bup
Git-based backup system with deduplication
TLDR
SYNOPSIS
bup command [options]
DESCRIPTION
bup is a backup system based on git's packfile format. It provides very efficient storage through deduplication, splitting large files into chunks, and using rolling checksums to find duplicate data.The tool is particularly effective for backing up large files with small changes.
PARAMETERS
init
Initialize bup repository. Uses ~/.bup by default or BUP_DIR if set.index path
Index files for backup. Must be run before save.save -n name path
Save indexed files as a named backup set.restore -C target backup-path
Restore files from a backup to the target directory.ls [backup]
List backups or files within a backup.fsck
Check repository integrity.fuse mountpoint
Mount backups as a read-only FUSE filesystem.damage
Deliberately damage a repository for testing fsck.margin
Report the maximum number of matching prefix bits between objects.midx
Create or display midx (multi-index) files.memtest
Test memory throughput.web
Start a web server to browse backups.
ENVIRONMENT
BUP_DIR
Path to the bup repository. Defaults to ~/.bup.BUP_FORCE_TTY
Force progress output even when not on a terminal.
WORKFLOW
export BUP_DIR=/backup/bup-repo
bup init
# Index files
bup index ~/Documents
# Save backup
bup save -n documents ~/Documents
# List backups
bup ls
# List files in backup
bup ls documents/latest/
# Restore
bup restore -C ~/restored documents/latest/
# Mount as filesystem
mkdir /mnt/bup
bup fuse /mnt/bup
# Remote backup over SSH
bup init -r myserver:
bup save -r myserver: -n documents ~/Documents
DEDUPLICATION
Uses rolling checksums to identify duplicate chunks even when shifted within files. Very efficient for:- Virtual machine images- Large archives- Database backups- Version control repositories
CAVEATS
Not encrypted by default. Repository can grow large without maintenance. Restore can be slow for many small files. Less mature than borg/restic. Documentation sometimes lacking. Not ideal for many tiny files. No built-in pruning of old backups.
HISTORY
bup was created by Avery Pennarun around 2010 to provide git-like backup with efficient deduplication for large files.
