FastCdcFs.Net.Shell 1.1.0-alpha06

FastCdcFS.Net

A .NET library for creating read-only virtual file systems backed by fast content-defined chunking. It allows efficient storage, deduplication, and retrieval of files and directories while keeping access fast and lightweight. Ideal for scenarios where immutable file systems are needed, such as archival storage, distribution, or data integrity�focused applications.

Usage

Create a file system

Create the file system with shell / terminal

Use FastCdcFs.Shell to a create virtual file-system:

fastcdcfs build -d "/directory/to/add" -o myfs.fastcdcfs

fastcdcfs build -f "/file/to/add"
fastcdcfs build -f "/another/file/to/add" --target "non/root/destination"
fastcdcfs build -d "/directory/to/add"
fastcdcfs build -d "/another/directory/to/add" --target "another/non/root/destination" --recursive
fastcdcfs build -o myfs.fastcdcfs

Create the file system programmatically

var writer = new FastCdcFsWriter(Options.Default);
writer.AddFile("/file/to/add");
writer.AddFile("/another/file/to/add", "non/root/destination");
writer.Build("/target/location");

Read a file system

Read the file system with shell / terminal

# list files (optional pass --directory to limit outputs to a specific directory)
fastcdcfs list myfs.fastcdcfs
fastcdcfs list myfs.fastcdcfs --directory "/non/root/destination"

# extract everything
fastcdcfs extract myfs.fastcdcfs --target "/extract/to"

# extract file
fastcdcfs extract --file "non/root/destination/file" myfs.fastcdcfs --target "/extract/to"

# extract directory (optional recursively)
fastcdcfs extract --directory "non/root/destination/file" myfs.fastcdcfs --target "/extract/to"

Read the file system programmatically

using var reader = new CdcFsReader("myfs.fastcdcfs");

foreach (var file in reader.List().Where(e => e.IsFile))
{
    // read all bytes
    var data = file.ReadAllBytes();

    // get stream
    var stream = file.OpenRead();
}

using var reader = new CdcFsReader("myfs.fastcdcfs");

foreach (var file in reader.List("a/known/directory"))
{
    // ...
}

using var reader = new CdcFsReader("myfs.fastcdcfs");

// read known entries immediately
var entry = reader.Get("some/known/file/path");
var stream = entry!.Open();

File System Format

Version 2 (Current)

• Improved zero file size handling
• Support for no compression dict even with compression

+--------------------------------------------------------------------------------------+
| [0x00..0x08]      magic: utf8 "FASTCDCFS"         | identifies file
| [0x09..0x09]      version: byte (2)               | identifies the file system version
| [0x0A..0x0A]      mode: byte                      | identifies the file system modes
| [0x0B..0x0C]      meta data length: u32           | length of metadata
| [0x0D..0x0E]  -   directory count: u32            | number of directories
| [0x0F..]     |C|  directory table: <repeated>
| [..]         |O|    parent id: u32                | parent id of directory
| [..]         |M|    name: utf8                    | name of directory
| [..]         |P|  files count: u32                | number of files
| [..]         |R|  files table: <repeated>
| [..]         |E|    directory id: u32             | id of directory
| [..]         |S|    name: utf8                    | name of file
| [..]         |S|    length: u32                   | length of file
| [..]         |E|    chunk count: u32              | number of chunk ids, not available when version > 1 and length = 0
| [..]         |D|    file chunk table: <repeated>
| [..]         |*|      chunk id: u32               | chunk id
| [..]         |.|  compression dict length: u32    | length of compression dict *
| [..]         |.|  compression dict: raw           | compression dict *
| [..]         |.|  chunk boundary count: u32       | number of chunk boundaries
| [..]         |.|  chunk boundary table: <repeated>
| [..]         |.|    chunk length: u32             | length of chunk
| [..]         | |    compressed chunk length: u32  | length of compressed chunk *
| [..]          -     chunk hash: u64               | xxHash64 of the chunk **
| [..]              meta data hash: u64             | xxHash64 of the meta data **
| [..]              chunk blobs: raw                | chunk blobs
+--------------------------------------------------------------------------------------+

* only available when the mode is not nozstd
** only available when the mode is not nohash

• the meta data is only compressed when the mode is not nozstd
• the utf8 string encoding uses a 7-bit encoded length prefix
• the id of a directory is the index in the directory table
• the id of a chunk is the index in the chunk boundary table

File System Modes

None:   0x0
NoZstd: 0x1
NoHash: 0x2