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-from-0.28.1.md">Changes in 0.29 as compared to 0.28.1</a>
90 - <a href="note/0.28.1-from-0.28.md">Changes in 0.28.1 as compared to 0.28</a>
91 - <a href="note/0.28-from-0.27.1.md">Changes in 0.28 as compared to 0.27.1</a>
92 - <a href="note/0.27.1-from-0.27.md">Changes in 0.27.1 as compared to 0.27</a>
101 - Check out the bup source code using git:
103 git clone https://github.com/bup/bup
105 - Install the required python libraries (including the development
108 On very recent Debian/Ubuntu versions, this may be sufficient (run
111 apt-get build-dep bup
113 Otherwise try this (substitute python2.6-dev if you have an older
116 apt-get install python2.7-dev python-fuse
117 apt-get install python-pyxattr python-pylibacl
118 apt-get install linux-libc-dev
119 apt-get install acl attr
120 apt-get install python-tornado # optional
122 On CentOS (for CentOS 6, at least), this should be sufficient (run
125 yum groupinstall "Development Tools"
126 yum install python python-devel
127 yum install fuse-python pyxattr pylibacl
128 yum install perl-Time-HiRes
130 In addition to the default CentOS repositories, you may need to add
131 RPMForge (for fuse-python) and EPEL (for pyxattr and pylibacl).
133 On Cygwin, install python, make, rsync, and gcc4.
135 If you would like to use the optional bup web server on systems
136 without a tornado package, you may want to try this:
140 - Build the python module and symlinks:
148 The tests should pass. If they don't pass for you, stop here and
149 send an email to bup-list@googlegroups.com. Though if there are
150 symbolic links along the current working directory path, the tests
151 may fail. Running something like this before "make test" should
152 sidestep the problem:
156 - You can install bup via "make install", and override the default
157 destination with DESTDIR and PREFIX.
159 Files are normally installed to "$DESTDIR/$PREFIX" where DESTDIR is
160 empty by default, and PREFIX is set to /usr/local. So if you wanted to
161 install bup to /opt/bup, you might do something like this:
163 make install DESTDIR=/opt/bup PREFIX=''
165 - The Python executable that bup will use is chosen by ./configure,
166 which will search for a reasonable version unless PYTHON is set in
167 the environment, in which case, bup will use that path. You can
168 see which Python executable was chosen by looking at the
169 configure output, or examining cmd/python-cmd.sh, and you can
170 change the selection by re-running ./configure.
175 Binary packages of bup are known to be built for the following OSes:
178 http://packages.debian.org/search?searchon=names&keywords=bup
180 http://packages.ubuntu.com/search?searchon=names&keywords=bup
181 - pkgsrc (NetBSD, Dragonfly, and others)
182 http://pkgsrc.se/sysutils/bup
183 http://cvsweb.netbsd.org/bsdweb.cgi/pkgsrc/sysutils/bup/
185 https://www.archlinux.org/packages/?sort=&q=bup
187 https://apps.fedoraproject.org/packages/bup
193 - Get help for any bup command:
202 - Initialize the default BUP_DIR (~/.bup):
206 - Make a local backup (-v or -vv will increase the verbosity):
209 bup save -n local-etc /etc
211 - Restore a local backup to ./dest:
213 bup restore -C ./dest local-etc/latest/etc
216 - Look at how much disk space your backup took:
220 - Make another backup (which should be mostly identical to the last one;
221 notice that you don't have to *specify* that this backup is incremental,
222 it just saves space automatically):
225 bup save -n local-etc /etc
227 - Look how little extra space your second backup used (on top of the first):
231 - Get a list of your previous backups:
235 - Restore your first backup again:
237 bup restore -C ./dest-2 local-etc/2013-11-23-11195/etc
239 - Make a backup to a remote server which must already have the 'bup' command
240 somewhere in its PATH (see /etc/profile, etc/environment, ~/.profile, or
241 ~/.bashrc), and be accessible via ssh.
242 Make sure to replace SERVERNAME with the actual hostname of your server:
244 bup init -r SERVERNAME:path/to/remote-bup-dir
246 bup save -r SERVERNAME:path/to/remote-bup-dir -n local-etc /etc
248 - Restore a backup from a remote server. (FAIL: unfortunately,
249 unlike "bup join", "bup restore" does not yet support remote
250 restores. See both "bup join" and "Things that are stupid" below.)
