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 - This is a very early version. Therefore it will most probably not work
71 for you, but we don't know why. It is also missing some
72 probably-critical features.
74 - It requires python >= 2.6, a C compiler, and an installed git
75 version >= 1.5.3.1. 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). Patches to support other
81 platforms are welcome.
83 - Any items in "Things that are stupid" below.
86 Notable changes introduced by a release
87 =======================================
89 - <a href="note/0.29.2-from-0.29.1.md">Changes in 0.29.2 as compared to 0.29.1</a>
90 - <a href="note/0.29.1-from-0.29.md">Changes in 0.29.1 as compared to 0.29</a>
91 - <a href="note/0.29-from-0.28.1.md">Changes in 0.29 as compared to 0.28.1</a>
92 - <a href="note/0.28.1-from-0.28.md">Changes in 0.28.1 as compared to 0.28</a>
93 - <a href="note/0.28-from-0.27.1.md">Changes in 0.28 as compared to 0.27.1</a>
94 - <a href="note/0.27.1-from-0.27.md">Changes in 0.27.1 as compared to 0.27</a>
97 Test status: [![Debian test status](https://api.cirrus-ci.com/github/bup/bup.svg?branch=tmp/test-cirrus&task=debian)](https://cirrus-ci.com/github/bup/bup)
98 [![FreeBSD test status](https://api.cirrus-ci.com/github/bup/bup.svg?branch=tmp/test-cirrus&task=freebsd)](https://cirrus-ci.com/github/bup/bup)
99 [![macOS test status](https://api.cirrus-ci.com/github/bup/bup.svg?branch=tmp/test-cirrus&task=macos)](https://cirrus-ci.com/github/bup/bup)
100 ==============================================================
110 - Check out the bup source code using git:
112 git clone https://github.com/bup/bup
114 - Install the required python libraries (including the development
117 On very recent Debian/Ubuntu versions, this may be sufficient (run
120 apt-get build-dep bup
122 Otherwise try this (substitute python2.6-dev if you have an older
125 apt-get install python2.7-dev python-fuse
126 apt-get install python-pyxattr python-pylibacl
127 apt-get install linux-libc-dev
128 apt-get install acl attr
129 apt-get install python-tornado # optional
131 On CentOS (for CentOS 6, at least), this should be sufficient (run
134 yum groupinstall "Development Tools"
135 yum install python python-devel
136 yum install fuse-python pyxattr pylibacl
137 yum install perl-Time-HiRes
139 In addition to the default CentOS repositories, you may need to add
140 RPMForge (for fuse-python) and EPEL (for pyxattr and pylibacl).
142 On Cygwin, install python, make, rsync, and gcc4.
144 If you would like to use the optional bup web server on systems
145 without a tornado package, you may want to try this:
149 - Build the python module and symlinks:
157 The tests should pass. If they don't pass for you, stop here and
158 send an email to bup-list@googlegroups.com. Though if there are
159 symbolic links along the current working directory path, the tests
160 may fail. Running something like this before "make test" should
161 sidestep the problem:
165 - You can install bup via "make install", and override the default
166 destination with DESTDIR and PREFIX.
168 Files are normally installed to "$DESTDIR/$PREFIX" where DESTDIR is
169 empty by default, and PREFIX is set to /usr/local. So if you wanted to
170 install bup to /opt/bup, you might do something like this:
172 make install DESTDIR=/opt/bup PREFIX=''
174 - The Python executable that bup will use is chosen by ./configure,
175 which will search for a reasonable version unless PYTHON is set in
176 the environment, in which case, bup will use that path. You can
177 see which Python executable was chosen by looking at the
178 configure output, or examining cmd/python-cmd.sh, and you can
179 change the selection by re-running ./configure.
