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>
103 - Check out the bup source code using git:
105 git clone https://github.com/bup/bup
107 - Install the required python libraries (including the development
110 On very recent Debian/Ubuntu versions, this may be sufficient (run
113 apt-get build-dep bup
115 Otherwise try this (substitute python2.6-dev if you have an older
118 apt-get install python2.7-dev python-fuse
119 apt-get install python-pyxattr python-pylibacl
120 apt-get install linux-libc-dev
121 apt-get install acl attr
122 apt-get install python-tornado # optional
124 On CentOS (for CentOS 6, at least), this should be sufficient (run
127 yum groupinstall "Development Tools"
128 yum install python python-devel
129 yum install fuse-python pyxattr pylibacl
130 yum install perl-Time-HiRes
132 In addition to the default CentOS repositories, you may need to add
133 RPMForge (for fuse-python) and EPEL (for pyxattr and pylibacl).
135 On Cygwin, install python, make, rsync, and gcc4.
137 If you would like to use the optional bup web server on systems
138 without a tornado package, you may want to try this:
142 - Build the python module and symlinks:
150 The tests should pass. If they don't pass for you, stop here and
151 send an email to bup-list@googlegroups.com. Though if there are
152 symbolic links along the current working directory path, the tests
153 may fail. Running something like this before "make test" should
154 sidestep the problem:
158 - You can install bup via "make install", and override the default
159 destination with DESTDIR and PREFIX.
161 Files are normally installed to "$DESTDIR/$PREFIX" where DESTDIR is
162 empty by default, and PREFIX is set to /usr/local. So if you wanted to
163 install bup to /opt/bup, you might do something like this:
165 make install DESTDIR=/opt/bup PREFIX=''
167 - The Python executable that bup will use is chosen by ./configure,
168 which will search for a reasonable version unless PYTHON is set in
169 the environment, in which case, bup will use that path. You can
170 see which Python executable was chosen by looking at the
171 configure output, or examining cmd/python-cmd.sh, and you can
172 change the selection by re-running ./configure.
177 Binary packages of bup are known to be built for the following OSes:
180 http://packages.debian.org/search?searchon=names&keywords=bup
182 http://packages.ubuntu.com/search?searchon=names&keywords=bup
183 - pkgsrc (NetBSD, Dragonfly, and others)
184 http://pkgsrc.se/sysutils/bup
185 http://cvsweb.netbsd.org/bsdweb.cgi/pkgsrc/sysutils/bup/
187 https://www.archlinux.org/packages/?sort=&q=bup
189 https://apps.fedoraproject.org/packages/bup
195 - Get help for any bup command:
204 - Initialize the default BUP_DIR (~/.bup):
208 - Make a local backup (-v or -vv will increase the verbosity):
211 bup save -n local-etc /etc
213 - Restore a local backup to ./dest:
215 bup restore -C ./dest local-etc/latest/etc
218 - Look at how much disk space your backup took:
222 - Make another backup (which should be mostly identical to the last one;
223 notice that you don't have to *specify* that this backup is incremental,
224 it just saves space automatically):
227 bup save -n local-etc /etc
229 - Look how little extra space your second backup used (on top of the first):
233 - Get a list of your previous backups:
237 - Restore your first backup again:
239 bup restore -C ./dest-2 local-etc/2013-11-23-11195/etc
241 - Make a backup to a remote server which must already have the 'bup' command
242 somewhere in its PATH (see /etc/profile, etc/environment, ~/.profile, or
243 ~/.bashrc), and be accessible via ssh.
244 Make sure to replace SERVERNAME with the actual hostname of your server:
246 bup init -r SERVERNAME:path/to/remote-bup-dir
248 bup save -r SERVERNAME:path/to/remote-bup-dir -n local-etc /etc
250 - Restore a backup from a remote server. (FAIL: unfortunately,
251 unlike "bup join", "bup restore" does not yet support remote
252 restores. See both "bup join" and "Things that are stupid" below.)
254 - Defend your backups from death rays (OK fine, more likely from the
255 occasional bad disk block). This writes parity information
256 (currently via par2) for all of the existing data so that bup may
257 be able to recover from some amount of repository corruption:
261 - Use split/join instead of index/save/restore. Try making a local
264 tar -cvf - /etc | bup split -n local-etc -vv
266 - Try restoring the tarball:
268 bup join local-etc | tar -tf -
270 - Look at how much disk space your backup took:
274 - Make another tar backup:
276 tar -cvf - /etc | bup split -n local-etc -vv
278 - Look at how little extra space your second backup used on top of
283 - Restore the first tar backup again (the ~1 is git notation for "one
284 older than the most recent"):
286 bup join local-etc~1 | tar -tf -
288 - Get a list of your previous split-based backups:
290 GIT_DIR=~/.bup git log local-etc
292 - Make a backup on a remote server:
294 tar -cvf - /etc | bup split -r SERVERNAME: -n local-etc -vv
296 - Try restoring the remote backup tarball:
298 bup join -r SERVERNAME: local-etc | tar -tf -
300 That's all there is to it!
306 - FreeBSD's default 'make' command doesn't like bup's Makefile. In order to
307 compile the code, run tests and install bup, you need to install GNU Make
308 from the port named 'gmake' and use its executable instead in the commands
309 seen above. (i.e. 'gmake test' runs bup's test suite)
311 - Python's development headers are automatically installed with the 'python'
312 port so there's no need to install them separately.