252 - Defend your backups from death rays (OK fine, more likely from the
253 occasional bad disk block). This writes parity information
254 (currently via par2) for all of the existing data so that bup may
255 be able to recover from some amount of repository corruption:
259 - Use split/join instead of index/save/restore. Try making a local
262 tar -cvf - /etc | bup split -n local-etc -vv
264 - Try restoring the tarball:
266 bup join local-etc | tar -tf -
268 - Look at how much disk space your backup took:
272 - Make another tar backup:
274 tar -cvf - /etc | bup split -n local-etc -vv
276 - Look at how little extra space your second backup used on top of
281 - Restore the first tar backup again (the ~1 is git notation for "one
282 older than the most recent"):
284 bup join local-etc~1 | tar -tf -
286 - Get a list of your previous split-based backups:
288 GIT_DIR=~/.bup git log local-etc
290 - Make a backup on a remote server:
292 tar -cvf - /etc | bup split -r SERVERNAME: -n local-etc -vv
294 - Try restoring the remote backup tarball:
296 bup join -r SERVERNAME: local-etc | tar -tf -
298 That's all there is to it!
304 - FreeBSD's default 'make' command doesn't like bup's Makefile. In order to
305 compile the code, run tests and install bup, you need to install GNU Make
306 from the port named 'gmake' and use its executable instead in the commands
307 seen above. (i.e. 'gmake test' runs bup's test suite)
309 - Python's development headers are automatically installed with the 'python'
310 port so there's no need to install them separately.
312 - To use the 'bup fuse' command, you need to install the fuse kernel module
313 from the 'fusefs-kmod' port in the 'sysutils' section and the libraries from
314 the port named 'py-fusefs' in the 'devel' section.
316 - The 'par2' command can be found in the port named 'par2cmdline'.
318 - In order to compile the documentation, you need pandoc which can be found in
319 the port named 'hs-pandoc' in the 'textproc' section.
322 Notes on NetBSD/pkgsrc
323 ----------------------
325 - See pkgsrc/sysutils/bup, which should be the most recent stable
326 release and includes man pages. It also has a reasonable set of
327 dependencies (git, par2, py-fuse-bindings).
329 - The "fuse-python" package referred to is hard to locate, and is a
330 separate tarball for the python language binding distributed by the
331 fuse project on sourceforge. It is available as
332 pkgsrc/filesystems/py-fuse-bindings and on NetBSD 5, "bup fuse"
335 - "bup fuse" presents every directory/file as inode 0. The directory
336 traversal code ("fts") in NetBSD's libc will interpret this as a
337 cycle and error out, so "ls -R" and "find" will not work.
339 - There is no support for ACLs. If/when some entrprising person
340 fixes this, adjust t/compare-trees.
346 - There is no support for ACLs. If/when some enterprising person
347 fixes this, adjust t/compare-trees.
349 - In t/test.sh, two tests have been disabled. These tests check to
350 see that repeated saves produce identical trees and that an
351 intervening index doesn't change the SHA1. Apparently Cygwin has
352 some unusual behaviors with respect to access times (that probably
353 warrant further investigation). Possibly related:
354 http://cygwin.com/ml/cygwin/2007-06/msg00436.html
360 - There is no support for ACLs. If/when some enterprising person
361 fixes this, adjust t/compare-trees.
370 bup stores its data in a git-formatted repository. Unfortunately, git
371 itself doesn't actually behave very well for bup's use case (huge numbers of
372 files, files with huge sizes, retaining file permissions/ownership are
373 important), so we mostly don't use git's *code* except for a few helper
374 programs. For example, bup has its own git packfile writer written in
377 Basically, 'bup split' reads the data on stdin (or from files specified on
378 the command line), breaks it into chunks using a rolling checksum (similar to
379 rsync), and saves those chunks into a new git packfile. There is at least one
380 git packfile per backup.
382 When deciding whether to write a particular chunk into the new packfile, bup
383 first checks all the other packfiles that exist to see if they already have that
384 chunk. If they do, the chunk is skipped.
386 git packs come in two parts: the pack itself (*.pack) and the index (*.idx).
387 The index is pretty small, and contains a list of all the objects in the
388 pack. Thus, when generating a remote backup, we don't have to have a copy
389 of the packfiles from the remote server: the local end just downloads a copy
390 of the server's *index* files, and compares objects against those when
391 generating the new pack, which it sends directly to the server.
393 The "-n" option to 'bup split' and 'bup save' is the name of the backup you
394 want to create, but it's actually implemented as a git branch. So you can
395 do cute things like checkout a particular branch using git, and receive a
396 bunch of chunk files corresponding to the file you split.
398 If you use '-b' or '-t' or '-c' instead of '-n', bup split will output a
399 list of blobs, a tree containing that list of blobs, or a commit containing
400 that tree, respectively, to stdout. You can use this to construct your own
401 scripts that do something with those values.
406 'bup index' walks through your filesystem and updates a file (whose name is,
407 by default, ~/.bup/bupindex) to contain the name, attributes, and an
408 optional git SHA1 (blob id) of each file and directory.
410 'bup save' basically just runs the equivalent of 'bup split' a whole bunch
411 of times, once per file in the index, and assembles a git tree
412 that contains all the resulting objects. Among other things, that makes
413 'git diff' much more useful (compared to splitting a tarball, which is
414 essentially a big binary blob). However, since bup splits large files into
415 smaller chunks, the resulting tree structure doesn't *exactly* correspond to
416 what git itself would have stored. Also, the tree format used by 'bup save'
417 will probably change in the future to support storing file ownership, more
418 complex file permissions, and so on.