184 Binary packages of bup are known to be built for the following OSes:
187 http://packages.debian.org/search?searchon=names&keywords=bup
189 http://packages.ubuntu.com/search?searchon=names&keywords=bup
190 - pkgsrc (NetBSD, Dragonfly, and others)
191 http://pkgsrc.se/sysutils/bup
192 http://cvsweb.netbsd.org/bsdweb.cgi/pkgsrc/sysutils/bup/
194 https://www.archlinux.org/packages/?sort=&q=bup
196 https://apps.fedoraproject.org/packages/bup
202 - Get help for any bup command:
211 - Initialize the default BUP_DIR (~/.bup):
215 - Make a local backup (-v or -vv will increase the verbosity):
218 bup save -n local-etc /etc
220 - Restore a local backup to ./dest:
222 bup restore -C ./dest local-etc/latest/etc
225 - Look at how much disk space your backup took:
229 - Make another backup (which should be mostly identical to the last one;
230 notice that you don't have to *specify* that this backup is incremental,
231 it just saves space automatically):
234 bup save -n local-etc /etc
236 - Look how little extra space your second backup used (on top of the first):
240 - Get a list of your previous backups:
244 - Restore your first backup again:
246 bup restore -C ./dest-2 local-etc/2013-11-23-11195/etc
248 - Make a backup to a remote server which must already have the 'bup' command
249 somewhere in its PATH (see /etc/profile, etc/environment, ~/.profile, or
250 ~/.bashrc), and be accessible via ssh.
251 Make sure to replace SERVERNAME with the actual hostname of your server:
253 bup init -r SERVERNAME:path/to/remote-bup-dir
255 bup save -r SERVERNAME:path/to/remote-bup-dir -n local-etc /etc
257 - Restore a backup from a remote server. (FAIL: unfortunately,
258 unlike "bup join", "bup restore" does not yet support remote
259 restores. See both "bup join" and "Things that are stupid" below.)
261 - Defend your backups from death rays (OK fine, more likely from the
262 occasional bad disk block). This writes parity information
263 (currently via par2) for all of the existing data so that bup may
264 be able to recover from some amount of repository corruption:
268 - Use split/join instead of index/save/restore. Try making a local
271 tar -cvf - /etc | bup split -n local-etc -vv
273 - Try restoring the tarball:
275 bup join local-etc | tar -tf -
277 - Look at how much disk space your backup took:
281 - Make another tar backup:
283 tar -cvf - /etc | bup split -n local-etc -vv
285 - Look at how little extra space your second backup used on top of
290 - Restore the first tar backup again (the ~1 is git notation for "one
291 older than the most recent"):
293 bup join local-etc~1 | tar -tf -
295 - Get a list of your previous split-based backups:
297 GIT_DIR=~/.bup git log local-etc
299 - Make a backup on a remote server:
301 tar -cvf - /etc | bup split -r SERVERNAME: -n local-etc -vv
303 - Try restoring the remote backup tarball:
305 bup join -r SERVERNAME: local-etc | tar -tf -
307 That's all there is to it!
313 - FreeBSD's default 'make' command doesn't like bup's Makefile. In order to
314 compile the code, run tests and install bup, you need to install GNU Make
315 from the port named 'gmake' and use its executable instead in the commands
316 seen above. (i.e. 'gmake test' runs bup's test suite)
318 - Python's development headers are automatically installed with the 'python'
319 port so there's no need to install them separately.
321 - To use the 'bup fuse' command, you need to install the fuse kernel module
322 from the 'fusefs-kmod' port in the 'sysutils' section and the libraries from
323 the port named 'py-fusefs' in the 'devel' section.
325 - The 'par2' command can be found in the port named 'par2cmdline'.
327 - In order to compile the documentation, you need pandoc which can be found in
328 the port named 'hs-pandoc' in the 'textproc' section.
331 Notes on NetBSD/pkgsrc
332 ----------------------
334 - See pkgsrc/sysutils/bup, which should be the most recent stable
335 release and includes man pages. It also has a reasonable set of
336 dependencies (git, par2, py-fuse-bindings).
338 - The "fuse-python" package referred to is hard to locate, and is a
339 separate tarball for the python language binding distributed by the
340 fuse project on sourceforge. It is available as
341 pkgsrc/filesystems/py-fuse-bindings and on NetBSD 5, "bup fuse"
344 - "bup fuse" presents every directory/file as inode 0. The directory
345 traversal code ("fts") in NetBSD's libc will interpret this as a
346 cycle and error out, so "ls -R" and "find" will not work.
