2 bup: It backs things up
3 =======================
5 bup is a program that backs things up. It's short for "backup." Can you
6 believe that nobody else has named an open source program "bup" after all
9 Despite its unassuming name, bup is pretty cool. To give you an idea of
10 just how cool it is, I wrote you this poem:
13 What rhymes with awesome?
15 But that's irrelevant.
17 Hmm. Did that help? Maybe prose is more useful after all.
20 Reasons bup is awesome
21 ----------------------
23 bup has a few advantages over other backup software:
25 - It uses a rolling checksum algorithm (similar to rsync) to split large
26 files into chunks. The most useful result of this is you can backup huge
27 virtual machine (VM) disk images, databases, and XML files incrementally,
28 even though they're typically all in one huge file, and not use tons of
29 disk space for multiple versions.
31 - It uses the packfile format from git (the open source version control
32 system), so you can access the stored data even if you don't like bup's
35 - Unlike git, it writes packfiles *directly* (instead of having a separate
36 garbage collection / repacking stage) so it's fast even with gratuitously
39 - Data is "automagically" shared between incremental backups without having
40 to know which backup is based on which other one - even if the backups
41 are made from two different computers that don't even know about each
42 other. You just tell bup to back stuff up, and it saves only the minimum
43 amount of data needed.
45 - Even when a backup is incremental, you don't have to worry about
46 restoring the full backup, then each of the incrementals in turn; an
47 incremental backup *acts* as if it's a full backup, it just takes less
50 - It's written in python (with some C parts to make it faster) so it's easy
51 for you to extend and maintain.
54 Reasons you might want to avoid bup
55 -----------------------------------
57 - This is a very early version. Therefore it will most probably not work
58 for you, but we don't know why. It is also missing some
59 probably-critical features.
61 - It requires python >= 2.4, a C compiler, and an installed git version >=
64 - It currently only works on Linux, MacOS X >= 10.4, or Windows (with
65 Cygwin). Patches to support other platforms are welcome.
67 - It has almost no documentation. Not even a man page! This file is all
74 - check out the bup source code using git:
76 git clone git://github.com/apenwarr/bup
78 - install the python 2.5 development libraries. On Debian or Ubuntu, this
80 apt-get install python2.5-dev
82 - build the python module and symlinks:
90 (The tests should pass. If they don't pass for you, stop here and send
93 - Try making a local backup as a tar file:
95 tar -cvf - /etc | bup split -n local-etc -vv
97 - Try restoring your backup tarball:
99 bup join local-etc | tar -tf -
101 - Look at how much disk space your backup took:
105 - Make another backup (which should be mostly identical to the last one;
106 notice that you don't have to *specify* that this backup is incremental,
107 it just saves space automatically):
109 tar -cvf - /etc | bup split -n local-etc -vv
111 - Look how little extra space your second backup used on top of the first:
115 - Restore your old backup again (the ~1 is git notation for "one older than
118 bup join local-etc~1 | tar -tf -
120 - get a list of your previous backups:
122 GIT_DIR=~/.bup git log local-etc
124 - make a backup on a remote server (which must already have the 'bup' command
125 somewhere in the PATH, and be accessible via ssh; make sure to replace
126 SERVERNAME with the actual hostname of your server):
128 tar -cvf - /etc | bup split -r SERVERNAME: -n local-etc -vv
130 - try restoring the remote backup tarball:
132 bup join -r SERVERNAME: local-etc | tar -tf -
134 - try using the new (slightly experimental) 'bup index' and 'bup save'
135 style backups, which bypass 'tar' but have some missing features (see
136 "Things that are stupid" below):
139 bup save -n local-etc /etc
141 - do it again and see how fast an incremental backup can be:
144 bup save -n local-etc /etc
146 (You can also use the "-r SERVERNAME:" option to 'bup save', just like
147 with 'bup split' and 'bup join'. The index itself is always local,
148 so you don't need -r there.)
150 That's all there is to it!
158 bup stores its data in a git-formatted repository. Unfortunately, git
159 itself doesn't actually behave very well for bup's use case (huge numbers of
160 files, files with huge sizes, retaining file permissions/ownership are
161 important), so we mostly don't use git's *code* except for a few helper
162 programs. For example, bup has its own git packfile writer written in
165 Basically, 'bup split' reads the data on stdin (or from files specified on
166 the command line), breaks it into chunks using a rolling checksum (similar to
167 rsync), and saves those chunks into a new git packfile. There is one git
170 When deciding whether to write a particular chunk into the new packfile, bup
171 first checks all the other packfiles that exist to see if they already have that
172 chunk. If they do, the chunk is skipped.
174 git packs come in two parts: the pack itself (*.pack) and the index (*.idx).
