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 >= 2.5, 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, MacOS X >= 10.4,
80 NetBSD, Solaris, or Windows (with Cygwin). Patches to support
81 other platforms are welcome.
83 - Any items in "Things that are stupid" below.
93 - Check out the bup source code using git:
95 git clone git://github.com/bup/bup
97 - Install the required python libraries (including the development
100 On very recent Debian/Ubuntu versions, this may be sufficient (run
103 apt-get build-dep bup
105 Otherwise try this (substitute python2.5-dev if you have an older
108 apt-get install python2.6-dev python-fuse
109 apt-get install python-pyxattr python-pylibacl
110 apt-get install linux-libc-dev
111 apt-get install python-tornado # optional
113 On CentOS (for CentOS 6, at least), this should be sufficient (run
116 yum groupinstall "Development Tools"
117 yum install python python-devel
118 yum install fuse-python pyxattr pylibacl
119 yum install perl-Time-HiRes
121 In addition to the default CentOS repositories, you may need to add
122 RPMForge (for fuse-python) and EPEL (for pyxattr and pylibacl).
124 On Cygwin, install python, make, rsync, and gcc4.
126 If you would like to use the optional bup web server on systems
127 without a tornado package, you may want to try this:
131 - Build the python module and symlinks:
139 (The tests should pass. If they don't pass for you, stop here and send
140 an email to bup-list@googlegroups.com.)
142 - You can install bup via "make install", and override the default
143 destination with DESTDIR and PREFIX.
145 Files are normally installed to "$DESTDIR/$PREFIX" where DESTDIR is
146 empty by default, and PREFIX is set to /usr. So if you wanted to
147 install bup to /opt/bup, you might do something like this:
149 make install DESTDIR=/opt/bup PREFIX=''
155 Binary packages of bup are known to be built for the following OSes:
158 http://packages.debian.org/search?searchon=names&keywords=bup
160 http://packages.ubuntu.com/search?searchon=names&keywords=bup
161 - pkgsrc (NetBSD, Dragonfly, and others)
162 http://pkgsrc.se/sysutils/bup
163 http://cvsweb.netbsd.org/bsdweb.cgi/pkgsrc/sysutils/bup/
169 - Get help for any bup command:
178 - Initialize the default BUP_DIR (~/.bup):
182 - Make a local backup (-v or -vv will increase the verbosity):
185 bup save -n local-etc /etc
187 - Restore a local backup to ./dest:
189 bup restore -C ./dest local-etc/latest/etc
192 - Look at how much disk space your backup took:
196 - Make another backup (which should be mostly identical to the last one;
197 notice that you don't have to *specify* that this backup is incremental,
198 it just saves space automatically):
201 bup save -n local-etc /etc
203 - Look how little extra space your second backup used (on top of the first):
207 - Get a list of your previous backups:
211 - Restore your first backup again:
213 bup restore -C ./dest-2 local-etc/2013-11-23-11195/etc
215 - Make a backup on a remote server (which must already have the 'bup' command
216 somewhere in the server's PATH (see /etc/profile, etc/environment,
217 ~/.profile, or ~/.bashrc), and be accessible via ssh.
218 Make sure to replace SERVERNAME with the actual hostname of your server):
220 ssh SERVERNAME bup init
222 bup save -r SERVERNAME: -n local-etc /etc
224 - Restore a backup from a remote server. (FAIL: unfortunately,
225 unlike "bup join", "bup restore" does not yet support remote
226 restores. See both "bup join" and "Things that are stupid" below.)
228 - Defend your backups from death rays (OK fine, more likely from the
229 occasional bad disk block). This writes parity information
230 (currently via par2) for all of the existing data so that bup may
231 be able to recover from some amount of repository corruption:
235 - Use split/join instead of index/save/restore. Try making a local
238 tar -cvf - /etc | bup split -n local-etc -vv
240 - Try restoring the tarball:
242 bup join local-etc | tar -tf -
244 - Look at how much disk space your backup took:
248 - Make another tar backup:
250 tar -cvf - /etc | bup split -n local-etc -vv
252 - Look at how little extra space your second backup used on top of
257 - Restore the first tar backup again (the ~1 is git notation for "one
258 older than the most recent"):
260 bup join local-etc~1 | tar -tf -
262 - Get a list of your previous split-based backups:
264 GIT_DIR=~/.bup git log local-etc
266 - Make a backup on a remote server:
268 tar -cvf - /etc | bup split -r SERVERNAME: -n local-etc -vv
270 - Try restoring the remote backup tarball:
272 bup join -r SERVERNAME: local-etc | tar -tf -
274 That's all there is to it!
280 - FreeBSD's default 'make' command doesn't like bup's Makefile. In order to
281 compile the code, run tests and install bup, you need to install GNU Make
282 from the port named 'gmake' and use its executable instead in the commands
283 seen above. (i.e. 'gmake test' runs bup's test suite)
285 - Python's development headers are automatically installed with the 'python'
286 port so there's no need to install them separately.
