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 version >=
77 - It currently only works on Linux, MacOS X >= 10.4,
78 NetBSD, Solaris, or Windows (with Cygwin). Patches to support
79 other platforms are welcome.
81 - Any items in "Things that are stupid" below.
91 - Check out the bup source code using git:
93 git clone git://github.com/bup/bup
95 - Install the required python libraries (including the development
98 On very recent Debian/Ubuntu versions, this may be sufficient (run
101 apt-get build-dep bup
103 Otherwise try this (substitute python2.5-dev if you have an older
106 apt-get install python2.6-dev python-fuse
107 apt-get install python-pyxattr python-pylibacl
108 apt-get install linux-libc-dev
109 apt-get install python-tornado # optional
111 On CentOS (for CentOS 6, at least), this should be sufficient (run
114 yum groupinstall "Development Tools"
115 yum install python python-devel
116 yum install fuse-python pyxattr pylibacl
117 yum install perl-Time-HiRes
119 In addition to the default CentOS repositories, you may need to add
120 RPMForge (for fuse-python) and EPEL (for pyxattr and pylibacl).
122 On Cygwin, install python, make, rsync, and gcc4.
124 If you would like to use the optional bup web server on systems
125 without a tornado package, you may want to try this:
129 - Build the python module and symlinks:
137 (The tests should pass. If they don't pass for you, stop here and send
138 an email to bup-list@googlegroups.com.)
140 - You can install bup via "make install", and override the default
141 destination with DESTDIR and PREFIX.
143 Files are normally installed to "$DESTDIR/$PREFIX" where DESTDIR is
144 empty by default, and PREFIX is set to /usr. So if you wanted to
145 install bup to /opt/bup, you might do something like this:
147 make install DESTDIR=/opt/bup PREFIX=''
153 Binary packages of bup are known to be built for the following OSes:
156 http://packages.debian.org/search?searchon=names&keywords=bup
158 http://packages.ubuntu.com/search?searchon=names&keywords=bup
159 - pkgsrc (NetBSD, Dragonfly, and others)
160 http://pkgsrc.se/sysutils/bup
161 http://cvsweb.netbsd.org/bsdweb.cgi/pkgsrc/sysutils/bup/
167 - Get help for any bup command:
176 - Initialize the default BUP_DIR (~/.bup):
180 - Make a local backup (-v or -vv will increase the verbosity):
183 bup save -n local-etc /etc
185 - Restore a local backup to ./dest:
187 bup restore -C ./dest local-etc/latest/etc
190 - Look at how much disk space your backup took:
194 - Make another backup (which should be mostly identical to the last one;
195 notice that you don't have to *specify* that this backup is incremental,
196 it just saves space automatically):
199 bup save -n local-etc /etc
201 - Look how little extra space your second backup used (on top of the first):
205 - Get a list of your previous backups:
209 - Restore your first backup again:
211 bup restore -C ./dest-2 local-etc/2013-11-23-11195/etc
213 - Make a backup on a remote server (which must already have the 'bup' command
214 somewhere in the server's PATH (see /etc/profile, etc/environment,
215 ~/.profile, or ~/.bashrc), and be accessible via ssh.
216 Make sure to replace SERVERNAME with the actual hostname of your server):
218 ssh SERVERNAME bup init
220 bup save -r SERVERNAME: -n local-etc /etc
222 - Restore a backup from a remote server. (FAIL: unfortunately,
223 unlike "bup join", "bup restore" does not yet support remote
224 restores. See both "bup join" and "Things that are stupid" below.)
226 - Defend your backups from death rays (OK fine, more likely from the
227 occasional bad disk block). This writes parity information
228 (currently via par2) for all of the existing data so that bup may
229 be able to recover from some amount of repository corruption:
233 - Use split/join instead of index/save/restore. Try making a local
236 tar -cvf - /etc | bup split -n local-etc -vv
238 - Try restoring the tarball:
240 bup join local-etc | tar -tf -
242 - Look at how much disk space your backup took:
246 - Make another tar backup:
248 tar -cvf - /etc | bup split -n local-etc -vv
250 - Look at how little extra space your second backup used on top of
255 - Restore the first tar backup again (the ~1 is git notation for "one
256 older than the most recent"):
258 bup join local-etc~1 | tar -tf -
260 - Get a list of your previous split-based backups:
262 GIT_DIR=~/.bup git log local-etc
264 - Make a backup on a remote server:
266 tar -cvf - /etc | bup split -r SERVERNAME: -n local-etc -vv
268 - Try restoring the remote backup tarball:
270 bup join -r SERVERNAME: local-etc | tar -tf -
272 That's all there is to it!
278 - FreeBSD's default 'make' command doesn't like bup's Makefile. In order to
279 compile the code, run tests and install bup, you need to install GNU Make
280 from the port named 'gmake' and use its executable instead in the commands
281 seen above. (i.e. 'gmake test' runs bup's test suite)
283 - Python's development headers are automatically installed with the 'python'
284 port so there's no need to install them separately.