348 - There is no support for ACLs. If/when some entrprising person
349 fixes this, adjust t/compare-trees.
355 - There is no support for ACLs. If/when some enterprising person
356 fixes this, adjust t/compare-trees.
358 - In t/test.sh, two tests have been disabled. These tests check to
359 see that repeated saves produce identical trees and that an
360 intervening index doesn't change the SHA1. Apparently Cygwin has
361 some unusual behaviors with respect to access times (that probably
362 warrant further investigation). Possibly related:
363 http://cygwin.com/ml/cygwin/2007-06/msg00436.html
369 - There is no support for ACLs. If/when some enterprising person
370 fixes this, adjust t/compare-trees.
379 bup stores its data in a git-formatted repository. Unfortunately, git
380 itself doesn't actually behave very well for bup's use case (huge numbers of
381 files, files with huge sizes, retaining file permissions/ownership are
382 important), so we mostly don't use git's *code* except for a few helper
383 programs. For example, bup has its own git packfile writer written in
386 Basically, 'bup split' reads the data on stdin (or from files specified on
387 the command line), breaks it into chunks using a rolling checksum (similar to
388 rsync), and saves those chunks into a new git packfile. There is at least one
389 git packfile per backup.
391 When deciding whether to write a particular chunk into the new packfile, bup
392 first checks all the other packfiles that exist to see if they already have that
393 chunk. If they do, the chunk is skipped.
395 git packs come in two parts: the pack itself (*.pack) and the index (*.idx).
396 The index is pretty small, and contains a list of all the objects in the
397 pack. Thus, when generating a remote backup, we don't have to have a copy
398 of the packfiles from the remote server: the local end just downloads a copy
399 of the server's *index* files, and compares objects against those when
400 generating the new pack, which it sends directly to the server.
402 The "-n" option to 'bup split' and 'bup save' is the name of the backup you
403 want to create, but it's actually implemented as a git branch. So you can
404 do cute things like checkout a particular branch using git, and receive a
405 bunch of chunk files corresponding to the file you split.
407 If you use '-b' or '-t' or '-c' instead of '-n', bup split will output a
408 list of blobs, a tree containing that list of blobs, or a commit containing
409 that tree, respectively, to stdout. You can use this to construct your own
410 scripts that do something with those values.
415 'bup index' walks through your filesystem and updates a file (whose name is,
416 by default, ~/.bup/bupindex) to contain the name, attributes, and an
417 optional git SHA1 (blob id) of each file and directory.
419 'bup save' basically just runs the equivalent of 'bup split' a whole bunch
420 of times, once per file in the index, and assembles a git tree
421 that contains all the resulting objects. Among other things, that makes
422 'git diff' much more useful (compared to splitting a tarball, which is
423 essentially a big binary blob). However, since bup splits large files into
424 smaller chunks, the resulting tree structure doesn't *exactly* correspond to
425 what git itself would have stored. Also, the tree format used by 'bup save'
426 will probably change in the future to support storing file ownership, more
427 complex file permissions, and so on.
429 If a file has previously been written by 'bup save', then its git blob/tree
430 id is stored in the index. This lets 'bup save' avoid reading that file to
431 produce future incremental backups, which means it can go *very* fast unless
432 a lot of files have changed.
435 Things that are stupid for now but which we'll fix later
436 ========================================================
438 Help with any of these problems, or others, is very welcome. Join the
439 mailing list (see below) if you'd like to help.
441 - 'bup restore' can't pull directly from a remote server.
443 So in one sense "save -r" is a dead-end right now. Obviously you
444 can use "ssh SERVER bup restore -C ./dest..." to create a tree you
445 can transfer elsewhere via rsync/tar/whatever, but that's *lame*.
447 Until we fix it, you may be able to mount the remote BUP_DIR via
448 sshfs and then restore "normally", though that hasn't been
451 - 'bup save' and 'bup restore' have immature metadata support.
