1 bup: It backs things up
2 =======================
4 bup is a program that backs things up. It's short for "backup." Can you
5 believe that nobody else has named an open source program "bup" after all
8 Despite its unassuming name, bup is pretty cool. To give you an idea of
9 just how cool it is, I wrote you this poem:
12 What rhymes with awesome?
14 But that's irrelevant.
16 Hmm. Did that help? Maybe prose is more useful after all.
19 Reasons bup is awesome
20 ----------------------
22 bup has a few advantages over other backup software:
24 - It uses a rolling checksum algorithm (similar to rsync) to split large
25 files into chunks. The most useful result of this is you can backup huge
26 virtual machine (VM) disk images, databases, and XML files incrementally,
27 even though they're typically all in one huge file, and not use tons of
28 disk space for multiple versions.
30 - It uses the packfile format from git (the open source version control
31 system), so you can access the stored data even if you don't like bup's
34 - Unlike git, it writes packfiles *directly* (instead of having a separate
35 garbage collection / repacking stage) so it's fast even with gratuitously
36 huge amounts of data. bup's improved index formats also allow you to
37 track far more filenames than git (millions) and keep track of far more
38 objects (hundreds or thousands of gigabytes).
40 - Data is "automagically" shared between incremental backups without having
41 to know which backup is based on which other one - even if the backups
42 are made from two different computers that don't even know about each
43 other. You just tell bup to back stuff up, and it saves only the minimum
44 amount of data needed.
46 - You can back up directly to a remote bup server, without needing tons of
47 temporary disk space on the computer being backed up. And if your backup
48 is interrupted halfway through, the next run will pick up where you left
49 off. And it's easy to set up a bup server: just install bup on any
50 machine where you have ssh access.
52 - Bup can use "par2" redundancy to recover corrupted backups even if your
53 disk has undetected bad sectors.
55 - Even when a backup is incremental, you don't have to worry about
56 restoring the full backup, then each of the incrementals in turn; an
57 incremental backup *acts* as if it's a full backup, it just takes less
60 - You can mount your bup repository as a FUSE filesystem and access the
61 content that way, and even export it over Samba.
63 - It's written in python (with some C parts to make it faster) so it's easy
64 for you to extend and maintain.
67 Reasons you might want to avoid bup
68 -----------------------------------
70 - It's not remotely as well tested as something like tar, so it's
71 more likely to eat your data. It's also missing some
72 probably-critical features, though fewer than it used to be.
74 - It requires python >= 2.6, a C compiler, and an installed git
75 version >= 1.5.6. It also requires par2 if you want fsck to be
76 able to generate the information needed to recover from some types
79 - It currently only works on Linux, FreeBSD, NetBSD, OS X >= 10.4,
80 Solaris, or Windows (with Cygwin, and maybe with WSL). Patches to
81 support other platforms are welcome.
83 - Any items in "Things that are stupid" below.
86 Notable changes introduced by a release
87 =======================================
89 - <a href="note/0.30.1-from-0.30.md">Changes in 0.30.1 as compared to 0.30</a>
90 - <a href="note/0.30-from-0.29.3.md">Changes in 0.30 as compared to 0.29.3</a>
91 - <a href="note/0.29.3-from-0.29.2.md">Changes in 0.29.3 as compared to 0.29.2</a>
92 - <a href="note/0.29.2-from-0.29.1.md">Changes in 0.29.2 as compared to 0.29.1</a>
93 - <a href="note/0.29.1-from-0.29.md">Changes in 0.29.1 as compared to 0.29</a>
94 - <a href="note/0.29-from-0.28.1.md">Changes in 0.29 as compared to 0.28.1</a>
95 - <a href="note/0.28.1-from-0.28.md">Changes in 0.28.1 as compared to 0.28</a>
96 - <a href="note/0.28-from-0.27.1.md">Changes in 0.28 as compared to 0.27.1</a>
97 - <a href="note/0.27.1-from-0.27.md">Changes in 0.27.1 as compared to 0.27</a>
103 | branch | Debian | FreeBSD | macOS |
