.Digest.String(), img.Target.Digest.String(), events.ActionLoad)
}
return nil
}