453 On the plus side, they actually do have support now, but it's new,
454 and not remotely as well tested as tar/rsync/whatever's. However,
455 you have to start somewhere, and as of 0.25, we think it's ready
456 for more general use. Please let us know if you have any trouble.
458 Also, if any strip or graft-style options are specified to 'bup
459 save', then no metadata will be written for the root directory.
460 That's obviously less than ideal.
462 - bup is overly optimistic about mmap. Right now bup just assumes
463 that it can mmap as large a block as it likes, and that mmap will
464 never fail. Yeah, right... If nothing else, this has failed on
465 32-bit architectures (and 31-bit is even worse -- looking at you,
468 To fix this, we might just implement a FakeMmap[1] class that uses
469 normal file IO and handles all of the mmap methods[2] that bup
470 actually calls. Then we'd swap in one of those whenever mmap
473 This would also require implementing some of the methods needed to
474 support "[]" array access, probably at a minimum __getitem__,
475 __setitem__, and __setslice__ [3].
477 [1] http://comments.gmane.org/gmane.comp.sysutils.backup.bup/613
478 [2] http://docs.python.org/2/library/mmap.html
479 [3] http://docs.python.org/2/reference/datamodel.html#emulating-container-types
481 - 'bup index' is slower than it should be.
483 It's still rather fast: it can iterate through all the filenames on my
484 600,000 file filesystem in a few seconds. But it still needs to rewrite
485 the entire index file just to add a single filename, which is pretty
486 nasty; it should just leave the new files in a second "extra index" file
489 - bup could use inotify for *really* efficient incremental backups.
491 You could even have your system doing "continuous" backups: whenever a
492 file changes, we immediately send an image of it to the server. We could
493 give the continuous-backup process a really low CPU and I/O priority so
494 you wouldn't even know it was running.
496 - bup only has experimental support for pruning old backups.
498 While you should now be able to drop old saves and branches with
499 `bup rm`, and reclaim the space occupied by data that's no longer
500 needed by other backups with `bup gc`, these commands are
501 experimental, and should be handled with great care. See the
502 man pages for more information.
504 Unless you want to help test the new commands, one possible
505 workaround is to just start a new BUP_DIR occasionally,
506 i.e. bup-2013, bup-2014...
508 - bup has never been tested on anything but Linux, FreeBSD, NetBSD,
509 OS X, and Windows+Cygwin.
511 There's nothing that makes it *inherently* non-portable, though, so
512 that's mostly a matter of someone putting in some effort. (For a
513 "native" Windows port, the most annoying thing is the absence of ssh in
514 a default Windows installation.)
516 - bup needs better documentation.
518 According to a recent article about bup in Linux Weekly News
519 (https://lwn.net/Articles/380983/), "it's a bit short on examples and
520 a user guide would be nice." Documentation is the sort of thing that
521 will never be great unless someone from outside contributes it (since
522 the developers can never remember which parts are hard to understand).
524 - bup is "relatively speedy" and has "pretty good" compression.
526 ...according to the same LWN article. Clearly neither of those is good
527 enough. We should have awe-inspiring speed and crazy-good compression.
528 Must work on that. Writing more parts in C might help with the speed.
532 Actually, that's not stupid, but you might consider it a
533 limitation. See the ["Related Projects"](https://bup.github.io/)
534 list for some possible options.
539 bup has an extensive set of man pages. Try using 'bup help' to get
540 started, or use 'bup help SUBCOMMAND' for any bup subcommand (like split,
541 join, index, save, etc.) to get details on that command.
543 For further technical details, please see ./DESIGN.
549 bup is a work in progress and there are many ways it can still be improved.
550 If you'd like to contribute patches, ideas, or bug reports, please join the
553 You can find the mailing list archives here:
555 http://groups.google.com/group/bup-list
557 and you can subscribe by sending a message to:
559 bup-list+subscribe@googlegroups.com
561 Please see <a href="HACKING">./HACKING</a> for
562 additional information, i.e. how to submit patches (hint - no pull
563 requests), how we handle branches, etc.