314 - To use the 'bup fuse' command, you need to install the fuse kernel module
315 from the 'fusefs-kmod' port in the 'sysutils' section and the libraries from
316 the port named 'py-fusefs' in the 'devel' section.
318 - The 'par2' command can be found in the port named 'par2cmdline'.
320 - In order to compile the documentation, you need pandoc which can be found in
321 the port named 'hs-pandoc' in the 'textproc' section.
324 Notes on NetBSD/pkgsrc
325 ----------------------
327 - See pkgsrc/sysutils/bup, which should be the most recent stable
328 release and includes man pages. It also has a reasonable set of
329 dependencies (git, par2, py-fuse-bindings).
331 - The "fuse-python" package referred to is hard to locate, and is a
332 separate tarball for the python language binding distributed by the
333 fuse project on sourceforge. It is available as
334 pkgsrc/filesystems/py-fuse-bindings and on NetBSD 5, "bup fuse"
337 - "bup fuse" presents every directory/file as inode 0. The directory
338 traversal code ("fts") in NetBSD's libc will interpret this as a
339 cycle and error out, so "ls -R" and "find" will not work.
341 - There is no support for ACLs. If/when some entrprising person
342 fixes this, adjust t/compare-trees.
348 - There is no support for ACLs. If/when some enterprising person
349 fixes this, adjust t/compare-trees.
351 - In t/test.sh, two tests have been disabled. These tests check to
352 see that repeated saves produce identical trees and that an
353 intervening index doesn't change the SHA1. Apparently Cygwin has
354 some unusual behaviors with respect to access times (that probably
355 warrant further investigation). Possibly related:
356 http://cygwin.com/ml/cygwin/2007-06/msg00436.html
362 - There is no support for ACLs. If/when some enterprising person
363 fixes this, adjust t/compare-trees.
372 bup stores its data in a git-formatted repository. Unfortunately, git
373 itself doesn't actually behave very well for bup's use case (huge numbers of
374 files, files with huge sizes, retaining file permissions/ownership are
375 important), so we mostly don't use git's *code* except for a few helper
376 programs. For example, bup has its own git packfile writer written in
379 Basically, 'bup split' reads the data on stdin (or from files specified on
380 the command line), breaks it into chunks using a rolling checksum (similar to
381 rsync), and saves those chunks into a new git packfile. There is at least one
382 git packfile per backup.
384 When deciding whether to write a particular chunk into the new packfile, bup
385 first checks all the other packfiles that exist to see if they already have that
386 chunk. If they do, the chunk is skipped.
388 git packs come in two parts: the pack itself (*.pack) and the index (*.idx).
389 The index is pretty small, and contains a list of all the objects in the
390 pack. Thus, when generating a remote backup, we don't have to have a copy
391 of the packfiles from the remote server: the local end just downloads a copy
392 of the server's *index* files, and compares objects against those when
393 generating the new pack, which it sends directly to the server.
395 The "-n" option to 'bup split' and 'bup save' is the name of the backup you
396 want to create, but it's actually implemented as a git branch. So you can
397 do cute things like checkout a particular branch using git, and receive a
398 bunch of chunk files corresponding to the file you split.
400 If you use '-b' or '-t' or '-c' instead of '-n', bup split will output a
401 list of blobs, a tree containing that list of blobs, or a commit containing
402 that tree, respectively, to stdout. You can use this to construct your own
403 scripts that do something with those values.
408 'bup index' walks through your filesystem and updates a file (whose name is,
409 by default, ~/.bup/bupindex) to contain the name, attributes, and an
410 optional git SHA1 (blob id) of each file and directory.
412 'bup save' basically just runs the equivalent of 'bup split' a whole bunch
413 of times, once per file in the index, and assembles a git tree
414 that contains all the resulting objects. Among other things, that makes
415 'git diff' much more useful (compared to splitting a tarball, which is
416 essentially a big binary blob). However, since bup splits large files into
417 smaller chunks, the resulting tree structure doesn't *exactly* correspond to
418 what git itself would have stored. Also, the tree format used by 'bup save'
419 will probably change in the future to support storing file ownership, more
420 complex file permissions, and so on.
