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 3.7 or newer, 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 WSL). Patches to support
81 other platforms are welcome.
83 - Until resolved, a [glibc bug](https://sourceware.org/bugzilla/show_bug.cgi?id=26034)
84 might cause bup to crash on startup for some (unusual) command line
85 argument values, when bup is configured to use Python 3.
87 - Any items in "Things that are stupid" below.
90 Notable changes introduced by a release
91 =======================================
93 - <a href="note/0.32-from-0.31.md">Changes in 0.32 as compared to 0.31</a>
94 - <a href="note/0.31-from-0.30.1.md">Changes in 0.31 as compared to 0.30.1</a>
95 - <a href="note/0.30.1-from-0.30.md">Changes in 0.30.1 as compared to 0.30</a>
96 - <a href="note/0.30-from-0.29.3.md">Changes in 0.30 as compared to 0.29.3</a>
97 - <a href="note/0.29.3-from-0.29.2.md">Changes in 0.29.3 as compared to 0.29.2</a>
98 - <a href="note/0.29.2-from-0.29.1.md">Changes in 0.29.2 as compared to 0.29.1</a>
99 - <a href="note/0.29.1-from-0.29.md">Changes in 0.29.1 as compared to 0.29</a>
100 - <a href="note/0.29-from-0.28.1.md">Changes in 0.29 as compared to 0.28.1</a>
101 - <a href="note/0.28.1-from-0.28.md">Changes in 0.28.1 as compared to 0.28</a>
102 - <a href="note/0.28-from-0.27.1.md">Changes in 0.28 as compared to 0.27.1</a>
103 - <a href="note/0.27.1-from-0.27.md">Changes in 0.27.1 as compared to 0.27</a>
109 | master | 0.30.x | 0.29.x |
110 |--------|--------|--------|
111 | [![master branch test status](https://api.cirrus-ci.com/github/bup/bup.svg?branch=master)](https://cirrus-ci.com/github/bup/bup) | [![0.30 branch test status](https://api.cirrus-ci.com/github/bup/bup.svg?branch=0.30.x)](https://cirrus-ci.com/github/bup/bup) | [![0.29 branch test status](https://api.cirrus-ci.com/github/bup/bup.svg?branch=0.29.x)](https://cirrus-ci.com/github/bup/bup) |
119 - Check out the bup source code using git:
122 git clone https://github.com/bup/bup
125 - This will leave you on the master branch, which is perfect if you
126 would like to help with development, but if you'd just like to use
127 bup, please check out the latest stable release like this:
133 You can see the latest stable release here:
134 https://github.com/bup/bup/releases.
136 - Install the required python libraries (including the development
139 For `bup fuse` you will need to install
140 [python-fuse](https://github.com/libfuse/python-fuse) rather than
141 [fusepy](https://github.com/fusepy/fusepy). For example, in
142 Debian, install python3-fuse rather than python3-fusepy.
144 On very recent Debian/Ubuntu versions, this may be sufficient (run
148 apt-get build-dep bup
154 apt-get install python3.7-dev python3-fuse
155 apt-get install python3-pyxattr python3-pytest
156 apt-get install python3-distutils
157 apt-get install pkg-config linux-libc-dev libacl1-dev
158 apt-get install gcc make acl attr rsync
159 apt-get isntall python3-pytest-xdist # optional (parallel tests)
160 apt-get install par2 # optional (error correction)
161 apt-get install libreadline-dev # optional (bup ftp)
162 apt-get install python3-tornado # optional (bup web)
166 On Cygwin, install python, make, rsync, and gcc4.
168 If you would like to use the optional bup web server on systems
169 without a tornado package, you may want to try this:
187 or if you're in a bit more of a hurry:
193 If you have the Python xdist module installed, then you can
194 probably run the tests faster by adding the make -j option (see <a
195 href="HACKING">./HACKING</a> for additional information):
201 The tests should pass. If they don't pass for you, stop here and
202 send an email to bup-list@googlegroups.com. Though if there are
203 symbolic links along the current working directory path, the tests
204 may fail. Running something like this before "make test" should
205 sidestep the problem:
211 - You can install bup via "make install", and override the default
212 destination with DESTDIR and PREFIX.
