package image import ( "bytes" "encoding/json" "io" "io/ioutil" "os" "path/filepath" "time" "github.com/Sirupsen/logrus" "github.com/containers/image/docker/reference" "github.com/containers/image/image" is "github.com/containers/image/storage" "github.com/containers/image/types" "github.com/containers/storage" "github.com/containers/storage/pkg/archive" "github.com/docker/docker/pkg/ioutils" "github.com/kubernetes-incubator/cri-o/cmd/kpod/docker" digest "github.com/opencontainers/go-digest" specs "github.com/opencontainers/image-spec/specs-go" "github.com/opencontainers/image-spec/specs-go/v1" "github.com/pkg/errors" ) // CopyRef handles image references used for copying images to/from remotes type CopyRef struct { store storage.Store compression archive.Compression name reference.Named names []string layerID string addHistory bool oconfig []byte dconfig []byte created time.Time createdBy string annotations map[string]string preferredManifestType string exporting bool } type copySource struct { path string ref *CopyRef store storage.Store layerID string names []string addHistory bool compression archive.Compression config []byte configDigest digest.Digest manifest []byte manifestType string exporting bool } // NewImage creates a new image from the given system context func (c *CopyRef) NewImage(sc *types.SystemContext) (types.Image, error) { src, err := c.NewImageSource(sc, nil) if err != nil { return nil, err } return image.FromSource(src) } func selectManifestType(preferred string, acceptable, supported []string) string { selected := preferred for _, accept := range acceptable { if preferred == accept { return preferred } for _, support := range supported { if accept == support { selected = accept } } } return selected } // NewImageSource creates a new image source from the given system context and manifest func (c *CopyRef) NewImageSource(sc *types.SystemContext, manifestTypes []string) (src types.ImageSource, err error) { // Decide which type of manifest and configuration output we're going to provide. supportedManifestTypes := []string{v1.MediaTypeImageManifest, docker.V2S2MediaTypeManifest} manifestType := selectManifestType(c.preferredManifestType, manifestTypes, supportedManifestTypes) // If it's not a format we support, return an error. if manifestType != v1.MediaTypeImageManifest && manifestType != docker.V2S2MediaTypeManifest { return nil, errors.Errorf("no supported manifest types (attempted to use %q, only know %q and %q)", manifestType, v1.MediaTypeImageManifest, docker.V2S2MediaTypeManifest) } // Start building the list of layers using the read-write layer. layers := []string{} layerID := c.layerID layer, err := c.store.Layer(layerID) if err != nil { return nil, errors.Wrapf(err, "unable to read layer %q", layerID) } // Walk the list of parent layers, prepending each as we go. for layer != nil { layers = append(append([]string{}, layerID), layers...) layerID = layer.Parent if layerID == "" { err = nil break } layer, err = c.store.Layer(layerID) if err != nil { return nil, errors.Wrapf(err, "unable to read layer %q", layerID) } } logrus.Debugf("layer list: %q", layers) // Make a temporary directory to hold blobs. path, err := ioutil.TempDir(os.TempDir(), "kpod") if err != nil { return nil, err } logrus.Debugf("using %q to hold temporary data", path) defer func() { if src == nil { err2 := os.RemoveAll(path) if err2 != nil { logrus.Errorf("error removing %q: %v", path, err) } } }() // Build fresh copies of the configurations so that we don't mess with the values in the Builder // object itself. oimage := v1.Image{} err = json.Unmarshal(c.oconfig, &oimage) if err != nil { return nil, err } dimage := docker.V2Image{} err = json.Unmarshal(c.dconfig, &dimage) if err != nil { return nil, err } // Start building manifests. omanifest := v1.Manifest{ Versioned: specs.Versioned{ SchemaVersion: 2, }, Config: v1.Descriptor{ MediaType: v1.MediaTypeImageConfig, }, Layers: []v1.Descriptor{}, Annotations: c.annotations, } dmanifest := docker.V2S2Manifest{ V2Versioned: docker.V2Versioned{ SchemaVersion: 2, MediaType: docker.V2S2MediaTypeManifest, }, Config: docker.V2S2Descriptor{ MediaType: docker.V2S2MediaTypeImageConfig, }, Layers: []docker.V2S2Descriptor{}, } oimage.RootFS.Type = docker.TypeLayers oimage.RootFS.DiffIDs = []digest.Digest{} dimage.RootFS = &docker.V2S2RootFS{} dimage.RootFS.Type = docker.TypeLayers dimage.RootFS.DiffIDs = []digest.Digest{} // Extract each layer and compute its digests, both compressed (if requested) and uncompressed. for _, layerID := range layers { omediaType := v1.MediaTypeImageLayer dmediaType := docker.V2S2MediaTypeUncompressedLayer // Figure out which media type we want to call this. Assume no compression. if c.compression != archive.Uncompressed { switch c.compression { case archive.Gzip: omediaType = v1.MediaTypeImageLayerGzip dmediaType = docker.V2S2MediaTypeLayer logrus.Debugf("compressing layer %q with gzip", layerID) case archive.Bzip2: // Until the image specs define a media type for bzip2-compressed layers, even if we know // how to decompress them, we can't try to compress layers with bzip2. return nil, errors.New("media type for bzip2-compressed layers is not defined") default: logrus.Debugf("compressing layer %q with unknown compressor(?)", layerID) } } // If we're not re-exporting the data, just fake up layer and diff IDs for the manifest. if !c.exporting { fakeLayerDigest := digest.NewDigestFromHex(digest.Canonical.String(), layerID) // Add a note in the manifest about the layer. The blobs should be identified by their // possibly-compressed blob digests, but just use the layer IDs here. olayerDescriptor := v1.Descriptor{ MediaType: omediaType, Digest: fakeLayerDigest, Size: -1, } omanifest.Layers = append(omanifest.Layers, olayerDescriptor) dlayerDescriptor := docker.V2S2Descriptor{ MediaType: dmediaType, Digest: fakeLayerDigest, Size: -1, } dmanifest.Layers = append(dmanifest.Layers, dlayerDescriptor) // Add a note about the diffID, which should be uncompressed digest of the blob, but // just use the layer ID here. oimage.RootFS.DiffIDs = append(oimage.RootFS.DiffIDs, fakeLayerDigest) dimage.RootFS.DiffIDs = append(dimage.RootFS.DiffIDs, fakeLayerDigest) continue } // Start reading the layer. rc, err := c.store.Diff("", layerID, nil) if err != nil { return nil, errors.Wrapf(err, "error extracting layer %q", layerID) } defer rc.Close() // Set up to decompress the layer, in case it's coming out compressed. Due to implementation // differences, the result may not match the digest the blob had when it was originally imported, // so we have to recompute all of this anyway if we want to be sure the digests we use will be // correct. uncompressed, err := archive.DecompressStream(rc) if err != nil { return nil, errors.Wrapf(err, "error decompressing layer %q", layerID) } defer uncompressed.Close() srcHasher := digest.Canonical.Digester() reader := io.TeeReader(uncompressed, srcHasher.Hash()) // Set up to write the possibly-recompressed blob. layerFile, err := os.OpenFile(filepath.Join(path, "layer"), os.O_CREATE|os.O_WRONLY, 0600) if err != nil { return nil, errors.Wrapf(err, "error opening file for layer %q", layerID) } destHasher := digest.Canonical.Digester() counter := ioutils.NewWriteCounter(layerFile) multiWriter := io.MultiWriter(counter, destHasher.Hash()) // Compress the layer, if we're compressing it. writer, err := archive.CompressStream(multiWriter, c.compression) if err != nil { return nil, errors.Wrapf(err, "error compressing layer %q", layerID) } size, err := io.Copy(writer, reader) if err != nil { return nil, errors.Wrapf(err, "error storing layer %q to file", layerID) } writer.Close() layerFile.Close() if c.compression == archive.Uncompressed { if size != counter.Count { return nil, errors.Errorf("error storing layer %q to file: inconsistent layer size (copied %d, wrote %d)", layerID, size, counter.Count) } } else { size = counter.Count } logrus.Debugf("layer %q size is %d bytes", layerID, size) // Rename the layer so that we can more easily find it by digest later. err = os.Rename(filepath.Join(path, "layer"), filepath.Join(path, destHasher.Digest().String())) if err != nil { return nil, errors.Wrapf(err, "error storing layer %q to file", layerID) } // Add a note in the manifest about the layer. The blobs are identified by their possibly- // compressed blob digests. olayerDescriptor := v1.Descriptor{ MediaType: omediaType, Digest: destHasher.Digest(), Size: size, } omanifest.Layers = append(omanifest.Layers, olayerDescriptor) dlayerDescriptor := docker.V2S2Descriptor{ MediaType: dmediaType, Digest: destHasher.Digest(), Size: size, } dmanifest.Layers = append(dmanifest.Layers, dlayerDescriptor) // Add a note about the diffID, which is always an uncompressed value. oimage.RootFS.DiffIDs = append(oimage.RootFS.DiffIDs, srcHasher.Digest()) dimage.RootFS.DiffIDs = append(dimage.RootFS.DiffIDs, srcHasher.Digest()) } if c.addHistory { // Build history notes in the image configurations. onews := v1.History{ Created: &c.created, CreatedBy: c.createdBy, Author: oimage.Author, EmptyLayer: false, } oimage.History = append(oimage.History, onews) dnews := docker.V2S2History{ Created: c.created, CreatedBy: c.createdBy, Author: dimage.Author, EmptyLayer: false, } dimage.History = append(dimage.History, dnews) } // Encode the image configuration blob. oconfig, err := json.Marshal(&oimage) if err != nil { return nil, err } logrus.Debugf("OCIv1 config = %s", oconfig) // Add the configuration blob to the manifest. omanifest.Config.Digest = digest.Canonical.FromBytes(oconfig) omanifest.Config.Size = int64(len(oconfig)) omanifest.Config.MediaType = v1.MediaTypeImageConfig // Encode the manifest. omanifestbytes, err := json.Marshal(&omanifest) if err != nil { return nil, err } logrus.Debugf("OCIv1 manifest = %s", omanifestbytes) // Encode the image configuration blob. dconfig, err := json.Marshal(&dimage) if err != nil { return nil, err } logrus.Debugf("Docker v2s2 config = %s", dconfig) // Add the configuration blob to the manifest. dmanifest.Config.Digest = digest.Canonical.FromBytes(dconfig) dmanifest.Config.Size = int64(len(dconfig)) dmanifest.Config.MediaType = docker.V2S2MediaTypeImageConfig // Encode the manifest. dmanifestbytes, err := json.Marshal(&dmanifest) if err != nil { return nil, err } logrus.Debugf("Docker v2s2 manifest = %s", dmanifestbytes) // Decide which manifest and configuration blobs we'll actually output. var config []byte var manifest []byte switch manifestType { case v1.MediaTypeImageManifest: manifest = omanifestbytes config = oconfig case docker.V2S2MediaTypeManifest: manifest = dmanifestbytes config = dconfig default: panic("unreachable code: unsupported manifest type") } src = ©Source{ path: path, ref: c, store: c.store, layerID: c.layerID, names: c.names, addHistory: c.addHistory, compression: c.compression, config: config, configDigest: digest.Canonical.FromBytes(config), manifest: manifest, manifestType: manifestType, exporting: c.exporting, } return src, nil } // NewImageDestination creates a new image destination from the given system context func (c *CopyRef) NewImageDestination(sc *types.SystemContext) (types.ImageDestination, error) { return nil, errors.Errorf("can't write to a container") } // DockerReference gets the docker reference for the given CopyRef func (c *CopyRef) DockerReference() reference.Named { return c.name } // StringWithinTransport returns the first name of the copyRef func (c *CopyRef) StringWithinTransport() string { if len(c.names) > 0 { return c.names[0] } return "" } // DeleteImage deletes an image in the CopyRef func (c *CopyRef) DeleteImage(*types.SystemContext) error { // we were never here return nil } // PolicyConfigurationIdentity returns the policy configuration for the CopyRef func (c *CopyRef) PolicyConfigurationIdentity() string { return "" } // PolicyConfigurationNamespaces returns the policy configuration namespace for the CopyRef func (c *CopyRef) PolicyConfigurationNamespaces() []string { return nil } // Transport returns an ImageTransport for the given CopyRef func (c *CopyRef) Transport() types.ImageTransport { return is.Transport } func (cs *copySource) Close() error { err := os.RemoveAll(cs.path) if err != nil { logrus.Errorf("error removing %q: %v", cs.path, err) } return err } func (cs *copySource) Reference() types.ImageReference { return cs.ref } func (cs *copySource) GetSignatures() ([][]byte, error) { return nil, nil } func (cs *copySource) GetTargetManifest(digest digest.Digest) ([]byte, string, error) { return []byte{}, "", errors.Errorf("TODO") } func (cs *copySource) GetManifest() ([]byte, string, error) { return cs.manifest, cs.manifestType, nil } func (cs *copySource) GetBlob(blob types.BlobInfo) (reader io.ReadCloser, size int64, err error) { if blob.Digest == cs.configDigest { logrus.Debugf("start reading config") reader := bytes.NewReader(cs.config) closer := func() error { logrus.Debugf("finished reading config") return nil } return ioutils.NewReadCloserWrapper(reader, closer), reader.Size(), nil } layerFile, err := os.OpenFile(filepath.Join(cs.path, blob.Digest.String()), os.O_RDONLY, 0600) if err != nil { logrus.Debugf("error reading layer %q: %v", blob.Digest.String(), err) return nil, -1, err } size = -1 st, err := layerFile.Stat() if err != nil { logrus.Warnf("error reading size of layer %q: %v", blob.Digest.String(), err) } else { size = st.Size() } logrus.Debugf("reading layer %q", blob.Digest.String()) closer := func() error { layerFile.Close() logrus.Debugf("finished reading layer %q", blob.Digest.String()) return nil } return ioutils.NewReadCloserWrapper(layerFile, closer), size, nil }