422 If a file has previously been written by 'bup save', then its git blob/tree
423 id is stored in the index. This lets 'bup save' avoid reading that file to
424 produce future incremental backups, which means it can go *very* fast unless
425 a lot of files have changed.
428 Things that are stupid for now but which we'll fix later
429 ========================================================
431 Help with any of these problems, or others, is very welcome. Join the
432 mailing list (see below) if you'd like to help.
434 - 'bup restore' can't pull directly from a remote server.
436 So in one sense "save -r" is a dead-end right now. Obviously you
437 can use "ssh SERVER bup restore -C ./dest..." to create a tree you
438 can transfer elsewhere via rsync/tar/whatever, but that's *lame*.
440 Until we fix it, you may be able to mount the remote BUP_DIR via
441 sshfs and then restore "normally", though that hasn't been
444 - 'bup save' and 'bup restore' have immature metadata support.
446 On the plus side, they actually do have support now, but it's new,
447 and not remotely as well tested as tar/rsync/whatever's. However,
448 you have to start somewhere, and as of 0.25, we think it's ready
449 for more general use. Please let us know if you have any trouble.
451 Also, if any strip or graft-style options are specified to 'bup
452 save', then no metadata will be written for the root directory.
453 That's obviously less than ideal.
455 - bup is overly optimistic about mmap. Right now bup just assumes
456 that it can mmap as large a block as it likes, and that mmap will
457 never fail. Yeah, right... If nothing else, this has failed on
458 32-bit architectures (and 31-bit is even worse -- looking at you,
461 To fix this, we might just implement a FakeMmap[1] class that uses
462 normal file IO and handles all of the mmap methods[2] that bup
463 actually calls. Then we'd swap in one of those whenever mmap
466 This would also require implementing some of the methods needed to
467 support "[]" array access, probably at a minimum __getitem__,
468 __setitem__, and __setslice__ [3].
470 [1] http://comments.gmane.org/gmane.comp.sysutils.backup.bup/613
471 [2] http://docs.python.org/2/library/mmap.html
472 [3] http://docs.python.org/2/reference/datamodel.html#emulating-container-types
474 - 'bup index' is slower than it should be.
476 It's still rather fast: it can iterate through all the filenames on my
477 600,000 file filesystem in a few seconds. But it still needs to rewrite
478 the entire index file just to add a single filename, which is pretty
479 nasty; it should just leave the new files in a second "extra index" file
482 - bup could use inotify for *really* efficient incremental backups.
484 You could even have your system doing "continuous" backups: whenever a
485 file changes, we immediately send an image of it to the server. We could
486 give the continuous-backup process a really low CPU and I/O priority so
487 you wouldn't even know it was running.
489 - bup only has experimental support for pruning old backups.
491 While you should now be able to drop old saves and branches with
492 `bup rm`, and reclaim the space occupied by data that's no longer
493 needed by other backups with `bup gc`, these commands are
494 experimental, and should be handled with great care. See the
495 man pages for more information.
497 Unless you want to help test the new commands, one possible
498 workaround is to just start a new BUP_DIR occasionally,
499 i.e. bup-2013, bup-2014...
501 - bup has never been tested on anything but Linux, FreeBSD, NetBSD,
502 OS X, and Windows+Cygwin.
504 There's nothing that makes it *inherently* non-portable, though, so
505 that's mostly a matter of someone putting in some effort. (For a
506 "native" Windows port, the most annoying thing is the absence of ssh in
507 a default Windows installation.)
509 - bup needs better documentation.
511 According to a recent article about bup in Linux Weekly News
512 (https://lwn.net/Articles/380983/), "it's a bit short on examples and
513 a user guide would be nice." Documentation is the sort of thing that
514 will never be great unless someone from outside contributes it (since
515 the developers can never remember which parts are hard to understand).
517 - bup is "relatively speedy" and has "pretty good" compression.
519 ...according to the same LWN article. Clearly neither of those is good
520 enough. We should have awe-inspiring speed and crazy-good compression.
521 Must work on that. Writing more parts in C might help with the speed.
525 Actually, that's not stupid, but you might consider it a
526 limitation. See the ["Related Projects"](https://bup.github.io/)
527 list for some possible options.
532 bup has an extensive set of man pages. Try using 'bup help' to get
533 started, or use 'bup help SUBCOMMAND' for any bup subcommand (like split,
534 join, index, save, etc.) to get details on that command.
536 For further technical details, please see ./DESIGN.
542 bup is a work in progress and there are many ways it can still be improved.
543 If you'd like to contribute patches, ideas, or bug reports, please join the
546 You can find the mailing list archives here:
548 http://groups.google.com/group/bup-list
550 and you can subscribe by sending a message to:
552 bup-list+subscribe@googlegroups.com
554 Please see <a href="HACKING">./HACKING</a> for
555 additional information, i.e. how to submit patches (hint - no pull
556 requests), how we handle branches, etc.