214 Files are normally installed to "$DESTDIR/$PREFIX" where DESTDIR is
215 empty by default, and PREFIX is set to /usr/local. So if you wanted to
216 install bup to /opt/bup, you might do something like this:
219 make install DESTDIR=/opt/bup PREFIX=''
222 - The Python version that bup will use is determined by the
223 `python-config` program chosen by `./configure`, which will search
224 for a reasonable version unless `BUP_PYTHON_CONFIG` is set in the
225 environment. You can see which Python executable was chosen by
226 looking at the configure output, or examining
227 `config/config.var/bup-python-config`, and you can change the
228 selection by re-running `./configure`.
230 - If you want to specify your own `CPPFLAGS`, `CFLAGS`, or `LDFLAGS`,
231 you can set them for individual `make` invocations, e.g. `make
232 CFLAGS=-O0 check`, or persistently via `./configure` with
233 `CFLAGS=-O0 ./configure`. At the moment, `make clean` clears the
234 configuration, but we may change that at some point, perhaps by
235 adding and requiring a `make distclean` to clear the configuration.
240 Binary packages of bup are known to be built for the following OSes:
243 http://packages.debian.org/search?searchon=names&keywords=bup
245 http://packages.ubuntu.com/search?searchon=names&keywords=bup
246 - pkgsrc (NetBSD, Dragonfly, and others)
247 http://pkgsrc.se/sysutils/bup
248 http://cvsweb.netbsd.org/bsdweb.cgi/pkgsrc/sysutils/bup/
250 https://www.archlinux.org/packages/?sort=&q=bup
252 https://apps.fedoraproject.org/packages/bup
254 https://formulae.brew.sh/formula/bup
260 - Get help for any bup command:
271 - Initialize the default BUP_DIR (~/.bup -- you can choose another by
272 either specifying `bup -d DIR ...` or setting the `BUP_DIR`
273 environment variable for a command):
279 - Make a local backup (-v or -vv will increase the verbosity):
283 bup save -n local-etc /etc
286 - Restore a local backup to ./dest:
289 bup restore -C ./dest local-etc/latest/etc
293 - Look at how much disk space your backup took:
299 - Make another backup (which should be mostly identical to the last one;
300 notice that you don't have to *specify* that this backup is incremental,
301 it just saves space automatically):
305 bup save -n local-etc /etc
308 - Look how little extra space your second backup used (on top of the first):
314 - Get a list of your previous backups:
320 - Restore your first backup again:
323 bup restore -C ./dest-2 local-etc/2013-11-23-11195/etc
326 - Make a backup to a remote server which must already have the 'bup' command
327 somewhere in its PATH (see /etc/profile, etc/environment, ~/.profile, or
328 ~/.bashrc), and be accessible via ssh.
329 Make sure to replace SERVERNAME with the actual hostname of your server:
332 bup init -r SERVERNAME:path/to/remote-bup-dir
334 bup save -r SERVERNAME:path/to/remote-bup-dir -n local-etc /etc
337 - Make a remote backup to ~/.bup on SERVER:
341 bup save -r SERVER: -n local-etc /etc
344 - See what saves are available in ~/.bup on SERVER:
350 - Restore the remote backup to ./dest:
353 bup restore -r SERVER: -C ./dest local-etc/latest/etc
357 - Defend your backups from death rays (OK fine, more likely from the
358 occasional bad disk block). This writes parity information
359 (currently via par2) for all of the existing data so that bup may
360 be able to recover from some amount of repository corruption:
366 - Use split/join instead of index/save/restore. Try making a local
370 tar -cvf - /etc | bup split -n local-etc -vv
373 - Try restoring the tarball:
376 bup join local-etc | tar -tf -
379 - Look at how much disk space your backup took:
385 - Make another tar backup:
388 tar -cvf - /etc | bup split -n local-etc -vv
391 - Look at how little extra space your second backup used on top of
398 - Restore the first tar backup again (the ~1 is git notation for "one
399 older than the most recent"):
402 bup join local-etc~1 | tar -tf -
405 - Get a list of your previous split-based backups:
408 GIT_DIR=~/.bup git log local-etc
411 - Save a tar archive to a remote server (without tar -z to facilitate
415 tar -cvf - /etc | bup split -r SERVERNAME: -n local-etc -vv
418 - Restore the archive:
421 bup join -r SERVERNAME: local-etc | tar -tf -
424 That's all there is to it!
430 - FreeBSD's default 'make' command doesn't like bup's Makefile. In order to
431 compile the code, run tests and install bup, you need to install GNU Make
432 from the port named 'gmake' and use its executable instead in the commands
433 seen above. (i.e. 'gmake test' runs bup's test suite)
435 - Python's development headers are automatically installed with the 'python'
436 port so there's no need to install them separately.