104 |--------|------------------------------------------------------------------------------------------------------------------------------------------------|--------------------------------------------------------------------------------------------------------------------------------------------------|----------------------------------------------------------------------------------------------------------------------------------------------|
105 | master | [![Debian test status](https://api.cirrus-ci.com/github/bup/bup.svg?branch=master&task=debian)](https://cirrus-ci.com/github/bup/bup) | [![FreeBSD test status](https://api.cirrus-ci.com/github/bup/bup.svg?branch=master&task=freebsd)](https://cirrus-ci.com/github/bup/bup) | [![macOS test status](https://api.cirrus-ci.com/github/bup/bup.svg?branch=master&task=macos)](https://cirrus-ci.com/github/bup/bup) |
106 | 0.30.x | [![Debian test status](https://api.cirrus-ci.com/github/bup/bup.svg?branch=0.30.x&task=debian)](https://cirrus-ci.com/github/bup/bup) | [![FreeBSD test status](https://api.cirrus-ci.com/github/bup/bup.svg?branch=0.30.x&task=freebsd)](https://cirrus-ci.com/github/bup/bup) | [![macOS test status](https://api.cirrus-ci.com/github/bup/bup.svg?branch=0.30.x&task=macos)](https://cirrus-ci.com/github/bup/bup) |
107 | 0.29.x | [![Debian test status](https://api.cirrus-ci.com/github/bup/bup.svg?branch=0.29.x&task=debian)](https://cirrus-ci.com/github/bup/bup) | [![FreeBSD test status](https://api.cirrus-ci.com/github/bup/bup.svg?branch=0.29.x&task=freebsd)](https://cirrus-ci.com/github/bup/bup) | [![macOS test status](https://api.cirrus-ci.com/github/bup/bup.svg?branch=0.29.x&task=macos)](https://cirrus-ci.com/github/bup/bup) |
115 - Check out the bup source code using git:
118 git clone https://github.com/bup/bup
121 - This will leave you on the master branch, which is perfect if you
122 would like to help with development, but if you'd just like to use
123 bup, please check out the latest stable release like this:
129 You can see the latest stable release here:
130 https://github.com/bup/bup/releases.
132 - Install the required python libraries (including the development
135 On very recent Debian/Ubuntu versions, this may be sufficient (run
139 apt-get build-dep bup
142 Otherwise try this (substitute python2.6-dev if you have an older
146 apt-get install python2.7-dev python-fuse
147 apt-get install python-pyxattr python-pylibacl
148 apt-get install linux-libc-dev
149 apt-get install acl attr
150 apt-get install libreadline-dev # optional (bup ftp)
151 apt-get install python-tornado # optional (bup web)
154 On CentOS (for CentOS 6, at least), this should be sufficient (run
158 yum groupinstall "Development Tools"
159 yum install python python-devel
160 yum install fuse-python pyxattr pylibacl
161 yum install perl-Time-HiRes
162 yum install readline-devel # optional (bup ftp)
163 yum install python-tornado # optional (bup web)
166 In addition to the default CentOS repositories, you may need to add
167 RPMForge (for fuse-python) and EPEL (for pyxattr and pylibacl).
169 On Cygwin, install python, make, rsync, and gcc4.
171 If you would like to use the optional bup web server on systems
172 without a tornado package, you may want to try this:
178 - Build the python module and symlinks:
190 or if you're in a bit more of a hurry:
196 The tests should pass. If they don't pass for you, stop here and
197 send an email to bup-list@googlegroups.com. Though if there are
198 symbolic links along the current working directory path, the tests
199 may fail. Running something like this before "make test" should
200 sidestep the problem:
206 - You can install bup via "make install", and override the default
207 destination with DESTDIR and PREFIX.
209 Files are normally installed to "$DESTDIR/$PREFIX" where DESTDIR is
210 empty by default, and PREFIX is set to /usr/local. So if you wanted to
211 install bup to /opt/bup, you might do something like this:
214 make install DESTDIR=/opt/bup PREFIX=''
217 - The Python executable that bup will use is chosen by ./configure,
218 which will search for a reasonable version unless PYTHON is set in
219 the environment, in which case, bup will use that path. You can
220 see which Python executable was chosen by looking at the
221 configure output, or examining cmd/python-cmd.sh, and you can
222 change the selection by re-running ./configure.