288 - To use the 'bup fuse' command, you need to install the fuse kernel module
289 from the 'fusefs-kmod' port in the 'sysutils' section and the libraries from
290 the port named 'py-fusefs' in the 'devel' section.
292 - The 'par2' command can be found in the port named 'par2cmdline'.
294 - In order to compile the documentation, you need pandoc which can be found in
295 the port named 'hs-pandoc' in the 'textproc' section.
298 Notes on NetBSD/pkgsrc
299 ----------------------
301 - See pkgsrc/sysutils/bup, which should be the most recent stable
302 release and includes man pages. It also has a reasonable set of
303 dependencies (git, par2, py-fuse-bindings).
305 - The "fuse-python" package referred to is hard to locate, and is a
306 separate tarball for the python language binding distributed by the
307 fuse project on sourceforge. It is available as
308 pkgsrc/filesystems/py-fuse-bindings and on NetBSD 5, "bup fuse"
311 - "bup fuse" presents every directory/file as inode 0. The directory
312 traversal code ("fts") in NetBSD's libc will interpret this as a
313 cycle and error out, so "ls -R" and "find" will not work.
315 - There is no support for ACLs. If/when some entrprising person
316 fixes this, adjust t/compare-trees.
322 - There is no support for ACLs. If/when some enterprising person
323 fixes this, adjust t/compare-trees.
325 - In t/test.sh, two tests have been disabled. These tests check to
326 see that repeated saves produce identical trees and that an
327 intervening index doesn't change the SHA1. Apparently Cygwin has
328 some unusual behaviors with respect to access times (that probably
329 warrant further investigation). Possibly related:
330 http://cygwin.com/ml/cygwin/2007-06/msg00436.html
336 - There is no support for ACLs. If/when some enterprising person
337 fixes this, adjust t/compare-trees.
346 bup stores its data in a git-formatted repository. Unfortunately, git
347 itself doesn't actually behave very well for bup's use case (huge numbers of
348 files, files with huge sizes, retaining file permissions/ownership are
349 important), so we mostly don't use git's *code* except for a few helper
350 programs. For example, bup has its own git packfile writer written in
353 Basically, 'bup split' reads the data on stdin (or from files specified on
354 the command line), breaks it into chunks using a rolling checksum (similar to
355 rsync), and saves those chunks into a new git packfile. There is one git
358 When deciding whether to write a particular chunk into the new packfile, bup
359 first checks all the other packfiles that exist to see if they already have that
360 chunk. If they do, the chunk is skipped.
362 git packs come in two parts: the pack itself (*.pack) and the index (*.idx).
363 The index is pretty small, and contains a list of all the objects in the
364 pack. Thus, when generating a remote backup, we don't have to have a copy
365 of the packfiles from the remote server: the local end just downloads a copy
366 of the server's *index* files, and compares objects against those when
367 generating the new pack, which it sends directly to the server.
369 The "-n" option to 'bup split' and 'bup save' is the name of the backup you
370 want to create, but it's actually implemented as a git branch. So you can
371 do cute things like checkout a particular branch using git, and receive a
372 bunch of chunk files corresponding to the file you split.
374 If you use '-b' or '-t' or '-c' instead of '-n', bup split will output a
375 list of blobs, a tree containing that list of blobs, or a commit containing
376 that tree, respectively, to stdout. You can use this to construct your own
377 scripts that do something with those values.
382 'bup index' walks through your filesystem and updates a file (whose name is,
383 by default, ~/.bup/bupindex) to contain the name, attributes, and an
384 optional git SHA1 (blob id) of each file and directory.
386 'bup save' basically just runs the equivalent of 'bup split' a whole bunch
387 of times, once per file in the index, and assembles a git tree
388 that contains all the resulting objects. Among other things, that makes
389 'git diff' much more useful (compared to splitting a tarball, which is
390 essentially a big binary blob). However, since bup splits large files into
391 smaller chunks, the resulting tree structure doesn't *exactly* correspond to
392 what git itself would have stored. Also, the tree format used by 'bup save'
393 will probably change in the future to support storing file ownership, more
394 complex file permissions, and so on.
396 If a file has previously been written by 'bup save', then its git blob/tree
397 id is stored in the index. This lets 'bup save' avoid reading that file to
398 produce future incremental backups, which means it can go *very* fast unless
399 a lot of files have changed.
402 Things that are stupid for now but which we'll fix later
403 ========================================================
405 Help with any of these problems, or others, is very welcome. Join the
406 mailing list (see below) if you'd like to help.
