Borg Backup

"The holy grail of backup software"

Thomas Waldmann @ shack-a-days 2015

Borg - a fork of Attic

• Attic: about 5y old, good design, proven code

• but:

  • going very slowly, main dev has little time

  • not very open to new developers

• Borg Backup: fork of Attic stable codebase

• and:

  • a community project, bus_factor++

  • has a lot of fixes and good PRs merged

  • open and inviting to new contributors

  • we want your ideas, feedback, contribs!

  • faster paced

Feature Set (1)

• easy and fast
• content-defined chunking (*)
• chunk deduplication (*)
• lz4, zlib, lzma compression

• encryption with AES256

• authentication with HMAC-SHA2

• simple backend (k/v, fs, via ssh)

Feature Set (2)

• FOSS (BSD license)

• good docs

• good platform / arch support

• xattr / acl support

• FUSE support ("mount a backup")

• Python 3.x + a little Cython & C

• good test coverage, CI

Deduplication (1)

• No problem with:
  • VM images (sparse file support)
  • (physical) disk images
  • renamed huge directories/trees
  • inner deduplication of data set
  • historical deduplication
  • deduplication different machines

Deduplication (2)

• Content defined chunking:
  • "buzhash" rolling hash
  • cut data when hash has specific bit pattern,
    yields chunks with 2^n bits target size
  • n + other chunker params configurable now
  • seeded, to avoid fingerprinting chunk lengths
     
• Store chunks under id into store:
  • id = HASH(chunk)
  • id = MAC(mac_key, chunk)

Borg Backup - Links

borgbackup.github.io
 

#borgbackup on chat.freenode.net

Questions / Feedback?

• Just grab me at the conference!

• Thomas J Waldmann @ twitter