227 Binary packages of bup are known to be built for the following OSes:
230 http://packages.debian.org/search?searchon=names&keywords=bup
232 http://packages.ubuntu.com/search?searchon=names&keywords=bup
233 - pkgsrc (NetBSD, Dragonfly, and others)
234 http://pkgsrc.se/sysutils/bup
235 http://cvsweb.netbsd.org/bsdweb.cgi/pkgsrc/sysutils/bup/
237 https://www.archlinux.org/packages/?sort=&q=bup
239 https://apps.fedoraproject.org/packages/bup
245 - Get help for any bup command:
256 - Initialize the default BUP_DIR (~/.bup -- you can choose another by
257 either specifying `bup -d DIR ...` or setting the `BUP_DIR`
258 environment variable for a command):
264 - Make a local backup (-v or -vv will increase the verbosity):
268 bup save -n local-etc /etc
271 - Restore a local backup to ./dest:
274 bup restore -C ./dest local-etc/latest/etc
278 - Look at how much disk space your backup took:
284 - Make another backup (which should be mostly identical to the last one;
285 notice that you don't have to *specify* that this backup is incremental,
286 it just saves space automatically):
290 bup save -n local-etc /etc
293 - Look how little extra space your second backup used (on top of the first):
299 - Get a list of your previous backups:
305 - Restore your first backup again:
308 bup restore -C ./dest-2 local-etc/2013-11-23-11195/etc
311 - Make a backup to a remote server which must already have the 'bup' command
312 somewhere in its PATH (see /etc/profile, etc/environment, ~/.profile, or
313 ~/.bashrc), and be accessible via ssh.
314 Make sure to replace SERVERNAME with the actual hostname of your server:
317 bup init -r SERVERNAME:path/to/remote-bup-dir
319 bup save -r SERVERNAME:path/to/remote-bup-dir -n local-etc /etc
322 - Make a remote backup to ~/.bup on SERVER:
326 bup save -r SERVER: -n local-etc /etc
329 - See what saves are available in ~/.bup on SERVER:
335 - Restore the remote backup to ./dest:
338 bup restore -r SERVER: -C ./dest local-etc/latest/etc
342 - Defend your backups from death rays (OK fine, more likely from the
343 occasional bad disk block). This writes parity information
344 (currently via par2) for all of the existing data so that bup may
345 be able to recover from some amount of repository corruption:
351 - Use split/join instead of index/save/restore. Try making a local
355 tar -cvf - /etc | bup split -n local-etc -vv
358 - Try restoring the tarball:
361 bup join local-etc | tar -tf -
364 - Look at how much disk space your backup took:
370 - Make another tar backup:
373 tar -cvf - /etc | bup split -n local-etc -vv
376 - Look at how little extra space your second backup used on top of
383 - Restore the first tar backup again (the ~1 is git notation for "one
384 older than the most recent"):
387 bup join local-etc~1 | tar -tf -
390 - Get a list of your previous split-based backups:
393 GIT_DIR=~/.bup git log local-etc
396 - Save a tar archive to a remote server (without tar -z to facilitate
400 tar -cvf - /etc | bup split -r SERVERNAME: -n local-etc -vv
403 - Restore the archive:
406 bup join -r SERVERNAME: local-etc | tar -tf -
409 That's all there is to it!
415 - FreeBSD's default 'make' command doesn't like bup's Makefile. In order to
416 compile the code, run tests and install bup, you need to install GNU Make
417 from the port named 'gmake' and use its executable instead in the commands
418 seen above. (i.e. 'gmake test' runs bup's test suite)
420 - Python's development headers are automatically installed with the 'python'
421 port so there's no need to install them separately.
423 - To use the 'bup fuse' command, you need to install the fuse kernel module
424 from the 'fusefs-kmod' port in the 'sysutils' section and the libraries from
425 the port named 'py-fusefs' in the 'devel' section.
427 - The 'par2' command can be found in the port named 'par2cmdline'.
429 - In order to compile the documentation, you need pandoc which can be found in
430 the port named 'hs-pandoc' in the 'textproc' section.