420 If a file has previously been written by 'bup save', then its git blob/tree
421 id is stored in the index. This lets 'bup save' avoid reading that file to
422 produce future incremental backups, which means it can go *very* fast unless
423 a lot of files have changed.
426 Things that are stupid for now but which we'll fix later
427 ========================================================
429 Help with any of these problems, or others, is very welcome. Join the
430 mailing list (see below) if you'd like to help.
432 - 'bup restore' can't pull directly from a remote server.
434 So in one sense "save -r" is a dead-end right now. Obviously you
435 can use "ssh SERVER bup restore -C ./dest..." to create a tree you
436 can transfer elsewhere via rsync/tar/whatever, but that's *lame*.
438 Until we fix it, you may be able to mount the remote BUP_DIR via
439 sshfs and then restore "normally", though that hasn't been
442 - 'bup save' and 'bup restore' have immature metadata support.
444 On the plus side, they actually do have support now, but it's new,
445 and not remotely as well tested as tar/rsync/whatever's. However,
446 you have to start somewhere, and as of 0.25, we think it's ready
447 for more general use. Please let us know if you have any trouble.
449 Also, if any strip or graft-style options are specified to 'bup
450 save', then no metadata will be written for the root directory.
451 That's obviously less than ideal.
453 - bup is overly optimistic about mmap. Right now bup just assumes
454 that it can mmap as large a block as it likes, and that mmap will
455 never fail. Yeah, right... If nothing else, this has failed on
456 32-bit architectures (and 31-bit is even worse -- looking at you,
459 To fix this, we might just implement a FakeMmap[1] class that uses
460 normal file IO and handles all of the mmap methods[2] that bup
461 actually calls. Then we'd swap in one of those whenever mmap
464 This would also require implementing some of the methods needed to
465 support "[]" array access, probably at a minimum __getitem__,
466 __setitem__, and __setslice__ [3].
468 [1] http://comments.gmane.org/gmane.comp.sysutils.backup.bup/613
469 [2] http://docs.python.org/2/library/mmap.html
470 [3] http://docs.python.org/2/reference/datamodel.html#emulating-container-types
472 - 'bup index' is slower than it should be.
474 It's still rather fast: it can iterate through all the filenames on my
475 600,000 file filesystem in a few seconds. But it still needs to rewrite
476 the entire index file just to add a single filename, which is pretty
477 nasty; it should just leave the new files in a second "extra index" file
480 - bup could use inotify for *really* efficient incremental backups.
482 You could even have your system doing "continuous" backups: whenever a
483 file changes, we immediately send an image of it to the server. We could
484 give the continuous-backup process a really low CPU and I/O priority so
485 you wouldn't even know it was running.
487 - bup only has experimental support for pruning old backups.
489 While you should now be able to drop old saves and branches with
490 `bup rm`, and reclaim the space occupied by data that's no longer
491 needed by other backups with `bup gc`, these commands are
492 experimental, and should be handled with great care. See the
493 man pages for more information.
495 Unless you want to help test the new commands, one possible
496 workaround is to just start a new BUP_DIR occasionally,
497 i.e. bup-2013, bup-2014...
499 - bup has never been tested on anything but Linux, FreeBSD, NetBSD,
500 OS X, and Windows+Cygwin.
502 There's nothing that makes it *inherently* non-portable, though, so
503 that's mostly a matter of someone putting in some effort. (For a
504 "native" Windows port, the most annoying thing is the absence of ssh in
505 a default Windows installation.)
507 - bup needs better documentation.
509 According to a recent article about bup in Linux Weekly News
510 (https://lwn.net/Articles/380983/), "it's a bit short on examples and
511 a user guide would be nice." Documentation is the sort of thing that
512 will never be great unless someone from outside contributes it (since
513 the developers can never remember which parts are hard to understand).
515 - bup is "relatively speedy" and has "pretty good" compression.
517 ...according to the same LWN article. Clearly neither of those is good
518 enough. We should have awe-inspiring speed and crazy-good compression.
519 Must work on that. Writing more parts in C might help with the speed.
523 Actually, that's not stupid, but you might consider it a
524 limitation. See the ["Related Projects"](https://bup.github.io/)
525 list for some possible options.
530 bup has an extensive set of man pages. Try using 'bup help' to get
531 started, or use 'bup help SUBCOMMAND' for any bup subcommand (like split,
532 join, index, save, etc.) to get details on that command.
534 For further technical details, please see ./DESIGN.
540 bup is a work in progress and there are many ways it can still be improved.
541 If you'd like to contribute patches, ideas, or bug reports, please join the
544 You can find the mailing list archives here:
546 http://groups.google.com/group/bup-list
548 and you can subscribe by sending a message to:
550 bup-list+subscribe@googlegroups.com
552 Please see <a href="HACKING">./HACKING</a> for
553 additional information, i.e. how to submit patches (hint - no pull
554 requests), how we handle branches, etc.