438 - To use the 'bup fuse' command, you need to install the fuse kernel module
439 from the 'fusefs-kmod' port in the 'sysutils' section and the libraries from
440 the port named 'py-fusefs' in the 'devel' section.
442 - The 'par2' command can be found in the port named 'par2cmdline'.
444 - In order to compile the documentation, you need pandoc which can be found in
445 the port named 'hs-pandoc' in the 'textproc' section.
448 Notes on NetBSD/pkgsrc
449 ----------------------
451 - See pkgsrc/sysutils/bup, which should be the most recent stable
452 release and includes man pages. It also has a reasonable set of
453 dependencies (git, par2, py-fuse-bindings).
455 - The "fuse-python" package referred to is hard to locate, and is a
456 separate tarball for the python language binding distributed by the
457 fuse project on sourceforge. It is available as
458 pkgsrc/filesystems/py-fuse-bindings and on NetBSD 5, "bup fuse"
461 - "bup fuse" presents every directory/file as inode 0. The directory
462 traversal code ("fts") in NetBSD's libc will interpret this as a
463 cycle and error out, so "ls -R" and "find" will not work.
465 - There is no support for ACLs. If/when some enterprising person
466 fixes this, adjust dev/compare-trees.
472 - There is no support for ACLs. If/when some enterprising person
473 fixes this, adjust dev/compare-trees.
475 - In test/ext/test-misc, two tests have been disabled. These tests
476 check to see that repeated saves produce identical trees and that
477 an intervening index doesn't change the SHA1. Apparently Cygwin
478 has some unusual behaviors with respect to access times (that
479 probably warrant further investigation). Possibly related:
480 http://cygwin.com/ml/cygwin/2007-06/msg00436.html
486 - There is no support for ACLs. If/when some enterprising person
487 fixes this, adjust dev/compare-trees.
496 bup stores its data in a git-formatted repository. Unfortunately, git
497 itself doesn't actually behave very well for bup's use case (huge numbers of
498 files, files with huge sizes, retaining file permissions/ownership are
499 important), so we mostly don't use git's *code* except for a few helper
500 programs. For example, bup has its own git packfile writer written in
503 Basically, 'bup split' reads the data on stdin (or from files specified on
504 the command line), breaks it into chunks using a rolling checksum (similar to
505 rsync), and saves those chunks into a new git packfile. There is at least one
506 git packfile per backup.
508 When deciding whether to write a particular chunk into the new packfile, bup
509 first checks all the other packfiles that exist to see if they already have that
510 chunk. If they do, the chunk is skipped.
512 git packs come in two parts: the pack itself (*.pack) and the index (*.idx).
513 The index is pretty small, and contains a list of all the objects in the
514 pack. Thus, when generating a remote backup, we don't have to have a copy
515 of the packfiles from the remote server: the local end just downloads a copy
516 of the server's *index* files, and compares objects against those when
517 generating the new pack, which it sends directly to the server.
519 The "-n" option to 'bup split' and 'bup save' is the name of the backup you
520 want to create, but it's actually implemented as a git branch. So you can
521 do cute things like checkout a particular branch using git, and receive a
522 bunch of chunk files corresponding to the file you split.
524 If you use '-b' or '-t' or '-c' instead of '-n', bup split will output a
525 list of blobs, a tree containing that list of blobs, or a commit containing
526 that tree, respectively, to stdout. You can use this to construct your own
527 scripts that do something with those values.
532 'bup index' walks through your filesystem and updates a file (whose name is,
533 by default, ~/.bup/bupindex) to contain the name, attributes, and an
534 optional git SHA1 (blob id) of each file and directory.
536 'bup save' basically just runs the equivalent of 'bup split' a whole bunch
537 of times, once per file in the index, and assembles a git tree
538 that contains all the resulting objects. Among other things, that makes
539 'git diff' much more useful (compared to splitting a tarball, which is
540 essentially a big binary blob). However, since bup splits large files into
541 smaller chunks, the resulting tree structure doesn't *exactly* correspond to
542 what git itself would have stored. Also, the tree format used by 'bup save'
543 will probably change in the future to support storing file ownership, more
544 complex file permissions, and so on.
546 If a file has previously been written by 'bup save', then its git blob/tree
547 id is stored in the index. This lets 'bup save' avoid reading that file to
548 produce future incremental backups, which means it can go *very* fast unless
549 a lot of files have changed.
552 Things that are stupid for now but which we'll fix later
553 ========================================================
555 Help with any of these problems, or others, is very welcome. Join the
556 mailing list (see below) if you'd like to help.