433 Notes on NetBSD/pkgsrc
434 ----------------------
436 - See pkgsrc/sysutils/bup, which should be the most recent stable
437 release and includes man pages. It also has a reasonable set of
438 dependencies (git, par2, py-fuse-bindings).
440 - The "fuse-python" package referred to is hard to locate, and is a
441 separate tarball for the python language binding distributed by the
442 fuse project on sourceforge. It is available as
443 pkgsrc/filesystems/py-fuse-bindings and on NetBSD 5, "bup fuse"
446 - "bup fuse" presents every directory/file as inode 0. The directory
447 traversal code ("fts") in NetBSD's libc will interpret this as a
448 cycle and error out, so "ls -R" and "find" will not work.
450 - There is no support for ACLs. If/when some enterprising person
451 fixes this, adjust t/compare-trees.
457 - There is no support for ACLs. If/when some enterprising person
458 fixes this, adjust t/compare-trees.
460 - In t/test.sh, two tests have been disabled. These tests check to
461 see that repeated saves produce identical trees and that an
462 intervening index doesn't change the SHA1. Apparently Cygwin has
463 some unusual behaviors with respect to access times (that probably
464 warrant further investigation). Possibly related:
465 http://cygwin.com/ml/cygwin/2007-06/msg00436.html
471 - There is no support for ACLs. If/when some enterprising person
472 fixes this, adjust t/compare-trees.
481 bup stores its data in a git-formatted repository. Unfortunately, git
482 itself doesn't actually behave very well for bup's use case (huge numbers of
483 files, files with huge sizes, retaining file permissions/ownership are
484 important), so we mostly don't use git's *code* except for a few helper
485 programs. For example, bup has its own git packfile writer written in
488 Basically, 'bup split' reads the data on stdin (or from files specified on
489 the command line), breaks it into chunks using a rolling checksum (similar to
490 rsync), and saves those chunks into a new git packfile. There is at least one
491 git packfile per backup.
493 When deciding whether to write a particular chunk into the new packfile, bup
494 first checks all the other packfiles that exist to see if they already have that
495 chunk. If they do, the chunk is skipped.
497 git packs come in two parts: the pack itself (*.pack) and the index (*.idx).
498 The index is pretty small, and contains a list of all the objects in the
499 pack. Thus, when generating a remote backup, we don't have to have a copy
500 of the packfiles from the remote server: the local end just downloads a copy
501 of the server's *index* files, and compares objects against those when
502 generating the new pack, which it sends directly to the server.
504 The "-n" option to 'bup split' and 'bup save' is the name of the backup you
505 want to create, but it's actually implemented as a git branch. So you can
506 do cute things like checkout a particular branch using git, and receive a
507 bunch of chunk files corresponding to the file you split.
509 If you use '-b' or '-t' or '-c' instead of '-n', bup split will output a
510 list of blobs, a tree containing that list of blobs, or a commit containing
511 that tree, respectively, to stdout. You can use this to construct your own
512 scripts that do something with those values.
517 'bup index' walks through your filesystem and updates a file (whose name is,
518 by default, ~/.bup/bupindex) to contain the name, attributes, and an
519 optional git SHA1 (blob id) of each file and directory.
521 'bup save' basically just runs the equivalent of 'bup split' a whole bunch
522 of times, once per file in the index, and assembles a git tree
523 that contains all the resulting objects. Among other things, that makes
524 'git diff' much more useful (compared to splitting a tarball, which is
525 essentially a big binary blob). However, since bup splits large files into
526 smaller chunks, the resulting tree structure doesn't *exactly* correspond to
527 what git itself would have stored. Also, the tree format used by 'bup save'
528 will probably change in the future to support storing file ownership, more
529 complex file permissions, and so on.
531 If a file has previously been written by 'bup save', then its git blob/tree
532 id is stored in the index. This lets 'bup save' avoid reading that file to
533 produce future incremental backups, which means it can go *very* fast unless
534 a lot of files have changed.
537 Things that are stupid for now but which we'll fix later
538 ========================================================
540 Help with any of these problems, or others, is very welcome. Join the
541 mailing list (see below) if you'd like to help.
