代码：git@github.com:google/crfs.git

概念

TOC ：Table of Contents
�
TOCEntry：每个文件都至少有一个TOCEntry，如果一个文件被分成多个chunk，则会有多个TOCEntry

In summary:

  • That traditional .tar.gz format is: *Gzip(TarF(file1) + TarF(file2) + TarF(file3) + TarFooter))
  • Stargz’s format is: Gzip(TarF(file1)) + Gzip(TarF(file2)) + Gzip(TarF(file3_chunk1)) + Gzip(F(file3_chunk2)) + Gzip(F(index of earlier files in magic file), TarFooter), where the trailing ZIP-like index contains offsets for each file/chunk’s GZIP header in the overall stargz file.

代码结构解析

crfs.go

layerHostOwner 对应镜像project /crfs/layers/gcr.io/foo-proj/
�
�layerHostOwnerImage 对应镜像名 /crfs/layers/gcr.io/foo-proj/ubuntu

layerHostOwnerImageReference 对应某个镜像 /crfs/layers/gcr.io/foo-proj/ubuntu/sha256-7de52a7970a2d0a7d355c76e4f0e02b0e6ebc2841f64040062a27313761cc978

  1. manifest 镜像的manifest，包含layers信息

�

�
�

  1. type node struct {
  2.     fs     *FS
  3.     te     *stargz.TOCEntry
  4. // 这个Reader是通过stargz.Open创建的，缓存了所有文件名和对应的TOCEntry
  5.     sr     *stargz.Reader
  6.     f      *os.File // non-nil if root & in debug mode
  7.     opaque bool     // true if this node is an overlayfs opaque directory
  9.     mu sync.Mutex // guards child, below
  10.     // child maps from previously-looked up base names (like "foo.txt") to the
  11.     // fspkg.Node that was previously returned. This prevents FUSE inode numbers
  12.     // from getting out of sync
  13.     child map[string]fspkg.Node
  14. }

node 对应 a CRFS node in the FUSE filesystem. 事实上就是镜像的一个layer
—》node.Open 打开一个文件，返回一个 nodeHandle
�
�nodeHandle
�—》Read 读取文件
�——》 chunkData 通过httpRangeRead读取部分数据