408 - 'bup restore' can't pull directly from a remote server.
410 So in one sense "save -r" is a dead-end right now. Obviously you
411 can use "ssh SERVER bup restore -C ./dest..." to create a tree you
412 can transfer elsewhere via rsync/tar/whatever, but that's *lame*.
414 Until we fix it, you may be able to mount the remote BUP_DIR via
415 sshfs and then restore "normally", though that hasn't been
418 - 'bup save' and 'bup restore' have immature metadata support.
420 On the plus side, they actually do have support now, but it's new,
421 and not remotely as well tested as tar/rsync/whatever's. However,
422 you have to start somewhere, and as of 0.25, we think it's ready
423 for more general use. Please let us know if you have any trouble.
425 Also, if any strip or graft-style options are specified to 'bup
426 save', then no metadata will be written for the root directory.
427 That's obviously less than ideal.
429 - bup is overly optimistic about mmap. Right now bup just assumes
430 that it can mmap as large a block as it likes, and that mmap will
431 never fail. Yeah, right... If nothing else, this has failed on
432 32-bit architectures (and 31-bit is even worse -- looking at you,
435 To fix this, we might just implement a FakeMmap[1] class that uses
436 normal file IO and handles all of the mmap methods[2] that bup
437 actually calls. Then we'd swap in one of those whenever mmap
440 This would also require implementing some of the methods needed to
441 support "[]" array access, probably at a minimum __getitem__,
442 __setitem__, and __setslice__ [3].
444 [1] http://comments.gmane.org/gmane.comp.sysutils.backup.bup/613
445 [2] http://docs.python.org/2/library/mmap.html
446 [3] http://docs.python.org/2/reference/datamodel.html#emulating-container-types
448 - 'bup index' is slower than it should be.
450 It's still rather fast: it can iterate through all the filenames on my
451 600,000 file filesystem in a few seconds. But it still needs to rewrite
452 the entire index file just to add a single filename, which is pretty
453 nasty; it should just leave the new files in a second "extra index" file
456 - bup could use inotify for *really* efficient incremental backups.
458 You could even have your system doing "continuous" backups: whenever a
459 file changes, we immediately send an image of it to the server. We could
460 give the continuous-backup process a really low CPU and I/O priority so
461 you wouldn't even know it was running.
463 - bup currently has no way to prune *old* backups.
465 Because of the way the packfile system works, backups become "entangled"
466 in weird ways and it's not actually possible to delete one pack
467 (corresponding approximately to one backup) without risking screwing up
470 git itself has lots of ways of optimizing this sort of thing, but its
471 methods aren't really applicable here; bup packfiles are just too huge.
472 We'll have to do it in a totally different way. There are lots of
473 options. For now: make sure you've got lots of disk space :)
475 Until we fix this, one possible workaround is to just start a new
476 BUP_DIR occasionally, i.e. bup-2013-10, bup-2013-11...
478 - bup has never been tested on anything but Linux, MacOS, and Windows+Cygwin.
480 There's nothing that makes it *inherently* non-portable, though, so
481 that's mostly a matter of someone putting in some effort. (For a
482 "native" Windows port, the most annoying thing is the absence of ssh in
483 a default Windows installation.)
485 - bup needs better documentation.
487 According to a recent article about bup in Linux Weekly News
488 (https://lwn.net/Articles/380983/), "it's a bit short on examples and
489 a user guide would be nice." Documentation is the sort of thing that
490 will never be great unless someone from outside contributes it (since
491 the developers can never remember which parts are hard to understand).
493 - bup is "relatively speedy" and has "pretty good" compression.
495 ...according to the same LWN article. Clearly neither of those is good
496 enough. We should have awe-inspiring speed and crazy-good compression.
497 Must work on that. Writing more parts in C might help with the speed.
501 Actually, that's not stupid, but you might consider it a limitation.
502 There are a bunch of Linux GUI backup programs; someday I expect someone
503 will adapt one of them to use bup.
509 bup has an extensive set of man pages. Try using 'bup help' to get
510 started, or use 'bup help SUBCOMMAND' for any bup subcommand (like split,
511 join, index, save, etc.) to get details on that command.
513 For further technical details, please see ./DESIGN.
519 bup is a work in progress and there are many ways it can still be improved.
520 If you'd like to contribute patches, ideas, or bug reports, please join the
523 You can find the mailing list archives here:
525 http://groups.google.com/group/bup-list
527 and you can subscribe by sending a message to:
529 bup-list+subscribe@googlegroups.com
531 Please see <a href="HACKING">./HACKING</a> for
532 additional information, i.e. how to submit patches (hint - no pull
533 requests), how we handle branches, etc.