543 - 'bup save' and 'bup restore' have immature metadata support.
545 On the plus side, they actually do have support now, but it's new,
546 and not remotely as well tested as tar/rsync/whatever's. However,
547 you have to start somewhere, and as of 0.25, we think it's ready
548 for more general use. Please let us know if you have any trouble.
550 Also, if any strip or graft-style options are specified to 'bup
551 save', then no metadata will be written for the root directory.
552 That's obviously less than ideal.
554 - bup is overly optimistic about mmap. Right now bup just assumes
555 that it can mmap as large a block as it likes, and that mmap will
556 never fail. Yeah, right... If nothing else, this has failed on
557 32-bit architectures (and 31-bit is even worse -- looking at you,
560 To fix this, we might just implement a FakeMmap[1] class that uses
561 normal file IO and handles all of the mmap methods[2] that bup
562 actually calls. Then we'd swap in one of those whenever mmap
565 This would also require implementing some of the methods needed to
566 support "[]" array access, probably at a minimum __getitem__,
567 __setitem__, and __setslice__ [3].
569 [1] http://comments.gmane.org/gmane.comp.sysutils.backup.bup/613
570 [2] http://docs.python.org/2/library/mmap.html
571 [3] http://docs.python.org/2/reference/datamodel.html#emulating-container-types
573 - 'bup index' is slower than it should be.
575 It's still rather fast: it can iterate through all the filenames on my
576 600,000 file filesystem in a few seconds. But it still needs to rewrite
577 the entire index file just to add a single filename, which is pretty
578 nasty; it should just leave the new files in a second "extra index" file
581 - bup could use inotify for *really* efficient incremental backups.
583 You could even have your system doing "continuous" backups: whenever a
584 file changes, we immediately send an image of it to the server. We could
585 give the continuous-backup process a really low CPU and I/O priority so
586 you wouldn't even know it was running.
588 - bup only has experimental support for pruning old backups.
590 While you should now be able to drop old saves and branches with
591 `bup rm`, and reclaim the space occupied by data that's no longer
592 needed by other backups with `bup gc`, these commands are
593 experimental, and should be handled with great care. See the
594 man pages for more information.
596 Unless you want to help test the new commands, one possible
597 workaround is to just start a new BUP_DIR occasionally,
598 i.e. bup-2013, bup-2014...
600 - bup has never been tested on anything but Linux, FreeBSD, NetBSD,
601 OS X, and Windows+Cygwin.
603 There's nothing that makes it *inherently* non-portable, though, so
604 that's mostly a matter of someone putting in some effort. (For a
605 "native" Windows port, the most annoying thing is the absence of ssh in
606 a default Windows installation.)
608 - bup needs better documentation.
610 According to an article about bup in Linux Weekly News
611 (https://lwn.net/Articles/380983/), "it's a bit short on examples and
612 a user guide would be nice." Documentation is the sort of thing that
613 will never be great unless someone from outside contributes it (since
614 the developers can never remember which parts are hard to understand).
616 - bup is "relatively speedy" and has "pretty good" compression.
618 ...according to the same LWN article. Clearly neither of those is good
619 enough. We should have awe-inspiring speed and crazy-good compression.
620 Must work on that. Writing more parts in C might help with the speed.
624 Actually, that's not stupid, but you might consider it a
625 limitation. See the ["Related Projects"](https://bup.github.io/)
626 list for some possible options.
631 bup has an extensive set of man pages. Try using 'bup help' to get
632 started, or use 'bup help SUBCOMMAND' for any bup subcommand (like split,
633 join, index, save, etc.) to get details on that command.
635 For further technical details, please see ./DESIGN.
641 bup is a work in progress and there are many ways it can still be improved.
642 If you'd like to contribute patches, ideas, or bug reports, please join the
645 You can find the mailing list archives here:
647 http://groups.google.com/group/bup-list
649 and you can subscribe by sending a message to:
651 bup-list+subscribe@googlegroups.com
653 Please see <a href="HACKING">./HACKING</a> for
654 additional information, i.e. how to submit patches (hint - no pull
655 requests), how we handle branches, etc.