558 - 'bup save' and 'bup restore' have immature metadata support.
560 On the plus side, they actually do have support now, but it's new,
561 and not remotely as well tested as tar/rsync/whatever's. However,
562 you have to start somewhere, and as of 0.25, we think it's ready
563 for more general use. Please let us know if you have any trouble.
565 Also, if any strip or graft-style options are specified to 'bup
566 save', then no metadata will be written for the root directory.
567 That's obviously less than ideal.
569 - bup is overly optimistic about mmap. Right now bup just assumes
570 that it can mmap as large a block as it likes, and that mmap will
571 never fail. Yeah, right... If nothing else, this has failed on
572 32-bit architectures (and 31-bit is even worse -- looking at you,
575 To fix this, we might just implement a FakeMmap[1] class that uses
576 normal file IO and handles all of the mmap methods[2] that bup
577 actually calls. Then we'd swap in one of those whenever mmap
580 This would also require implementing some of the methods needed to
581 support "[]" array access, probably at a minimum __getitem__,
582 __setitem__, and __setslice__ [3].
584 [1] http://comments.gmane.org/gmane.comp.sysutils.backup.bup/613
585 [2] http://docs.python.org/3/library/mmap.html
586 [3] http://docs.python.org/3/reference/datamodel.html#emulating-container-types
588 - 'bup index' is slower than it should be.
590 It's still rather fast: it can iterate through all the filenames on my
591 600,000 file filesystem in a few seconds. But it still needs to rewrite
592 the entire index file just to add a single filename, which is pretty
593 nasty; it should just leave the new files in a second "extra index" file
596 - bup could use inotify for *really* efficient incremental backups.
598 You could even have your system doing "continuous" backups: whenever a
599 file changes, we immediately send an image of it to the server. We could
600 give the continuous-backup process a really low CPU and I/O priority so
601 you wouldn't even know it was running.
603 - bup only has experimental support for pruning old backups.
605 While you should now be able to drop old saves and branches with
606 `bup rm`, and reclaim the space occupied by data that's no longer
607 needed by other backups with `bup gc`, these commands are
608 experimental, and should be handled with great care. See the
609 man pages for more information.
611 Unless you want to help test the new commands, one possible
612 workaround is to just start a new BUP_DIR occasionally,
613 i.e. bup-2013, bup-2014...
615 - bup has never been tested on anything but Linux, FreeBSD, NetBSD,
616 OS X, and Windows+Cygwin.
618 There's nothing that makes it *inherently* non-portable, though, so
619 that's mostly a matter of someone putting in some effort. (For a
620 "native" Windows port, the most annoying thing is the absence of ssh in
621 a default Windows installation.)
623 - bup needs better documentation.
625 According to an article about bup in Linux Weekly News
626 (https://lwn.net/Articles/380983/), "it's a bit short on examples and
627 a user guide would be nice." Documentation is the sort of thing that
628 will never be great unless someone from outside contributes it (since
629 the developers can never remember which parts are hard to understand).
631 - bup is "relatively speedy" and has "pretty good" compression.
633 ...according to the same LWN article. Clearly neither of those is good
634 enough. We should have awe-inspiring speed and crazy-good compression.
635 Must work on that. Writing more parts in C might help with the speed.
639 Actually, that's not stupid, but you might consider it a
640 limitation. See the ["Related Projects"](https://bup.github.io/)
641 list for some possible options.
646 bup has an extensive set of man pages. Try using 'bup help' to get
647 started, or use 'bup help SUBCOMMAND' for any bup subcommand (like split,
648 join, index, save, etc.) to get details on that command.
650 For further technical details, please see ./DESIGN.
656 bup is a work in progress and there are many ways it can still be improved.
657 If you'd like to contribute patches, ideas, or bug reports, please join the
658 <a href="mailto:bup-list@googlegroups.com">bup mailing list</a>:
660 You can find the mailing list archives here:
662 http://groups.google.com/group/bup-list
664 and you can subscribe by sending a message to:
666 bup-list+subscribe@googlegroups.com
668 You can also reach us via the
669 \#bup IRC channel at ircs://irc.libera.chat:6697/bup
670 on the [libera.chat](https://libera.chat/) network or via this
671 [web interface](https://web.libera.chat/?channels=bup).
673 Please see <a href="HACKING">./HACKING</a> for
674 additional information, i.e. how to submit patches (hint - no pull
675 requests), how we handle branches, etc.