175 The index is pretty small, and contains a list of all the objects in the
176 pack. Thus, when generating a remote backup, we don't have to have a copy
177 of the packfiles from the remote server: the local end just downloads a copy
178 of the server's *index* files, and compares objects against those when
179 generating the new pack, which it sends directly to the server.
181 The "-n" option to 'bup split' and 'bup save' is the name of the backup you
182 want to create, but it's actually implemented as a git branch. So you can
183 do cute things like checkout a particular branch using git, and receive a
184 bunch of chunk files corresponding to the file you split.
186 If you use '-b' or '-t' or '-c' instead of '-n', bup split will output a
187 list of blobs, a tree containing that list of blobs, or a commit containing
188 that tree, respectively, to stdout. You can use this to construct your own
189 scripts that do something with those values.
193 'bup index' walks through your filesystem and updates a file (whose name is,
194 by default, ~/.bup/bupindex) to contain the name, attributes, and an
195 optional git SHA1 (blob id) of each file and directory.
197 'bup save' basically just runs the equivalent of 'bup split' a whole bunch
198 of times, once per file in the index, and assembles a git tree
199 that contains all the resulting objects. Among other things, that makes
200 'git diff' much more useful (compared to splitting a tarball, which is
201 essentially a big binary blob). However, since bup splits large files into
202 smaller chunks, the resulting tree structure doesn't *exactly* correspond to
203 what git itself would have stored. Also, the tree format used by 'bup save'
204 will probably change in the future to support storing file ownership, more
205 complex file permissions, and so on.
207 If a file has previously been written by 'bup save', then its git blob/tree
208 id is stored in the index. This lets 'bup save' avoid reading that file to
209 produce future incremental backups, which means it can go *very* fast unless
210 a lot of files have changed.
213 Things that are stupid for now but which we'll fix later
214 --------------------------------------------------------
216 Help with any of these problems, or others, is very, very welcome. Join the
217 mailing list (see below) if you'd like to help.
219 - 'bup save' doesn't know about file metadata.
221 That means we aren't saving file attributes, mtimes, ownership, hard
222 links, MacOS resource forks, etc. Clearly this needs to be improved.
224 - There's no 'bup restore' yet.
226 'bup save' saves files in the standard git 'tree of blobs' format, so you
227 could then "restore" the files using something like 'git checkout'. But
228 that's a git command, not a bup command, so it's hard to explain and
229 doesn't support retrieving objects from a remote bup server without first
230 fetching and packing an entire (possibly huge) pack, which could be very
231 slow. Also, like 'bup save', you would need extra features in order to
232 properly restore file metadata. And files that bup has split into
233 chunks would need to be recombined somehow.
235 - 'bup index' is slower than it should be.
237 It's still rather fast: it can iterate through all the filenames on my
238 600,000 file filesystem in a few seconds. But sometimes you just want to
239 change a filename or two, so this is needlessly slow. There should be
240 a way to binary search through the file list rather than always going
241 through it sequentially. And if you only add a couple of filenames,
242 there's no need to rewrite the entire index; just leave the new files
243 in a second "extra index" file or something.
245 - bup could use inotify for *really* efficient incremental backups.
247 You could even have your system doing "continuous" backups: whenever a
248 file changes, we immediately send an image of it to the server. We could
249 give the continuous-backup process a really low CPU and I/O priority so
250 you wouldn't even know it was running.
252 - bup currently has no features that prune away *old* backups.
254 Because of the way the packfile system works, backups become "entangled"
255 in weird ways and it's not actually possible to delete one pack
256 (corresponding approximately to one backup) without risking screwing up
259 git itself has lots of ways of optimizing this sort of thing, but its
260 methods aren't really applicable here; bup packfiles are just too huge.
261 We'll have to do it in a totally different way. There are lots of
262 options. For now: make sure you've got lots of disk space :)
264 - bup has never been tested on anything but Linux, MacOS, and Linux+Cygwin.
266 There's nothing that makes it *inherently* non-portable, though, so
267 that's mostly a matter of someone putting in some effort. (For a
268 "native" Windows port, the most annoying thing is the absence of ssh in
269 a default Windows installation.)
271 - bup has no GUI. Actually, that's not stupid, but you might consider it
272 a limitation. There are a bunch of Linux GUI backup programs; someday
273 I expect someone will adapt one of them to use bup.
279 bup is a work in progress and there are many ways it can still be improved.
280 If you'd like to contribute patches, ideas, or bug reports, please join the
283 You can find the mailing list archives here:
285 http://groups.google.com/group/bup-list
287 and you can subscribe by sending a message to:
289 bup-list+subscribe@googlegroups.com