286 - To use the 'bup fuse' command, you need to install the fuse kernel module
287 from the 'fusefs-kmod' port in the 'sysutils' section and the libraries from
288 the port named 'py-fusefs' in the 'devel' section.
290 - The 'par2' command can be found in the port named 'par2cmdline'.
292 - In order to compile the documentation, you need pandoc which can be found in
293 the port named 'hs-pandoc' in the 'textproc' section.
296 Notes on NetBSD/pkgsrc
297 ----------------------
299 - See pkgsrc/sysutils/bup, which should be the most recent stable
300 release and includes man pages. It also has a reasonable set of
301 dependencies (git, par2, py-fuse-bindings).
303 - The "fuse-python" package referred to is hard to locate, and is a
304 separate tarball for the python language binding distributed by the
305 fuse project on sourceforge. It is available as
306 pkgsrc/filesystems/py-fuse-bindings and on NetBSD 5, "bup fuse"
309 - "bup fuse" presents every directory/file as inode 0. The directory
310 traversal code ("fts") in NetBSD's libc will interpret this as a
311 cycle and error out, so "ls -R" and "find" will not work.
313 - There is no support for ACLs. If/when some entrprising person
314 fixes this, adjust t/compare-trees.
320 - There is no support for ACLs. If/when some enterprising person
321 fixes this, adjust t/compare-trees.
323 - In t/test.sh, two tests have been disabled. These tests check to
324 see that repeated saves produce identical trees and that an
325 intervening index doesn't change the SHA1. Apparently Cygwin has
326 some unusual behaviors with respect to access times (that probably
327 warrant further investigation). Possibly related:
328 http://cygwin.com/ml/cygwin/2007-06/msg00436.html
334 - There is no support for ACLs. If/when some enterprising person
335 fixes this, adjust t/compare-trees.
344 bup stores its data in a git-formatted repository. Unfortunately, git
345 itself doesn't actually behave very well for bup's use case (huge numbers of
346 files, files with huge sizes, retaining file permissions/ownership are
347 important), so we mostly don't use git's *code* except for a few helper
348 programs. For example, bup has its own git packfile writer written in
351 Basically, 'bup split' reads the data on stdin (or from files specified on
352 the command line), breaks it into chunks using a rolling checksum (similar to
353 rsync), and saves those chunks into a new git packfile. There is one git
356 When deciding whether to write a particular chunk into the new packfile, bup
357 first checks all the other packfiles that exist to see if they already have that
358 chunk. If they do, the chunk is skipped.
360 git packs come in two parts: the pack itself (*.pack) and the index (*.idx).
361 The index is pretty small, and contains a list of all the objects in the
362 pack. Thus, when generating a remote backup, we don't have to have a copy
363 of the packfiles from the remote server: the local end just downloads a copy
364 of the server's *index* files, and compares objects against those when
365 generating the new pack, which it sends directly to the server.
367 The "-n" option to 'bup split' and 'bup save' is the name of the backup you
368 want to create, but it's actually implemented as a git branch. So you can
369 do cute things like checkout a particular branch using git, and receive a
370 bunch of chunk files corresponding to the file you split.
372 If you use '-b' or '-t' or '-c' instead of '-n', bup split will output a
373 list of blobs, a tree containing that list of blobs, or a commit containing
374 that tree, respectively, to stdout. You can use this to construct your own
375 scripts that do something with those values.
380 'bup index' walks through your filesystem and updates a file (whose name is,
381 by default, ~/.bup/bupindex) to contain the name, attributes, and an
382 optional git SHA1 (blob id) of each file and directory.
384 'bup save' basically just runs the equivalent of 'bup split' a whole bunch
385 of times, once per file in the index, and assembles a git tree
386 that contains all the resulting objects. Among other things, that makes
387 'git diff' much more useful (compared to splitting a tarball, which is
388 essentially a big binary blob). However, since bup splits large files into
389 smaller chunks, the resulting tree structure doesn't *exactly* correspond to
390 what git itself would have stored. Also, the tree format used by 'bup save'
391 will probably change in the future to support storing file ownership, more
392 complex file permissions, and so on.
394 If a file has previously been written by 'bup save', then its git blob/tree
395 id is stored in the index. This lets 'bup save' avoid reading that file to
396 produce future incremental backups, which means it can go *very* fast unless
397 a lot of files have changed.
400 Things that are stupid for now but which we'll fix later
401 ========================================================
403 Help with any of these problems, or others, is very welcome. Join the
404 mailing list (see below) if you'd like to help.
406 - 'bup restore' can't pull directly from a remote server.
408 So in one sense "save -r" is a dead-end right now. Obviously you
409 can use "ssh SERVER bup restore -C ./dest..." to create a tree you
410 can transfer elsewhere via rsync/tar/whatever, but that's *lame*.
412 Until we fix it, you may be able to mount the remote BUP_DIR via
413 sshfs and then restore "normally", though that hasn't been
416 - 'bup save' and 'bup restore' have immature metadata support.
418 On the plus side, they actually do have support now, but it's new,
419 and not remotely as well tested as tar/rsync/whatever's. However,
420 you have to start somewhere, and as of 0.25, we think it's ready
421 for more general use. Please let us know if you have any trouble.
423 Also, if any strip or graft-style options are specified to 'bup
424 save', then no metadata will be written for the root directory.
425 That's obviously less than ideal.
427 - bup is overly optimistic about mmap. Right now bup just assumes
428 that it can mmap as large a block as it likes, and that mmap will
429 never fail. Yeah, right... If nothing else, this has failed on
430 32-bit architectures (and 31-bit is even worse -- looking at you,
433 To fix this, we might just implement a FakeMmap[1] class that uses
434 normal file IO and handles all of the mmap methods[2] that bup
435 actually calls. Then we'd swap in one of those whenever mmap
438 This would also require implementing some of the methods needed to
439 support "[]" array access, probably at a minimum __getitem__,
440 __setitem__, and __setslice__ [3].
442 [1] http://comments.gmane.org/gmane.comp.sysutils.backup.bup/613
443 [2] http://docs.python.org/2/library/mmap.html
444 [3] http://docs.python.org/2/reference/datamodel.html#emulating-container-types
446 - 'bup index' is slower than it should be.
448 It's still rather fast: it can iterate through all the filenames on my
449 600,000 file filesystem in a few seconds. But it still needs to rewrite
450 the entire index file just to add a single filename, which is pretty
451 nasty; it should just leave the new files in a second "extra index" file
454 - bup could use inotify for *really* efficient incremental backups.
456 You could even have your system doing "continuous" backups: whenever a
457 file changes, we immediately send an image of it to the server. We could
458 give the continuous-backup process a really low CPU and I/O priority so
459 you wouldn't even know it was running.
461 - bup currently has no way to prune *old* backups.
463 Because of the way the packfile system works, backups become "entangled"
464 in weird ways and it's not actually possible to delete one pack
465 (corresponding approximately to one backup) without risking screwing up
468 git itself has lots of ways of optimizing this sort of thing, but its
469 methods aren't really applicable here; bup packfiles are just too huge.
470 We'll have to do it in a totally different way. There are lots of
471 options. For now: make sure you've got lots of disk space :)
473 Until we fix this, one possible workaround is to just start a new
474 BUP_DIR occasionally, i.e. bup-2013-10, bup-2013-11...
476 - bup has never been tested on anything but Linux, MacOS, and Windows+Cygwin.
478 There's nothing that makes it *inherently* non-portable, though, so
479 that's mostly a matter of someone putting in some effort. (For a
480 "native" Windows port, the most annoying thing is the absence of ssh in
481 a default Windows installation.)
483 - bup needs better documentation.
485 According to a recent article about bup in Linux Weekly News
486 (https://lwn.net/Articles/380983/), "it's a bit short on examples and
487 a user guide would be nice." Documentation is the sort of thing that
488 will never be great unless someone from outside contributes it (since
489 the developers can never remember which parts are hard to understand).
491 - bup is "relatively speedy" and has "pretty good" compression.
493 ...according to the same LWN article. Clearly neither of those is good
494 enough. We should have awe-inspiring speed and crazy-good compression.
495 Must work on that. Writing more parts in C might help with the speed.
499 Actually, that's not stupid, but you might consider it a limitation.
500 There are a bunch of Linux GUI backup programs; someday I expect someone
501 will adapt one of them to use bup.
507 bup has an extensive set of man pages. Try using 'bup help' to get
508 started, or use 'bup help SUBCOMMAND' for any bup subcommand (like split,
509 join, index, save, etc.) to get details on that command.
511 For further technical details, please see ./DESIGN.
517 bup is a work in progress and there are many ways it can still be improved.
518 If you'd like to contribute patches, ideas, or bug reports, please join the
521 You can find the mailing list archives here:
523 http://groups.google.com/group/bup-list
525 and you can subscribe by sending a message to:
527 bup-list+subscribe@googlegroups.com
529 Please see <a href="HACKING">./HACKING</a> for
530 additional information, i.e. how to submit patches (hint - no pull
531 requests), how we handle branches, etc.